Probing The Sky

Curtis Peebles
With Contributions
by Richard P. Hallion
Library of Congress Cataloging-in-Publication Data
Peebles, Curtis.
Probing the sky : selected NACA/NASA research airplanes and their
contributions to flight / by Curtis Peebles.
pages cm
Includes bibliographical references and index.
1. Research aircraft--United States--History. 2. United
States.--National Aeronautics and Space Administration--Research. I. Title.
TL567.R47P4349 2015
629.133’3--dc23
2014039331
Copyright © 2014 by the National Aeronautics and Space Administration.

The opinions expressed in this volume are those of the authors and do not
necessarily reflect the official positions of the United States Government or
of the National Aeronautics and Space Administration.
This publication is available as a free download at

http://www.nasa.gov/ebooks.
National Aeronautics and Space Administration

Washington, DC
Table of Contents
Introduction: Toward—and Into—the Unknown v
Chapter 1: Confronting the “Sound Barrier”: The Bell XS-1................................ 1
Chapter 2: Flying Test Tube: The Douglas D-558-1 Skystreak........................... 45
Chapter 3: Proving the Swept Wing: The Douglas D-558-2 Skyrocket.............. 85
Chapter 4: U
nfulfilled Promise, Serendipitous Success:
The Douglas X-3 Stiletto................................................................. 135
Chapter 5: Versatile Minimalist: The Northrop X-4 Bantam............................. 167
Chapter 6: Transformative Pioneer: The Bell X-5............................................... 207
Chapter 7: Progenitor of the Delta: The Convair XF-92A................................... 243
Round One: A Reflection 269

Appendix: Technical Specifications for the Round One Aircraft 280
Acknowledgments 282
Selected Bibliography 283
Index 307
iii
The scale of what’s involved in undertaking research flights is made clear in this iconic image
of the D-558-2, the B-29 launch plane, and the ground support personnel and equipment. The
photo was taken in front of the NACA hangar at South Base shortly before the move to the pres-
ent facilities. (NASA Photo)
Introduction
Toward—and Into—the Unknown

In the decades since the Wright brothers’ first flights, a body of knowledge
and tools, created in an evolutionary process of small steps, had been built
up to guide engineers and researchers in developing new aircraft. The early
wood-and-fabric biplanes had given way to all-metal monoplanes. Aircraft
size, range, and payload had also grown, until the oceans could be spanned in
a fraction of the time a ship would take. Speed became the critical factor in
both commercial and military operations.
But by the early 1940s, speed itself had become the problem. With aircraft
flying near the speed of sound, the old rules of subsonic aerodynamics no longer
applied and the old tool of aeronautical research—the wind tunnel—no longer
worked. Consequently, engineers lacked the means to determine if their designs
would withstand actual flight conditions.
The technology, tools, and procedures needed in this new realm of flight
completely transformed the fields and practices of aerodynamics, propulsion,
structures, design theory, materials, flight control systems, life-support systems,
escape systems, safety procedures, wind tunnels, and data collection systems
and methodologies. The body of knowledge for the supersonic era was effec-
tively recast and made anew.
Thomas Kuhn’s term “paradigm shift” is now much overused both as an
expression and as an intellectual concept, but, in the case of the midcentury
high-speed revolution, it is certainly appropriate. What aeronautical science
had accrued for subsonic flight following the Wrights’ first flight in 1903
was rendered almost entirely irrelevant by the late 1940s and the onset of the
supersonic era.
The engine of transformation was the turbojet. It spelled the end of one era
in aviation and the dawn of another, pushing people and machines across the
sonic divide faster than they had gone before, forcing transonic and supersonic
researchers to address the mysteries that were affecting airplanes entering this
new and unknown realm.
Ironically, while it was this new form of propulsion that drove the super-
sonic breakthrough, the turbojet itself could not yet take the researchers and
pilots far enough into the realm to conduct the vital research data they sought.
v
Probing the Sky
The turbojet of the 1940s was a new and still immature technology. The
temperatures in its combustion chambers (burner cans) and which the spin-
ning turbine blades endured approached the limits of contemporary metal-
lurgical science, making them unsuited then for supersonic flight. In order to
explore the supersonic regime, researchers turned to rocket planes, relegating
turbojets to slower speeds. Some of these turbojets became famous, such as
the XS-1 (later designated simply the X-1, the first of the postwar “X-series”
research airplanes). Others are little remembered today—but each played its
own distinctive part in taking aviation from the subsonic to the supersonic era.
Collectively, these first research aircraft are known as “Round One,” the tran-
sonic and supersonic probers of the 1940s and 1950s that preceded “Round
Two,” the hypersonic North American X-15 of the 1960s.
This story is not simply one of strange-looking airplanes and their coura-
geous test pilots. It offers, as well, insights into the history of aerospace tech-
nology and science and of shifting technology and (yes) paradigms within that
field. The turbojet and the rocket changed the rules. Together, they led to the
X-1 and the other research airplanes. These airplanes were also why a small
group of engineers came to a vast dry lakebed in the Mojave Desert, establish-
ing there, arguably, the world’s premier center for aeronautical research and
development, the Air Force Test Center–NASA Armstrong Flight Research
Center complex (and the associated airspace) at Edwards Air Force Base, CA.
Creating the new era, with its own rules and tools, took place against the
dynamic background of the Cold War, forcing and feeding the need for ever-
higher performance. But at the onset of that era, lacking reliable tools to deter-
mine what would work and what would not, engineers had few means to decide
what paths they should follow. In a real sense, they faced the same situation
as the Wright brothers, when they were building their first kites and gliders in
1899–1902, with little reliable information, much of which was contradictory
and misleading. They discovered that little of the available information was reli-
able. They discovered that they had to reject treasured assumptions and awake
to new realities. They had to rethink their concepts of stability and control.
For the test pilots, there was yet something more: since the potential of failure
and death was very real, when they went aloft, their skill had to be matched
by even greater courage. Knowledge, dedication, expertise, and courage: of the
mix of such was the crafting of the transonic and supersonic revolution, made
manifest in the skies over the Mojave over a half century ago.
vi
A semispan airplane model on the wing of a P-51 in January 1946. In an adaptation of the
wing-flow technique, a half-scale model of an airplane was attached to the upper surface of
a wing. As the airplane made a dive, airflow over the wing accelerated to high transonic or
supersonic velocity. (NASA)
viii
CHAPTER 1
Confronting the “Sound Barrier”:
The Bell XS-1
The resistance of a wing shoots up like a barrier as we approach the speed of sound.
—W.F. Hilton1
In 1935, when British aerodynamicist W.F. Hilton inadvertently coined the

phrase “sound barrier,” the piston engine aircraft was nearing its performance
peak, producing approximately one horsepower per pound of engine weight.
But aeronautical engineers were already outrunning the knowledge they had
acquired in the three decades since the Wright brothers’ first flight at Kitty
Hawk, NC, in December 1903. Up to this time, designers had always treated
the airflow over wings as if it were an “inviscid” or incompressible fluid. This
simplified their calculations, and the errors induced by this convenient (if false)
assumption were too small to be of significance.
With aircraft now exceeding 400 miles per hour (mph), the compressibility
of air could no longer be ignored. But engineers and scientists lacked the insight
on how the required revolutionary changes could be accomplished. The wind
tunnel, the aerodynamicist’s standard tool since before the Wright brothers, was
of little help. As the airflow neared the speed of sound, shock waves formed on
the models and supports and reflected onto the tunnel walls, rendering the data
questionable from Mach 0.8, just below the speed of sound, to approximately
Mach 1.2, just beyond the speed of sound. While tunnels could hint at some
of the flow changes induced by high-speed flight—for example, a dramatic
loss of lift and simultaneous sharp increase in drag—they could not furnish
precise quantified data that would permit accurate analysis.
Thus, very suddenly, a new realm of flight loomed—mysterious, unknown,
dangerous, and destructive. Pilots making high-speed dives found that control
surfaces would not move or did so to little effect. In some cases, their airplanes
broke up or dove uncontrollably into the ground. Others, more fortunate,
found that the controls would respond normally once the airplanes reached
lower altitudes. The phrase “sound barrier” encapsulated the mystique of this
realm. Very quickly, a mythology developed regarding supersonic flight. Myths
unsupported by scientific fact gained traction, some of which bordered on the
1
Probing the Sky
bizarre: Some held that a pilot’s voice would become caught in the throat as the
plane “broke” the sound barrier, while others alleged that time would reverse
and the pilot would become younger after going supersonic.2
The Onset of Transonic Research

The first indications of these problems appeared in 1918. Aircraft of that time
had top speeds of about 100 mph. However, the tips of their propellers were
rotating at speeds as high as 300 to 650 mph. It was Frank W. Caldwell and
Elisha N. Fales, of the Army Air Service Engineering Division, who made the
first realistic measurements of propeller aerodynamics. Using a wind tunnel
able to reach speeds of 450 mph, they discovered that “[w]hen the stress reaches
a certain value, two adjoining layers of air begin to slide past each other and
the character of the flow is changed.”3
They had observed flow separation, which occurs as airflow passing over an
airfoil accelerates. As a consequence, the airfoil’s drag increases, reducing its
efficiency. Because a propeller’s airfoil and that of a wing are similarly affected,
the propeller tests gave the first indications of what would occur as aircraft
reached the same speeds.4
Lyman J. Biggs and Dr. Hugh L. Dryden of the Bureau of Standards
expanded upon this initial work over the next
decade. Between 1924 and 1931, Biggs and
Dryden undertook four separate studies. The
first study used the same six airfoils Caldwell
and Fales had tested, but at speeds of 375 mph
to 682 mph, confirming their earlier results.5
Biggs and Dryden’s second set of tests also
involved the six airfoils, but at speeds of 383
mph (Mach 0.5) to 830 mph (Mach 1.08). This
reconfirmed the earlier results, determined that
flow separation was the cause of the increased
drag, and expanded the speed range.6
Hugh L. Dryden undertook This was followed by a third, much more
some of the first research into expansive study of 24 propeller airfoil shapes,
aerodynamics at transonic speeds tested at speeds of Mach 0.5, 0.65, 0.8, 0.95,
during his research on propeller and 1.08. The study’s results showed a general
airfoils. The aerodynamic effects
rule: thin airfoils had lower drag at higher speeds,
the propeller airfoils experienced
gave indications as to what would making them more effective. In contrast, thick
occur as aircraft flew close to the airfoils showed greatly increased drag at higher
speed of sound. (NASA) speeds. Another general observation was that
2
Confronting the “Sound Barrier”: The Bell XS-1
when flow separation occurred, all the airfoils’ efficiencies were reduced “nearly
to the same level irrespective of their efficiency when the flow is smooth.”7 As
a result, “[t]he efficient sections therefore suffer most.”8 Because of the differ-
ent Mach numbers that were used for each test, Biggs and Dryden were able
to show that the rate of the increase in drag rose abruptly at speeds well below
Mach 1, while the flow separation occurred above Mach 0.8 in most cases.
Biggs and Dryden’s fourth and final research study was on the advantages
of “circular arc” airfoils. One of the airfoils used in the third study was of this
type. A circular-arc airfoil has a flat lower surface, while the upper surface is a
segment of a circle. Despite its being a thick airfoil, Biggs and Dryden found
that the circular-arc shape was more efficient than conventional airfoil shapes
at high speeds. They wanted to see if this applied as a general rule. The tests
involved eight circular-arc airfoils at Mach 0.5, 0.65, 0.8, 0.95, and 1.08. By
using the same speeds as the earlier tests, they could make a systematic com-
parison. The results were surprising—one circular-arc airfoil was “extremely
inefficient” at low speeds but had only 80 percent of the drag of a conventional
airfoil at speeds between Mach 0.85 and 1.08. Biggs and Dryden realized that
this result offered an elegant solution to the loss of propeller efficiency. They
wrote, “These results indicate that it would be beneficial to use circular-arc sec-
tions for the outer part of a propeller blade intended for use at high tip speeds,
retaining sections of the conventional type nearer the hub where the thickness
ratio is large and the speed low. It seems not unreasonable to expect that a
circular-arc section can be used profitably over the outer third of the blade.”9
Although these studies were conducted by the Army Air Corps and the
Bureau of Standards, their results were all published by the National Advisory
Committee for Aeronautics (NACA), which was established in 1915 “to super-
vise and direct the scientific study of the problems of flight with a view to their
practical solution.”10 The NACA was created in response to the advances in
aeronautical research and aircraft design being made in Europe. Over time,
the NACA increasingly focused its efforts on finding “practical solutions” that
could advance aeronautics, establishing a pattern of inquiry that constitutes
one of the foundational underpinnings of the more recent NASA, which was
created in 1958 to address the challenges of the emergent post-Sputnik space
age. The propeller-airfoil research represented milestones in the understand-
ing of compressibility effects at transonic and supersonic speeds. Although it
was not done to make supersonic flight possible, but rather for the seemingly
mundane goal of improving propeller efficiency, it nevertheless had profound
significance for the future assault on the “sound barrier.”11
This focus on practical applications and incremental improvements would
serve the NACA well in the 1920s and the Depression years of the 1930s.
It reflected the evolutionary nature of aviation technology in this period.
3
Probing the Sky
Wood-and-fabric biplanes were giving way to all-metal monoplanes, aircraft

ranges and speeds were steadily increasing, and passenger comfort and safety
were improving.12 The wooden Fokker trimotors of 1929 had, a decade later,
been replaced by all-metal DC-3s.
In all of these endeavors, the NACA played a significant role. Its work in airfoil
development transformed aircraft design practice and established a global stan-
dard. While agency researchers did miss—as many others in America and else-
where did as well—the potential significance of the gas turbine engine, the agency
did support studies on reaction propulsion systems, and its many investigations
of conventional aircraft propulsion and propulsion systems design resulted in
much more efficient engine installations on aircraft, typified by nacelle and
cowling location, and the nearly ubiquitous NACA cowling of Fred Weick. As
well, the NACA’s detailed studies on flight structures, bolstered by its persistent
analysis of foreign research, hastened the transformation of the airplane from
a wooden, braced biplane to a metal, cantilever, and monocoque monoplane.
The Turbojet Revolution

Serious interest within the NACA in reaction propulsion began at the same
time as the propeller-efficiency studies and likewise involved piston-engines,
not pure turbojets. In both America and abroad, early concepts for a “jet”
engine involved using a piston engine to power a reciprocating air compressor.
The resulting high-pressure air was then fed into a combustion chamber, where
fuel was added and ignited. The hot gas efflux exited via a convergent-divergent
nozzle to produce thrust.13
Edgar Buckingham of the Bureau of Standards undertook a study of jet
propulsion for the Army Air Service using this conceptual engine. (No actual
hardware or subscale engine existed.) His conclusions were published in 1923
by the NACA.14 Buckingham found that the hybrid piston engine/jet concept
had numerous flaws, which included high fuel consumption, a much heavier
weight than a piston engine, an increased number of moving parts, and the
resulting lower reliability and higher maintenance requirements. Buckingham
concluded, “There does not appear to be, at present, any prospects whatever
that jet propulsion of the sort here considered will ever be of practical value,
even for military purposes.”15 Buckingham’s conclusions were incorrect, both
for this hybrid concept and, more significantly, for the higher-performance
(and more technologically demanding) turbojet. It should be noted, however,
that the Bureau of Standards and the NACA were far from alone in such judg-
ments, which, generally, were accepted more broadly by the global engineering
and scientific community as a whole.
4
Subsequent NACA studies in the 1920s and early 1930s used the same
flawed, conceptual engine and reached the same flawed conclusions. In part,
this was due to the inefficient hybrid design, incorrect assumptions about
the large size and weight of the compressor, and the research culture at the
NACA. The Langley Memorial Aeronautical Laboratory, which was the only
NACA research facility until 1941, had only a small engine research division,
and its members were largely focused on the piston engine. Thus, it was three
European countries—Nazi Germany, Great Britain, and Fascist Italy—and
not the United States, that made the most significant jet-related technology
breakthroughs of this era. (It should be noted, however, that the same compla-
cency afflicting American aeropropulsion experts was likewise generally found
overseas. Only after inventors outside the aeropropulsion mainstream—Hans
von Ohain in Germany and Frank Whittle in Great Britain—demonstrated
the value of the gas turbine was it effectively “seized” by the aeropropulsion
community and subsequently made a normative element of aircraft design.)16
Neither von Ohain nor Whittle was aware of the other’s work until after the
outbreak of the Second World War. Hans von Ohain, a young physicist, began
his work on jet engines in 1933. Von Ohain’s engine was a more advanced axial-
flow turbojet engine design than Whittle’s concept, but, even so, he finished
first, thanks in large measure to the support of Ernst Heinkel, a leading aircraft
manufacturer in Germany. Von Ohain’s turbojet engine powered the He 178,
which made the world’s first flight by a pure jet aircraft on August 24, 1939.17
Frank Whittle, von Ohain’s equivalent in Great Britain, began working on a
turbojet engine design earlier, in 1928. He persevered in the face of official
skepticism, metallurgical challenges, and a lack of serious financial support,
and finally he successfully ground-tested a centrifugal-flow jet engine in 1937,
following this with a flying example four years later, the Gloster E28/39, which
flew on May 15, 1941, as the first British jet aircraft. At first glance, Whittle’s
engine resembled a Moss or Rateau-type turbosupercharger in design, but it
had very great differences, notably, of course, in its direct connection, via a
single drive shaft, of the “hot” turbine and the “cold” compressor.18
Despite growing interest in gas turbine propulsion, the NACA remained
focused on the original ducted fan design, assisted by downstream burning
with the jet efflux passing through a convergent-divergent nozzle. In early
1940, Albert E. Sherman undertook a second look at Buckingham’s report on
jet propulsion. Sherman examined a jet’s feasibility in terms of flight at 500
mph. His conclusion was that jet propulsion offered the possibility of high
speeds without the heavy machinery Buckingham had thought necessary 17
years previously. Sherman’s report was discussed at an April 11, 1940, meeting
in the office of Henry Reid, the head of the Langley Aeronautical Laboratory.
5
Probing the Sky
The senior staff in attendance accepted the report and concluded that Langley
should undertake jet research.
Dr. George W. Lewis, the NACA Director of Aeronautical Research, sent
a letter to Langley on April 22, 1940, approving the start of work on a com-
bustion test rig for what was now called the “Jeep” engine, but which was
still, at heart, simply a refined version of the conceptual design Buckingham
had examined in 1923. This was a piston-engine/combustion-chamber design,
rather than a true turbojet. It was built using sheet iron and a salvage engine,
but the low priority and demands of other work meant 2 years passed before
Jeep test runs would begin.
Europe was now engulfed by war, and, while the United States was still
officially neutral, a close watch was kept on Germany’s aircraft developments.19
General H.H. “Hap” Arnold, chief of the U.S. Army Air Corps (shortly to be
reorganized as the U.S. Army Air Forces [AAF]), sent a letter on February 25,
1941, to Dr. Vannevar Bush, the chairman of both the NACA and the National
Defense Research Committee. Arnold warned that the Germans had made
considerable progress on rocket propulsion, which threatened to make exist-
ing fighter aircraft obsolete. He continued, “Further investigation by a large
group of able scientists is immediately needed.”20 Bush formally established the
NACA Special Committee on Jet Propulsion, with retired professor William F.
Durand as chairman. Durand had been a member of the NACA since it was
organized and was still active at the age of 82. Bush told him in a March 18,
1941, letter about the Jeep engine and commented, “It seems to have great
possibilities and I cannot find any flaw in their arguments.”21
Eastman N. Jacobs, an engineer in Langley’s airflow research branch who
had been studying reaction propulsion and thrust-augmented systems since the
late 1920s, wanted to do more than ground-test the Jeep engine. He wanted
to use the Jeep to power a specially designed high-speed research aircraft. His
concept, which emerged from the drawing tables in July 1942, had a cylindrical
fuselage, a nose-inlet, a shoulder-mounted straight wing, and a V-tail, and it
sat on tricycle landing gear. The pilot’s cockpit was just behind the inlet. The
research plane used a wing with a 15-percent thickness-chord ratio (i.e., the
wing’s depth from top to bottom was 15 percent of the distance from the wing’s
leading edge to its trailing edge) and an NACA high-speed airfoil. The power
came from an 825-horsepower Pratt & Whitney R-1535 radial piston engine,
to compress the incoming air, and the downstream fuel-injection and burning
Jeep propulsion system. The drag due to compressibility effects above 550 mph
made the airplane’s top speed uncertain. However, the team felt confident that
the airplane could reach transonic speeds.
Ironically, the driving force behind the proposal was not propulsion
experimentation itself, but the loss of aircraft in high-speed dives. Lockheed
6
P-38 Lightnings, Republic P-47

Thunderbolts, and Bell P-39
Airacobras had all suffered fatal tail
failures. The Navy’s Curtiss SB2C
Helldiver had similar problems
(and some foreign aircraft did as
well). Tests at both Langley and
the new Ames Research Center
(the second NACA center, cre-
ated in 1941 to support the West
Coast aircraft industry) showed
that the P-47 and P-39 needed to
have their tail structures strength-
ened to withstand the higher-than-
predicted air loads at high speeds.
With the SB2C, compressibility at
high speeds was distorting the wing
surface, causing flow separation
and buffeting turbulence, which in
turn caused the horizontal stabiliz-
ers to flutter, exceeding their struc- This design, proposed by Eastman Jacobs
tural strength and breaking them in 1941, constituted the first NACA research
off. To cure the problem, the wing aircraft. The airplane was powered by a Jeep
engine, which combined a piston engine to
structure was stiffened, preventing compress the air and a combustion chamber
the distortions. Additionally, the to produce thrust. The Jeep engine had already
tail had to be redesigned to with- been made obsolete by British and German
stand the 13-g loads experienced turbojet engines. (NASA)
in dives.22
Despite the need, Jacobs’s proposed research aircraft and the Jeep engine
would never be built. In April 1941, Hap Arnold had visited Britain and been
informed of British work with the Whittle engine. He immediately arranged
for the importation of Whittle engine technology to the United States. Out
of this came the first American “black” aircraft program: to develop a highly
secret jet engine test bed. Bell Aircraft Corporation was awarded a contract to
develop the plane, and General Electric another contract to develop its engines.
The plane itself was assigned an abandoned propeller-driven fighter designation
(P-59) used for another program to mask its own development.
On October 1, 1942, at Muroc Army Air Field, California (now Edwards
Air Force Base), the Bell XP-59A jet made its first flight. The aircraft was pow-
ered by a pair of General Electric I-A Whittle turbojet engines. Durand was on
hand for the airplane’s first flight at Muroc and had known about the airplane
7
Probing the Sky
from its beginning. Due to the secrecy of the project, however, he could not
discuss it with the other committee members. With an advanced turbojet
already in production, the research aircraft and the Jeep engine project were
obsolete. The Jeep was formally canceled in March 1943, marking the end of
formal American interest in ducted fan approaches. From now on, America
would be firmly fixed on the turbojet—and the rocket.
During the summer of 1943, Jacobs went to England and visited the Royal
Aircraft Establishment, where he learned about the state of British turbojet
engine technology. When Jacobs returned to Langley, he drafted a letter strongly
protesting the military services’ refusal to fund a research aircraft. In the letter,
he stated that the United States was making the same mistake that the British
had, namely, applying a revolutionary engine to a conventional airframe. But it
was his final comment that carried the most weight. In his warning, Jacobs dem-
onstrated uncharacteristic perception for someone from an agency increasingly
being seen as having failed to meet its primary mission of advancing American
aviation technology. “The development of the jet power units themselves,” he
warned any and all, “had progressed beyond the development of suitable air-
planes to employ them.”23
Jacobs realized that the unknowns of compressibility were no longer the
single issue facing designers. The revolutionary turbojet engine had to be
matched with an equally revolutionary airframe design. He argued that a uni-
fied effort involving the military services, aircraft manufacturers, and the NACA
was needed to solve the multiple unknowns of high-speed flight. This effort
would be done “with a view to producing quickly extreme-performance aircraft
of several types to be developed around existing units and suitable to exploit to
the full the capabilities of those existing jet-propulsion power plants.”24
Two Roads Toward the Transonic

The Jacobs research aircraft was not the only such proposal being discussed.
John Stack, the head of the Compressibility Research Division, met with Lewis
in the spring of 1942 to discuss the development of such an aircraft. Lewis
did not like the idea of a research aircraft, but he did respect Stack’s ability,
self-assurance, ambitions, and skills as a researcher. Lewis did not approve
development work on a research airplane due to wartime pressures on the
NACA. Stack came away from the meeting with a sense that Lewis did not
object to a low-priority effort to identify the design features that a transonic
research aircraft should have.
As Jacobs had done, Stack also assembled a team to define the research air-
craft. The members included engineer Milton Davidson and junior engineering
8
aide Harold Turner, Jr. (Turner had

worked on Jacobs’s aircraft design.)
Stack’s design had a top speed of
Mach 0.9, as he was not interested
in actual supersonic flight. The lim-
ited performance of early turbojet
engines made a successful Mach
0.9 flight marginal in any event. By
early summer of 1944, the team had
completed a preliminary design of
a small turbojet-powered research
aircraft capable of Mach 0.8 to John Stack, the NACA researcher who master-
Mach 1 speeds.25 minded the design and construction of the D-558
Stack informed selected friends series. He was determined to get a jet-powered/
within the AAF and Navy aeronau- transonic research airplane and was opposed
tical communities of the study; the to the rocket-powered/air-launched/supersonic
XS-1. In this 1946 photo, a D-558-1 wind tunnel
only military officer who expressed model is in the background and Stack is holding
strong support for his research air- a swept-wing D-558 model. This eventually led
craft was Captain Walter Stuart to the D-558-2 Skyrocket. (NASA)
Diehl of the U.S. Navy, a noted
aeronautical engineer, author of an influential aerodynamics textbook, expert
on seaplane design, and senior naval representative to the NACA. During
late 1943, Diehl met several times with the chief of the Navy’s Bureau of
Aeronautics’ structural branch. He argued that a research airplane was the only
means to convince people that the “sound barrier” was “only a steep hill.”26
Support for a research airplane also came from an industrial source. On
December 18, 1943, Durand held a special meeting at the NACA Headquarters
on the subject of jet propulsion. At the meeting, he asked a basic question:
What should the United States do with the turbojet propulsion technology
being developed?
Among those in attendance was Robert W. Wolf of Bell Aircraft, who was
the designer of the XP-59A’s airframe, cockpit, and propulsion system. Wolf
had also traveled to Britain in 1943 to review its progress in turbojet technol-
ogy. He knew more powerful engines were under development that would
allow aircraft to reach transonic speed in level flight.
During the discussion, Wolf suggested that a high-speed, jet-powered
research aircraft be designed. These data were urgently needed, as new jets
would soon be facing the same structural and control problems in level flight
that had befallen the P-38, P-39, P-47, and SB2C during dive tests. After the
meeting concluded, Wolf put his ideas into a letter sent to Lewis on December
29, 1943. He wrote:
9
Probing the Sky
It appears quite possible to construct a single engine aircraft based

on available gas turbine jet power plants which will fly at speeds in
level flight exceeding the critical Mach numbers of currently used
types of wings. If this aircraft were designed with enough inherent
versatility to changes in control surfaces, wings, etc., it should be
possible to develop usable control surfaces such as ailerons, dive
control flap, tail surfaces, etc., which would work satisfactorily at
or above the critical speeds of the wings. Furthermore, this could
be done in level flight and would not be subject to the dangers
and difficulties associated with the high accelerations encountered
in current dive programs.27
Lewis’s response was more positive than that he had given Stack in 1942. He
wrote Wolf that the NACA was highly interested in learning more regarding
compressibility and aircraft stability and control at transonic speeds by means
of a turbojet-powered research aircraft. Lewis continued that the NACA was
giving this project “our very serious concern.”28
The AAF had some intelligence
reports on Germany’s rocket and jet-
powered aircraft, though these often
offered sketchy and in some cases contra-
dictory information. It was also clearly
interested in transonic flight research.
In mid-January 1944, the Development
Engineering Branch of the Materiel
Division at AAF Headquarters autho-
rized a study of a research airplane for
investigating aerodynamic phenomena
between 600 and 650 mph. At the same
time, Ezra Kotcher requested that the
Design Branch of the Aircraft Laboratory
at Wright Field conduct a comparative
A 1944 photo of John Stack when he
was the head of Langley’s Compressibility
study of two different research airplane
Research Division. With wind tunnels inca- concepts. The first would be powered
pable of providing reliable transonic-speed by a General Electric TG-180 axial flow
data, attention began to focus on building turbojet engine with a thrust of 4,000
a research aircraft to provide the needed pounds. The other aircraft design would
data. This resulted in a conflict between
use a proposed 6,000-pound-thrust
the NACA and the Army Air Forces, as
well as within the NACA over the aircraft’s Aerojet rocket engine.29
design—subsonic and jet powered versus While this study was underway, two
supersonic and rocket powered. (NASA) joint military-NACA conferences took
10
place at Langley on March 15 and 16, 1944, to deal with the issue of a
transonic research aircraft. One meeting was chaired by Navy Captain Diehl
and the other by Colonel Carl Greene of the Materiel Command. Greene
was Diehl’s AAF equivalent, functioning as the permanent AAF liaison officer
at Langley. While the NACA leadership had now accepted the need for a
specialized research aircraft, the military services—with numerous responsi-
bilities for waging a global war against ruthless foes—were still divided over
the need, design, and objectives: was, for example, the aircraft to be purely
experimental? Or might it have operational features or even be the basis of an
operational design?30
The AAF’s Design Branch completed its studies for the supersonic research
airplanes in April 1944. Kotcher’s design was dubbed the “Mach 0.999 study,”
a reference to the popular belief in an impenetrable sound barrier. A rocket-
powered aircraft was seen as offering the best approach to collect supersonic
flight data. The thrust of the rocket engine meant higher speeds could be
reached at higher altitudes, eliminating the need to make dive tests, as well as
the associated risks for both pilot and plane.
This question of using a turbojet versus a rocket motor for the research
airplane sharply divided research airplane proponents in the AAF and the
NACA. Kotcher had presented the preliminary results of the Mach 0.999 study
at a May 15–16 meeting with the NACA. Stack was critical of the design in
large part because of the rocket engine. NACA researchers believed this form
of propulsion would be unsafe. If the rocket failed during the takeoff from a
runway, they argued, the heavily loaded airplane would crash and explode.
A rocket-powered aircraft, Stack believed, could not meet the research
requirements. Its flight time would be brief compared to that of a turbojet-
powered aircraft (he did not envision the air-launching that would make them
highly productive). This meant a rocket-powered aircraft would not be able
to collect the volume and quality of data the researchers needed. While the
AAF vehicle was designed to fly above Mach 1, the NACA’s research airplane
was limited to flying at transonic speed. Finally, the NACA objected that the
performance of a rocket-powered vehicle would not be representative of an
operational aircraft’s. In contrast, the experience of flying a turbojet research
aircraft would be applicable. Kotcher countered that the XP-59A had shown
that existing turbojet engines lacked the thrust needed to fly at transonic speeds,
while the Mach 0.999 study showed that rockets could reach such speeds.
The AAF gave its reply in a final round of meetings at Langley on December
13–14, 1944. Not surprisingly, the AAF rejected the NACA jet-powered air-
craft proposal, as the subsonic concept was too conservative. The AAF wanted
the airplane to reach Mach 1.2 (approximately 800 mph, depending on the
altitude), and, as NACA researcher John V. Becker later observed, “The Army
11
Probing the Sky
was putting up the money and they decided to do it their way.”31 In late
December 1944, 1 week after the meeting, the AAF began negotiations with
Bell Aircraft Corporation to build the XS-1 (Experimental Sonic-1) rocket-
powered research aircraft, the result of a fortuitous chance meeting at Wright
Field between Kotcher and Bell chief engineer Robert J. Woods.32
Stack was committed to acquiring a turbojet-powered research aircraft, for if
it did not have the fullest performance of the rocket-powered one, it would nev-
ertheless have greater endurance and might well, in his view, be safer and more
reliable as well. Thus, after Kotcher (whom Stack highly respected) made it clear
that the AAF was determined to build a rocket-powered research aircraft based
on his Mach 0.999 study, Stack turned to the Navy’s Bureau of Aeronautics.
His message was dire: he told Navy personnel that the AAF’s rocket-powered
aircraft was unlikely to survive enough flights to provide significant data. His
efforts had the desired results.
Abraham Hyatt, a Marine Corps officer and aeronautical engineer, proposed
a research aircraft design in September 1944 that matched the NACA concept.
It would have a top speed of 650 mph at sea level, use a turbojet engine, have
thin 10-percent thickness-chord wings, take off and land under its own power,
have room for flight instrumentation, and have excellent low-speed handling.
These requirements were not surprising, as the Bureau of Aeronautics stated
that the Navy “was only interested in obtaining an airplane which met with
the full approval of the NACA.”33
In late December 1944, Diehl, Hyatt, and Captain William “Bill” Sweeney,
of the Bureau of Aeronautics, showed L. Eugene “Gene” Root, of the Douglas
Aircraft Corporation, the preliminary specifications for the jet-powered
research aircraft. Douglas was a major contractor for Navy aircraft, and Gene
Root took the Navy specification to Douglas’s El Segundo plant and met with
its chief engineer, Edward Henry “Ed” Heinemann. Heinemann formed a
small design team and, by February 1945, had completed a preliminary design
for the Douglas Model 558 High-Speed Test Aircraft.
This aircraft would be the first of a three-part effort. The D-558 phase 1
would have a top speed of about Mach 0.89. In the phase 2 effort, two of the
aircraft would be modified to have both a Westinghouse 24C turbojet engine
and a Reaction Motors Incorporated rocket. This would allow them to fly at
speeds between Mach 0.89 and about Mach 1 in level flight. The data from
the phase 1 and 2 aircraft would be used by Douglas to produce engineering
drawings and a mockup of a phase 3 operational Navy combat aircraft.
The Bureau of Aeronautics approved the new design of the D-558 phase 1 and
2 aircraft. To differentiate the two aircraft designs, the phase 1 aircraft were given
the Douglas model number D-558-1, while the phase 2 aircraft were called the
12
The scale of what is involved in undertaking research flights is made clear in this iconic image
of the D-558-2, the B-29 launch plane, and the ground support personnel and equipment. The
photo was taken in front of the NACA hangar at South Base shortly before the move to the pres-
ent facilities. (NASA)
D-558-2. The Bureau of Aeronautics also outlined a development program that

gave the NACA a major role in the management of the flight research program.
The NACA, the Navy, and Douglas would each have access to one of the
D-558-1 aircraft. The NACA research pilots would use their aircraft to col-
lect data on air loads, flutter, engine performance, and stability and control
at transonic speeds. Douglas test pilots would fly the company’s D-558-1 for
information useful for the D-558-2 operational aircraft. In early March 1945,
the NACA sent a letter to the Navy giving its observations: “The Committee
is certain that the procurement of these two models of high-speed research
airplanes will permit making of a large advance in aerodynamic knowledge
in the transonic region of flight and every attempt should be made to make
these aircraft available to the NACA for flight research as soon as possible.”
After several inspections of the D-558-1 mockup and revisions of the design to
13
Probing the Sky
correct what NACA representatives saw as shortcomings, the D-558-1 design

was approved.34
While Navy-NACA-Douglas relations were smooth, the same was not
completely true of the AAF–NACA–Bell XS-1 effort. The initial disagree-
ment was within the Langley staff and concerned the best wing to use on the
aircraft. Both groups knew that a thin wing was preferable to a thick wing for
high-speed flight. The reason was that thick wings have a lower “critical Mach
number,” the velocity at which a shock wave forms, leading to an immediate
large increase in drag, causing the so-called “shock-stall.”
Stack and fellow wind tunnel researchers wanted the XS-1 to have a wing
with a 12-percent thickness-chord-ratio wing. The group took this position
because the thicker 12-percent wing would encounter flow changes at speeds
of Mach 0.75 to 0.9. The aerodynamicists wanted data in this speed range to
correlate with their wind tunnel data. As a result, they recommended that Bell
fit this thicker wing to the XS-1 to deliberately cause it to fly within the risky
shock-stall flight regime.
The opposite viewpoint was expressed by Robert R. Gilruth, a researcher
at Langley, who argued for a very thin wing, with a thickness as low as 5 per-
cent. He believed the XS-1 and its pilot
would need every advantage they could
get to survive the unknowns of transonic
flight. If a thin wing was used, the air-
plane would be much safer to fly.
The thick/thin wing debate had a
larger importance. If the 12-percent-
thickness wing was selected, the XS-1
would only be able to reach Mach 0.9;
it could not go supersonic. In contrast,
Gilruth was arguing that to increase
safety, a very thin wing should be used to
lessen the inherent unknowns and dan-
gers of transonic flight. Ironically, this
Robert R. Gilruth invented the wing-flow safe approach also meant that the XS-1
technique to collect transonic data from could exceed Mach 1.
an airplane in a dive, and he supported Floyd Thompson reviewed the argu-
the rocket-powered XS-1. Stack effec- ments and evidence and decided that
tively designed the jet-powered D-558-1. Gilruth’s thin-wing concept was prefer-
Their activities following the establishment
of NASA reflected their outlooks. Gilruth
able. Stack responded by suggesting that
worked on piloted space flight activities, the XS-1 be fitted with two different sets
and Stack retired when the NACA ceased of wings. The first would be a 10-percent-
to exist. (NASA) thickness wing, less than what Stack had
14
wanted but more than what Gilruth recommended. The second wing would
have an 8-percent thickness-chord ratio. This was slightly thicker than Gilruth
had argued for, but much less than what Stack wanted. The logic was clear to
Thompson and other Langley managers, and, in March 1945, they decided to
have Bell build the two sets of wings.35
A far more serious set of objections arose as Bell Aircraft began construc-
tion of the XS-1. The original design had cylindrical internal fuel tanks for
the liquid oxygen and alcohol/water fuel positioned fore and aft of the wing
carry-through structure. A turbopump fed the oxidizer and fuel into the rocket
engine. Difficulties in building this turbopump led to the cylindrical tanks’
being replaced by two spherical tanks. High-pressure nitrogen gas would now
feed the fuel into the engine. The nitrogen gas was carried in multiple spherical
tanks. To withstand the pressure of the nitrogen gas, both the storage tanks and
the oxygen and alcohol/water fuel tanks had to be made of thick metal. The
result was that the XS-1’s landing weight was increased by a ton and the fuel
supply dropped due to the lesser volumetric efficiency of spherical tanks. As
a result, a ground takeoff would reduce the XS-1’s maximum speed to below
Mach 1.
Larry Bell had decided at an early stage that the XS-1 should be air-launched
from a four-engine mother ship such as a Douglas C-54 Skymaster transport
or Boeing B-29 Superfortress. Air-launching avoided the risk of the heavily
loaded XS-1’s suffering an engine failure on takeoff, undoubtedly resulting
in the loss of both plane and pilot. With the war still underway, however, no
aircraft were available. Langley researchers, in contrast, expected the airplane to
take off from a runway, fly the speed run, and make a normal runway landing.
Due to the lack of a mother ship, the launch issue was still open.
With the redesign of the fuel system, the only option to reach the AAF’s
planned Mach 1.2 top speed was an air launch from a mother ship. This would
carry the XS-1 to an altitude of 30,000 feet, at which point the XS-1 would be
released. Although an air-launch profile had earlier been considered as a safety
measure, it was now mandatory to meet the AAF’s speed goals. An additional
problem was that the smaller fuel supply meant that the rocket’s burn time had
dropped from 5.4 minutes to around 2.6 minutes. These changes reopened the
disagreements between the NACA and the AAF.36
Upon hearing of the changes, Stack sent a memo to Langley’s chief of
research. The XS-1 “may prove to be unsuitable.”37 Stack listed five basic
requirements the research airplane had to meet before he would be satisfied:
• Speed greatly in excess of the critical Mach number.
• Duration at full power for complete observations in level flight at
steady conditions.
• Takeoff, flight, and landing with self-contained power units.
15
Probing the Sky
• Flexibility to permit changes of all principal components such as

wings, tail surfaces, and canopies.
• Space for adequate instrumentation.38
Stack then noted that the XS-1 fell short of these basic requirements. In
particular, the aircraft lacked the needed duration at full power as well as the
ability to take off, fly, and land with self-contained engines. It also probably
could not now meet the requirement for adequate instrumentation.
Because the XS-1 did not meet the required duration at full power, Stack
continued, it also now failed to meet the requirement for speed in excess of
the critical Mach number. The fuel supply was so limited that the XS-1 would
burn its entire fuel supply to reach 35,000 feet and then be unable to make
the level speed run. While speed runs could be made at lower altitudes, their
research value was minimal, as it was only at altitudes of around 35,000 feet
that speeds in excess of the critical Mach number could be reached. In closing,
he urged that a major effort be made to develop the turbopump needed for the
original fuel system design.
Another factor that disturbed engineers at Langley, perhaps Stack most of
all, was control of the program. The AAF began looking for possible test sites
other than Langley. Langley was a busy airport located in a heavily populated
area, and any emergency ran the risk of killing and injuring people on the
ground. A possible move meant the Langley researchers would not have the
direct control over the flight research program that they wanted and were
accustomed to.39
The initial test sites considered for the XS-1 other than Langley were Muroc
Army Air Field in California and Wendover Field in Utah. Both were in iso-
lated areas ideally suited for safety and security. The move to a remote site
prompted objections from both NACA Headquarters and Langley personnel.
Stack dashed off an angry memo to John Crowley, chief of research at Langley,
on June 14, 1945, which said, “This airplane originated here as did the P-80
program. If we are to do research of this kind we must have the airplane here. I
do not believe we should again be treated as a service as was in the case with the
P-80. If the shifting of this aircraft to a western station materializes I propose
that we transfer all work beginning right now so we can free our people to do
research with our present equipment.”40
By late summer of 1945, the issue of the test site was still unresolved. The
candidate sites being checked by Bell test pilot Jack Woolams included Muroc
Army Air Field; Wendover Field, UT; Salina Field, KS; Daytona Beach, FL;
Marietta Field, OH; and several airfields in Texas. Stan Smith of Bell thought
Muroc, deep in the Mojave desert, 100 miles from Los Angeles and over 2,100
miles from Buffalo (Bell Aircraft headquarters), was too far away and recom-
mended Daytona Beach.41
16
Why Not Swept Wings?

By this time, a much larger issue had arisen that damaged the NACA’s repu-
tation even more than its earlier neglect of jet-engine technology or any dis-
agreements with the AAF over rocket planes and tests sites. Ironically, it also
represented a major breakthrough in the design of supersonic aircraft. It was
the result of theoretical calculations by Robert T. Jones at Langley.
Jones’s accomplishment began with a study of a dart-shaped missile in
late 1944. The missile’s slim delta shape required a new approach to calculate
its lift. Jones saw that he could apply the mathematical studies of the airflow
around airships that Max Munk had made in 1924 to analyze the flow around
the missile, buttressed by studies by Hsue-Shen Tsien, a von Kármán associate
from the Guggenheim Aeronautical Laboratory at the California Institute of
Technology, for supersonic flow around projectiles and slender bodies.
Jones began studying the mathematics of “potential flows” at supersonic
speeds in early 1945. This term refers to fluid motion where there is no
rotation of the fluid element. Jones realized he was deriving the same equa-
tions as he had using his earlier lifting theory. These new equations now also
included compressible flow. Jones recalled that Tsien had said that some
slender projectiles exhibited no effects of compressibility when rotated. Jones
took his earlier calculations out of the desk drawer and incorporated the
compressible flow equations. Jones found that very slender wings lacked
compressibility effects.
Trying to find a physical explanation, Jones undertook a series of elaborate
calculations and eventually found that this lack of compressibility effects was
the result of sweepback on the lift generated by large-span wings. Again, it was
a paper by Munk that gave Jones an understanding of the physical process.
Munk’s paper had dealt with the effects on aircraft stability due to wing dihe-
dral and sweepback at low subsonic speeds. Munk noted that for an airplane
in level flight, only the component of velocity at a “normal angle” (i.e., at a
perpendicular angle) to the wing’s leading edge affected the production of lift.
Although Munk’s paper only applied to subsonic aerodynamics, Jones believed
it could apply to slender wings in supersonic flows.
When Jones finished his calculations, he had a unified theory that encom-
passed swept wings with angles ranging from 0 degrees to 90 degrees and
covered all possible wing configurations rather than solely slender wings.
By late April of 1945, Jones had completed a report on his calculations.
Jones noted, “The analysis indicates that for aerodynamic efficiency, wings
designed for supersonic speed should be swept back at an angle greater than
the Mach angle and the angle of sweepback should be such that the compo-
nent of velocity normal to the leading edge is less than the critical speed of
17
Probing the Sky
the airfoil sections. This principle may also be applied to wings designed for
subsonic speeds near the speed of sound, for which the induced velocities
resulting from the thickness might otherwise be sufficiently great to cause
shock waves.”42
The physical explanation of these results was complex, but the results were
astonishing. With the Mach line ahead of the wing, the “streamlines,” a smooth,
nonturbulent flow over the upper surface of a wing following a set pattern,
would curve and follow paths consistent with a flow at subsonic speeds, even
though the velocity was actually supersonic.
This meant the effective Mach number was reduced compared to the actual
Mach number of the airplane. Jones had suspected that this would occur, but
he was surprised by the size of the reductions his calculations indicated, which
showed that the effective Mach number would be three to five times smaller
than that on a straight wing. This limited the amount of wave drag, which was
due to changes in velocity through a shock wave. Also reduced were the effects
of compressibility shock, which involved significant changes in the patterns of
airflow pressures, densities, and temperatures.
The advantages of swept wings were not limited to supersonic flight, Jones
found. He noted the sudden increase in drag when the airflow “may be avoided
by increasing the angle of sweepback so that the normal component of the
velocity not only is subsonic but is also less than the critical speed of the airfoil
sections. This principle may also be applied to wings designed for subsonic
speeds near the speed of sound.”43
Jones wrote a memo to Crowley telling him he had “made a theoretical
analysis which indicated that a V-shaped wing would be less affected by com-
pression than other planforms. In fact, if the angle of the V was kept small
relative to the Mach angle, the lift and center of pressure remain the same at
speeds both above and below the speed of sound.”44
The next step was to test the theory in actual flight. Jones concluded the
memo by asking Crowley to approve wing-flow and drop-body testing of
different wing shapes “designed to minimize compressibility effects.”45 These
alternative techniques had been developed in response to the inaccurate data
from wind tunnels at transonic speeds. Gilruth had developed the wing-flow
technique. A model connected a balance mechanism located in the starboard
gun bay of a P-51 Mustang protruded above the upper surface of the wing, and
an NACA pilot then dove the plane to over Mach 0.80. While the aircraft itself
was still subsonic, the airflow over the model was supersonic, to as high as about
Mach 1.4. While this produced useful results, the short duration of the test
methodology—and the obvious dangers attending diving into the dense lower
atmosphere at compressibility-inducing Mach numbers—limited its appeal.
18
A semispan airplane model on the wing of a P-51 in January 1946. In an adaptation of the
wing-flow technique, a half-scale model of an airplane was attached to the upper surface of a
wing. As the airplane made a dive, airflow over the wing accelerated to high transonic or super-
sonic velocity. (NASA)
A second approach was drop-body tests. These were streamlined shapes with
either rectangular or swept wings, dropped from 40,000 feet by a B-29. During
their fall, the shapes reached Mach 0.9 to Mach 1.27. The forces acting on the
wings were measured using onboard instruments, and the measurements were
transmitted to ground stations. Their trajectories were followed by radar and
optical tracking. Crowley gave his approval, and Gilruth’s flight research section
began testing to verify or refute the advantages of swept wings.
While these tests were still underway, Jones’s draft paper met with strong
opposition from Dr. Theodore Theodorsen, head of the Physical Research
Division, a leading theoretical physicist, and chairman of the Langley publica-
tion committee. He had had a contentious dispute with Eastman Jacobs in the
1930s over the use of wind tunnels versus theoretical calculations. Theodorsen
was similarly critical of Jones’s results.
Theodorsen did not agree with Jones’s arguments or his conclusions, call-
ing them “hocus-pocus” and demanding that they be clarified with some “real
mathematics.” Worse, he dismissed Jones’s results as “a snare and a delusion.” At
the end of the publication committee meeting, Theodorsen demanded that the
material on Jones’s swept-wing theory be removed from the paper. Considering
the weight he carried because of his stature and term of service, Langley man-
agement agreed with Theodorsen’s objections and decided publishing the paper
without experimental proof could not be justified.46
19
Probing the Sky
The debate was settled by the end of

May 1945, and Jones’s reputation was
redeemed with the completion of the
wing flow and drop-body tests. They
showed that Jones’s predictions were
valid. The data showed a reduction of
wing drag by a factor of almost four
using a swept-wing configuration. The
results were also confirmed by Macon
C. Ellis and Clinton Brown, using a
section of wire at a large angle of sweep A 35° swept-wing model mounted on the wing
in Langley’s supersonic wind tunnel.47 of a P-51 Mustang. The “wing flow” technique
Langley’s engineer-in-charge, provided data on airflow at transonic speeds
that were otherwise unobtainable through wind
Henry Reid, sent Jones’s completed tunnel tests. (NASA)
paper to NACA Headquarters in early
June 1945. Reid noted that Theodorsen still did not agree with Jones’s argu-
ments and conclusions and had refused to participate in the paper’s editing. The
paper, titled “Wing Plan Forms For High-Speed Flight,” was issued on June
21, 1945, as a Confidential Memorandum Report. A second edition was issued
3 weeks later, as an Advanced Confidential Report, with a wider distribution
to both the military services and specific personnel in the aviation industry.48
At this same time, Allied intelligence teams were searching for documents
and other data on advanced technology developed by Germany. They soon
discovered that extensive work on high-speed airfoils and sweptback wings
had been done. In contrast, the NACA had done virtually no research in these
fields during the 1930s and 1940s. The failure to anticipate the advantages of
swept wings was seen by the military and aircraft contractors as still another
NACA failure to keep pace with European aviation advances, and it caused
an angry response.
Brigadier General Alden R. Crawford, chief of the AAF’s Production
Division, asked NACA Chairman Jerome Hunsaker in October 1945 why
no mention of Jones’s swept-wing theory had been made during earlier XS-1
design reviews. Had the AAF known about this, General Crawford claimed, the
XS-1 design could have been changed to incorporate swept wings. The general
was incorrect—the AAF had been told about Jones’s paper before the first XS-1
mockup review. The AAF had specified straight wings for the XS-1 because its
planned jets used straight wings. For its part, the NACA had rejected using a
swept planform because, in the minds of Langley researchers (as they informed
the AAF), logically, it made little sense to further burden the research airplane
with yet another unproven design concept. Even so, the NACA’s reputation
20
was again damaged, and, for the record, Jones himself had hoped that the
XS-1 might incorporate a swept-wing planform.49
Flight Tests and Unresolved Issues

As 1945 drew to a close, the XS-1 project neared the flight-test phase. Colonel
George F. Smith, chief of the Engineering Division’s Experimental Aircraft
Project Division at Wright Field, who oversaw the XS-1 project, sent a letter
to Lewis on November 23 to outline the AAF’s position on several issues. The
initial Bell glide flights would be made at Pinecastle Army Air Field in Florida.
This base had a 10,000-foot runway; was being equipped with an SCR-584
radar and optical tracking equipment; had predictable weather; and was nearer
to the Bell, AAF, and NACA facilities than Muroc. Pinecastle’s remote location
also provided security. Colonel Smith’s letter also made clear that the inability
to develop a turbopump would not be allowed to delay the project and that
the XS-1 would be flown using the high-pressure fuel system. Not surprisingly,
Stack had very different ideas about the NACA’s role and was not willing to
change them.50
Stack wrote a three-page memo on December 28, 1945, to research direc-
tor Crowley regarding the NACA’s views on XS-1 flight research program
management. Stack wrote that the Pinecastle flights were “to determine the
feasibility of operating from Langley Field,” a notion he and other NACA
engineers still harbored. Stack
added that after the Bell contractor
flights were completed, Langley
would take over pilot responsi-
bilities for the tests. His earlier
objections to the XS-1 design still
remained. He wrote that the NACA
should not take possession of the
XS-1 project until the aircraft was
flown with the turbopump system
and fuel for a powered landing
and was capable of a ground take-
off. Finally, he emphasized in the Senior officials with the XS-1 research program
memo to Crowley that the NACA included, from left, Joseph Vensel (head of
had to insist that these conditions operations), Gerald Truszynski (head of instru-
mentation), Capt. Charles Yeager (XS-1 pilot),
be met.51 Walter C. Williams (head of NACA Muroc Flight
On January 8, 1946, NACA Test Unit), Maj. Jackie Ridley (XS-1 project engi-
Headquarters sent a reply to Colonel neer), and De E. Beeler (chief engineer). (NASA)
21
Probing the Sky
Smith’s letter. Although the NACA letter supposedly agreed with the AAF posi-
tion on the XS-1 program, in fact, it was “accepting” Stack’s rejection of the
Army position. The NACA letter stated the following:
• The final Army acceptance would occur after the rocket motor was
demonstrated and without the need for the B-29 launch aircraft.
• The Pinecastle flights were to determine if Langley Field operations
were feasible.
• Bell Aircraft would supply the pilot until the NACA decided to take
over the program.52
By the time the NACA’s letter was sent, the Bell glide flights were under-
way at Pinecastle Field. Bell test pilot Jack Woolams made 2 captive flights
and 10 free flights between January 25 and March 6, 1946, flying the first
Bell XS-1 (AAF serial 46-062), equipped with a 10-percent thickness-chord
wing and 8-percent thickness-chord tail (changed subsequently to an 8-percent
thickness-chord wing and 6-percent thickness-chord tail before it began its
powered flights). The Pinecastle glide flights provided the answers to several
issues. The XS-1 had good low-speed flight characteristics but poor visibility
through the flush windshield, and the air-launched concept was practical.
More important, the Pinecastle flights eliminated the use of either Langley
Field or Niagara Airport for subsequent XS-1 powered test flights. A much
wider landing area was required for safe operations. Of the 10 glide flights,
the first landed short of the runway. On the fourth flight, the left main gear
retracted after touchdown and the XS-1 went off the runway and suffered
minor wing and fuselage damage. On the fifth landing, the nose gear retracted,
causing minor damage. Even a 10,000-foot runway lacked the qualities needed
for a consistently acceptable unpowered touchdown. Additionally, bad weather
caused numerous canceled flights. The XS-1 required a site with better and
more predicable weather conditions.53
Despite the hopes of both Langley’s Stack and Bell Aircraft’s Smith, the only
site that met the requirement was Rogers Dry Lake at Muroc Army Air Field
in California. With a surface area of 44 square miles, Rogers was the largest
dry lake in the world. Its surface was dried silt, able to support 250 pounds
per square inch. The lakebed’s north-south distance was 11 miles. It enjoyed
clear skies some 350 days a year. The lakebed was located in the middle of the
Mojave Desert, far from prying eyes, yet was less than a hundred miles from
Los Angeles and its aircraft contractors. The XP-59A had first flown there, and
the advantages of the site for flight testing were clear to project personnel.54
The next step was the powered flights, using the Reaction Motors, Inc., four-
chamber XLR-11 6,000-pound-thrust rocket engine.55 As part of its contract
with the AAF, Bell Aircraft had to demonstrate that the XS-1 was control-
lable to a speed of Mach 0.8 and could withstand an 8-g pullout. Although
22
this concluded Bell’s obligations under the contract, company management

assumed they would also make the first Mach 1 flight. The longstanding prac-
tice was that contractor test pilots made the envelope expansion flights, up to
the airplane’s maximum speed.
The powered XS-1 flights would not be made by Woolams, who had been
killed while practicing for the 1946 Thompson Trophy air race. His replacement
was Chalmers H. “Slick” Goodlin. He was only 23 years old when assigned
to be the XS-1 test pilot, but he had amassed extensive flight experience. He
had soloed before turning 17; he had served in the Royal Canadian Air Force
as a flight instructor, in the Royal Air Force as a Spitfire pilot, and in the U.S.
Navy as a test and ferry pilot—all before joining Bell in December 1943 as a
production test pilot.56
The NACA was also preparing for the Muroc tests. In September, Hartley
Soulé, chief of the Stability Research Branch at Langley, selected Walter C.
Williams to oversee a small team, including engineers, instrument technicians,
and technical personnel, who were to support the Bell XS-1 powered flights.
They were called the NACA Muroc Flight Test Unit. Williams and four other
NACA engineers arrived at Muroc in September 1946. Six more arrived in
Group photo of the groundbreaking for the new main building at the High-Speed Flight Research
Station. From left to right are Gerald Truszynski (head of instrumentation), Joseph Vensel (head
of operations), Walter C. Williams (director of the High-Speed Flight Research Station), Marion
Kent (administration officer), and California state official Arthur Samet. (NASA)
23
Probing the Sky
October, and the two women “computers,” who reduced the data from the
onboard instrumentation into charts and graphs, arrived in December. The 13
personnel were initially considered to be a unit on a temporary assignment to
the desert base. The NACA Muroc Flight Test Unit gained permanent status
on September 7, 1947. It became the High-Speed Flight Research Station on
November 14, 1949, and was later renamed simply the NACA High-Speed
Flight Station.
They found themselves in a place very different from Virginia. Facilities
at Muroc were meager, with the initial NACA facilities consisting of a single
hangar shared with the AAF, plus two rooms. Darkroom facilities were lack-
ing. Support facilities were the loading pit for the XS-1, the liquid-oxygen and
liquid-nitrogen tanks, and a fuel trailer to mix the alcohol/water fuel.
Housing was also marginal. Williams found an apartment in Palmdale, 40
miles from Muroc. He was the exception. Single engineers and technicians
lived in “Kerosene Flats,” the decaying wartime housing in the town of Muroc
near the edge of the lakebed. In late 1946, the Marine air station in Mojave
was closed, and married NACA personnel moved into the base housing. In
both cases, the housing was low-quality wartime construction, cold at night,
and in dilapidated condition.
The biggest change from Langley was the landscape. Muroc was in the
wide-open spaces of the southwestern desert, with hot, dry weather during the
summer, while winter saw freezing temperatures and occasional snow. It was so
bleak that some Langley engineers and their families outright refused to leave
the familiar shores of Tidewater, VA. Of those who came west, some adapted
and thrived, while others soon returned to the Chesapeake.57
The second XS-1 (AAF 46-063) arrived at Muroc on October 7, 1946, under
the belly of the B-29 launch aircraft. The aircraft had the thicker 10-percent
thickness-chord wings and 8-percent thickness-chord horizontal tail flown
on the first airplane at Pinecastle 9 months earlier. Goodlin initially made a
series of glide flights to become familiar with the XS-1’s handling and flight
characteristics. The first glide-flight attempt was made 2 days later but was
aborted before launch. It was rescheduled and successfully made on October
11. A total of four glide flights were made by December 2, 1946. The way
was now clear for the initial powered flights. The first attempt, on December
9, was aborted before launch. The next try, on December 20, was successful.
Goodlin made a total of 12 powered flights in the second XS-1 over the next
3 months, with a final flight on February 21, 1947. Two more attempts were
aborted before launch.
24
Settling Differences, Building a Program

While Bell’s powered flight tests were going smoothly, the internal differences
continued. The Bell and NACA personnel had different concepts of the role
of the contractor XS-1 glide and powered flights. For Bell, these were to prove
the basic airworthiness of the XS-1, within the contract specifications. This
was to be a quick program, which Richard Frost of Bell Aircraft noted did not
“envision any lengthy series of scientific tests to investigate all the byroads of
stability in its various forms.”58
In contrast, Williams told Frost and Robert Stanley (also with Bell) during
a September 1946 meeting at Muroc that the NACA wanted complete stability
and control data from the upcoming powered flights, as well as aerodynamic
and structural load data. The data needed, he continued, included longitudinal
stability characteristics, control in both steady and accelerated flight condi-
tions, and the buffet boundary.59 Frost responded that no special flights would
be made to collect such data. Bell would accept its test pilot’s assessment of
stability and control issues to meet the contract requirements. There would
be no delays in the flight to accommodate the NACA’s research requirements.
On April 11, 1946, Colonel Smith wrote to Lewis to outline the AAF’s
position regarding its understanding and plans for the XS-1 project. Once the
Bell test flights were completed, the NACA would undertake the high-speed
flights and supply a pilot and data collection system and personnel. If the
NACA concluded the aircraft was unsafe, the XS-1 could be returned to AAF
control. The letter also stated that the AAF position was for the air launch to
continue, as this was the most practical and safest method of flight testing for
the XS-1. Colonel Smith made it clear that the AAF wanted a “firm under-
standing” between the AAF and the NACA on these issues before the end of
the Bell powered flights.
Williams wrote a memo to Melvin N. Gough, chief of the Flight Research
Division at Langley, in response to the AAF letter. Williams stated that the
NACA should not accept the XS-1 until the turbopump had been successfully
tested. He also rejected the use of the B-29 launch aircraft. Williams also noted
that Bell was planning to make the initial powered flight at Muroc, having
rejected operating from Langley.
Gough had his own requirements, which included the capability to take
off and climb to 35,000 feet under its own power. While he suggested that
the NACA might waive this requirement once the B-29 operations gained
experience, Gough insisted that the NACA not accept the XS-1 without
the turbopump and reserve fuel capacity for emergency landings. Langley’s
engineer-in-charge, Henry Reid, in a letter to NACA Headquarters written
on April 29, 1946, noted that unless these requirements were met, the NACA
25
Probing the Sky
“will not undertake to supply a pilot and

operate the airplane.” Reid asked NACA
Headquarters to press the AAF to begin
the construction of XS-1 fueling and
handling facilities at Langley. Doing so
would allow the transonic research flights
to begin soon after the acceptance testing
was completed. Despite the results of the
Pinecastle glide flights, Langley person-
nel still intended to fly their XS-1 from
their facility.
As for the XS-1 test plan itself,
Williams outlined a two-phase approach
on June 7, 1946, to Gough. The first phase
would determine stability and control at
high Mach numbers. Flight would be
made in increasing Mach number incre- Melvin N. Gough, chief test pilot at Langley,
ments, with eight flights made for each was among the group of staff members
who wanted the XS-1 to take off and land
speed increment. A total of 48 “successful on the Langley runway and who opposed
flights” would be made in the first phase the air-launch technique, supporting devel-
to reach the XS-1’s “operational limits.” opment of the D-558-1. (NASA)
If a change in wing thickness was needed,
the entire sequence would have to be repeated, requiring another 48 flights. A
“successful flight” was defined as one in which all systems functioned correctly,
the pilot flew the mission exactly as planned, and all the instrumentation oper-
ated properly. The odds of this were about 50/50, indicating that as many as
100 flights might actually be needed.
The second phase would be to measure aerodynamic loads on the wings
and tail and collect additional drag and performance data along with stabil-
ity and control measurements. This phase would be much shorter, with only
16 flights being planned for each wing thickness. Reid forwarded the plan to
NACA Headquarters on June 24, 1946. What was not in the plan was just
as important. The XS-1 would be flown to its “operational limits.” Williams
never mentioned making supersonic flights. His plan was also a long, deliber-
ate, step-by-step effort, requiring a year or more to complete.60
The differing viewpoints between Bell, the NACA, and the AAF were such
that insignificant matters gave rise to anger. As 1946 ended and the first pow-
ered flights were made, both the AAF and Bell began to publicize the XS-1
project (even though the XS-1 was a classified project and was being flown at
a restricted base). Stories were carried in the New York Times and later in Time
magazine, which focused on Bell, the AAF, and Goodlin.
26
On December 12, 1946, the day after the New York Times article was pub-
lished, Williams sent a letter to Soulé at Langley. Williams complained that
the Bell-AAF publicity was about an airplane that, he claimed, did not really
exist. They “should admit that the airplane was not designed as a supersonic
aircraft but rather a high transonic airplane.”61 He continued that the super-
sonic capability was only because they were “forced” to use the B-29 launch
plane to achieve the altitude and flight time. The issue of straight wings versus
swept wings also raised Williams’s ire. He dismissed comments that straight
wings were on the XS-1 because it was not known if swept wings could be used
“for the birds.” The real reason was that no one knew much about swept wings
when the aircraft was being designed.62
Both Williams and James Voyles, the civilian AAF representative for the
XS-1 project, believed Stanley might go as far as to authorize Goodlin to “acci-
dentally” exceed Mach 1. While the suspicions were unjustified, they reflected
the sharp disagreements emerging among all players over the program’s goals
and future direction.
Even the test site still seemed subject to debate. In January 1947, a group of
Ames researchers came to Muroc and were shown around by Gough. They were
impressed by the XS-1 support facilities, including the loading pit. Lawrence
Clousing of Ames commented about it, and Gough replied that a similar
pit would be built at Langley. Clousing was surprised, but Gough told him
Goodlin had landed the XS-1 on the paved Muroc runway. Operating the
rocket plane from Langley, Gough said, would not be a problem.63
The NACA’s plans for the XS-1 had also undergone a change from those
in Williams’s June 7, 1946, letter. NACA management expected to acquire
one of the XS-1s in July 1947, which would mark the start of phase 1 of the
three-phase research effort. The first phase was built around the XS-1 and the
D-558-1 aircraft. Reflecting Stack’s preferences, phase 1 involved only tran-
sonic flights. Phase 2, using the as yet unbuilt Bell XS-2 and Douglas D-558-2
swept-wing research aircraft, would reach speeds up to supersonic. Not until
phase 3 would flights above Mach 1 be made by the NACA. The third phase
would use the XS-3, a jet-powered aircraft with low-aspect straight wings, and
a proposed third derivative of the D-558-1 (which was never built).64
Both Bell Aircraft and the AAF felt that the NACA approach was too slow
and cautious, and indeed it was. Robert Stanley told Larry Bell that the NACA
was only marking time until the D-558 was ready. Stanley also feared that the
NACA would follow the earlier pattern of testing done with the P-80 and
P-84 aircraft. Stanley believed the NACA planned to fly the XS-1 at altitudes
of 20,000 to 30,000 feet until the pilot encountered trim changes or buffeting,
at which point the NACA flights would stop. Stanley pointedly added, “At
these low altitudes, they could do the same with a P-84 since it reaches Mach
27
Probing the Sky
trouble at part[ial] throttle.” Little wonder, then, that Stanley saw no point in
testing the XS-1 in this manner.65
The AAF had its own issues with both Bell and the NACA. From the begin-
ning, reaching Mach 1 was the Army’s primary goal for the XS-1 project. This
was to be accomplished in the shortest time possible. The NACA seemed,
at best, reluctant to take on the challenge. Williams’s initial flight plan and
the later three-phase effort were both going to be slow, long-term efforts. In
contrast, the AAF’s focus was on the near term. The swept-wing XP-86, then
under development, would be approaching supersonic speeds as early as late
1947, and the AAF had to have data on the possible risks it faced and needed
the data sooner, not in a year or more.
The AAF also had management and monetary issues regarding Bell’s efforts.
The contractor’s proposal included no specifics on the length of its project or
any guarantees of results. Colonel Smith also objected to Goodlin’s “bonus
money.” The practice itself was not unusual, and contractor test pilots had long
been paid extra money for risky flights. Bell management and Goodlin had
reached a “handshake agreement” that he would receive $150,000 for making
the first supersonic flight in the XS-1, a not-unreasonable sum given the risks
involved. But the AAF was suffering from the postwar funding cutbacks, and
money for the XS-1 was running short. The AAF was unable to meet the pro-
posed payment, and Smith would not agree to so large a pilot bonus.
Instead, he offered Bell a fixed-price contract, which included specific
requirements Bell had to meet during the tests. At the same time, the NACA’s
perceived foot-dragging and continued objections to AAF decisions convinced
Smith and other AAF personnel that they had to pursue a new approach to
meet the AAF’s research needs.66
The new AAF plan was a two-part complementary program. The AAF
would reach Mach 1 in a minimum number of flights, using a service test pilot.
The NACA would conduct a detailed, incremental flight research effort using
its pilots. Bell’s involvement would end once the company’s original contractual
obligations were completed.
By the spring of 1947, as the AAF was debating the future of the XS-1
flight program, the Flight Test Division personnel were confident that AAF test
pilots were ready for the challenges that the XS-1 posed. Colonel Osmond J.
Ritland recommended to Colonel Smith that the Flight Test Division be given
responsibility for the supersonic test flights.67 Based on Ritland’s assessment,
Smith asked Colonel Albert Boyd, chief of the Flight Test Division at Wright
Field, if his pilots could undertake the accelerated flight test plan for the XS-1.
Boyd’s reply was, “You bet.”
Bell turned down the proposed fixed-price contract, and on May 1, 1947,
Smith notified AAF Headquarters of the company’s decision, adding, “As a
28
result of this notification, discussion is now underway with [the] view of having
this program taken over by [the AAF Air Mobility Command] Flight Test
Division.”68 Larry Bell appealed to General Carl Spaatz, the AAF commander,
to reverse the decision (ultimately to no avail).
Following this interchange, the Flight Test Division began planning for
the AAF program. The instrumentation on the XS-1 would be the minimum
required for measuring the speeds and altitudes reached during its flights. As
many as five glide and powered flights were planned just to reach Mach 0.8.
While the AAF flights would be done in parallel with the NACA’s effort, the
Army would focus on achieving a Mach 1.1 flight in the shortest time possible.
The first XS-1, fitted with the thinner 8-percent thickness-chord wings and
6-percent thickness-chord horizontal stabilizers, had a higher critical Mach
number and speed capability than the thick-wing second XS-1 and was thus
better suited for the AAF’s accelerated assault on Mach 1. A key decision made
by the AAF flight planners was to make the supersonic flights at high altitudes,
thereby reducing the dynamic pressure the XS-1 would experience and mini-
mizing the loads encountered should it experience abrupt transonic pitching
and buffeting. Boyd and Lieutenant Colonel Fred Ascani now reviewed the
records of the 125 pilots at the Flight Test Division and compiled a list of
candidates to fly the plane.69
They selected Captain Charles E. “Chuck” Yeager as the primary XS-1 pilot.
Yeager was then 24 years old, married, and a P-51 ace from World War II. While
Yeager lacked both a college degree and formal engineering training, Boyd
considered him the best instinctive pilot he had ever known. Yeager was the
engineers’ choice because he had an uncanny ability to return from a test flight
and tell the engineers exactly what the airplane had done in response to what
control input at what stage of the flight, in language the engineers immediately
understood, and all after a flight was over, no matter what else had transpired
during the flight. Yeager was clearly a standout, even among the very select
group of test pilots composing the AAF’s Flight Test Division at Wright Field.
For Yeager’s backup, Boyd selected First Lieutenant Robert A. “Bob”
Hoover, another outstanding intuitive fighter test pilot. The AAF’s engineer-
in-charge/project manager was test pilot Captain Jackie L. “Jack” Ridley,
holder of a master’s degree in aeronautical engineering from the California
Institute of Technology, who could furnish the engineering support the two
pilots required.70
On June 24, 1947, General Spaatz concurred with Smith’s accelerated plan
for XS-1 testing, and the following day, Yeager, Hoover, and Ridley joined
personnel from the Flight Test Division and the Aircraft Projects Section to
develop the accelerated test plan, preparing as well for a meeting with NACA
personnel at the end of the month to finalize the two-phase test plan.
29
Probing the Sky
Langley’s personnel proved less than enthusiastic in the change of program

direction. In a June 13, 1947, letter to NACA Headquarters, Henry Reid indi-
cated that Langley researchers had never liked the Bell Aircraft Corporation’s
XS-1 flight program and had only agreed to it in order to establish a good work-
ing relationship with the company, thereby ensuring they got the modifications
that agency engineers wanted before finally accepting the aircraft. The Bell
Aircraft Corporation’s test program would also have shown that the XS-1 was
safe—as far as the NACA was concerned—before the NACA flew the airplane.
Neither of these goals really mattered now that Bell was out of the project.
The result, Reid drily noted, “is considered unfortunate by the Laboratory.”71
Reid was also skeptical about the AAF’s plans. While promising NACA
Headquarters that Langley would cooperate with the joint effort and that the
agency had no objection to the AAF’s taking over the project, he did want
the AAF to acknowledge that Langley disagreed with the expedited flight
plan and still preferred its own plan, as it would provide the largest amount
of research data.72
It was this last point that turned out to be the redeeming element for the
NACA in all this, although no one knew it at the time. In the rush to Mach
1, the AAF made it clear that the details were of secondary concern, while to
the NACA those details were the only concern, and this had been the source
of so much friction between the two groups. In time, knowing exactly what
was happening all along the way turned out to be essential, at which point
the NACA’s methodical, ever-so-slow approach to increased speeds was its
trump card. But in mid-1947, the logic of more rapidly accelerating the drive
through Mach 1 was unassailable, particularly given the urgency of new high
subsonic and transonic aircraft programs such as the XF-86.
AAF and NACA representatives met at Wright Field on June 30 and July
1, 1947, to discuss the coordination of their respective XS-1 plans. Colonel
P.B. Klein opened the meeting by stating that the AAF planned to undertake
an accelerated flight plan with the XS-1 to achieve speeds of Mach 1.1 at alti-
tudes of 50,000 to 60,000 feet in the shortest possible time. AAF personnel
were aware of Reid’s comments and stressed that the XS-1 was a joint effort.
Colonel Boyd emphasized that the Flight Test Division would “appreciate all
of the assistance the NACA personnel could give them in conducting this
program.”73 AAF personnel also asked Bell Aircraft for the help of Richard
Frost, who had been part of the Bell XS-1 team. He had provided technical
advice during the Bell flights and urged that all the AAF tests parties cooper-
ate with him.
Hartley Soulé presented a summation of what had been learned thus far,
the possible technical problems that lay ahead, and the NACA-recommended
procedures that should be used. He noted that no significant compressibility
30
effects on the XS-1 had been identified up to Mach 0.82. He added, “It is
apparent from the flight and wind tunnel data that above M-0.85 large
changes in stability and control and vibrational changes are to be expected.
These have been anticipated in the Langley flights, which will be made at an
altitude of 30,000 feet, and plans have been made to increase speed cautiously
in small increments and to explore conditions at each increment thoroughly
before proceeding to a higher speed.”74
NACA researchers had concerns about the AAF’s plan to make the tests at
high altitudes, and Soulé focused on this in his remarks. He noted that the
Mach 0.8 flight had been made at 30,000 feet, which resulted in a dynamic
pressure on the XS-1 of about 250 pounds per square foot. “At 60,000 feet
for the same Mach number of 0.8, the dynamic pressure would be about 65
pounds per square foot. It appears doubtful, therefore, that any inadvertent
attitude to which the airplane might go as a result of stability and control
changes could result in any structural failures at 60,000 feet.”75
Reflecting how little was then known of transonic flying, Soulé cautioned,
“The pilot should avoid prolonged glides to lower altitudes where the density
is higher because conditions may change critically between the acceleration
and deceleration phase of the flight, and consequently such glides may be
extremely dangerous.”76 He also recommended that AAF personnel read
several NACA flight-data reports, the results of wind tunnel testing, and
other testing results.
He also provided AAF personnel with a copy of the NACA flight plan
and asked for a copy of theirs in return. The Flight Test Division represen-
tatives told them no flight plan had been developed listing specific Mach
numbers and altitudes. Decisions on such matters would not be made until
the flights were underway and would be based on data analysis and Yeager’s
recommendations after each flight.
Aware of the rising tension in the room, Boyd assured Soulé and the rest
of the NACA contingent that the Flight Test Division flight planning was
to be guided by common sense, sound engineering experience, and a focus
on safety. He stressed, however, that the flight plan would be progressive and
brief; the goal remained to fly supersonic in the shortest time, and they would
not be diverted from this objective. The official start of the AAF accelerated
plan was July 10, 1947. The time required to reach Mach 1 was estimated to
be 60 to 90 days. As events would prove, the high end of this estimate was
right on the mark.77
31
Probing the Sky
Toward—and Beyond—the Unknown

Following briefings at Bell Aircraft, Yeager, Hoover, and Ridley traveled to
Muroc, arriving there on July 27. Frost began a 4-day XS-1 “familiarization
schooling” for the new arrivals. Frost later summed up the unknowns of
supersonic flight in the summer of 1947. Regarding an impregnable sound
barrier, “At best, I was ambivalent about it. In hindsight, people may well say
that the so-called ‘sound barrier’ really didn’t prove to be a barrier at all…but
let me assure you, the conditions Chuck was facing going into those flights
were very much a barrier in our minds at that time…. It wasn’t within our
power to give Chuck any real assurances about what might happen in any
one of a multitude of different circumstances.”78
As with Woolams and Goodlin, Yeager began by making glide flights
in the first XS-1. Three of these were made between August 6 and August
9, 1947. Yeager commented that the little saffron speedster was “graceful,
responsive, and beautiful to handle.”79
Due to a shortage of B-29 parts, it was 3 weeks before the powered flights
could begin. Yeager made his first powered flight on August 29 and reached
Mach 0.85. This was followed by four more powered flights in September. Yeager
increased the XS-1’s speed in small increments, going from Mach 0.89 to 0.91,
and then to Mach 0.92. In the process, Yeager was assessing the XS-1’s stability
and control, elevator and stabilizer effectiveness, and buffet. This continued with
another two flights, on October 5 and 8. While there had been problems, the
accelerated effort made good progress. The October 8 flight reached Mach 0.925.
The next flight was made on October 10, 1947, reaching an indicated
airspeed of Mach 0.94 at 45,000 feet. Yeager then pulled back on the control
column and was surprised when the aircraft failed to respond. The shock
wave on the horizontal stabilizers had moved aft until it was on the elevator
hinge line. The elevators had lost all effectiveness, and, as a result, Yeager had
no pitch control of the XS-1. He wrote in his flight report, “[T]he control
column could be moved to the limits of travel each way with little force and
very slow response in airplane attitude.”80
In fact, things were far less bleak than this report suggests. Aerodynamicists
had predicted that the XS-1 would pitch up or down at Mach 1 and that
without the elevator’s pitch control, the aircraft might be lost. That had been,
after all, the reason that the XS-1 had an adjustable horizontal tail installed
in the first place. Significantly, NACA testing had already indicated that the
XS-1 had to use the adjustable stabilizers in order to fly through the speed
of sound. At the request of the AAF, the NACA had tested a 1/16-scale Bell
model of the XS-1 in the Langley 8-foot high-speed tunnel at speeds of up to
Mach 0.945 and Reynolds numbers of 1.18 × 106.
32
The tests clearly revealed the steady decrease in elevator control effective-
ness above Mach 0.87, Langley researcher Axel T. Mattson noted in a report
issued in May 1947 (5 months before Yeager’s flight): “At a Mach number of
0.9, however, the airplane, because of an indicated diving tendency with loss and
reversal in elevator control, will require the use of the stabilizer as a trim control.
Control by the use of the stabilizer is effective at least up to a Mach number
of 0.93, the limit for these tests.”81 (Emphasis added.) So available data clearly
indicated both the necessity for using stabilizer trim in the transonic region
and the likelihood that it would resolve any pitch control problems. In any
case, Yeager’s associate Jack Ridley reassured the airman that by moving the
stabilizers in very small increments of ¼ to ⅓ of a degree, Yeager could retain
pitch control without the elevators.
A genuine surprise came from John Mayer while he was reducing the data
from the flight. After correcting for errors in the XS-1’s airspeed system, he
found that the aircraft had reached a true airspeed of Mach 0.957. After work-
ing through the weekend, he again revised his analysis, concluding the XS-1’s
true airspeed was Mach 0.997 at 37,000 feet. Only the narrowest of margins
still remained.
On the morning of October 14, 1947, the B-29 with the XS-1 attached
underneath took off from the Muroc runway. Due to the loss of pitch control
on the previous flight, NACA engineers told Yeager to limit his speed to Mach
0.96 unless he was certain he could safely fly faster. Yeager himself was not in
the best condition as he climbed into the XS-1 cockpit. The previous weekend,
he had fallen from a horse at aviatrix Florence “Pancho” Barnes’s notorious
Happy Bottom Riding Club and had broken two ribs. In pain, Yeager went
to a civilian doctor in Rosamond to have them taped up, rather than go to an
Air Force flight surgeon and risk being grounded. Ridley, by now a firm Yeager
friend, knew about the mishap and cut a 10-inch length of broom handle to
help Yeager lock the hatch from inside the rocket plane, despite his broken ribs.
The drop was made at 10:26 a.m., at an altitude of 20,000 feet and an
airspeed of 250 mph. This was slower than expected, and Yeager had to lower
the XS-1’s nose to avoid stalling. He fired the four rocket chambers in rapid
succession and began to accelerate and climb. He shut down two of the rocket
chambers and began using the movable stabilizer for pitch control, finding it
“very effective.”
Yeager leveled out at approximately 42,000 feet and fired a third rocket
cylinder. The XS-1 accelerated rapidly to an indicated airspeed of Mach 0.98.
The Machmeter needle fluctuated and jumped off scale. There was no violent
buffeting. The aircraft did not pitch up or down. There were no indications
that something unusual had happened. Yeager held the speed for 20 seconds
before shutting off the engine. Yeager radioed, “Ridley! Make another note.
33
Probing the Sky
There’s something wrong with this Machmeter. It’s gone screwy!” Ridley
replied, “If it is, we’ll fix it. Personally, I think you’re seeing things.”82 NACA
personnel analyzed the flight data and determined that the XS-1 had reached
a speed of Mach 1.06. The task had taken 96 days since the AAF took over
the program on July 10, 1947. Within hours, the achievement was classified,
and no public announcement was made.83 Not quite three months later, at
the express direction of Air Force Secretary Stuart Symington, the Army Air
Forces held a secret conference at Wright Field to present the results of the
program to a select audience of the Nation’s leading designers, engineers, and
aeronautical researchers.84 NACA’s Muroc Flight Test Unit harvested the tech-
nical results of the flight in a series of analytical reports issued over the next
several years, buttressing the quick look afforded by the Air Force’s accelerated
flight test program.85
Beyond Yeager’s landmark achievement awaited a new series of unknowns.
There were many design ideas about how a supersonic airplane should be con-
figured. However, there was little solid information and, with wind tunnels still
unable to provide reliable data, few means to find out more. There remained
an interrelated series of problems in aerodynamics, propulsion, and aircraft
configurations to bedevil and bother the aeronautical community in the years
ahead. Nevertheless, if there were many challenges and unknowns ahead, there
was one major bedrock accomplishment: the myth of an impenetrable barrier
had been forever vanquished.
34
Endnotes
1. Opening quote from Louis Rotundo, Into the Unknown: The X-1
Story (Washington, DC: Smithsonian Institution Press, 1994), p. 6.
2. James R. Hansen, Engineer in Charge: A History of the Langley
Aeronautical Laboratory, 1917–1958 (Washington, DC: NASA
SP-4305, 1987), p. 253; Richard P. Hallion, Supersonic Flight:
Breaking the Sound Barrier and Beyond—The Story of the Bell X-1
and Douglas D-558 (London: Brassey’s, 1997), p. x. When asked
about these stories, John McTigue, a former X-15 engineer, recalled
reading such claims in newspaper articles and letters to the edi-
tor circa 1944–45. The stuck-voice story was false, as the air in the
cockpit was not moving with respect to the pilot. The time-reversal
story represents a misunderstanding of the theory of relativity. As an
object nears the speed of light, time slows down. The crew of a star-
ship traveling at nearly the speed of light would experience a flight
as lasting only a few months. Simultaneously, back on Earth, years,
decades, or even centuries would pass. In no case, however, can time
actually be “reversed.”
3. F.W. Caldwell and E.N. Fales, “Wind Tunnel Studies in
Aerodynamic Phenomena at High Speeds,” Report No. 83 (NACA,
1920), p. 59.
4. Ibid., pp. 60, 77, 86, 89.
5. L.J. Briggs, G.F. Hull, and H.L. Dryden, “Aerodynamic
Characteristics of Airfoils at High Speeds,” Report No. 207 (NACA,
1924), pp. 465, 478.
6. L.J. Briggs and H.L. Dryden, “Pressure Distribution Over Air Foils
at High Speeds,” Report No. 255 (NACA, 1926), p. 555.
7. L.J. Briggs and H.L. Dryden, “Aerodynamic Characteristics of
Twenty-Four Airfoils at High Speeds,” Report No. 319 (NACA,
1929), pp. 327–328, 345–346.
8. Ibid.
9. L.J. Briggs and H.L. Dryden, “Aerodynamic Characteristics of
Circular-Arc Airfoils at High Speeds,” Report No. 365 (NACA,
1931), pp. 67, 69.
10. NASA, “The National Advisory Committee for Aeronautics: Tracing
NASA’s 95-Year-Old Roots,” http://www.nasa.gov/centers/ames/news/
features/2010/95_anniversary_prt.htm, updated March 3, 2010.
11. Hansen, Engineer in Charge, pp. xxviii, 2, 126–133.
12. Eric Schatzberg, Wings of Wood, Wings of Metal (Princeton: Princeton
University Press, 1998). This book looks at the social factors, as
35
Probing the Sky
opposed to engineering requirements, behind the switch from

wooden construction toward all-metal aircraft. These include a per-
ception that metal was “modern,” while wood was a “preindustrial”
material. Another factor was the influence of the U.S. military on
aircraft design.
13. For a discussion of these, see G. Geoffrey Smith, Gas Turbines and
Jet Propulsion for Aircraft (New York: Aircraft Books Inc., 1946), pp.
34–50. For the jet engine more generally, see Edward W. Constant
II, The Origins of the Turbojet Revolution (Baltimore: Johns Hopkins
University Press, 1980), pp. 63, 64, 70–72, 77, 89, 93, 95; and
Robert Schlaifer and S.D. Heron’s Development of Aircraft Engines
and Fuels (Boston: Harvard Business School, 1950).The best-
known of these hybrid concepts was the “Campini engine,” after
the Italian engineer Secondo Campini, who successfully flew one in
1940–41, but many other concepts for such engines existed as well.
In England, the term “motorjet” was also used. For Campini’s work,
see Gregory Alegre, Campini Caproni, no. 5 of the Ali d’Italia series
(Turin: La Bancarella Aeronautica, n.d.).
14. For American work, see James St. Peter, The History of Gas Turbine
Development in the United States…A Tradition of Excellence (Atlanta,
GA: International Gas Turbine Institute of the American Society of
Mechanical Engineers, 1999).
15. Edgar Buckingham, “Jet Propulsion for Airplanes,” Report No. 159
(NACA, 1923), p. 85.
16. For a particularly useful introduction to the history and devel-
opment of the jet engine, see Walter J. Boyne and Donald S.
Lopez, eds. The Jet Age: Forty Years of Jet Aviation. Washington,
DC: National Air and Space Museum in association with the
Smithsonian Institution Press, 1979.
17. Margaret Connor, Hans von Ohain: Elegance in Flight (Reston, VA:
American Institute of Aeronautics and Astronautics, 2001).
18. For Whittle’s view of his work, see his autobiographical Jet: The
Story of a Pioneer (New York: Philosophical Library, 1954) and John
Golley’s (in association with Whittle and with technical assistance
from Bill Gunston) semiautobiographical Whittle: The True Story
(Washington, DC: Smithsonian Institution Press, 1987). See also
John Grierson’s memoir-history Jet Flight (London: Samson Low,
Marston & Co., Ltd., 1944); and Andrew Nahum’s thought-
provoking Frank Whittle: Invention of the Jet (Duxford, U.K.: Icon
Books, 2004).
36
19. For the impact (in both directions) of turbojet development and
combat requirements, see Sterling Michael Pavelec, The Jet Race and
the Second World War (Annapolis, MD: Naval Institute Press, 2007).
20. Hansen, Engineer in Charge, pp. 230–231.
21. Hansen, Engineer in Charge, pp. 231–232. Dr. Vannevar Bush’s
interest in the Jeep engine concept also reflected his own technologi-
cal conservatism and academic mindset. Despite his accomplish-
ments as a scientist, he endorsed the same jet concept the NACA
was studying. For the record, Bush consistently displayed a pattern
of being too quick to dismiss the possibilities of technological break-
throughs. In late 1945, for example, he told a congressional com-
mittee that a long-range rocket armed with a nuclear weapon was
impractical and that the Congress and the American people should
leave it out of their thinking. In 1949, he predicted it would be
many years before the USSR would build an A-bomb. The Soviets
tested their first A-bomb that year.
22. Ibid., p. 260; Hallion, Supersonic Flight, pp. 19–20.
23. Hansen, Engineer in Charge, p. 244.
24. Ibid., pp. 236–245; Hallion, Supersonic Flight, pp. 18–21. The
XP-59A itself was an example of what Jacobs was describing. It was
initially intended to evolve into an operational combat aircraft rather
than remain a technology demonstrator. The goal was to have a com-
bat aircraft ready sooner than first building an experimental aircraft,
followed by an operational fighter. (Much like the Durand Committee
had proposed for Jacobs’s research aircraft.) While the turbojets were
revolutionary, the XP-59A’s airframe was a late-1930s high-drag
design. This limited it to a top speed of 389 mph. As a result, the few
P-59s built were used as jet trainers to introduce pilots into the higher-
performance Lockheed P-80A that succeeded it.
25. John V. Becker, The High-Speed Frontier: Case Studies of Four NACA
Programs, 1920–1950 (Washington, DC: NASA SP-445, 1980), p. 90.
27. Hallion, Supersonic Flight, p. 22.
28. Ibid., pp. 20–23.
29. Ibid., pp. 24–25.
31. Becker, The High-Speed Frontier, pp. 91–92.
32. Becker, The High-Speed Frontier, p. 91; Hansen, Engineer in Charge,
pp. 271–273. “XS-1” continued to be used into 1948, when the Air
Force revised its designation system. The “S” was dropped for all of
37
Probing the Sky
the research aircraft (XS-1, XS-2, XS-3, and XS-4), and they became
the X-1, X-2, X-3, and X-4.
33. Hallion, Supersonic Flight, pp. 31–32, 61, quotes the Hyatt memo
at length, which was USN Bureau of Aeronautics AER-E-225-AH,
“Proposed High-Speed Research Airplane” (September 22, 1944).
See also Rotundo, Into the Unknown, p. 21; and Hansen, Engineer in
Charge, pp. 273–274.
34. Ibid., pp. 65–67, 69, 70; Hansen, Engineer in Charge, pp. 290–291.
Unlike the AAF or the Air Force, the U.S. Navy never had a spe-
cific designation for a purely experimental aircraft. Until September
1962, the Navy and Air Force had separate aircraft designation
systems, even for the same aircraft.
35. Hansen, Engineer in Charge, pp. 275–279; Hallion, Supersonic
Flight, pp. 71–72. Stack and Gilruth also jointly recommended that
the thickness-chord ratio of the XS-1’s horizontal stabilizer be less
than that of the wing. This was to prevent them from both expe-
riencing shock wave formation at the same time, with attendant
changes such as turbulence, loss of control effectiveness, etc. They
also recommended attaching the horizontal stabilizer midway up
the vertical fin to avoid a loss of effectiveness due to the wing wake
impinging upon it. A final joint recommendation was to have the
horizontal stabilizer be adjustable, effectively making the combined
stabilizer and elevator surface almost an “all-moving tail.” As an
airplane’s speed increased, the shock wave moved aft on the hori-
zontal stabilizer. When it reached the hinge line of the movable
elevator, the control surface would be blanked out and the pilot
would lose pitch control to such a degree that, in many cases, he
could freely move the elevator without his control inputs having any
effect upon flight path control. To avoid this, the stabilizer’s angle
could be changed by the pilot, giving a measure of control. The
same design approach was also applied to the D-558-1 and the Air
Force’s F-86E Sabre jet fighter. Later supersonic aircraft such as the
American F-100A and Soviet MiG-19 had a one-all-moving hori-
zontal stabilizer.
36. Hallion, Supersonic Flight, pp. 52–53.
38. Ibid., p. 288.
39. Ibid., pp. 288–289.
40. Rotundo, Into the Unknown, p. 31.
41. Ibid., pp. 35–36. Stan Smith and Bell Aircraft management had
a private reason for rejecting Muroc Army Air Field in California.
38
The West Coast was the home, by the mid-1940s, of many U.S.
aviation contractors. Had the XS-1 gone to Muroc, many of the
engineers might have decided to forego the Buffalo winters for the
California sunshine.
42. Robert T. Jones, “Wing Plan Form for High-Speed Flight,” in
Collected Works of Robert T. Jones, TM X-3334 (Washington, DC:
NASA, 1976), p. 379.
43. Ibid., p. 379; Richard P. Hallion, “Lippisch, Gluhareff, and
Jones: The Emergence of the Delta Planform and the Origins
of the Sweptwing in the United States,” Aerospace Historian 26,
no. 1 (spring 1979), passim; Hansen, Engineer in Charge, pp.
282–283; MSN Encarta, “Sound Barrier,” http://uk.encarta.msn.
com/text_781533139__3/Sound-Barrier.html, accessed November
30, 2009 (cached page). The development of Robert T. Jones’s wing
planform paper shows the complex process by which technological
and scientific breakthroughs are made. It started with Jones’s efforts
to develop a lift theory for a dart-shaped missile. This, in turn, led
Jones to a pair of very different studies. The first, by Max Munk, was
of low-speed airflow around dirigibles, while the other, by Hsue-Shen
Tsien, was on supersonic airflow around spinning projectiles, which
indicated a lack of compressibility effects on these bodies. When Jones
factored Tsien’s equations into his calculations, it appeared that very
slender wings would also show no compressibility. Next, Jones recalled
another paper by Munk, which dealt with the effects on aircraft stabil-
ity due to dihedral and wing sweep. This was the first paper to describe
the basic effects of sweepback on a wing. But because it only dealt
with low subsonic speeds, the implications for transonic and super-
sonic flight went unrealized. Once Jones incorporated Munk’s second
paper into the calculations, he was able to create a general theory that
was valid for all wing-sweep angles.
What Jones did not know was also important. Adolf Busemann, a
German aerodynamicist, was among those giving papers at the Volta
Congress on High-Speed Aeronautics in 1935. Part of his highly
theoretical paper dealt with “arrow wings” as a means of reducing
wave drag. Busemann’s theory did not include the idea of position-
ing the wing behind the Mach line, however. As a result, the cross
flow over the wing was still supersonic. Three American scientists
attended the Volta Congress—Eastman Jacobs, Hugh Dryden, and
Theodore von Kármán. None of them viewed Busemann’s idea as
important. Not until after Jones had written an initial draft of his
paper was a 1942 British translation of Busemann’s paper found in
39
Probing the Sky
the Langley library. Jones incorporated a citation listing it in the

revised text. See Richard P. Hallion, “Sweep and Swing: Reshaping
the Wing for the Jet and Rocket Age,” in Richard P. Hallion, ed.,
NASA’s Contributions to Aeronautics, vol. 1: Aerodynamics, Structures,
Propulsion, Controls, SP-2010-570-Vol 1 (Washington, DC: National
Aeronautics and Space Administration, 2010), pp. 5–17.
Of the three papers by Munk and Tsien, written in the 1920s and
1930s, only one was related to supersonic flight. Despite this, they
played a role in Jones’s development of this swept-wing paper. This
influence was due to Jones’s ability to realize their relevance to the
problems he was investigating, despite their subject matter. The use
of swept wings on the German Me 262 jet fighter and the Me 163
rocket fighter also had no influence on Jones’s paper, as these aircraft
were subsonic and the swept wings were used primarily to maintain
their center of gravity. Indeed, the Me 163 had profoundly poor—
even dangerous—high-speed stability and control characteristics
induced by its tailless swept-wing planform, as discussed subse-
quently in the chapter on the Northrop X-4 Bantam.
45. Ibid., p. 284.
46. Ibid., pp. 105–107, 284–285.
47. Ibid., pp. 281–284; Hallion, Supersonic Flight, p. 45.
49. Ibid., pp. 289–290; Hallion, Supersonic Flight, pp. 46–47.
50. Rotundo, Into the Unknown, pp. 43–44.
51. Ibid., p. 50.
52. Ibid., pp. 51–59, 62. Langley’s location near the ocean and nearby
rivers, as well as its humid climate, all ruled out operating the XS-1
from the site. Pinecastle had better weather than Langley, but this
still interfered with the glide flights. Given the poor visibility from
the XS-1 cockpit and the landing accidents during the Pinecastle
glide flights, it seems the loss of the airplanes would have been nearly
certain had the flights operated from Langley’s runway.
54. John Ball, Jr., Edwards: Flight Test Center of the USAF (New York:
Duell, Sloan and Pearce, 1962), pp. 38–41.
55. For the story of this remarkable engine, see Frank H. Winter,
“‘Black Betsy’: The 6000C4 Rocket Engine, 1945–1989,” a paper
presented at the 23rd Symposium on the History of Astronautics,
40th International Astronautical Congress of the International
Astronautical Federation, Malaga, Spain, October 1989.
40

57. Ibid., pp. 91–92, 98–102; Richard P. Hallion and Michael H. Gorn,
On the Frontier: Experimental Flight at NASA Dryden (Washington,
DC: Smithsonian Books, 2003), pp. 21–24.
58. Rotundo, Into the Unknown, pp. 102–103.
59. Ibid., pp. 118–119.
60. Ibid., pp. 104–109, 120, 125–126; James O. Young, Meeting the
Challenge of Supersonic Flight (Edwards Air Force Base, CA: Air
Force Flight Test Center History Office, 1997), p. 35.
61. Rotundo, Into the Unknown, p. 151.
62. Ibid., p. 152.
63. Ibid., p. 162. The reality was that the airplane Williams described
was the one that did not exist. The turbopump would not be ready
for several more years, no provisions for a reserve fuel supply were
ever included in the aircraft, and a ground takeoff capability was
impractical and unsafe.
64. Ibid., pp. 39, 181.
65. Ibid., pp. 191, 193–194; Young, Meeting the Challenge of Supersonic
Flight, p. 36.
66. Young, Meeting the Challenge of Supersonic Flight, pp. 34–37.
67. Ibid., pp. 38–39.
68. Ibid., p. 40.
69. Young, Meeting the Challenge of Supersonic Flight, pp. 39–41.
70. Ibid., pp. 41–42.
71. Rotundo, pp. 230–231.
72. Ibid., pp. 230–231.
73. Young, Meeting the Challenge of Supersonic Flight, p. 45.
74. Ibid.
75. Ibid.
76. Ibid.
77. Ibid., pp. 45–46.
78. Ibid., p. 47.
79. Ibid., p. 48.
80. James O. Young, Supersonic Symposium: The Men of Mach 1, Air
Force Special Code (AFSC) Special Study, September 1999, p. 219.
81. Axel T. Mattson, “Force and Longitudinal Control Characteristics
of a 1/16-scale Model of the Bell XS-1 Transonic Research Airplane
at High Mach Numbers,” NACA RM L7A03 (May 21, 1947),
p. 9–10. All NACA Research Memorandums in this book can be
found on the NASA Technical Reports Server (http://ntrs.nasa.gov/
search.jsp) or within Dryden Flight Research Center’s archives.
41
Probing the Sky
82. From flight transcript, reprinted in Young, Meeting the Challenge of

Supersonic Flight, p. 73.
83. Hallion, Supersonic Flight, pp. 113–119, 124–125; Rotundo, Into the
Unknown, pp. 274–279; Young, Meeting the Challenge of Supersonic
Flight, pp. 52–59. The secret held until December 22, 1947, when
the Los Angeles Times carried a front-page story. The headline read,
“U.S. Mystery Plane Tops Speed of Sound.” The story had a number
of significant errors: Captain Charles “Yaeger” had broken the sound
barrier at 70,000 feet, NACA research pilots Herbert Hoover (true)
and Howard Lilly (false) had also flown supersonic, and Goodlin
had made the first 30 flights (false). Not until June 15, 1948, was
the flight officially confirmed in a joint press conference by General
Hoyt S. Vandenberg, Air Force Chief of Staff, and Dr. Hugh L.
Dryden, NACA Director of Research. On December 17, 1948,
the Robert J. Collier Trophy, aviation’s most prestigious award,
was given to the XS-1 project. Its citation read: “To Robert Stack,
Research Scientist, NACA, for pioneering research to determine the
physical laws affecting supersonic flight, and for his conception of
transonic research airplanes; to Lawrence D. Bell, President, Bell
Aircraft Corporation, for design and construction of the special
research aircraft X-1; and to Captain Charles E. Yeager, U.S. Air
Force, who, with that airplane, on October 14, 1947, first achieved
human flight faster than sound.”
It is ironic that Stack, who preferred the D-558 over the X-1 and
who objected to its propulsion system, its launch method, and its
test site, should win the Collier Trophy for its achievements. In con-
trast, Ezra Kotcher, whose “Mach 0.999 study” and whose support
of a rocket engine over a turbojet led directly to the X-1, was most
unfortunately ignored, at least in the Collier citation.
84. The symposium papers were subsequently published by the USAF
as Air Force Supersonic Research Airplane XS-1 Report No. 1 (Wright
Field, Dayton, OH: USAF Air Materiel Command, January 9,
1948). This historic report was subsequently reprinted and reissued
in October 1997 by Richard P. Hallion, then the Air Force Historian
at the Air Force History and Museums Program, Headquarters Air
Force, Washington, DC, in commemoration of the 50th anniversary
of the first piloted supersonic flight.
85. These reports included the following: Ellwyn E. Angle and Euclid C.
Holleman, “Determination of Longitudinal Stability of the Bell X-1
Airplane from Transient Responses at Mach Numbers Up to 1.12 at
Lift Coefficients of 0.3 and 0.6,” NACA RM L50I06a (November 7,
42
1950); De E. Beeler and John P. Mayer, “Measurements of the

Wing and Tail Loads During the Acceptance Tests of Bell XS-1
Research Airplane,” NACA RM L7L12 (April 13, 1948); L. Robert
Carman and John R. Carden, “Lift and Drag Coefficients for the
Bell X-1 Airplane (8-Percent-Thick Wing) in Power-Off Transonic
Flight,” NACA RM L51E08 (June 25, 1951); Hubert M. Drake
and John R. Carden, “Elevator-Stabilizer Effectiveness and Trim of
the X-1 Airplane to a Mach Number of 1.06,” NACA RM L50G20
(November 1, 1950); Hubert M. Drake, Harold R. Goodman,
and Herbert H. Hoover, “Preliminary Results of NACA Transonic
Flights of the XS-1 Airplane with 10-Percent-Thick Wing and
8-Percent-Thick Horizontal Tail,” NACA RM L8I29 (October 13,
1948); Hubert M. Drake, Milton D. McLaughlin, and Harold R.
Goodman, “Results Obtained During Accelerated Transonic Tests
of the Bell XS-1 Airplane in Flights to a Mach Number of 0.92,”
NACA RM L8A05a (April 19, 1948); Harold R. Goodman and
Hubert M. Drake, “Results Obtained During Extension of U.S.
Air Force Transonic-Flight Tests of XS-1 Airplane,” NACA RM
L8I28 (November 16, 1948); Harold R. Goodman and Roxanah B.
Yancey, “The Static-Pressure Error of Wing and Fuselage Airspeed
Installations of the X-1 Airplanes in Transonic Flight,” NACA RM
L9G22 (August 12, 1949); Walter C. Williams and De E. Beeler,
“Results of Preliminary Flight Tests of the XS-1 Airplane (8-Percent
Wing) to a Mach Number of 1.25,” NACA RM L8A23a (April 6,
1948); and Walter C. Williams, Charles M. Forsyth, and Beverly
P. Brown, “General Handling-Qualities Results Obtained During
Acceptance Flight Tests of the Bell X-1 Airplane,” NACA RM
L8A09 (April 19, 1948).
43
The third D-558-1 parked on the South Base ramp with three ground personnel. The Skystreak
was one of two concepts for how best to gain data on transonic flight. It represented a tradi-
tional design, had straight wings, and was jet-powered, with a top speed just below Mach 1. It
was funded in part by the Navy and built by Douglas to NACA specifications. In contrast, the
Army Air Forces XS-1 was rocket-powered, with a top speed well above Mach 1. (NASA)
44
CHAPTER 2
Flying Test Tube:
The Douglas D-558-1 Skystreak
The Model D-558 gives the impression of being an outstandingly excellent job of
design and engineering, and a very sound airplane for research purposes….
—Captain Frederick M. Trapnell, U.S. Navy1
If the D-558-1 could have been promoted in the early forties, it would have been
timely. But coming into the flight picture as it did in 1947, it was unnecessary.
—John V. Becker, NACA
The Douglas D-558-1 Skystreak always flew in the shadow of the Bell XS-1.2
If radical by the standards of conventional propeller-driven airplanes, then by
the standards of the rocket-powered XS-1 the Skystreak was a very conservative
design, reflecting the NACA’s desire for a turbojet-powered research airplane
that could effectively “loiter” in the transonic regime, nibbling at the sonic fron-
tier. If overshadowed by its flashier rocket-powered contemporary, it admirably
fulfilled the expectations of the NACA, playing a significant role in the early years
of high-speed research with its capability to undertake a wide range of research
activities over a 6-year period, effectively “freeing up” its higher-performance
rocket-powered rivals to explore the frontiers of the supersonic regime while it
generated detailed information on the high-subsonic and transonic.
In performance, appearance, mode of operation, and systems, the D-558-1
appeared little different in design from the first generation of U.S. jets. Chief
engineer Edward H. Heinemann, L. Eugene Root, Kermit Van Every, A.M.O.
Smith, Robert C. Donovan, R.G. Smith, and the other members of the Douglas
design team faced a number of challenges if the D-558-1 was to be capable
of accomplishing the research goals that the NACA and the Navy’s Bureau of
Aeronautics had set for it.3
The first of these tasks was to minimize the fuselage’s frontal area so as to
maintain a high fineness ratio, the ratio of the fuselage’s length to its diameter.
Drop-body tests showed that a high fineness ratio reduced drag at transonic
speeds. At the same time, the fuselage shape had to prevent airflow from being
accelerated to supersonic speeds, which would cause a flow-disrupting shock
45
Probing the Sky
wave to form in the vicinity of the wing. To address these issues, the best solu-
tions were also the simplest.
The team selected a cylindrical fuselage, one with a diameter just big enough
to hold the TG-180 turbojet engine; indeed, afterwards, A.M.O. Smith (who
determined the wing planform and airfoil section and did much of the work on
the aerodynamics of the nose inlet) recalled the design process as “a case of wrap-
ping the smallest airplane around the largest jet engine that was available.”4 This
shape also minimized airflow acceleration over the wing. A final design require-
ment was that of the fuselage/wing fillet, which needed a critical Mach number
as high as that of the wing but which would not exhibit poor stall characteristics.
Meeting this specification would require considerable NACA wind tunnel testing.
The wings used the NACA’s 65-110 airfoil (a symmetrical “65-series” wing
section with a thickness-chord ratio of 10 percent), as it had a high critical
Mach number, had good high-speed characteristics, and had been used previ-
ously by the Douglas engineers for their A-26 medium bomber, the highest-
performance twin-engine medium bomber developed during the Second
World War. As with the XS-1, the Douglas team employed a straight-wing
planform. But in contrast to the relatively high aspect ratio of the XS-1’s wings,
the D-558-1 had low-aspect-ratio wings (a short wingspan with a wide chord).
Further, all the D-558-1s employed a 10-percent thickness-chord section, not
the lower 8-percent thickness-chord section employed on the first of the XS-1s.
The horizontal stabilizer was similar to that on the XS-1. It had a thinner airfoil
than did the wing (6 percent for the stabilizer versus 10 percent on the wing)
and was mounted higher up on the vertical tail, with the forward section made
movable to retain control at high transonic speeds. The pilot had a switch on
the control wheel for moving the stabilizer up and down.5
The D-558-1 structural design was described by Heinemann as being com-
paratively conventional and straightforward. It did have some unusual features,
however, that were different from those of contemporary aircraft structures.
In an era dominated by aluminum skinning, the Skystreak’s fuselage skin was
a magnesium alloy slab 1/10-inch thick, which eliminated the need for form-
ers or stringers, both used in conventional fuselage designs, thus ensuring a
light yet rigid and strong fuselage tube, saving at least 60 pounds in structural
weight over a conventional rib-stringer-skinning design. The only internal
reinforcement used was at points of concentrated structural loads. As a result,
the D-558-1’s internal volume was relatively large.
The wings and the tail were of standard aluminum construction, using
high-strength alloy in a conventional rib and spar matrix. To ensure that their
external skins were as smooth as possible, sheet-metal fabricators and crafts-
people attached the fuselage and wing skins to rigid contoured frames and then
attached the internal framework directly to the skin.
46
Flying Test Tube: The Douglas D-558-1 Skystreak
The thin wings posed several design problems. First was fuel capacity, for
the aircraft had to carry enough fuel for a research flight lasting about 30 min-
utes from takeoff through landing. The solution designers chose was to make
the front half of the wing into an integral fuel tank, a so-called “wet wing.”
Technicians sealed the wing interior using a synthetic rubber compound, which
required 5 weeks to cure and dry. Once ready, the tank formed by this process
held 230 gallons of kerosene fuel. To extend flight time and altitude, a pair of
50-gallon tip tanks could also be carried, which when used added another half
hour to flight duration (and which were, in fact, employed at various points
during the test programs flown on the three Skystreak aircraft).
The D-558-1 had tricycle landing gear, but, here, too, the thin wing cre-
ated a challenge. While the nosewheel could retract into the forward fuselage
behind the cockpit, the tightly packed fuselage—occupied by the engine,
controls, instrumentation, and other systems—lacked any room for the twin
main wheels. Accordingly, the only place for the main landing gear was inside
the wings, but, given the thin airfoil, the wheels had to be much thinner and
smaller than would have been typical for an aircraft of the D-558-1’s size.
Douglas engineers approached the Goodrich Corporation to produce special
20-by-4.4, 8-ply nylon tires. Originally, the tires were intended to operate at
230 pounds per square inch (psi), though this was later reduced to a more
forgiving 175 psi. As Heinemann later noted, “It was realized from the outset
that this wheel and tire size was much smaller than desired for the load, but
the selection was considered justified due to the serious effect larger wheels
would have had upon the size of the airplane.”6
Additionally, Douglas engineers had the issue of pilot escape from the
D-558-1 at high speeds to consider. Aeromedical experts were asked about
escape options and indicated they doubted a human could withstand a normal
bailout at high speeds. The air blast was considered too great for a pilot to sur-
vive, and the acceleration needed to propel both an ejection seat and the pilot
clear of the vertical tail was in excess of human anatomical limits.
The approach taken was to use a capsule escape system. The capsule’s nose
was attached to the fuselage at four points. In an emergency, the pilot would
pull a handle to release the four attachments, freeing the nose section. Once
it fell away and slowed, the pilot pulled a second handle that dropped the
seat back, then fell out of the nose section and opened his parachute. Similar
capsules were later fitted to the D-558-2 and the Bell X-2 research aircraft. In
both X-2 crashes, the capsules proved ineffective, and both pilots died.7
To aid optical tracking from the ground, the D-558-1 was painted insignia
red (though this was subsequently changed to white, as the red actually made
the aircraft far less visible against the dark blue Mojave sky). The aircraft carried
standard U.S. “star-and-bar” markings, and the words “Douglas Skystreak” were
47
Probing the Sky
painted on the nose in a stylized typeface suggesting speed and modernity. With
its red paint finish, cigarette-like cylindrical fuselage, nose inlet, and rounded silver
jet nozzle, the aircraft was not surprisingly nicknamed the “Crimson Test Tube.”8
“The arrangement of the airplane was quite conventional,” Heinemann sub-
sequently recalled, “as it was believed the best way to obtain the largest amount
of useful data in the shortest period of time was to employ only design features
that were well known and did not involve uncertainty.”9 He also noted, “It is
considered necessary that all of the high speed aerodynamic data be obtained
in level flight instead of vertical or near vertical dives as in the past.”10 These
stipulations were a reflection of the NACA’s viewpoint, goals, and influence.
First Flights and Speed Records

In January of 1947, the D-558-1 #1 (Bureau of Aeronautics number [BuNo]
37970) was completed and several months of ground testing was begun. Two
trucks then transported the disassembled aircraft from the Douglas plant at
El Segundo over the San Gabriel Mountains to the Mojave Desert and Muroc
Army Air Field. After reaching Muroc on April 10, the aircraft was reassembled,
the engine and other systems were tested, and preparations were made for the
first flight.
Veteran Douglas test pilot Eugene F. “Gene” May, who had flown for the
company since 1941, was selected for the initial flights. The airplane was ready
by April 15 for the first flight. May took off from the lakebed and immediately
ran into trouble. The TG-180 engine suffered a partial power loss, and May
landed straight ahead on the lakebed. When he applied the brakes, the left
brake disintegrated and he had to hold the left rudder to keep rolling forward
in a straight line. On April 21, similar problems plagued the second attempt,
grounding the D-558-1 until the end of May 1947.
Landing gear problems were encountered on the next six flights, with the
gear either not retracting or locking in place. Even so, by the 12th flight, on July
12, 1947, the aircraft had demonstrated satisfactory low-speed flight charac-
teristics. Douglas engineers began modifications needed for transonic research
flights. The low-speed, clear-bubble canopy was replaced with a V-shaped,
reinforced high-speed hooded windscreen. The new canopy design was found
by several pilots to be too small to allow them to wear a helmet, and they flew
without one. With the work complete, the aircraft made an airspeed calibra-
tion check on its 13th flight.
May now began a buildup to transonic speeds, beginning with the 14th
flight made on July 17. During the course of the next six flights, May reached a
speed of Mach 0.85 on August 5. During this same period, the second D-558-1
48
From left to right, Eugene May (Douglas Aircraft) and Howard Lilly (NACA research pilot) pose in
front of the second D-558-1, which was destroyed in a crash on May 3, 1948. Lilly was killed in
the accident, the first NACA research pilot lost in the line of duty. (NASA)
(BuNo 37971) was completed and delivered to Muroc Army Air Field, where
it was to be used by NACA pilots. The next step, however, was an attempt at
the world’s airspeed record.11
The Navy was interested in using the second D-558-1 to best the exist-
ing airspeed record of 615.778 mph, set by British Royal Air Force Group
Captain E.M. Donaldson in a modified Gloster Meteor in September 1946.
The rules for an official record recognized by the Fédération Aéronautique
Internationale (FAI) required that an aircraft fly at an altitude below 250 feet
(75 meters) and make four passes along a 1.864-mile (3-kilometer) course.
While the discussions were underway, Donaldson’s record had fallen to the
AAF. Colonel Albert Boyd broke the record on June 17, 1947, in a modified
P-80R jet, reaching a speed of 623.738 mph, the last speed record set by the
service before it transformed into the independent United States Air Force in
September of that year. The higher speed increased the challenges both the
airplane and pilots would have to overcome.
With the second D-558-1 delivered and test-flown, the way was clear for the
record attempts. Navy Commander Turner F. Caldwell made the first record
attempt on August 20, 1947, in the first D-558-1 and achieved an average
speed in the four passes of 640.663 mph, breaking the standing record by
16.924 mph. Marine Major Marion E. Carl’s turn came on August 25. He
had an idea for a way to squeeze a little more speed from the second D-558-1.
49
Probing the Sky
On Caldwell’s flight, the TG-180 engine had produced 100 percent rotations
per minute (rpm) on the ground, but power readings dropped to 98 percent
rpm once he took off. Carl convinced the Douglas ground crew to raise the
engine’s rpm to 102 percent on the ground. As expected, the same rpm drop
occurred after takeoff, leaving Carl with 100 percent rpm in flight. When his
four passes were averaged, he had achieved 650.796 mph, raising the record
by 10.133 mph, small but enough.12
With the world speed record now in Navy/Marine hands, attention shifted
back to the test-and-research effort. The first D-558-1 resumed its contractor
flights, piloted by May. Another 18 flights were flown before it was turned over
to the NACA on October 23, 1947, a week after Chuck Yeager and the rival
XS-1 had broken the sound barrier.
The NACA D-558-1, the second of the three Skystreaks, had not been
fitted with data instrumentation, so the first step in preparing the aircraft was
to install a standard NACA recording package, overseen by the Muroc Flight
Test Unit’s chief instrumentation and telemetry tracking engineer, Gerald M.
“Gerry” Truszynski:13
Instrumentation Data Collected
Airspeed-altitude recorder Indicated airspeed and pressure altitude
Three-component accelerometer Normal, longitudinal, and transverse acceleration
Angular-velocity recorder Rolling velocity
Sideslip-angle recorder Sideslip measurement
Wheel-force recorder Aileron and elevator forces
Pedal-force recorder Rudder-pedal force
Control-surface position recorder Aileron, elevator, rudder, and stabilizer position
Wing-bending moment/sheer load and horizontal

Consolidated oscilloscope
tail sheer load
Common timing circuit Synchronize collected data
It was not until late November that the work had progressed far enough
for a flight to be made. NACA research pilot Howard C. “Tick” Lilly had
50
been selected to fly the second D-558-1. The first NACA flight was made on
November 25, 1947, for pilot familiarization, but it was cut short after instru-
mentation problems. Lilly made a second flight the following day, but it also
had to be aborted due to both instrumentation problems and the failure of the
landing gear to lock properly after retracting. With the onset of the winter rainy
season, Rogers Dry Lake was flooded. Engine modifications also had to be made.
Taken together, all these factors halted flight operations for several months.14
Project managers used the winter downtime to make another modification.
As mentioned previously, on the ground and at low altitudes over the desert the
Skystreak was eye-catching in its dark red finish, but optical and photographic
trackers had a hard time spotting it when it flew at higher altitudes, against the
dark blue Mojave sky. Walter C. Williams later wrote,
It was found…that very little photographic contrast was being

obtained between the red airplane and the relatively dark blue
sky conditions prevalent in this area[,] with the result that photo-
graphs could not be obtained to ranges greater than the order of
25 to 30 thousand yards. It was reasoned that the photographic
contrast could be increased by using the lightest color possible
against the darker sky. On this basis, the aircraft was test painted
white and both visibility and photographability were found to be
greatly increased. With proper filtering techniques, photographs
of the airplanes are now taken to greater than 60,000 yards and
are generally visible over their entire test flight range.15
Disaster: The NACA Loses Its First Pilot

The second D-558-1 had a history of landing gear problems. On November
26, 1947, on the second NACA flight, the landing gear door would not lock.
Between March 31 and April 7, 1948, the problem caused NACA flights
4 through 7 to be aborted. After several successful research flights, the gear
problem reoccurred on April 28 when the right landing gear did not retract,
forcing an abort of the 16th NACA flight.16
The problem reoccurred on May 3, 1948. At noon, after takeoff on the
18th NACA flight, the landing gear failed to lock in the full-up position. Lilly
landed the aircraft, and the next several hours were spent troubleshooting the
problem. When that was finished, the ground crew towed the D-558-1 to the
west end of the runway; Lilly started the engine and took off heading east. The
landing gear retracted normally, and the aircraft accelerated to a speed of about
51
Probing the Sky
250 mph at an altitude of 100 to 150 feet above the runway. Several ground
crewmembers continued watching the D-558-1 and saw “a large piece of white
material” separate from the fuselage.17
Smoke and flames began coming from the fuselage as the aircraft main-
tained level flight for several seconds. The D-558-1 began a left yaw and right
sideslip roll, which continued until the left wingtip, canopy, and vertical tail
struck the lakebed. The impact point was about 1,800 feet beyond where the
large piece of fuselage skin had landed. The aircraft bounced into the air and
broke up, hitting the ground about 400 feet farther along the flightpath and
scattering debris over a wide area. The fire was extinguished by Air Force per-
sonnel. Lilly was killed on impact, the first NACA research pilot killed in the
line of duty since the agency’s founding in 1915.
Crash investigators began by carefully searching for evidence along the takeoff
path, locating and tagging debris and marking it on a grid chart to establish the
sequence of events. The first pieces found were fragments of the engine compres-
sor case and blades, bits of the fuselage skin from the top of the engine section,
and paint chips, evidence of in-flight engine disintegration. These were located
less than 2 miles from where takeoff had begun. About 0.2 miles farther down
the flightpath was the 4-foot-square section of white fuselage skin spotted by
witnesses. Examination showed no evidence of fire. Instead, it was determined
that compressor case and blade debris had torn through the fuselage skin.
These fragments struck a vulnerable spot on the top of the fuselage. The
single set of rudder and elevator cables passed between this section of skin and
the compressor case. As the fuselage skin separated, the control cables were
“plucked,” causing a leftward movement of the rudder, a right sideslip, and
a left roll. The right rudder cable and the up elevator cable were severed. In
contrast, the left rudder cable, down elevator cable, and aileron cables survived
the disintegration of the compressor case but broke on ground impact.
Examination of the debris revealed that the fire had broken out after the
compressor case had broken apart. The left side of the rudder was badly burned,
but the right side showed no sign of fire, confirming that the rudder had moved
to the left before impact. The fixed vertical tail, however, had heavy fire damage
on both sides. The tail cone bore a burn pattern consistent with flames at an
angle approximately that of a full left rudder.
The investigation board examined the TG-180 engine and determined that
the rapidly spinning compressor rotor had suddenly stopped turning, causing
the compressor shaft to fail. Many of the compressor blades were torn out,
and the compressor casing had broken up with apparently explosive force. The
board learned that three other TG-180 engines had suffered similar failures in
earlier aircraft accidents, a pattern that pointed to fatigue failure of the blades as
a potential primary cause. The board noted other possible causes: blades damaged
52
by objects being sucked into the engine, or a failure in a steel spacer ring between
the 10th and 11th rotor stages.
The board also learned that three TG-180 engines assigned for use in the
D-558-1 aircraft had been rejected by NACA and Navy inspectors after numerous
large nicks had been found on the edges of the engines’ turbine blades. Inspection
of the compressor for such damage is practically impossible as a routine preflight
check, but if present, it could lead to blade failure. Additionally, investigators
learned that the steel spacer ring had been replaced in later-production TG-180
engines with an aluminum ring to reduce the potential for failure. This change
had not been made in the engines for the NACA aircraft.
The investigation board made a number of recommendations aimed at improv-
ing flight safety. The board wanted only newly manufactured TG-180 engines
fitted in the two remaining D-558-1 research aircraft before flights were resumed.
Board members also wanted improvements in preflight inspections, as well as in
specifications and manufacturing and inspection procedures for turbine blades.
Further, the recommendation was made that controls, fuel lines and pumps,
and electrical circuits located near the compressor section be protected and
that control cables should be armored, shielded, or duplicated to protect them
from damage.
The board was also critical of the D-558-1 canopy and cockpit design. Lilly had
not worn a crash helmet in the aircraft because the narrower high-speed canopy
did not leave sufficient room for one. He had earlier removed the shoulder straps
because he found them inconvenient. The board also discovered that the cockpit
space was restrictive and quite dark. The pilot had to lower his head to see some of
the instruments and could not see both outside and inside the cockpit at the same
time. The board urged that a study and redesign of the cockpit be undertaken.
Given that the idea of air-launching the XS-1 had been controversial among
Langley researchers, and in light of their repeated demands that the aircraft
be ground-launched, one of the board’s recommendations was ironic: “As a
safeguard for personnel and valuable research equipment, the fairly well proven
use of air launching should be given more consideration. The great difference
in wing loading between take-off and landing, as well as the magnitude of
the loading, and the desirability for maneuvering capability at the start of an
uncertain flight emphasize the value of air launching.”18
First Research Results

As the investigation of Lilly’s crash was underway, Williams wrote up the first
research results from the limited longitudinal-stability and control data taken
during the second D-558-1’s short operational life. The data, taken during the
53
Probing the Sky
Skystreak’s airspeed calibration flights, up to Mach 0.85, were published less

than 2 months after the crash.19 The flights involved making level flights at a
30,000-foot-pressure altitude and increasing the aircraft’s speed from Mach
0.55 to Mach 0.85. During the flight, changes in elevator position and force
required to trim the aircraft were recorded. These were used to determine the
aircraft’s longitudinal stability and any trim changes caused by the effects of
transonic speed, such as the formation and movement of shock waves on the
wings and stabilizers. The flights were done at stabilizer incidence angle set-
tings of 1.95° and 2.32°.
Williams wrote, “The results of measurements of the elevator angle and
force required for trim at Mach numbers up to 0.85 show that below a Mach
number of 0.80 the D-558-1 airplane possesses positive static longitudinal
stability. Above a Mach number of 0.82, there is a nose-down trim change.”20
Charting of the elevator force, specifically how many pounds of force the pilot
had to use to counter any trim changes, made clear what had occurred. Between
Mach 0.6 and 0.75, only minor pressure was needed to keep the D-558-1 level.
Once the aircraft reached Mach 0.8, however, the amount of pressure required
sharply increased. At Mach 0.85, the pilot had to pull back on the control yoke
with nearly 10 pounds of force to keep the airplane level.
As the NACA realized, the report constituted a “quick look,” rather than
a detailed, incisive examination. The airspeed calibration data had not been
completely evaluated, but the error margin was estimated to be 1 percent or
less. This estimate was based on comparisons with the airspeed calibrations
done with the XS-1, also at Mach 0.85. There were also instrumentation issues;
no elevator-position data were obtained above Mach 0.8 for a stabilizer setting
of 2.32°, as the recorder’s film had run out. Finally, the stabilizer incidence
angles were close together, making it difficult to determine the relative elevator
effectiveness over various Mach numbers.
The results were considered valid, however, as the elevator force and angle
data showed the aircraft had positive longitudinal stability up to a Mach
number of about 0.80 with the control wheel both fixed and free. Above a speed
of about Mach 0.82, the data showed that a trim change occurred, the first
indication of compressibility effects in level flight. A similar trim change also
appeared in the data derived with the second XS-1, which also had a 10-percent
wing. NACA researchers subsequently issued further reports likewise based on
the initial flights of the second Skystreak, following these reports with a more
extensive study by Williams examining the aircraft’s stability characteristics
during sideslips, dated April 18, 1949.21
The technique used by the pilot to measure the static directional stability
of the D-558-1 was to slowly deflect the ailerons, creating gradually increasing
sideslips. At the same time, he added enough rudder and elevator to maintain
54
level flight. The Skystreak was flying level at a constant altitude, but with its
nose angled to one side. Two series of tests were flown. One was made at a
10,000-foot-pressure altitude and at indicated airspeeds between Mach 0.50
and 0.80. The second series was made at a 30,000-foot-pressure altitude, at
Mach 0.50 and 0.84. (The indicated airspeed was above that where the nose-
down trim change occurred.)
Once the data were collected, the rudder, aileron, and elevator positions
and the forces and angles of bank were plotted on graphs as functions of
sideslip angle. Bank angles were obtained from measurements of transverse
acceleration. Variations of rudder position and force with the sideslip angle
gave a measure of the aircraft’s static directional stability, with both fixed and
free control. The dihedral effect, which is the rolling moment of an aircraft
caused by the spanwise inclination of the wings, was measured with both fixed
and free controls. This was illustrated in the variations of aileron position and
force with the sideslip angle. The pitching moment due to the sideslip was
indicated by the variation of the elevator position and force with sideslip angle.
The variation of angle of bank with sideslip angle gave a measure of crosswind
force characteristics. The report concluded:
• “The apparent directional stability of the D-558-1 was high
throughout the speed range covered, but was greater at low altitudes
than at high altitudes at any given Mach number. There was also an
increase in directional stability with an increase in Mach number.”
• “The increase in directional stability at lower altitudes was probably
due to a decrease in rudder efficiency, caused by distortions of the
vertical tail and fuselage by higher dynamic pressure.”
• “The dihedral effect was positive but low over the speed range tested.”
• “There was little to no change in pitching moment with sideslip and
the cross-wind force was positive.”22
Several issues related to data collection had been neglected. No measure-
ments of rudder forces were recorded at 10,000 feet. There were also dis-
continuities in the variation of both the aileron force and the position with
sideslip angles near zero. Williams concluded that the problems with the force
measurements were due to friction in the control system. He believed the
position errors were caused by play in the linkage between the aileron and the
point of measurement. Despite this, Williams believed the slope of the curves
should provide a good measure of dihedral effect.23 These two reports contained
preliminary data drawn from a few early flights. A more detailed report was
issued on April 22, 1949, and contained information on the D-558-1’s high-
speed characteristics, up to Mach 0.89, that was not included in Williams’s
two early reports.24
55
Probing the Sky
The stability measurements used in the report were primarily derived

from two of Lilly’s high-speed flights. The two flights, made at an altitude
of 40,000 feet, reached Mach 0.89 and used stabilizer incidence angles of
2.3° and 1.4°. The results of the different settings were significant. The 2.3°
stabilizer-incidence angle resulted in the aircraft’s becoming increasingly
nose-heavy as the Mach number increased above Mach 0.80. During the
initial phase of the recovery maneuver, Lilly had to pull back hard on the
control yoke to decrease the Mach number. As the Mach number decreased
over a 10-second period, nose heaviness was also reduced, and Lilly had to
reduce the pull force to avoid too great an acceleration.
The Skystreak’s behavior was very different at a stabilizer incidence angle
of 1.4°. Above Mach 0.83, the Skystreak became increasingly tail-heavy.
During the 24-second recovery phase, Lilly simply reduced the push force;
the airplane’s speed dropped from Mach 0.88 to 0.834, and he completed a
normal recovery. When the sideslip angle and the control forces and positions
were plotted as a function of Mach number, the differences were glaringly
apparent. Depending on the stabilizer setting, a pull- or push-force of 30
pounds was required to correct trim.
During both high-speed flights, Lilly reported that above Mach 0.84 the
right wing became very heavy, and he had to make increased left aileron
inputs to correct it. He added that the wing heaviness was not continuous,
which made the aircraft’s lateral stability feel uncertain at its highest speed.
Lilly found it difficult to determine the amount of lateral control needed
for trim. His control movements resulted in lateral oscillations. The report’s
author believed some of this problem to be due to aileron fraction.
Some stability and control data were also collected during several incre-
mentally increasing turns made at an altitude of 30,000 feet, at Mach num-
bers between 0.50 and 0.80. A single turn was made at 10,000 feet and
Mach 0.71. The data from the turns showed positive longitudinal stability
throughout the speed range. The lowest value was recorded at Mach 0.675,
and beyond this point stability increased with increasing Mach number.
These results were also consistent with those from the XS-1 flights.
The single test at 10,000 feet indicated that the apparent stability was
higher at this altitude than at 30,000 feet. The report noted, “Some of this
difference can be accounted for by the effects of altitude but it is also pos-
sible that, because of the higher dynamic pressure at the lower altitudes, the
apparent stability is altered by distortion effects.”25 This increase in apparent
stability at lower altitude was also noted in Williams’s second report, also
issued in April 1949.
Buffeting occurred on both of the high-speed runs, beginning at a speed
of about Mach 0.85. The D-558-1’s buffet boundary was determined not
56
only during speed runs, but also in straight stalls and turns. The normal-force
coefficients necessary to cause buffeting were plotted as functions of Mach
number, to define the combination at which buffeting began.
Data from the XS-1 flights confirmed the Skystreak results. The XS-1 was
also flown with the same two stabilizer incidence angles as the D-558-1, and
in both cases the aircraft became nose-heavy at higher speeds. The XS-1 also
showed similar wing-heaviness behavior. This similarity was to be expected,
as both had 10-percent wings and 65-110 wing sections.26
Modifications and Flights Resume

With the death of Lilly and the destruction of the second D-558-1, the two
remaining aircraft were grounded until the accident investigation was com-
pleted. During this period, the D-558-1 #1 (the Douglas aircraft) underwent
modifications to fix the problems identified by the accident board and was
returned to Muroc.
Gene May resumed the Douglas flight-test program, which entered a dan-
gerous phase during which stability and control data for speeds of up to the
aircraft’s maximum Mach number were to be collected. Because of the D-558-
1’s limited performance capability, May had to make risky dive flights, similar
to those done in the P-38s and other aircraft, rather than level speed runs, as
were done with the X-1.
The test program involved 10 dive tests, 5 with wingtip fuel tanks and 5
without. A dive would begin at 40,000 feet, with a pullout following at around
30,000 feet. On September 29, 1948, May exceeded Mach 1 in a 35° dive
in D-558-1 #1, constituting the only time a Skystreak was to fly supersonic.
Stability and control deteriorated badly at Mach 0.84, with the aircraft oscil-
lating laterally. As speed increased, the left wing became heavy. Longitudinal
stability decreased above Mach 0.94, with the aircraft “tucking under.” This
was similar to what had been experienced with piston-powered aircraft during
transonic dive flights.
The dive tests with the tip tanks followed in early November 1948. With
the stock tank configuration, drag was reduced and range increased. However,
top speed was also reduced, and the takeoff roll was longer. The tanks were
later fitted with endplates in a study of wing airflow. On November 4, 1948,
Gene May made another dive test to Mach 0.945. The flight data showed the
endplates had stopped spanwise flow separation over the wings, which dis-
rupted lift. Despite this, the endplates were not used again on the D-558-1.
These endplates, in some respects, anticipated the development of the vortex-
reducing winglet two decades later.27
57
Probing the Sky
The third D-558-1, which was assigned to the NACA, had been trucked
to Muroc on November 4, 1947, but, with attention focused on the first and
second aircraft, it had made only four flights by early 1948. Following Lilly’s
crash, the third D-558-1 was disassembled and trucked back to the Douglas
plant for modifications. The work involved adding duplicate control cables and
¼-inch stainless steel armor to protect the emergency fuel pump and fuel lines.
Engineers also tested the vulnerability of the high-pressure fuel hoses to shrap-
nel from an exploding engine. This was done by firing .22-caliber rifle bullets
into the standard fuel hose. The bullets easily punctured the standard hose. As a
result, these were replaced by wire-wound fuel hoses, which were better able to
withstand high-velocity impacts.
The final change Douglas made was to repaint the third D-558-1 in an
overall white finish. This posed a complication with the control surfaces. A
letter from R.B. Cox, of Douglas, to Walter C. Williams, dated October 11,
1948, noted:
When the aircraft was sprayed with the white undercoat…the

control surfaces were painted. I had seen a schedule of work,
issued by the chief engineer for this division, which called for a
white paint coat on the entire ship, except for the control sur-
faces. I called this to the attention of the aircraft project engineer
and investigation proved the shop order was in conflict with the
original order. Further investigation showed the reason for not
painting the surfaces involved a problem of weight and balance.
With the original red color the surfaces were just within the allow-
able margins and the addition of the white undercoat threw them
over the limits. The solution was to remove the white, rubbing
down into the red just slightly and then fog a light mist coat of
red on the units.28
The modified and repainted third D-558-1 was returned to Muroc Army
Air Field by Douglas in early November 1948. After the aircraft was reas-
sembled, Douglas pilot Gene May completed a demonstration flight in the
modified aircraft. He accomplished this on January 3, 1949, in a flight that
included a 6.8-g pullout, left and right maximum sideslips at 580 mph, and a
low-level pass at 605 mph indicated airspeed.
The flight was apparently without mishap. However, an NACA safety rep-
resentative making a postflight inspection discovered a damaged brass safety
wire that had passed through the jet engine. Pulling the engine and conduct-
ing an inspection would have delayed turning the aircraft over to the NACA.
Instead, Douglas simply installed a new TG-180 engine. The NACA took
58
formal delivery of the third D-558-1 on January 22, 1949. The ground crew
then began preparing it for research flights, a process that lasted until April.
As this was underway, the Douglas flight
tests with the first Skystreak were con-
cluded. A total of 101 flights had been
completed. Douglas transferred the air-
craft to the NACA, which put it in “dead
storage” for use as spare parts. It never
flew again. The third D-558-1 was the
only one of the three Skystreaks still
in use.29
Preparing the NACA Skystreak for
research flights took much of 1949–50.
Robert A. Champine was the first to
fly the aircraft after it was turned over
to the NACA. He had arrived at Muroc
from Langley in October 1948. He made
two D-558-1 pilot proficiency flights in
April 1949. After Champine’s two flights, NACA research pilot Robert Champine
the aircraft was grounded for an engine climbs out of the D-558-1 after a flight.
Champine became an NACA research
change and remained grounded during pilot in December 1947, and he retired in
the spring and summer of 1949. It would 1979. On December 2, 1948, he became
not be ready to fly until August. the sixth person to fly supersonic. (NASA)
NACA Research Flights Continue

Once the Skystreak was restored to flight status, the initial research focus was
on its handling qualities. Piloting duties were split between Bob Champine
and John H. Griffith, an NACA research pilot assigned to Muroc in August of
1949. Between August and September of 1949, the two made seven flights.30
One issue explored during these flights was the effectiveness of the D-558-1’s
ailerons. NACA flight 8 was made by Champine on August 31 and involved
22 aileron rolls, of which 4 were at Mach 0.86. On NACA flight 9, made by
Griffith on September 28, 16 aileron rolls were made, 4 of them above Mach
0.875. The rolls were abrupt, with the rudder held in a fixed position, and
made at speeds between Mach 0.6 and 0.89.
The amount of aileron deflection was between one-eighth and one-half
the total available deflection of ±15°. Most of the rolls were made at pressure
altitudes of about 35,000 feet, although some were as low as 15,000 feet. The
rolls were made in both directions. To make the research pilot’s task easier, a
59
Probing the Sky
mechanical stop was placed in the cockpit to allow the pilot to hold a constant
aileron deflection until a constant rolling velocity was established.
As the Skystreak rolled, a yawing oscillation occurred, much like that of
a wobbling top as it spins. This yawing was most apparent when the rolling
velocity was increasing. Once the rolling velocity reached maximum value,
however, the yawing damped out. Several of the roll maneuvers were studied
to determine what effects the yawing had on maximum rolling velocity. No
significant change was noted.
Complicating the tests was the fact that the ailerons were slightly warped.
With no load on them, this amounted to as much as a 1° difference between
the aileron-cord line and the wing-cord line at different points along the span
of each aileron. Ideally, there should have been a continuation of the airfoil.
The results of the flights were described in a May 1950 NACA research
memorandum. Altitude had no effect on the ailerons’ ability to control the
aircraft. None of the test rolls were executed with more than half the avail-
able aileron deflection. The pilot believed the rolling velocities achieved with
one-half deflection were enough to meet the maximum requirements for
either test or military operations with the aircraft. Indeed, a full aileron deflec-
tion at Mach 0.85 at 35,000 feet would result in a complete revolution in
0.95 seconds.
The aileron forces required at a given deflection increased with indicated
airspeed. The indicated value of force at maximum rolling velocity depended
on the time required to reach maximum rolling velocity and the fraction in
the aileron control system (estimated at ±5 pounds). When the force data were
graphed, an approximately straight line resulted. This indicated that the total
hinge-moment coefficient for a given aileron deflection was independent of
the Mach number, at least for the speed range of the tests.
From the earlier stability and control research flights, as well as the aileron
tests, a clearer understanding of the Skystreak’s lateral trim and handling
characteristics at high Mach numbers was now available. The report noted
the following:
As the speed of the airplane is increased a right-wing heaviness

becomes apparent to the pilot at about the same time as general
buffeting of the airplane is encountered. As the speed is further
increased the wing heaviness increases, a change in aileron trim
force of about 7 pounds and a corresponding change in total
aileron deflection of about one-half degree being required to
trim…at a Mach number of about 0.88. The trim force and
deflection for the wing-heaviness example quoted are typical,
although in a few cases trim changes could not be detected on the
60
recording instruments and were not noticeable to the pilot. The

lateral unsteadiness of the airplane at high speeds is evident on
the time history from the rapid variations of force and deflection
applied by the pilot in attempting to trim. It was also evident…
that the airplane has a short-period rolling-yawing oscillation of
small amplitude. In addition to the wing heaviness, pilots have
reported an intermittent “wing dropping” which occurs above a
Mach number of about 0.86. This sudden rolling of the airplane
occurs above a Mach number of about 0.86. This sudden roll-
ing of the airplane occurs in either direction and appears to be
associated with the general lateral unsteadiness of the airplane at
high Mach numbers.31
As earlier reports had noted, the wing heaviness of the X-1 and D-558-1
were similar, but more details were now apparent. On the X-1, the heaviness
occurred at about Mach 0.85 and appeared to be related to an abrupt reduction
in aileron effectiveness. The stability and control data on the Skystreak showed
that aileron effectiveness did not decrease at speeds of up to Mach 0.89. Aileron
effectiveness had not been investigated for small deflections, however. The
report also noted, “A possible contributing cause of both the wing heaviness
and wing dropping is probably asymmetric location and movement of shock
waves on the wing resulting from construction asymmetry.”32
During September and October of 1949, two 60-cell manometers were
installed to record right-wing surface-pressure-differential measurements. Six
rows of orifices were cut into the upper and lower surfaces of the right wing,
running along the chord from front to back. Row 1 was only 6 inches from
the fuselage; row 6 was close to the tip. The pressure differential between the
upper and lower wing surfaces was measured from rows 1, 2, 3, 4, and 6. Row 5
measured individual surface pressures relative to the instrument compartment,
and the instrument compartment pressure was measured relative to the static
pressure, which was corrected to the free-stream static pressure by using radar-
tracking data. This process calibrated the pressure measurements, eliminating
errors and ensuring accuracy.
The “plumbing” installed within the wing was extensive. The flush-type
orifices in the wing skin were connected to the instrument compartment
with ⅛-inch-inside-diameter aluminum tubing, which was connected to the
manometer cells with 3/16-inch rubber tubing. The length of the aluminum
tubing ranged from 6 feet at the wing-root stations to about 14 feet at row 6,
near the wingtip. About 4 feet of rubber tubing was used on each line.
Given the length of tubing, the maneuvers that were planned, and the need
for precision data, the effects of instrument lag had to be considered. Ground
61
Probing the Sky
testing indicated that any lag attributable to tubing length would be negligible.
The lag in the airspeed recording system was calculated using established pro-
cedures and corrections made in the data. Considerable instrument lag in the
airspeed recording system occurred during speed runs and the windup turns.
Corrections were calculated and added to Mach-number and dynamic-pressure
measurements. For the 1-g-stall measurements, lag was negligible because the
pilot used a separate airspeed system.33
With the installation of the manometers now complete, the aircraft was
returned to flight status on October 28, 1949, with Griffith making the first
pressure-distribution flight. Two more pressure-distribution flights, one each
by Griffith and Champine, were made by late November, before the year’s
activities came to a close.
More than 2 months passed before D-558-1 flights resumed. Technical
problems continued, grounding the aircraft in February 1950 following an
engine malfunction. Repairs took significant time, and it was not until April
5, 1950, that pressure-distribution flights resumed. Griffith flew the third
D-558-1 as Champine had returned in 1950 to Langley, where he continued as
a research pilot. Following a flight on April 11, 1950, the aircraft experienced
hydraulic problems after landing.34
Pressure-distribution research involved a large number of flights and a wide
range of test maneuvers and procedures. Early research activities involved a
1-g stall at a subcritical Mach number and 15,000 feet, a speed run to Mach
0.90, and a windup turn at Mach 0.86. The 1-g stall was executed by gradu-
ally slowing the aircraft until it stalled. Other maneuvers were more complex.
The speed run started with the Skystreak at 37,000 feet and Mach 0.70.
The pilot dove to 33,000 feet and a Mach number of 0.90 and then began a
gradual left turn, which he tightened until maximum allowable buffeting was
reached. During the turn, airspeed dropped to around Mach 0.86. Once at a
near-constant Mach number and an increasing normal-force coefficient, the
pilot collected several data points. The ailerons were held near neutral during
the maneuvers, and rolling velocities due to lateral oscillations were low.35
The third D-558-1 was also used to investigate a simple solution for reduc-
ing or delaying a range of effects resulting from compressibility, including buf-
feting, lateral instability, changes in trim, and reduction in control efficiency.
Engineers attached “vortex generators” to the upper wing surfaces. Despite
their impressive name, these were small airfoils with an NACA 0012 section
and a chord of 0.5 inches, and they were positioned at 2-inch intervals. The
generators were alternately tilted toward and away from the fuselage. They did
not change the wings’ section profile and were added following Griffith’s April
11, 1950, flight. The work took less than a month.
62
The first vortex generator research flight was made on May 5 as a continu-
ation of pressure-distribution studies. The vortex generators on the first flight
extended only from the wing’s mid-flap to the mid-aileron. The flight entailed
takeoff and landing followed by a flight to altitude. The goal was to collect data
on low-speed handling characteristics with the vortex generators installed. The
initial configuration was modified for the second flight, with the vortex genera-
tors extending from the mid-flap section out to the wingtip. The final design,
for the third flight, stretched the full width of the wing.
The portions of the research flights made at altitude involved low-speed
stalls, pull-ups in the buffet regions at speeds of up to Mach 0.89, and abrupt
aileron rolls above Mach 0.7. Once the flights were completed, the vortex
generators were removed and the flight conditions were repeated to collect
baseline measurements. The sixth and final vortex generator flight was made
on June 13, during which Griffith reached a speed of Mach 0.98. This marked
the end of the pressure-distribution research flights.
NACA researchers analyzed the data and found that the vortex generators
produced a number of effects. At speeds above Mach 0.85, the areas of flow
separation were reduced. At Mach numbers greater than 0.85, the flow separa-
tion and the forward movement of the shock wave on the wing’s upper surface
were reduced, though no change in the small-amplitude lateral oscillations
could be identified. The buffet boundary and wing drop were both delayed by
about Mach 0.05. The pilot reported that the intensity of the buffeting was
“appreciably reduced” below the stall.36 No detrimental effects from the vortex
generators were found on the Skystreak’s longitudinal and lateral control, at
least for the conditions of the tests. The only negative result was an increase
in drag.
The NACA vortex generator test series had a major impact on aeronautics
and aircraft design. Boeing was the first to use vortex generators on production
airplanes. Rows of the little metal tabs soon appeared on the wings of B-47 and
B-52 bombers, KC-135 tankers, and thousands of airliners.37 Other companies
followed suit as well.
The next research effort undertaken was the measurement of the Skystreak’s
buffet boundary, which is the combination of speed and lift coefficients at
which an aircraft experiences irregular shaking or oscillation due to tur-
bulent flow or flow separation. The Skystreak was grounded for modifica-
tion during the summer and early fall of 1950. A high-speed photographic
manometer was added to measure wing-pressure distribution over a spanwise
station, and a downwash vane for airflow measurements was added to deter-
mine the wing’s contribution to buffeting. A nose boom was also added for
angle-of-attack measurements.
63
Probing the Sky
While the aircraft was grounded, the engineers replaced its stabilizer with
the first D-558-1’s instrumented stabilizer. The engineers had noticed a trim
change during pressure-distribution flights that was due to a loss of elevator
effectiveness. They were uncertain whether this was a result of changes in
the actual pressure distribution or of physical distortion of the stabilizer and
elevator. The solution was to replace the stabilizer. They completed the work in
mid-October 1950, and D-558-1 flights resumed on October 26 with a check
flight by Griffith, clearing the way for buffeting, tail-load, and longitudinal-
stability flights, which would be made over the course of the coming year.38
The check flight would be Griffith’s last in the D-558-1. He left the NACA in
the fall of 1950 to become a senior test pilot on the troubled Chance Vought
F7U Cutlass flight test program.
His replacement was A. Scott Crossfield, a World War II Navy fighter
pilot and gunnery instructor. Crossfield was a thorough-going aeronautical
professional, both a consummate engineer and consummate pilot. He had
attended the University of Washington, earning bachelor’s and master’s degrees
in engineering during 1949 and 1950. Soon after, he joined the NACA as an
aeronautical research pilot. His arrival reflected the new demands on test pilots
wrought by transonic and supersonic flight. It was no longer enough to be a
hot-shot fighter pilot with plenty of stick-and-rudder time. Both the new and
the old test and research pilots were accomplished aviators, but for the new
breed such as Crossfield, engineering knowledge was at least as important as
flying skill.39
Crossfield made his first D-558-1 flight on November 29, 1950. It was
both a pilot checkout flight and the beginning of the buffet, tail-load, and
longitudinal-stability research project. He made a total of five flights before
year’s end. The effort continued into the new year with Crossfield making
another four research flights in January 1951.
Another new pilot now joined the D-558-1 project. Walter P. Jones had
arrived at the High-Speed Flight Research Station in September 1950 with
both undergraduate and graduate degrees in aeronautical engineering from
Purdue University. Jones made his first flight in the aircraft on February 13,
1951—a pilot check flight that also included the collection of some buffet,
tail-load, and longitudinal-stability data. Jones’s second Skystreak flight, on
February 20, highlighted the risks of research flying. His oxygen regulator was
faulty, and Jones began to suffer from anoxia. He nevertheless recognized his
situation, aborted the flight, and landed safely.
These were not the only dangers Crossfield and Jones faced. Engineers
discovered that the elevator was twisting by as much as 2° during pull-ups at
about Mach 0.80 and was also experiencing undesirable vibrations. Flight data
indicated that these vibrations increased in amplitude in direct proportion to
64
the increase in Mach number and occurred in all high-lift conditions over the
entire Mach number range, even if the airplane was in a stall. Elevator force
and position indicators did not reflect the vibrations until they had reached an
undesirable level. Though a precautionary x-ray examination of the elevators
showed no sign of fatigue cracks, concerns persisted.
NACA engineers contacted Douglas, as the latter had recorded similar prob-
lems during its flight program with the first Skystreak. During a Mach 0.94
dive test, the first D-558-1 had experienced vibrations so severe that a “never-
exceed” speed of Mach 0.92 had been established for the aircraft. Douglas
engineers said they had attempted to reduce the vibrations by changing the
position of the outboard elevators’ balance weights. The Douglas contractor
flight program was ending, and the engineers admitted that they had not
evaluated the change. More serious was the fact that the Douglas engineers
had never informed NACA engineers of the speed limitation.
The team at Douglas analyzed data from the NACA flights and discovered
that the elevator vibrations had reached nearly 70 percent of the maximum
stress limits of the design and could eventually cause a fatigue failure. Between
the Douglas and NACA tests, more than a hundred flights had been made
with the elevators from the first D-558-1, which had been fitted subsequently
to the third D-558-1, and it was questionable how many more flights could
be made without a failure. As a precaution, the first D-558-1’s elevators were
removed and the third D-558-1’s elevators were reinstalled, since these had
been used on only 22 flights and had a much longer service life. The switch
took a significant amount of time, and it was not until late April 1951 that
the work was completed.40
Jones made the first flight in the reequipped D-558-1 on May 2, 1951.
Between May and late June, Crossfield and Jones made a total of six buffet,
tail-load, and longitudinal-stability flights at speeds between Mach 0.835 and
Mach 0.86. Once more, there was a change in personnel. During the remainder
of the NACA Skystreak project, Crossfield made only three more flights, with
Jones making a single additional D-558-1 flight. Subsequent research activities
were undertaken once several new pilots joined the effort.
Joseph A. Walker was the first of these. An Army Air Forces pilot in World
War II, he had flown reconnaissance missions in P-38 Lightnings over Austria,
the Black Sea, and southern France. At the war’s end, and already holding
a bachelor’s degree in physics from Washington and Jefferson College, he
joined the NACA’s Lewis Flight Propulsion Laboratory in 1945 as a physi-
cist. He transferred to the High-Speed Flight Research Station in 1951 as a
research pilot.41
Walker’s first pilot checkout flight came on June 28, 1951, during which
he reached a speed of Mach 0.82. He made a second checkout flight on July 3,
65
Probing the Sky
The third D-558-1 on the lakebed, being prepared for a flight. A ground crewman is adjusting
the pilot’s parachute straps, which the pilot cannot see. The aircraft is also positioned close
to the lakebed shoreline to allow the maximum distance for a takeoff run. When the NACA
contingent first arrived at Muroc Army Air Field in late 1946, they found the facilities of a war-
time training base very different from those at Langley. Housing was subpar, the facilities were
limited, and the climate was usually either too hot or too cold to work comfortably. Some NACA
personnel left, but others adapted. (NASA)
which also included tests of buffeting and tail loads. Walker made a total
of 14 D-558-1 flights between June 28 and October 18, 1951. Although
the D-558-1 flights represented a significant part of the High-Speed Flight
Research Station’s activities, they were never routine.
On Walker’s third flight in the D-558-1, on July 17, he had to cut the
mission short due to low fuel. Two flights later, on July 26, weather forced an
inflight abort. Clouds prevented tests at altitudes above 15,000 feet. The next
two flights were successful, but on August 10 the airplane suffered a fuel leak
due to a malfunctioning vent valve, cutting short still another mission. After
a successful flight on August 20, on which Walker reached a true airspeed of
Mach 0.9, problems struck again. A hydraulic failure caused his August 22
flight to be aborted. Walker’s final D-558-1 flight, on October 18, was for
the collection of lateral-stability and landing data. Walker then moved on to
other projects.42
Stanley P. Butchart took over research duties with the D-558-1. Like
Crossfield and Walker, he was a World War II combat pilot, having flown TBM
Avenger torpedo planes in the Pacific.43 Butchart entered the University of
Washington after the war, earning bachelor’s degrees in aeronautical engineering
66
and mechanical engineering. While there, he and Crossfield served in the same
Naval Reserve squadron before both joined the NACA in May 1951. Butchart
made his first D-558-1 flight, for pilot checkout, on October 19, 1951. This
was followed on November 9 by his first research flight, which was the last
flight of the buffeted, tail-load, lateral-stability, and landing-study project.
Skystreak flights made during this period resulted in the publication of
several NACA research memorandums between early 1951 and early 1952.
Additional tables of wing-pressure measurements were released in January
1951. Like the data issued in December 1950, the new data were collected in
windup turns. Unlike in the earlier tests, however, the new information was
not limited to collection at a single speed of Mach 0.86. Rather, it covered
Mach 0.67, 0.74, 0.78, and 0.82 at an altitude of 35,000 feet. Both reports
consisted of tabulated measurements but lacked detailed analysis, in the interest
of making flight-test data available to designers as quickly as possible.44
During the flights made with the second D-558-1, measurements had been
made of the effects of different stabilizer incidents on the aircraft’s longitudinal-
stability and control characteristics. These preliminary results showed that
minor changes in the stabilizer incidents caused major changes in longitudinal
trim characteristics. Once flights resumed with the third Skystreak, the issue of
the effects of the stabilizer incidents was revisited in a more thorough investiga-
tion. The tests were made using shallow dives, pullouts, and windup turns at
altitudes ranging from 37,000 to 27,000 feet, at Mach numbers between 0.60
and 0.89. The stabilizer incidences used during the research flights were 1.6°,
2.2°, 2.6°, 2.7°, 2.9°, and 3.3°. These tests were considerably more complete
than those made during initial efforts with the first D-558-1.
In newspaper accounts of test pilots’ experiences, the focus is more often on
the drama and danger of the flights than on the test results. In NACA research
memorandums, the text is dry and to the point. The results of the flights are
depicted in charts and graphs and are often unclear to a lay reader. But some-
times, even dry text and charts make clear what has transpired. Regarding a dive
to about Mach 0.89, “The data were obtained in a dive from about 37,000 feet
with a stabilizer setting of 3.3°. At about 48 seconds, as the pilot attempted to
pull out, the elevator angle and stick forces necessary to execute the maneuvers
became excessive and the stabilizer had to be used to recover from the dive.
The time history for this run was not extended beyond 48 seconds because the
subsequent data were not satisfactory for analysis…. [I]t is evident from the
figures that large changes in longitudinal trim occur at Mach numbers above
about Mach 0.84.”45
The pull-stick force needed to recover from the dive with the 3.3° stabilizer
setting was about 80 pounds. More important, the recovery was only pos-
sible by also using the adjustable stabilizer. The conventional fixed horizontal
67
Probing the Sky
The third D-558-1 parked on the South Base ramp with three ground personnel. The Skystreak was
one of two concepts for how best to gain data on transonic flight. It represented a traditional design,
had straight wings, and was jet-powered, with a top speed just below Mach 1. It was funded in part
by the Navy and built by Douglas to NACA specifications. In contrast, the Army Air Forces XS-1 was
rocket-powered, with a top speed well above Mach 1. (NASA)
stabilizer and movable elevator, used since the early days of flight, were defini-
tively shown to be obsolete for transonic and supersonic flight. The two-part
design of the X-1 and D-558 was the origin of the all-moving tail fitted to the
later model F-86 and subsequent high-performance aircraft and is still in use
today. The report also noted, “The results indicate that large and rapid changes
in elevator deflection and force were required for balance at Mach numbers
above 0.84. At Mach numbers above about 0.84, a sharp decrease in the rela-
tive elevator-stabilizer effectiveness was shown and analysis indicated that a
major part of the observed trim changes was explained by this decrease…. The
increase in apparent stick-fixed stability parameter was attributed to a decrease
of relative elevator effectiveness together with an increase of the stability of the
airplane by a factor of 4 between Mach numbers of 0.75 and 0.89.”46
Though the bulk of research undertaken with the D-558-1 was similar to
that done with conventional aircraft to understand stability and control issues,
another study using the Skystreak looked to the future, toward the dawning
computer revolution. By the early 1950s, engineers realized that automatically
stabilizing dynamic systems were necessary for future supersonic aircraft. To
design such a system, however, engineers needed aircraft frequency-response data
68
at high subsonic, transonic, and supersonic speeds. Acquiring these data involved
the transfer functions of both the aircraft and the control systems.
The procedure used to collect the data was to make several stick-fixed 2°
elevator pulses, each lasting 0.5 to 1.0 second. The aircraft was in stabilized 1-g
flight at a speed between Mach 0.52 and 0.90 and altitudes of 30,000 to 37,000
feet. These pulses produced an initial aircraft oscillation of approximately ±½
g to ±1 g and a pitching velocity of ±0.1 radians per second. As with other
tests, precision was central to ensuring that the data were valid. A restricting
device was attached to the elevator control that returned it to approximately
the original position following the pulses and also maintained the fixed elevator
condition as the oscillation subsided. The other controls—ailerons, rudder, and
stabilizers—were fixed during the maneuvers. The stabilizer incident was fixed
throughout the tests at a nose-down angle of –2°. Another factor affecting the
data was the twist of the elevator, which amounted to about 0.4°. To produce
an average value, the elevator position was recorded at four positions along the
external span of the fuselage. The four control positions were averaged, and
this average was used as the input function.
When the flights were completed, Fourier transform (a mathematical pro-
cess) was applied to the input and output functions, establishing the longitudi-
nal frequency response of the aircraft as a function of Mach number. The report
noted, “A comparison of the response data estimated from wind tunnel data
with the experimental results showed good agreement. It was found that the
maximum response amplitude was a minimum at a Mach number of 0.88. At
lower Mach numbers (0.52 to 0.66) the effects of lift coefficient on frequency
response are indicated.”47
1952–1953: The Skystreak’s Twilight Years

Much had changed at the High-Speed Flight Research Station since the D-558-1
had been delivered in the spring of 1947. New X-planes had been built, result-
ing in a heavy workload for the limited number of engineers, technicians, and
computers at the remote site. The Skystreak was also showing its age. Following
Butchart’s second flight, it was grounded for maintenance work that involved
repairing major fuel leaks and correcting engine-ignition problems. Work was
completed in January 1952, but winter rains flooded Rogers Dry Lake, prevent-
ing flight operations. The research plan for the D-558-1 in early 1952 called for
an investigation of vertical-tail loads. To support the tail-loads study, strain-gauge
instrumentation was installed. Before this research activity could begin, however,
or the strain gauges could even be calibrated, a study of lateral stability and aileron-
roll effectiveness was undertaken with the D-558-1.
69
Probing the Sky
The D-558-1 resumed flight on June 25, 1952, with Crossfield as pilot,
beginning the lateral-stability and aileron-effectiveness research flights.
Crossfield made two flights before turning the project over to Butchart. The
pilots made abrupt aileron rolls from Mach 0.4 to the Skystreak’s limiting
Mach number at 10,000 feet, 25,000 feet, and 35,000 feet. Results showed
that aileron effectiveness dropped rapidly above Mach 0.88. Butchart made a
total of five flights between July 17 and August 12, 1952. He later recalled the
cramped Skystreak’s cockpit. During takeoff on July 17, 1952, he reached for
the landing gear handle but could not squeeze his hand between the control
yoke column and the emergency oxygen bailout bottle strapped to his left leg.
It took Butchart three tries to get the gear retracted.48
Horizontal stabilizer issues continued to plague the program. During a
postflight inspection, the horizontal stabilizer mountings were found to be
loose. High-Speed Flight Research Station engineers judged this to be due
to wear over a long period of time, rather than to one-time damage resulting
from a severe maneuver. The horizontal stabilizers were originally fitted to
the first D-558-1, and about 100 flights already had been made with them.
They were then sent to Langley to be instrumented. Once this was completed,
the stabilizers were returned to the High-Speed Flight Research Station and
reinstalled on the third Skystreak. By the time the loose mountings were dis-
covered, 41 more flights had been made with the stabilizers. The risk posed by
the loose mountings was considerable. The horizontal stabilizer was known to
vibrate during buffeting, and during a Mach 0.90 dive, the vibration became
quite pronounced. High-Speed Flight Research Station engineers decided
to repair the vertical-load links and the pins that attached the links to the
horizontal stabilizer.
Repairing the horizontal stabilizers first required removing their skins. Once
this was done, the strain gauges were inspected and found to need exten-
sive repairs; they would have to undergo a complete recalibration. The strain
gauges were critical for measurements of vertical-tail load as the stabilizers were
mounted directly on the tail.
The engineers estimated that repairs to the horizontal stabilizers and recali-
brations of the tail-load gauges would take 3 months. This resulted in a man-
agement decision to omit the vertical-tail-load research from future D-558-1
activities. Managers at NACA Headquarters learned of the decision and sent
a letter to Hartley Soulé at Langley seeking his input on the advisability of
abandoning the tail-loads study. A letter was also sent to the High-Speed Flight
Research Station asking for information on the circumstances of the decision.49
Donald Bellman, an engineer at the High-Speed Flight Research Station,
wrote the reply to NACA Headquarters. He noted that the 3-month delay
needed to make the repairs and complete the recalibration “eliminated the
70
possibility of starting the vertical tail load program in the 1952 flying season.
Prospective work of greater importance precluded the program from the 1953
flying season, so the program was abandoned.”50
Research with the D-558-1 did not resume until January 29, 1953.
Dynamic-stability measurements were the goals of the new flights. These would
be accomplished by making elevator and rudder pulses at transonic speeds at
25,000- and 35,000-foot altitudes. Although some dynamic-stability data that
had been collected in earlier missions focused on lateral stability, about two-
thirds of the elevator pulse data and nearly all the rudder pulses had yet to be
collected. Butchart made five dynamic-stability flights, the last on March 27.
As before, a new pilot now joined the project. He was John B. “Jack” McKay, a
Navy F6F Hellcat pilot during World War II. After the war, he attended Virginia
Polytechnic Institute and graduated in 1950 with a bachelor’s degree in aeronau-
tical engineering. McKay joined the NACA in January 1951, initially working at
Langley for a brief period as an engineer before transferring to the High-Speed
Flight Research Station. His assignment on the D-558-1 was to provide dynamic-
stability “fill-in” data.
McKay made his pilot checkout on March 27, 1953. This was followed on
April 1 and 2 by the two fill-in flights. At the time of his assignment, these
flights were supposed to mark the end of research with the D-558-1. But
High-Speed Flight Research Station managers had approved a new research
project designed to investigate the effects of tip tanks on the Skystreak’s buffet
characteristics, so McKay also piloted these flights.
The first tip-tank flight, made on May 7, 1953, was aborted because a
fuel-vent failure caused a leak in the left tank. The second flight, on May 12,
was no more successful. The data recorder had not been turned on, so McKay
came home with no data. Despite McKay’s shaky start on the project, the next
four flights, on May 13, May 20, June 2, and June 3, 1953, were successful.
Crossfield made the 78th and final NACA research flight in the third D-558-1,
an investigation of low-speed stability and control in coordinated turns, on
June 10, 1953. With his landing, the Skystreak passed into history.51
An Assessment
The Skystreak’s importance in the exploration of the supersonic frontier was, as
the two opening quotations indicate, conflicted. This was due to many factors.
The first was the aircraft’s limitations. John Stack did not want a supersonic
airplane, but one that could fly at high-transonic speeds and provide the data
that existing wind tunnels could not. The NACA’s large wind tunnels, then
the best in the world, began to choke at the speeds at which compressibility
71
Probing the Sky
The third D-558-1 Skystreak parked outside an NACA hangar at Edwards South Base in 1949.
The aircraft was painted overall white to make photo tracking easier against the dark blue desert
sky. The original dark red paint was intended to be highly visible but actually made the Skystreak
difficult to spot. (NASA)
problems were appearing with aircraft. Makeshift efforts designed to get around
this stumbling block, such as wing flow and rocket-boosted models, had limita-
tions, so building research aircraft remained the only option.
In a broader context, the D-558-1’s limitations reflected the NACA’s cau-
tious approach to the sonic frontier. The agency’s approach in 1944–46 was
characterized by a focus on research conducted through small steps rather than
bold leaps and through incremental improvements in aircraft technology rather
than wholesale revolutionary breakthroughs. Much of the NACA’s war work
was in drag reduction for existing production aircraft and was focused on small
increases in performance. The surprisingly ambivalent response to Robert T.
Jones’s paper on swept wings (until confirmed by evidence of Nazi wartime
work) was another reflection of this mentality; it continued with the NACA’s
proposed X-planes research plan, which called for numerous flights with several
different aircraft stretching over a long period, with supersonic flight achieved
only at the very end—a program plan that led to the AAF ramming through
its accelerated XS-1 assault.
The Douglas engineers designing the D-558-1 were driven by the NACA’s
recommendations for a simple first-generation transonic research aircraft, but
also by the Navy’s (and their own) interest in possibly using the aircraft to at
72
least contribute to an operational naval fighter. The Skystreak represented a

conservative design for a land-based jet aircraft, but a radical one for a naval
fighter, in the age of straight-deck carriers more suited to straight-wing slow-
approach-speed aircraft than to the “hotter” swept-wing jets. It is important to
note that the advent of the swept wing did eventually result in pressures to build
a swept-wing derivative of the D-558-1, which led to the Mach 2–breaking
D-558-2 Skyrocket, first flown in February 1948. Compared to this pointed,
streamlined (indeed elegant) design, the “original” D-558-1, with its cylindrical
constant-diameter fuselage and straight wings and tail surfaces, looked archaic
rather than futuristic.
Despite all of this, the Skystreak was a remarkably productive aircraft.
During the D-558-1 research flights, a wide range of activities was undertaken.
The tests included high-Mach-number dives; research on the effects of tip tanks
and on aileron effectiveness; directional-stability rudder kicks and side slips
made to collect stability data; measurements of pressure distribution and lift
pressure; and checks of vortex generators, buffeting and tail loads, longitudinal
stability, and dynamic and lateral stability in the transonic range.52
The significance of the Skystreak’s accomplishments lies not in the tests
themselves, many of which were the standard tests used on prototype and
production aircraft, but rather in the database assembled across the range of
speeds at which they were flown. This made the D-558-1 important. In this
context, at least, it did not matter that the Skystreak lacked swept wings or
rocket power.
The capabilities and limitations of the D-558-1 and the XS-1 (renamed the
X-1 in June 1948) resulted in a division of research methodology. Because the
Skystreak could cruise for (relatively) prolonged periods at transonic speeds,
it was used for such research, freeing the shorter-duration XS-1 for use in
collecting data at supersonic speeds. The XS-1 was ill-suited to undertake sus-
tained flights at transonic speeds because of the rocket engine’s ravenous fuel
consumption, which left it with, at most, 2½ minutes of powered flight time.
Paradoxically, the D-558-1’s contributions were also more limited for the
same reason. The X-1s and swept-wing D-558-2 continued to be successfully
used for research activities into the late 1950s. The lower performance of the
D-558-1 meant its useful lifespan was shorter; Soulé’s December 19, 1952,
letter to NACA Headquarters regarding the cancellation of the vertical-tail-
loading studies with the Skystreak noted, “It should be understood that the
abandonment of the proposed D-558-1 vertical-tail-load flight tests does not
mean the abandonment of the study of vertical-tail loads at Edwards. Vertical-
tail load data are being obtained during flights of the X-5 and D-558-2 air-
planes, and it is believed that the resulting information will be of more interest
than such information on the D-558-1 airplane.”53 (Emphasis added.)
73
Probing the Sky
The clear implication was that the D-558-1 was nearing the end of its use-
fulness and that the faster, rocket-powered Skyrocket had more to offer than
the jet-powered Skystreak.
Over time, as memories of the debates that attended the development of
the XS-1 and D-558-1 faded, the myth that their separate roles had been
intentional from the start developed in the research community. This mythol-
ogy held that the two airplanes were deliberately planned to undertake the two
disparate sets of research activities. While not true, and while later aircraft such
as the F-86 were soon made available to the NACA for high-speed research, it
is certainly true that the availability of the Skystreak as a complementary test
system to the flashier XS-1 benefited postwar aeronautical research. As John
Becker noted, “It was the D-558-1’s and not the advanced service aircraft that
were used for extensive flight research at high subsonic speeds by [the] NACA,
complementing coverage of the higher transonic speeds by the X-1s. It is quite
understandable how some NACA managers by hindsight can see a logic in
the way those two vehicles were used that did not really exist when they were
promoted in 1944 and 1945.”54
For NACA researchers, pilots, and engineers, the Skystreak also provided
initial experience with the new demands of research aircraft operations, but
at a terrible price. The death of “Tick” Lilly (whose portrait hangs in Dryden
Flight Research Center to this day) in the second D-558-1 crash highlighted
design flaws in the aircraft, as well as in the approach taken to the project.
These oversights included the fact that the engine lacked the latest upgrades and
modifications, as well as the vulnerability of the control cables, fuel lines, and
other components to damage. Acting on the recommendations of the accident
board in the wake of Lilly’s crash, the Douglas X-3, Northrop X-4, and Bell
X-5 research airplanes also received modifications, benefiting them greatly.55
There was a final accomplishment of the X-1 and D-558-1 programs that
is often overlooked. The research aircraft were built in part because the exist-
ing wind tunnels choked between Mach 0.80 and low supersonic speeds.
Researchers saw research aircraft as being a substitute for wind tunnel testing
as drop bodies, rocket-boosted models, and the wing flow technique had
been. From 1947 to the early 1950s, the D-558-1 and the X-1 represented
the only tools available for collecting transonic data. The irony was that the
construction of the research aircraft generated pressures forcing wind tunnel
researchers to seek ways to fix the choking problem.
Overcoming the wind tunnel limitations required several steps. The first
was reducing the size of the model to just one-tenth of 1 percent of the tunnel
throat area. Choking still occurred, but its onset was delayed from Mach 0.80
to Mach 0.95. Langley researchers realized that in order to take advantage of
74
The third D-558-1 in flight with scattered clouds in the distance and the desert below. The
Skystreak did not have the impact that the XS-1 had on aviation technology, but it did contribute
to the understanding of transonic flight. (NASA)
the small-model technique, the support-structures’ designs would have to be

changed, as their surface area was now much larger than that of the models.
They developed the technique of attaching the model to a long rod, called
a “sting,” which was placed farther downstream in the tunnel. A specially
contoured insert on the tunnel’s wall was also added ahead of the sting. The
two features both corrected the blockage of the tunnel and created a more
uniform airflow. Langley researchers used an early version of the sting/liner
beginning in the spring of 1946 for wind tunnel testing of the XS-1 and
D-558-1 designs at speeds as high as Mach 0.92.
Langley engineer Ray H. Wright made the next significant breakthrough.
He proposed putting lengthwise slots in the throat of a wind tunnel test section,
originally to eliminate the effects of wall interference. This technique evolved
into using the slots to reduce the choking at high transonic speeds, a practice
to which Italian aerodynamicist Antonio Ferri and the German inventor of
the swept wing, Adolf Busemann, both expatriate theoreticians working at
Langley in the postwar years, objected to at a September 1947 meeting with
John Stack. They argued that though the slots would reduce the choking, at
Mach 1 the data were unlikely to be valid.
75
Probing the Sky
Stack was not overly concerned about the small remaining gap. His response
to Ferri and Busemann’s objections was to say that if the slotted tunnel worked
at Mach 0.995 and Mach 1.005, the gap in the middle was meaningless. It
took several more years before the slotted tunnel was an operational reality. The
initial modifications to Langley’s 8-Foot High Speed Tunnel (HST) exceeded
Mach 1 in late 1948, but the airflow was “rough and uneven.” The slots had to
be carefully shaped to achieve smooth transonic airflow. On October 6, 1950,
transonic research operations in the 8-Foot HST began. The 16-Foot HST
also began operation with a slotted throat 3 months later. The significance
of the achievement was recognized in 1951 when Stack and his associates
were awarded the Collier Trophy for the development of the slotted tunnel,
indirectly another accomplishment attributable to the onset of the postwar
X-series aircraft.56
76
Endnotes
1. Opening quotations are from Hallion, Supersonic Flight, p. 141, and

Becker, The High-Speed Frontier, p. 96.
2. During its development and service life, the Douglas Skystreak was
variously referred to as the “D-558-I” and the “D-558-1.” The latter
usage is more prevalent in Douglas, Navy, and NACA documenta-
tion, and, accordingly, this work uses D-558-1 for consistency.
3. See Edward H. Heinemann and Rosario Rausa, Ed Heinemann:
Combat Aircraft Designer (Annapolis, MD: Naval Institute Press,
1980), pp. 141–150.
4. Tuncer Cebeci, Legacy of a Gentle Genius: The Life of A.M.O. Smith
(Long Beach, CA: Horizons Publishing, Inc., 1999), p. 28.
5. Hallion, Supersonic Flight, pp. 69, 71–73.
6. Edward H. Heinemann, “The Development of the Navy-Douglas
Model D-558 Research Project” (El Segundo, CA: Douglas Aircraft
Company, November 17, 1947), pp. 3–7.
7. Heinemann, “The Development of the Navy-Douglas Model
D-558 Research Project,” p. 7; Curtis Peebles, ed., The Spoken
Word: Recollections of Dryden History, The Early Years, SP-2003-4530
(Washington, DC: NASA, 2003), pp. 84, 110. Scott Crossfield,
who flew the D-558-1 and -2, did not have a high opinion of their
escape systems. He later noted in The Spoken Word, “This is the way to
commit suicide to keep from getting killed. They never did have the
development on them that they should have had, and they weren’t any
good anyway. If you could make a capsule that was good enough to
live through the emergency, you might as well fly it and throw away
the airplane.” Stanley P. Butchart, who also flew the D-558-1 and -2,
echoed Crossfield’s comments. Butchart added, also in The Spoken
Word, “When you stop to think of it, [at] the higher speeds, and you
drop the nose off, you’re going to get a very big negative g as you come
out of there. So that restricts you as to how fast you can be going and
still use that escape method.” Just such a problem killed Air Force
Captain Milburn Apt on September 27, 1956, when he attempted to
escape from the first Bell X-2 following his flight to Mach 3.2.
8. “Supersonic Douglas Skystreak To Race Speed of Sound for Navy”
(Douglas press release), pp. 1–2; Hallion, Supersonic Flight, pp.
74–75; Dryden Flight Research Center public movie Web page,
http://www.dfrc.nasa.gov/Gallery/Movie/D-558-1/index.html, accessed
February 11, 2010.
77
Probing the Sky
9. Heinemann, “The Development of the Navy-Douglas Model D-558

Research Project” (El Segundo, CA: Douglas Aircraft Company,
November 17, 1947), pp. 2–3.
10. Ibid.
11. Dryden Flight Research Center public movie Web page, http://www.
dfrc.nasa.gov/Gallery/Movie/D-558-1/index.html, accessed February
11, 2010; Hallion, Supersonic Flight, pp. 140, 145.
12. Hallion, Supersonic Flight, pp. 140–142; Douglas Aircraft Company,
“Skystreak World’s Speed Records,” September 19, 1947; Becker,
The High-Speed Frontier, p. 96. The D-558-1 record stood until an
F-86A broke the record on September 15, 1948, reaching a speed of
671 mph. By then, F-86s were routinely exceeding Mach 1 in dives.
The low-altitude speed run was originally designed for wood-and-
fabric biplanes, so pilots could not artificially increase their speed by
making a dive from high altitude before beginning each run. By the
late 1940s, with jets flying at speeds in excess of 600 mph in the dense
lower atmosphere, this approach was very dangerous. A small control
error would cause the aircraft to hit the ground. Ironically, while Carl’s
average speed was higher, both he and Caldwell had matching Mach
numbers of 0.828. This was because Mach number depended on
altitude and air temperature. Caldwell’s flights were made at a tem-
perature of a little over 75 °F. Carl’s flight, in contrast, was made at an
air temperature of nearly 95 °F.
13. Walter C. Williams, “Limited Measurements of Static Longitudinal
Stability in Flight of Douglas D-558-1 Airplane” (BuAero No.
37971), NACA RM L8E14 (June 24, 1948), p. 2. The date on an
NACA Research Memorandum was when it was issued.
14. Hallion, Supersonic Flight, pp. 142–144; Richard P. Hallion and
Michael Gorn, On the Frontier: Experimental Flight at NASA Dryden
(Washington, DC: Smithsonian Books, 2003), p. 386.
15. Letter, Walter C. Williams to NACA, “Choice of Color for Research
Aircraft at Edwards,” December 3, 1951, reprinted in J.D. Hunley,
Toward Mach 2: The Douglas D-558 Program (Washington, DC:
NASA, 1999), pp. 116–117. The word “photographability” is used
in the text. The D-558-1s eventually had pronounced differences
in their finishes. The first Douglas D-558-1 ended its career with a
white fuselage and lower wings, and the canopy, wing tanks, rud-
der, vertical-tail tip, horizontal stabilizers, elevators, ailerons, speed
brakes, and (apparently) upper wing surfaces and landing gear doors
were red. The NACA’s third D-558-1 was all-white with the excep-
tion of the red rudder, elevators, and ailerons. The second D-558-1
78
had an NACA shield on its vertical tail. The third D-558-1 bore the
NACA shield initially, then a yellow band with “NACA” in black
and “NACA” on the upper right wing. Late in its service, a large
black “X” was painted on both sides of its forward and aft fuselages
for photo tracking.
17. “NACA Aircraft Accident Investigation Report, Douglas D-558-1
Airplane” (BuNo 37971), Muroc Air Force Base, Muroc, CA, May
3, 1948, p. 2.
18. “NACA Aircraft Accident Investigation Report, Douglas D-558-1
Airplane” (BuNo 37971), Muroc Air Force Base, Muroc, CA, May
3, 1948. The D-558-1 was originally fitted with a bubble canopy
for the Douglas flights. In videos of early flights, May, who was
small in stature, can be seen wearing a silver crash helmet. When the
V-shaped canopy replaced the original design, it was then too small
to allow most pilots to wear a crash helmet. The accident report
noted that the lack of a helmet and the shoulder straps would have
made no difference in Lilly’s crash. During the speed-record flights,
Caldwell had worn a helmet. Carl was taller than Caldwell, however,
and he could not close the canopy while wearing a crash helmet.
Carl had to wear a World War II–type cloth helmet.
19. Walter C. Williams, “Limited Measurements of Static Longitudinal
Stability in Flight of Douglas D-558-1 Airplane, BuAero No.
37971,” NACA RM L8E14, (June 24, 1948).
20. Ibid., p. 4.
21. Williams, “Limited Measurements of Static Longitudinal Stability in
Flight of Douglas D-558-1 Airplane,” pp. 3–4, 10.
22. Walter C. Williams, “Flight Measurement of the Stability
Characteristics of the D-558-1 Airplane (BuAero No. 37971) in
Sideslips,” NACA RM L8E14a (April 18, 1949), pp. 1–4.
23. Ibid., p. 4.
24. William H. Barlow and Howard C. Lilly, “Stability Results
Obtained with Douglas D-558-1 Airplane (BuAero No. 37971) in
Flights up to a Mach Number of 0.89,” NACA RM L8K03 (April
22, 1949).
25. Ibid., p. 4.
26. Ibid., pp. 3–5.
27. Scott Libis, Skystreak, Skyrocket, & Stiletto: Douglas High-Speed
X-Planes (North Branch, MN: Specialty Press, 2005), pp. 26–28.
Credit for the later development of the winglet goes primarily to
79
Probing the Sky
the remarkable NACA-NASA aeronautical scientist Richard T.

Whitcomb of the Langley Research Center.
28. Hallion, Supersonic Flight, pp. 146, 260.
29. Ibid., pp. 146–147.
30. Hallion and Gorn, On the Frontier, p. 387.
31. Jim Rogers Thompson, William S. Roden, and John M. Eggleston,
“Flight Investigation of the Aileron Characteristics of the Douglas
D-558-1 Airplane (BuAero No. 37972) at Mach Numbers Between
0.6 and 0.89,” NACA RM L50D20 (May 26, 1950), p. 5.
32. Ibid. pp. 1–6.
33. Earl R. Keener and Mary Pierce, “Tabulated Pressure Coefficients
and Aerodynamic Characteristics in Flight on the Wing of the
Douglas D-558-1 Airplane for a 1g Stall, a Speed Run to a Mach
Number of 0.90, and a Wind-Up Turn at a Mach Number of 0.86,”
NACA RM L50J10 (December 15, 1950).
34. Hallion, Supersonic Flight, pp. 147–148; Libis, Skystreak, Skyrocket,
& Stiletto, p. 121.
35. Keener and Pierce, “Tabulated Pressure Coefficients and
Aerodynamic Characteristics in Flight on the Wing of the Douglas
D-558-1 Airplane for a 1g Stall, a Speed Run to a Mach Number of
0.90, and a Wind-Up Turn at a Mach Number of 0.86,” p. 5.
36. De E. Beeler, Donald R. Bellman, and John H. Griffith, “Flight
Determination of the Effects of Wing Vortex Generators on the
Aerodynamic Characteristics of the Douglas D-558-1 Airplane,”
NACA Research Memorandum RM L51A23, August 14, 1951,
p. 8.
37. Ibid., pp. 4–7; Hallion and Gorn, On the Frontier, pp. 386–387.
Vortex generators were also tested on the final flight of NACA
X-1 #2 on October 23, 1951. The rocket cut out after two igni-
tion attempts, and the aircraft glided to a landing. The aircraft was
subsequently grounded due to possible fatigue failure of the nitrogen
spheres.
38. Ibid., p. 149; Hallion and Gorn, On the Frontier, pp. 386–387.
39. Peebles, ed., Spoken Word, pp. 52–53, 77.
41. Ibid., p. 135; Hallion and Gorn, On the Frontier, p. 388.
42. Ibid., p. 151; Peebles, ed., The Spoken Word: Recollections of Dryden
History, The Early Years, pp. 103–104.
43. Butchart served with another outstanding young naval airman
who became a close friend—then-Ensign George H.W. Bush, later
President of the United States.
80
44. Earl R. Keener, James R. Peel, and Julia B. Woodbridge, “Tabulated

Pressure Coefficients and Aerodynamic Characteristics Measured in
Flight on the Wing of the Douglas D-558-1 Airplane Throughout
the Normal-Force Range at Mach Numbers of 0.67, 0.74, 0.78, and
0.82,” NACA RM L50L12a (January 29, 1951), pp. 1, 4–5. One
possible reason for the lack of analysis in the reports is the outbreak
of the Korean War in June 1950. The straight-wing U.S. Air Force
and Navy jet fighters were outclassed by the swept-wing Soviet MiG-
15, which entered combat in November 1950, forcing immediate
deployment of the swept-wing USAF F-86A.
45. Melvin Sadoff, William S. Roden, and John M. Eggleston,
“Flight Investigation of the Longitudinal Stability and Control
Characteristics of the Douglas D-558-1 Airplane (BuAero No.
37972) at Mach Numbers up to 0.89,” NACA RM L51D18 (June
25, 1951), p. 4.
46. Ibid., pp. 1, 7, 16–18. The all-moving stabilizer on the F-86 gave
U.S. pilots an advantage over the Soviet MiG-15 in air combat
during the Korean War. An F-86 could dive away from an attack-
ing MiG-15 and retain control despite reaching transonic and
low supersonic speeds. This was because the MiG-15 still used
fixed-stabilizer and movable elevators, making it unstable at tran-
sonic speeds and incapable of reaching Mach 1 under any flight
conditions.
47. Ellwyn E. Angle and Euclid C. Holleman, “Longitudinal Frequency-
Response Characteristics of the D-558-1 Airplane as Determined
from Experimental Transient-Response Histories to a Mach Number
of 0.90,” NACA RM L51K28 (February 11, 1952), pp. 1, 4, 7–8.
& Stiletto, p. 32.
49. Letter, Hartley A. Soulé to NACA Headquarters, “Request for
Comments on Abandonment of D-558-1 Vertical-Tail Load
Investigation,” December 19, 1952.
50. Letter, Donald R. Bellman to NACA, “Looseness of Stabilizer
Mountings for the D-558-1 Airplane,” November 5, 1952; letter,
Donald R. Bellman to NACA, “Reply to Request for Information
on D-558-1 Vertical-Tail Load Investigation,” January 12, 1953.
& Stiletto, p. 122; Hallion and Gorn, On the Frontier, p. 388.
The fates of the different research aircraft reflected their perceived
importance. The X-1 #1, in which Yeager made the first supersonic
flight, was given to the Smithsonian Institution after its retirement
81
Probing the Sky
in May 1950. It is currently on display at the National Air and Space

Museum in Washington, DC. It shares the main entrance with the
Wright Flyer, the Spirit of St. Louis, the X-15 #1, and the Apollo 11
Command Module. After Crossfield’s last flight in the third D-558-
1, it joined the first D-558-1 in dead storage at Edwards Air Force
Base. Both aircraft were subsequently transferred. The first D-558-1
was sent to California State Polytechnic College in San Luis Obispo,
CA, in 1957. It was transferred to the National Museum of Naval
Aviation in 1964, where it was restored and remains today. The
third D-558-1 was sent for display at the Marine Corps Air-Ground
Museum at Quantico, VA. The aircraft was displayed outdoors and
rapidly deteriorated. Fortunately, it was transferred to the Carolina
Aviation Museum, restored, and put on display in 2001.
52. Libis, Skystreak, Skyrocket, & Stiletto, pp. 116–122.
53. Soulé, “Request for Comments on Abandonment of D-558-1
Vertical-Tail Load Investigation.”
54. Becker, The High-Speed Frontier, p. 96.
56. Hanson, Engineer in Charge, pp. 312–327; Becker, The High-Speed
Frontier, pp. 98–114.
82
The second D-558-2, NACA 144, parked on the ramp at South Base. Modifications made to
allow air launch included the removal of the jet engine and fuel tanks. Between the air launch
and the increased rocket burn time, the craft’s maximum speed was doubled. On November 20,
1953, this aircraft made the first Mach 2 flight. (NASA)
84
CHAPTER 3
Proving the Swept Wing:

The Douglas D-558-2 Skyrocket
It was actual flight time that was the real education—five minutes in
the air with the experimental ship was worth ten hours of study on the ground,
and gradually I understood the magnitude of the horizon that lay out there,
unknown, waiting to be probed in the rocket ship.
—William “Bill” Bridgeman1
Ed Heinemann and the Douglas design team had always envisioned that the
D-558 project would entail multiple aircraft configurations, consistent with
what the Navy’s Bureau of Aeronautics envisioned coming from the program.
Originally, the Navy contract awarded to Douglas called for six D-558 phase 1
aircraft, with different combinations of side and nose inlets and straight wings
with three different airfoil sections. Phase 2 involved adding rocket boosters
and replacing the TG-180 jet engine with a smaller Westinghouse 24C turbojet
on three of the aircraft. Phase 3 originally called for a mockup of an operational
combat aircraft.2
This plan underwent a radical change due both to the capture of research
material from Germany, after the war, and to R.T. Jones’s paper on swept
wings. In the early summer of 1945, John Stack suggested to Douglas that the
D-558-1 incorporate a 35° swept wing. However, it was not until after analy-
sis of German documents brought back to America by the Naval Technical
Mission to Europe (“NavTechMisEu”), one of whose members was Douglas
engineer A.M.O. Smith, that the NACA, the Bureau of Aeronautics, and
Douglas agreed to examine a swept-wing derivation of the D-558 program
effort. The formal request by the Navy and the NACA for the new configura-
tion came in mid-August 1945.
A practical consideration was the fact that the existing turbojet engines lacked
the thrust capacity to reach high transonic speeds. As a result, Douglas, the Navy,
and the NACA agreed from the program’s outset that a rocket-powered research
aircraft would be needed to evaluate the full potential of swept wings at tran-
sonic and supersonic speeds. Fortunately, with the development of the Reaction
85
Probing the Sky
NACA 144 at South Base. The “Turbine Ex” is the exhaust for the turbopump, which fed fuel
to the rocket engine. The “LOX Prime” was part of the liquid-oxygen vent/jettison system. The
X’s on the forward and aft fuselages were photo reference marks. South Base was built during
World War II as a training field for B-24 and P-38 pilots. It was the home for the NACA contin-
gent from 1946 to 1954, when it moved to the current facility. (NASA)
Motors, Inc., (RMI) 6000C4 engine—which the AAF designated the XLR-11,
and the Navy the XLR-8—just such an engine existed that could be applied to
the D-558. The swept-wing D-558 would take off and climb to test altitude using
a Westinghouse 24C jet engine. The pilot would then employ the RMI rocket
engine to accelerate into the high transonic and low supersonic regime. Once
this was completed, the rocket would shut down, and the pilot would restart
the jet engine for the descent to a powered landing. Two separate engines and
fuel systems would have to fit within the fuselage. Douglas engineers found it
impossible to squeeze both a rocket and a jet engine into the D-558-1’s already-
narrow fuselage, so they had to start from scratch in designing the D-558-2. The
result was a much shapelier and elegant design, which became one of the iconic
symbols of aeronautical progress in the postwar era.
The D-558 Phase 2 Design Process

Kermit E. Van Every was assigned the task of designing the new aircraft, and he
faced a number of design requirements. Among these was the Navy managers’
requirement that the swept-wing version should have the low-speed and relatively
benign stall characteristics of the straight-wing D-558-1. Douglas engineers faced
86
Proving the Swept Wing: The Douglas D-558-2 Skyrocket
a difficult task in meeting this requirement, as swept wings were known to have
poor low-speed and stall behavior. At the same time, they had to ensure that the
aircraft would be stable at high speeds.
The D-558-1’s cylindrical fuselage and nose inlet were abandoned in the
new design. Instead, the new aircraft had an elongated fuselage; a flush wind-
shield similar to that of the XS-1 enhanced the plane’s sleek lines. Flush air
inlets were located low on the sides of the forward fuselage. The fuselage diam-
eter had to be increased, as compared with that of the D-558-1, for it contained
the Westinghouse 24C jet engine and two tanks holding 250 gallons of jet
fuel, the LR-11 rocket engine and tanks holding 195 gallons of water/alcohol
fuel, 180 gallons of liquid oxygen (LOX), 11 gallons of 90-percent hydrogen
peroxide to power the turbopump, and the helium used to pressurize the fuel
system. The nose and main landing gear also retracted into the fuselage. Finally,
the flight data instrumentation, totaling between 800 and 1,100 pounds, com-
pleted the payload.
The rocket engine was mounted at the rear of the fuselage. The jet engine
was positioned in the middle, with the air-inlet ducts passing around the liquid-
oxygen tank to reach the engine. The two tanks for the jet fuel were above
the jet engine. The water/alcohol fuel tank was located above the jet engine’s
exhaust pipe, which was angled slightly downward and exited under the aft
fuselage. Saying that the airplane was oddly arranged—to say nothing of the
mix of power plants—would be an understatement. In fact, the NACA never
had another such amalgamation.
The D-558-1’s straight cylindrical fuselage cross section was retained for the
D-558-2 at the wing-fuselage intersection, which gave it a measure of “area
ruling” before the advent of the concept. Had the fuselage been an ogival body,
like the X-1’s, local airflow velocity at the wing-fuselage juncture would have
increased, causing early shockwave formation and, therefore, an increase in
transonic drag. Its wing had a 35° sweep and a span of 25 feet, representing a
compromise between low- and high-speed requirements. This also resulted in
a lower aspect ratio than the D-558-1’s straight wings, further reducing drag.
But it also complicated making the airplane safe to fly at low speed. Kermit Van
Every blended a variety of elements to ensure that the D-558-2’s swept wing
was suitable. He selected a modest 35° sweep angle since this was already, by
the beginning of the postwar era, a much-studied planform, tested by German
wartime researchers, employed on the postwar Bell L-39 swept-wing test bed (a
Navy-funded research program), and intended for the AAF’s XP-86 (prototype
for the F-86) and the planned Boeing XB-47 jet bomber. Van Every replaced
the NACA 65 series airfoil used on the D-558-1 with a higher-lift NACA 63
series airfoil. The latter had better low-speed/stall characteristics yet did not
sacrifice much in high-speed performance. Next, he increased the wing’s area to
87
Probing the Sky
175 square feet compared to the D-558-1’s 150.7 square feet. He gave the wing
a “reverse taper” (from a 10-percent thickness-chord ratio at the root to a more
lift-friendly 12-percent thickness-chord ratio at the tip). Finally, he added flaps,
wing fences, and automatic Handley Page wing-leading-edge slats to improve
its low-speed behavior. The slats could either operate automatically or be locked
open or closed. As with its predecessor, the D-558-2’s horizontal stabilizers were
thinner than the wings, could be moved independently of the elevators for high-
speed control, and were positioned high on the vertical tail to remain out of the
wing wake. Van Every increased the stabilizer’s critical Mach number further by
giving it a sharper 40° sweep, rather than the 35° sweep of the wings.3
First Flights
Douglas rolled out the first D-558-2 Skyrocket (BuNo 37973) on November
10, 1947, less than a month after Yeager’s Mach 1 flight. The aircraft was
not complete, however, as the LR-8 rocket engine had not yet been installed
because the turbopump needed to supply it with propellants was not yet ready.4
In its place, a cone-shaped fairing was added for the initial flights. Once the cer-
emonies were finished, ground tests of the aircraft began. The turbojet engine
underwent test runs on November 21. The aircraft had a flush canopy (like
the XS-1) that added to its racy looks but afforded its pilot minimal forward
visibility during approach and landing. Consequently, by December this had
been replaced with a standard raised canopy with a V-shaped windshield. This
provided much better visibility for the pilot even if it did spoil the aircraft’s
sleek lines. Once these final details were complete, the D-558-2 was loaded
on a flatbed truck and wrapped in tarps to conceal its shape. The convoy left
for Muroc on December 10, 1947.
After arriving at Muroc, the aircraft underwent further checkouts and the
instrumentation was installed, which took the rest of December. The pilot
selected for the early flights of the first D-558-2 was John F. Martin, a Douglas
test pilot since 1940 who had a background as a United Airlines pilot. He had
served as test pilot on the A-20, A-26, and C-54.
Initial taxi tests were made on January 5 and February 2, 1948, clearing the
way for the first flight on February 4. Problems were experienced on the first
flight; the already-anemic Westinghouse J34 engine was sluggish during startup
and acceleration. Things did not improve much when the takeoff roll across the
lakebed took 15,000 feet, and, once aloft, Martin discovered the airplane suffered
from a persistent “Dutch roll” lateral-directional (roll-yaw) oscillation, an early
indication that Douglas needed to increase the height (and hence area) of the
vertical fin, which it subsequently did, by 18 inches.5
88
A D-558-2 jet and JATO rocket take off in 1949. Because the LR-8 rocket engine was not initially
fitted, the program had to make do with this interim propulsion system to carry out research with
the aircraft. The takeoff roll was long, hindering maximum performance. Modifications were subse-
quently made to allow the D-558-2 to be air-launched from a B-29. (NASA)
To correct the long takeoff roll, Douglas engineers prudently decided to

attach two jet-assisted-takeoff (JATO) solid rocket boosters to the aircraft for
an added boost. The first JATO-boosted flight was made on July 13, 1948. The
rockets were fired midway through the roll and jettisoned after takeoff. This
shortened the takeoff run to 8,210 feet, a little over half the distance required
with the jet alone. As a result, the use of JATO rockets became standard for
the jet-powered, ground-takeoff flights; they not only shortened the takeoff
roll, but they also conserved fuel and improved safety. Various combinations of
two, three, or four rockets were tested, with four rockets deemed best. Martin’s
last D-558-2 flight was on August 25, 1948. Skystreak veteran Gene May
now took over the Douglas test duties, making his first flight in the airplane
on September 16, 1948. May began an extensive flight-test effort that lasted
more than a year.6
By this time, the second D-558-2 (BuNo 37974), intended for the NACA,
had been delivered to Muroc, and May made a pair of dive demonstration
flights on November 2 and 7, 1948, after which the aircraft was formally turned
over to the NACA on December 1, 1948. Like the first Skyrocket, the second
D-558-2 also lacked the LR-8 rocket engine when it was delivered because the
planned turbopump to feed it was still unavailable. In spite of this, the NACA
accepted the aircraft since initial flights were to be made for general stability
and control and air-load measurements at Mach 0.85, both of which could be
89
Probing the Sky
done with the jet engine alone. Once the LR-8 engine was available, the second
Skyrocket would be returned to Douglas and the rocket installed.
Engine problems and the installation and calibration of instrumentation
lasted through the winter and spring of 1949, delaying the start of research
operations. NACA research pilot Robert Champine made the first NACA
flight on May 24 for pilot familiarization, instrumentation checkout, and
general handling characteristics. Champine made two more research flights in
June for data on stability, control, and wing bending, as well as for wing twist
measurements. During the second flight, the cockpit camera caught fire, filling
the cockpit with smoke. Two airspeed calibration flights were made in July,
followed by a lateral-control-exploration flight in August.7
The dynamic-lateral-stability data collected in the first two NACA
Skyrocket flights were analyzed. Test pilots made sudden control inputs, and
the onboard instrumentation evaluated aircraft response. One test involved “a
lateral oscillation of the airplane resulting from abrupt deflection and release of
the rudder.” The research memorandum noted, “This maneuver was made at a
Mach number of 0.63 and an altitude of 12,000 feet with the airplane in the
clean condition. The data show the oscillation is slow to damp out especially
at small amplitudes where the oscillation is practically of constant amplitude.
The period of the oscillation is 1.6 seconds.”8
A second test maneuver made was a lateral oscillation while in the landing
condition (gear and flaps down). The memorandum noted, “This oscillation
was again induced by abrupt deflection and release of the rudder. In the land-
ing condition, the airplane performs a constant-amplitude oscillation with a
period of approximately 2.7 seconds.”9
A more complicated test maneuver measured the D-558-2’s behavior
during part of the landing approach. The report described the events as fol-
lows: “During the first part of this time history, between 30 and 44 seconds,
the pilot did not attempt to stop the oscillation by use of the ailerons or rudder
and the airplane performed a constant-amplitude oscillation. From 44 seconds
to 60 seconds the pilot used the ailerons and was able to damp the oscillation.
Even though the pilot can damp the oscillation, the oscillation is objectionable
particularly during landing approaches and landings because the controls must
be moved almost constantly. The rough-air handling qualities of the airplane
would probably be particularly objectionable.”10
The conclusion of the memorandum was that the D-558-2 showed differ-
ing dynamic-lateral-stability characteristics, depending on the situation. In a
“clean” condition (flaps and landing gear up, slats locked), the airplane’s lateral
oscillations were lightly damped, particularly when they were of small ampli-
tudes. With the D-558-2 in a landing condition, however, the airplane had
neutral oscillatory stability, which was indicated by the airplane undergoing a
90
The NACA 144 undergoing wing-strain-gauge calibration in a hangar. Sandbags were piled on
the upper wing to simulate aerodynamic forces. Strain-gauge readings were compared to calcu-
lated forces to determine any errors. These were used by the “computers” who reduced onboard
data after a flight. The number of engineers required for the calibrations was impressive. (NASA)
constant-amplitude oscillation. The pilot could damp this oscillation, but doing
so required vigorous control inputs, activity that was considered objectionable by
the pilot, as it required almost continuous control-surface movements. These, in
turn, rendered the D-558-2’s handling poor in less than ideal conditions, such
as turbulence.
This was not the only handling problem the D-558-2 would experience. A
bigger issue soon came to light.
Pitch-Up: The Insidious Threat

The seventh NACA flight, with Champine as the pilot, came on August 8,
1949. The plan involved making a 4-g turn at Mach 0.6. Without warning,
the Skyrocket’s nose pitched up with an acceleration of 6 g’s during the turn.
Champine countered with full-down elevator, and the aircraft recovered. As
a precaution, he immediately landed. What he had experienced was the first
91
Probing the Sky
proof that a swept-wing aircraft would pitch up during a hard turn. Tests
made using models had hinted at the possibility that longitudinal instability
would occur, resulting in the nose’s pitching up, but what the models had not
indicated was the severity of the problem.11
The source of the pitch-up problem was a combination of several factors.
A. Scott Crossfield, an NACA research pilot who made 62 D-558-2 flights,
described the sequence: “The air we fly in doesn’t like high sweep angles. It
doesn’t like severe taper ratios. And it doesn’t like low aspect ratios. And the
D-558-2 had a little bit or a lot of every one of these. And it was classic in what
it did as a swept wing…. The tips of the [D-558-2] wings tended to stall before
the roots of the wings. And if that’s aft of the center of gravity, the airplane
wants to pitch-up.”12
Pitch-up posed a number of risks, not least of which was loss of control.
In a turning dogfight, a swept-wing fighter experiencing a pitch-up could be
overstressed by g forces; the sudden loss of airspeed resulting from a pitch-
up could make an aircraft vulnerable to an enemy fighter’s attack or could
cause it to enter a spin and be unable to recover. Pitch-up posed risks in
day-to-day operations. A swept-wing aircraft maneuvering during a landing
approach might pitch up, stall, and have insufficient altitude and airspeed to
recover before crashing. In an extreme case, structural failure could occur. With
swept-wing F-86s entering service and a new generation of swept-wing fighters
and bombers in development, understanding pitch-up assumed critical and
urgent importance.
The NACA research effort with the second D-558-2 now focused on the
pitch-up issue. Joining the project was another NACA research pilot, John
Griffith. After Griffith’s checkout flight on September 12, 1949, he and
Champine split flying duties. Griffith experienced an inadvertent pitch-up
on his November 1, 1949, flight. The sequence was similar to one Champine
had experienced on a flight made nearly 3 months before. Griffith made
a 4-g turn at Mach 0.6, and the aircraft became longitudinally unstable,
triggering a pitch-up. Griffith attempted to fly beyond the instability; the
D-558-2’s angle of attack increased, and the aircraft yawed and rocked, then
snap-rolled, turning completely upside down. Griffith recovered, confirmed
that the Skyrocket had not been damaged, and continued the flight.13
Champine later described the experience of a D-558-2 pitch-up: “If
you pulled up and got 4 or 5 gs, it would suddenly stall in such a manner
that the lift distribution on the wing would cause it to pitch-up violently.
It would go to extremely high angles of attack, between 45 and 60 degrees,
and then it would start to roll violently, so the aircraft became completely
and totally out of control—just spinning around in the sky. Once you fell
into it you had no way of controlling it.”14
92
Griffith’s pitch-up experience was analyzed in a 1951 NACA research

memorandum as follows:
[T]he pilot attempted to fly the airplane in the normal-force-

coefficient range where the airplane was unstable (times between
11 and 14 sec). After the airplane pitched up, the airplane first
performed an unsteady rolling motion. The pilot used the rud-
der in attempting to control this rolling motion and caused the
airplane to perform a 360° snaproll. The data indicate that very
large angles of sideslip were reached during the snaproll. No
sideslip-angle measurements were obtained during the maneu-
ver, but integration of the yawing velocity indicates that sideslip
angles on the order of 30° to 35° were approached. A maximum
lateral acceleration of about 1.1g occurred during the snaproll.
This lateral acceleration corresponds to a side force on the airplane
of about 10,800 pounds. The distribution of side force on the
airplane between the fuselage and vertical tail is not known, but
it is likely that the load on the vertical tail, during the oscillatory
yawed flight, approached the vertical tail design limit of 8,700
pounds. The maximum rolling velocity which occurred during
the snaproll is not known as the 2.6-radian-per-second range of
the rolling-velocity recorder was exceeded, but it is likely that the
maximum rolling velocity was on the order of 3.5 to 4.0 radians
per second. In recovering from the snap-roll the airplane again
reached a negative normal acceleration of 3.0g.
The pilot reported and the recording instruments showed that
airplane buffeting occurred at normal accelerations slightly less
than the acceleration at which the airplane became unstable. This
buffeting served as a warning of the approach of instability. If the
elevator is moved down when the airplane buffeting occurs, the
response of the airplane is good and the instability can be avoided.
In the pilot’s opinion the airplane is unflyable in accelerated flight
in the lift-coefficient range in which it is unstable. If the pitch-up
resulting from the instability is not checked by moving the eleva-
tor down as soon as it is noticed by the pilot, the angle of attack
increases very rapidly and violent rolling and yawing motions
occur when the high angles of attack are reached.15
Between May 1949 and January 1950, 21 NACA research flights had been
made with the second D-558-2. The bulk of these were for longitudinal- and
lateral-stability, lateral-control, and stall data. The results of these flights were
93
Probing the Sky
summarized in a series of NACA research memorandums.16 Beyond describ-

ing the pitch-up characteristics of the D-558-2, one of the reports also gave
a preliminary assessment of the D-558-2’s longitudinal-stability and control
characteristics based on these flights. It noted the following:
With the slats locked and the flaps up, the airplane was longitu-
dinally unstable at normal-force coefficients greater than approxi-
mately 0.8 in steady flight at low speeds and in maneuvering flight
at Mach numbers up to at least 0.65. No data were obtained
at high normal-force coefficients at Mach numbers greater than
about 0.65 because of the power limitations of the airplane with
only the jet engine installed. The instability proved objective to the
pilots, particularly in accelerated flight because of the tendency
for the airplane to pitch to high angles of attack very rapidly and
because violent rolling and yawing motions sometimes occurred
when the high angles of attack were reached. The instability prob-
ably resulted from a large increase in the rate of change of effective
downwash at the tail with increase in angle of attack at moderate
and high angles of attack.
With the flaps down and the slats locked the longitudinal sta-
bility characteristics in steady flight at low speeds were very similar
to the characteristics with the flaps up and the slats locked except
that the instability occurred at a higher normal-force coefficient.
The degree of instability present with the slats unlocked and
the flaps up or down was much less than with the slats locked and
the pilots had only minor objections to the longitudinal charac-
teristics of the airplane.
In steady flight in the Mach number range from 0.50 to 0.87
the airplane is stable longitudinally and no abrupt trim changes
occurred up to the highest Mach number reached, 0.87. The
data indicate that only a slight reduction in the relative elevator-
stabilizer effectiveness occurred in going from a Mach number of
about 0.55 to 0.85.17
Into the Supersonic: The Skyrocket Turns to Air Launch

With the initial jet-powered NACA research flights completed, the second
D-558-2 was returned to Douglas for modifications. The shortcomings of
the ground takeoffs, even with JATO rockets, were now apparent. Bridgeman
later described the experience of a JATO takeoff in the Skyrocket:
94
A mile of runway is eaten up and she hasn’t hit 100 [mph]. Up to

80, now 90, there it is, 100. Hit the first two JATOs. A second,
and a kick in the fanny, and another kick as the remaining JATOs
fire off. She’s up to 180, roaring down the lakebed. This is the
time; pull her nose up…this is no F-80! Without her rockets,
climbing on jet alone, I can feel that she is far underpowered; she
handles like a truck, heavy and large. I’ve got hold of something
new all right.18
Clearly, something as simple as a blown tire from overheating during the

long takeoff runs could spell disaster. Something had to be done, and the sim-
plest solution was modifying the D-558-2s for air launching from a Boeing
B-29 mother ship.
Though the first and second D-558-2s had been delivered without the LR-8
rocket engine, the third (BuNo 37975) was equipped with the rocket power
plant (as well as its J34 turbojet) from the start. Gene May made the first flights
in the third D-558-2 beginning on January 8, 1949. On June 24, 1949, May
flew this aircraft through the speed of sound, noting later that the “flight got
glassy smooth, placid, quite the smoothest flying I had ever known.”19
But despite this, the D-558-2’s performance was still disappointing, a
byproduct of having both the Westinghouse 24C turbojet and LR-8 rocket
engine. Given the large amount of kerosene needed for the J34 turbojet for the
climb to altitude and descent to landing, the D-558-2 had insufficient rocket
propellant—the diluted alcohol/water fuel and liquid-oxygen oxidizer—to
permit the Skyrocket to penetrate far into the supersonic regime. An NACA
study showed that a jet-only ground takeoff was limited to a top speed of Mach
0.9. Using the LR-8 for an added boost during ground takeoff resulted in a
limiting top speed of Mach 0.95. NACA engineers concluded that the small
increase in speed was not justified by the added risk of an LR-8 rocket ground
takeoff. Following the fourth Douglas flight, on March 27, 1949, May had
noted, “[The] ultimate performance of the aircraft will never be achieved with
the limited supply of rocket fuel.”20
Air launch enabled getting the most out of the aircraft by eliminating the
need for using the scarce rocket propellants to help kick the Skyrocket off the
ground. Better still was turning the Skyrocket into an all-rocket boost-glider
like the Bell X-1. Douglas engineers calculated that by removing the jet engine
and adding more fuel for the rocket (using the space previously taken up by
the jet engine and its inlets and exhaust system), an air launch would produce
a top speed of Mach 1.46 to 1.6, dramatically increasing the Skyrocket’s value
as a research tool. The NACA’s chief of research, Hugh L. Dryden, supported
the Douglas proposal, suggesting the company modify the second aircraft to
95
Probing the Sky
an all-rocket/air-launch configuration at the same time the LR-8 rocket was

installed. The third could be modified at the same time and used by Douglas
in an air-launch demonstration before it was turned over to the NACA.
Dryden sent a letter to the Navy on September 1, 1949, strongly recom-
mending the modifications. Navy managers agreed, and, on November 25,
1949, the Navy Bureau of Aeronautics issued a contract change ordering the
second and third D-558-2 aircraft to be modified to an air-launch configuration
and a B-29 modified as the mother ship, which, by the Navy’s nomenclature
system, was designated a P2B-1S Superfortress. (The B-29 already modified
for the XS-1 could not be used to drop the Skyrocket, as the sweptback wings
and tail surfaces of the D-558-2 necessitated a completely different geometric
set of modifications to the bomb bay.)
The third D-558-2 would retain the J34 jet engine and inlets, the LR-8
rocket engine, and the existing fuel system. The only modifications would be
the addition of retractable launch hooks. The aircraft was trucked back to the
Douglas plant in El Segundo. Bridgeman made the first air launch of a D-558-2
on September 8, 1950, using only the turbojet. Five more air launches fol-
lowed, the last two using both jet and rocket power. Bridgeman described his
experiences on his November 27, 1950, flight:
Ten…nine…eight…seven…. My hand is on the data switch…

six…five. There. One hand on the wheel, one wrapped around
the throttle. Four…three…two…one…drop!
From out of the dark, protective belly of the [B-]29 the world
bursts over me in bright light. She’s free. I’m away from the mother
ship clean and free. It works…. Rapidly I click on the four rocket
tubes…. She heads out, still losing altitude but trimmed nose-
up…. I check the rocket pressures. In the green…. I pull it up
and begin the climb…. Too much…. I drop the nose a fraction.
Too far. Up again; it is now a matter of calculation and feel…. I
push her nose over for the speed run. She accelerates rapidly in
the thin sky-meadow of the higher altitude.
The buffet! Okay, that’s .91. But she doesn’t stop; she still
buffets although we are through into .92, the other side, where
she smooths out always. The buffeting continues… .93, .94…. I
search the instrument panel and there’s the answer—the turbine
out temperature is overboard by 500[°]. At once I jerk the throttle
to idle, but it is not fast enough; the engine gives up. She flames
out, loses thrust, and pitches over, throwing me in the harness
and against the panel…. At once, in rapid succession, the rocket
tubes fail. The altitude is too much for the jet engine. When she
96
pitched forward she unported the rocket fuel, starving off the
tubes—all in a matter of ten seconds.21
With the completion of the company’s acceptance flights, the aircraft was
turned over to the NACA on December 15, 1950, and given the designa-
tion NACA 145. Scott Crossfield became the project pilot. He made pilot-
familiarization and instrumentation check flights on December 22 and 27,
1950. These were jet-only flights, as were the initial two research flights in
March and April 1951. Crossfield made the first NACA jet-rocket flight on
May 17, but he shut down the jet engine during the flight because of combus-
tion instability. Joining the project at this point was NACA pilot Walter P.
Jones, who made his first Skyrocket flight on July 20, 1951. Early activity in
the series focused on lateral and longitudinal stability evaluated via aileron rolls
and elevator pulses, as well as accelerated turns and pitch maneuvers.
Longitudinal-stability characteristics were determined with the aircraft in a
clean condition and in turning flight at Mach numbers between 0.5 and 0.96,
at altitudes between 19,000 and 36,000 feet. The Skyrocket was stable up to
moderate values of normal-force coefficients. At higher coefficients, however,
a constant elevator deflection triggered a rapid pitching of the aircraft. At
the start of the pitch-up, the stick force lightened and the pilot reversed the
elevator control in an attempt to stop the pitch-up. This was unsuccessful,
and both the angle of attack and normal-force coefficient increased until a
recovery was made.
In an effort to control the pitch-up, an outboard wing fence was added to
both wings. Jones made the first flight on October 18, 1951, with the wing
fences, at Mach 0.7. A second flight was made on November 9 at Mach 0.95.
The pitch-up flights were halted until the following summer.22
The all-rocket second D-558-2 required more extensive changes before it
returned to the air. It was sent back to the Douglas plant in El Segundo in early
1950 and disassembled. The two jet fuel tanks were removed, and, in their
place, a liquid-oxygen tank and an alcohol/water tank were added. This almost
doubled the fuel supply to 345 gallons of liquid oxygen and 378 gallons of
alcohol/water. The jet engine was removed, and its exhaust in the lower aft fuse-
lage was faired over. Air inlets were also removed and flush panels added. The
LR-8 rocket engine and its turbopump were mounted in the aft fuselage. The
result was an aircraft optimized for high-speed/high-altitude research flights.23
On November 8, 1950, the second D-558-2 was loaded under the P2B-1S
mother ship and flown back to what was now Edwards Air Force Base after
Muroc had been renamed. As with the third D-558-2, Bridgeman would
make the initial flights with the second one before it was turned over to the
NACA for research work. This proved difficult. Several attempts were aborted
97
Probing the Sky
A jet-powered D-558-2 parked on the ramp at South Base in 1949. The jet intakes are located
low on the forward fuselage, and the exhaust is on the underside of the rear fuselage. The
aircraft has been fitted with wing fences, and the wing slats are extended. The original NACA
shield appears on the vertical tail but was subsequently replaced by a yellow band with a winged
NACA insignia. (NASA)
shortly before launch. At one point, a sign appeared on the hangar wall: “Old
Skyrockets Never Die—They Just Jettison Away.” Bridgeman later said, “An
experimental test pilot has to be right on the edge, at the peak of his perfor-
mance; and the delay was sapping this capacity.”24 Finally, on January 26, 1951,
the Douglas team’s luck seemed about to change. Then, as the countdown
reached less than a minute to launch, Bridgeman saw that a rocket pressure
gauge was dropping. Bridgeman radioed the P2B-1S’s pilot, George Jansen:
“No drop. This is an abort,” and began shutting down the systems. Bridgeman
was startled to hear Jansen begin the 10-second countdown. Bridgeman radi-
oed: “Don’t drop me, George!” but Jansen had his thumb on the microphone
key and could not hear Bridgeman’s calls.
Bridgeman frantically pushed the circuit breakers back into place and
centered the control stick, hoping he had not missed anything. When the
launch count reached zero, the Skyrocket separated; once clear of the bomber,
Bridgeman fired the four rockets in rapid sequence. They ignited despite the
pressure readings, and the Skyrocket accelerated upward.
By now, 35 seconds had passed since he had been dropped, and Bridgeman
made his first radio call: “George, I told you not to drop me!” Colonel Frank
“Pete” Everest, the Air Force chase pilot, replied, “You got keen friends,
98
Bridgeman.” Everest added, “He’s accelerating away from me in a climb. Looks

like he’s doing all right…all four rockets appear to have lit off.” Because of
the higher speed, the elevators were ineffective and Bridgeman had to use the
movable stabilizers for pitch control. The peak speed reached was Mach 1.28
in a slight dive.25
To prevent a repetition, a green light was rigged up in the P2B-1S cockpit to
indicate that the D-558-2 was ready for launch, to be triggered by the research
pilot in the test vehicle when he was ready in case there was another sticky
microphone. The launch countdown was also shortened. It was not until April
5, 1951, that the second Douglas flight was made, reaching a peak speed of
Mach 1.36 at 46,500 feet. Bridgeman again experienced severe lateral oscilla-
tions, forcing him to shut off all four chambers of the engine before burnout.
The fix for the oscillations was to add a rudder lock, which prevented it from
moving at speeds above Mach 1. The all-rocket D-558-2 had been proven
airworthy and had proven that an air launch was feasible. The Navy, Douglas,
and Bridgeman now began a series of maximum-speed and -altitude flights.
Bridgeman began the speed buildup on May 18, 1951. After launch, he
climbed to about 55,000 feet and made a –0.8-g pushover to begin the speed
run. He reached a maximum speed of Mach 1.72 at 62,000 feet before the
engine shut down. This was followed on June 11 with a flight that reached
Mach 1.79. Bridgeman was now the fastest man on Earth. Both these flights
exceeded the most optimistic performance estimates made by Douglas for an
all-rocket/air-launched Skyrocket.
The Douglas engineers decided to try for even faster speeds. Doing this,
however, required that a lower 0.25-g pushover be made. The flight was sched-
uled for June 23. Bridgeman was successfully launched; he ignited the rocket
and began the climb. Reaching 60,000 feet, he made the 0.25-g pushover and
accelerated in a shallow dive. When the aircraft reached Mach 1.5 it began
rolling violently, throwing Bridgeman from side to side as the wings rocked
back and forth as much as 70° in less than a second. Bridgeman’s attempts to
bring the aircraft back under control actually aggravated the problem, and, as
a last resort, he shut the engine down.
But instead of calming, the rocking motion increased. Bridgeman pulled
back on the control wheel and made a 4-g pullout. Now headed toward the
safety of the lakebed and accompanied by the F-86 chase plane, he made a
successful landing. NACA engineers determined that the aircraft had reached
a top speed of Mach 1.85 at 63,000 feet. This represented a new speed record
for a rocket-powered aircraft and was well beyond what both the NACA and
the Navy had originally envisioned as the upper-end of D-558-2 performance.
But NACA managers were troubled by the continuation of the Douglas
flight test project. The all-rocket aircraft had been proven airworthy, and NACA
99
Probing the Sky
An in-flight photo of a Skyrocket descending toward the lakebed with an F-86 chase plane
following behind. Chase planes were key elements in undertaking research flights safely. The
chase pilot provides an external set of eyes to warn of problems and calls out the research
airplane’s altitude above the lakebed. (NASA)
researchers did not view record flights favorably. Speed and altitude records
set during research flights were not significant. When setting records became
the reason for a flight, airplanes crashed and pilots died. Research aircraft, their
view held, were for use in collecting data for future aircraft development, and
it was time—even beyond time—for the agency to have received the aircraft
back for its own extensive research. Ultimately, the Navy agreed, but only after
signing off on two more Douglas flights.
The first of these came on August 7, 1951, as a maximum-speed flight. To
avoid the instability experienced on previous flights, the initial pushover load
factor was 0.8 g. Bridgeman then reduced this to 0.6 g. The aircraft became
left-wing-heavy, however, and his attempt to correct this with the ailerons was
unsuccessful, so he raised the loading back to 0.8 g, which restored lateral sta-
bility. Bridgeman then reduced it to 0.6 g until the rocket engine shut down.
When the data reduction was complete, engineers found that the Skyrocket
had reached Mach 1.87, with a possible error of ±0.05, at a pressure altitude
of 67,300 feet. The maximum airspeed was calculated to be 1,243 mph ± 33
mph. Bridgeman believed the aircraft capable of even higher speeds, stating,
“[M]agic Mach 2 is attainable if lateral control can be maintained.”26
100
Bridgeman’s final Skyrocket flight was on August 15, a maximum-altitude

flight. Charles Pettingall, chief aerodynamicist in the Douglas testing division,
calculated the flightpath this time. It was designed to exceed the 72,395-foot
altitude record set by the U.S. Army Air Corps–National Geographic Society
Explorer II piloted balloon flight in 1935. After launch, Bridgeman ignited
the rocket and reached a peak speed of Mach 1.35, slower than the maximum
speeds attained in earlier launches. Passing through 63,000 feet, the D-558-2
began rolling to the left. Even with full opposite aileron input, the airplane was
slow to recover. The rocket engine shut down, and the aircraft coasted upward
under its own momentum. Bridgeman described the view from an altitude
no other human had reached: “Out of the tiny window slits there is the earth,
whipped clean of civilization, a vast relief map with papier-mâché mountains
and mirrored lakes and seas. The desert is not the same desert I have seen for
two years. The coastline is sharply drawn with little vacant bays and inlets, a
lacy edge to the big brown pieces of earth.”27
When calibration corrections were made to the air pressure data, results
indicated that the Skyrocket had reached a pressure altitude of 77,500 feet ±
500 feet. But the radar data had to be corrected for errors in the radar slant
range and elevation angle, as well as for such factors as beacon delay, atmo-
spheric refraction, and Earth’s curvature. Once done, the Skyrocket’s altitude
above sea level was shown to have been 79,500 feet ± 65 feet. Bridgeman was
now both the fastest and highest-flying man on Earth.28
Beyond setting the new speed and altitude records, the Douglas demonstra-
tion flights significantly increased knowledge of the D-558-2 flying qualities
and handling. On one of the first supersonic flights, an uncontrollable lateral
oscillation occurred, forcing the pilot to abandon the speed run to Mach 1.4.
The magnitude of the oscillations was inversely related to the angle of attack:
the lower the angle of attack, the more violent the oscillations. At higher
speeds, such as on the Mach 1.85 flight, a different problem appeared. If the
pilot made the pushover to begin the speed runs at too low a g-force, severe
oscillations were triggered.29
The all-rocket second D-558-2 was turned over to the NACA on August
31, 1951, and given the designation NACA 144. Research flights with both
of the Skyrockets could now begin.30
The First NACA Supersonic Skyrocket Research Flights

There were two different Skyrocket configurations in NACA hands: the second
D-558-2/NACA 144, with its all-rocket configuration, and the third D-558-2/
NACA 145, with a jet-rocket propulsion system. This dictated that each plane
101
Probing the Sky
The D-558-2 was not only used for aerodynamic research, but also for the collection of opera-
tionally oriented data. The third Skyrocket, NACA 145, in which the jet/rocket propulsion system
was retained, was fitted with pylons and simulated bombs to test their effects. (NASA)
would be used for different types of research. Once the Douglas speed- and alti-
tude-record flights had been completed, NACA engineers decided to explore
the second Skyrocket’s operational limitations, with Crossfield as pilot. He had
his first pilot check flight in the new aircraft on September 28, 1951, during
which he experienced rough engine operation but still reached Mach 1.2.
He made three more flights in October and November. The first two flights
provided data on longitudinal and lateral stability and control, loads, and aile-
ron effectiveness at low supersonic speeds (Mach 1.28 and Mach 1.11). The
final flight, on November 16, 1951, saw Crossfield reach a speed of Mach 1.65
at 60,000 feet. The winter rains closed the lakebed until the spring of 1952.
In contrast, the third Skyrocket, with its dual jet-rocket propulsion system,
had a much more limited performance capability and thus was more suitable
for transonic research, exploring such issues as swept-wing pitch-up. Once the
lakebed dried out, these flights resumed on June 19, 1952, with Crossfield
and Jones as project pilots.31 The initial flights were made to test aircraft
response with an inboard and outboard fence on each wing. The first flight,
by Crossfield, reached a speed of Mach 0.7. Subsequent flights were made at
Mach 0.96. Four flights were made in this configuration, with one aborted
due to a failed cockpit heater. The last flight in the series, on August 14, was
made with the inboard wing fence removed.32
102
A 1951 photo showing one of a number of wing-fence configurations tested to counter the pitch-
up caused by swept wings. The inboard fence is mounted on top of the wing, and the outboard
fence extends around the wing leading edge. (NASA)
High-Speed Flight Research Station engineers Jack Fischel and Jack Nugent
summarized the effectiveness of the wing fences in a research memorandum.
The tests were made with the Skyrocket in a clean configuration in turning
flight at speeds of Mach 0.5 to 0.96, at altitudes between 19,000 and 36,000
feet. The flights were made both with the original wing configuration and with
the addition of an outboard fence on each wing since wind tunnel testing had
indicated that this fence would alleviate the pitch-up problem.
Fischel and Nugent found the flight results less clear-cut than wind tunnel
data had suggested they would be. They concluded: “The addition of wing
fences appeared to provide only a slight improvement over the original con-
figuration, inasmuch as the pitch-up occurred at only slightly higher values of
normal-force coefficient for the modified airplane configuration.”33
Their analysis also included the pilots’ impressions, which were quite blunt:
In the pilots’ opinion, the airplane is uncontrollable for a range

of normal acceleration of about 1 g and 1½ g above the value at
which the reported change in stability occurs; this behavior is
very objectionable. At low speeds, if the pilot does not check the
pitch-up by use of the elevator as soon as it is noticed, the angle
103
Probing the Sky
of attack increases rapidly and violent rolling and yawing motions

are experienced at large values of [angle of attack]. At high speeds
the pitch-up appeared to be more severe and more abrupt.
Throughout the speed range covered, the occurrence of a reduc-
tion in stick-free stability, almost simultaneously with the reduction
in stick-fixed stability, tended to accentuate the pitch-up to the
pilot. The pilot felt that even with improved control, as would
result from an all-movable tail, flight above the stability bound-
ary would not be sufficiently steady for gunnery or other precise
maneuvering.”34
For the next series of flights, engineers tested several different wing configura-
tions. The goal was to find a means of preventing instability and pitch-ups during
accelerated longitudinal maneuvers at speeds of up to Mach 1 and altitudes of
10,000 and 35,000 feet. The aircraft was flown with wing slats fully extended,
both with and without inboard wing fences, and, finally, with the wing slats half
extended and wing fences removed. Another change made, in order to improve
the stick-force characteristics at moderate and large angles of attack, was to attach
two bungee cords to the control column.
Crossfield made the initial flight on October 8, 1952, at a speed of Mach
0.97. During the flight, and while performing a turn at high speed, the
Skyrocket pitched up 36° followed by the now-anticipated sharp roll-off and
near loss of control. The final pitch-up research flight for the year was made
on October 22, 1952, also by Crossfield. In this instance, the aircraft was
configured without wing fences. During the flight, pitch-up occurred when
the aircraft made turns.35
The results from the extensive series of tests were summed up in a 1954
research memorandum written by Fischel:
Opening the wing slats to the fully extended position improved

the stability characteristics of the airplane by alleviating pitch-
up at Mach numbers of below approximately 0.8; however, at
Mach numbers between 0.80 and 0.85 the severity of the pitch-up
remained unaltered. At Mach numbers of about 0.98 and 1.00,
maneuvers performed up to relatively high values of normal-
force coefficient with slats fully extended exhibited no evidence
of pitch-up; however, this effect has since been duplicated with
the clean-wing configuration (no fences, slats retracted).
Removing the wing fences from the airplane configuration
with slats fully extended caused the reduction in stick-fixed
stability to become slightly more pronounced, and generally,
104
to occur at approximately the same or slightly lower values of

normal-force coefficient.
With wing slats half extended and no wing fences, the airplane
exhibited instability characteristics and pitch-up similar to that
exhibited by the airplane with slats retracted and with wing fences.
With slats fully extended and wing fences removed, use of a
bungee in the control system to alleviate or eliminate the stick-free
instability caused the airplane to appear more controllable to the
pilot and caused the decrease in stick-fixed stability to become
less apparent and less objectionable.36
Fischel also described the pilots’ impressions of the aircraft’s behaviors:
In general, the pilots’ reports corroborated the data and conclu-

sions reached for the maneuvers performed. With the slats fully
extended at all Mach numbers below M ≈ 0.8, it is the pilot’s
opinion that the airplane stability did not deteriorate appreciably
after the initial decay, and as a consequence control was regained
more rapidly than in the original slats-retracted configuration. At
M ≈ 0.98 (wing fences on) and M ≈ 1.0 (wing fences removed),
the airplane appeared controllable up to the maximum value of
[angle of attack] attained. In both fence configurations with the
slats fully extended the stability change most apparent to the pilot
was lightening of stick forces at moderate angles of attack. The
pilots reported a stick-fixed stability change at moderate angles of
attack which became somewhat more apparent when the inboard
wing fences were removed from the slats fully extended configura-
tion. The pilots thought that the airplane configuration with slats
fully extended were [sic] a definite improvement over the airplane
configurations flown with the slats retracted.
Because the soft bungee had little or no effect on the stick
forces, the character of the stick-free and stick-fixed instability of
the airplane appeared to the pilot to be about the same as when
no bungee was used. In both instances, the lightening of the stick
forces at moderate angles of attack tended to increase the control
rate, which in turn would aggravate any pitching. With the stiff
bungee, however, the character of the decay in stability appeared
much improved, for now the stick-free stability was improved and
the airplane appeared to have a lower pitch divergence rate than
previously; thus, the change in stick-fixed stability was somewhat
less apparent and less objectionable.
105
Probing the Sky
In the configuration with slats half extended, the pilot thought

the airplane behavior was similar to that encountered with slats
retracted, and the pitch-up encountered was equally uncontrol-
lable. Although data obtained in this configuration were limited
to two high-speed maneuvers, the pilot reported that pitch-up was
also encountered in other maneuvers performed at lower speeds
(down to M < 0.7), at lower values of normal acceleration as the
Mach number was decreased.
…If the stick-fixed stability is made acceptable, the provision
of a bob weight, bungee, or artificial feel system to supply more
satisfactory stick-force characteristics would be desirable.37
When pitch-up flights resumed in February 1953, with Crossfield as the

sole project pilot, a new wing modification had been made to the airplane.38
Wind tunnel tests indicated that a chord increase on the outer 32 percent of
the wing might eliminate the tendency to pitch up under high lift conditions
at around Mach 1.2. During the winter of 1952–53, new outer wing panels
were added, effectively giving the Skyrocket a “sawtooth” leading edge.
The first flight with the chord extensions was made on February 27, 1953,
in a jet-only flight. Crossfield made windup turns and 1-g stalls. The maneuvers
were halted when longitudinal or lateral instability occurred. Additional tests of
the modified wing shape were made on April 8 and 10 at speeds between Mach
0.45 and Mach 1.0 and altitudes from 18,000 to 34,000 feet. Crossfield made
windup turns, aileron rolls, sideslips, and 1-g stalls during the flights. The tests were
repeated on the final flight of the series, on June 15, bringing the tests to a close.
Jack Fischel and Cyril D. Brunn subsequently wrote a research memoran-
dum on the chord extensions. Their summery concluded:
Addition of wing chord-extensions had only a minor effect on the

decay in stick-fixed stability (pitch-up) and stick-free stability expe-
rienced by the airplane at moderate angles of attack. The chord-
extensions alleviated the pitch-up to a small degree, but the pilot
still considered the airplane unsatisfactory for controlled acceler-
ated flight in this region. However, at higher angles of attack, the
airplane appeared to retrim and regain some stability….
The buffeting of the airplane was of such a nature that the
increase in buffeting intensity induced by the chord-extensions
became a major problem, in addition to the longitudinal instabil-
ity problem.
A comparison of wind tunnel with flight data showed good
agreement in the reduction of stability evident at moderate angles
106
One approach tested on the D-558-2 was that of a sawtooth chord extension on the outer lead-
ing edge designed to prevent spanwise airflow across the wing. (NASA)
of attack. The results indicate that an abrupt reduction of stabil-

ity to a region of neutral or even slight stability could tend to
cause pitch-up.39
Following the April 10, 1953, flight, the chord extensions were removed
from the aircraft and the slats were reinstalled and locked in an open position.
The pitch-up research flights resumed on June 15, using both the jet and rocket
engines. Accelerated longitudinal-stability maneuvers began in earnest, and the
aircraft showed decay in stability at all speeds except Mach 1. Following the
flight, a stiff bungee was installed on the control stick. This change was tested
on the June 25 jet-only flight and showed improvement, sufficient at least to
make the Skyrocket controllable at high angles of attack. The reduction in
stability was judged by Crossfield to be less objectionable. After this flight, the
aircraft was restored to its basic configuration.
Crossfield made five more flights between September 9 and December 22,
1953, in the third D-558-2, with both the jet and rocket propulsion systems.
These looked at lateral, longitudinal, and directional stability and control at
transonic speeds (Mach 0.4 to Mach 1.08), as well as turns and stalls. Two
of the flights suffered rocket malfunctions; on the September 22 flight only
two rocket cylinders fired, and on the final flight the rockets failed to ignite
107
Probing the Sky
altogether. This completed pitch-up flight testing. After that final attempt, the
group terminated tests with that aircraft.
A summation of the different wing modifications applied to the Skyrocket
was published in 1956. The research memorandum noted:
None of the wing modifications had an appreciable effect on the

decay in stick-fixed stability (pitch-up) exhibited by the airplane
at moderate angles of attack, particularly over a Mach number
range from about 0.8 to 0.95. All configurations were consid-
ered unsatisfactory and uncontrollable in the pitch-up region by
the pilots. On the basis of these tests and other flight and tun-
nel investigations, it is felt the position of the horizontal tail on
this airplane should be lowered appreciably to obtain substantial
improvements in longitudinal handling qualities.
Wing fences had no apparent effects on the buffeting char-
acteristics with slats retracted; however, unlocking the wing slats
raised the buffet boundary, below a Mach number of 0.70, above
that for the retracted slats condition for the basic-wing, one-fence,
and two-fence configurations. Wing chord-extensions lowered the
buffet boundary, compared with unmodified airplane configura-
tion, up to a Mach number of 0.80 and caused an increase in
buffet intensity which was objectionable to the pilot. Moderate
buffeting appeared to exist over most of the lower and moderate
lift range with the slats fully extended; however, this configuration
did alleviate some of the pitch-up divergent rate and appeared to
the pilots to provide the greatest improvement in the longitudinal
handling characteristics of the airplane.40
None of the wing modifications had an appreciable effect on
the trim-stability characteristics of the airplane and all configu-
rations exhibited similar trends over the Mach number range.
The airplane was stable at Mach numbers below about 0.82, and
exhibited characteristic nose-down and nose-up trim changes
between Mach numbers of about 0.87 and 1.03.41
Clearly, the various wing modifications failed to solve the pitch-up issue.
For all the advantages of sweptback wings, pitch-up represented a major prob-
lem that required a solution. This was found not exclusively in modifying the
wings, but rather in moving the location of the horizontal tail. The original
placement of the horizontal stabilizer midway up on the D-558-2’s vertical
tail was to avoid the wing wake. This, however, placed the stabilizer in the
downwash from the wings when the pilot pulled back on the stick, initiating
108
a sharp pull-up maneuver that increased the aircraft’s angle of attack relative
to the airflow around it. The report continued:
On the basis of wind-tunnel tests performed on a model of

the D-558-2 airplane, as well as other wind-tunnel and flight
investigations, it has been concluded that with the present tail
configuration of the D-558-2 airplane, a real cure of the pitch-
up is not feasible. Lowering the horizontal tail to approximately
the height of the wing-chord plane extended would be required
to obtain substantial improvement in airplane longitudinal
handling qualities.42
The combined results of the D-558-2 research flights and wind tunnel test-
ing altered subsequent aircraft designs. The first generation of jet fighters had
horizontal stabilizers located at the base of the vertical fin. Later aircraft had
their stabilizers located on the lower aft fuselage, even with or below the wing
centerline. This cured the pitch-up issue, and it is why the low-placed horizon-
tal tail is a standard feature of transonic and supersonic combat aircraft design.
Exploring Operational Limits, Lateral

Stability, and Vertical-Tail Loads
At the same time that the third D-558-2, NACA 145, was making its pitch-
up research flights, the all-rocket second D-558-2, NACA 144, was also
flying research missions. After the latter had been turned over to the NACA,
Crossfield had made a series of flights to explore its operational limits. Once
these were completed, NACA researchers undertook a new effort focused on
the aircraft’s lateral stability and control at high Mach numbers. This had been
the cause of the violent rolling motions experienced by Bridgeman during the
Douglas contractor flights in 1951. Additionally, loads on the vertical tail were
also to be recorded.
Flights in the new research effort began on June 13, 1952. The first of these
reached a speed of Mach 1.36. Two more flights were made in June, reaching
low supersonic speeds of Mach 1.05 and Mach 1.35. The July 10 flight reached
Mach 1.68 at 55,000 feet, the peak speed for this initial test series. As with
the other Skyrocket, the flights did not always go smoothly. The July 15 flight
reached only Mach 1.05 because of an engine malfunction, and the August
13 flight was aborted after launch from the P2B-1S when a liquid-oxygen
valve stuck in the open position. The aircraft did not return to flight until
October 10, 1952. This flight measured longitudinal stability at speeds of up
109
Probing the Sky
The second D-558-2, NACA 144, parked on the ramp at South Base. Modifications made to
allow air launch included the removal of the jet engine and fuel tanks. Between the air launch
and the increased rocket burn time, the craft’s maximum speed was doubled. On November 20,
1953, this aircraft made the first Mach 2 flight. (NASA)
The modified B-29 used to launch the Skyrockets had striking nose art. This included its nick-
name, “Fertile Myrtle,” and a scoreboard that tallied the number of launches made with each of
the D-558-2s. Each of the launch aircraft was used for a single rocket plane, as modifications
made to the mother ships’ underside were specific to each aircraft type. Fertile Myrtle survived
damage it sustained when the no. 4 propeller tore free. (NASA)
110
to Mach 1.65. The year’s activities ended with a flight on October 23, and the
research effort was then placed on hold until the following spring. The first
flight of 1953, made on March 26 by Crossfield, brought the longitudinal-
stability flights to a close.
The level of D-558-2 flight activity was lower than that of any new jet
fighter undergoing testing at Edwards. This highlighted the difference between
specialized flight research and production and acquisition-oriented test opera-
tions. Extensive effort went into planning for each research flight. Research
aircraft also required extensive ground testing before a flight, including careful
calibration of instrumentation. Once the flight was completed, malfunctions
had to be identified, understood, and corrected before the next flight was
attempted. Onboard systems had to be inspected and tested after the flight
was completed. Finally, the data collected during the flight had to be reduced
and then analyzed to assess results, detect any problems, and identify data
points that would have to be repeated or would be used in preparing the next
flight’s plan.
The next phase in research activity with the second D-558-2 was a series of
flights to collect data on supersonic lateral stability at low and moderate angles
of attack, in both pushover and straight flight, at altitudes between 50,000
and 70,000 feet. Crossfield made six flights between April 2 and August 5,
1953. One of the flights, on June 18, was aborted after launch due to a rough
running rocket engine. During the August 5 flight, Crossfield reached a speed
of Mach 1.878, just below Bridgeman’s maximum speed flight.43
The research memorandum noted:
The first flights to Mach numbers near 1.8 were performed by

climbing the airplane to altitudes of about 55,000 feet and push-
ing over to nearly zero lift (angle of attack ≈ –2°) with the rudder
locked. During these flights, violent lateral oscillations occurred,
with side-slip angles reaching ±6° and roll angles reaching about
±60°…. It was decided that the low angle of attack, and conse-
quent low inclination of the principal axis of inertia with respect
to the flight path, may have aggravated the motion. A subsequent
flight was made in which the pilot did not push over to low lift
but maintained an angle of attack greater than 0°. In this condi-
tion, the lateral oscillation was much less pronounced. It should
be noted that the two flights were made at different altitudes and
at different constant angles of attack.44
111
Probing the Sky
The memorandum also noted the Skyrocket pilots’ experiences:
During one of the flights with the rudder locked, the pilot
attempted to hold the ailerons fixed for a period of time in order
to determine the lateral motion that would ensue…. The airplane
tended to roll off at a fairly rapid rate until the roll angle reached
about 90°, at which time the pilot stopped the motion with the
ailerons. During another flight at a low angle of attack with the
rudder locked…the lateral motion was of such a nature that the
airplane rolled nearly 140° although the pilot was trying to control
the airplane with the ailerons. At one time during the flight when
the pilot felt the controls were ineffective in stopping the rolling
motion…the pilot reversed the aileron control in order to make
the airplane complete the roll to 360° in order to recover from the
inverted attitude. This action was also ineffective as the airplane
at the same time apparently began to recover of its own accord
against the control motion supplied by the pilot. It is pertinent to
point out here that this condition was not caused by a complete
lack of aileron effectiveness. Unpublished flight data indicate that
aileron effectiveness at these speeds is indeed low but that an
appreciable amount still remains….
The motions obtained during these two flights suggest the
possibility that the control-fixed transient oscillation at high
supersonic speeds at low angles of attack may be one in which
the roll angles reach values in excess of 90°, whereas the transient
sideslip angles remain at relatively low values. However, because
of the difficulty of flying the airplane at these speeds and angles
of attack, it has not been possible to check this hypothesis further
in flight.45
The Road to Mach 2

Crossfield’s flights were approaching twice the speed of sound. Though reach-
ing Mach 2 was the next landmark in the quest for higher altitudes and faster
speeds, it did not have the significance of breaking the sound barrier. Unlike
the transition from transonic to supersonic flow, no fundamentally differ-
ent aerodynamic phenomena appeared going from Mach 1.88 to Mach 2.0.
The D-558-2 had been reaching speeds of about Mach 1.8 to 1.9 during
both Bridgeman’s and Crossfield’s flights made between 1951 and 1953. But
112
although a Mach 2 flight seemed to be only an incremental increase over earlier

flights, it was, in fact, far more difficult, and, if nothing else, it had tremendous
symbolic importance as the next important milestone on a path that, increas-
ingly, many saw as eventually leading to flight into space.
Crossfield noted that “even at Mach 1.8 we were pressing [the Skyrocket]
far beyond rational limits.”46 That the NACA team could even reach Mach
1.8 was due to experience gathered in earlier flights. Recalled Crossfield, “The
plane was by now almost completely debugged.”47 The team had also learned
the “many little tricks to save time and gain an edge on the unknown.”48 Even
so, Crossfield said, “the best any ordinary team could hope for, with luck, was
a speed of Mach 1.9.”49
A second factor working against the prospect of a Mach 2 flight was the
traditional NACA mindset. NACA flights were made to collect data, not set
speed records. Crossfield’s flights in the D-558-2 had exceeded the existing
world speed record, but the NACA had not highlighted the achievement in
any public announcements. The research memorandums on the flights were
classified “Confidential” and were limited to distribution among the NACA,
contractors, and the military. Most important, Hugh L. Dryden had told
Crossfield not to attempt a Mach 2 flight.
But much larger issues were at hand than whether research for its own sake
was justified. The cost of research activities was going up, and construction of
new, more advanced facilities was required to explore new realms of flight. The
NACA required political support for this, as money was short and the new
Eisenhower administration, elected in November 1952, was skeptical about
devoting scarce resources into research activities at a time when the Nation
was at war in Korea and facing extremely serious global challenges in Europe
and Asia.50
The NACA contributions were highly technical ones, easily appreciated
by the science and engineering community but not so readily discerned by
others outside those fields. What others could understand, however, was a new
world airspeed record. Aviation had always been about flying faster, higher,
and farther. In the 1930s, air races had been major spectator events. In both
commercial activities and military aviation, better aircraft performance was
equated with better aircraft. Additionally, aircraft technology continued to
undergo rapid change in the early 1950s. In late 1953, the F-80 Shooting Star,
the hottest American airplane in 1945, was being phased out of service; the
F-86 Sabre was approaching old age; the supersonic F-100 Super Sabre was
beginning production; and the Mach 2 F-104 was in advanced development.51
So Hugh Dryden gave Walt Williams approval for a single attempt at a
Mach 2 flight. Exceeding Mach 2 in the Skyrocket would require careful plan-
ning and special preparations. Herman O. Ankenbruck, the project engineer,
113
Probing the Sky
Aft view of NACA 144. The four nozzles of the LR-8 rocket engine are visible. This engine was
essentially the same one used in the X-1 series of aircraft. (NASA)
developed a flight plan that would offer the best chance of reaching the goal.
Success would depend on Crossfield’s ability to fly the plan, the thrust of the
rocket engine, and the amount of fuel that could be carried.
None of this would prove easy. Lieutenant Colonel Marion Carl, a superb
Marine test pilot, had made two attempts in the summer of 1953 to reach
Mach 2 and fallen short both times. Crossfield noted that “the slightest over-
pressure on the stick would cut the speed back drastically.” Crossfield did have
an advantage over Carl, however. The D-558-2 had been fitted with cone-like
nozzle extensions following Carl’s Mach 2 attempts, to increase thrust at high
altitude. These allowed Crossfield to reach Mach 1.96 during an October 14,
1953, flight, on the sixth anniversary of the first supersonic flight by Chuck
Yeager in the first XS-1.52
The nozzle extensions were critical to Crossfield’s eventual success in reach-
ing Mach 2. The reason literally came down to rocket science. High-pressure
gas in the LR-8 combustion chambers expanded out the nozzles to produce
an equal and opposite force, which accelerated the D-558-2. As the exhaust
gas left the nozzles, it expanded and its pressure dropped. In an ideal expan-
sion, exhaust pressure drops to that of the outside atmospheric pressure, but
no lower. These conditions would achieve maximum thrust. But if the exhaust
gas over-expanded, it created drag.
114
Because atmospheric pressure varied according to the altitude at which the

D-558-2 was flying, the extended nozzle was shorter than the length needed
for ideal expansion. The excess energy was lost, as the pressure was not fully
recovered. If Crossfield were to reach Mach 2, however, he would have to fly
at a higher altitude than that for which the rocket nozzles had originally been
designed. Moreover, he would need to get every bit of thrust he could from
the rocket. He needed the larger, extended nozzles to capture more of the ideal
expansion of exhaust gas.53
This was not the only advantage that would be needed to reach Mach 2.
The second D-558-2 underwent special preparation. To minimize drag, the
ground crew sanded and polished the aircraft, and every panel seam was taped
over. The two stainless-steel fuel-jettison tubes were replaced with aluminum
tubes, which were bent into the rocket’s exhaust. Once the rockets fired, the
tubes were no longer needed and
would burn off, reducing drag and
weight by a small amount. But
even with the reductions in drag
and weight, a longer rocket burn
time would also be needed. To
enable this, the water/alcohol fuel
would be cold-soaked in a refrig-
erator the night before the flight,
increasing the amount that could
be carried by 10 or 15 gallons.
The liquid oxygen was also loaded
into the aircraft the night before,
cold-soaking the tank and airframe From left are Walter C. Williams, High-Speed
Flight Research Station director; A. Scott
around it, which enabled the crew Crossfield, D-558-2 pilot; and Joe Vensel, HSFRS
to add a little more oxidizer. Some official in charge of operations, in front of a
of the liquid oxygen would boil off D-558-2, circa 1953. (NASA)
during ground preparations and
the flight under the P2B-1S to the launch point, they knew; such was always
the case. Just before the drop, the plane’s liquid-oxygen tank would be topped
off from a tank on the P2B-1S, replacing the lost oxidizer.
The attempt was scheduled for November 20, 1953. Crossfield was not
in the best condition. He had a bad case of the flu but was determined to fly.
With the Skyrocket’s preflight preparations completed, the P2B-1S launch
aircraft took off. Stanley Butchart piloted the P2B-1S and was in charge of a
precise sequence of events. Each of the rocket planes had an individual launch
area. The second D-558-2 (NACA 144) was typically launched over Lake
Elizabeth. The jet-and-rocket-powered third D-558-2 was usually launched
115
Probing the Sky
The rocket-powered X-planes underwent regular engine test firings. The shock diamonds in the
rocket exhaust are visible. Engineers monitoring the test are standing close to the D-558-2. In
the early 1950s, ear protectors, blast shields, and similar safety measures were not required.
The panels attached to the horizontal stabilizers were there to prevent any damage. (NASA)
west of Rosamond. The X-1s were mostly launched over Victorville, southeast
of Edwards.
Butchart later described the Skyrocket launch procedures. After takeoff, he
would head out over Big Bear Lake in the long, slow climb to launch altitude.
With no control room then in use, the launch plane pilot directed the opera-
tion. The first step in the sequence came after about 20 to 30 minutes, when
Butchart called for fire trucks to be deployed on the lakebed. This was followed
a few minutes later by a call for the chase planes to take off and join up with the
P2B-1S. When the launch plane reached 10,000 feet, Crossfield climbed into
the D-558-2, closed the canopy, and began launch preparations. The maximum
power of the P2B-1S engines was used for the climb, for which an hour to an
hour and a half was allotted. This put a heavy strain on the engines. Once the
launch altitude was reached, Butchart began the final 6-minute sequence before
launch. He would then fly out for 2 minutes, make a turn lasting 2 minutes,
and take the final 2 minutes to return to the launch point. The release would
then be pulled.
On November 20, 1953, Crossfield was successfully launched from the
P2B-1S, and he followed the flightpath Ankenbruck had calculated. Reaching
72,000 feet, he gently nosed over into a shallow dive, reaching the maximum
Mach number at 62,000 feet. In doing so, Crossfield avoided too low a g-force
116
on the aircraft, which had caused the uncontrolled rolls Bridgeman had experi-
enced 2 years before. Following burnout, he made a sweeping approach to the
lakebed for a deadstick landing, the usual chase plane trailing behind.
Flight data indicated that Crossfield had reached a speed of Mach 2.005 ±
0.02. Calculated in a standard atmosphere, the true airspeed was determined
to be 1,327 miles per hour. However, weather balloon data indicated the air
temperature was 22° lower than that of the standard atmosphere. With this
correction, the D-558-2 had reached 1,291 miles per hour ± 17 miles per hour.
This was the highest speed yet reached in the Skystreak, and no subsequent
attempts were made to exceed Mach 2. Crossfield later recalled that it was
decided that the extraordinary efforts needed to exceed Mach 2 would not be
justified by the value of the data.54
After the brief interlude of the Mach 2 flight, research with the second
D-558-2 resumed. For the next 3 years, engineers and pilots put the aircraft
through a wide range of research activities. These included examinations of
dynamic stability and control, structural temperature and loads, wing and tail
pressure distribution, buffeting data, and vertical-tail loads.
Though Crossfield made most of those flights, he left the NACA High-
Speed Flight Station in the late summer of 1955 to join North American
A D-558-2 being loaded into its launch aircraft. The B-29 was lifted up on jacks inside a hangar,
the Skyrocket was rolled underneath, and the launch plane was then lowered. Executing the
process inside a hangar eliminated problems posed by desert winds. (NASA)
117
Probing the Sky
Aviation as a technical adviser and eventually became a corporate test pilot on

the X-15 program.55 His replacements on the Skyrocket were NACA research
pilots Joseph A. Walker and John B. “Jack” McKay. And while the D-558-2
research flights had been an ongoing NACA project since 1949, they were never
routine. On McKay’s pilot-familiarization flight, made on September 16, 1955,
the landing gear failed to extend, and he had to use the emergency hydraulic
system to lower the wheels.56
McKay also experienced a far more dangerous situation that nearly resulted in
the loss of his own life, the second D-558-2, the P2B-1S launch aircraft, and the
P2B-1S crew. On March 22, 1956, the P2B-1S had reached drop altitude when
its number 4 engine abruptly failed. The flight engineer reported the problem
to Butchart, who responded by feathering the propeller, that is, adjusting the
pitch of the blades so that it would stop and present a minimal-drag profile to
the airflow. The propeller slowed to a stop but then unfeathered and began spin-
ning again. Butchart feathered the propeller again and again, but it continued
to unfeather and spin back up. At this speed and altitude, the propeller was
spinning so fast that centrifugal force was inevitably going to cause the blades to
disintegrate. On his third try, Butchart was still unable to feather the propeller.
(He later recalled that the airplane’s manual warned that there was only enough
hydraulic fluid for three in-flight feathering attempts.) At that point, McKay
called out that he had a problem with the Skyrocket and directed the P2B-1S
crew not to drop him.
But Butchart and P2B-1S copilot Neil Armstrong had no choice, as the
propeller would soon disintegrate. They nosed the P2B-1S down, and Butchart
pulled the emergency jettison handle. The D-558-2 separated; McKay jettisoned
the fuel and glided to a lakebed landing. That aircraft was undamaged and McKay
was not harmed.
Just 10 to 15 seconds after Butchart dropped the Skyrocket, the P2B-1S
propeller disintegrated. A blade from the number 4 engine passed through the
number 3 engine, destroying it, then sliced all the way through the fuselage,
exiting to strike the number 2 engine. Had Butchart not released the D-558-2,
the blade would have hit the Skyrocket, likely triggering an explosion that would
have destroyed both planes and all aboard them. Shrapnel from the disintegrat-
ing propeller hub and blades severely damaged engine number 3 and severed
the aileron cables to Butchart’s yoke, leaving him with no control. Armstrong’s
aileron cables had only one or two remaining strands. The severed wires rendered
the yoke difficult to move. Despite the loss of engines 3 and 4 and the damaged
controls, Armstrong was able to retain control and limp home to land on the
north lakebed. The P2B-1S was grounded for nearly 5 months of repairs.57
The second D-558-2 returned to flight on August 17, 1956, with McKay
as pilot. This was the first of eight flights following the near tragedy. The final
118
flight was on December 20, 1956, and it was the last flight by a Skyrocket.
Today, the historic second D-558-2 is suspended in honor at the National Air
and Space Museum of the Smithsonian Institution, Washington, DC.
Denouement: Skyrocket External

Stores Research Flights
The rocket-powered X-planes were commonly viewed by the public as exotic
aircraft being used to probe the boundaries of the unknown, traveling ever
higher and faster. In reality, their role also included the collection of flight
data on more mundane aspects of high-performance aircraft. While the second
D-558-2 was flying at high Mach numbers, the third Skyrocket was testing the
effects of “external stores,” such as bombs and fuel tanks, on aircraft handling
qualities at subsonic and transonic speeds. Fighter aircraft had long carried
these items under their wings. These payloads increased drag and affected flight
characteristics. As part of the development effort of operational fighters, exten-
sive wind-tunnel testing had been done to verify the safety and aerodynamics
of these fixtures. In contrast, little effort had been made to assess the effects of
external stores on longitudinal and lateral handling.
To explore these effects, the third Skyrocket was used for a series of flight
tests involving three different streamlined configurations: with an empty
A close-up of the pylon/bomb on NACA 145. Data collected on their effects was applied to
operational Air Force and Navy strike aircraft. (NASA)
119
Probing the Sky
mid-semispan under-slung pylon on each wing; with a simulated 1,000-pound

bomb shape attached to the pylon; and with a simulated 150-gallon fuel tank
attached. The three were respectively referred to as the “pylon configuration,”
the “small store configuration,” and the “large store configuration,” all built
by Douglas Aircraft Company.58
The aircraft was modified with the pylons in mid-May 1954 and made its
first flight in the stores project on June 2, 1954, with Crossfield as pilot. The
aircraft was in the pylon configuration for the test and used the jet engine only.
Top speed was Mach 0.72. The flight was repeated on June 16, again using only
the jet engine and also at Mach 0.72, but with the small store configuration.
The flight went smoothly, with no apparent complications.
The same was not true for the next flight, on July 8. Top speed was Mach
1, using both the jet and rocket engines. The small store configuration resulted
in a decrease in transonic performance and an increase in buffeting. These
problems caused the stores project to temporarily be put on hold, resuming
on October 8. The first flight made once research resumed used the large store
configuration’s simulated 150-gallon fuel tank. As on the initial flights, the jet
engine only was used for this mission, and speed was limited to Mach 0.74.
The way was now cleared for testing the large store configuration at higher
speeds. Crossfield reported no adverse effects but noted a rise in drag and heavier
buffeting in longitudinal maneuvers. More serious, the data from the wing strain
gauges indicated problems. The stores flights were again halted until Douglas
engineers completed rechecking the strength of both the wings and pylon.
The stores flights resumed in December 1954 with McKay in the third
D-558-2. These were made in the clean configuration, without the pylons
or stores shapes, and with the jet engine only. Once this series was complete,
however, another 4 months would pass before more stores flights would be
made. Flights resumed again on April 27, 1955, using the pylon configuration
and both the jet and rocket engines. Reaching a speed of Mach 1, McKay made
sideslips, rolls, and elevator and rudder pulses to collect data on the Skyrocket’s
handling qualities, wing and pylon loading, and buffet levels.
Nearly a month later, on May 23, McKay repeated the maneuvers, this
time with the large stores configuration. He noted that the simulated fuel
tanks caused high buffet levels compared to those of the pylon configuration.
June 1955 was a busy month, with five flights—two with the pylon configura-
tion and three with the large stores configuration. Identical maneuvers were
performed on all five.59
Data on the static longitudinal-stability and control characteristics were
obtained during speed runs made at about 35,000 feet, along with windup
turns between Mach 0.50 and 1.03 and altitudes ranging from 22,200 and
39,400 feet. Elevator pulses provided data on the aircraft’s longitudinal-dynamic
120
characteristics. These were made at Mach numbers between 0.49 and 0.75,
at altitudes between 22,000 and 27,000 feet. Data were also collected on
low-speed characteristics during 1-g stalls with the slats both locked and
unlocked and in the landing configuration with the slats unlocked and flaps
and gear down.
The D-558-2’s static lateral- and directional-stability characteristics were
determined through incremental increases in constant flightpath sideslips,
abrupt aileron rolls, and trim runs at Mach numbers of 0.44 to 1.04 and alti-
tudes of 22,000 to 36,000 feet. Lateral damping characteristics were derived
from rudder pulses made at Mach numbers between 0.50 and 0.87, at 20,000
and 32,000 feet. Both the longitudinal and lateral pulses were abrupt inputs,
with the pilot attempting to return and hold the controls at the trim position
while the oscillation damped out.60
Following the series of flights in June, tests were halted and did not resume
until August 30. As with previous flights, these were made with large store
The Skyrocket following a landing on the lakebed. Ground support personnel would remove the
film used to record the flight data and prepare the aircraft to be towed back to the hangar. The
dress code at the High-Speed Flight Research Station was relaxed but ran the gamut from suits
and ties to work clothes, short-sleeve shirts, baseball caps and sun hats, and Hawaiian shirts.
Crossfield was the pilot for the flight. (NASA)
121
Probing the Sky
configurations, and, though the maneuvers performed were the same, the
outcome was different. McKay landed nose high, and the underside of the
aft fuselage was damaged when it scraped on the lakebed. The aircraft was
grounded for 2 months for repairs.
Butchart flew the next mission with the D-558-2 on November 2, and
McKay made the next three flights on November 8 and 17 and December 8,
1955. As with all the flights since May, the large store configuration was used
and the same maneuvers were performed. This brought the stores research
project to a close. Two flights remained, however. The Skyrocket was stripped
of the pylons and restored to the original clean configuration. This took place
during the winter months, while the lakebed was typically closed to flight
anyway. Once the work was completed, the two additional flights would be
made to collect wing-load data. The lateral, directional, and longitudinal data
from the two clean flights would identify the effects caused by the stores.
The first of these, on February 1, 1956, was successful, reaching a speed of
Mach 1. The final flight, planned for Mach 0.9, was attempted on February
3. However, a turbopump overspeed aborted the flight. Because the flight data
were needed to complete the analysis of the stores data, a final flight of the third
D-558-2 was added. It did not take place until August 28, 1956, more than 6
months after the initial attempt. This time the flight was successful, reaching
Mach 0.96. The aircraft was permanently grounded, and the stores tests ended.
The results of the external stores tests were summarized in a research memo-
randum issued in October 1957:
The results presented herein are somewhat limited—particularly

in regard to the dynamic characteristics of the airplane—because
only a few flights were obtained with each external-store con-
figuration. Sufficient data were obtained, however, to establish
trends of the effects of the stores on the stability and control char-
acteristics of the airplane at subsonic and transonic speeds. Since
most of the tests were performed in the large-stores configuration
and only small differences were noted in the data for the various
configurations, most of the data presented here are for the large-
store configuration.
The incremental lift and drag effects for the large-store con-
figuration…indicate the appreciable increment of drag pro-
duced by the stores, especially at M > 0.9. In addition, the pilots
commented that they detected an increased penalty in airplane
performance with an increase in size of the stores investigated,
particularly at transonic speeds.61
122
The report’s concluding remarks noted:
The results of a flight investigation, from a Mach number of

approximately 0.45 to a Mach number of approximately 1.05,
of the Douglas D-558-2 research airplane equipped with several
configurations of external stores have indicated that the smaller
mid-semispan store installations generally had a small or negli-
gible effect on the handling qualities of the airplane. With the
large-store configuration, the trends exhibited in the longitudinal
and lateral trim, stall approach, dynamic and static stability and
control, and buffeting characteristics were generally the same as
for the clean airplane; however, significant changes in the mag-
nitude of the parameters measured were sometimes apparent,
particularly at the higher speeds.
The large-store configuration effected: a decrease in the sub-
sonic static longitudinal stability; an improvement in the lateral
damping characteristics at subsonic speed; an appraisable left-
wing drop at a Mach number greater than 0.9; an appreciable
decrease in lift-curve slope between a Mach number of about
0.85 and 0.96; a slight decrease in the level of lift for the onset
of buffeting at all subsonic speeds; and an appreciable increase
in apparent dihedral parameter dδa / dβ particularly at a Mach
number greater than 0.94. The side force parameter CY β increased
and the aileron control parameter (pb / 2V) / δα decreased with an
increase in the size of the store.62
Originally, this was not to be the end of the stores testing. A second series
of tests, at supersonic speeds, was planned. The aircraft to be used for these
flights was to have been the first D-558-2, used in the Douglas test flights.
As the others were, NACA 143 (as it was designated) had been transferred
by the Navy to the NACA in August 1951. It spent the following 3 years in
storage at Edwards. NACA managers decided in 1954 to have it modified to
an all-rocket/air-launched configuration. (The same as the second D-558-2.)
The Skyrocket was returned to Edwards on November 15, 1955, and check-
out began. It was not until the fall of 1956 that preparations with NACA 143
were complete. McKay made the modified aircraft’s first flight on September
17, 1956. And though the flight was successful, time had run out. The NACA
canceled the planned supersonic stores test project. As a result, this was the first
and only flight NACA 143 would ever make following modifications. The third
Skyrocket had made its final flight in August 1956. The second D-558-2 would
123
Probing the Sky
continue to fly until December 1956 before being grounded. This brought the
8-year Skyrocket project to a close.63
The NACA’s focus was shifting, and despite the sad irony of the end of the
Skyrocket project after many productive years, a new set of challenges was
on the horizon. Flights with the X-15 were due to begin in 3 years. The goals
of the X-15 program were not simply to fly slightly faster, but to triple both
the speeds and altitudes reached with the D-558-2. The X-15 would not just
reach Mach 6 and fly to the edge of Earth’s atmosphere. It would enter space.
An Assessment
Although the D-558-1 Skystreak’s achievements in the development of high-
speed flight were mixed, the importance of the D-558-2’s accomplishments
was clear. Ironically, like its jet-powered predecessor, the D-558-2’s original
performance was designed only for Mach 0.9 to 0.95 due to having to make
a ground takeoff. Had the Skyrocket not been modified for air launch, its
accomplishments would have been minimal and would have probably been
seen as a major disappointment.
NACA researchers had opposed air-launching the XS-1, arguing it should
be capable of normal takeoff and landing from Langley’s runway. By the time
the Douglas test flights on the Skyrocket were starting, however, the advantages
of air-launching research aircraft had been made clear. Converting the second
and third Skyrockets to an air-launched configuration greatly expanded their
performance envelope and, as a result, their research value. Another conse-
quence of the decision to use an air launch was that the second and third
Skyrockets ultimately were assigned different research roles. The earlier objec-
tions of the Langley researchers were forgotten.
Removal of the second D-558-2’s jet engine allowed the aircraft to carry a
larger rocket fuel supply, doubling its maximum speed. It repeatedly flew at
speeds of Mach 1.8 to 1.9. This allowed testing of the swept wings to speeds
higher than could be reached with operational service aircraft. The third
D-558-2 retained the jet/rocket configuration, which reduced its top speed.
Work with this aircraft focused on transonic and low-supersonic research, with
top speeds ranging from Mach 0.72 to 1.04. As with the XS-1 and D-558-1,
this provided a complementary research capability.
An interrelated series of factors in the D-558-2 wing design contributed to
one of its major accomplishments. The idea that swept wings could reduce the
detrimental effects of transonic and supersonic flight was new. The means of
ground-testing the Skyrocket’s wing design were limited by wind tunnel chok-
ing, which referred to the inability of wind tunnels to provide reliable data at
124
Aft upper view of NACA 144 at South Base. Though a variety of methods were tried in attempts
to reduce or eliminate pitch-up, their effects were limited. The solution was to place the horizontal
stabilizer low on the fuselage, where it would not be affected by airflow from the wings. (NASA)
high transonic and supersonic speeds. Finally, the D-558-2 wing design focused
on good low-speed/stall characteristics without loss of high-speed performance.
Thanks to Van Every’s careful design, the result was a wing with good low-speed
handling, low drag, and excellent transonic and supersonic characteristics.
The Skyrocket wing’s susceptibility to pitch-up, coupled with its compre-
hensive instrumentation package, ironically rendered the aircraft ideal for
research on the problem. To find solutions, various configurations of wing
fences, slats, leading-edge chord extensions, and bungee cords were developed
in wind tunnel tests. These tests looked at ways of reducing the flow of air that
caused the separation of flow at the tips. Separation triggered both the pitch-up
and the tip stalls, which were then exacerbated by the swept-wing configura-
tion. Crossfield described the test procedures: “The technique would be to go
up there and pull g at a fairly constant rate, trying to maintain as constant an
air speed as possible. And incidentally, there was something we really re-learned
with these kinds of wings. And that is that the old CMCl was a bunch of garbage
as far as this goes. We had to go back to the CMa because CL was dropping so
fast that it looked like the airplane was going stable—when really it was going
quite unstable at the time. And it would pitch.”64
125
Probing the Sky
Pitch-up was the most signifi-

cant stability problem plaguing the
D-558-2, but it was not the only one.
The aircraft was originally designed
for transonic and low-supersonic
speeds due to the ground takeoff pro-
file. When Bridgeman made the early
air launches, he experienced severe
oscillations and directional diver-
gence from straight and level flight.
This became known as “supersonic
yaw,” and it had not been expected
as other high-speed phenomena such
as dynamic and static instability had
been. Crossfield said later, “Bill
Bridgeman found that by manipu-
lating the G, you could control the
rate of this divergence, and give your- The D-558-2 and other X-planes existed
self time to get in very soft controls expressly for the collection of flight data.
Shown here on a D-558-2 wing are some of
to hold it on almost a knife edge.… the instruments used to record measurements
And we managed to take the airplane of the forces and moments acting on the
out substantially beyond its expected aircraft. (NASA)
design speed.”65 Originally built to fly
Mach 0.9 after a ground takeoff, the second D-558-2 ultimately broke Mach 2.
The Skyrocket also served as a training ground for the engineers and pilots
who later worked on the X-15 rocket plane in the 1960s. Crossfield made a
total of 87 rocket-powered flights in the X-1 and D-558-2, along with another
12 D-558-2 jet-only flights, all of these between 1951 and 1956. He subse-
quently made another 14 X-15 flights (1 glide flight and 13 powered flights)
as the North American Aviation contractor pilot. This made him the most
experienced rocket-plane pilot in the Nation. Jack McKay was the second-most
experienced rocket pilot, with 46 flights in the X-1B, X-1E, and D-558-2. He
subsequently made another 29 X-15 flights. Joe Walker made 26 flights in the
X-1, X-1A, X-1E, and D-558-2 (3 glide flights and 23 powered), which he
followed with 25 X-15 flights. Additionally, McKay and Walker qualified for
astronaut wings, as they exceeded an altitude of 50 statute miles in the X-15.
Another measure of the Skyrocket’s success was that all three aircraft sur-
vived the unknowns of research work. In contrast, of the six X-1 aircraft com-
pleted, only three—two of the first X-1s (the NACA aircraft, the second XS-1,
was rebuilt as the X-1E) and the X-1B—survived. The third X-1, X-1A, and
X-1D were all lost in explosions caused by the use of so-called “Ulmer” leather
126
gaskets in the liquid-oxygen tanks. Both of the X-2 aircraft were lost, the first
in a gasket explosion during a captive flight, the second when its pilot lost
control during a Mach 3 flight.66
In the end, however, it was the data that counted. The Skyrocket research
activities covered a wide range of fields: aircraft control and stability at speeds
approaching Mach 2, measurements of structural temperature and loads, wing
and tail pressure distribution, buffeting data, vertical-tail loads, possible solu-
tions to pitch-up, and data on the effects of external stores. Though it was
overshadowed by both the X-1 and the X-15, the D-558-2 played a major
role in exploring the unknowns of supersonic flight during the critical years
between the late 1940s and the mid-1950s.
127
Probing the Sky
Endnotes
1. Opening quotation from William Bridgeman and Jacqueline

Hazard, The Lonely Sky (New York: Henry Holt and Company,
Inc., 1955), p. 154.
2. Hallion, Supersonic Flight, pp. 68–69, 75–76.
3. Ibid., pp. 76–79.
4. The turbopump delay was, of course, why the first two XS-1s had
been completed with high-pressure blow-down nitrogen fuel and
oxidizer-feed systems. The turbopump issue thus prevented both
the X-1 and the D-558-2 from achieving their greatest possible
performance, which, in both cases, was approximately Mach 2 and
more than 70,000 feet.
5. The change is very noticeable when comparing photos taken during
the initial versus the later Douglas flights.
& Stiletto, pp. 48–53, 123–126. According to legend, John Martin
became the first D-558-2 project pilot due to the Douglas test pilots’
being reluctant to fly the aircraft. After discussion, the story goes,
the Douglas test pilots all decided to submit very high bids that they
knew the company would turn down. Martin was away delivering
an aircraft and did not know about the plan. Martin submitted a
reasonable bid and “won” the job.
7. Sigurd A. Sjoberg, “Preliminary Measurements of the Dynamic
Lateral Stability Characteristics of the Douglas D-558-2 (BuAero
No. 37974) Airplane,” NACA RM L9G18 (August 18, 1949), p.
2; Hallion, Supersonic Flight, p. 159; Libis, Skystreak, Skyrocket, &
Stiletto, p. 128.
8. Sjoberg, “Preliminary Measurements of the Dynamic Lateral
Stability Characteristics of the Douglas D-558-2 (BuAero No.
37974) Airplane,” pp. 2–3.
9. Ibid., p. 3.
10. Ibid., p. 3.
12. J.D. Hunley, ed., Toward Mach 2: The Douglas D-558 Program
(Washington: DC: NASA SP-2003-4531, 1999), p. 48.
13. Peebles, ed., The Spoken Word: Recollections of Dryden History, The
Early Years, pp. 50–51; Hallion, Supersonic Flight, pp. 159–160.
128
14. Michael H. Gorn, Expanding the Envelope Flight Research at NACA

and NASA (Lexington, KY: The University Press of Kentucky, 2001),
p. 220.
15. S.A. Sjoberg, James R. Peele, and John H. Griffith, “Static
Longitudinal Stability and Control Characteristics at Mach
Numbers up to 0.87,” NACA RM L50K13 (January 17, 1951),
p. 11.
16. Those seeking more information on the results of early flights with
the D-558-2 can refer to these NACA research memorandums: S.A.
Sjoberg, “Static Lateral and Directional Stability Characteristics
as Measured in Sideslips at Mach Numbers up to 0.87,” NACA
RM 50C14 (May 19, 1950); J.V. Wilmerding, W.H. Stillwell,
and S.A. Sjoberg, “Lateral Control Characteristics as Measured in
Abrupt Aileron Rolls at Mach Numbers up to 0.86,” NACA RM
L50E17 (July 20, 1950); John P. Mayer and George M. Valentine,
“Measurements of the Buffet Boundary and Peak Airplane Normal-
Force Coefficients at Mach Numbers up to 0.90,” NACA RM
L50E31 (August 28, 1950); W.H. Stillwell, J.V. Wilmerding, and
R.A. Champine, “Low-Speed Stalling and Lift Characteristics,”
NACA RM L50G10 (September 5, 1950); John P. Mayer, George
M. Valentine, and Beverly J. Swanson, “Measurements of Wing
Loads at Mach Numbers up to 0.87,” NACA RM L50H16
(December 26, 1950); John P. Mayer and George M. Valentine,
“Measurements of the Distribution of the Aerodynamic Loads
Among the Wings, Fuselage, and Horizontal Tail at Mach Numbers
up to 0.87,” NACA RM L50J13 (January 19, 1951); and W.H.
Stillwell and J.V. Wilmerding, “Dynamic Lateral Stability,” NACA
RM L51C23 (June 18, 1951).
17. S.A. Sjoberg, James R. Peele, and John H. Griffith, “Static
Longitudinal Stability and Control Characteristics at Mach
Numbers up to 0.87,” NACA RM L50K13 (January 17, 1951), pp.
5, 11, 13–14.
18. Bridgeman and Hazard, The Lonely Sky, p. 145.
20. Libis, Skystreak, Skyrocket, & Stiletto, p. 133
22. Ibid., pp. 162–163.
23. Ibid.
24. Bridgeman and Hazard, The Lonely Sky, p. 248.
25. Ibid., pp. 257–261; Hallion, Supersonic Flight, p. 231.
26. Libis, Skystreak, Skyrocket, & Stiletto, p. 130.
129
Probing the Sky
27. Hallion, Supersonic Flight, pp. 165–167; Bridgeman and Hazard,

The Lonely Sky, pp. 304–305; Libis, Skystreak, Skyrocket, & Stiletto,
p. 130; Theodore E. Dahlen, “Maximum Altitude and Maximum
Mach Number Obtained with the Modified Douglas D-558-2
Research Airplane During Demonstrations Flights,” NACA RM
L53B24 (April 20, 1953), pp. 4–5.
28. Dahlen, “Maximum Altitude and Maximum Mach Number D-558-
2,” p. 4.
29. Herman O. Ankenbruck and Theodore E. Dahlen, “Some
Measurements of Flying Qualities of a Douglas D-558-2 Research
Airplane During Flights to Supersonic Speeds,” NACA RM L53A06
(March 10, 1953), pp. 2, 8–9.
30. Letter, Chief, Bureau of Aeronautics, “Transfer of Aircraft to the
NACA under Public Law 266 of the 81st Congress,” December
27, 1950; letter, Chief of Naval Operations, “Transfer of Aircraft to
the NACA under Public Law 266 of the 81st Congress,” January 5,
1951. These letters covered the transfer of the two D-558-1s, the
three D-558-2s, the ex-B-29–now–P2B-1S launch aircraft, and spare
parts to the NACA.
31. Rogers Dry Lake is often flooded by winter rains, turning it into an
actual lake.
32. Hallion, Supersonic Flight, pp. 175, 184–185, 232, 237.
33. Jack Fischel and Jack Nugent, “Flight Determination of the
Longitudinal Stability in Accelerated Maneuvers at Transonic Speeds
for the Douglas D-558-2 Research Aircraft Including the Effects of
an Outboard Wing Fence,” NACA RM L53A16 (March 13, 1953),
pp. 1, 5–8.
34. Ibid.
35. Hallion, Supersonic Flight, pp. 175, 184–185, 232, 237.
36. Jack Fischel, “Effects of Wing Slats and Inboard Fences on the
Longitudinal Stability Characteristics of the Douglas D-558-2 Research
Airplane in Accelerated Maneuvers at Subsonic and Transonic Speeds,”
NACA RM L53L16 (February 8, 1954), pp. 13–14.
37. Ibid., pp. 13–14.
38. This period also saw the High-Speed Flight Research Station pilot’s
office undergo a number of changes. Robert Champine transferred
back to Langley, and John Griffith took a job with Chance Vought
in Texas. Walter P. Jones left for a job as a test pilot with Northrop
in August 1952, during the pitch-up tests. Jones was killed in the
crash of a YF-89D prototype on October 20, 1953. Crossfield made
nearly all the Skyrocket flights for the next 2 years.
130
39. Jack Fischel and Cyril D. Brunn, “Longitudinal Stability

Characteristics in Accelerated Maneuvers at Subsonic and Transonic
Speeds of the Douglas D-558-2 Research Airplane Equipped with
a Leading Edge Wing Chord-Extension,” NACA RM H54H16
(October 27, 1954), pp. 1–2.
40. Jack Fischel and Donald Reisert, “Effects of Several Wing
Modifications on the Subsonic and Transonic Longitudinal
Handling Qualities of the Douglas D-558-2 Research Airplane,”
NACA RM H56C30 (June 5, 1956), p. 12.
41. Ibid., p. 13.
42. Ibid., p. 7.
43. Libis, Skystreak, Skyrocket, & Stiletto, p. 75.
44. Herman O. Ankenbruck and Chester H. Wolowicz, “Lateral
Motions Encountered with the Douglas D-558-2 All-Rocket
Research Airplane During Exploratory Flights to a Mach Number of
2.0,” NACA RM H54I27 (December 17, 1954), p. 7.
45. Ibid., pp. 8–9.
46. A. Scott Crossfield and Clay Blair, Jr., Always Another Dawn:
The Story of a Rocket Test Pilot (New York: Arno Press, 1972), pp.
171–172.
47. Ibid., p. 160.
48. Ibid., p. 172.
49. Ibid.
50. Ibid.
51. Marcelle Size Knaack, Post–World War II Fighters (Washington, DC:
Office of Air Force History, 1986), pp. 10, 60–64, 114, 175–176.
52. Crossfield and Blair, Always Another Dawn, pp. 160–161.
53. E-mail, Albion H. Bowers to Curtis Peebles, “Rocket Science,” July
14, 2010. In possession of the author.
54. Hunley, ed., Toward Mach 2: The Douglas D-558 Program,
pp. 54–56; Cyril D. Brunn and Wendell H. Stillwell, “Mach
Number Measurements and Calibrations During Flights at High
Speeds and at High Altitudes Including Data for the D-558-2
Research Airplane,” NACA RM H55J18 (March 12, 1956), p.
15. Crossfield was not the only person who was not in the best
condition that morning. Jack Moise, a B-29 launch panel operator,
was sprayed with highly concentrated hydrogen peroxide during
preflight preparations. A mechanic named Kinkaid hosed him
down with water, and both were taken to the flight-line dispensary.
Moise suffered several minor chemical burns on his face, which
cleared up within a few days. Crossfield saw Kinkaid sitting on a
131
Probing the Sky
bench, wearing wet clothing. When Crossfield asked if he was cold,

Kinkaid replied that he was warm. Crossfield realized Kinkaid was
also soaked in hydrogen peroxide and could catch fire. Crossfield
ripped off the clothing and found that chemical burns had bleached
Kinkaid’s arms and legs white.
55. Crossfield’s departure roughly coincided with another change in the
NACA’s High Desert test team: the High-Speed Flight Research
Station dropped “Research” from its name in 1954, becoming
simply the High-Speed Flight Station.
56. Hallion and Gorn, On the Frontier, p. 392.
57. Peebles, ed., The Spoken Word: Recollections of Dryden History, The
Early Years, pp. 121–123.
58. Jack Fischel, Robert W. Darville, and Donald Reisert, “Effects of
Wing-Mounted External Stores on the Longitudinal and Lateral
Handling Qualities of the Douglas D-558-2 Research Airplane,”
NACA RM H57H12 (October 28, 1957), pp. 1–2.
59. Hallion, Supersonic Flight, p. 240. Those seeking detailed
information on the results of later D-558-2 research missions
can refer to these NACA research memorandums: Jack Nugent,
“Lift and Drag Characteristics of the Douglas D-558-2 Research
Airplane Obtained in Exploratory Flights to a Mach Number of
2.0,” NACA RM L54F03 (August 4, 1954); Wendell H. Stillwell,
“Results of Measurements Made During the Approach and Landing
of Seven High-Speed Research Airplanes,” NACA RM H54K24
(February 4, 1955); Gareth H. Jordan and Earl R. Keener, “Flight-
Determined Pressure Distribution over a Section of the 35° Swept
Wing of the Douglas D-558-2 Research Airplane at Mach Numbers
up to 2.0,” NACA RM H55A03 (March 25, 1955); and Glenn
H. Robinson, George E. Cothren, Jr., and Chris Pembo, “Wing
Load Measurements at Supersonic Speeds of the Douglas D-558-2
Research Airplane,” NACA RM H54L27 (March 30, 1955).
60. Fischel, Darville, and Reisert, “Effects of Wing-Mounted External
Stores on the Longitudinal and Lateral Handling Qualities of the
Douglas D-558-2 Research Airplane,” pp. 2–3, 6. The apparent
dihedral parameter dδa / dβ is defined as the “rate of change of
aileron deflection with sideslip angle.” The side force parameter CYß
is defined as the “rate of change of lateral force coefficient with angle
of sideslip.” The aileron control parameter (pb / 2V) / δα is defined as
the “variation of wing-tip helix angle with total aileron deflection”
and is measured in radians/degrees.
61. Ibid., p. 4.
132
62. Ibid., p. 13.

63. Hallion, Supersonic Flight, p. 230. Once retired, the three D-558-2s
were put on display. The first D-558-2, used by Douglas and later
modified to an all-rocket configuration, is at the Planes of Fame
Museum in Chino, CA. The second D-558-2 (all-rocket configu-
ration) is in the Smithsonian Institution National Air and Space
Museum in Washington, DC. The third D-558-2 (jet-and-rocket
configuration) is mounted outdoors on a pole at Antelope Valley
College in Lancaster, CA.
64. J.D. Hunley, ed., Toward Mach 2: The Douglas D-558 Program, pp.
48–49. CMCl is the static stability in pitch of an aircraft. CM is the
pitching-moment coefficient. CMa is the partial derivative of the
pitching-moment coefficient with respect to the angle of attack.
65. Ibid., p. 54.
66. For details, see Hallion, Supersonic Flight, pp. 169–172, 175–177,
189–191.
133
Mockup of the X-3 at the Douglas factory. The X-3 was the sleekest of the early X-planes but
delivered the most disappointing performance. (USAF)
134
CHAPTER 4
Unfulfilled Promise,
Serendipitous Success:
The Douglas X-3 Stiletto
Boy, it’s hard to keep your speed and altitude in this thing.
—William “Bill” Bridgeman1
An ambitious jet-powered research aircraft meant to cruise at Mach 2, the

twin-engine Douglas X-3 Stiletto was the sleekest and most radical-appearing
of the early X-planes. Like the D-558-1, the Stiletto’s accomplishments were
mixed. The X-3 never achieved its intended design speed but did furnish critical
data on a new stability and control peril spawned by the design requirements
of supersonic flight. The X-3’s failure to attain Mach 2 stemmed largely from
the failure of Westinghouse, its engine manufacturer, to successfully develop its
Westinghouse J46 turbojet. As a consequence, the X-3 had to employ smaller
and far less powerful Westinghouse J34 engines, inhibiting its potential per-
formance and making a mockery of its sleek supersonic shape.
The X-3 had its origins in December 1943, when the Army Air Forces
requested that the Douglas Aircraft Company study the feasibility of an aircraft
capable of sustained speeds in excess of Mach 1. This was a bold request, as
many (as discussed previously) considered the “sound barrier” as presenting an
insurmountable challenge. A team of Douglas engineers working for Frank N.
Fleming took on the challenge, submitting a proposal in January 1945, and the
AAF responded by issuing Douglas a contract in June 1945 for the design of a
supersonic jet-powered research aircraft capable of attaining Mach 2 at 30,000
feet and having a flight endurance of 30 minutes. The project was designated
MX-656, the plane to be known as the XS-3 (later simply X-3).2 As with the
X-1 and D-558 aircraft, initial flights with the X-3 would be made by the
contractor to demonstrate that the aircraft could meet the basic performance
requirements. The Air Force would then conduct an abbreviated performance
evaluation up to the X-3’s maximum design speed and altitude. Finally, NACA
pilots would make a long-term series of research flights to extract the maximum
135
Probing the Sky
amount of data on aerodynamics, stability and control, handling qualities, and

other aspects of the aircraft.
Designing the Stiletto

Douglas Aircraft’s Frank N. Fleming was named project engineer on the X-3
and faced a demanding task. The design had to be a conventional one capable
of operating without special launch and handling equipment, taking off and
landing under its own power using standard landing gear, and flying its entire
speed range without external assistance, and it would be required to demon-
strate flying and handling qualities like those of modern high-speed aircraft.
Finally, it would have to have conservative thrust margins, fly for a reasonable
amount of time at supersonic speeds, perform satisfactorily at transonic speeds,
and withstand aerodynamic heating.3
Thanks largely to the visual power of the V-2 missile and Douglas’s shapely
Skyrocket, the popular image of a supersonic aircraft in the late 1940s was
one featuring an elongated, needle-shaped fuselage with highly swept wings
and tail surfaces. Actually making such airplanes was extremely challenging.
The basic requirement would call for the engines’ thrust capacity to be greater
than the airplane’s drag up to the Mach 2 design goal. If the thrust and drag
curves intersected below this speed, the X-3 would not perform as required.
The engines would need to have a frontal area that was as small as possible yet
still capable of producing the necessary amount of thrust. In the late 1940s,
this meant turbojet engines fitted with afterburners.
To minimize drag, the engines had to be enclosed within a fuselage likewise
having the minimum possible frontal area. But jet engines typically had a con-
siderable amount of external equipment attached to their cylindrical casings.
This necessarily increased frontal area. Designers also determined that in order
for the X-3 to achieve its performance goals, two engines would be needed, both
for safety and to improve thrust-to-drag ratio. The two engines represented the
greatest mass of the aircraft, so the center of lift would need to be near the engines.
This meant the wing spar had to be under the engines, which would further
enlarge the fuselage. The X-3’s main landing gear, like that of the D-558-2, also
had to be housed in the fuselage. And as with the wings, the main landing gear
had to be near the center of the greatest mass, reinforcing yet again the fact that
the fuselage would have to be made bigger.
Auxiliary equipment also had to be included. Air had to be supplied to the
engines. Protruding air inlets were added to the fuselage, and diffusers of the
proper length and expansion ratio connected these to the engines. The two
turbojet engines and afterburners generated considerable heat. Countering this
136
Unfulfilled Promise, Serendipitous Success: The Douglas X-3 Stiletto
effect required having ducting

sufficient for carrying a large
volume of cool air, the source
of which was the boundary
layer bled off ahead of the
air intake scoops. Air passed
through ducts between the
two engines. To make room
for the ducts, the two engines
were spread apart. Shrouds and
insulation were required to
direct the cooling air and keep
the aircraft structure at proper
temperatures. Still more space
would be needed to allow for
thermal expansion, structural
deformation, vibration, and
manufacturing tolerances. A
tunnel was added on top of the
fuselage to contain the flight
An overhead photo of the completed X-3. The aircraft
control cables, hydraulic lines, had a very long fuselage and a short wingspan. The
tank vents, wiring, instrumen- fuselage was tightly packed with the two engines,
tation tubing, and other items. fuel tanks, landing gear, and other systems required.
These changes resulted in an (USAF)
increase in the X-3’s frontal area
and thus the amount of drag. The ratio of the engine area as a percentage of total
fuselage frontal area had been 93 percent in the minimal fuselage design. With
all the added equipment, the ratio of engine area to total fuselage frontal area
dropped to 33 percent for the two-engine design.
The next design issue was the selection of a wing planform combining the
optimum values of high-speed drag, stability, weight, maximum lift, structural
strength, and minimum drag for a given amount of lift. The X-3’s wings would
need both a high critical Mach number and the ability to cope with the effects of
compressibility over the full speed range up to Mach 2. The design that emerged
was a slightly swept low-aspect-ratio wing that was very thin and had a sharp-
wedge airfoil shape and leading and trailing-edge flaps.
Beyond wind tunnel tests of the proposed wing shape, a series of free-flight
rocket launches were made by the NACA to compare the coefficient of drag for
various wing shapes. These shapes included the X-1 (high-aspect-ratio, straight
wings), D-558-2 (low-aspect-ratio, swept wings), X-2 (high-aspect-ratio, swept
wings), and X-3 (low-aspect-ratio, swept wings). All the wings had identical areas
137
Probing the Sky
and fuselage bodies, so the lone variable was wing drag due to compressibility.
Of the four wing shapes, the X-3s had the lowest drag coefficient over the range
of Mach 0.9 to Mach 1.7. The highest-drag wing was the X-1’s. Although the
main fuselage and wing designs were complete, the X-3 still lacked a cockpit,
fins, rudder or horizontal control surfaces, fuel tanks, or accommodations for
an instrument payload.
Early rocket-powered model flights and drop tests had indicated that drag
on an optimum ogival body was reduced when the shape had a high fineness
ratio (the body’s length divided by its diameter). The optimum ratio for a
subsonic aircraft was 3:1, versus 14:1 for a supersonic aircraft. With the fron-
tal area established, it was clear that the X-3’s fuselage length would need to
be extended to minimize supersonic drag. This was accomplished by adding
a long nose that held the pilot, data instrumentation, controls, radio, and
other equipment.
The next step was the addition of tail surfaces. There were long-established
relationships between the positions of the tail surfaces and the wing, relation-
ships shaped by structural and aerodynamic factors. In the early X-3 design
studies, the tail area was increased and moved downward and forward relative
to the wing. This also allowed the tailpipes and supporting structure to be
extended to support the tail. The flaw in this design was added weight, but
engineers reduced the weight by keeping the tailpipe short. Additionally, expe-
rience held that to maximize the tail surfaces’ stability and control effectiveness,
they should be located as far back as practical. As a result, a tail boom was used
on the final X-3 design, the Model 499D. The X-3’s horizontal stabilizer also
had an all-moving, one-piece design, rather than the two-part movable stabi-
lizer and elevator used on the X-1, D-558-1, and D-558-2. The X-3 marked
the first application of an all-moving “slab” tail to an American supersonic
aircraft. Testing in Langley’s 7-by-10-foot High-Speed Wind Tunnel resulted
in Douglas’s enlarging the horizontal tail by 40 percent.
The X-3 design was fitted with an air conditioning system installed between
the air intakes; additional flight data equipment placed below the wing; and the
hydraulic units, fire extinguishers, electrical components, and other items fixed
in various nooks and crannies. Finally, fuel tanks were fitted into the remaining
fuselage volume. A fuel tank holding 500 gallons of kerosene was located in
the lower fuselage, below the diffusers, and another 170-gallon fuel tank was
added to the aft fuselage, under the engine exhaust. The final 330-gallon fuel
tank was located in the tail boom, above the jet engines. Douglas engineers
estimated this would enable a 3-hour flight at high-subsonic speeds or 10
minutes at Mach 2. To protect the fuel tanks and control cables in the event of
an engine compressor or turbine wheel failure, armor plating was installed in
critical locations. This was similar to the modifications made to the D-558-1.4
138
Reaching a sustained speed of Mach 2 required considerable thrust.

Consequently, the X-3 had a pair of Westinghouse J46 axial-flow jet engines.
These were derivations of the firm’s earlier J34 engine, but with a greatly rede-
signed turbine and compressor section. Though the J34 jet could produce
a maximum thrust of 3,000 pounds, the J46’s basic thrust was to be 4,200
pounds. Use of the afterburner raised this to 6,600 pounds. Douglas engineers
estimated that the X-3 could reach Mach 2 at 35,000 feet using only 80.6 per-
cent of the J46’s anticipated maximum thrust. The top speed of the X-3 was
limited not by engine thrust, but by structural heating of its largely 24S-T81
aluminum alloy airframe.5
The result was an aircraft that seemed as if it had flown in from the future.
None of the other early X-planes had the X-3’s look of raw power. The bullet
shapes of the X-1 and D-558-2 looked conventional by comparison, while
the cylindrical D-558-1 was old-fashioned. Not until the hypersonic X-43A
a half century later was there another experimental aircraft that so embodied
the concept of speed.
In Thrust, Not Enough

Because its designers drastically underestimated the challenges they would face
creating several new high-performance engines, Westinghouse’s efforts to build
upon its combat-proven J34 experience failed miserably, crippling a number of
aircraft programs, one of which was the X-3. “The history of the Westinghouse
J34,” RAND analyst Thomas A. Marschak wrote in 1964, “[gives] a rather clear
impression of a ‘classic’ case of engine-airframe commitments being made in
the face of great uncertainty about all engine magnitudes.”6 By June 1951,
development of the J46 was a year behind schedule, and Westinghouse, in
desperation, decided to spin off a lower-powered variant of the J46 that could
be installed in the X-3, with the higher-performance J46 variant installed later.
But a year later, by August 1952, the problems had grown worse, and even the
lower-powered engine was now 14 months behind schedule and not expected
to be ready until 1955, 11 years after the X-3 program had begun.7
Douglas managers had already considered using the less-powerful J34-WE-
17 engines as an interim measure pending the availability of the more powerful
J46. These could produce 4,850 pounds of thrust with afterburner and had
dimensions similar to those envisioned for the more powerful J46, minimizing
the difficulties created by the switch. The assumption was that initial flights
would be made with these J34 engines. Then, once the J46 engines were ready,
they would be installed in the pair of X-3s being built. It was a reasonable
strategy, and one followed subsequently for other programs, such as the North
139
Probing the Sky
American X-15 (which used two RMI XLR-11 interim engines until the more
powerful Thiokol XLR-99 was ready) and the Lockheed A-12 Oxcart program
(which employed two Pratt & Whitney J75 engines until its more powerful
P&W J58s were available).
But now, the J46 was not only behind schedule, it was growing too large
to permit two to fit within the X-3, forcing Douglas to turn to an alternative:
the earlier J34. The transonic thrust-to-drag ratio of the X-3 with the interim
J34-WE-17 engines was much lower than that calculated for the J46 engine.
Various means of boosting the J34’s thrust were examined. NACA engineers
at the High-Speed Flight Station went even further and considered modifying
the X-3 to an all-rocket/air-launched configuration.8
This, however, would have involved removing the fuel tanks, air inlets, and
jet engines and replacing the latter with a pair of Reaction Motors 6000C4
four-chamber rocket engines. The eight rocket cylinders would also have to be
fitted with nozzle extensions, resulting in exhaust expansion to the ambient air
pressure at 50,000 feet. The amount of propellant to be carried, based on the
assumed consumption rate, was approximately equal to the available volume
of the jet engine compartment.
A second prohibitive issue was the choice of a launch aircraft for use with an
all-rocket/air-launched configuration. The candidates were the Douglas C-74
and C-124 transports and the B-36, B-52, and YB-60 bombers. The C-74
and C-124 were swiftly eliminated because their maximum launch speed was
slower than the X-3’s anticipated stall speed. (As well, the C-124 was also too
small to carry the X-3). The B-36 could carry and launch the X-3 but was
unavailable because the Air Force required all available examples of this huge
6- (and later 10-) engine bomber for its global strategic nuclear bombing
mission. The B-52’s wheelbase was too short to allow the X-3 to be mounted
beneath the fuselage (air-launch advocates did not yet conceive of attaching
the X-3 to a launch shackle under the wing, as with the later North American
X-15). The only possibility was the YB-60, a jet-powered version of the B-36
that met the performance requirements. Ultimately, however, both the acqui-
sition of a launch aircraft and the necessary modifications to the X-3 were
deemed too costly, ending any prospects for the program’s future entirely.9
The failure to produce the J46 engine in time and in a size to be incorpo-
rated on the aircraft crippled the X-3 program. The X-3’s best speed in level
flight with the J34-WE-17 engines was a meager Mach 0.96. If it entered a
dive, its smooth high-fineness-ratio fuselage and low transonic drag ensured
that it rapidly accelerated into the supersonic, to Mach 1.2—but the pilot
then had to pull out, and its speed would swiftly drop back to the “other
side” of the sonic divide. Although hopes lingered for a time that a rocket-
powered/air-launched conversion might be approved, the NACA Research
140
A view of the X-3 from the rear. The fuselage was painted white but the wings, horizontal stabi-
lizer, jet exhaust, and parts of the tail boom were left in bare metal. (NASA)
Aircraft Projects Panel rejected both options as too expensive and unjustified.
The panel reasoned that “other aircraft are coming along which operate in
the same speed range as that for which the X-3 was intended,” a reference to
the Mach 2 Lockheed F-104, which, ironically, benefited greatly from the
X-3’s development experience.10 By this time, the first X-3 was nearly ready
for delivery to Edwards Air Force Base; the second airframe was only partially
completed. The Air Force ended construction of the nearly complete vehicle
and transferred the unfinished airframe to the NACA for spare-parts support.
The X-3 research aircraft had been intended for the exploration of the
unknowns of sustained Mach 2 flight, such as aerodynamic heating, intake
scoop and air duct flow characteristics, and high-speed stability and control.
But the failure to produce a J46 that could fit in the airplane effectively trans-
formed the X-3 into a low-speed test bed for aircraft having low-aspect-ratio
wings joined to high-fineness-ratio fuselages, nothing more.11
141
Probing the Sky
The completed X-3 was delivered to Edwards on September 11, 1952, and
final checkout by Douglas engineers began in preparation for high-speed taxi
tests. This involved making engine test runs, replacing failed components,
and installing NACA instrumentation. The instruments recorded airspeed
and altitude; normal and transverse acceleration; roll, pitch, and yaw angular
velocity; angle of attack and sideslip; control column, control wheel, and
rudder pedal positions; stabilizer, aileron, and rudder positions and control
forces; and leading and trailing-edge flap positions.12
The X-3 high-speed taxi tests began on September 30, 1952. William
Bridgeman, having made the air-launched high-speed/high-altitude flights
in the second D-558-2 the previous summer for Douglas, now conducted
the manufacturer acceptance flights with the X-3. A series of problems were
encountered in early tests, including a failure of the afterburner hydraulic
actuator; a primary engine control system malfunction; a malfunction of the
drag chute, which was jettisoned immediately after deployment on several
runs; the loss of all tread on the nosewheel; and problems with the main
landing gear brakes and tires. By the sixth taxi test, made on October 14,
the problems had been largely resolved. The following day, Bridgeman lifted
the X-3 off the lakebed and remained airborne for about a mile and a half.
The way had now been cleared for the first flight, scheduled for October 20.
The eighth and final taxi test was made to check the X-3’s longitudinal control
with hydraulic power off. Bridgeman found the resulting control forces were very
high, and the dead spot around neutral made control difficult. Despite this, the
first flight went ahead. The weekly status report noted:
Flight No. 1 was made on 10-20-52 right after taxi test No. 8. Loss
of some power on the right-hand engine and loss of the right-hand
hydraulic pump output was caused by a malfunction of the right-
hand fire shut-off control. Longitudinal pitching of the airplane
was excessive. Duration of the flight was nineteen minutes.13
Troubleshooting of the problems took nearly 2 weeks, and the second flight
was made on October 31, 1952. Again, as the progress report noted, “[During
t]he longitudinal control sensitivity oscillations in which fairly large attitude
changes occurred, a buffet-like shutter was felt in the airplane. Flight duration
was 13 minutes.”14
A considerable amount of troubleshooting, repairs, testing, and modi-
fication had to be undertaken. The resulting delays were made worse when
a new nozzle actuator on the right engine afterburner failed twice. Then in
mid-November 1952, an inch of rain fell, closing the lakebed. The initial
estimate was that the lakebed would be closed for 3 weeks. But coupled with
142
Normally, aircraft were taxied or towed to the runway. The X-3’s tires had to withstand very high
takeoff and landing speeds. To avoid tire damage, a special vehicle was built to carry the aircraft
out to the lakebed. Here it is seen on the back ramp behind Hangar 4802, at the new NACA
facility at Edwards (still in use more than a half-century later). (NASA)
the weather, continuing repair work on the X-3 delayed the third flight until
late April 1953.
The X-3 was ready, but delays persisted. The first attempt was made on April
21, 1953. As with the first flight, a taxi run was made to check out the aircraft.
This was followed by the attempt at the third flight. The taxi test resulted in
a loss of tread from all three tires. The tires were replaced, but by the time the
work was finished, wind conditions had changed, and the flight was postponed.
The tire failures were later traced to the high takeoff speeds. Underpowered
engines and small wings meant the X-3 had to reach 260 knots before it could
take off. Existing tire technology and materials were incapable of withstanding
the stresses this produced.
A second try came on April 23, but high-altitude turbulence forced another
cancellation. The following day, Bridgeman attempted to take off, but just before
leaving the ground he felt the aircraft vibrating, so he aborted the takeoff. An
inspection showed that the nose tire’s tread and most of the main tires’ treads
were missing. Again, the high-speed takeoff had resulted in tire failure. Engineers
concluded that the loss of tread had caused the main tires to be unbalanced,
resulting in the vibration.15
Flight 3 was rescheduled for April 27, 1953, but rain fell the night before,
closing the lakebed for a day and a half. Bridgeman successfully made a
34-minute flight on April 30. Among changes that had been made to the X-3
in the months since its last flight was an increase of the elevator power-valve
143
Probing the Sky
linkage ratio to 5:1, which resulted in some improvement in longitudinal

control. However, the aircraft had been fitted with a “load feel control” that
allowed the control setting to be changed either automatically or manually. The
weekly report noted, “But it was still very difficult, if not impossible, to avoid
over-controlling with the load feel control operating automatically. Manual
increase of the load feel gradient to its maximum value improved control con-
siderably at the speeds encountered during the flight…. With the exception of
longitudinal control, airplane and engine operation was satisfactory.”16
A subsequent report expanded on the handling problems:
During the climb following take-off on Flight No. 3, while the

control system was still on automatic load feel, longitudinal oscil-
lations periodically appeared. These oscillations were similar to
those encountered on Flight No. 2 following afterburner shut-
down…. It is apparent that at oscillation frequencies as low as
½ cps [cycles per second], the time lag between aircraft normal
acceleration response and pilot’s elevator force makes it difficult
for the pilot to control these longitudinal oscillations.17
Again, extensive work and checkout were required before the fourth flight
could be made on June 5. This one lacked most of the problems of earlier
flights. The X-3 Weekly Status Report for the week ending on June 5, 1953,
stated, “Flight no. 4 was made on 6-5-53. A maximum indicated airspeed of
525 knots was reached. Airplane and control stability was good.”18
X-3 Flight 5 was made on June 11, 1953, and reached an altitude of 25,000
feet. In a level flight run at this altitude, a maximum observed Mach number
of 0.943 was reached. The X-3’s control and stability appeared satisfactory.19
Flight number 6 was made on June 25. The report for the week ending on
June 27, 1953, noted:
The airplane climbed to 34,000 feet and a level run was made at
approximately 33,000 feet. The maximum observed Mach No.
during the run was approximately .94. An accelerated turn at a
load factor of 2.0 was made at approximately 30,000 feet and an
observed Mach No. between .93 and .90. At approximately 28,000
feet, a pushover at a load factor between 0.2 and 0.5 was made at an
observed Mach No. of approximately .94. A lateral and directional
stability check made during the pushover indicated no reduction in
damping at the reduced load factor. During the recovery from the
pushover, a maximum observed Mach No. of 1.018 was reached….20
144
This cleared the way for supersonic dive flights. The NACA X-3 weekly
report noted:
Flight No. 7 was made on 7-15-53. The airplane climbed to

36,000 feet from which a dive at an angle of approximately
15° was started. During the dive, a rudder pulse was made at
an observed Mach No. of .96, an aileron pulse was made at an
observed Mach No. of 1.05 and an elevator pulse was made at an
observed Mach No. of 1.09. Immediately following the elevator
pulse, a +2.5g pullout was made and the minimum observed alti-
tude reached was 23,790 feet. A maximum Mach No. of M=1.10
was reached at 26,150 feet during the dive. All pulses during the
dive were satisfactory.21
Bridgeman commented after the flight in “Douglas Flight Report No. 7”:
In the initial part of the dive, the rudder was pulsed. The damping
of the lateral-directional oscillations was slow and, because of the
necessity of waiting for the damping before increasing the dive
angle, the requested twenty degree attitude was never reached.22
The first attempt to make flight 8 on July 21 was aborted during the takeoff
roll when the nose wheel again lost tread. The flight was successfully made
on July 22. Bridgeman climbed to about 36,160 feet, accelerated in level
flight to Mach 0.89, and pushed over into a 25° dive angle. During the dive,
Bridgeman made rudder and elevator pulses and reached Mach 1.19 before
pulling out of the dive. The X-3 pitched around the lateral axis several times.
Bridgeman commented, “This was mostly pilot induced [while] coping with
the tender longitudinal control.” Low-speed stability and control tests were
made before landing.23
Bridgeman made the X-3’s fastest flight on July 27, 1953. The weekly
report stated:
A dive was made from 36,440 feet to 19,320 feet at a maximum

dive angle of approximately 30°. A maximum observed Mach No.
of 1.21 was attained during the dive and a maximum load factor
of 6.7 was reached during the pullout. Low-speed stability and
control was investigated during three approaches to stall and two
stalls using “military” power with the leading edge flaps full down.
In each stall, the airplane rolled to the right.24
145
Probing the Sky
Despite the successful supersonic flight, it was now clear that the X-3 was
a disappointment in terms of performance. Even in a steep dive, its top speed
was little different from that of an F-86. Bridgeman continued to make X-3
flights, with his 25th and final flight coming on December 1, 1953, but most
were in the low-supersonic speed range. Nevertheless, given the X-3’s lack of
thrust and stability, the flights were never routine. On Bridgeman’s 23rd flight,
on October 21, he made a rolling pullout from a dive. Both afterburners were
already shut down due to high temperatures in the tail area. Bridgeman felt
severe vibrations, and the right engine switched to emergency mode, its revolu-
tions dropped, and the outlet temperature increased. He was unable to restart
the engine and had to use the left engine’s afterburner just to stay aloft. He
was able to make a lakebed landing on one engine without damaging the even
more severely underpowered aircraft, which itself constituted a tremendous
tribute to his extraordinary abilities as an experimental test pilot.25
A postflight report noted the scale of the engine malfunction:
The failure was due to the root failure of one blade in the second-
stage turbine and shearing at approximately the mid-span of an
adjoining blade. It is the opinion of the engine manufacturer’s rep-
resentative that the whole blade was ejected downstream through
the engine and the afterburner and subsequently lost to the free
stream but that the half-blade segment or a fragment of it traveled
forward, finally becoming wedged between the first-stage turbine
and the first-stage stator vanes. Further, it is his opinion that the
half-blade segment being wedged between the turbine wheel and
the nozzle vanes was responsible for the failure of 10 doweled
bolts which allowed all of the first-stage nozzle vanes to fall to the
bottom of the nozzle, from where they were recovered.26
A replacement engine was installed in the X-3, and Bridgeman made con-
tractor flights 24 and 25, completing Douglas’s obligations to the customer.
On December 8, 1953, Douglas turned the X-3 over to the Air Force.
Launching the Air Force–NACA Program

Originally, the Air Force had planned to conduct an extensive evaluation of the
aircraft. But the X-3’s disappointing performance caused Air Force testers to
drastically scale back their planned evaluation, particularly as the service had
major flight-testing responsibilities reflecting its participation in—and lessons
learned from—the Korean War, which had ended the previous July.
146
Charles E. Yeager beside the X-3, in which he made three flights. The aircraft was designed to
reach Mach 2, but the jet engines that powered it were never capable of delivering the required
amount of thrust. As a result, the X-3 barely exceeded Mach 1 in a dive. The aircraft’s design,
with its mass concentrated in the fuselage and small, razor-blade-like wings, made it suscep-
tible to inertial coupling. Ironically, data collected on this phenomenon constituted the project’s
most important results. (USAF)
Accordingly, Lieutenant Colonel Frank K. Everest and Major Charles E.

“Chuck” Yeager each made only three flights in the X-3, then washed their
hands of the project. Although these were made by Air Force pilots, they were
listed as NACA flights 1 through 6 because NACA instrumentation was aboard
the X-3, enabling the taking of flight research data. Everest made his first flight
on December 23, 1953, with Yeager following on December 29. The flights
included windup turns, stalls, level runs, dives, and rudder pulses. The top
speed reached during NACA flights 1 and 2 was Mach 1.09. NACA flights 3
147
Probing the Sky
through 6 were made between July 2 and July 29, 1954. Like the two initial
flights, these were primarily for evaluating handling qualities and performance.
Strain gauges were also checked out, and measurements were made of wing-
pressure as well as of wing, tail, and landing gear loads.27
Colonel Everest later recounted his experiences with the X-3, which he
called “one of the most difficult airplanes I have ever flown”:
Distinct control problems resulted from its combination of very

long fuselage and extra-short wings. Longitudinal control was
sensitive and the airplane pitched up and down on the slightest
provocation…. The wings had both leading and trailing edge
flaps to give it added lift during take-off and landing. In fact,
leading edge flaps were required after take-off until the airplane
had attained an air speed close to 350 knots. I found this out on
my first flight, when I retracted them at 300 knots and at the same
time began a turn back toward the base. The X-3 immediately
began to buffet and stall because the wings could not support the
weight, and I had to level out and continue flying straight ahead.
I climb[ed] to about 37,000 feet, where I made some maneuvers,
put the airplane in a dive and went supersonic…. After more
maneuvers I returned to Edwards for my landing. At 5,000 feet
I extended leading edge flaps and chopped the engines back to
a low power setting. Then I turned into my downwind leg, and
extended the trailing edge flaps and landing gear. As I did so I
began sinking like a rock. At once I applied full power and pulled
my gear up, and not until I was turning into my final approach
did I again extend the gear.28
On a more detailed level, the stability and control data from the Douglas
contractor flights and those made by Everest and Yeager were summarized in
an NACA research memorandum:
Longitudinal control deflection required to trim the airplane

over the Mach number range was generally similar to that of
other airplanes, characterized by a stable variation at Mach num-
bers below 0.92 and a slight nose-down trim change at Mach
numbers above 1.07. Data obtained during turns and pull-ups
indicated that throughout the Mach number range from 0.65
to 1.21, the apparent static longitudinal stability was positive at
low lifts and increased by a factor of about 2½ as Mach number
was increased from 0.9 to 1.2. The apparent stability exhibited a
148
gradual decrease as lift increased and mild pitch-ups occurred at

Mach numbers above 0.95….
Difficulty was experienced in performing smooth longitudi-
nal maneuvers. This condition appeared to result from the com-
bination of control system, pilot, airplane, and their dynamic
characteristics; however, additional tests are required to deter-
mine the primary cause of the lag and oscillations experienced.
Unaccelerated stalls appeared stable in all configurations tested,
except at large angles of attack in the landing configuration where
some instability was evident. Roll-off tendencies, which became
more severe as the speed was decreased, were apparent in all con-
figurations. Data obtained during sideslips at Mach numbers
from 0.84 to 0.98 showed the apparent directional stability to be
positive and to increase with increase in Mach number. A smaller
degree of apparent stability existed for smaller angles of sideslip
than existed for larger angles.
Meager aileron effectiveness data obtained at Mach num-
bers of 0.89 to 0.98 indicate that the control effectiveness was
generally linear with deflection and exhibited little change with
increase in Mach number. Comparison of flight data with wind-
tunnel and rocket-model tests showed similar trends and good
quantitative agreement.29
With the completion of the six Air Force flights, the X-3 was turned over
to the NACA for planned research flights. NACA pilot Joseph A. Walker was
assigned as X-3 project pilot. After an engine inspection, instrument calibra-
tion, and maintenance work, Walker made his first X-3 flight (NACA flight
7) on August 23, 1954. During the flight, he made two windup turns, stalls
in both clean and landing configurations, and a dive to Mach 1.05. After the
flight, the vertical-tail strain gauges were calibrated. With both Walker and the
X-3 checked out, the NACA research flights could begin.30
On September 3, Walker successfully made NACA flight 8, which involved
speed runs, longitudinal maneuvers, and stalls. Data were collected at speeds
between Mach 0.70 and 1.1. A second flight by Walker was planned but was
canceled due to lakebed cross winds. Walker then completed NACA flights 9
and 10 on September 9. Despite the flight series’ successful start, the basic unre-
liability of the X-3 continued to hamper progress. Two flights were planned
for September 16. Walker took off and began to climb on NACA flight 11
but noticed that the right afterburner’s fuel consumption was too high. This
suggested a problem with the afterburner’s fuel-flow control, and Walker cut
the afterburners. Despite the failure, some useful data were salvaged from the
149
Probing the Sky
The X-3 parked on the ramp after ownership was transferred to the NACA. The Air Force mark-
ings were removed and NACA insignia added. Joe Walker, who also flew the X-1E, X-4, X-5, and
X-15, among other aircraft, was the only NACA research pilot to fly the X-3. He made 20 flights
and experienced its demanding flight characteristics. (NASA)
aborted flight. He performed 1-g stalls in a clean configuration with the leading
edge flaps in the 10° and full-down 30° positions and landed on the lakebed
without mishap. To correct the fuel-consumption problem, the ground crew
had to replace the afterburner and repair the fuel-flow control.31
Walker made five NACA flights in late October, collecting data on stability
and control; wing pressure distribution pattern; lift and drag; buffeting; and
wing, horizontal stabilizer, and vertical-tail loading. Walker also made fixed-
rudder aileron rolling maneuvers at Mach 0.9 and 1.05. On the plane’s 10th
NACA flight, on October 27, 1954, Joe Walker and the X-3 entered aviation
safety’s history book, and their milestone came nearly at the price of both
plane and pilot.32
The Inertial Coupling Crisis

In June 1948, NACA Langley researcher William Hewitt Phillips pub-
lished a technical note innocuously entitled “Effects of Steady Rolling on
Longitudinal and Directional Stability.”33 It soon proved a remarkably prescient
150
and important report, one of particular significance for the future design of
practical high-speed aircraft.
In his introduction, Phillips set forth why studying the inertial character-
istics of the new generation of high-speed airplanes—airplanes that had long
fuselages and increasingly smaller wings—was important:
When an airplane rolls around an axis…not aligned with its

longitudinal axis, inertial forces are introduced which tend to
swing the fuselage out of line with the flight path. These forces are
ordinarily neglected when the usual theory of lateral stability of
aircraft is used to calculate the motion of an airplane in a roll. This
assumption is probably justified for the case of most conventional
airplanes because inertial forces involved are small compared with
aerodynamic forces on the airplane. Design trends of very high-
speed aircraft, however, which include short wing spans, fuselages
of high density, and flight at high altitude, all tend to increase the
inertial forces due to rolling in comparison with the aerodynamic
restoring forces provided by the longitudinal and directional sta-
bilities. It is therefore desirable to investigate the effects of rolling
on the longitudinal and directional stabilities of these aircraft.34
As a consequence of these phenomena, Phillips wrote, “The rolling

motion introduces coupling between the longitudinal and lateral motion of
the aircraft. An exact solution of this problem is very complicated because
of the large number of degrees of freedom involved.”35
At the same time Phillips was writing his paper, Fleming and the Douglas
engineers were in the midst of designing the X-3. The Douglas aircraft had a
fuselage 66.75 feet long containing two jet engines, fuel tanks, air condition-
ing, research instrumentation, ducts, control cables, landing gear, and a host
of other systems. Yet its wingspan was just 22.69 feet. Douglas, in short, was
building the airplane Phillips had imagined.36
Phillips’s idea was still considered a theoretical one at the time of his
writing, but there was already a crucial data point from the XS-1 program
that applied. In the summer of 1947, during falling-body tests of the XS-1,
a model of the rocket plane veered wildly off course and disappeared after
having been dropped from a high-flying B-29. Postflight analysis of opti-
cal and telemetric records indicated that, as Phillips recalled, “some kind
of gyroscopic effect” had taken place, causing the XS-1 to roll, pitch, and
tumble, falling so far off its predicted trajectory that it was never found.37
Out of this incident, Phillips began an analytical study that resulted in his
151
Probing the Sky
1948 technical note, which introduced the aeronautical world to the expres-
sion “inertial coupling.”38
A hint of problems to come was mentioned briefly in a memorandum by
Hartley Soulé about Douglas engineer Harold F. Kleckner’s visit to Langley
on September 5, 1951. Soulé’s text read:
Early computations of the lateral stability of the X-3 airplane at

supersonic speeds indicate that the relation between rolling and
yawing motions would be different from…past experience, pri-
marily because of the high airplane length to wingspan ratio with
the resulting large differences in the moment of inertia and rolling
as compared with yaw and pitching. For this reason, these charac-
teristics have been the subject of concern although it is not definitely
known that the different characteristics would be undesirable.39
(Emphasis added.)
Slightly less than a year later, in August 1952, technicians at the NACA’s
Pilotless Aircraft Research Division (PARD) located on Wallops Island, VA,
fired a rocket-boosted model of an early short-fin D-558-2 model fitted with
a small rocket thruster in its nose to induce combined rolling, pitching, and
yawing motions typical of inertial coupling as the model decelerated below the
speed of sound. After it fired, the D-558-2, as expected, “coupled” and expe-
rienced the combined roll, pitch, and yaw motions. But they did not dampen
out. Rather, they grew in severity, and the D-558-2 model was completely
out of control by the time it impacted in the Atlantic Ocean off the Virginia
coast. As the XS-1 model had shown 6 years earlier, and the D-558-2 model
confirmed, roll-coupling was a far from innocuous phenomenon.40
On December 12, 1953, during a Bell X-1A flight to Mach 2.44 (1,612
mph) at 74,200 feet, Air Force test pilot Major Chuck Yeager nearly perished
when the speeding rocket plane violently coupled, tumbling over 50,000
feet before Yeager managed (in a feat of superlative airmanship unequaled in
flight-testing history) to recover the airplane safely to level flight and return to
Edwards.41 Inertial coupling had struck a piloted aircraft and, had it not been
for its extraordinary pilot, likely would have destroyed it. In any case, during
his glide earthwards, Yeager confided to listeners that “if I’d had [an ejection]
seat, you wouldn’t still see me in this thing.”42 Because the X-1A was, like its
XS-1 predecessor, thoroughly instrumented, the NACA was able to analyze
the flight and the onset of the coupling departure in detail, issuing a thorough
report on the episode.43
While NACA analysts pondered their experience with models and the
X-1A, events were moving swiftly and ominously in the fast-paced world of
152
Another overhead shot of the X-3 on the back ramp. Unlike the D-558-I and D-558-II, the X-3
was fitted with a downward-firing ejection seat. The seat was lowered to allow the pilot to enter
the cockpit then returned to its original position once he was seated. A similar design was used
in the early F-104s, with poor results. (NASA)
fighter aircraft development. North American had designed a new jet fighter,
the F-100A Super Sabre, for the supersonic era. Highly streamlined, with an
afterburning J57 turbojet and a 45° swept wing, it first flew at the end of May
1953. Slightly over a year later, the Tactical Air Command activated its first
F-100A fighter wing, at George AFB, Victorville, CA, just 35 miles southeast
of Edwards, over the objections of Frank Everest and other Edwards test pilots,
who thought the plane needed more study and possible modifications to increase
its stability at supersonic speeds. Almost immediately, some pilots encountered
disturbing, unsettling motions at high speed. On October 12, 1954, disaster
struck. North American Aviation test pilot George “Wheaties” Welch took
off from Palmdale, CA, in an F-100A, climbed high over Rosamond, dove to
supersonic speed, and then began a rolling dive pullout. The F-100A abruptly
yawed, rolled, and pitched out of control, disintegrating and killing its pilot.
Inertial coupling had claimed its first, but not its last, victim.44
As accident investigators analyzed the wreckage of Welch’s ill-fated F-100A,
Joe Walker and the X-3 continued probing the behavior and performance
of the sleek research jet. On October 27, Walker initiated an abrupt left roll
153
Probing the Sky
at Mach 0.92 at an altitude of 30,000 feet. For 5 wild seconds, as the X-3
rolled, its nose pitched up and simultaneously yawed until its motions finally
damped out. As NACA researchers Richard E. “Dick” Day and Jack Fischel of
the High-Speed Flight Station dryly reported afterwards (describing the time
history of the maneuver):
While the ailerons are deflected for the aileron roll, a favorable
sideslip angle is generated, together with a rather large increase in
angle of attack. (The initial decrease in angle of attack is probably
attributable to pilot stabilizer control input.) At time 3.8 seconds,
even though the pilot is now applying 10° right aileron control,
left rolling velocity increased and exceeded 5 radians/sec accom-
panied by violent pitching and sideslipping motions. During this
uncontrollable phase of the maneuver, an angle of attack of 20° and
left sideslip angle of 16° were encountered. It might be of interest
to note that the onset of the violent maneuver coincided with the
attainment of the angle of attack (α = 80) at which unpublished
flight data indicates the occurrence of reduction of longitudinal
stability. Also, the angle of attack of 8° corresponds to the angle of
attack at which a reduction in the measured wing lift slope occurs;
therefore large wing loads were not experienced at the maximum
angles of attack. After the primary rolling motion has subsided
at t = 5 seconds, the large lateral and longitudinal motions damp
fairly well.45
Walker’s next foray with inertial coupling—during another left roll at Mach
1.05—was far more violent, and even the dry language of aeronautical engi-
neering cannot mask the danger and drama attending his encounter:
During this maneuver the favorable sideslip builds up rapidly

with roll velocity and peaks at 21° at the time the airplane ceases
rolling left. This large sideslip angle results in about 2g transverse
acceleration. Near the time at which maximum sideslip occurs
(t ≈ 4.0 sec) a large divergence in pitch develops in the negative
direction which attains about –6.7g. The pilot applied control to
stop this pitch-down and immediately reduced the control deflec-
tion, but was unable to avoid obtaining 7g when the airplane
pitched up. Maximum wing loads measured during the maneuver
did not approach or exceed the design limit load; however, the
fuselage load obtained from airplane weight and acceleration, and
horizontal tail loads, and wing load showed maximum values of
154
63,000 pounds. These maximum values approximated the limit

design total load of the fuselage. The measured horizontal tail
loads were near limit design load. The maximum measured verti-
cal tail loads reached approximately 50 percent of the limit load at
sideslip angles of 21°. It may be noted that a reduction in vertical
tail load with sideslip was experienced at sideslip angles about
8°. In this maneuver…when the rolling stopped, the airplane
motions quickly damped.46
Fortunately, the plane was at such a high pitch angle that the wing was
effectively robbed of lift, thus not exceeding its limit loads. But the fuselage,
as noted above, had reached, though not exceeded, its limit load. Following
this flight, the X-3 was grounded for an extensive structural inspection, not
flying again for 11 months.
In the wake of both the F-100A crash and the near loss of the X-3, Scott
Crossfield was assigned to fly inertial coupling research missions in an F-100A
delivered to the High-Speed Flight Station for research. Crossfield made a total
of 45 flights to probe the phenomenon. During one flight, the forces were so
severe that Crossfield suffered a cracked vertebra in his neck.47
In the research memorandum detailing the X-3’s inertial coupling incidents,
there is a description of an F-100 flight, this one made at 30,000 feet and Mach
0.70, initiated by an abrupt left aileron roll:
As peak aileron deflection is attained, there is a steady develop-

ment of left (adverse) sideslip and a progressive decrease in angle
of attack. Between t = 3 seconds and t = 4 seconds, the divergence
rates are accelerated considerably and negative angles of attack
greater than 16° (–4.4 g) and the left sideslip angles as large as
26° were reached in the more violent stages of this maneuver.
Maximum vertical tail loads of 5,500 pounds were measured at
a sideslip angle of 26°.48
Day and Fischel summed up the results of the X-3 and F-100 inertial cou-
pling tests by noting that “[t]he behavior of the two airplanes in the aileron
rolls is similar in that large cross-coupling effects are evidenced. The airplanes
are loaded primarily along the fuselage, particularly in the case of [the X-3],
so that considerable inertial coupling is expected.”49
In addition to analyzing the data from the X-3 and Crossfield’s F-100 flights,
a new tool was used for the first time in the inertial coupling investigation:
the first computer flight simulations. Analog computers were programmed by
Richard E. Day to simulate the inertial conditions that led to the incidents.
155
Probing the Sky
In an interview, Day commented on the value of Phillips’s theoretical paper

on inertial coupling:
It was almost essential. It was great, because nobody had any idea
what had happened…. So once the F-100 and X-3 got into roll
coupling, somebody either at Edwards or back at Langley said
Bill Phillips had written a theoretical report on this. So Walt
Williams sent Hubert Drake and Joe Weil back to Langley to
talk with Phillips.50
The inertial coupling problem with the F-100 was solved by making the
vertical tail taller and increasing the wingspan by 2 feet. The grounding order
was partially lifted in February 1955. Subsequently, automatic dampers and
stability augmentation systems were added to supersonic aircraft control systems
to prevent the inertial coupling from occurring. Even so, until the advent of
advanced stability augmentation and electronic flight controls, the challenge
of inertial coupling imposed serious handling-quality limitations on new high-
performance aircraft such as the McDonnell F-101 Voodoo and the Lockheed
F-104 Starfighter. The legacy of the inertial coupling crisis could be found in
the proliferation of ventral fins and large vertical fins on advanced fighters (and
eventually twin vertical fins, beginning with the MiG-25 and the Grumman
F-14A Tomcat, which are now found on many, though not all, of the world’s
high-performance supersonic military aircraft).51
The X-3 did not fly again until September 20, 1955, and when it did, it
marked the beginning of the final phase of the aircraft’s useful life. Walker
continued as project pilot. Research goals included collecting data on static
longitudinal stability and control, wing and tail loads, and pressure distribu-
tion, but NACA researchers were careful not to probe further into the inertial
coupling arena. Six more flights were made in October of 1955 focusing on
directional stability and control and tail loads. The October 12 flight was
marred by the inadvertent deployment of the drag chute in flight, although
the plane landed safely.
The X-3 made its next flight on December 13, 1955, but the flight was cut
short when a pressure probe broke off and damaged an engine. The aircraft
was grounded for repairs until the following spring. It next took to the air on
April 6, 1956, for pressure-distribution measurements. That day, there was an
in-flight abort when Walker smelled smoke in the cockpit and landed quickly.
The problem was traced to an electrical failure of test instrumentation in the
nose. The damage was limited to charred wiring, and the airplane was again
repaired. The final flight of the project was made May 23, 1956, to collect
lateral control data. With the research flights complete, the aircraft was retired
156
and sent to the Air Force Museum, where, even today, it impresses visitors with
its sleekness and purity of line.52
An Assessment
Before the X-3 flew, researchers already realized that the ambitious goals set for
the aircraft would never be met, not through the fault of its manufacturer or
its creators, but because of the regrettable failure of its engine manufacturer to
deliver suitable engines. Thus, the desired data on aerodynamic heating, inlet
duct and scoop airflow, stability and control of high-fineness-ratio fuselages,
and aerodynamics of low-thickness-chord-ratio wings at Mach 2—the raison
d’être of the program—were never collected.53
The X-3 was difficult to fly—John McTigue, a former High-Speed Flight
Station engineer, recalled that following one landing, Joe Walker had climbed
out of the aircraft and thrown his helmet to the lakebed in apparent frustration
with the plane—though this reflected as much the unknowns and challenges
of control system design in the early supersonic era as it did any inherent flaw
in the system itself. Specifically, the system suffered from control lag, a flawed
automatic load feel system, longitudinal oscillations that were difficult to con-
trol, and poor damping of oscillations. In light of these control system short-
comings, a question arises. If the J46 jet engines had been available or the pair
of LR-8 rocket engines had been fitted, could the pilot have even controlled
the aircraft at Mach 2? Inertial coupling came close to destroying the X-3 at
just over Mach 1, nearly destroyed the X-1A at Mach 2.44, and did destroy the
Bell X-2 at Mach 3. At Mach 2, the loads on the X-3 would have been much
higher; it is reasonable to wonder how long the airplane would have remained
intact at that speed at the first hint of roll coupling.
The relative “research worth” of the X-3 compared with that of the three
Douglas D-558-2 research aircraft can also be measured by the number of
flights made with each aircraft. The single X-3 accumulated only 52 flights,
and the three D-558-2s made 313 flights in total.54
Given all this, the X-3 might be easily dismissed as an outright failure—
underpowered, exhibiting poor control characteristics, and unable to meet
performance or research goals. Yet the X-3 provided critical data on inertial
coupling, a goal never planned and one involving an aircraft design issue no
one knew existed. Its contribution was a happy accident. The D-558-2’s swept-
wing design had inadvertently allowed the testing of methods for preventing
pitch-up. In the same way, Douglas’s efforts to build a Mach 2 aircraft resulted
in a design that concentrated mass in the fuselage. Entirely by accident, this
resulted in production of an aircraft susceptible to inertial coupling. This, in
157
Probing the Sky
The transport vehicle on the lakebed with the X-3. Once at the takeoff point, the X-3 was rolled
off the trailer and prepared for flight. (NASA)
turn, facilitated the discovery of an as yet unknown problem with high-speed

aircraft. Since modern jet fighters would be built to resemble the F-100 in
their concentration of mass in the fuselage, with swept, low-aspect wings, the
X-3’s contributions were invaluable to virtually every new fighter coming off
the drawing boards.
The best illustration of this is seen in the iconic 1950s fighter, the Lockheed
F-104, to whose design the X-3 contributed immediately. Both aircraft had
thin, short, razorbladelike wings and heavy fuselages. The development of the
F-104 overlapped with the X-3 research flights, and the Air Force insisted that
Douglas provide Lockheed with the X-3 plans. The F-104 was designed as a
lightweight air-superiority fighter capable of reaching high altitudes and engag-
ing Soviet fighters. The initial contract for two XF-104 aircraft was signed on
March 11, 1953, and the first XF-104 flight was made on February 28, 1954.
The speed with which the XF-104 went from drawing board to flying aircraft
was due to the X-3’s development, which proved the aerodynamic validity of
the low-aspect-ratio wing, performance predictions, and other details.
A RAND study of the F-104’s development underscored this assessment,
observing:
The F-104 history illustrates that research and development in one

program can have a great carry-over value in another. Lockheed’s
success in building and flying a prototype less than a year after
158
go-ahead would very probably not been possible without the

knowledge derived from the Douglas X-3 program. Although
the value of this experimental effort in the F-104 could hardly
have been anticipated when Air Force money was advanced to
finance the program, nevertheless the value to the Air Force of the
X-3 program extended far beyond the immediate results achieved
with it.55
The F-104 was used by the U.S. Air Force and by allied air forces in West
Germany, Canada, Belgium, the Netherlands, Italy, and Japan. More than
1,400 F-104s of all variants were built by Lockheed and under license. The low-
aspect-ratio wing design was incorporated into other fighter designs as well,
including the Northrop F-5, which saw service in 15 air forces, including those
of South Vietnam, Greece, Iran, South Korea, the Philippines, Nationalist
China, Turkey, Norway, and Libya.56 In its own way, this, too, represented a
legacy of the X-3.
159
Probing the Sky
Endnotes
1. Opening quotation from the pilot transcript of the X-3’s second
flight, October 31, 1952, DFRC archives.
2. The AAF development contract was W33-038-ac-10413, dated 30
June 1945, and the aircraft received project designation MX-656.
For the detailed development history of this airplane, see Richard
P. Hallion, “Serendipity at Santa Monica: The Story of the Douglas
X-3,” Air Enthusiast Quarterly 4 (October 1977): 129–136.
3. Douglas Aircraft Company, “Model X-3 Mock-Up Conference
Airplane Descriptions and Illustrations,” December 6, 1948, p. 7.
4. Ibid., pp. 8, 13–14, 38, 42–43; Frank N. Fleming, “Evolution of
the Configuration of the X-3 Supersonic Research Aircraft” (Santa
Monica, CA: Douglas Aircraft Company, August 1, 1949), pp. 1,
4–18.
5. Douglas Aircraft Company, “Model X-3 Mock-Up Conference
Airplane Descriptions and Illustrations,” pp. 8–10.
6. Thomas A. Marschak, The Role of Project Histories in the Study of
R&D, Report P-2850 (Santa Monica: The RAND Corporation,
January 1964), p. 43.
7. Marschak, Role of Project Histories, p. 44; see also James St. Peter,
The History of Gas Turbine Development in the United States…A
Tradition of Excellence (Atlanta, GA: International Gas Turbine
Institute of The American Society of Mechanical Engineers, 1999),
pp. 139–140. Nor was the X-3 the J46’s only “victim”: it helped
limit the Navy’s Vought F7U-3 Cutlass, rendered the entire first
production run of McDonnell F3H-1 Demons both dangerous
and useless, led to the abandonment of the Lockheed XF-90, nearly
crippled the Navy’s F3D-2 Skyknight, and forced Vought to lay off
over 2,500 production workers, resulting in widespread criticism
of Westinghouse. Though the company ostensibly remained in the
jet engine business for another decade, it dropped out of the first
rank of engine manufacturers and never matched the success it had
enjoyed with its J34, an excellent design. The tragedy is that the
X-3 appeared too early to employ a magnificent small afterburn-
ing turbojet, the General Electric J85 (the power plant for the later
Northrop F-5/T-38 family). At half the weight and smaller even
than the J34, the J85 had substantially higher performance. With
it, the X-3 could clearly have reached Mach 2—but the J85 was not
ready for flight testing until almost the end of the decade.
8. Hallion, “Serendipity at Santa Monica,” pp. 131–132.
160
9. Memorandum for Chief of Research, “Proposed Modifications of

X-3 Airplane to Rocket Power and Air Launch,” October 24, 1952.
10. Hallion, “Serendipity From Santa Monica,” p. 132; for the F-104
benefit, see Marschak, Role of Project Histories, pp. 85–86, which
notes, “Lockheed’s task was made substantially easier by the avail-
ability of information, particularly wind-tunnel data, gathered in
the course of the Douglas X-3 program [whose experience] had a
carry-over value for the F-104 program.” Richard P. Hallion recalls
a conversation with Melvin B. Zisfein, Deputy Director of the
National Air and Space Museum, who recalled seeing X-3 wind tun-
nel reports handed to Lockheed staff for reference as they worked on
the F-104’s design.
11. Hallion, “Serendipity at Santa Monica,” pp. 132, 136.
12. Richard E. Day and Jack Fischel, “Stability and Control
Characteristics Obtained During Demonstration of the Douglas X-3
Research Airplane,” NACA RM H55E16 (July 21, 1955), p. 6.
13. NACA HSFRS, “X-3 Status Report for Period 14 September to 18
October 1952,” DFRC Archives. Unless otherwise noted, all NACA
status reports are located in the DFRC archives.
14. NACA HSFRS, “X-3 Status Report for Week Ending 25 October
1952.”
15. NACA HSFRS, “X-3 Weekly Status Report for Week Ending 1
November 1953”; “X-3 Weekly Status Report for Week Ending
15 November 1953”; “X-3 Weekly Status Report Week Ending 25
April 1953.”
16. NACA HSFRS, “X-3 Weekly Status Report Week Ending 2 May
1953.”
17. NACA HSFRS, “X-3 Weekly Status Report Week Ending 20 June
1953,” p. 1.
1953,” p. 1.
1953,” p. 1.
1953,” p. 1.
21. NACA HSFRS, “X-3 Weekly Status Report Week Ending 18 July
1953.”
22. Douglas Aircraft Company, Flight Test Division, “Flight Report No.
7,” 1953, DFRC Archives.
161
Probing the Sky
23. NACA HSFRS, “X-3 Weekly Status Report Week Ending 25 July
1953”; Douglas Aircraft Company, Flight Test Division, “Flight
Report No. 8 (1953).”
24. NACA HSFRS, “X-3 Weekly Status Report Week Ending 1 August
1953.”
25. Douglas Aircraft Company Testing Division, “Flight Report No.
23,” 1953, DFRC Archives.
26. Memorandum for Research Airplane Projects Leader, “Progress
Report for the X-3 Research Airplane for the Period November 1
to November 30, 1953”; “Bridgeman,” http://www.astronautix.com/
astros/brigeman.htm, accessed October 6, 2010.
27. Memorandum for Research Airplane Projects Leader, “Progress
Report for the X-3 Research Airplane for the Period July 1 to July
31, 1954.”
28. Lt. Col. Frank K. Everest, as told to John Guenther, The Fastest Man
Alive (New York: E.P. Dutton, 1958), pp. 158–160.
29. Day and Fischel, “Stability and Control Characteristics Obtained
During Demonstrations of the Douglas X-3 Research Airplane,” pp.
1–2.
30. Memorandum for Research Airplane Projects Director, “Progress
Report for the X-3 Research Airplane for the Period August 1 to
August 31, 1954.”
Report for the X-3 Research Airplane for the Period September 1 to
September 30, 1954.”
Report for the X-3 Research Airplane for the Period October 1 to
October 31, 1954.”
33. NACA Technical Note 1627. For this remarkable aeronautical
scientist, see W. Hewitt Phillips, Journey in Aeronautical Research: A
Career at NASA Langley Research Center (Washington, DC: NASA,
1998).
34. William H. Phillips, “Effects of Steady Rolling on Longitudinal and
Directional Stability,” NACA Technical Note No. 1627, June 1948,
pp. 1–2.
35. Ibid.
36. Fleming, “Evolution of the Configuration of the X-3 Supersonic
Research Aircraft,” pp. 10, 20–23; High-Speed Flight Station,
“Flight Experience with Two High-Speed Airplanes Having Violent
Lateral-Longitudinal Coupling in Aileron Rolls,” NACA RM
H55A13 (February 4, 1955), pp. 7–8.
162
37. Quoted in Richard P. Hallion, “Sweep and Swing: Reshaping the

Wing for the Jet and Rocket Age,” in Richard P. Hallion, ed.,
NASA’s Contributions to Aeronautics, vol. 1: Aerodynamics, Structures,
Propulsion, Controls (Washington, DC: NASA SP-2010-570-Vol 1.
2010), p. 38.
38. As well, the more descriptive “roll coupling” was used.
39. Memorandum for Chief of Research, “X-3 Airplane Visit of Harold
F. Kleckner of the Douglas Company to Langley on October 18,
1951,” p. 2.
40. Hallion, “Sweep and Swing,” pp. 39-40; James H. Parks,
“Experimental Evidence of Sustained Coupled Longitudinal and
Lateral Oscillations from a Rocket-Propelled Model of a 35° Swept
Wing Airplane Configuration,” NACA RM L54D15 (May 28, 1954).
42. Yeager transcript and pilot report, 23 December 1953, and J.L.
Powell, “X-1A Airplane Contract W33-038-ac-20062, Flight Test
Progress Report No. 15, Period from 9 December through 20
December 1953,” Bell Aircraft Corporation Report No. 58-980-019
(February 3, 1954), in AFFTC History Office archives.
43. Hubert M. Drake and Wendell H. Stillwell, “Behavior of the Bell
X-1A Research Airplane During Exploratory Flights at Mach
Numbers Near 2.0 and at Extreme Altitudes,” NACA RM H55G25
(September 1, 1955).
44. Everest and Guenther, The Fastest Man Alive, pp. 19–22; Bill
Gunston, Early Supersonic Fighters of the West (New York: Charles
Scribner’s Sons, 1975), pp. 146–154; Knaack, Post–World War II,
pp. 115–116.
45. NACA High-Speed Flight Station (HSFS), “Flight Experience with
Two High-Speed Airplanes Having Violent Lateral-Longitudinal
Coupling in Aileron Rolls,” NACA RM H55A13 (February 4,
1955), p. 4.
46. Ibid.
47. Crossfield and Blair, Always Another Dawn, pp. 198, 199.
48. HSFS, “Flight Experience with Two High-Speed Airplanes,” p. 5.
49. Ibid.
50. Peebles, ed., The Spoken Word, pp. 68–69.
51. Ibid.
52. Hallion, “Serendipity at Santa Monica,” p. 136; Hallion and
Gorn, On the Frontier, p. 398. More information on the X-3
research results can be found in the following NACA research
memorandums: Harriet J. Stephenson, “Flight Measurements of
163
Probing the Sky
Horizontal-Tail Loads on the Douglas X-3 Research Airplane,”

NACA RM H56A23 (April 18, 1956); William L. Marcy, Harriet
J. Stephenson, and Thomas V. Cooney, “Analysis of the Vertical-Tail
Loads Measured During a Flight Investigation at Transonic Speeds
of the Douglas X-3 Airplane,” NACA RM H56H08 (November 9,
1956); Jack Fischel, Euclid C. Holleman, and Robert A. Tremant,
“Flight Investigation of the Transonic Longitudinal and Lateral
Handling Qualities of the Douglas X-3 Research Airplane,” NACA
RM H57I05 (December 5, 1957); William L. Marcy, “High-Speed
Loads Measured on the Douglas X-3 Research Airplane,” NACA
RM H57L08 (February 24, 1958); and Earl R. Keener, Norman
J. McLeod, and Norman V. Taillon, “Effect of Leading-Edge-Flap
Deflections on the Wing Loads, Load Distribution, and Flap Hinge
Moments of the Douglas X-3 Research Airplane at Transonic
Speeds,” NACA RM H58D29 (July 15, 1958).
53. Hallion, “Serendipity at Santa Monica,” p. 136.
54. “D-588 II Skyrocket,” http://www.globalsecurity.org/military/systems/
aircraft/d-588-ii.htm, accessed May 24, 2010; John McTigue inter-
view notes, March 9, 2010.
55. Marschak, Role of Project Histories, pp. 85–86; Knaack, Post–World
War II Fighters, pp. 175–176; Hallion and Gorn, On the Frontier,
pp. 57–58.
56. Knaack, Post–World War II Fighters, pp. 185–187, 288–289.
164
The X-4 Bantam, the second X-4 flown by the NACA between 1950 and 1953, after restoration
at the National Museum of the Air Force. The semi-tailless swept-wing design was influenced by
Northrop’s work on flying-wing aircraft, the belief that a lack of horizontal stabilizers would pre-
vent transonic handling problems, and the Me-163B Komet rocket-powered fighter developed
by Nazi Germany. (USAF)
166
CHAPTER 5
Versatile Minimalist:
The Northrop X-4 Bantam
We did a lot of things with the X-4 that

weren’t only involved in its being a tailless airplane.
—A. Scott Crossfield1
The tiny Northrop X-4 Bantam, one of the smallest piloted jet aircraft ever
built and flown, represented the confluence of several technological threads.
First was aviation pioneer and industrialist John Knudsen “Jack” Northrop’s
longstanding interest in flying-wing aircraft, dating to the late 1920s. By the
early postwar years, his dream had culminated in the graceful and futuristic
multiengine Northrop XB-35 and YB-49 flying wing bombers, the former
piston-powered and the latter a pure turbojet, undergoing flight testing at
Muroc. Another thread was the transonic revolution itself: studies by both
the Army Air Forces and the NACA indicated that a semi-tailless aircraft with
swept wings and a vertical tail, devoid of horizontal stabilizer and elevators,
might avoid the buffeting and controllability problems afflicting conventional
“tailed” aircraft (a byproduct of the deleterious interaction between the turbu-
lent flow streaming behind a wing and the tail surfaces located at the rear of
the aircraft). Instead, to compensate for the lack of conventional elevators for
pitch control, combined elevator and aileron control surfaces (called elevons)
were built into the wing. Operating symmetrically (both up or both down),
they controlled pitch, like elevators; operating differentially (one up, the other
down), they controlled roll, like ailerons.
Tailless aircraft (pure flying wings with no tail surfaces whatsoever) and
semi-tailless aircraft (ones having only a vertical tail) had a long heritage dating
to the pre–World War I era, and while (by the late 1940s) never securing a
dominant position among widely accepted aircraft configurations, they never-
theless had undergone significant technical evolution, rendering them increas-
ingly practicable.2 The advent of the swept-wing semi-tailless Messerschmitt
Me 163 Komet rocket-propelled interceptor had added to their allure. In an
era when 400-mph propeller-driven fighters climbing at 4,000 feet per minute
167
Probing the Sky
were considered remarkable, the 600-mph Komet—climbing to 30,000 feet

in little over 2 minutes and then streaking through Allied bomber formations
so rapidly that bomber gunners could not bring their weapons to bear—was
fantastic. At first glance, to a postwar world entranced by German wartime
aerodynamic and design accomplishments, the Me 163 seemed to point the
way to future flight. What was missed was that the Me 163 had demonstrated
extremely dangerous transonic handling qualities, as a result of its highly defi-
cient high-speed stability and control, and that this had severely limited its
military effectiveness. Coupled with other weaknesses, including an unreliable,
unstable, and highly dangerous propulsion system and limited endurance,
these deficiencies as a group ensured that the Me 163 remained more a curios-
ity than a serious threat.3
Fascination with the semi- or completely tailless configuration had spawned
some remarkable concepts for a variety of military and civil aircraft. In Britain,
this had led to the de Havilland DH 108 Swallow, a transonic swept-wing
research aircraft intended to function as a technology demonstrator for the
proposed DH 106 jet airliner.4 Inspired by the Me 163 and launched in late
July 1945, the Swallow took the fuselage and turbojet engine installation of
the de Havilland DH 100 Vampire jet fighter and joined it to a graceful swept
wing, with an equally graceful vertical fin installed on the aft fuselage. The
DH 108 had a more sharply swept wing compared to the Me 163’s gentler
sweep. Flight testing of the first of three DH 108s began in May 1946, with
the aircraft flown by Geoffrey de Havilland, Jr., son of the firm’s founder. The
first Swallow (TG 283) was reserved for low-speed trials, but the second (TG
306) was a high-speed machine. It achieved Mach 0.895 at 34,000 feet on its
fourth flight. During its early trials, it experienced a “short period” longitu-
dinal pitching oscillation at higher Mach numbers that was poorly damped.
On September 27, 1946, as young de Havilland practiced for a world airspeed
record attempt, TG 306 violently pitched at Mach 0.875 at only 7,500 feet (an
altitude that imposed very high structural loadings on the aircraft), breaking up
and killing him instantly. Though a third DH 108 Swallow (VW 120) eventu-
ally did exceed the speed of sound during a high-speed dive from 45,000 feet
to 23,500 feet on September 6, 1948 (becoming the first British-designed jet
airplane to exceed the sound barrier), it entailed another wild ride. At one point
(as test pilot John Derry noted), the plane exhibited an “extremely rapid and
completely unstable nose-down pitch” during which the VW 120 briefly went
past the vertical, attaining –3 g’s.5 Derry, an unusually skilled and courageous
airman (even by the standards of test pilots) managed to both exceed Mach 1
and then return safely to Earth. The Swallow had earned a more favorable place
in aviation history, but it still clearly was a dangerous airplane, and, indeed, all
three were eventually lost in uniformly fatal accidents.6
168
Versatile Minimalist: The Northrop X-4 Bantam
The first X-4 was used only for contractor demonstration flights, as it had a number of deficien-
cies. After the flights were completed, it was turned over to the NACA for spare parts. (USAF)
The X-4: Concept, Design, and Construction

On June 11, 1946, slightly over 3 months before Geoffrey de Havilland, Jr.’s
fatal crash in the second DH 108, Northrop and Army Air Forces’ representa-
tives had signed a contract to build and flight-test two examples of a semi-
tailless swept-wing aircraft. This marked the beginning of the XS-4 (later X-4)
program, which was undertaken as Air Force research and development project
MX-810. As with Britain’s luckless Swallow, the inspiration for the Northrop
design was largely that of the Me 163. Northrop constituted a natural contrac-
tor for the experimental airplane, as no other American company had such
extensive experience and insight into the special challenges and problems of
tailless and flying-wing aircraft. Unlike any of its X-series predecessors, the X-4
was specifically intended for subsonic testing.
Northrop entrusted design of the X-4 to a team led by chief project engi-
neer Arthur Lusk. Assisting Lusk were aerodynamicist Irving Ashkenas; chief
of structures A.M. Schwartz; and a team of specialists in weight and balance,
wing construction, propulsion, hydraulics, landing gear, instrumentation,
and avionics. Basic design work for the aircraft was completed at Northrop’s
Hawthorne, CA, plant by the fall of 1946, and the mockup was finished in
mid-November. Results of the inspection of the mockup by Air Force and
NACA engineers were generally favorable. The exception was one item found
by NACA personnel, who noted that no space had been allotted for spe-
cialized research instrumentation. In response, the Air Force and Northrop
169
Probing the Sky
(at the NACA’s suggestion) reduced the fuselage fuel tank’s size to provide the
requisite room.7
The aircraft employed two Westinghouse J30-WE-7-9 nonafterburning
turbojet engines, each producing just 1,600 pounds of thrust at sea level—the
same propulsion package used earlier in an unsuccessful Northrop flying-wing
fighter, the XP-79B, that had fatally crashed on its very first flight. Use of the
low-thrust engines underscored that the airframe had to be both small and
light; in its final configuration, a person could look into the cockpit with-
out having to use a ladder. To minimize weight, the X-4’s wings were made
of magnesium and incorporated integral fuel tanks that, when combined to
the tanks in the fuselage, carried 230 gallons of useable fuel, giving the little
jet a 45-minute endurance.8 In configuration, the X-4 resembled the British
Swallow, but with a more angular and lower-aspect-ratio 200-square-foot wing.
Its leading-edge sweep was 41.57°, it had an overall span of 26.83 feet (its
fuselage length was 23.25 feet), and it had an aspect ratio of 3.6. For safety,
the X-4 design team prudently incorporated hydraulically actuated split flaps
on the wing’s inner trailing edge. If the X-4 encountered dangerous transonic
pitching, its pilot could actuate the flaps, which would immediately extend
both above and below the wing, increasing its frontal area and generating such
drag that the X-4 would rapidly decelerate to firmly subsonic velocities before
it emulated the unfortunate DH 108 and pitched to destruction. Overall, the
little X-4 had a maximum weight of 7,050 pounds.9
Northrop completed the first X-4 (AF serial number 46-676) in June 1948,
and the aircraft underwent an engineering inspection with mixed results. Issues
surfaced regarding the landing gear up-and-down locks, the fuel system, and
the level of protection for control cables against engine failure or fire. Howard
Lilly had been killed the previous month in the second D-558-1 when its
engine disintegrated, causing inspectors to regard new aircraft more stringently.
The required changes to install armor shielding for its flight control system
and around the engines to prevent a disintegrating engine from puncturing
the fuselage or wing tanks took 4 months to make. The first X-4 finally arrived
at Muroc aboard a flatbed trailer on November 15, 1948. After taxi tests were
completed, it was ready for its first flight.10
Early Flights
Northrop test pilot Charles Tucker flew the customary proving flights before
the two X-4s were turned over to the Air Force and the NACA. The first
Northrop flight was made on December 16, 1948. Instrumenting the aircraft
promptly proved problematic due to the aircraft’s small size and Northrop’s
170
structural and engine-temperature-measurement requirements. As a result,

NACA stability and control instrumentation on the first flight was minimal.
Standard NACA instrumentation recorded altitude, airspeed, angle of side-
slip, right and left elevon position, and rudder position. The instrumentation
transmitted some data to a ground station, where they were recorded for later
analysis. (Monitoring in real time, however, was not possible.) Data taken
included rates of normal acceleration, altitude, airspeed, right and left elevon
position, and rudder position. The data were synchronized with a common
time stamp.
During the first flight, Tucker took data during takeoff and landing and
made in-flight records of the X-4’s speed as it went from 250 to 275 miles
per hour and back down to 225 miles per hour indicated air speed (IAS). A
subsequent research memorandum on the flight noted:
These data show that in the clean condition the airplane is slightly
unstable as shown by an upward deflection of the control required
for increasing speed. The pilot stated that it was impossible to trim
the airplane in the clean condition. With gear down the airplane is
stable for both center-of-gravity positions. There is an indication
that there may be some instability at high normal-force coef-
ficients with the gear down. However, the data at 145 miles per
hour were obtained in the landing approach just before contact
so there may be some effects of the proximity of the ground on
these data. Although these data are rather sketchy, they indicate
that the center of gravity should be moved forward.11
One issue that appeared on the first flight was the effectiveness of the rudder
control. The X-4 originally was fitted with an electronically operated system
with a four-speed actuator, an early example of applying electronic flight con-
trol (though not a computerized flight control) to an experimental airplane.
The maximum rate of rudder movement was 25° per second. Available evidence
indicates that this system was designed to serve as a yaw-damper stability-
augmentation system. Controlling yaw excursions was far from an innocuous
issue. Should too large a sideslip angle develop, the aircraft could stall and then
enter a dangerous spin.
Flying-wing aircraft had inherently poor spin characteristics, the danger
exacerbated by often violent longitudinal pitch changes that could result in
structural failure. Several Northrop flying wings already had been lost in spins
and pitching accidents, including an MX-324 glider, an N-9M subscale piloted
demonstrator for the XB-35, an experimental XP-56 tailless propeller-driven
fighter, and the sole XP-79B jet-powered flying-wing fighter. Then, on June 5,
171
Probing the Sky
1948, as Northrop readied the first X-4 for flight, came the dramatic loss of a
test crew aboard an eight-engine Northrop YB-49, resulting in the renaming
of Muroc as Edwards Air Force Base, after the ill-fated plane’s copilot, Captain
Glen Edwards. An investigation indicated that the YB-49 had stalled, entered
a spinning dive, and then had broken up as the crew tried to recover. Thus, in
the X-4, Northrop, the Air Force, and the NACA all had a compelling desire
to ensure that yaw rates could be limited, thus reducing the risk of the little
jet’s entering an unrecoverable or otherwise destructive spin.12
For that reason, researchers noted with concern the apparent lag test pilot
Charles Tucker reported in the X-4’s rudder response:
The pilot stated that the rudder control seemed to have consider-
able lag and the motion of the control was too slow…. [T]he rate
of rudder motion was about 25° per second, which [shows that
the] rate is considerably slower than the rate at which a pilot is able
to move the rudder pedals. The electrical system which operates
the rudder is arranged to give several rates of control movement
corresponding to the rate at which the pilot moves the pedals. The
rate of 25° per second is the maximum rate that is available to the
pilot at the present. The flight records showed this rate was used
in virtually all rudder applications indicating that motion to the
pedals was applied at a rate of 25° per second or greater. During
the first attempt for take-off, it was also indicated that this rate of
rudder movement is too slow for maintaining directional control.
The pilot reported excessive friction in the elevon control sys-
tem[,] which is an irreversible hydraulic system with artificial feel
for the pilot. He also reported that the aileron forces seemed very
heavy relative to the elevator forces[,] which on occasion caused
him to apply elevator control as well as aileron when attempting to
move only the ailerons. Since the aileron forces are about normal
it is believed that the pilot was given this impression by the exces-
sively light elevator forces. Since the elevator-force system depends
primarily upon elevator position, it could be expected that with a
change in center-of-gravity location sufficient to provide adequate
stick-fixed longitudinal stability, the aileron-elevator forces would
be proportioned satisfactorily.
Inspection of the sideslip records showed no evidence of snak-
ing oscillation over the range of speeds covered.13
Changes were made to the X-4 that reflected the results of the first flight.
Postflight analysis indicated that the aft position of the center of gravity was
172
Head-on view of the first X-4. The aircraft was the smallest of the early X-planes. The bubble
canopy was different from the V-shaped windshields of most other research aircraft. Like the
X-3, X-5, and XF-92A, the X-4 offered the advantage of an ejection seat. (NASA)
the cause of the poor longitudinal stability. Consequently, the airplane’s center
of gravity was shifted to 19.7 percent of the mean aerodynamic chord (MAC)
from 22.4 percent of the MAC. To resolve the rudder issues, the theoreti-
cally more advanced electrically operated rudder actuator was removed and
replaced with a much more reliable (and, somewhat counterintuitively, faster)
mechanical cable and bell-crank system, akin to the rudder control in a model
airplane. This doubled the rate of rudder travel to 50° per second. The control
system friction and springs were also checked by mechanics to eliminate the
excessive friction reported by the pilot. A check of engine vibrations during
ground run-up was planned. The date of the second flight was dependent on
lakebed conditions. In late January 1949, it was not expected that it would be
usable for about 4 weeks.14
In fact, because of lingering water on the lakebed, the X-4’s second flight
was not made until April 27, 1949. Tucker collected data during stabilized
speed runs at 170, 210, and 290 miles per hour IAS and at altitudes from
12,000 to 15,000 feet, but he had to terminate the flight prematurely because
of fuel transfer difficulties.15 The approach and landing data were also analyzed
from the second flight; the results reflected both good and bad characteristics.
Williams’s postflight memorandum noted, “The airplane possesses adequate
173
Probing the Sky
position stick-fixed longitudinal stability with the center of gravity at 19.7

percent mean aerodynamic chord as compared with the slight instability with
the center of gravity at 22 percent mean aerodynamic chord.”16
As for the approach and landing, Williams wrote,
[T]he airplane possessed slight positive longitudinal stability

with the landing gear down and flaps up as evidenced by the
increase in upward elevon deflection as the speed was decreased.
Approximately 14° of longitudinal control was used for landing[,]
which left adequate control for lateral motions of the airplane.
It should be pointed out, however, that this landing was made
well above minimum speed without flaps. It should be noted that
during the approach to landing, the small movements of the rud-
der caused a lateral oscillation that was slow to damp and even
continued at small amplitudes with the rudder held fixed…. [The
control was] held essentially fixed while the airplane oscillated in
sideslip…. The lateral oscillation has a period of approximately
2 to 3 seconds and…the oscillation damps to half-amplitude in
approximately 4 to 5 seconds.
The pilot was satisfied with the longitudinal stability with the
center of gravity at 19.7 percent mean aerodynamic chord. The
rudder control was considered adequate. Although the pilot did
not consider the poor damping of the lateral oscillation objec-
tionable at the speeds for which data are presented herein, he
encountered a poorly damped lateral oscillation at 290 miles per
hour which he considered very objectionable.17
Clearly, the X-4 had stability and control issues, reflecting its close-coupled
configuration, which virtually guaranteed poorly damped control response, low
inherent stability, and a greater than usual susceptibility to pilot-induced oscil-
lations (PIOs). But that it did is unsurprising since so much of what aeronauti-
cal engineers were doing at the time was new, and it was flying, as well, in both
the pre-stability augmentation and pre-fly-by-wire flight control eras. Between
April 27 and June 1, 1949, flights 2 through 6 were completed with the first
X-4. The flights generated data, but problems appeared with Northrop’s tem-
perature recorder and radio. After the fourth flight, the engineers initially
concluded that the interference was caused by the NACA telemeter system.
The telemeter was not used on the fifth and sixth flights, to verify this, but the
problems remained. Other difficulties were experienced during the sixth flight.
After takeoff, Tucker mistakenly left the recording instruments on through the
climb, leaving no recording film for the rest of the planned maneuvers, so they
174
Another view of the first X-4. The aircraft had a functional design, but it suffered from major
stability problems that the technology of the early 1950s could not correct. (NASA)
were aborted. Northrop engineers also reported the left engine had problems
during the flight, requiring that it be replaced. This grounded the aircraft for
2 months.
Such delays were often experienced with research aircraft. These were often
due, as in this case, to technical problems with equipment and systems. The
data instrumentation system was particularly critical. If it was not working,
the flight was pointless. Another cause was the flooding of the lakebed, which
occurred during most winters. Another was processing a backlog of data. This
was the job of female “computers,” who had to measure the traces on the film
manually, apply corrections and calibrations that transformed the traces into
numerical data, and plot these on graphs.
The first and second flights of the second X-4 (AF serial number 46-677)
were made in June. During the aircraft’s preflight preparation, all NACA instru-
ments were recalibrated. The telemeter in the X-4 burned out before the first
flight, however, and was not replaced before the flight was made. Some data
were collected, including a rudder release oscillation and a 2-g pull-up. But the
lack of telemetry didn’t matter since the primary goals for the flight were pilot
familiarization and instrumentation checkout. And in any event, both flights
were cut short due to fuel leaking from the wing filler caps.18
175
Probing the Sky
The second X-4 underwent modifications to correct the fuel siphoning.

That aircraft’s third flight was made on June 23 to test the fix. But as on the
two previous flights, fuel continued to be lost and the mission had to be cut
short. Despite the aborted flight, data were collected on a snaking oscillation
and two pull-ups of 2 g’s.
To identify the cause of the continuing fuel siphoning, instrumentation
was added to measure vent-line pressures under varying flight conditions.
Measurements were taken during the fourth X-4 flight, on June 30. In addi-
tion to diagnostic measurements, records were taken at speeds of 300 and 160
mph at different speed brake angles and at altitudes of 10,000 and 8,500 feet.19
While the first X-4 still remained grounded, Tucker made flights 5 and 6
in the second X-4 on July 8 and 12, respectively. The fifth flight provided data
on accelerated maneuvers at 30,000 feet and at 180, 260, and 300 mph; the
sixth was primarily a photographic mission, but data were also obtained on
different dive brake angles at 140 mph. Following this flight, the second X-4
was grounded, pending engine changes. The first X-4 resumed flights on July
26, 1949, the second aircraft on the following day. These constituted the sev-
enth flight for each aircraft. Again, a significant shortcoming was identified: a
series of directional stability runs was undertaken with the second X-4, but data
indicated that hands-off stability runs could not be made, as the time required
to trim the aircraft was excessive. As well, Northrop engineers revealed they
were “concerned about some component of the X-4 airplane which has proven
unsatisfactory in another installation.” But they “were unwilling to state at
this time what the exact trouble is.” A meeting between Northrop and NACA
personnel was held, and the Northrop contingent indicated that the landing
gear door locks were not satisfactory. The risk was that the doors could open
in flight at high speeds. Work began on developing a fix for the problem.20
To allow Tucker’s contractor flights to continue, the X-4s were restricted
from exceeding 300 mph IAS and 1.5 g’s of normal acceleration. Each aircraft
made its eighth and ninth flights in August. The landing gear door issue was
not the only problem the project faced. The second X-4 made a lateral-stability
flight on August 3 and a roll-rate measurement flight on August 5. When the
film from onboard instruments was developed following the flight, it was
discovered that the film drums had been improperly loaded and no data had
been collected. Additionally, the aircraft had experienced excess vibrations on
both flights, requiring that both engines be removed and replaced.21
This was successfully carried out, but Northrop managers requested fund-
ing to investigate the vibrations. Pending a decision on their request, the
second aircraft was temporarily grounded. The first X-4 remained on flight
status, but no flights were made through mid-September. This was due to the
removal of the NACA telemetry transmitter so as to allow the installation of
176
The research airplanes were not typically designed with ease of servicing in mind. The X-4 was
an exception. Its aft fuselage could be removed from the forward section, allowing access to the
jet engines. (NASA)
engine-temperature-measurement instruments requested by Northrop. Another

reason no flights were made was that Northrop engineers were still working
on a fix for the landing gear doors. Not until this was completed would the
aircraft make another flight.22
The second X-4 finally resumed flights on September 30. During flight 10,
Tucker made two stalls in the clean configuration and with the gear down.
Additionally, an accelerated stall was performed. Based on the flight data, it
was clear that modifications would have to be made to the landing gear doors
on both aircraft. The task was expected to take 3 weeks.
During the halt in flights, Northrop submitted a revised contractor test plan.
This sharply pared the original plan to a level sufficient only for proving the guar-
anteed performance requirements. This would involve six more flights totaling
about 15 hours. In reviewing Northrop’s proposed revision, Melvin Sadoff, an
NACA aeronautical research scientist at Muroc who wrote the X-4 report, noted
several requirements that would be necessary in order to provide the data that the
NACA needed for its research efforts: “With regard to the NACA’s final accep-
tance of the abbreviated Northrop program on the X-4, it is considered essential
that directional stability data be obtained up to the level flight high speed of the
177
Probing the Sky
airplane. The tests should be made at sufficiently large sideslip angles to provide
satisfactory proof of the structural integrity of the vertical tail at high speeds.”23
Sadoff also summarized the data from the second X-4’s 10th flight:
The results indicate that the 1 g “stalls” in both the clean and the
gear down configurations were mild and were accompanied by a
slight dropping of the right wing. The maximum lift coefficients
obtained were about 0.71 in both cases. There was no warning
prior to these stalls. Rapid recovery was effected with down ele-
vons. From the relative mildness of the airplane motion subsequent
to these stalls and because no appreciable buffeting was obtained,
it is believed that the stalls were not quite complete. Conceivably,
higher values of CLmax could be obtained by holding the right
wing up with the rudder control. An accelerated stall to about
1.6 g was made at about an indicated speed of 165 mph, (25 mph
higher than the 1 g stalls), and the data showed that the stall was
essentially complete. The right wing dropped fairly abruptly and
large up-elevons at the stall were ineffective in increasing the CLmax
above about 0.83. Moderate buffeting set in at the stall in the case
and persisted throughout the recovery. Recovery was again rapid
and complete with the down elevon movement.24
Sadoff also raised a more significant issue: whether the X-4 would be suitable
from an operational and maintenance standpoint. He listed a series of concerns:
(a) Engine vibration

(b) Engine availability and maintenance (including accessories)
(c) Fuel system peculiarities and sources of trouble
(d) Hydraulic system setup, including protection from vibration and
engine failure
(e) Availability of control in case of hydraulic system failure, etc.
Sadoff concluded, “From information available at present, item (a) will

not be remedied and will probably be a source of pilot discomfort, although
Westinghouse assures us the vibration does not affect the structural soundness
of the engine. Items (b) through (e) are currently being looked into by Messrs.
Collins, Griffith and myself.”25
The second X-4 did not make its 11th flight until November 29, 1949. The
results were a disappointment; after takeoff, Tucker retracted the landing gear,
but it did not lock properly. He aborted the flight, and the only data recorded
were of the approach and landing. The Northrop contractor test flights were
178
drawing to a close, with only four flights planned with the second X-4. Sadoff
stated that NACA and Air Force representatives needed to meet before the
completion of the Northrop flights and make a decision on the number of
additional flights the company would need to complete before the X-4s could
be accepted by the Air Force and the NACA.26
Northrop X-4 Flights Continue

After the delays in the early contractor flights, activity with the second X-4 picked
up in mid-December 1949. Flights 12 and 13 were both made on December 7,
flight 14 on December 9, and flight 15 on December 14. The most significant
event occurred on flight 15. As Sadoff summarized in the biweekly report:
During Flight 15 at about 0.80 Mach number and about 5g, an

inadvertent pitch-up occurred and the airplane acceleration built
up to about 6.5g before the pilot was able to regain control of the
airplane. During the recovery, the airplane went through a series
of violent pitching, rolling, and yawing oscillations, which appar-
ently were associated with the pilot’s manipulation of the controls
rather than an inherent dynamic instability of the airplane.
Flight 16 scheduled for Friday, December 16 was cancelled so
that the nose wheel door may be reworked preparatory to higher
speed flights. The interval before the next flight will also be used
to analyze the data obtained during Flight 15 so that we won’t
be proceeding to higher Mach numbers entirely ignorant of the
airplane’s behavior at lower speeds.27
The rapid flight schedule continued into the new year. With the nosewheel
door up-lock modified, the second X-4 returned to flight on January 13, 1950.
The day’s first mission, flight 16, was aborted soon after takeoff after the outside
canopy lock opened about an inch. The aircraft landed, the lock was repaired,
and flight 17 was made, which collected static-stability data at 35,000 feet at a
Mach number of about 0.8. Similar tests had been successfully made at 20,000
feet without incident. At the higher altitude, Tucker reported difficulty achieving
the higher acceleration in the test plan due to insufficient longitudinal control.
Once the test runs had been completed, Tucker began a rapid descent with the
dive brakes open and immediately encountered violent rolling and yawing oscil-
lations he could not control. Tucker closed the dive brakes, and the X-4 rapidly
sped up. The aircraft pulled away from the F-86 chase plane and reached a speed
of about Mach 0.91. Unfortunately, no data were recorded during the dive.
179
Probing the Sky
Engineers decided that the flight would have to be repeated, as the normal
acceleration data and the telemeter record were not useful. Further compli-
cations were engine malfunctions that occurred on both flights 16 and 17
and caused the aircraft to be grounded for 10 days pending an inspection.
Examination revealed that the right engine had sustained damage due to exces-
sive temperatures during flight 17.
While the second X-4 was grounded pending delivery of a new engine,
the first X-4 made its 10th and final flight on January 24, 1950. The flight’s
primary research goal was to check the aircraft’s climb performance at 25,000
feet. A secondary goal was the collection of limited longitudinal-stability data
in steady, straight flight at 25,000 and 10,000 feet. Finally, data on an oscil-
lation that had occurred with the dive brakes open was also on the flight-test
card during the descent. Tucker reported after landing that the friction in
all three controls was excessive and made the X-4 very unpleasant to fly. The
ground crew began an investigation of the control friction on both aircraft.28
Northrop flights with the second X-4 resumed in mid-February 1950 with
the completion of flights 18, 19, and 20. The two initial flights investigated the
aircraft’s longitudinal-stability characteristics in accelerated flight at speeds of
up to about Mach 0.84 at 30,000 feet. During the last run on flight 19, Tucker
reached Mach 0.88 in steady, straight flight. He reported “a very noticeable
buffeting, porpoising, and a yawing and rolling oscillation which were uncon-
trollable.”29 Tucker compared the buffeting to that of a washboard road, an oft-
repeated judgment that quickly became convenient shorthand to describe the
X-4’s transonic qualities. Sadoff noted in his report: “Although true buffeting
existed at this speed, what the pilot probably felt was the porpoising motion
which had a period of about 0.6 seconds. The yawing oscillation reached maxi-
mum double amplitude of about 6 degrees.”30
Flight 20 was made on February 17, 1950, but no data were collected
because the landing gear failed to lock up. At the time, Northrop and NACA
personnel expected the contractor flights to continue, but events intervened.
Changing of the Guard

Northrop contractor flights in the X-4 ended abruptly when the Air Force
issued an order on February 20, 1950, halting the flights. The following day,
Colonel E.W. Richardson (author of the halt order) and Major L.K. Cox came
to the High-Speed Flight Research Station to discuss with NACA engineers
their desires regarding the acceptance and future of the two aircraft.
The NACA engineers shared their concerns about remaining work.
Pull-up tests had been made by Tucker to the NACA’s satisfaction during
180
the contractor flights. The strength of the vertical tail, however, had not yet
been demonstrated. This involved making steady sideslips up to maximum
level-flight speed at 20,000 feet and was accomplished using either full rudder
deflection or 300 pounds of rudder pedal force.
NACA engineers felt that these sideslips would still not give a true indica-
tion of the vertical tail’s structural integrity until the diameter of the rudder
cables was increased. One-sixteenth-inch-diameter cables had been used in the
X-4s and were prone to excessive stretching, making it very difficult to obtain
reasonable values of sideslip at high speeds. While Northrop engineers argued
that the small cables were sufficiently strong, Colonel Richardson agreed with
NACA engineers who felt that larger-diameter cables were necessary. As was
often the case, money entered the picture. Colonel Richardson told the NACA
that the Air Force lacked the funding it would require to change the rudder
cables. The NACA agreed to install the larger cables but insisted the sideslip
tests had to be completed.31
The X-4 rudder cables were not the only safety issue. Ralph Sparks, a
Northrop engineer working on the aircraft, later recalled his concern about
the two dive brakes, each of which was operated by a separate hydraulic system.
If one of the systems failed and the pilot activated the speed brakes, one would
remain closed while the other opened, inducing dangerous asymmetrical loads
and forces that could not only throw the X-4 out of control, but perhaps break
it up as well. Sparks feared the tail might be torn off the aircraft. He discussed
his concerns with Walt Williams; Williams told Colonel Richardson and Major
Cox that it would be a simple matter to fix the problem and the repair could
be done at the same time the rudder cables were replaced.32
The repeated problems with the X-4 engines were also a concern. Williams
noted that the NACA was not happy about the “burping” of the J30 engines
or their excessive vibrations. However, discussions with Westinghouse and the
Air Force’s Power Plant Laboratory clarified that these conditions were not
inherently dangerous. And so the NACA engineers decided that the engines
could be operated without modification. They did insist that all engines used
on the X-4 have annealed turbine shafts, as both Westinghouse and the Navy
had issued technical orders requiring shafts to be annealed to reduce failures.
By that time, about one-half of the X-4’s engine stock had the annealed shafts.
Major Cox asked the NACA engineers if they had a preference as to which
of the two X-4 aircraft they would receive. They replied that “it would be highly
desirable” to get the second X-4. The primary reason for their choice was that
the second aircraft had already been fitted with the NACA instrumentation,
so no time would be lost reequipping it. Perhaps a more significant reason was
that the first X-4 was in poor mechanical condition compared to the second,
and the NACA personnel knew it.33
181
Probing the Sky
The second X-4, photographed from a chase plane. As with Northrop’s other flying wings, the
X-4 had a sleek appearance. It also shared the stability problems that these aircraft displayed.
Not until development of the computer fly-by-wire control system would semi-tailless aircraft be
practical. (USAF)
By the spring of 1950, the status of the X-4 aircraft was settled. The Air
Force originally requested that the NACA maintain both aircraft on flight
status, one to be flown by Air Force pilots and the other by the NACA, but the
NACA rejected this proposal. The first X-4 would be grounded and used for
spare parts, and the second would be kept on flight status. In addition, NACA
personnel would replace the existing 1/16-inch rudder cables with 1/8-inch cables
and modify the dive brakes’ hydraulic system on the second X-4. Once that
work was complete, Air Force test pilots would make a short series of evalua-
tion flights and make the high-speed sideslip tests of the vertical fin’s structural
integrity. NACA pilots would then begin research missions.34
Work on the aircraft modifications began in mid-June 1950. In addition to
the cable replacement and modifications to the hydraulic system, the engines
were replaced. The only major problem was a delay in the engine installa-
tion that was due to a fuel leak in the left wing. This was fixed in August,
and the installation was nearly complete. Separate from the X-4 modification
was preparation for the vertical fin structural tests. Before making any flights,
ground load tests would be made using the first X-4. Flight tests of the vertical
fin structure would be made once these tests were completed.35
The Air Force X-4 evaluation began on August 18 with flight 1 by Captain
Charles “Chuck” Yeager. (As with the X-3, these were listed as NACA flights
since NACA instrumentation was used.) The second and third flights, by Major
182
Frank Everest, were made on August 22. The third flight was aborted when
the main landing gear did not lock up, a problem the team thought it had
resolved. (The solution was to increase the cycling time of the main landing
gear). Flight 4 was made without incident. On flight 5, both Everest and Yeager,
the chase pilot, noticed oscillations of the wingtips and the outboard portions
of the elevons. This was especially noticeable on the left side. After Everest landed,
an examination showed the left elevon had about 1½° of play, the right elevon
½°. The X-4’s elevons were not mass balanced, and this caused them to flutter.
NACA ground personnel began work to eliminate the play in the elevons, for
aerodynamic flutter could pose an insidious threat, particularly if encountered at
higher speeds and with such magnitude that the ailerons were torn from the plane.
The Air Force evaluation now moved to tests of the X-4’s stall behavior. As
a safety measure, an 8-foot-diameter spin chute was added to the aircraft so
that if the X-4 got into a spin from which the pilot could not recover, he could
deploy the spin chute. As the chute opened, it pulled the tail up, forcing the
nose down into a dive. Airflow over the wings would be restored and speed
increased, allowing the pilot to recover.36
The Air Force flights were made in rapid succession. Flights 6 through 11
came between September 13 and 22, 1950, all of them made by Yeager. Flight
6 involved 1-g stalls, an accelerated stall to about 2 g’s, stall approaches with
the dive brakes at 20° and 30°, and some longitudinal and lateral dynamic
stability. Due to instrumentation failure, the flight had to be repeated. After
that flight, the spin chute was removed.
The Air Force pilots did have problems: flight 7 had to be aborted when the
canopy and landing gear did not lock closed, and flight 8 was a repeat of flight
6. On flight 9, a maximum-speed run was made to Mach 0.89, static-stability
data were collected, and an accelerated maneuver was made to the instability
boundary at Mach 0.7. Flight 10 suffered an instrumentation failure and had to
be repeated as flight 11. This involved accelerated maneuvers to the instability
boundary at Mach numbers of 0.6, 0.65, and 0.76; directional, longitudinal,
and lateral dynamic-stability data at Mach 0.5 and 0.7; and aileron rolls to the
right and left at Mach 0.7. Flight 11 marked the effective end of the Air Force
X-4 flights, although eight more flights were made, short ones on which no data
were taken. NACA pilot John Griffith made flight 15. The Air Force evalua-
tion was completed with flight 19 and totaled about 13 hours of flight time.37
Data from the Northrop and Air Force flights of both X-4 aircraft were
summarized in a December 1950 research memorandum:
The airplane was almost neutrally stable in straight flight at low

Mach numbers with the center of gravity located at about 21.4 per-
cent of the mean aerodynamic chord for the clean configuration.
183
Probing the Sky
Lowering the landing gear had no significant effect on the lon-

gitudinal stability. There was some indication that the stability
tended to increase for both configurations as the normal-force
coefficient was increased.
The airplane was longitudinally stable in accelerated flight
over a Mach number range of 0.44 to about 0.84 up to a normal-
force coefficient of about 0.4. At higher values of normal-force
coefficient and at Mach numbers of about 0.8 a longitudinal
instability was experienced.
The airplane does not meet the Air Force specifications for the
damping of the longitudinal oscillations. The pilot, however, did
not object to the low damping for small amplitude oscillations.
However, an objectionable undamped oscillation about all three
axes was experienced at the highest test Mach number of about
0.88[,] which may well limit the X-4 to this speed.
The theory predicted the period of short-period longitudinal
oscillation fairly well, while, in general, the theoretical damping
indicated a higher degree of stability than was actually experi-
enced. This disagreement was traced to a large error in the estima-
tion of the rotational damping factor.
The directional stability of the airplane was high and essentially
constant over the speed range considered, while the effective
dihedral increased considerably with an increase in normal-force
coefficient. The lateral- and directional-stability characteristics
estimated from wind-tunnel data compared favorably with the
flight results.
The damping of the lateral oscillation does not meet the Air
Force requirements for satisfactory handling qualities.
The dynamic lateral-stability characteristics were estimated
fairly well by the theory at low Mach numbers at a pressure alti-
tude of 10,000 feet. At 30,000 feet, however, and at [a] Mach
number above about 0.6, the theory indicated a higher degree of
stability than was actually experienced.
For the conditions covered in these tests, the stalling charac-
teristics of the airplane at low Mach numbers were, in general,
satisfactory. The stall was characterized by a roll-off to the right
and by moderate buffeting[,] which served as a stall warning.
The buffet boundary for the X-4 airplane, which was almost
identical to that for the D-558-2 airplane, showed a sharp drop-
off in the normal-force coefficient for the onset of buffeting as the
Mach number exceeded about 0.8.38
184
In addition to the Northrop/Air Force/NACA flights, the ground static-

load tests of the first X-4’s vertical tail, fin, and rudder were also completed on
September 21, 1950. The fin was tested to its design limit of 6,600 pounds.
The rudder was loaded to 1,500 pounds at the hinge line near the center hinge.
This was approximately 65 percent of a hypothetical combination of maximum
loads on the three rudder hinges and about equal to the maximum load on any
one hinge in a critical load condition. The results showed that the vertical tail
and rudder could withstand the aerodynamic forces they would experience on
future NACA research flights.39
The NACA’s X-4 Research Flights

With the Air Force flights completed, the second X-4 was grounded for instru-
mentation changes. In addition, several engine problems had appeared that
required both engines to be replaced, including governor and starter troubles.
This work grounded the aircraft until early November.40
Flights resumed on November 7, 1950, with NACA flight 20 flown by
Griffith. The pilot made a maximum-speed run to about Mach 0.88 to deter-
mine whether elevon motion was contributing to the aircraft’s porpoising.
Additional test maneuvers included gradual turns to the stall or instability
boundary at Mach 0.60 to 0.80, to further define that boundary, and rudder
kicks at Mach 0.70 to 0.80, to confirm deterioration of lateral damping. All
the tests were made at 30,000 feet. Though instrument malfunctions occurred
on the flight, they did not impair the collection of data.
NACA flight 21 was not made until November 17, 1950, again with Griffith
as pilot. This time, tests included turns to the instability boundary at speeds of
up to Mach 0.88. In addition, aileron rolls were made at Mach 0.40, 0.50, and
0.60. Problems were found in the postflight inspection: minor instrumenta-
tion errors had occurred, fuel leaks appeared in the wings, and the left engine
suffered from governor problems. These issues were fixed by early December.41
Flight 22, on December 6, 1950, was a pilot check for Major R.L. Johnson,
and no data were collected. NACA research pilot A. Scott Crossfield made
flight 23, which was for aircraft familiarization, the same day. Crossfield later
described his first experience with the X-4:
As the X-4 wobbled down the long, bumpy runway, I gingerly

felt out the controls. Then churning jets took hold, and the small
X-4 abruptly lunged into the air. Backing off the stall point, I
nosed her over gently and leveled out. Then I eased back on the
stick and the tiny tailless craft zoomed skyward like a winged
185
Probing the Sky
rocket. As predicted, at Mach .88 the X-4 broke into its gentle
but potentially dangerous porpoising motion. I opened the speed
brakes, and the X-4 slowed instantly, throwing me forward against
my shoulder restraining straps.
After about fifteen minutes in the air, I felt at home in the
X-4. The plane responded so well, in fact, that it was hard for me
to keep in mind that I was flying a marginally stable, experimental
race horse.42
The flight was cut short when the left engine suffered a flameout. The X-4’s
biweekly report stated that Crossfield was making a prolonged 4-g maneuver
at an altitude of 23,000 feet when the problem occurred. In his biography,
Crossfield wrote that he had pulled into a loop and, when going over the top,
both engines flamed out. He was able to restart the right engine but not the
left one. He made an emergency landing on the lakebed rather than the South
Base runway.43
Despite these difficulties, Crossfield began a series of flights to explore
the X-4’s basic handling qualities. The series began with NACA flight 24
on December 15, 1950. Sideslip and aileron roll data were supposed to
be collected, but the flight was cut short due to instrument problems. The
year’s activities came to a close with flight 25, made on December 28, for
lateral directional-stability, accelerated longitudinal-stability, and lateral con-
trol data at Mach numbers between 0.50 and 0.84. This was the first of 20
flights made to collect data on flight qualities, a series that lasted until May
29, 1951. All but three were made by Crossfield, the others by NACA pilot
Walter P. Jones.44
The initial results were written up by Walt Williams and Scott Crossfield
in 1952 as part of an overview of handling qualities of the X-1, D-558-1,
D-558-2, X-4, XF-92A, and F-86A. By this time, Crossfield understood
how difficult the X-4 was to fly. He and Williams described the X-4’s most
serious shortcoming:
With the X-4 airplane, an undamped oscillation about all three

axes at a Mach number of 0.88 has been experienced…. The
pitching appears predominant to the pilot. The small undamped
yawing oscillation at this speed induces a pitching oscillation at
twice the frequency of the yawing oscillation. The pitching is
apparently amplified because the natural frequency in pitch is
twice that of yaw…. These data were obtained from rudder kicks
and stick impulses, and incidentally, in all maneuvers where there
is yawing there is pitching at twice the frequency…. [A]t 0.9
186
Mach number the yawing oscillation diverges, rolling becomes

large, and the whole motion is intolerable. The violence is attested
by the control motions which result from accelerations on the
pilot. Lateral accelerations reached ±1 g. The irregularity of pitch-
ing is probably caused by the control motion.”45
This handling characteristic was judged to be a “Category I” problem,

as it was dangerous and imposed serious limitations on the operation of the
aircraft.46 The research memorandum continued, “Also in the first category,
the X-4 oscillations about the three axes are determined as having their origin
in very low to zero damping in yaw and by the fact that the ratios of natural
frequencies and coupled oscillations are similar. Largely because of these oscilla-
tions it was considered unreasonable to extend the speed beyond the maximum
Mach number reached, nearly 0.93.”47
A more detailed report of the X-4’s handling was written by Melvin Sadoff
and Scott Crossfield in 1954. The conclusions stated:
At low speeds marginal stability restricted the aft center-of-gravity

travel to 19 percent mean aerodynamic chord and low longitu-
dinal control power restricted the forward limit to 16.5 percent
mean aerodynamic chord yielding less than 3 percent permis-
sible center-of-gravity travel. The low longitudinal control power
within this center-of-gravity range limited the approach to 1g
stalls which was characterized by mild instability roll-off and nor-
mal response to recovery control.
Throughout the speed range, typical swept-wing instability
and buffet characteristics occurred at lower normal-force coef-
ficients than with tail-on airplanes of similar sweep.
At high speeds the X-4 characteristics deteriorated as follows:
• At Mach numbers above 0.76 a residual yawing and rolling motion

persisted at all times.
• At Mach numbers above 0.75 loss of total elevon effectiveness with
speed and acceleration severely restricted maneuverability and maxi-
mum attainable lift.
• At Mach numbers above 0.85 elevon effectiveness began to decline
rapidly in rolling maneuvers.
• At a Mach number of 0.88 the yawing and rolling coupled with
the longitudinal motions resulting in persistent oscillation about
three axes.
187
Probing the Sky
• At a Mach number of 0.90 a high-frequency short-period longitudi-

nal oscillation appeared at normal acceleration greater than 1g.
• At Mach numbers above 0.90 elevon effectiveness had virtually
disappeared, angles required for trim in level flight were high and
maneuverability was only slight.
• Also, at Mach numbers above 0.90 the lateral-directional oscillation
diverged to unsafe values. The tests were limited by the lack of con-
trol power to trim and maneuver and the divergent oscillation.48
Additional Research Activities

The NACA basic handling qualities program was completed with flight 45,
made on May 29, 1951. This included a climb to about 42,300 feet, a speed
run at about 38,000 feet, and windup turns at 30,000 feet over a speed range
from Mach 0.40 to 0.88. After the flight, the aircraft was grounded for the
repair of an oil leak and installation of new turbine blades. The work involved
the assembly of a new right engine. During its thrust-stand check, the new
engine suffered a failure of the accessory drive.49
In the wake of the problems with the right engine, plans for the aircraft
were changed. Project managers decided a modification of the wing shape
would be made. Like those on most aircraft, the X-4 wing had a rounded
leading edge and a thin, knife-edge trailing edge. The modification increased
the dive brake’s trailing-edge thickness to half that of the hinge-line thickness.
Engineers thought a blunt trailing edge on the speed brakes would improve
aircraft stability. Accordingly, they blocked the dive brakes open at an angle
of ±5°. The brakes could be opened to angles greater than ±5° but could no
longer be closed completely.50 The modifications to the X-4’s dive brakes
were completed in late July 1951. Strain gauges were also added to the upper
and lower segments of the left dive brakes to record hinge-moment data. The
research flights were delayed when a test run of the replacement right engine
indicated that the throttle mechanism required adjustment. This pushed the
start of research flights back to late August.51
Walter P. Jones made the first tests of the dive brake modifications on
NACA flight 46 on August 20, 1951. The tests involved an unaccelerated
stall at 35,000 feet and aileron rolls and turns at Mach 0.71, 0.83, and 0.86.
Jones then dove and made a turn and aileron roll at Mach 0.91 at 28,000 feet.
Jones reported a general improvement in the aircraft’s handling. In particu-
lar, the porpoising and lateral oscillations that had limited the X-4 at speeds
of Mach 0.88 to 0.90 were not apparent, with the thickened trailing edges,
to a Mach number of 0.91.52
188
Once more, engine problems appeared, delaying flight 47 until October

2, 1951. Crossfield made this flight, which continued the investigation of the
effects of the thickened dive brakes on stability and control. The improve-
ments were significant; the X-4’s longitudinal oscillations with the thin trail-
ing edges were undamped above Mach 0.88. In contrast, with the thickened
speed brakes at a speed of Mach 0.90, the longitudinal oscillations damped
out after about two cycles.
Crossfield executed windup turns at Mach numbers from 0.79 to 0.90 for
static-stability data and reported after the flight that slight porpoising and
instability occurred at about Mach 0.86. When he added a rudder impulse at
Mach 0.90, it caused an undamped lateral oscillation of about ±5½° of sideslip
and ±10° of angle of bank, with slight porpoising. A similar maneuver had
been made with the thinner trailing edges at about the same altitude and Mach
number, but subsequent comparison showed that the longitudinal oscillations
were less pronounced with the thicker trailing edges.53
With initial tests of the thickened dive brake complete, focus shifted
to obtaining lift-to-drag ratios for various dive brake settings during land-
ings. Crossfield piloted NACA flights 48 through 51 between October 5
and October 12, 1951. The dive brake deflections ranged from 0° to ±40°
and provided landing data on lift-to-drag ratios between 8 to 1 and 3.5
to 1. Jones made flight 52 on October 17, making constant speed runs to
determine lift-to-drag ratios at dive brake deflections of 0° to ±60° in the
landing configuration.
Joe Walker now joined the X-4 research effort, making NACA flights
53 and 54 on October 18 and 19. These were both pilot-familiarization
and research flights involving speed runs, rudder kicks, low-speed stalls at
10,000 feet, dive brake deflections, and the collection of landing pattern data.
Research with the X-4 ended for 1951 with NACA flight 55, made by Jones
on October 24. This flight provided additional stability and control flight data
on the thickened dive brakes.
After the flight, an inspection found that fuel leaks had reappeared.
Engineers initially estimated that these could be quickly fixed and flights could
be resumed on November 9. The leaks were more serious than first thought,
however, and the aircraft was grounded indefinitely.54
After considerable effort, NACA flight 56 was finally attempted on January
29, 1952, but was aborted before takeoff due to excessive fuel overflow. The
fuel regulator was replaced and set at a lower limiting pressure. This was not
the only maintenance problem X-4 personnel faced in the winter of 1951–52.
The sideslip transmitter had to be repaired, and the left engine was replaced due
to “torching” between the tailpipe and the shroud. This was traced to leaking
oil around the front bearing seal.55
189
Probing the Sky
NACA research pilot Joe Walker discusses his upcoming X-4 flight with a ground crewman.
Walker made only one flight in the aircraft, on October 18, 1951. The bulk of the NACA flights
were made by Crossfield and Walter P. Jones. (NASA)
Not until March 6, 1952, did Jones finally complete NACA flight 56.
This involved measurements of lift-over-drag characteristics at different drag
brake deflections, as well as unaccelerated stalls and constant speed runs at
various throttle settings. The engine problems reoccurred, this time on the
right engine. Flame instabilities, which required investigation, also occurred
during the flight. Unlike previous engine difficulties, these were soon fixed,
and NACA flight 57 was made on March 13. It was primarily an engine check
flight, but data were recorded on directional trim changes at Mach 0.79 and
during the landing.
Flight 58, on March 17, was intended to be a continuation of the lift-over-
drag studies, but an open circuit in the dive brake actuating solenoid prevented
the dive brakes’ use. Flight 59, during which lift-and-drag data were collected
with the dive brakes at ±50° deflections, was completed March 21. NACA
flight 60 was made on March 26 and entailed rudder-fixed aileron rolls, as well
as testing of stick and rudder impulses at true Mach numbers between 0.80 and
0.90. Flight 61 was made on March 27, with stick impulses recorded at true
Mach numbers of 0.50, 0.60, and 0.70. Landing data were also collected on
190
the flight. In all, Jones made seven consecutive X-4 research flights in March
1952. This brought the dive brake tests to a close.56
The X-4’s hectic March flight schedule was concluded with a pilot checkout
flight on March 27. Stanley P. Butchart made NACA flight 62. The dive brake
research was now completed, the X-4 was grounded, and the dive brake data
were under analysis.57
Landing data initially seemed a less consequential area of study but emerged
as a significant issue. Supersonic flight required an entirely new set of design
features. These included thin wings, low-aspect-ratio wings, and swept wings,
all with high wing loading. As a result, the aircraft had lower lift-to-drag ratios
and higher stall speeds than earlier designs. The resulting vertical velocities
made it difficult for a pilot to land safely and accurately.
The low lift-to-drag ratio and high stalling speed of high-performance air-
craft meant the excess speed ratio required at the start of the flare increased
considerably as lift-drag ratio decreased. The flare also had to begin at a rela-
tively high altitude. Additionally, past flight tests had shown that a vertical
velocity in excess of 25 feet per second at the start of the flare put too high a
demand on pilot skill and was regarded to be impractical. The X-4’s large dive
brakes enabled a landing maneuver at lift-drag ratios from 8 to 1 down to 3.5
to 1. A research memorandum discussed the results:
These landings were started at an altitude of approximately 3,000

feet with the engines maintaining zero thrust and with a con-
stant dive-brake angle during the landing maneuver. The patterns
become smaller as the lift-drag ratio decreases[,] which requires
an increase in acceleration during the approach turn from 1.1g
at a lift-drag ratio of 8 to about 1.5g at a lift-drag ratio of 3.5.
The higher acceleration results also from the fact that part of the
landing flare is made during the final approach turn at the lower
lift-drag ratios. This has prevented landings from being made at
dive-brake settings greater than 35° because the largest portion
of the flare is made during the turn at these settings and there is
insufficient elevon control to enable the maneuver to be accom-
plished at larger dive-brake settings. One factor noted by the
pilots was the short length of time, 50 seconds, at an approach
lift-drag ratio of 3.5 as compared with about 140 seconds at a lift-
drag ratio of 8 during which the pilot could correct and modify
his landing approach.
The poor longitudinal control at large dive-brake settings was
the pilots’ greatest complaint during these flights. They felt that,
if sufficient longitudinal control were available, landings could
191
Probing the Sky
be performed at still lower lift-drag ratios. Landings at the lowest

lift-drag ratios were not felt to require exceptional piloting skill
or a great deal of practice. However, it should be remembered
that for these landings the lift-drag ratio increased with decreas-
ing speed and although landings were started at a lift-drag ratio
of 3.5 the lowest lift-drag ratio at contact was about 6.2 even
neglecting ground effect. At high lift-drag ratios ground effect was
very noticeable to the pilots, whereas at the lower lift-drag ratios
ground effect was not nearly so pronounced.58
The memorandum concluded, “Landings of the X-4 airplane to determine

the effect of lift-drag ratio showed that the largest portion of the landing flare
was made at altitudes above 50 feet at low lift-drag ratio and that, although the
vertical velocities during the approach varied from 30 to 90 feet per second,
the vertical velocities at contact were less than 5.5 feet per second.”59
The X-4’s Twilight

In the spring of 1952, the X-4 was again grounded for a prolonged period.
Several modifications had to be completed before flights resumed. One of these
was the fabrication of jigs to support the aircraft during planned moment-of-
inertia measurements. A more significant activity was the modification of the
elevons’ trailing edges. As with the dive brakes, their thickness was increased to
50 percent of hinge-line thickness. Again, this was done to determine whether
blunt trailing edges would improve aircraft handling and was completed by
the end of April. However, a May 1 ground engine run was unsuccessful. The
left engine overheated, requiring replacement.60
A successful engine run was made on May 16, and NACA flight 63 was
made May 19 with Jones as pilot. Of the planned test maneuvers, the stick
impulses and one-half-deflection aileron rolls, performed at 30,000 feet and at
Mach numbers of 0.55 and 0.60, were successfully completed. At this point,
the right engine began to overheat. Jones aborted the rest of the tests and
landed safely. This flight also marked the end of Jones’s research flying with the
NACA. He left the High-Speed Flight Research Station in July 1952 to work
as a Northrop test pilot. Tragically, he died the following October in the crash
of a Northrop YF-89D Scorpion.
The malfunctioning engine was replaced, but a ground run on May 26 was
unsuccessful due to a stuck fuel valve. The problem was fixed, and a second
ground test was made on May 29. Excessive vibrations caused this test to fail
as well. Their cause was unknown, and the X-4 remained grounded until a
192
replacement engine could be installed and successfully tested. This proved

more difficult than expected. All of the spare J30 engines proved faulty when
ground tested, due either to vibration or to excessive bearing heating. To pro-
duce a functional engine, maintenance personnel had to assemble parts from
spare jets.61 The effort proved frustrating, as a progress report noted: “In each
instance the reassembled engine has experienced excessive vibration. Assembly
of another engine is underway.”62 Finally, on July 25, an engine was tested that
seemed to be fault-free.
After being grounded for more than 2 months, NACA flight 64 was made
on August 6 with Crossfield as pilot. The first test of the thickened elevons was
made without incident. Crossfield’s opinion was that elevon control had been
much improved at higher Mach numbers by the modification. Postflight exami-
nation showed the windup turn data were unusable, however, due to instru-
mentation failure. The maneuvers would have to be repeated on a later flight.
More bad news followed: NACA flight 65 was delayed following a preflight
engine run. An investigation showed that an object had passed through the right
engine, presumably during landing on flight 64. The aircraft was grounded
until the engine was rebalanced and a successful ground run completed. While
the aircraft was being serviced, the thickening material was removed from the
dive brakes, restoring their original configuration. Subsequent research flights
would be made with the thickened elevon configuration and the original dive
brake configuration.63
After the series of problems, the X-4 made NACA flights 65 through 68
during September 1952. Data from the initial flight were lost due to instru-
mentation malfunction. The three other flights were successful, with windup
turns and aileron rolls at Mach 0.65 to 0.94 and stick and rudder impulses from
Mach 0.70 and 0.90. Postflight analysis indicated that the increased elevon
thickness improved aileron effectiveness over the entire Mach number range.64
Research missions with the X-4s had been flown without serious mishaps,
despite each aircraft’s having exhibited poor stability at high transonic speeds
and unreliable engines. It was therefore ironic that the most serious damage
either aircraft suffered happened on the ground, at the hands of a person
assigned to protect it. Over the weekend of October 4–5, an Air Force guard
on duty at the NACA hanger at South Base accidentally drove an airplane tug
into the second X-4’s right wing. The impact damaged the wing’s underside.
The extent of the damage was not immediately evident; an inspection was
made to determine whether the wing could be repaired and the X-4 returned to
flight, or if the damage was so severe that the aircraft would have to be retired.65
Inspection determined that the damage to the aircraft was limited to the
wing’s skin. Repair was complete by mid-December 1952. The X-4 remained
off flight status for several weeks while the moments of inertia about the
193
Probing the Sky
three axes at various fuel levels and attitudes were determined. This work
continued through the end of February 1953, and the aircraft then resumed
research flights.66
The first of these was made on March 27 with Crossfield as pilot. NACA
flight 69 included a speed run to Mach 0.91 and successful windup turns
at Mach 0.78 and 0.88. At that point, the fuel gauge malfunctioned and
Crossfield had to cut the flight short.67
NACA flights 70 and 71 followed at the end of April. The first flight
involved tower passes at a low level to check airspeed calibration. The second
flight undertook speed runs and windup turns between Mach 0.77 and
Mach 0.90, both to check the airspeed calibration and to collect additional
data for the longitudinal control study with the thickened elevons. Flight
70 marked the end of this investigation. After it was concluded, the elevons
were restored to their original configuration. Work was completed by late
May, when Crossfield made NACA flight 72. This flight provided dynamic-
stability data without the thickened elevons, using rudder pulses at Mach
0.50 and 0.88 speeds.68
X-4 research activity continued with NACA flights 73 and 74 on July 1
and 3. Both flights were made by Butchart to acquire additional data on the
aircraft’s dynamic stability without the thickened elevons. This was accom-
plished using elevon and aileron pulses at Mach numbers from 0.50 to 0.88.69
Although the X-4 project was entering its final phase, activity increased
substantially. During August 1953, seven flights were made beginning with
flight 75 on August 4, a checkout flight for Ames Aeronautical Laboratory
research pilot George Cooper. Flight 76 on August 11 was a checkout flight
for John B. McKay, a new High-Speed Flight Research Station research pilot.
McKay also made NACA flights 77 and 78, on August 13 and 19, collecting
dynamic-stability data from aileron pulses. However, data from several runs
were lost due to instrumentation failures. Once again, this required another
research flight to be added to the schedule.
The fast pace continued to the end of the X-4 project. Two days after his
previous mission, McKay again piloted the X-4 on NACA flight 79. Crossfield
went aloft on August 24 for flight 80. The X-4 was checked out over the next
2 days and flown by McKay on August 26 for the 81st NACA mission. These
flights were for dynamic-stability data. The 4 years of X-4 flights by Northrop,
the Air Force, and the NACA then came to a close.
The final research mission, NACA flight 82, was made on September 29,
1953. McKay’s flight plan involved aileron pulses across the Mach number
range and low-speed turns to 2 g’s at different dive brake settings. This brought
the X-4 project to an end. At the time the mission was flown, there were plans
to use the aircraft to check out a number of Air Force and NACA pilots, but
194
plans were dropped, and the second X-4 joined the first in storage. Eventually,
the first X-4 was put on display at the U.S. Air Force Academy in Colorado
Springs, CO. It subsequently became part of the Air Force Flight Test Museum
collection at Edwards Air Force Base, where it awaits restoration. The second
X-4 was sent in 1955 to Maxwell Air Force Base, Montgomery, AL. It remained
there until March 17, 1972, when it was transferred to the U.S. Air Force
Museum at Wright-Patterson Air Force Base, OH, where it is on display, having
been magnificently restored.70
An Assessment
The X-4 Bantam was a highly productive research aircraft, though one that,
at the time, was of impracticable configuration for commercial or military
use. More research flights were made with the X-4 than with the X-3, and the
X-4 provided data in a wide range of disciplines. Yet the X-3 influenced the
design and development of the F-104, one of the most widely used fighters of
the mid–20th century. In contrast, the X-4’s semi-tailless swept-wing design
did not see significant use until much later, following the advent of electronic
fly-by-wire flight control technology. The only U.S. aircraft to feature a similar
configuration in the 1950s was the Navy’s severely underpowered and, indeed,
highly dangerous F7U Cutlass, which blended prolonged development with
a brief operational service life, entering service in 1954 and retiring almost
exactly 5 years later. Of the 320 F7Us built, some 80—fully 25 percent—were
lost in accidents.71
Crossfield summed up the X-4 project during a 1998 interview:
Oh, the X-4 was a whole buildup of stability and handling, stabil-
ity and control, and handling qualities of a tailless airplane. With
the X-4, we pretty much found out why we may not have solved
the problem…. The X-4 was a very productive program. We did
a lot of things with the X-4 that weren’t only involved in it’s [sic]
being a tailless airplane. It was a good test bed for other purposes,
as we did thickened trailing edges on the X-4. We did variable
Lift over Drag (L over D) landings from probably an L over D of
two to an L over D of nine with the airplane, which gave us good
insight into pilots preferences and that sort of thing, which were
useful in coming up with new airplanes like the X-15 and some
of the other low L over D airplanes [like] the lifting bodies. But
the X-4 was a very good research airplane, for purposes of looking
into a tailless configuration, and for other reasons.72
195
Probing the Sky
Though the configurations of the D-558-2, X-3, X-5, and XF-92A became the basis for numer-
ous U.S. aircraft, only the F7U Cutlass featured the X-4’s semi-tailless swept-wing configuration.
An early prototype XF7U-1 is shown here on the Langley ramp in 1948. The aircraft entered U.S.
service after a prolonged development, but it suffered a high loss rate and was soon retired from
service. (NASA)
But for all the flights made with it and all the data returned, the X-4 was
conceptually a flawed aircraft. Its instability was described by research pilots
as being akin to driving on a washboard road. The X-4 displayed increasingly
severe combined yaw, pitch, and roll motions, as well as inadequate damping
as Mach number increased. In an attempt to correct the instability, NACA
engineers decided to modify the dive brakes and later the elevons, giving
them a blunt trailing edge. X-4 research flights indicated that this produced a
25-percent increase in roll rate and improved longitudinal control effectiveness.
Above Mach 0.9, however, instability reoccurred. The severity of the porpois-
ing increased rapidly; at Mach 0.94, vertical accelerations reached ±1½ g’s.
The basic cause of the instability could not be solved by Band-Aid fixes such
as thickening the trailing edges of dive brakes and elevons.73
Until the advent of the fly-by-wire Northrop B-2A Spirit stealth bomber,
the various direct or hydromechanically controlled semi-tailless and flying-
wing designs, such as the Horten Ho 229; the Northrop N-9M, XP-79B, and
YB-49; and the British de Havilland DH 108 and Armstrong Whitworth AW
196
52, all suffered from serious stability and control deficiencies that limited their
practical use and even endowed them with vicious, life-endangering quirks
(indeed, between them, the six types listed here killed a dozen highly skilled
airmen). For all their perceived advantages from an aerodynamic viewpoint,
all were thus impractical from the start. The semi-tailless X-4 had been built to
determine whether eliminating the horizontal stabilizer would prevent interac-
tions between the shock waves formed by the wings and horizontal stabilizers.
In fact, the lack of horizontal stabilizers made the instability worse compared
with that of a conventional swept-wing/horizontal stabilizer design. The DH
108 demonstrated the risks of what deficiencies could do if encountered at
high speed and low level. The X-4 demonstrated the problem but, flown at
higher altitudes and lower dynamic pressures, it avoided the disastrous experi-
ence of its British cousin. Nevertheless, it clearly had such deficiencies that it
did not warrant further development. Northrop played around with the swept
semi-tailless configuration, even employing it on a long-range, jet-propelled,
nuclear-tipped cruise missile (the disappointing Northrop SM-62 Snark), but,
overall, swept-wing semi-tailless aircraft remained impractical for nearly three
decades, until the advent of computer-controlled flight.
Ironically, the original X-4 rudder control system gave a hint of what would
be needed to correct the aircraft’s control problems. Rather than using a cable
and bell-crank system, the rudder was electrically operated by a four-speed
actuator. This was a proto-fly-by-wire system, which served as a yaw-damping
stability-augmentation system. The X-4 system, as noted earlier, was limited to
a maximum speed of 25° per second. This system was too slow to the point of
being dangerous, and it was replaced with a conventional mechanical system.
As long as an aircraft had to be inherently stable, the swept-wing semi-tailless
configuration remained impractical. The technology for providing a solution
did not exist in the early 1950s.74
It was not until the 1970s that practical digital fly-by-wire systems were
developed. These systems supplied an artificial inherent stability by providing
control inputs faster than human pilots could. The faster inputs eliminated
washboard and other unstable motions, allowing the pilot to maneuver the
aircraft rather than struggle to control it. The first aircraft to use a digital fly-
by-wire control system was a modified F-8 flown at the Dryden Flight Research
Center (now called the Armstrong Flight Research Center). In an important
innovation, the airplane’s mechanical control system was removed. Should
the digital system fail, an analog backup computer would take over. NASA
engineers realized that fly-by-wire would never be accepted as proven as long
as a mechanical backup was still being used.
Once the aircraft was totally dependent on a fly-by-wire system, inherent
stability was no longer a design limitation. Among the first applications of this
197
Probing the Sky
A photo illustrating the rapid changes in aviation technology at Edwards in the mid-1950s. In
the background is a pair of B-36 bombers, originally designed for the bombing of Nazi Germany
from U.S. bases. The XB-52 prototype, with the original tandem cockpit design, is in the
foreground. Its descendants are still in service a half century later. Beside the XB-52 is the tiny,
futuristic X-4. Visible are its control surfaces, which were designed to overcome the absence of
horizontal stabilizers. (USAF)
was the Lockheed XST Have Blue stealth test aircraft, first flown in December
1977. This was a swept-wing, semi-tailless aircraft with two inwardly canted tail
fins mounted at mid-span of each wing. The design was unstable in the pitch
axis and relied on a modified F-16 fly-by-wire system for stability.75 Even so,
for various reasons, its successor, the Lockheed F-117A stealth fighter, had a
different vertical fin configuration, effectively giving it a V-tail (like the Beech
Bonanza general aviation airplane).
Over the decades to follow, fly-by-wire control systems were fitted to aircraft
ranging from business jets, airliners, fighters, and research aircraft to bombers
and heavy transports. But for the X-4, such a solution was not yet available.
A computer in the early 1950s took up a large room, was slow and unreli-
able, and required a large maintenance staff to keep it operating. Today, the
basic X-4 configuration can be discerned in the lines of a variety of remotely
piloted, computer-controlled aircraft. Impractical in the 1950s, the little X-4
was ahead of its time.
198
A famous photo showing the family of X-planes that were being flown at the High-Speed Flight
Research Station in the early and mid-1950s. In the center is the X-3, with the X-1A, D-558-1
Skystreak, XF-92A, X-5, D-558-2 Skyrocket, and X-4 arrayed around it. (NASA)
199
Probing the Sky
Endnotes
1. Opening quote from A. Scott Crossfield interview transcript, DFRC
Archives.
2. Richard P. Hallion, “Before B-2: The History of the Flying Wing,” 2
parts, Air Power History 41, nos. 3 and 4 (fall–winter 1994): 4–13,
40–51.
3. Richard P. Hallion, “Rocket Dreams,” World War II 23, no. 4
(October–November 2008): 57–58. For reminiscences by mem-
bers of the Komet team, see Wolfgang Späte, Top Secret Bird: The
Luftwaffe’s Me 163 Comet [sic] (Missoula, MT: Pictorial Histories
Publishing Company, 1989); and Mano Ziegler, Rocket Fighter
(Garden City, NY: Doubleday and Co., Inc., 1961).
4. The DH 106, which evolved into the Comet I jetliner, was initially
conceived as a swept-wing semi-tailless design. However, it ulti-
mately emerged as a conventional wing-fuselage-tail configuration,
with four jet engines buried in its wing roots.
5. Quote from Derry’s pilot notes, reprinted in Charles Burnet,
Three Centuries to Concorde (London: Mechanical Engineering
Publications Limited, 1979), p. 112.
6. See also Captain Eric Brown, RN FAA, “An Ill-Fated ‘Swallow’…
but a Harbinger of Summer,” Air Enthusiast 10 (July 1979): 1–7;
and Brian Rivas, A Very British Sound Barrier: DH 108—A Story
of Courage, Triumph, and Tragedy (Newark, U.K.: Red Kite/Wing
Leader Publishers, 2012).
7. For the detailed history of the X-4’s design, development, and flight
testing, see Richard P. Hallion, “X-4: The Bantam Explorer,” Air
Enthusiast Quarterly 3 (June 1977): 18–25.
8. Total fuel load was 240 gallons, but 10 gallons were “trapped”
within the vehicle’s plumbing and pump system and unusable.
9. Technical details from Hubert M. Drake, “Stability and Control
Data Obtained from First Flight of X-4 Airplane,” NACA RM
L9A31 (February 7, 1949), Table 1, “Physical Characteristics of X-4
Airplane,” p. 6.
10. Jay Miller, The X-Planes X-1 to X-45 (Hinckley, England: Midland
Publishing, 2001), pp. 79–81; see also Hallion, “X-4: The Bantam
Explorer,” pp. 20–21.
11. Drake, “Stability and Control Data Obtained from First Flight of
X-4 Airplane,” p. 3.
12. Discussion with NASA research pilot Mark Pestana, Dryden Flight
Research Center, July 19, 2010.
200
13. Drake, “Stability and Control Data Obtained from First Flight of
X-4 Airplane,” p. 4.
14. Memorandum for Chief of Research, “Progress Report on
Acceptance Tests of X-4 Airplanes from January 1 to January 14,
1949.” Readers are advised that this and all other flight-test docu-
ments and reports mentioned in this chapter are from the DFRC
historical archives unless otherwise noted.
15. Hallion, “X-4: The Bantam Explorer,” p. 20.
16. Walter C. Williams, “Results Obtained from Second Flight of X-4
Airplane (A.F. No. 46-676),” NACA RM L9F21 (July 18, 1949), p. 3.
17. Ibid., pp. 3–4.
18. Memorandum for Chief of Research, “X-4, January 14, 1949 to
June 10, 1949”; Memorandum for Chief of Research, “X-4, June
20, 1949.”
19. Memorandum for Chief of Research, “X-4, June 17 through July 1,
1949.”
20. Memorandum for Chief of Research, “X-4, July 16 through July 29,
1949”; Memorandum for Chief of Research, “X-4 Airplanes, July 30
through August 12, 1949.”
21. Memorandum for Chief of Research, “X-4 Airplanes, July 31
through August 12, 1949.”
22. Memorandum for Chief of Research, “X-4, August 27 to September
14, 1949.”
23. Memorandum for Chief of Research, “X-4 Airplanes, October 8 to
October 21, 1949.”
24. Ibid.
25. Ibid.
26. Memorandum for Chief of Research, “X-4, November 19 to
December 2, 1949.”
27. Memorandum for Chief of Research, “X-4, December 3 to
28. Memorandum for Chief of Research, “X-4, December 31, 1949
to January 1, 1950”; Memorandum for Chief of Research, “X-4,
January 14 to January 27, 1950.”
29. Memorandum for Chief of Research, “X-4, January 27 to February
24, 1950.”
30. Ibid.
31. Memorandum for Research Airplanes Project Leader, “Acceptance of
the X-4 Airplane.”
32. Ibid.; Peebles, The Spoken Word, p. 28.
201
Probing the Sky
33. Memorandum for Research Airplanes Project Leader, “Acceptance of

the X-4 Airplane.”
34. Memorandum for Chief of Research, “X-4, February 24, 1950 to
March 24, 1950”; Memorandum for Chief of Research, “X-4, May
5 to May 19, 1950”; Memorandum for Chief of Research, “Test
Program for X-4 Aircraft Evaluation, May 3, 1950.”
35. Memorandum for Chief of Research, “X-4, July 15 to July 28,
1950.”
36. Memorandum for Chief of Research, “X-4, August 26 to September
8, 1950.”
37. Memorandum for Chief of Research, “X-4, September 22 to
October 6, 1950.”
38. Melvin Sadoff and Thomas R. Sisk, “Summary Report of Results
Obtained During Demonstration Tests of the Northrop X-4
Airplanes,” NACA RM A50I01 (December 13, 1950), pp. 11–12.
September 22, 1950.”
40. Memorandum for Chief of Research, “X-4, October 6 to October
20, 1950.”
November 17, 1950”; Memorandum for Chief of Research, “X-4,
November 17 to December 1, 1950.”
42. Quote from Crossfield and Blair, Always Another Dawn, pp. 41–44.
44. Memorandum for Chief of Research, “X-4, December 15, 1950 to
45. W.C. Williams and A.S. Crossfield, “Handling Qualities of High-
Speed Airplanes,” NACA RM L52A08 (January 28, 1952), p. 5.
46. Indeed, tailless aircraft in general, and the X-4 in particular, inher-
ently had very low damping-in-pitch compared to conventional
“tailed” aircraft. The conventional F-86 had a damping-in-pitch
factor of +10. A 60° delta like the XF-92A or F-102 had a damping-
in-pitch of +3. The X-4 had a damping-in-pitch factor less than +1,
and as one study concluded, it “dropped to almost zero at Mach
numbers slightly below 1.0.” See Joseph Adams Shortal, A New
Dimension—Wallops Island Flight Test Range: The First Fifteen Years,
NASA RP 1028 (Washington, DC: NASA, 1978), p. 278.
47. Williams and Crossfield, “Handling Qualities of High-Speed
Airplanes,” p. 7.
202
48. Melvin Sadoff and A. Scott Crossfield, “A Flight Evaluation of the

Stability and Control of the X-4 Swept-Wing Semitailless Airplane,”
NACA RM H54G16 (August 30, 1954), pp. 13–14.
49. Memorandum for Chief of Research, “X-4, May 18 to June 1,
1951”; Memorandum for Chief of Research, “X-4, June 16 to June
29, 1951.”
50. Memorandum for Chief of Research, “X-4, June 30 to July 13,
1951.” The concept of a blunt trailing wing edge was also tested at
supersonic speeds using rocket-launched models launched from the
Pilotless Aircraft Research Division over the Atlantic. The test-wing
configuration had an NACA 0010-64 airfoil (identical to that of
the X-4) with flat-side ailerons with a trailing-edge thickness equal
to one-half that of the hinge line. NACA researchers H. Kurt Strass
and Edison M. Fields analyzed the data from the rocket flights and
found that the increased trailing-edge thickness eliminated con-
trol reversal and prevented a large and abrupt decrease in aileron
effectiveness in the transonic speed range. The downside was that
the aileron modifications increased wing drag at subsonic speeds.
Above Mach 1, however, the drag increase was small. See H. Kurt
Strass and Edison M. Fields, “Flight Investigation of the Effect
of Thickening the Aileron Trailing Edge on Control Effectiveness
for Sweptback Tapered Wings Having Sharp- and Round-Nose
Sections,” NACA RM L9L19 (May 2, 1950), pp. 1–2, 5.
51. Memorandum for Chief of Research, “X-4, July 14 to July 27,
1951”; Memorandum for Chief of Research, “X-4, July 28 to
August 10, 1951.”
52. Memorandum for Chief of Research, “X-4, August 11 to August 25,
1951.”
October 5, 1951.” Initially, flight 47 was made for further investiga-
tion of the effect of the thickened dive brakes. Subsequently, this
was changed to obtaining data on the X-4’s landing characteristics.
Finally, the plan was again changed back to collecting data on the
thickened dive brakes. The tests of the landing configurations fol-
lowed the completion of the dive brake tests.
54. Memorandum for Chief of Research, “X-4, October 6 to October
19, 1951”; Memorandum for Chief of Research, “X-4, October 20
to November 2, 1951”; Memorandum for Chief of Research, “X-4,
November 3 to November 16, 1951.”
203
Probing the Sky
55. Memorandum for Chief of Research, “X-4, January 26 to February

8, 1952”; Memorandum for Chief of Research, “X-4, February 9 to
February 22, 1952.”
56. Memorandum for Chief of Research, “X-4, March 8 to March 21,
1952”; Memorandum for Chief of Research, “X-4, March 22 to
April 4, 1952.”
57. Memorandum for Chief of Research, “X-4, March 22 to
April 4, 1952.”
58. Wendell H. Stillwell, “Results of Measurements Made During the
Approach and Landing of Seven High-Speed Research Airplanes,”
NACA RM H54K24 (February 4, 1955), pp. 1–2, 9–10.
59. Ibid.
60. Memorandum for Chief of Research, “X-4, April 5 to April 18,
1952”; Memorandum for Chief of Research, “X-4, April 19 to May
2, 1952.”
61. Memorandum for Chief of Research, “X-4, May 17 to May 30,
1952”; Memorandum for Chief of Research, “X-4, May 31 to June
13, 1952.”
62. Memorandum for Chief of Research, “X-4, June 28 to July 11,
1952”; Memorandum for Chief of Research, “X-4, July 12 to July
25, 1952.”
63. Memorandum for Chief of Research, “X-4, August 9 to August 22,
1952”; Memorandum for Chief of Research, “X-4, September 5 to
October 1, 1952.”
64. Memorandum for Chief of Research, “X-4, September 5 to October
1, 1952.”
65. Memorandum for Chief of Research, “X-4, October 1 to November
1, 1952.”
December 1, 1952”; Memorandum for Chief of Research, “X-4,
January 1, 1953 to January 31, 1953”; Memorandum for Chief of
Research, “X-4, February 1 to February 28, 1953.”
67. Memorandum for Chief of Research, “X-4, March 1 to March 31,
1953.”
68. Memorandum for Chief of Research, “X-4, April 1 to April 30,
1953”; Memorandum for Chief of Research, “X-4, May 1 to May
31, 1953.”
69. Memorandum for Chief of Research, “X-4, July 1 to July 31, 1953.”
The monthly report included an item regarding analysis of the X-4
flight data. It read, “Frequency response analysis of the last two data
flights is awaiting access to automatic computing equipment here at
204
Edwards Air Force Base.” Before this, data reduction had been done
by women who were referred to as “computers.” It was at this point
that the term’s meaning was changed from a human to a machine.
70. Memorandum for Chief of Research, “X-4, August 1 to August
31, 1953”; Memorandum for Chief of Research, “X-4, September
1 to September 30, 1953.” The August 1 to August 31, 1953,
report indicates that NACA flights 79, 80, and 81 were pilot-
familiarization flights. However, Crossfield had made numerous X-4
flights, and McKay made the majority of the final flights. It appears
that the familiarization flights entry was an error and that these were
research flights.
71. “Vought F7U Cutlass,” http://aviationtrivia.info/Vought-F7U-Cutlass.
php, and “F7U Cutlass,” http://www.absoluteastronomy.com/topics/
F7U_Cutlass, accessed October 5, 2010.
72. Peebles, The Spoken Word, p. 79.
73. Hallion and Gorn, On the Frontier: Experimental Flight at NASA
Dryden, p. 50.
74. Sadoff and Crossfield, “A Flight Evaluation of the Stability and
Control of the X-4 Swept-Wing Semitailless Airplane,” p. 4;
Malcolm J. Abzug and E. Eugene Larrabee, Airplane Stability
and Control: A History of the Technology that Made Aviation
Possible (Cambridge, U.K.: Cambridge University Press, 1997),
pp. 175, 311.
75. David C. Aronstein and Albert C. Piccirillo, Have Blue and the
F-117: Evolution of the “Stealth Fighter” (Reston, VA: American
Institute of Aeronautics and Astronautics, 1997).
205
The initial wind tunnel tests of variable sweep wings were made using an X-1 fitted with mov-
able wings. The design proved impractical, but the results were sufficient to encourage the
approval of the X-5. (NASA)
206
CHAPTER 6
Transformative Pioneer:
The Bell X-5
It was like having a whole stable of swept-wing airplanes in one.

The development of swept wings was a major step in making routine supersonic
flight a reality.2 But while swept wings reduced the drag that occurs at transonic
and supersonic speeds, they also introduced new problems, principally at low
speeds and at high angles of attack, particularly during takeoff and landing.
The wing design had to balance the conflicting demands of the two ends of the
performance envelope. Engineers tried to capture the beneficial characteristics
of both low- and high-speed wing shapes. This, though, could potentially
make the aircraft difficult to fly, increasing the pilot’s workload and the risks
that aviators faced.
Additionally, jet air combat required multiple mission profiles during a
single flight. A long-range strike mission might involve a long-range cruise,
requiring maximum fuel efficiency, followed by a brief supersonic dash to bomb
a target and escape, and, finally, a long cruise back to base. The Navy also faced
additional problems. The high landing speeds of swept-wing aircraft would
make already-dangerous carrier landings on narrow straight-deck carriers even
more difficult. Tactically, swept wings posed a problem for fighters on combat
air patrol over a naval task force. A capability for low-speed loitering was needed
to maximize coverage. If incoming bombers were detected, fighters would have
to accelerate to high speeds to intercept the bombers before the bombers got
close enough to attack the carriers.
The shape of aircraft carriers complicated matters further; the vessels were
still being designed with straight decks, as they had been from their origins.
Before the advent of powerful steam catapults, launching these fighters was
difficult. Deck activities were also more difficult since the new, larger aircraft
occupied more of the available deck space, which was limited. Of course, the
new jet fighters approached and landed at faster speeds than had their propeller-
driven antecedents, and this complicated matters on straight decks even more,
207
Probing the Sky
to say nothing of circumstances created for pilots making a landing approach in

the rain, at night, to an airport rolling and pitching in the middle of the ocean.
A potential solution was changing the sweep angle of the wings during
flight. For takeoff, landing, and cruise flight, the wings’ sweep angle could
be nearly straight. As an aircraft’s speed increased, the wings could be swept
back to reduce drag. The aircraft’s aerodynamics could be adjusted to meet the
demands of speed and altitude at any time during a flight. Initial research on
this innovation had begun previously, as the X-1 was being prepared for flight.
After the Second World War, the NACA moved rapidly to study the appli-
cation of variable-sweep wings to aircraft. Early NACA studies examined both
symmetrical wing sweeping (as with the X-5 and subsequent aircraft such
as the F-111) and asymmetrical “oblique” wing sweeping (as later incorpo-
rated on the Ames-Dryden AD-1 research aircraft).3 In early 1946, Charles
J. Donlan and William C. Sleeman tested an X-1 model fitted with variable-
sweep wings in Langley’s 7-by-10-foot High-Speed Wind Tunnel. The wings
could be adjusted to sweep angles of 0°, 15°, 30°, and 45°. The design assumed
that the wings moved on a fixed pivot point on the fuselage centerline. Results
were mixed. Tests showed that variable-sweep wings reduced the inherent
low-speed handling problems experienced with high-speed aircraft. Engineers
also saw limited success in controlling longitudinal stability by changing the
sweep angles. A problem was identified regarding the pivot design, however: A
fixed wing-pivot point was not workable. This was because both the airplane’s
center of gravity and center of pressure shifted as the wings changed position.
The net result was too much stability, resulting in excessive trim drag and
limited maneuverability.4
The wing-pivot system would have to be capable of changing the wing-
sweep angle and translating it forward and aft to compensate for the resulting
changes in center of gravity and center of pressure caused by the wing-sweep
movement. The amount of translation was also dependent on the mass distri-
bution of the airplane and the location of the wing pivot point. The weight of
the wings, the locations of the wing and fuselage fuel tanks, and the procedures
for emptying the fuel tanks in flight also had to be factored into any assess-
ments of the aircraft’s stability. This was a major issue in the view of researchers.
In their conclusions, they wrote, “It seems unlikely that satisfactory stability
for all flight conditions can be achieved with a variable-sweep wing without
recourse to relative translation between the wing and the center of gravity of
the airplane.”5
208
Transformative Pioneer: The Bell X-5
Nose view of the X-5 with the wings at the 20° sweep angle. Unlike those of later variable-sweep
aircraft, the X-5’s mechanism moved the wings forward as the sweep angle increased. On the
F-111, B-1, and F-14, the pivot point was in a stub wing outboard of the fuselage. This eliminated
the need for dual motions. (NASA)
The Evolution of the X-5

Just as the X-4 was influenced by the Me 163, Bell’s X-5 design was influenced
by an unsuccessful aircraft design that originated in Nazi Germany. This was
the P.1101, designed by Waldemar Voigt and built by Messerschmitt’s advanced
projects development group outside Oberammergau, Bavaria, near the end of
the war. Unlike the Me 163, however, the P.1101 never flew. Rather, having
been rejected by the Luftwaffe as unsuitable for further development, it had
been retained by Messerschmitt as a possible test bed for swept wings. The
single prototype, nearing completion, was captured in April 1945 when the
Oberammergau development center was overrun by U.S. Army forces. Shortly
afterwards, a U.S. technical intelligence team arrived, including Robert Woods
of Bell Aircraft. Woods arranged to have the unfinished aircraft shipped to
the United States for analysis and eventually turned over to Bell Aircraft in
August 1948.
209
Probing the Sky
As originally conceived, the wing-sweep angle of the P.1101 could be

adjusted on the ground to three fixed settings of 35°, 40°, and 45°, and Bell
engineers originally suggested to the Army Air Forces that the P.1101 be used
in similar fashion, recommending that it be repaired and fitted with a U.S.
turbojet engine and then test-flown. A potential next step would have been
to modify the P.1101 wings to allow the sweep range to increase from 20° to
50°, again at three fixed positions. But Bell’s P.1101 advocates also suggested
redesigning the P.1101 so that its sweep angle could be changed in flight. This
capability would be useful in an operational fighter, as the wings could be fully
extended for takeoff and landing, then swept back to achieve high-speed flight.
Another Bell proposal was to use the P.1101 as an engine test bed. The jet
engine was attached beneath the aircraft, at its center of gravity. This made engine
changes relatively easy. Among the engines proposed for tests on the P.1101 were
the J34, J35, J46, and J47. Another possibility was a 36-inch-diameter ramjet.
Finally, a 4,000-pound liquid-fuel rocket also could be tested, by placing a fairing
over the nose inlet and fitting the aircraft with new fuel tanks.
In the end, use of the P.1101 itself as a research aircraft proved impractical.
The aircraft was too small, major technical problems existed with the propos-
als, and the aircraft had been damaged when it was dropped during delivery
to Bell Aircraft. Rather, Woods and Bell engineers concluded that designing
a new aircraft based upon its basic configuration would be preferable. The
proposed model would be similar in design configuration to the P.1101 but
larger and more advanced. Bell also hoped that the resulting research plane
could be modified into a lightweight production jet fighter. (Eventually, the
X-5 would be proposed as a NATO export fighter but rejected on grounds
remarkably similar to those given by the Luftwaffe in its rejection of the earlier
P.1101.) Given the collapse in aircraft orders following the end of World War
II, however, such approval was unlikely.6
Bell’s Robert Wood and his design engineers began work on the new aircraft.
A major problem was the development of a practical wing-sweep mechanism.
The sweep angle could be varied from 20.25° to 58.7°. Limit switches prevented
the wings from exceeding these angles, which would cause interference between
the wing root and fuselage. As the wings swept back and forward, the X-5’s
center of gravity and center of pressure shifted. These characteristics had to be
kept in close proximity to prevent stability problems.
The solution was complex. As the wings’ sweep angle increased (reducing
the wingspan and aspect ratio), the X-5’s center of gravity moved aft. To coun-
ter this, the complete wing-root assembly moved forward along a track inside
the fuselage. When the sweep angle was reduced (increasing the wingspan
and aspect ratio), the wing assembly moved backward as the wingtips moved
forward. The net result was to keep both the center of gravity and center of
210
Nose view of the X-5 with the wings swept back to 60°. The aircraft was a low-speed demon-
strator for variable-swept wings, as it lacked the supersonic-speed capacity necessary to take
full advantage of the maximum sweep angle. (NASA)
pressure within limits, no matter the sweep angle. As the wing’s sweep angle
changed, so did the wing characteristics. At a 20° wing sweep, the X-5 had a
wing area of 184.3 square feet and an aspect ratio of 6.09. With the wings swept
back to the maximum 58.7° angle, these values were reduced to a 166.9-square-
foot area and 2.16 aspect ratio.7
Each wing also required a special fairing “glove” where the leading edge of
the wing met the fuselage. This was designed to be movable to ensure that the
wing had a smooth airfoil whatever the sweep angle. The wings’ leading edges
also had slats, which could be extended to increase lift and reduce stall speed.
One unusual feature was the speed brakes. Normally, these are located on the
aft fuselage, as on the D-558-1. The X-5 had its speed brakes on the nose, just
behind the inlet. This proved to be a very poor location, however, and thus the
brakes proved generally ineffective.
As on the P.1101, the X-5’s engine, an Allison J35-A-17A axial flow jet
engine, was slung under the fuselage. The engine produced 4,900 pounds of
thrust. Unlike the exotic flush inlets on the D-558-2 or the wing inlets of the
211
Probing the Sky
X-4, the X-5 had a simple, straight nose inlet that ran directly to the front of the
engine. “This design,” a Bell Aircraft press release noted, “holds air-duct losses
to a minimum, scooping greater quantities of air at high altitudes where the
decreased oxygen content of the atmosphere lowers jet engine performance.”8
Like the X-3 and X-4, the X-5 had an emergency ejection seat. The long main
landing gear struts retracted into the fuselage above the engine, and the nose
gear retracted into the nose, under the air inlet tunnel.
Unlike the other early X-planes, the X-5 was not particularly sleek, having
a configuration more akin to then-contemporary transonic jet fighters than
exotic supersonic research airplanes. Bell described it as having a “flying guppy”
configuration, with an underslung engine, fat forward fuselage, and boom tail.
The X-5 was 33.6 feet long and stood 12.2 feet tall from the ground to the tip
of the fin. Its wing employed an NACA 64(10)A011 airfoil at the wing pivot
point, changing to an NACA 64(08)A008.28 section at the wingtip. Maximum
wingspan was 32.75 feet; with the wings fully swept back, the wingspan was
just 19.4 feet. The aircraft weighed 10,006 pounds fully loaded and 7,894
pounds without fuel.9
Bell Aircraft made a formal proposal to the Air Force for building a pair
of X-5 research aircraft on February 1, 1949. (The first aircraft was to be
operated by the NACA, the second by the Air Force.) The response was rapid;
just 3 days later, Air Force Headquarters directed that a contract be written
and signed. The “X-5” was to be used to demonstrate the best sweep angle for
lightweight interceptors and the tactical advantages of variable-sweep wings.
But the USAF had no interest at that time in funding an operational aircraft,
and, indeed, variable sweep would not appear on an Air Force combat aircraft
for another 15 years.10
The contract signing initiated a series of reviews of Bell’s design. Engineers
from the Air Force Power Plant Laboratory issued their report on May 9,
1949. They objected to the placement of the fuel tank directly over the
jet engine and requested a redesign. This would have meant designing an
entirely new aircraft, delaying the project significantly. And so the Power
Plant Laboratory engineers withdrew their objections, instead agreeing on
August 5 to approve the existing configuration. (However, they asked that
shielding of the fuel tank be improved, which it was.) They stressed that their
approval applied only in the case of the X-5’s use as a research aircraft and
not as a production fighter.
These were not the only doubts about the X-5 design. Engineers at the
Wright Field Aircraft Laboratory in Ohio issued a critical report on August
29, 1949, in which they urged numerous design changes to make the aircraft
both more airworthy and compliant with Air Force design specifications.
More serious was the engineers’ questioning of the project’s value. The 20°
212
minimum angle was judged to be too low to be of any real use. More sub-
stantive, it seemed, was that the X-5 data would duplicate data from aircraft
either already available or under construction. The engineers concluded, “In
other words, the X-5 does not fill any particular gap in either high-speed or
sweep-back research.”11 What was missed in this was that the greatest value
of the X-5 would be in simply demonstrating the act of in-flight variable
wing sweep itself.12 The NACA, in contrast, strongly supported the X-5’s
development. NACA researchers had decided by mid-1948 that the aircraft
would complement their earlier wind tunnel work. Following Bell’s initial
proposal that a variable-sweep research aircraft be designed and built, a team
from Langley went to Wright Field to argue for Air Force approval. This did
much to win needed military support for the project at a time when the
service faced many competing demands for scarce funding.
Contractor and Air Force Proving Flights

Ground tests with the first X-5 (AF Serial 50-1838) began on February 15,
1951, at Bell’s Niagara Falls facility. The next several months entailed aircraft
system testing, load and structural testing, and taxi runs. The Air Force also
completed engineering inspections. On June 9, 1951, the disassembled aircraft
was crated and loaded onto a Fairchild C-119 cargo plane for the long flight to
Edwards. The test flights and research plan for the pair of X-5s were similar to
those of earlier X-planes. Initial demonstration flights would be made by Jean
L. “Skip” Ziegler, Bell’s chief test pilot. Once the first X-5 had been proven
airworthy, Air Force familiarization flights would follow. With these complete,
the first X-5 would then be turned over to the NACA, which would begin
research flights. The second X-5 (AF Serial 50-1839) would be flown by Air
Force pilots in a separate test effort.
Ziegler took the first X-5 aloft on June 20, 1951. After takeoff, he climbed
to 15,000 feet, using reduced power to keep his speed below the wheels-down
limit. Once at altitude, he tried out the X-5’s control response and found it
to be stiff. He reached a top speed of Mach 0.56 and a maximum altitude of
15,974 feet. So far, the flight had been a success, and Ziegler began to descend.
At that point, he discovered the fuel cells were showing negative pressure.
This meant their internal pressure was lower than that of outside air. If the
reading was valid, as the X-5 descended the increasing outside air pressure could
cause the fuel cells to rupture. Fuel would then spill onto the engine, triggering
a fire. Ziegler made maneuvers to determine whether the reading was correct,
and after several minutes he determined that the fuel-cell pressure reading was
false. He landed several minutes later without incident.
213
Probing the Sky
The fuel-cell pressure gauge error and several other problems had to be
addressed, and the aircraft was grounded for 5 days. The June 25 flight went
off but yielded no data. The next two flights, on June 27 and 28, reached Mach
0.638 and 0.696. Zeigler undertook the first wing-sweep test on July 27. This
involved only a partial movement, rather than the full range of 20.25° to 58.7°.
The Bell contractor tests continued into late August 1951, with nine flights made.
Brigadier General Albert Boyd, the Air Force Flight Test Center com-
mander, received permission from the Wright Air Development Center to
make an evaluation flight in the X-5 on August 23. General Boyd took off,
climbed to 40,000 feet, and moved the wings back to their full 58.7° sweep
angle. He then descended to 30,000 feet, reached a speed of Mach 0.92, and
made several more test maneuvers. Boyd’s total flight time was 28 minutes.
The NACA progress report on the X-5 stated that the primary goal of Boyd’s
flight was to demonstrate a landing with a 40° sweepback. Data were obtained
on three landing configurations:
• Landing at alternate gross weight, slats out, and with wings at a
20° sweep
• Landing with slats out and wings at a 20° sweep
• Landing with flaps at 35°, slats out, and at a sweepback of 40°
The first two landings were made without incident, but during a landing
approach at 40°, Boyd accidently raised the flaps by striking the flap switch.
The X-5 experienced a stall roll-off, striking the ground on its nose and right
main wheels before Boyd could recover. The impact caused minor damage to
the tire tread. No structural damage was listed.13
The Bell test effort with the first X-5 ended on October 8, 1951, when
Ziegler made his final flight in the NACA aircraft. The first X-5 was grounded
pending the installation of NACA instrumentation. The initial payload mea-
sured the following quantities:
• Vertical, longitudinal, and transverse acceleration
• Sensitive longitudinal acceleration
• Rolling angular velocity
• Pitching angular velocity and acceleration
• Yawing angular velocity and acceleration
• Airspeed and altitude
• Angle of sideslip and angle of attack
• Control positions
• Wing-sweep angle
• Elevator and aileron stick forces14
Along with instrument installation, the X-5 was inspected. The fuel cells
were replaced, and functional checks of them were made. These preparations
were expected to be completed by the week of December 17, 1951. Once the
214
work was complete, Air Force demonstration flights could begin. These were to
be made by Lieutenant Colonel Frank K. “Pete” Everest.15
Everest’s first two flights were made on December 20. The first was for pilot
familiarization, and no data were collected. A second flight the same day reached
Mach 0.844 and 22,380 feet. Everest made two more flights on December
21 and 27. His final pair of flights came on January 7 and 8, 1952. The tests
involved clean stalls with wing sweeps of 20°, 40°, and 60° and multiple accel-
erated turns at the three sweep settings of clean, slats out and flaps down, and
flaps down. Thomas W. Finch, NACA X-5 project manager, summarized the
initial results of the flights in two research memorandums. In the first, he said:
At low speed for 20° sweepback an increase in longitudinal stability

occurred as maximum lift was approached. At 40° and 60° sweep a
reduction in stability occurred at the higher lift coefficients. During
the tests at 30,000 feet to investigate the effects of lift coefficients
at Mach number of approximately 0.8—for the 20° sweep, there
occurred a reduction in longitudinal stability at lift coefficient of
approximately 0.4. No change in stability was noticed at 40° sweep
up to lift coefficient of 0.5 (the limit of tests to date). For 60° sweep
an abrupt pitch-up occurred at a lift coefficient of 0.6. Stick forces
were of the order of 20 and 35 lbs/g at the high Mach numbers.
Structural integrity demonstration required 5.86 g at Mach of
0.8 at 12,000 feet which corresponds to a lift coefficient of approxi-
mately 0.5. Pull-ups made at 20° and 60° sweep at this speed and
altitude were limited to load factors less than 4gs due to the high
stick forces.16
And in the second, he said:
There was a slight increase in stability at 20° sweep from M = 0.75

to M = 0.79, and the stability at 60° sweep remained essentially
the same for a Mach number range of 0.7 to 0.8. At 40° sweep
there is an appreciable increase in stability at M = 0.81 above that
for M = 0.71 and M = 0.76. During the tests to investigate the
effects of lift coefficients up to M ≈ 0.8, instability was found for
the following conditions: (1) for 40° sweep there was a decrease in
stability at M ≈ 0.75 and at M ≈ 0.81 a mild pitch up occurred at
a CAn [airplane normal-force coefficient] of 0.5. (2) for 60° sweep
an abrupt pitch up occurred at a CAn of 0.6 and M ≈ 0.8.
The only indication of instability occurred at 20° sweep and
20,800 feet with a decrease in stability above a CAn [of ] 0.4.
215
Probing the Sky
A load factor of 5.86g could not be demonstrated at M = 0.8

and 12,000 feet because of high stick forces. At lower Mach num-
bers of 0.7 and 0.75 the stick forces were essentially the same,
and the lift coefficients for instability were being approached or
exceeded. In view of these results it was decided to accept the air-
plane since it became apparent that the longitudinal stability eval-
uation required a more complete analysis, and the flight testing
was entering into the research program planned for the airplane.17
NACA X-5 Research Flights Begin

With the X-5 officially transferred to the High-Speed Flight Research Station,
Joe Walker was selected as the initial NACA X-5 pilot. He made his checkout
flight on January 9, 1952. Walker’s first research flight came on January 14,
1952, and dealt with static and dynamic longitudinal and lateral stability and
control. Ultimately, the bulk of the flights in the X-5 focused on this type of
research. Walker was kept busy flying the X-5, completing 17 flights between
January 9 and April 1. Though most of the flights were for stability and control,
three were for airspeed calibration and another was for a gust loads investigation
at 20° and 60° sweep angles. Finally, on February 12, Air Force pilot J.C. Meyer
had a check flight in the NACA’s X-5. (This was the only flight not made by
Walker during this period.)
Not everything went smoothly. The March 13 flight was aborted after take-
off, and Walker’s first flight on March 17 was aborted as well due to “poor volt-
age, no data.”18 The problem was corrected, and a second flight that day was
made for lateral and longitudinal stability and control data at a 60° sweep angle.
The test data from the Bell acceptance flights had been analyzed, and a
research memorandum emerged. The report offered a more detailed look at
the data than had initial observations in Finch’s progress reports. The memo-
randum on the X-5’s horizontal-tail load measurements noted:
During sweep changes at an altitude of 20,000 feet, the trends

of the balancing tail loads variation with sweep angle at Mach
numbers of 0.50, 0.56, and 0.85 were similar with the larger
down tail load occurring at approximately 36° sweep angle at each
Mach number tested. The largest balancing tail load in a down
direction over the entire sweep range from 20° to 59° occurred at
a Mach number of 0.85.
216
During level-flight-trim points at 20° sweepback at an alti-

tude of about 20,000 feet and an airplane weight of about 8,800
pounds, the tail load increased in a down direction as the Mach
number was increased from 0.54 to 0.84.
During pull-ups at a Mach number of about 0.83 and at
sweep angles of 20°, 45°, and 59°, the wing-fuselage combination
was stable at 20° and 45° sweepback and unstable at 59° sweep-
back at normal-force coefficients less than 0.3. At normal-force
coefficients near 0.3 the stability of the wing-fuselage combina-
tion changed for these sweep angles, becoming unstable at a sweep
angle of 20°, experiencing a reduction in stability at 45°, and
became stable at 59° as the normal-force coefficient increased.
For the normal-force-coefficients range covered in these tests the
aerodynamic center of the wing-fuselage combination moved
rearward as the sweep angle was increased from 20° to 59°. A larger
change in the wing-fuselage aerodynamic center was experienced
with the sweep for the high normal-force-coefficient range than
was experienced for the low normal-force-coefficient range.19
In April 1952, A. Scott Crossfield and Walter P. Jones began making X-5
flights. Crossfield’s check flights came on April 2, Jones’s on April 29. Crossfield
had already talked with Walker about the X-5’s handling. Crossfield recalled
later, “The X-5 handled in the air like a three-wheeled automobile. It was loose
and danced crazily. Even so, we thought it would make a fine research tool.”20
The X-5’s most serious shortcoming was its spin characteristics. Both Walker
and Crossfield would experience these in full fury.
Crossfield’s flight plan called for making several accelerated stalls. After
reaching 25,000 feet, he pulled back on the throttle and the stick. He wrote later:
As the X-5 slowed, she began to buffet. Suddenly her nose veered
sharply to the left. In a split second, the X-5 turned 180 degrees.
Then she dropped precipitously into a spin.
A kaleidoscope of brown desert, blue sky, and white clouds
passed dizzily in review in my windshield as the X-5 wound up
steadily towards the desert floor. I pressed the stick hard to for-
ward left and bent on full right rudder—the prescribed spin recov-
ery maneuver—but the X-5 stubbornly refused to conform. After
a drop of over 10,000 feet, the X-5 pulled out.21
Walker had experienced the X-5’s spin behavior the month before; he had
entered a spin at 40,000 feet and had not recovered until reaching 20,000
217
Probing the Sky
feet. He described the experience in his pilot report: “As the airplane pitches,
it yaws to the right and causes the airplane to roll to the right. At this stage
aileron reversal occurs; the stick jerks to the right and kicks back and forth from
neutral to full right deflection if not restrained. It seems that the airplane goes
longitudinally, directionally, and laterally unstable in that order.”22
During the spring and summer of 1952, pilots flew the X-5 research plane for
longitudinal control effectiveness, static lateral stability and control characteris-
tics, and dynamic longitudinal and lateral stability and control data. Between
May 3 and May 16, four research flights were made, three by Crossfield and
one by Jones. Maneuvers included speed runs, aileron rolls, sideslips, elevator
and rudder pulses, and stall approaches. The bulk of the maneuvers were flown
with the wings at a 60° sweep angle. Speeds ranged from Mach 0.47 up to
0.94, altitudes from 36,000 to 42,000 feet. The initial results were as follows:
Preliminary evaluation of the records and pilot’s notes indicated

that all speed runs were easily controllable with elevator. The nose
down trim change reported in the speed run…occurred about
0.93 which is in agreement with previously reported data.
The aileron effectiveness is low and approximately constant
up to M = 0.9. The rudder effectiveness determined from side-
slip data is noticeably improved below M = 0.70 as compared
to an approximate constant effectiveness between M = 0.70 and
M = 0.90. Almost total aileron control available is required for
sideslip angles of about 5° at Mach numbers below 0.70.
The values of the period and time-to-damp of the longitudinal
and lateral oscillations obtained from pulses below M = 0.90 were
in agreement with previously reported data. Data have not yet
been reduced from maneuvers performed above M = 0.9.23
Crossfield and Jones continued to make longitudinal and lateral stability

data flights into late June 1952. Walker made his next flight on June 25, 1952.
These flights continued static and dynamic longitudinal and lateral stability
data collection. Walker made flights on July 2 and 10. The July 2 flight and
most of the July 10 flight collected additional static longitudinal and lateral
stability data at 59° wing-sweep angles. These involved speed runs, pushovers
and pull-ups, aileron rolls, and gradually increasing sideslips.
Walker’s July 10 flight also generated data on lift and drag at a 20° sweep
angle. He made speed runs as well as pushovers and pull-ups with the wings at
a 20° sweep. The pushover was made to a g-meter reading of 0 g, followed by a
pull-up into the longitudinal instability boundary. These runs were made to
obtain accelerated maneuver data starting from zero lift. During this flight,
218
Walker also made fast-rate stabilizer pull-ups into the instability region while
flying in a speed range of Mach 0.70 to 0.96. He performed these maneuvers
to get baseline data, which would be compared to data collected following
the installation of modified wing-root fillets.
By July 25, Walker had made another three research flights, with all the tests
conducted at a 59° sweep angle. As on his earlier flights, he performed gradu-
ally increasing sideslips, pushovers and pull-ups, and a speed run. One of the
flights was made with the modified wing-root fillets. A wind tunnel test with
a model that resembled the X-5 indicated that changing the wing-root shape
would prevent the reduction of longitudinal stability that characterized the X-5.
For these tests, stabilizer pull-ups were made from Mach 0.70 to 0.965, in
a 2° dive from 40,000 to 38,000 feet. Mindful of how fast the aircraft could
lose altitude in a spin, pushovers and pull-ups were made at Mach 0.70 to
0.84 and 40,000 feet, pull-ups at Mach 0.965 and 38,000 feet. The results
were summarized as follows:
Static longitudinal instability was encountered on all pull-ups

both in the original configuration and with modified wing root
fillets. The pilot reported increased stability at lower normal-force
coefficients with the modified wing root fillets at the higher Mach
numbers tested while the instability seemed to occur at slightly
higher load factors.
Data from the speed run performed in a 20° degree dive
has not been reduced but the pilot’s Machmeter reading of 0.93
indicates a Mach number greater than 0.98 has been reached.
The speed run was easily controlled with elevator with the pilot
reporting maximum stick forces of about 10 pounds push at the
trim change and about 15 to 20 pounds pull at the maximum
Mach number.24
Walker expanded the longitudinal and lateral stability and control flights
in late July and early August 1952. To extend the speed range of the X-5,
shallow dives were made to exceed Mach 1. He began the dives at angles
of 20° and 30°, starting at 45,000 feet, and pulled out at 38,000 feet at a
wing-sweep angle of 59°. The test maneuvers were much the same as those
on earlier flights: aileron rolls, gradually increasing sideslips, pushovers and
pull-ups, and stabilizer pull-ups. (The latter were done at both 59° and 20°
wing-sweep angles.) During the July 25 flight, Walker reached a maximum
speed of Mach 0.986. He achieved Mach 1.006 on both the August 1 and 7
flights. The preliminary results of the three flights were summarized in the
X-5 progress report:
219
Probing the Sky
Preliminary inspection of the records and pilots comments indi-

cate that near M = 1.0 the maneuvers were performed without
any difficulties. About 35 pounds aileron force was required for
a full deflection roll and about two-thirds the available aileron
control was required in sideslips to β ≈ 1.5°.
The dives were easily controlled with elevator with a maxi-
mum push force of about 60 pounds being encountered in the
dive at…–3°. Elevator forces up to about 180 pounds was required
in the pullout from the dives near M = 1.0.
Longitudinal instability was encountered in all pull-ups. The
trend of the instability boundary above M = 0.96 will not be
known until the data has been reduced from the pullouts.25
Additional research flights were made in September at the 59° sweep angle
before being halted. The X-5 did not fly again until October 21, with Walker
at the controls. Pilots of the era would likely have described Walker’s experience
that day as “eventful.” The progress report noted:
At 59° sweepback a 30° dive was made to M = 1.04 with a sta-

bilizer setting of –2.5°. Trim force for 1g was about 45 pounds
push at maximum speed. An elevator pull-up to about 2g was
completed above M = 1.0 without encountering any reduction
in stability. A stabilizer pull-up was started above M = 1.0 with
a reduction of stability being encountered at better than 2g as
the Mach number decreased. The remainder of the flight was to
be composed of pushovers to 0g followed by pull-ups through
the reduction of stability boundary starting from the trimmed
stabilizer deflection. However, a violent spin occurred following
the reduction of stability during a pull-up at a Mach number near
0.7. As the airplane yawed to the right following the reduction
in longitudinal stability it abruptly snap-rolled and went into a
spin at an altitude of about 44,000 feet. There was a high rate
of roll (greater than 3 radians/sec) about the longitudinal axis.
The elevator hinge moments caused the elevator to float up; the
forces were too high for the pilot to apply full down elevator. The
recovery was not affected until the stabilizer control had been in
the full leading-edge up position (+4.4°) for several seconds with
ailerons and rudder opposing the spinning direction. Recovery
was completed at an altitude of about 37,000 feet.26
220
Following the spin, mechanics inspected the aircraft and the flight data
were analyzed for any indication that the aircraft had been overstressed. The
progress report noted:
Strain gage data indicated that compression loads imposed on

the sweep mechanism jackscrews approached limit load. It may
be necessary to examine and test the various components of the
sweep mechanism. The engine was removed for inspection and
cracks were found in the nozzle diaphragm as well as evidence
of rubbing between the turbine wheel and the diaphragm. The
engine will be replaced as soon as the oil tank can be modified to
accommodate a –B engine.
A “g” limitation that has existed for the X-5 wing is being
lifted by replacement of a number of rivets on each wing panel
aft of the pivot point by high sheer bolts. The present limitation
of 5.65g will be then lifted to 7.33g.27
Given the complexity of the inspection and the resulting repair work, it is
not surprising that the X-5 was grounded until December 5, 1952. The aircraft
made a total of seven flights in December, another five in January 1953. These
flights, unlike those made earlier in the year, had a wider range of goals. Only
three of the December flights dealt with longitudinal- and lateral-stability and
control data; two were checkout flights for Stanley P. Butchart; another was a
photographic flight; and one was aborted after takeoff due to an inoperative
stabilizer motor. The January 1953 flights involved two made for vertical-tail
load measurements during maneuvers, another for stalls and maneuvers at a
20° wing sweep, and the check flight of Major Arthur A. “Kit” Murray. The
last was made to compare the behavior of the first X-5 with that of the second,
which was being flown by the Air Force.28
The other two flights in January and the first flight in February all had an
unusual pair of goals: “investigation of flight at speeds below minimum drag of
the test airplane and testing the controllability of a highly swept-wing aircraft
in the wingtip wake of a straight-wing aircraft.”29 The tests originated with a
verbal request made by Wright-Patterson engineers to Walter Williams during a
visit to Edwards. The data they sought would be used in the design of a tanker
capable of refueling an interceptor aircraft in flight. Originally, the delta-wing
XF-92A was suggested as the test aircraft. Analysis indicated that preliminary
results could be obtained at a much earlier date using the X-5 flying with a 59°
sweepback behind an F-80 and a B-29 bomber.
221
Probing the Sky
Multiple exposures showing the full range of travel of the X-5 wings. The sweep angle could be
adjusted in flight depending on test requirements. One unusual effort undertaken with the air-
craft involved the X-5 flying behind a B-29 in a test of how a swept-wing aircraft would handle
while taking on fuel from a straight wing/propeller-powered tanker. (NASA)
Walker made the research flights. The first, on January 27, 1953, was flown
with the X-5 trailing an F-80 Shooting Star, the first production U.S. jet fighter.
The X-5’s wing sweep was 59° at an altitude of 25,000 feet. The first part of
the formation tests involved stabilized test points recorded at intervals from
290 to 190 mph. A second set was done in a continuingly decreasing speed
run between 290 and 190 mph. Finally, a continuously increasing speed run
was made from 190 to 290 mph. In the X-5, Walker took a position 10 to 15
feet behind the F-80’s wingtip, which he used as a reference point. His pilot
report on the January 27 flight noted:
Accurate position could be maintained at any of the test speeds

and during the accelerating and decelerating runs once the posi-
tion was established. Getting into position was difficult at these
test speeds. Throttle control of speed was effective only for minor
222
position changes once position was established. At other times,

except at high speed, power change was so slow that speed changes
were difficult and slow. A fast acting effective speed brake com-
bined with a control switch on the throttle is the only practi-
cal means of speed control for join up. If the speed brakes were
capable of variable positioning, throttle regulation would be made
easier in case of two dissimilar aircraft attempting to maintain an
accurate position relative to each other.
The variations in power required to maintain position at dif-
ferent speeds was most apparent during the deceleration and accel-
eration speed runs. At minimum speed of 190 mph there was still
ample reserve power. The power required had not started increas-
ing rapidly down to 190 mph. It was relatively less difficult to
maintain position at 190 mph than at 250 mph because of larger
airplane drag changes with slight changes of angle of attack.30
The other two flights, on January 29 and February 6, 1953, were with a
B-29 at 40,000 feet. The X-5’s wing sweep was again 59°, with the stabilizer
incidence at –2°. Level formation runs were made with the B-29 at speeds of
230, 220, 210, 200, and 190 mph. Walker commented later that he experi-
enced “less difficulty” maintaining position with the B-29 than with the F-80,
as the bomber “didn’t bounce around as much,” adding, “It was not difficult to
maintain fore and aft and vertical position. There was plenty of elevator and
rudder control. However, lateral control with the ailerons gradually deterio-
rated as speed was reduced.”31 A problem appeared at 190 mph, when Walker
extended the X-5’s wing slats. He later wrote, “This configuration is not recom-
mended. As soon as slat extension started, extreme sensitivity in roll developed.
The airplane gained nose up trim. The ailerons and elevator became noticeably
stiffer. It appeared that the static longitudinal stability stick-fixed and stick-free
decreased markedly. The result was an all out effort to maintain the wings level
and increased difficulty maintaining vertical position…. It should be possible to
reduce speed below 190 mph in the clean configuration without encountering
as much control difficulty as with the slates extended at 190 mph.”32
Walker flew the third and final formation test on February 6, 1953, again
trailing a B-29. The flight went smoothly, with Walker’s postflight report noting:
Because the airplane [X-5] buffeted at 176 mph and the stability
decreased at that speed, 180 mph was set as the minimum speed
for tests, clean configuration, of ability to fly formation below
minimum drag. A record was obtained while getting into position
on the lead airplane which was flying at 190 mph. No particular
223
Probing the Sky
difficulties were encountered which would not be present at any

speed during the join up.
The only noticeable handicap to maintaining position at 180
mph was the small amount of control remaining and large deflec-
tion required. This was minimized for the X-5 because of the
component of thrust acting as an additional longitudinal control.
It actually appeared easier to maintain position at 180 mph
than at 240 mph because of greater drag at trim and larger drag
changes with small angles of attack changes. It was found, for
example, that a slight tendency to overshoot could be quickly
stopped merely by raising the nose of the airplane slightly.
Based upon observations during these flights, for the pur-
poses of aerial refueling it would be feasible to fly below the speed
for minimum drag of a highly swept wing airplane and may be
more easily done than at the speed for minimum drag, providing
reserve power is available for maneuvering.33
Initial X-5 Research Results

As NACA research flights with the X-5 began, the flight data were being
simultaneously analyzed. This was a complex process that began with data
reduction and was followed by the writing of a rough draft, a review by
High-Speed Flight Research Station staff, and further reviews and comments
from other NACA laboratories. Not until early 1953 were the initial research
memorandums published.
The first of these dealt with flight measurements of the X-5’s stability char-
acteristics in sideslips with the wings at a 59° sweep angle. The flight condi-
tions were at Mach 0.62 to 0.97 between 35,000 and 40,000 feet. The results
indicated the following:
Throughout the Mach number range the apparent directional

stability is positive. It is constant from a Mach number of 0.62
to 0.90 and increases to a value about 60 percent higher as the
Mach number increases to 0.97.
The apparent effective dihedral is positive and high, doubling
in value between Mach numbers of 0.75 and 0.97. The dihedral
is so high at the highest Mach number that the ailerons can only
trim out about 2.5° of sideslip and can be easily overpowered by
the rudder.
224
The cross-wind force coefficient per degree of sideslip is stable

and constant to a Mach number of 0.94, above which it increases
rapidly to a Mach number of 0.97.
There is little or no change in pitching moment due to sideslip.34
Another research memorandum looked at the lift and drag measurements

at different sweep angles, including a comparison of the lift and drag at 59°
and 20° sweep angle configurations. The data indicated the following:
At an altitude of about 42,000 feet the drag force for the 20°
sweep configuration in unaccelerated flight was considerably
less than that for the 59° sweep configuration for flight speeds
below a Mach number of 0.81; above a Mach number of 0.82
the reverse is true.
At an altitude of 42,000 feet and a Mach number of 0.74 the
total drag force for the 59° configuration was more than twice
that for the 20° configuration at any given lift coefficient below
maximum lift.35
Other important research was devoted to determining the X-5’s buffeting

characteristics, as these produced bending stresses and moments, shear loads,
and torque on the aircraft. Buffeting also affected controllability. Flight mea-
surements of the X-5’s buffeting were made at a 58.7° sweep angle, at a speed
range of between Mach 0.65 and 1.03, and at altitudes between 37,000 and
43,000 feet. Maximum airplane normal-force coefficients were reached for
Mach numbers up to 0.96. The research memorandum’s concluding remarks
on the X-5’s buffeting read:
At all airplane normal-force coefficients the horizontal tail was

found to experience buffeting…. At angles of attack below maxi-
mum airplane normal-force coefficient the peak buffet-induced
tail-bending stresses were 20 percent of the maximum observed
tail steady-state bending stress; the peak buffet-induced tail shear
loads were approximately 5 percent of the tail design limit load.
Wing buffeting began at moderate angles of attack but above
the wing buffet boundary there was no appreciable increase in
wing buffet intensity with increase in Mach number or angle of
attack. At angles of attack below maximum airplane normal-
force coefficient, the peak buffet-induced wing bending stresses
were 5 percent of the maximum observed wing steady-state
bending stress.
225
Probing the Sky
Coefficients of incremental normal acceleration greater than

±0.05, considered to be high-intensity buffeting on other research
airplanes, were not experienced by the X-5 airplane below normal-
force coefficient. The pilot considered the buffeting to be “unob-
jectionable” throughout the entire test region.36
Though most of the research was focused on the X-5’s basic aerodynamics,
stability and control, along with the effects of the swept wings, changing the
shape of the wing roots to improve stability was also examined. This effort
illustrated the value of actual flight testing as a means of verifying wind tunnel
results. Replacing the original 52.5° sweepback leading edges with rounded
leading-edge fillets showed improved stability in the wind tunnel, but flight
tests produced different results. The research memorandum noted:
A comparison was made between two configurations of the Bell

X-5 research airplane at 59° sweepback, one with the original
wing-root fillets and the other with the wing-root fillets shown by
low-speed wind-tunnel to eliminate the loss of stability at high-
lift coefficients. The data obtained from the flight investigations,
however, show that the longitudinal stability characteristics, as
well as the buffet and drag characteristics, were essentially unaf-
fected by the modification.37
Dark Day: The Loss of Ray Popson and the Second X-5
While NACA research pilots were flying missions in the first X-5, Air Force test
pilots were conducting their own flight tests with the second aircraft. As the first
X-5 underwent contractor flights in 1951, the second X-5 (AF serial number
50-1839), meant for use in the Air Force test effort, was under construction. It
was completed in early October and arrived at Edwards aboard a C-119 cargo
plane on October 9, 1951. It was test-flown by Ziegler on December 10 and
delivered to the Air Force 8 days later. The Air Force test effort was an exten-
sion of Bell Phase I testing and involved stall testing with wing slats closed.
Subsequently, the X-5 would make takeoffs with slats closed. The goal was to
determine whether they could be closed permanently. Other tests involved the
nose-mounted speed brakes. The X-5 had experienced severe buffeting during
the Bell flights at speeds at which the aircraft were to be used. The Air Force
wanted more data on the cause.38
226
On October 13, 1953, Major Raymond A. “Ray” Popson took off from
Edwards for a familiarization flight in the second X-5. Such flights were
common Air Force practice for test pilots at Edwards. Popson was an expe-
rienced and skilled test pilot. Major Arthur A. “Kit” Murray was the chase
pilot, flying an F-86. The takeoff and climb to 40,000 feet were uneventful.
Popson then began a series of stalls with the wing sweep at 20°, 40°, and 60°,
as well as level flight accelerations. These maneuvers were followed by a dive
that incorporated a 20° sweep angle and pull-up. During the pull-up, Popson
swept the wings back to 60°. He then entered a dive, reaching a speed of Mach
0.96 and leveling off at 35,000 feet.
The first indication of trouble appeared at that point. Popson radioed
Murray that the X-5 had lost oxygen pressure. Murray answered that Popson
should immediately descend, which he did. During the descent, made with the
speed brakes extended, Popson indicated he had 25 to 50 pounds of oxygen
pressure. This was sufficient for continuing the flight.
The NACA X-5 parked on the ramp at South Base. The X-5 aircraft had a bulky design that led
to poor flight characteristics. Spins were a particular problem and resulted in the loss of the
second X-5 as well as the death of Air Force pilot Ray Popson. (NASA)
227
Probing the Sky
The two aircraft leveled out at 10,000 feet. Popson radioed Murray that the
X-5 had 120 gallons of gas remaining and he wanted to “feel the aircraft out
some more” before landing. The pair climbed to 12,000 feet above sea level
(about 10,000 feet above ground level), and Popson began a series of stalls at
20° and 40° sweep angles. He then began a third series of stalls at 60°, but the
X-5 slowly rolled over onto its back and entered a spin to the right. Murray
radioed Popson to eject during the third turn, but received no reply. The X-5
struck the ground after four to six turns, in a 60° diving turn. Popson made
no ejection attempt and died on impact.39
Although the accident investigation board concluded that the cause of the
crash was loss of control leading to an unintentional spin at an altitude too low
for successful recovery, board members felt that several factors had contributed
to the loss of control. One of these factors was that analysis of crash debris indi-
cated that the wings had been at a 45° to 47° sweep angle when the X-5 hit the
ground. As the sweep angle was known to be 60° at spin entry, the board believed
Popson had attempted to move the wings to a 20° sweep angle. The reaction to
sweeping the wings forward during a spin was unknown, but in straight and
level flight, a trim change had been noted during wing-sweep changes.
The X-5’s cockpit pressurization system was not operational at the time
of the flight. This meant Popson would have been using 100-percent oxygen
from takeoff until reporting the loss of oxygen pressure. However, the supply
of oxygen would have lasted about 35 minutes. Popson reported the loss of
oxygen pressure after 26 minutes of flight time. Roughly 2 minutes later, he
radioed that he had 25 to 50 pounds of oxygen pressure. The X-5’s crew chief
told the accident board he was certain a full supply of oxygen was on board
the aircraft before it took off. If the numbers Popson reported were an accurate
reading, this would indicate excessive leakage or a partial malfunction within
the oxygen system.40
The medical official on the accident panel expressed the opinion that the
conditions described could induce some degree of hypoxia, but to an extent that
was unknown. Although Popson’s actions and radio communications during
the 20° and 40° sweep-angle stalls had seemed normal, hypoxia could have been
a contributing factor in the accident, he asserted. Popson’s intentions for the
60° stalls at 12,000 feet were unknown; they had not been part of the original
familiarization flight plan. Full exploration of the X-5’s stall characteristics at
so low an altitude “would have been unwise,” the accident report noted.41 But
Popson had extensive experience and should have been able to recover success-
fully. The board considered the possibility that Popson had made mistakes in
attempting to recover from the spin, such as holding the ailerons against the
spin rather than with it, or using excessive elevator trim. The accident report
stated, “This is considered less likely, however, than the assumption of partial
228
In-flight photo of the X-5 over Edwards Air Force Base with the wings in the midrange position.
As many of the other X-planes did, the X-5 suffered from poor stability and control. The stall/
spin behavior was very poor, requiring 10,000 feet to make a recovery. This was the cause of
the loss of Major Popson during a familiarization flight in the second aircraft. (NASA)
hypoxia[,] which would reduce the pilot’s ‘keenness’ or ‘edge’ of reaction.” (But
the report did not officially endorse the hypoxia theory.)42
As was routine in such incidents, panel members questioned Major Murray.
He was asked if Popson’s voice had sounded normal during radio communica-
tion, or if there had been any indication he was suffering from hypoxia. Murray
responded, “At the time I did not [think there was anything unusual], he was
perfectly normal, it seemed….” At a later point, a panel member asked, “Again,
he seemed normal, his words were used properly, no slurring of speech?” Murray
replied, “When talking through an oxygen mask a voice sounds garbled, at the
time I was under the impression that he was in a normal condition, gave no
hesitation in answering.”
One member also asked Murray: “Have you ever thought there was any
indication that the flight characteristics of this aircraft differed from the
NACA aircraft?”
Murray, who had about 40 hours of flight time piloting both of the X-5s,
replied, “They differ quite a bit in the high speeds in that [the Air Force’s]
airplane tends to become uncontrollable [at] .93 to .96 and we have noticed a
great difference in the stall characteristics. In accelerated stalls and unacceler-
ated stalls the airplane will recover as soon as the elevator was put forward and
after speed was regained lateral control came back.”
229
Probing the Sky
Joe Walker also testified before the accident panel. One of the members
noted that Popson had stalled the X-5 with a 60° wing sweep, in a clean con-
figuration, and entered a spin. He asked Walker if he had experienced any
similar conditions. Walker replied:
In a 1g stall, clean configuration, 60° sweep I have experienced

several different conditions. The airplane exhibited increasing loss
of lateral control and abrupt yaw to the left. The most vicious stall
approach was with slats extended and occurred with no warning.
This took the form of a half snap roll to the left. If excess right
rudder were used it would probably have been to the right. It
seemed like I had perfect control of the airplane before the snap
roll. Application of trim by actuation of the stabilizer with the
electrical motor results in excess control which pitches the airplane
into a spin condition near the stall.43
Walker was asked later if he had anything he wanted to add to his testimony.
He replied, “I would like to say a 1g stall approach at 60° is accompanied by
large loss of altitude. You get an increase of drag with an increase of angle of
attack. Below 170 mph use of full throttle still requires about 5,000 feet altitude
to gain sufficient speed to pull out of the dive.”44
The final conclusion was, “The Board is of the opinion that this accident
was the result of loss of control of the aircraft resulting in an unintentional
spin at insufficient altitude for recovery but it is felt that several other factors or
conditions could have contributed to this loss of control. They are as follows:
‘Wing Sweep Control and Oxygen System and Pressurization.’ The panel rec-
ommended that stalls only be performed above 20,000 feet in research aircraft
with unknown qualities.”45
Although not mentioned in the accident investigation, a breakdown in com-
munications between the Air Force and the NACA had occurred. Crossfield
recalled later that following Walker’s and his own X-5 spin, a note had been
added to the X-5 pilot’s handbook prohibiting any maneuvers that could result
in a spin at an altitude below 20,000 feet. When Crossfield learned that Popson
had made aggravated stalls at 12,000 feet, he later recalled, “I was sick. We
had failed in a basic NACA mission—getting information to the right place
in time…. If there had been better coordination between [the] Air Force and
NACA, [Popson] might be alive today.”47
The medical safety division had its own comments: “One fact apparently
overlooked is that even after descending to 12,000 feet…where stalls were prac-
ticed for about 10 minutes, the cabin pressure would have been near 13,000
feet to 14,000 feet and hypoxia possibly induced at a higher altitude would not
230
The first X-5 in January 1952, after it was transferred to the NACA. A total of 122 NACA
research flights were made with the X-5, most of them piloted by Joe Walker. In contrast,
A. Scott Crossfield made 10 flights between 1952 and 1954. He later commented that the low-
slung engine resulted in a misalignment of the drag axis, the principal axis, and the thrust axis.
“So,” he later said, “it could get into some interesting maneuvers and motions….”46 (NASA)
be completely recovered from since mild hypoxia is actually induced at these

altitudes. Concur generally with findings.”48
The X-5’s Later Years—1953–1955

The NACA continued to fly the surviving X-5 following the Popson crash. The
number of flights made in the aircraft each year gradually declined, however.
In 1953, a total of 31 flights were made with the X-5 (four by Crossfield, one
by Murray, and the rest by Walker). Research goals included collecting data
on vertical-tail loads, gust loads, effects of wing transition on aircraft trim,
longitudinal stability and control, and wing twisting and bending tail loads.
The X-5 made a total of 27 flights in 1954, with Walker, Crossfield, Stanley
P. Butchart, and John B. “Jack” McKay as pilots. Much of the research work
continued to focus on longitudinal and lateral stability and control, as well
as vertical-tail loads. Additional research included measuring the effects of
dynamic pressure on buffet. Most of the flights were made between January and
June of that year. The X-5 then was grounded, awaiting delivery and installation
of a new nose gear housing from the Electrol Company. The aircraft did not fly
231
Probing the Sky
again until December 14, when Butchart conducted an instrumentation check

flight. The following day, McKay made the first of four pilot checkout flights.
The X-5 project made its final research flights in 1955, totaling 19 flights
in all. The first flight, made on January 27, was an instrumentation check by
Butchart and reached Mach 0.922. The next eight flights, made by Butchart
and McKay between January 28 and March 23, were for longitudinal-stability
and control data. Six more flights, starting on March 23 and continuing
through April 8, were for lateral-stability and control measurements. Flights
then halted until the fall.49 A limited number of additional flights were still
necessary to collect lateral control effectiveness data and to obtain an angle-of-
attack calibration. McKay made these on October 19, 20, and 24, 1955. The
134th and final NACA flight was made on October 25 by Neil A. Armstrong.
During the pilot check flight, a landing gear door separated. Armstrong landed
safely, but the aircraft never flew again.50 The X-5 was put into storage for the
next several years. In March 1958, it was sent to the U.S. Air Force Museum
at Wright-Patterson Air Force Base. The aircraft was subsequently restored
and now is on display in the museum’s Research and Development Gallery.51
As the X-5 flights were winding down, the final set of research memoran-
dums on the project was completed, reviewed, and approved. The first of these,
which dealt with wing loads at a sweep angle of 58.7°, stated:
Flight measurements on the Bell X-5 research airplane at a sweep

angle of 58.7° have shown the wing loads exhibit nonlinear trends
over the angle-of-attack range from zero to maximum wing lift.
These nonlinearities were, in general, more pronounced at angles
of attack above the “pitch-up” where there is a reduction in the
wing-panel lift-curve slope and an inboard and forward move-
ment in the center of load. These characteristics have been found
to exist in the results of wind-tunnel tests of swept wings and
emphasize the need of model testing for accurate wing design
data when nonlinearities exist.
No apparent effects of altitude on the wing loads were evident
over the comparable lift ranges of these tests at altitudes from
40,000 to 15,000 feet.52
The second memorandum looked at the loads on the X-5’s horizontal tail.
Its conclusions read:
An investigation of the horizontal-tail loads on the Bell X-5

research airplane at a sweep angle of 58.7° has indicated that the
horizontal-tail loads are nonlinear with lift throughout the lift
232
ranges tested at all Mach numbers except at a Mach number of

approximately 1.00. The balancing tail loads reflected changes
which occur in the wing characteristics with increasing angle of
attack. The nonlinearities of the horizontal-tail loads were gener-
ally more pronounced at the higher angles of attack near the pitch-
up where the balancing tail loads indicate that the wing-fuselage
combination becomes unstable.
No apparent effects of altitude on the balancing tail loads were
evident over the comparable lift ranges of these tests at altitudes
from 40,000 feet to 15,000 feet.
Comparisons of balancing tail loads obtained from flight and
wind-tunnel tests indicated discrepancies in absolute magnitudes,
but the general trends of the data agree. Some differences in abso-
lute magnitude may be accounted for by the tail load carried
inboard of the strain-gage station and the load induced on the
fuselage by the presence of the tail. These loads were not measured
in flight.53
Research with the X-5 was primarily focused on longitudinal and lateral
stability and control. Most of the research flights were made for these purposes,
and the resulting research memorandums were both lengthy and detailed. One
such memorandum noted:
An investigation of the dynamic stability of the X-5 research air-

plane at 58.7° sweepback at altitudes of 40,000 feet and 25,000
feet over a Mach number range of 0.5 to 0.97 shows the following:
The longitudinal motions were well damped over the entire Mach
number range tested except for residual oscillations resulting prin-
cipally from engine gyroscopic coupling with the lateral oscilla-
tory mode. This engine gyroscopic coupling results in motion on
both the longitudinal and lateral oscillatory modes for either a
longitudinal or lateral disturbance. The lateral oscillatory mode
exhibits moderately good damping except for nonlinear damping
characteristics above a Mach number of 0.80 where the motion is
well damped at large amplitudes and poorly damped for double
amplitude sideslip angles less than 2°. Small inadvertent aileron
control motions often produced apparently undamped small
amplitude oscillations in the Mach number range above 0.80.54
Another memorandum expanded on this, describing lateral-stability

issues such as roll coupling, pitching characteristics aggravated by directional
233
Probing the Sky
divergence and aileron overbalance, an abrupt wing-dropping tendency, wing

heaviness, and rudder flutter. Direction divergence, for example, meant the X-5
had suddenly entered a sideslip of 25°, resulting in a spin. The memorandum
also noted, “The problem of aileron overbalance occurred less frequently but
was no less disconcerting to the pilot because the stick would jerk from side
to side unless restrained.”55
Perhaps the most damning criticism of the airplane was reserved for the
section “Pilots’ Impressions,” which made clear the demands of flying the X-5
(emphasis added):
The X-5 airplane at 58.7° sweepback is considered to have the least

desirable lateral stability and control characteristics of any of the
airplanes tested, including straight-wing, swept-wing, semitailless,
and delta-wing configurations. One pilot, while checking out in
the X-5 airplane, discontinued a speed run at M = 0.85 and an
altitude of 35,000 feet because he strongly doubted his ability to keep
the airplane right side up.56
The outstanding deficiency of the X-5 airplane is the lateral-
directional oscillation or “Dutch roll” caused by the high posi-
tive dihedral effect. This oscillation is annoying but tolerable for
research flying over the entire speed range at 40,000 feet, except
over the range of M = 0.86 to M = 0.88 where the residual,
small amplitude, virtually undamped oscillation is most notice-
able. The dihedral effect decreases with a decrease of altitude but
never reaches a satisfactory value. The airplane exhibits positive
lateral stability during sideslip maneuvers and requires large aile-
ron deflections for small rudder deflections; however, it is impos-
sible to maintain a steady sideslip without rolling oscillations.
Normal turning maneuvers tend to be jerky with abrupt increases
and decreases of bank angle, apparently caused by small yawing
motions and angle-of-attack changes. In straight and level flight,
lateral-directional oscillations can be initiated by control motions,
power changes, or turbulent air.
The aileron effectiveness is low at all Mach numbers and,
except for the adverse dihedral effects in some conditions, the roll-
ing characteristics are normal with rolling velocities proportional
to aileron deflection and increase as Mach number increases.
The rolling characteristics improve with decrease of altitude, but
maximum rolling velocity is limited because of the excessive force
necessary to obtain large aileron deflections. Near the 1g stall there
is little or no lateral control and nearly zero stick force.57
234
An Assessment
The X-5 left a mixed legacy. Like its progenitor, the Messerschmitt P.1101,
the aircraft was flawed, with poor stability and vicious stall/spin behavior that
cost the life of one of its pilots. At the same time, while the wing-sweeping
mechanism used on the X-5 was impractical, the aircraft had proven the basic
fundamental value of in-flight variable wing sweeping (in contrast to the X-4,
whose stability problems had proven the semi-tailless configuration impracti-
cal in the pre-fly-by-wire era). Additionally, because it could vary its sweep,
researchers gained a wide range of data on wing-sweep at various angles; NACA
test pilot Scott Crossfield memorably recalled that having the X-5 was like
“having a whole stable of swept-wing airplanes in one.”58 Absent the X-5,
collecting such a quantity and variety of data would have required multiple
aircraft with different wings.
Though the X-5, despite its stability problems, had demonstrated that the
concept of variable-sweep wings was workable, manufacturers were slow to
embrace the idea of variable-sweep wings, in part because of traditionalist
design tendencies, but in large measure because the technology was still imma-
ture and not yet ready for practical application. The mechanism needed for
the complex wing movements added weight and complexity, and the benefits
did not justify this burden.
A lone prototype U.S. fighter was built in the 1950s using variable-sweep
wings of a configuration similar to that of the X-5. This was the Grumman
XF10F-1 Jaguar, which began flight testing in early 1952. A “shoulder wing”
design like the X-5, it, too, used a single pivot point and shifted the wing roots
fore and aft, though coupling this with an imaginative delta-wing “flying tail”
to help control trim changes. But again like the X-5, it was impractical for
routine use. Worse, the XF10F-1 was in any case a thoroughly bad design,
plagued by miserable flying qualities and an unreliable engine. These deficien-
cies, coupled with changes in aircraft carrier design (including the angled deck
and mirror landing system and the steam catapult) and better aircraft exempli-
fied by the forthcoming Vought XF8U-1 Crusader, resulted in its termination.
The cancellation must have come as a relief to company test pilot Corwin H.
“Corky” Meyer, who stated revealingly afterward, “I had never attended a
test pilots’ school, but for me the XF10F-1 flight test program provided the
complete curriculum.”59
The NACA, and later NASA, played the leading role in transforming
the variable-sweep wing concept into a viable configuration.60 The change
came about as the result of a British proposal by famed Vickers corporation
designer Barnes Wallis for a supersonic airliner with variable-sweep wings
called the “Swallow.” Wallis’s imaginative study was largely impractical but did
235
Probing the Sky
incorporate a novel form of pivot location that ultimately hastened the develop-
ment of practical, operational variable-sweep aircraft. Vickers requested that
NASA review the design, and at the behest of John Stack and project engineer
Charles Donlan, four different Swallow configurations were eventually tested.
Configurations I through III had better stability margins than the X-1 model
had shown but were unacceptable. Configuration IV incorporated a different
concept for the variable-sweep wing. Unlike on the X-5 and the XF10F-1, the
pivot point was outboard of the fuselage, at the extremities of a small, fixed
forward wing. The outboard pivot, combined with the geometry of the fixed
forward wing, favorably changed the way the aerodynamic load shifted as the
wing sweep changed. The means of building a variable-sweep aircraft without
the complications of the X-5 or XF10F-1 wing designs was now at hand.61
The first U.S. variable-sweep aircraft to enter service was the F-111A, first
flown in 1964.62 It had originally been designed as a strike aircraft for the U.S.
Air Force and a fleet air defense interceptor for the Navy. (The Navy version
was subsequently canceled.) It was followed by the Navy’s F-14 fleet air defense
fighter and the Air Force B-1B strategic bomber. The NATO and Saudi air
forces operated the Panavia Tornado strike aircraft. The Soviets also operated
and exported a number of variable-sweep aircraft, including the MiG-23/27
fighter, the Su-17/22 and Su-24 strike aircraft, and the Tu-22M and Tu-160
strategic bombers. The original Boeing supersonic transport (SST) design also
featured variable-sweep wings, though fixed-sweep wings replaced these during
the design process. (The American SST project ended in 1971 when Congress
withdrew the funding.)
But variable-sweep wings subsequently proved to be of largely transitory
value, however, for they added cost and weight to a design and complicated the
designer’s challenge in the era of stealth. Today, while “legacy” variable-sweep
aircraft such as the B-1B and Tu-160 remain in service, few designers look to
them as viable options for future aircraft. As Richard P. Hallion noted in 2010:
On the whole, the variable-geometry wing has not enjoyed the

kind of widespread success that its adherent hoped. While it may
be expected that, from time to time, variable-sweep aircraft will
be designed and flown for particular purposes, overall the fixed
conventional planform, outfitted with all manner of flaps and
slats and blowing, sucking, and perhaps even warping technology,
continues to prevail.63
236
Endnotes
1. Opening quote from Crossfield and Blair, Always Another Dawn,
p. 155.
2. The best survey of the advent and significance of variable sweep
is Robert L. Perry’s Innovation and Military Requirements: A
Comparative Study, RAND Report RM-5182PR (Santa Monica,
CA: The RAND Corporation, 1967). See also Richard P. Hallion,
“Sweep and Swing: Reshaping the Wing for the Jet and Rocket Age,”
in Richard P. Hallion, ed., NASA’s Contributions to Aeronautics, vol.
1: Aerodynamics, Structures, Propulsion, Controls, NASA SP-2010-
570-Vol 1 (Washington, DC: NASA, 2010), pp. 57–64; and James
R. Hanson, The Bird Is on the Wing: Aerodynamics and the Progress of
the American Airplane (College Station, TX: Texas A&M University
Press, 2004), p. 123.
3. NACA-NASA interest in variable-sweep wings is documented
in NASA Langley Research Center, “Summary of NACA/NASA
Variable-Sweep Research and Development Leading to the F-111
(TFX),” NASA LRC Working Paper LWP-285, 1966; and Edward
C. Polhamus and Thomas A. Toll, “Research Related to Variable
Sweep Aircraft Development,” NASA TM 83121, 1981, pp. 5–6.
For the earliest example of oblique-wing wind tunnel testing, see
John P. Campbell and Hubert M. Drake, “Investigation of Stability
and Control Characteristics of an Airplane Model with Skewed
Wing in the Langley Free-Flight Laboratory,” NACA TN No. 1208,
May 1947.
4. Charles J. Donlan and William C. Sleeman, “Low-Speed Wind-
Tunnel Investigation of the Longitudinal Stability Characteristics of
a Model Equipped with a Variable-Sweep Wing,” RM L9B18, May
23, 1949.
5. Hanson, The Bird Is on the Wing, p. 123; Donlan and Sleeman,
“Low-Speed Wind-Tunnel Investigation,” pp. 1, 3, 8; Jay Miller, The
X-Planes: X-1 to X-45 (Hinckley, U.K.: Midland Publishing, 2001),
pp. 87–88.
6. Hanson, The Bird Is on the Wing, p. 124; Miller, The X-Planes: X-1 to
X-45; Bell Aircraft Corporation, “Bell X-5 Ready For Flight Tests,”
for release June 14, 1951.
7. John T. Rogers and Angel H. Dunn, “Preliminary Results of
Horizontal-Tail Load Measurements of the Bell X-5 Research
Airplane,” NACA RM L52G14 (August 15, 1952), p. 4; Donald
R. Bellman, “Lift and Drag Characteristics of the Bell X-5 Research
237
Probing the Sky
Airplane at 59° Sweepback for Mach Numbers from 0.60 to 1.03,”

NACA RM L53A09c (February, 17, 1953), p. 7. It should be noted
that reports and documents frequently give X-5 wing-sweep angles
of 59° or 60°. This was apparently rounding up, rather than an
actual modification to the wings.
8. Bell Aircraft Corporation, “Bell X-5 Ready for Flight Tests,”
press release.
9. Data from Jack Nugent, “Lift and Drag of the Bell X-5 Research
Airplane in the 45° Sweptback Configuration at Transonic Speeds,”
NACA RM H56E02 (July 11, 1956), Table I, pp. 10–11.
10. The Air Force role in developing the X-5 is minutely and defini-
tively detailed in Warren E. Green’s The Bell X-5 Research Airplane
(Wright-Patterson AFB, OH: Wright Air Development Center,
March 1954).
11. Miller, The X-Planes: X-1 to X-45, p. 89.
12. See Perry’s Innovation and Military Requirements: A Comparative Study.
13. Memorandum for Chief of Research, “Progress Report of
Engineering and Instrumentation for the X-5 Airplane During the
Period August 25 to September 7, 1951.”
14. Thomas W. Finch and Donald W. Biggs, “Preliminary Results
of Stability and Control Investigation of the Bell X-5 Research
Airplane,” NACA RM L52K18b (February 11, 1953), p. 4.
Period December 1 to December 14, 1951.”
Period December 15, 1951 to December 28, 1951.”
Period December 29, 1951 to January 11, 1952.”
18. Hallion and Gorn, On the Frontier, p. 402 (quoting flight records).
19. Rogers and Dunn, “Preliminary Results of Horizontal-Tail Load
Measurements of the Bell X-5 Research Airplane,” pp. 8–9.
20. Crossfield and Blair, Always Another Dawn, p. 155.
21. Ibid.
22. Hallion and Gorn, On the Frontier, pp. 47–48.
23. Memorandum for Chief of Research, “Progress Report for the X-5
Research Airplane for the Period May 3 to May 16, 1952.”
24. Memorandum for Chief of Research, “Progress Report for the X-5
Research Airplane for the Period July 12 to July 25, 1952”; James
238
A. Martin, “Longitudinal Flight Characteristics of the Bell X-5

Research Airplane at 59° Sweepback with Modified Wing Root,”
NACA RM L53E28 (August, 10, 1953), p. 1.
25. Memorandum for Chief of Research, “Progress Report for the X-5-1
Research Airplane for the Period July 26 to August 8, 1952.”
26. Memorandum for Chief of Research, “Progress Report for the X-5-1
Research Airplane for the Period October 1 to November 1, 1952.”
27. Ibid. At the time, 7.33 g’s was the Air Force loads design standard.
28. Hallion and Gorn, On the Frontier, p. 403 (flight records).
29. De Elroy Beeler, Acting Chief, NACA High Speed Flight Research
Station, to NACA Liaison Officer, Wright Patterson Air Force Base,
“Preliminary Investigation of Handling the X-5 Airplane Below
Minimum Drag in Close Proximity with Other Aircraft,” March
5, 1953, and Joseph A. Walker, NACA HSFRS, “Investigation of
Flight at Speeds Below Minimum Drag, Flight No. 60,” January 27,
1953, p. 1, DFRC Archives. These tests were not an investigation
specifically into wingtip vortices. Later, pilots and engineers would
realize that the wingtip vortices from a jumbo jet could cause a
smaller airplane to lose control and crash during a landing approach.
This led to increases in separation between aircraft on final approach.
30. Joseph A. Walker, NACA HSFRS, “Investigation of Flight at Speeds
Below Minimum Drag, Flight No. 60,” January 27, 1953, p. 1,
DFRC Archives.
Below Minimum Drag, Flight No. 61,” January 29, 1953, p. 1,
DFRC Archives.
32. Ibid.
Below Minimum Drag, Flight No. 63,” February 6, 1953, p. 1,
DFRC Archives. Aerial refueling became a standard practice in both
the Air Force and Navy. In particular, aerial refueling had become
critical to Strategic Air Command operations within a few years of
the X-5 tests. Aerial refueling extended the range of bombers and
allowed airborne alert missions. Subsequently, tactical aircraft and
even transport aircraft were outfitted for aerial refueling.
34. Joan M. Childs, “Flight Measurements of the Stability
Characteristics of the Bell X-5 Research Airplane in Sideslips at 59°
Sweepback,” NACA RM L52K13b (February 11, 1953), pp. 4–5.
35. Donald R. Bellman, “Lift and Drag Characteristics of the Bell X-5
Research Airplane at 59° Sweepback for Mach Numbers from 0.60
to 1.03,” p. 8.
239
Probing the Sky
36. Donald W. Briggs, “Flight Determination of the Buffeting

Characteristics of the Bell X-5 Research Airplane at 58.7°
Sweepback,” NACA RM L54C17 (May 24, 1954), pp. 1, 10–11.
37. Martin, “Longitudinal Flight Characteristics of the Bell X-5
Research Airplane at 59° Sweepback with Modified Wing Root,”
pp. 6–7.
38. Miller, The X-Planes: X-1 to X-45, p. 91; Hallion and Gorn, On the
Frontier, p. 401 (flight records).
39. USAF, “Report of Major Aircraft Accident—X-5, 50-1839A,
Section 0—Description of Accident” (Edwards AFB, CA: AFFTC,
October 27, 1953), first page. This section of the accident report has
no page numbers. It is typed text on a blank page.
40. Ibid.
41. Ibid., first and second pages.
42. Ibid., second page.
43. Ibid., eighth page.
44. Ibid., ninth page.
45. Ibid., “Accident Board Action Findings, Recommendations,” no
page number.
46. NASA, Fact Sheet “X-5,” http://www.nasa.gov/centers/dryden/news/
FactSheets/FS-081-DFRC_prt.htm, modified July 3, 2013.
47. Crossfield and Blair, Always Another Dawn, pp. 155–156.
48. USAF, “Report of Major Aircraft Accident—X-5, 50-1839A,”
October 27, 1953, “Remarks Medical Safety Division,” no page
number.
49. Hallion and Gorn, On the Frontier, pp. 402–404 (flight records).
50. Ibid., p. 404. See also Memorandum for Research Aircraft Project
Leader, “Progress Report for the X-5-1 Research Airplane for the
Period July 1 to July 31, 1955.”
51. X-5 Fact Sheet, http://www.nationalmuseum.af.mil/factsheets/factsheet.
asp?id=630 (accessed October 5, 2010).
52. Richard D. Banner, Robert D. Reed, and William L. Marcy, “Wing-
Load Measurements of the Bell X-5 Research Airplanes at a Sweep
Angle of 58.7°,” NACA RM H55A11 (April 5, 1955), p. 8.
53. Robert D. Reed, “Flight Measurements of Horizontal-Tail Loads on
the Bell X-5 Research Airplane at a Sweep Angle of 58.7°,” NACA
RM H55E20a (April 4, 1955), pp. 6–7.
54. Edward N. Videan, “Flight Measurements of the Dynamic Lateral
and Longitudinal Stability of the Bell X-5 Research Airplane at 58.7°
Sweepback,” NACA RM H55H10 (October 6, 1955), p. 1.
240
55. Thomas W. Finch and Joseph A. Walker, “Flight Determination of

the Lateral Handling Qualities of the Bell X-5 Research Airplane
at 58.7° Sweepback,” NACA RM H56C29 (May 31, 1956), pp.
12–13.
56. Ibid., p. 12.
57. Ibid., p. 13.
58. Crossfield and Blair, Always Another Dawn, p. 155.
59. Corwin H. Meyer, Corky Meyer’s Flight Journal: A Test Pilot’s Tales
of Dodging Disasters—Just in Time (North Branch, MN: Specialty
Press, 2006), p. 216.
60. The NACA, which existed as an agency from 1915 to 1958, gave
way to the National Aeronautics and Space Administration (which
had a broader mandate encompassing both aeronautics and astro-
nautics) in the reorganization of Federal aerospace taking place
after the shock of Sputnik. See Roger D. Launius, “An Unintended
Consequence of the IGY: Eisenhower, Sputnik, and the Founding
of NASA,” AIAA Paper No. 2008-860, 2008; and T. Keith Glennan
(edited by J.D. Hunley and in association with Roger D. Launius),
The Birth of NASA: The Diary of T. Keith Glennan, NASA SP-4105
(Washington, DC: NASA, 1993).
61. Polhamus and Toll, “Research Related to Variable Sweep Aircraft
Development,” pp. 9–12.
62. Robert W. Kress, “Variable-Sweep Wing Design,” AIAA Paper No.
83-1051 (1983) is an excellent survey of this and other efforts.
63. See the previously cited Hallion, “Sweep and Swing: Reshaping the
Wing for the Jet and Rocket Age,” p. 64.
241
The XF-92A is shown parked on the lakebed soon after final assembly. The aircraft spent most
of its research career in a bare-metal finish. The fuselage was wider at its midpoint than at the
nose or tail, which was long the standard design used in high-speed aircraft. (USAF)
242
CHAPTER 7
Progenitor of the Delta:
The Convair XF-92A
The pilot never really flew that airplane, he corralled it.

The Consolidated-Vultee (Convair) XF-92A was not built as a research air-

plane but as a flying mockup for a proposed supersonic interceptor, hence
its prototype fighter designation. It was not, however, intended as a combat
aircraft, and thus it occupies a similar place in research aircraft development
as the Bell XP-59A, the first American turbojet, which itself had a fighter des-
ignation while actually functioning as a technology demonstrator. Its design
evolution shows the ambitious plans of the immediate post–World War II
period, the unknowns of supersonic flight, and the impractical concepts pro-
posed to overcome them. From this environment came the birth of the semi-
tailless delta-wing aircraft. This configuration would eventually be used by
nations around the world in fighters, strike aircraft, reconnaissance airplanes,
and bombers, as well as in a pair of supersonic airliners—but it was proven
first by the Convair XF-92A.
The Birth of the XF-92A

The war against Imperial Japan had barely ended in the first week of September
1945 when the Army Air Forces announced a design competition for three
different fighters: the first was for an all-weather fighter capable of attack-
ing enemy aircraft at night or in bad weather; the second was for a long-
range escort fighter to protect bombers; and the third was for a short-range
point-defense supersonic interceptor, intended as a last-ditch defense against
enemy bombers. This third project became the XP-92. Consolidated-Vultee
Aircraft Corporation (CVAC, subsequently Convair) was among potential
contractors the Army approached on the three projects. In response, the firm
assembled a small design team at their Downey, CA, plant. The team consisted
243
Probing the Sky
of Jack Irvine, Frank W. Davis, Adolph Burstein (chief of design), Thomas

Hemphill, and Ralph H. Shick, chief of aerodynamics.2
Out of this team came the Convair Model 7002, later known as the XF-92A.
The first challenge the four men faced was the point-defense interceptor’s pro-
pulsion system. Existing jet engines lacked the thrust necessary for reaching
supersonic speeds, and rocket engines lacked sufficient burn time to take off,
climb to altitude, engage an enemy bomber, and land. As a result, engineers
came up with a hybrid propulsion system that used a ramjet combined with
an auxiliary jet engine, similar to that employed a decade later on the French
Mach 2+ Nord 1500-02 Griffon II experimental fighter. Several ducted rocket
motors were mounted within the XP-92’s ramjet both to provide thrust and
to act as igniters and flame holders. The auxiliary jet engine was fitted in the
aft fuselage for the return to base and landing.
The initial XP-92 airframe design featured a cylindrical fuselage, large
swept wings, and a V-tail. The wing sweep was 35° on the leading edge. The
Consolidated-Vultee design was submitted to the Army Air Forces, where
it was selected for development in May 1946. The first step in refining the
design involved wind tunnel testing. These tests showed a number of prob-
lems. The wing suffered from tip stall at angles of attack as low as 5°, as well
as poor lateral control characteristics. The design team set to work modifying
the XP-92 configuration.
Replacing the V-tail with a conventional design resolved some of the sta-
bility problems, but the wings required more work. The Convair design team
looked at wing slots, fences, and gloves as possibilities. The solution proved
far simpler. The wing’s trailing edge was changed to a straight line, making the
wing a 45° triangle rather than a conventional swept wing. This was the first
time the shape had been used on a jet aircraft. The design was further refined
to a pure 60° delta wing in a series of July 1946 wind tunnel tests.3
The adaptation of a delta wing for the XP-92 thus was a byproduct of swept-
wing refinement, not a choice from the outset based on theoretical work or
emulation, despite NACA research aerodynamicist Robert T. Jones’s having
enunciated the thin, sharply swept delta more than a year previously.4 More
importantly, in contrast to popular myth, the XP-92 (and the XF-92A and, for
that matter, all subsequent Convair delta aircraft) owed nothing to the work
of German aerodynamicist Alexander Lippisch, who had designed the Me 163
and who had conceived a very thick-wing 60° delta glider, the Lippisch DM-1,
intended as a low-speed test bed for a proposed ramjet-powered delta-wing
fighter, the P.13. The Convair team rejected his concept of a delta wing, which
envisioned a very thick wing with a very thick vertical fin (in contrast, the delta
wing that they chose had—for its time—a very thin 6.5-percent thickness-
chord ratio). Lippisch’s DM-1 had more than twice the thickness-chord ratio
244
Progenitor of the Delta: The Convair XF-92A
Another early photo of the XF-92A. The canopy was considered problematic by pilots because
the framework hampered visibility. The sensitivity of the flight-control system, which required a
light touch by pilots, also presented difficulties. (USAF)
of the XP-92, and its vertical fin was so large that the pilot sat within it, the
clear cockpit windscreen being formed by its leading edge. More significantly,
when the DM-1 glider was tested in Langley’s Full-Scale Tunnel in 1946, the
results were “very disappointing; the lift coefficient was low, the drag was high,
the directional stability was unsatisfactory, and the craft was considered unsafe
for flight tests.”5 The Convair thin-wing approach was altogether more satisfac-
tory, and for it, Convair’s XP-92 design team deserves great—and sole—credit.
Model 7002—From Flying Mockup

to Research Airplane
The final design of the XP-92 was striking, with the 60° delta wing, a triangu-
lar vertical fin, and a tubular fuselage that extended well behind the fin.6 The
cockpit was in a bullet-shaped nose mounted within the circular ramjet-intake
duct. In the event of an emergency, the forward fuselage could be jettisoned
as an escape capsule in a manner reminiscent of the D-558-1 and -2. After
separation, a parachute would slow the capsule, and the pilot would then bail
245
Probing the Sky
A research aircraft in wolf’s clothing. During the Air Force test program, the XF-92A was painted
in bogus Soviet markings as a “MiG 23” for the movie Jet Pilot. The aircraft apparently did not
appear in the final film. (USAF)
out using his own parachute. Ironically, the airplane, which resembled a craft
straight out of the world of Buck Rogers, was armed with conventional weap-
ons—four 20-millimeter cannons.
The new design retained the rocket-augmented ramjet engine. Because a
ramjet must be in flight before it can produce thrust, a set of eight 1,500-pound-
thrust rockets, burning liquid oxygen and alcohol/water, would be used for the
takeoff. The XP-92 did not make use of conventional landing gear for takeoff
but rather was mounted on an eight-wheel cart containing the liquid-oxygen
and alcohol/water tanks necessary for the takeoff run. Once the plane lifted
off, a pair of large wing drop tanks supplied the fuel to boost the XP-92 to a
top speed of Mach 1.65.
At that point, 16 50-pound ducted rockets fueled by liquid oxygen and
gasoline and mounted inside the ramjet ignited. The rockets started the ramjet
and stabilized the flame within the ramjet combustion chamber. The ramjet
increased the airplane’s speed to Mach 1.75 at 50,000 feet, speed and alti-
tude that could be maintained for 5.4 minutes. The XP-92 was designed as a
last-ditch/point-defense interceptor to destroy high-speed bombers and cruise
missiles before they could reach their target. Once the mission was complete,
246
the XP-92 would fly back using the auxiliary jet to land on a runway with its
tricycle landing gear.
A full-scale metal mockup of the XP-92 was built, but Army Air Forces
personnel had already begun to have doubts about its design. The rocket-
augmented ramjet engine was seen as too complex for use in a near-term
project, and the sled launch system seriously limited its utility. Other doubts
were raised about Convair’s proposed hydraulic-powered, nonreversible control
system, and weight increases had reduced performance. Even if these issues
could be dealt with, the Air Force was in the midst of postwar budget cuts,
and the number of personnel and aircraft were at a low point. No funding was
available for so advanced and risky a concept as the XP-92. The interceptor
program ended in June 1948.
But construction of the subscale demonstrator went ahead, for it had value
even if completely separated from the futuristic interceptor program that had
spawned it. The research aircraft that eventually became the XF-92A originated
in meetings between Air Force and Convair personnel in September 1946.
Building a subscale transonic test vehicle would provide data on the behavior
of the delta wing at low and high speeds. Delta wings offered a number of
advantages, including large wing area, a thin airfoil cross section, and a low
aspect ratio. In addition, the wings were light in weight relative to their surface
area while being very strong, an added benefit in any aircraft design.
On November 7, 1946, the Army Air Forces contract with Convair was
amended. Two XP-92 prototypes were to be built. In addition, a single Model
7002 would be constructed to test the wing and tail configuration. The Model
7002 that emerged had little in common with the exotic XP-92. The cylindrical
fuselage had a small diameter at the nose and tail, but the center section was
much larger in diameter to accommodate its centrifugal-flow jet engine. The
Model 7002’s wing was the same size, airfoil, and configuration as that planned
for the XP-92, though its tail fin was larger than the planned interceptor’s. To
keep development costs down, existing components and systems were used in
building the aircraft. The Model 7002 was powered by an Allison J33 jet engine
from a Lockheed P-80 and had the canopy and ejection seat from Convair’s
YP-81, the nose gear of a Bell P-63, the brakes of a Lockheed P-80, the pilot
stick and rudder pedals from a Vultee BT-13, and the main landing gear of a
North American FJ-1.
But in other ways, the Model 7002 was as exotic as its ramjet-powered
predecessor. The Model 7002 had the 100-percent hydraulically powered, non-
reversible flight controls that were planned for the XP-92. Because the Model
7002 was a semi-tailless configuration—lacking a conventional horizontal
stabilizer—the plane employed elevons combining pitch and roll control, simi-
lar to the X-4 then under development. Convair closed its Downey facility in
247
Probing the Sky
the summer of 1947, as a result of tightening military budgets following the

end of World War II, and transferred the unfinished Model 7002 airframe to
the company’s San Diego facility between June and July. The final assembly
of the Model 7002 was completed in San Diego on October 31, 1947. In
November, Convair sent the aircraft by Navy cargo ship to San Francisco for
wind tunnel testing in the full-size tunnel at the Ames Aeronautical Laboratory.
This testing was completed in early January 1948, and the aircraft returned to
San Diego for final testing and installation of its J33-A-21 turbojet engine. The
completed Model 7002, now designated the XF-92A, was then shipped from
the San Diego harbor aboard the Navy landing craft LST-827 and arrived at
the harbor in Los Angeles on March 26, 1948. The aircraft was then loaded
on a truck and shipped over the San Gabriel Mountains to Muroc Air Force
Base, where it arrived on March 31, 1948.7
Phase I Flight Tests

Convair’s chief test pilot, Ellis D. “Sam” Shannon, started tests of the XF-92A
on May 25, 1948, by making high-speed taxi runs across the Rogers lakebed.
The first “hop” came on June 9, 1948, during the fifth taxi test. The aircraft
reached 10 or 15 feet above the
lakebed, and he flew straight for
about 2 miles before touching
down again. Shannon found the
controls were very sensitive to
the smallest inputs. Touching
the stick and rudder pedals was
enough to create control prob-
lems. With the first taxi tests
and short hop behind them, the
team took the XF-92A back to
the hangar, where the plane was
fitted with a more powerful J33-
A-23 jet engine that had water
injection to increase thrust.
The aircraft resumed taxi tests
in September. These were suc-
cessfully completed, and with
A delta-wing aircraft model attached to a rocket
the aircraft systems checked booster for a high-speed test. Such tests were used
out, Shannon could make the for preliminary research on new wing configura-
first flight. tions. (NASA)
248
Shannon made his first actual flight in the XF-92A on September 18, 1948,
and it lasted 18 minutes. He found that the hydraulic control system was
overly sensitive and had a lag in its control movements. Tiny movements of the
stick and rudder pedals caused the XF-92A to wallow in all three axes, he told
engineers following the flight. During a second flight, on September 29, he
found the same conditions. After another flight, on October 13, the XF-92A
was grounded through November 1948 while the ground crew harmonized
its flight controls.
Testing resumed once sufficient modifications had been made, and by the
end of 1948, Shannon had made a total of 10 flights. A second Convair test
pilot, Bill Martin, joined the Phase I test effort, making his first taxi test on
December 21, 1948. Martin’s first flight was on February 9, 1949. The con-
tractor flights initially focused on general handling qualities and performance.
An unusual characteristic of the XF-92A that the pilots quickly identified
was its remarkable low-speed stability. Shannon discovered to his surprise that
even with the indicated airspeed below 90 mph and the angle of attack as high
as 30°, the aircraft would not stall. This was due to the delta wing’s lack of an
abrupt break in its lift curve at the stall. Instead, the XF-92A exhibited a stabi-
lized high sink rate due to the increase in drag. This loss of altitude continued
despite the use of full throttle.
Another unusual characteristic was that the aircraft would not spin. Tests
of a model in the Langley spin tunnel had indicated that if the Model 7002’s
center of gravity was between 24 and 29 percent of the mean aerodynamic
chord, a spin was impossible. By August 1949, the Phase I tests moved to
transonic flights. To reach these speeds, the pilots had to put the underpowered
XF-92A into steep dives. In the end, neither pilot exceeded Mach 0.925, as
the nonafterburning J33 could not power the XF-92A through the speed of
sound. Phase I tests ended on August 26, 1949, after a total of 47 flights by
Shannon and Martin.
With the contractor flights completed, the Air Force now began Phase II
flights. The two pilots assigned to the tests were Captain Charles E. Yeager
and Major Frank Everest. Phase II flights began on October 13, 1949, and
concluded on December 28. The two pilots completed a total of 25 flights and
were highly enthusiastic about the little delta. The Phase II flights also included
very steep dives attempting to reach speeds slightly above Mach 1. Even assisted
by a dive, however, the J33-A-23 engine lacked sufficient thrust to reach Mach
1. Despite this, Everest was impressed with the aircraft. He later noted, “We
found that the delta planform handled very well during the transition from
subsonic to supersonic speed, in comparison with the straight-wing F-94 and
the swept-wing F-86, which encountered severe buffet and loss of control
effectiveness in this speed range.”8
249
Probing the Sky
While Phase II flights were underway, the Air Force issued a contract to
Convair to replace the J33-A-23 engine with an afterburner-equipped J33-
A-29 engine and lengthen the aft fuselage. This would increase the available
thrust with afterburner to 7,500 pounds, compared to the 4,250 pounds with
water injection available with the J33-A-23. Convair engineers estimated that
the XF-92A’s level speed would be increased to Mach 0.98, higher than the
maximum dive speed with the J33-A-23 engine. Moreover, the maximum dive
speed would be Mach 1.2. To improve visibility for the photo-tracking cameras,
the XF-92A’s bare metal exterior was also painted white. Original plans called
for Yeager to make a ferry flight of the XF-92A from Edwards to Convair’s San
Diego facility. Yeager took off on May 12, 1950, but suffered an engine failure
requiring an immediate emergency landing on the lakebed. The aircraft was
not damaged but had to be taken south by truck. The aircraft spent the next
14 months in San Diego undergoing the engine change. The XF-92A did not
return to Edwards until July 1951, and Yeager made the first flight with the
new engine on July 20.
Yeager and Everest now embarked on a second series of flights to evaluate
the XF-92A’s performance with the new engine, which proved to be a disap-
pointment. They found that the afterburner would flame out above 38,000
feet, and the J33-A-29’s increased thrust failed to raise the aircraft’s Mach
The XF-92A in flight over a cloud deck late in the Air Force test flight program. The airplane had
been painted white, as had other X-planes, to allow the aircraft to be photo-tracked against the
dark blue desert sky. The dark orange or dark red finishes originally used on the X-1 and the
D-558-1 Skystreak were difficult to spot against the dark sky. (USAF)
250
number. In fact, the XF-92A’s performance was not better than it had been
with the J33-A-23 engine. To compound matters, while the J33-A-23 engine
had few maintenance issues, the J33-A-29 was beset by reliability problems.
Completing the initial 18 flights of the power plant demonstration program
required six engine changes.
Problem areas included the fuel control system, the improper sequencing
of the afterburner, and the backward flow of the turbine cooling air. The last
issue resulted in molten metal in the aft compressor inlet, which triggered fire
warning lights in the afterburner and faulty turbine wheel seal clearances.9
Despite the unreliable engine, Yeager and Everest made 10 research flights
in the XF-92 by the end of the year. Peak altitude achieved with the aircraft was
41,443 feet on August 8, 1952, and the top speed was Mach 1.01 on February
2, 1953. The only serious failure occurred on September 14, 1951, when the
engine suffered a turbine failure shortly after takeoff, forcing an emergency
landing. The aircraft was repaired and test flights resumed. Another 11 flights
were made in 1952 before the Air Force flights ended.10
Initial Results
As they had with earlier contractor and Air Force test flights of research aircraft,
NACA personnel wrote research memorandums on the XF-92A’s handling and
flight qualities. The NACA had access to the data collected and conducted the
data reduction and analysis. From this, the XF-92A’s characteristics were deter-
mined. The first of these, on preliminary measurements of static longitudinal
stability and trim, covered the power plant demonstration flights following
the replacement of the original J33-A-23 engine with a J33-A-29 fitted with
an afterburner, as well as Air Force performance testing. The flights covered
Mach numbers up to 0.75 to 0.97, at altitudes from 11,000 to 40,000 feet.
The research memorandum noted:
The airplane had a stable variation of control angle with speed

up to some Mach number between 0.75 and 0.87. A nose-down
trim change extended from a Mach number of 0.87 to a Mach
number of about 0.93, at which point a nose-up trim change was
encountered.
The apparent longitudinal stability was constant up to a Mach
number of about 0.75. As the Mach number increased above
0.75 the apparent stability increased rapidly to approximately five
times the low-speed value at a Mach number of 0.96. The static
longitudinal stability increased threefold as the Mach number
251
Probing the Sky
increased from approximately 0.75 to 0.94; this result indicates

that most of the increase in apparent stability was caused by an
increase in static longitudinal stability.11
Air Force Phase II flights tested the XF-92A’s performance, and, as a result,
no maneuvers were made during Phase II flights specifically to obtain dynamic-
stability data. However, several random disturbances occurred that could be
analyzed to identify dynamic longitudinal-stability characteristics. The situ-
ation was complicated by the fact that under certain flight conditions, pilots
observed “undesirable lateral and longitudinal oscillations” during this stage of
the flights.12 NACA researchers thought these could indicate “the possibility
of cross coupling between the lateral and longitudinal modes of motion.”13
To obtain the stability data, two different techniques were used. The first was
simple: the period of the oscillations and time it took to damp out the aircraft’s
motions were measured. The other technique, involving the use of a Reeves
Electronic Analog Computer (REAC), represented a glimpse of the future.
The actual control deflections were used as the REAC input. A solution, in the
form of a time history, was produced for a transfer-function equation with a
particular set of transfer-function coefficients. The coefficients were varied on
each run until the output most nearly matched the actual flight record.
Though the REAC represented an advance in data processing, it was sig-
nificantly limited by the primitive computer technology of the early 1950s.
The effects of trim changes due to Mach number and altitude changes were
not included in the calculations. Agreement was considered good between
the period and static-stability measurements acquired using the two methods,
though there were differences in the time and cycles needed to damp and in
the damping factor. These differences were thought to be due to such factors
as the method of analysis, small control motions during the maneuver, or
nonlinear damping.
An example of the analysis and the information that could be determined
from it was a time history of an XF-92A dive from 38,000 feet to 34,000 feet
at Mach 0.94. The control deflections were small during the dive. The REAC
was used to analyze the dive from 2 to 16 seconds, using only the longitudinal
control inputs. The REAC calculations showed good agreement with actual
flight data. From this, the NACA researchers concluded that no serious cou-
pling was occurring between lateral and longitudinal modes, as the aircraft
underwent both lateral and longitudinal motion. (An earlier concern had been
raised over the possibility that the XF-92A’s lateral and longitudinal motions
were coupled, indicating major stability problems.)14
After Phase II flights were completed, several Air Force test pilots made
familiarization flights in the XF-92A. These included Colonel Albert Boyd,
252
Major Jack Ridley, Captain J.E. Wolfe, Captain Arthur “Kit” Murray, and
Lieutenant Colonel Richard L. Johnson. After one landing, Colonel Boyd,
the Edwards base commander at the time, was asked by Shannon, “Colonel,
didn’t you find the airplane a little sensitive laterally?” Boyd replied, “No, not
sensitive, but if I find the guy that didn’t show me where the lateral trim switch
is, I’ll shoot him. I had to hold five pounds pressure on the stick for the whole
flight, and my arm is about to fall off.”15
Following XF-92A tests with the J33-A-29 engine, the Air Force test team
noted that, “Since the XF-92A was designed purely as a research airplane and
never intended for operational use, the aircraft systems, maintenance, and
accessibility were not stressed in the design.”16 The Air Force report also stated
that “the ‘delta-wing’ configuration exhibits characteristics that could be used
advantageously in future Air Force tactical aircraft.” Given the advantages
shown by the delta wing, the report recommended improvements for a new
delta-wing aircraft:
• Install a higher rated power plant in conjunction with the design of
a more efficient inlet duct
• Improve the centering and breakout forces of the power
control system
• Improve the over-sensitivity of the longitudinal and lateral
control system
• Improve the directional stability characteristics and decrease the
excessive dihedral effect (the rolling moment of an aircraft due
to sideslip)
• Decrease the high side forces produced by the vertical tail being
located too near the airplane’s center of gravity
• Improve the poor low-speed flying characteristics during approach
and landing
• Modify the directional control system to eliminate the trim changes
caused by a change in temperature
• Improve directional control on the ground and cross-wind handling
characteristics by increasing the landing gear tread and the wheel size
• Improve general ground handling characteristics and reduce landing
ground roll by installation of an adequate brake system
• Improve forward visibility at high angles of attack
• Improve the restricted visibility resulting from the
“greenhouse” canopy
• Relocate the landing gear position relative to the center of gravity to
eliminate tail heaviness on the ground
• Redesign the aft fuselage to permit engine accessibility
for maintenance
253
Probing the Sky
• Redesign the landing gear control system for easier operation by

the pilot
• Install an improved hydraulic system17
Crossfield Flies the First Delta

The final three flights of the Air Force power plant demonstration program were
completed in early 1953. Crossfield and the NACA engineers then began a
separate series of research flights. The NACA personnel were eager to try out the
new wing configuration. They were also aware of the aircraft’s shortcomings,
in particular, those of the J33-series engines. One of the first things ground
personnel did was replace the J33-A-29 jet with an afterburner-equipped J33-
A-16 engine capable of producing a maximum thrust of 8,600 pounds. The
modification gave the XF-92A a lengthened tail cone enclosing the afterburner.
Crossfield was named NACA project pilot, and his introduction to the air-
craft was not encouraging. Crossfield had a “first flight jinx”; he always seemed
to have difficulties on his first flight of a new airplane, and the XF-92A “hop”
was no exception. Pete Everest was originally scheduled to make the final Air
Force flight before the XF-92A was turned over to the NACA. The XF-92A was
towed out onto the lakebed and pointed into the wind. But the wind shifted and
quickly grew too strong, so the final Air Force flight was canceled. Crossfield
and Everest were meeting with Williams at the time, and Williams suggested
that Crossfield taxi the airplane back to the NACA hangar at the main base and
lift it off the lakebed to get a feel for its handling. Crossfield agreed.
He climbed into the XF-92A, started the engine, closed the canopy, and
began a fast taxi toward the main base. The XF-92A was heavy with fuel and
wobbled into the air. By this time, the lakebed shoreline was fast approaching.
Crossfield pulled back on the stick to raise the nose and slow the airplane. The
decrease in speed was only slight, and he thought the XF-92A was going to
hit a sand dune and be destroyed. Crossfield then spotted a dirt road heading
into the desert and pressed the left brake to turn the aircraft toward it. The
brake seized up, then failed, but not before turning the aircraft toward the
road. When the XF-92A reached the road, the tires burst, but the airplane
continued straight ahead. After about a hundred yards, it came to a stop.
Crossfield triggered the fire extinguishers and jumped out. Crossfield and the
XF-92A were unharmed. The dirt road was promptly, if unofficially, dubbed
“Crossfield Pike.”18
The NACA XF-92A research effort involved 25 flights focused on two
areas. The first 15 flights would collect the stability and control characteristics
of the basic aircraft design. The subsequent 10 flights would then look at the
254
The XF-92A in the desert after Crossfield’s first high-speed taxi test. The dirt road was soon
named “Crossfield Pike.” The aircraft sustained only minor damage. (NASA)
effectiveness of wing fences positioned at 60 percent of the wing semispan.

Data on drag; static longitudinal, lateral, and directional stability and control;
buffeting; wing loads; and airfoil section characteristics would be obtained
both with and without the fences.19
Crossfield made his first XF-92A flight on April 9, 1953, for pilot checkout
and static longitudinal-stability data. He then flew the airplane three times
in April, including one flight that reached the highest altitude by an NACA
pilot in the airplane—41,364 feet. Following one flight in May, the research
program hit its stride. A total of eight flights were made between June 3 and
June 24 and included the fastest NACA XF-92A flight, which reached Mach
0.962. All these flights were made for longitudinal-stability and control data.
The procedure was to make sideslips, aileron rolls, and rudder pulses at altitudes
from 18,000 to 30,000 feet at indicated airspeeds of 160 to 420 mph.
In the case of the rudder pulses, Crossfield stabilized the aircraft at the
planned test speed and altitude and then made an abrupt pulse of the rudder
while the other control surfaces were held in a fixed position until the aircraft
returned to stabilized flight. Test maneuvers were done in both a clean con-
figuration and with landing gear down. The month’s final flight, on June 26,
was for low-speed stability and control data.20
255
Probing the Sky
After that, the aircraft was grounded to undergo the installation of the
wing fences. Wing fences had been installed on the D-558-2 as one of several
potential cures for pitch-up. The fence itself was simple, running from the
wing’s front edge back to the 60-percent chord. The leading and trailing edges
of the fence were fairing curves.21
The first two flights made with the fences installed took place on July 3.
The fences were then modified and reinstalled on the XF-92A, which was
again flown on July 22. The changes proved unsuccessful; one of the fences
buckled under the flight stresses. Following the incident, the aircraft remained
grounded for nearly a month. Crossfield took the airplane aloft on August
17, but the engine suffered a malfunction and forced the research flight to be
cut short.
In contrast to the long delays the Air Force endured during test flights, the
engine problem was repaired in 3 days and the XF-92A was back in the air on
August 20. Crossfield made two flights that day, both for longitudinal-stability
and control data with the modified fences.
Not until September 30 did the aircraft again go aloft. The two flights on
this date marked the start of flights to collect data on the airplane’s low-speed
lateral and directional control with the original wing fences. Two more flights
followed with the same research goals, on October 2 and October 5. The
fences were removed as the research program drew to a close. The earlier test
maneuvers were repeated to identify what effects the fences had on the aircraft’s
low-speed lateral and directional control.
A pair of flights was made on October 14, 1953. At the end of the second
mission, Crossfield touched down on the lakebed. As the XF-92A rolled out,
the nose landing gear collapsed. The aircraft ground-looped and came to a
stop on its nose and one wingtip. Once Crossfield was sure the aircraft would
not tip over, he climbed out of the airplane. Inspection revealed consider-
able damage to the nosewheel, forward fuselage, and right wingtip. Taken
together, the extent of the damage, the amount of time and money that would
have been needed to restore the XF-92A to flight status, and the fact that
the research program was nearly complete meant that the aircraft would be
grounded permanently.22
Even as the XF-92A project ended, the data generated during the NACA
research flights were being analyzed. There was quite a bit of information, and
because data reduction in that era took time, results were somewhat slow in
coming. But come they did. And since the XF-92A was the first delta-wing
aircraft from which any data had been collected in flight, what the Air Force
and NACA engineers and pilots produced was of considerable value. A 1955
research memorandum on the aircraft’s lateral-stability and control character-
istics noted:
256
The static lateral stability characteristics as measured in sideslips

appear satisfactory although there is a rapid increase in apparent
dihedral effect at the lower speeds. There is adequate rudder power
over the entire speed range tested.
The lateral control as measured in aileron rolls appears ade-
quate over the entire speed range tested although there is a con-
siderable reduction in aileron effectiveness at the lower speeds.
The dynamic lateral stability of the airplane was generally
in the unsatisfactory region when compared to U.S. Air Force
requirements for satisfactory values of reciprocal of cycles to damp
to half amplitude and ratio of roll angle to sideslip velocity.
Although the airplane appears to have satisfactory static lateral
stability and control characteristics, the pilots object to the over-
all lateral handling characteristics primarily because of the high
roll-to-sideslip ratios which probably resulted from the relatively
low static directional stability and relatively high effective dihe-
dral. These adverse characteristics were aggravated at low speeds
by high adverse yaw and rough air and at high speeds by high
airplane response to small control deflections. The apparent high
side force and poor hydraulic control system added to the objec-
tionable characteristics.23
The research memorandum also included comments on the control system

issue: “Part of the difficulties encountered on the XF-92A may be attrib-
uted to the poor hydraulic control system. Evaluation of ground calibration
has shown the control system to have high friction and breakout forces and
appreciable lag of surface-to-stick motion. These characteristics are particu-
larly objectionable at low speeds where large control deflections are required
to maneuver and also at high speeds where the airplane is sensitive to small
control displacements.”24
Though the bulk of the NACA flights had been made for longitudinal-
and lateral-stability and control data, the tests of the wing fences proved
successful. The research memorandum stated: “Installing wing fences on the
airplane improved the handling characteristics at the lower speeds, in the
pilots’ opinion, probably because of the increase in static directional stability
attributed to the fences.”25
The XF-92A was eventually patched up as a static display for air shows
and was later sent to the University of the South, at Sewanee, TN, where it
remained until 1969. It was returned to the U.S. Air Force Museum at Wright-
Patterson AFB, OH, and thoroughly restored; it is now on display, along with
the X-3, X-4, and X-5, in the museum’s Research and Development Gallery.26
257
Probing the Sky
In-flight photo of the XF-92A during one of its NACA flights. The only significant difference in fin-
ish was the X’s on the nose and tail, which were put on for photo reference. The right wing had
tufts attached to the upper surface to make visible changes in the airflow that occurred under
different flight conditions. (USAF)
An Assessment
The XF-92A occupies an honored place among the world’s pioneering air-
craft, but it was not an airplane that was popular with its pilots. An NACA
research memorandum noted: “All the pilots who have flown the XF-92A
(during the joint NACA-Air Force program and the NACA research program)
have reported that the airplane exhibited poor lateral handling characteris-
tics, particularly at low speed.”27 Crossfield later said of the airplane, “It was
under-powered, under-geared, under-braked, and overweight.” After the gear
collapsed on Crossfield’s last flight, he noted, “It was finished and no one shed
any tears.”28
At the same time, the potential that the XF-92A represented was clear. The
Air Force technical report noted, albeit in understated tones, “The ‘delta-wing’
configuration exhibits characteristics that could be used advantageously in
future tactical aircraft.”29 Yeager was impressed with the high roll rate of the
aircraft, which would be useful in a dogfight. He commented, “When I’m
258
doing a maximum rate of roll, it feels like

the seat of my pants are pushing down
into the seat while at the same time the
centrifugal force is pulling my head up
toward the canopy.”30 Though Crossfield
did not lament the airplane’s demise, his
Air Force counterparts found elements
of it to their liking. And the Air Force
itself saw the planform’s potential.
On another note, the shortcomings
of the XF-92A’s control system under-
scored that the building of a successful
supersonic aircraft was not merely an
issue of aerodynamics. A proper control
system had to provide handling quali-
ties that would allow the pilot to fly his
aircraft to its limits rather than struggle
against it. Some of the most skilled pilots
in America flew the XF-92A, and even
they found it a difficult airplane to con-
trol in spite of its desirable qualities. It is
clear in retrospect that control systems
evolved in the same way other aircraft
elements did in this new era of avia-
tion and that fits and starts were part of
the process.
Of course, the story of delta-wing
aircraft development did not end with
XF-92A. On October 24, 1953, just
10 days after the XF-92A’s final flight,
Convair’s YF-102 prototype took off The solution to performance problems that
from Edwards. The new design was based plagued the F-102 was a major redesign
of the aircraft. On the top is the original
on what Convair engineers had learned YF-102. Like the XF-92A, it had a fuselage
from the XF-92A. The new delta-wing that was widest at the midsection. This
aircraft featured a pointed nose to house produced excessive drag. The solution,
the radar system along with side air developed by Richard T. Whitcomb, was
inlets. The NACA was eager to test the the “area rule.” The fuselage was reduced
in width, forming a shape referred to as a
new aircraft and had begun negotiations
“Coke-bottle” or “wasp-waist.” The produc-
with the Air Force to receive an early pro- tion F-102 saw considerable improvement
totype. But like the XF-92A, the YF-102 in performance as a result of the modifica-
had its problems. The first prototype tion. (NASA)
259
Probing the Sky
Ground photo of the XF-92A during NACA research flights. The poor visibility from the canopy is
evident. So is the forward position of the main landing gear, which made the aircraft tail-heavy,
and a pole, visible under the aft fuselage, that was necessary to prevent the aircraft from resting
on its tail. (USAF)
The YF-102 prototype interceptor was developed using data acquired with the XF-92A. The
new interceptor was beset by performance problems and crashed only 9 days after its first
flight. (NASA)
260
crashed on November 2, only 2 weeks after its first flight, and the brief test
program was just long enough to demonstrate the design’s poor performance.
Test flights with the second YF-102A began on January 11, 1954, and
although Convair had expected it, they found that the plane could not exceed
Mach 0.98 and had a
maximum ceiling of only
48,000 feet, well below
the required performance
set by the Air Force.
The design underwent
extensive changes, which
included cambered wing
leading edges and new
wingtips. The wings were
also moved farther aft
on the fuselage, which
itself was lengthened by
11 feet. The vertical fin
was also moved aft, and,
most important, the
center fuselage was rede-
signed (thanks to NACA
research aerodynamicist
Richard Whitcomb) with
a “Coke-bottle” shape to
reduce drag. This last fea-
ture represented one of
Two NACA ground personnel with a prototype F-102. The
the great breakthroughs family resemblance between the XF-92A and the F-102 is
made during the early apparent. Both had a bulky fuselage that resulted in exces-
supersonic era.31 sive drag and poor performance. (NASA)
261
Probing the Sky
Endnotes
1. Opening quotation from Crossfield and Blair, Always Another Dawn,
p. 167.
2. For the XF-92A’s evolution and flight research program, see Richard
P. Hallion, “Convair’s Delta Alpha,” Air Enthusiast Quarterly 2 (June
1976): 177; Robert E. Bradley, “The Birth of the Delta Wing,”
American Aviation Historical Society (winter 2003): 243.
3. Hallion, “Convair’s Delta Alpha,” pp. 177–178; Bradley, “The Birth
of the Delta Wing,” pp. 245–247.
4. See Hallion, “Lippisch, Gluhareff, and Jones,” passim.
5. Letter, Major Howard C. Goodall, USAF, to Paul E. Garber, “DM-1
Glider Disposal,” November 28, 1949, in the Gluhareff Dart
accession file, Registrar’s Office, National Air and Space Museum,
Smithsonian Institution, Washington, DC.
6. For details, see Hallion, “Convair’s Delta Alpha,” passim.
7. Bradley, “The Birth of the Delta Wing,” pp. 254–258. The point-
defense role would eventually be filled by surface-to-air missiles
(SAMs), which provided quicker response and could be built,
deployed, and operated far more cheaply than aircraft.
8. Hallion, “Convair’s Delta Alpha,” p. 181.
9. “High-Speed Flight Research Station: Research Airplane Project
Panel 1952 Annual Report,” pp. 8–9.
10. Hallion, “Convair’s Delta Alpha,” pp. 181–182; Bradley, “The Birth
of the Delta Wing,” p. 259. The following chronology of reports
offers some idea of the engine problems faced by the XF-92-A:
10/3/1951: Installed in airframe.
10/4–8/1951 Ground runs—would not start on “normal,” had
to use emergency fuel supply.
10/10/1951 AF flight 10—afterburner would not light normally.
10/11/1951 Engine removed—turbine rotor #5-6-7 replaced.
11/28/1951–1/9/1952 Reinstalled—experienced fuel control
problems, changed controls several times/two small fuel fires
(no damage).
1/10/1952 Removed from aircraft—nozzle vanes cracked/dia-
phragm replaced.
6/4/1952 AF flight 11—two cracks found in aspirator.
6/5/1952 AF flight 12—molten metal found in aft impeller area.
6/9–11/1952 Ground runs—cracks in afterburner.
6/11/1952 Engine removed from aircraft—cooling air diffuser
overheated.
262
7/18/1952 Test cell run.

7/22/1952–8/5/1952 Installed in airframe/ground runs.
8/6–8/1952 AF flights 13–14
8/15/1952 AF flights 15–16
8/18/1952 Removed from aircraft—air baffles cracked—metal
found on aft impeller area.
11. Thomas R. Sisk and John M. Mooney, “Preliminary Measurements
of Static Longitudinal Stability and Trim for the XF-92A Delta-
Wing Research Airplane in Subsonic and Transonic Flight,” NACA
RM L53B06 (March 27, 1953), p. 5.
12. Euclid C. Holleman, John H. Evans, and William C. Triplett,
“Preliminary Flight Measurements of the Dynamic Longitudinal
Stability Characteristics of the Convair XF-92A Delta-Wing
Airplane,” NACA RM L53E14 (June 30, 1953), p. 1.
13. Ibid., pp. 4–5.
14. Ibid., p. 5.
15. Letter, William F. Chana to James Young, April 4, 2001;
“Comments on Derek Horne Convair XF-92A Website,” Edwards
AFB History Office F-92 file. The quoted text is from the letter.
16. Joseph W. Redd, Jr., Major Charles E. Yeager, and Lieutenant
Colonel Frank K. Everest, “Performance Flight Tests of the XF-92A
Airplane, USAF S/N 46-682, with a J33-A-29 Power Plant,”
AFFTC Technical Report No. 53-11, Edwards AFB, CA: Air Force
Flight Test Center, 1953, pp. 4–5.
17. Ibid.
18. Crossfield and Blair, Always Another Dawn, pp. 165–167. When a
pilot was killed in a crash at Edwards Air Force Base, a street was
named for the pilot as a memorial. Crossfield was the only living
pilot at the time with a street named after him.
Panel 1953 Annual Report,” pp. 10–11.
20. Hallion and Gorn, On the Frontier, p. 405 (flight records); Thomas
R. Sisk and Duane O. Muhleman, “Lateral Stability and Control
Characteristics of the Convair XF-92A Delta-Wing Airplane as
Measured in Flight,” NACA RM H55A17 (May 26, 1955), p. 1;
Euclid C. Holleman, “Flight Measurements of the Lateral Response
Characteristics of the Convair XF-92A Delta-Wing Airplane,”
NACA RM H55E26 (August 5, 1955), p. 5.
21. Sisk and Muhleman, “Lateral Stability and Control Characteristics
of the Convair XF-92A Delta-Wing Airplane as Measured in Flight,”
p. 19.
263
Probing the Sky

Panel 1953 Annual Report,” p. 10; Hallion and Gorn, On the
Frontier, p. 405 (flight records).
p. 12. Other research memorandums on the XF-92A include
Clinton T. Johnson, “Flight Measurements of the Vertical-Tail
Loads on the Convair XF-92A Delta-Wing Airplane,” NACA RM
H55H25 (October 27, 1955); and Earl R. Keener and Gareth H.
Jordon, “Wing Pressure Distributions over the Lift Range of the
Convair XF-92A Delta-Wing Airplane at Subsonic and Transonic
Speeds,” NACA RM H55G07 (November 30), 1955.
24. Ibid.
25. Ibid.
26. Crossfield and Blair, Always Another Dawn, p. 167; Convair XF-92A
Fact Sheet, National Museum of the United States Air Force.
p. 1.
28. Crossfield and Blair, Always Another Dawn, pp. 165, 167.
29. Redd, Yeager, and Everest, “Performance Flight Tests of the XF-92A
Airplane, USAF S/N 46-682, with a J33-A-29 Power Plant,” p. 4.
30. Fred Stoliker, Bob Hoey, and Johnny Armstrong, Flight Testing at
Edwards: Flight Test Engineers’ Stories 1946–1975 (Edwards, CA:
Flight Test Historical Foundation, n.d.), pp. 29–30.
31. In late 1949 and early 1950, Langley researchers conducted wind
tunnel tests of models of aircraft with swept wings and optimized
fuselage shapes. The results were disappointing. The models had
a higher total drag than transonic theory had predicted for the
combined drag of the fuselage and wings. Richard T. Whitcomb,
one of the researchers working on the study, realized he needed to
understand the details of the flow fields around different wing/fuse-
lage combinations at transonic speeds. The first clue was found in
Schlieren photos, which showed shock waves forming at the model’s
nose; a second where the wing and fuselage pushed the air out of
their way; and a third at the wings’ trailing edge. This investiga-
tion was followed in November 1951 by another test series focused
on measuring the transonic drag caused by interference between
the wings and fuselage. Whitcomb drew two conclusions from
this series. The first was that even small changes in the shape of the
fuselage caused disproportionate changes in the amount of drag. The
264
second was that the drag from the wings and fuselage was the result
of both interacting together and not, as had been assumed, the result
of simple addition. In thinking about these conclusions, Whitcomb
realized that if the middle of the fuselage were pinched in, the air
would flow over the area and would not form the strong shock waves
shown in the Schlieren photos.
Convair had learned of the “area rule” and hurriedly applied it to
their YF-102A prototype. Whitcomb assisted with the effort. On
December 20, 1954, a YF-102A modified with an area-rule fuse-
lage went supersonic without difficulty. The modifications proved
successful, and the Air Force accepted its first production F-102A
on June 24, 1955. The F-102A showed that a delta-wing aircraft
could fulfill the promise hinted at in the XF-92A. These planes were
soon followed by Convair’s F-106 interceptor and B-58 medium
bomber, which were also semi-tailless delta-wing aircraft. Other U.S.
semi-tailless delta-wing aircraft were Lockheed’s Blackbird fam-
ily—the A-12, YF-12, M-21, SR-71, and D-21, as well as North
American Aviation’s XB-70 Valkyrie. France undertook the devel-
opment of a family of semi-tailless delta-wing aircraft, beginning
with the Dassault MD 550 Mirage I test aircraft, the Mirage II and
Mirage III fighters, and the Mirage IV strategic bomber (the same
sequence of a test aircraft, two fighters, and a bomber that Convair
had followed). Another French delta aircraft, the Nord 1500-02
Griffon II, was designed to be a ramjet-powered fighter similar to
the original XP-92 concept and, flown by Andre Turcat, successfully
demonstrated the concept. British researchers extensively devel-
oped deltas, beginning with the Avro Type 707 research aircraft,
which served as a flying mockup of the Vulcan delta-wing strategic
bomber. The Boulton Paul P.111 and P.120 served as transonic delta-
wing research aircraft. The Fairey Delta 2 extended this research
to supersonic speeds and, in modified form, influenced the “ogee”
wing planform of the Anglo-French Concorde SST, one of only two
commercial deltas, which introduced airline passengers to Mach 2
flight. The Soviet Tu-144 also featured a delta wing, of the so-called
“double delta” planform. Sweden introduced a remarkably advanced
multipurpose fighter, the delta-wing Saab 35 Draken, following this
with the JA-37 Viggen and the JAS-39 Gripen. With the end of the
Cold War, several air forces now fly the Eurofighter, a combination
of a classic semi-tailless delta planform with small canards. Finally,
on the horizon are several proposed delta-shaped Unmanned Aerial
Vehicles (UAVs), which eliminate the vertical tail. Thus, across a
265
Probing the Sky
span of six decades, the descendants of the XF-92A fly, marking it as

one of the most successful of the NACA’s configuration demonstra-
tor X-planes.
Regarding these aircraft, see Bill Gunston, Early Supersonic Fighters
of the West (New York: Charles Scribner’s Sons, 1975); Marcelle
Size Knaack, Post–World War II Fighters 1945–1973 (Washington,
DC: Office of Air Force History, 1986); and John W.R. Taylor,
Jane’s Pocket Book of Research and Experimental Aircraft (New York:
Collier Books, 1976). Regarding Whitcomb and his “Area Rule,”
see Richard T. Whitcomb, “A Study of the Zero-Lift Drag-Rise
Characteristics of Wing-Body Combinations Near the Speed of
Sound,” NACA RM L52H08 (September, 3, 1952); Richard P.
Hallion, “Richard Whitcomb’s Triple Play,” Air Force Magazine
93, no. 2 (February 2010): 68–72; and Jeremy Kinney, “Richard
Whitcomb and the Quest for Aerodynamic Efficiency,” in Richard
P. Hallion, ed., NASA’s Contributions to Aeronautics, vol. 1:
Aerodynamics, Structures, Propulsion, Controls, NASA SP-2010-570-
Vol 1 (Washington, DC: NASA, 2010), pp. 89–133.
266
The third D-558-1 in flight with scattered clouds in the distance and the desert below. The
Skystreak did not have the impact the XS-1 had on aviation technology, but it did contribute to
the understanding of transonic flight. (NASA)
268
Round One: A Reflection
The United States research-airplane program has been a fruitful one

which resulted in a systematic and orderly approach to extending
man’s flight capabilities in both the atmosphere and space.
—Kenneth S. Kleinknecht1
In the final years of the Second World War, knowledgeable authorities recog-
nized that aviation was poised on the cusp of a revolution. The piston-powered
propeller airplane was in decline, the jet was in the ascendant, and the rocket
and missile were an uncertain emerging technology. Aircraft were already enter-
ing the transonic region, and the ballistic missile had already traversed it and
was even nudging the hypersonic. The next clear challenge in piloted, winged
flight was the achievement of supersonic flight.
But supersonic flight required a reliable knowledge base, and at this pivotal
moment in aeronautical history, the NACA and other Federal institutions
lacked the ability to produce it using conventional means. The standard tool
of aeronautical research—the wind tunnel—was unable to cope with the dis-
torted flows caused by shock impingement at transonic velocities. Theory was
contradictory. Flight experience was replete with examples of aircraft that had
already crashed and pilots who had already died in the grips of “compress-
ibility.” It was in this environment that the military services, the NACA, and
the industry formed a robust partnership to seek a new kind of research tool,
one that would use the sky itself as a laboratory—the transonic and supersonic
research airplane.
The decade following the initial research flights of the X-1 and D-558-1 was
a dynamic period in the history of aviation, arguably exceeded only by the first
decade after the invention of the airplane by the Wright brothers in 1903. It
marked a great divide between what came before and what came after. When
the decade began, jet aircraft represented a new and unreliable technology and
Mach 1 seemed to many an unattainable goal. By decade’s end, supersonic
flight was routine, humanity had placed small satellites in orbit, and space
now beckoned. In the same fashion that aviation progressed in the first years
after the Wrights, researchers, designers, engineers, and pilots were feeling their
269
Probing the Sky
way into the supersonic era. Not surprisingly, aircraft of both eras had severe
shortcomings in control, handling, stability, and performance.
Years after the First World War, an authentic reproduction of a World War I
Sopwith Camel was built and test-flown to see how well the plane flew. By
then, designers, engineers, and pilots well understood how an airplane should
fly and what handling qualities it should have. The description of the Camel’s
flight characteristics was telling:
Once in the air…the pilot is faced with almost total control dis-
harmony. The Camel is mildly unstable in pitch and considerably
unstable in yaw, and both elevator and rudder are extremely light
and sensitive…the ailerons are in direct and quite awe-inspiring
contrast. The Camel…has four enormous…barn doors [for
ailerons] which require an equally enormous force to be moved
quickly. And when you have moved them, the wing section is so
degraded…that the roll response is very slow indeed…. At the
same time, aileron drag is quite staggering.2
While the wording of NACA research memorandums relating to the Round

One research aircraft was not as “colorful,” one need only compare this descrip-
tion of the Sopwith Camel to the shortcomings of the X-planes of the early tran-
sonic and supersonic era to understand that both were flying into the unknown:
D-558-1: Landing gear and brake problems, vulnerability of the

control cables to damage, an unreliable capsule escape system,
nose-heavy above Mach 0.80, wing heaviness at high Mach num-
bers, rolling-yawing oscillation, and elevator vibrations approach-
ing the maximum load limit.
D-558-2: Long and slow takeoff using the original jet engine
and JATO rockets, pitch-up during maneuvers leading to loss of
control, poor high-speed stability resulting in uncontrolled rolls, an
unreliable capsule escape system, and oscillations during landing.
X-3: Unreliable and underpowered jet engines, tire failures,
very high takeoff and landing speeds, failure to reach design per-
formance, flawed control system, and dangerous combined oscil-
lations in flight. The most serious issue was inertial coupling.
Walker encountered forces during the inertial coupling flight
that reached the fuselage’s maximum-load limit. Had they been
slightly greater, both plane and pilot would have been lost.
X-4: Severe stability and control issues, flight control problems,
jet engine problems, and poor handling at speeds near Mach 1.
270
X-5: Complexity and weight of the wing-sweep mechanism,

undesirable lateral-stability and control characteristics, violent
stall/spin characteristics, and slow recovery. The aircraft was so
unstable under some conditions that one pilot doubted he could
keep it flying right side up.
XF-92A: Overly sensitive longitudinal and lateral controls
combined with a lag in the response, unreliable and underpow-
ered engines, poor directional stability, poor low-speed control
during approach and landing, and poor brakes.3
Indeed, somewhat ironically, having shortcomings proved in some respects

to offer an advantage. They allowed the NACA, the Air Force, the Navy, and
contractors to anticipate and correct deficiencies before operational aircraft
were built using their configurations.
It is well worthwhile, then, to review these aircraft from the point of view
of what they contributed:
D-558-1: With a largely similar configuration to the more famous

XS-1, this aircraft complemented the results achieved with the
X-1. The rocket-powered X-1 provided data at speeds above Mach
1, and the jet-powered D-558-1 covered the high transonic speed
range. As a result, the X-1 did not have to make low-speed flights;
this accelerated the overall collection of data. Additionally, this
aircraft was the first to test vortex generators, which improved
wing aerodynamics at transonic speed and became standard on
jet aircraft.
D-558-2: Research with this pioneering swept-wing air-
craft showed how to avoid the stability problems inherent in
these wings, in particular that the horizontal stabilizers should
be mounted low on the fuselage. When this aircraft was being
designed, all operational U.S. and British jet aircraft had straight
wings. Swept wings introduced the potential for much higher
performance. The D-558-2 was also used in external stores testing,
which allowed combat aircraft to fly at high speeds while carrying
streamlined weapons and fuel tanks. Information derived from
the D-558-2 program made all swept-wing aircraft that came after
it safer and certainly more controllable.
X-3: In terms of performance, this was the biggest disappoint-
ment among the early X-planes. It never came close to reaching
its design speed of Mach 2 due to its underpowered jet engines.
But it did show that low-aspect-ratio wings were practical, and
271
Probing the Sky
such wings were subsequently used on aircraft such as the F-104,

F-5/T-38, and X-15. The X-3’s most valuable contribution was to
reveal inertial coupling to be anything but theoretical. The fully
instrumented X-3 greatly enhanced the NACA’s understanding
of the phenomenon and was passed to industry as well.
X-4: The semi-tailless swept wing was the only configuration
tested in the 1950s that proved impractical. It would be another
three decades before computer fly-by-wire systems would make
the semi-tailless swept-wing configuration a workable reality, in
the first generation of stealth aircraft and in many other aircraft
and remotely piloted systems that have followed.
X-5: The aircraft pioneered variable-sweep wings, but at a
terrible price. While its wing pivot concept was impractical, it
did validate the concept of variable sweep, encouraging further
development that spawned the fixed outboard pivot point, used
on subsequent aircraft such as the F-111, MiG-23, B-1, F-14,
and Tu-160.
XF-92A: This accidental research airplane had more than its
share of poor flying characteristics, control system flaws, design
shortcomings, and engine deficiencies. Unloved by its pilots, it
nevertheless demonstrated the value of the delta-wing configura-
tion, which became one of the iconic planforms of the super-
sonic—and indeed hypersonic—era. The XF-92A directly led to
a notable family of delta successors, and from this sprang as well
(though not solely because of it) Richard Whitcomb’s concept of
area ruling.
The X-series family was generally hailed for its contributions to aeronauti-
cal science but was not without some critics. For example, Clarence L. “Kelly”
Johnson, head of the famed Lockheed “Skunk Works,” dismissed the X-planes
as drawing resources and technical effort that could best be focused on the
development of practical aircraft designs. But in retrospect it is hard to imag-
ine, given the utter unknowns of supersonic flight, that any wind tunnel or
rocket-boosted model tests could alone have produced the volume of data
needed to determine which designs would work and which would not. The
X-planes also provided practical experience operating exotic aircraft with unfa-
miliar designs and as-yet-unidentified problems. Had there been no X-planes,
designers would have had to try out different configurations on prototype
production aircraft. This would have been a waste of time and money and
would have interfered with the building of U.S. air power. And despite his
public dismissal, Kelly Johnson himself quietly took full advantage of NACA
272
research when he designed the Lockheed F-104. That task would have taken
considerably longer and cost much more than it did had the NACA not under-
taken X-plane research.4
This book has traced the development and significance of seven of those
research airplanes. Each had its own distinctive “personality,” and each contrib-
uted to aeronautical advancement in its own way. While each program had its
own trials, tribulations, and difficulties, in great measure all worked together
to advance the cause of flight.
Beyond these “Round One” airplanes was “Round Two”: the hypersonic
X-15 that took piloted, winged flight across the divide from the atmosphere
into space. If more glamorous, more exciting, and better known than its prede-
cessors, it was nevertheless largely dependent for its success upon the structure,
organizational culture, and expertise that the military services, the NACA (and
NASA thereafter), and the industry had developed during the earlier Round
One effort.
As the Round One aircraft wound down and Round Two began its own ges-
tation, a shocking event occurred that dramatically reshaped not only American
but global perspectives on air and space: the Soviet Union’s launch of Sputnik,
the world’s first artificial satellite, on October 4, 1957. The significance of
Sputnik’s launch became even more apparent in ensuing weeks and was the
impetus for a permanent U.S. space program and a new agency to run it.5 Less
than a week after Sputnik I was launched, the American Rocket Society called
for the establishment of an “Astronautical Research and Development Agency”
that would be responsible for nonmilitary space activities. The Rocket and
Satellite Research Panel of the National Academy of Sciences made a similar
suggestion on November 21, 1957.
President Dwight D. Eisenhower responded with several actions. These
included announcing the establishment of the position of the Special Assistant
to the President for Science and Technology, and he named Massachusetts
Institute of Technology (MIT) president James R. Killian, Jr., to the post.
Soon after, the President’s Scientific Advisory Committee (PSAC) was estab-
lished, with Killian named its chairman. Killian’s first assignment in his new
position was to review the alternatives for a U.S. space program. His report,
titled “Memorandum on Organizational Alternatives for Space Research and
Development,” was completed on December 30, 1957, and offered two options.
The first was for the DOD to establish a central space laboratory that could
undertake basic space research as well as military-related work. This laboratory
might also have the authority to sponsor research in civilian institutions.6 His
second was using the NACA as an alternative to a military-sponsored central
space laboratory, an idea first proposed by James McCormack, a vice presi-
dent of MIT, and James B. Fisk, of Bell Telephone Laboratories.7 In Killian’s
273
Probing the Sky
proposal, space activities would be split between the DOD and “some other
agency or agencies external to the DOD” that “might engage in basic research.”8
Killian continued, “One obvious way of doing this would be to encourage
the NACA to extend its space research and to provide it with the necessary
funding to do so.”
In February 1958, Eisenhower asked Killian to study options for organiz-
ing the U.S. space effort. Killian appointed a PSAC panel headed by Nobel
laureate Edward Purcell that included the University of California’s Herbert
York, NACA chairman James H. “Jimmy” Doolittle, and Polaroid’s Edwin
Land. The Purcell committee members were favorably disposed to accepting
Killian’s recommendation regarding the NACA, particularly since other con-
tenders (including the National Science Foundation and the Atomic Energy
Commission) had various shortcomings. Accordingly, in early March, the com-
mittee recommended a reorganized NACA as the organization best prepared
to manage civilian space activity. Draft legislation was soon written, leading
to the establishment of the National Aeronautics and Space Administration,
the follow-on agency to the NACA, and began operations on October 1,
1958. The “new-old” Agency absorbed the NACA’s existing laboratory struc-
ture—the Langley, Lewis, Ames, and Edwards facilities—joining them to the
Development Operations Division of the Army Ballistic Missile Agency (sub-
sequently the Marshall Space Flight Center) and partnering with the California
Institute of Technology’s Jet Propulsion Laboratory (JPL).
Each of the Centers brought different skills, traditions, cultures, and
experiences to the new Agency. These had been shaped in the course of the
transonic-supersonic revolution, which had made the NACA the only organi-
zation realistically qualified to take America into space. On October 5, 1958,
4 days after opening its doors, NASA Administrator T. Keith Glennan said,
“All right. Let’s get on with it!”9 Thus began the Mercury program, first to put
an American astronaut into orbit.
Mercury was where the legacy of the Round One research programs paid
off. In contrast to the academic research culture of Ames, Langley, and Lewis,
the High-Speed Flight Station was more operationally oriented. This was due
to its focus on flying research aircraft, the lack of wind tunnels or similar
ground research facilities, and the close ties it maintained with the military
services. The experiences of Round One airplanes, and particularly the High-
Speed Flight Station pilots, ground crews, and researchers, became the basis
for planning Mercury.10
From December 1946 onward, the High-Speed Flight Research Station
personnel had been flying piloted rockets, with hundreds of successful flights
to the facility’s credit. No one else on Earth, in the United States, Europe, or
the U.S.S.R., had such experience. The Round One pilots—Air Force, Navy,
274
NACA, Marine, or contractor—were the first to see the sky turn black and
see Earth’s curvature, its hallmark as a planet. Their flights reached speeds and
altitudes not exceeded until the first Soviet and American space missions. They
wore the first partial pressure suits and then, later, full pressure suits anticipat-
ing the spacesuits of the 1960s in order to survive cockpit depressurization at
the altitudes at which they flew. Their ground crews developed and mastered
the procedures for checking out the aircraft; maintaining the temperamental
rocket engines; and handling liquid oxygen, pure hydrogen peroxide, alcohol/
water mixtures, anhydrous ammonia, and the other dangerous chemicals used
in the engines. High-Speed Flight Station engineers had done the first test
work with the reaction controls necessary to stabilize an orbiting spacecraft
operating in a vacuum. Indeed, the very design developed for testing on the
X-1B became the basis for the Mercury spacecraft’s thrusters. And it was the
High-Speed Flight Research Station’s control room that the Mercury project
copied when casting about for a way to monitor the spacecraft.
In many respects, the NACA’s human contribution to the X-series program
was its most significant. The transonic and supersonic revolution was, like his-
tory itself, the working of men and women over time. Those of the NACA who
had advanced flight through the speed of sound were, to a great degree, the
men and women who now advanced flight into space.11 Robert Gilruth, who
pioneered the interim “wing flow” method of transonic research and who had
favored a thin wing for the XS-1, became Director of the Space Task Group,
responsible for America’s first piloted space program, Project Mercury, and later
Director of the Manned Spacecraft Center, now the NASA Lyndon B. Johnson
Space Center. Walter C. Williams, who had gone to Muroc from Langley to
head the Muroc Flight Test Unit, remaining there to lead the High-Speed
Flight Station into the NASA era, became Operations Director for Project
Mercury, and later NASA’s Chief Engineer, a singular distinction. Hartley
Soulé, the NACA’s Research Airplane Projects Leader, supervised establishment
of its tracking network. Gerald Truszynski, who had instrumented and tracked
the XS-1 at Pinecastle and Muroc, remaining through the Round One era at
Edwards as chief of instrumentation at the HSFS, eventually became NASA
Associate Administrator for Tracking and Data Acquisition in the Apollo and
Shuttle eras.
Fittingly, one of their colleagues, a young research pilot who flew the X-1,
the X-5, and later (in the early years of NASA) the X-15, went on to fly into
space and become the first human to set foot on another celestial body: Neil
A. Armstrong. He was humanity’s link from the surface of Earth through the
atmosphere and into space and from the subsonic to the transonic, supersonic,
hypersonic, and onwards to escape velocity and footprints on other worlds.12
275
Probing the Sky
Progress always extracts a price, and flight has always extracted a particularly
high tribute from those who have sought to advance its frontiers. The early
Round One X-series was no different: between 1948 and 1957, accidents and
incidents related to the Round One program claimed nine airplanes—seven
research airplanes and two launch aircraft—and the lives of five aircrew (two in
airplanes discussed in this work). It is to them—pilots Howard Lilly, Raymond
Popson, Jean Ziegler, Milburn Apt, and aircrewman Frank Wolko—that this
book is dedicated.13
276
Endnotes
1. Kenneth S. Kleinknecht, “The Rocket Research Aircraft,” in Eugene
M. Emme, ed., The History of Rocket Technology: Essays on Research,
Development, and Utility (Detroit: Wayne State University Press,
1964), p. 210.
2. Brian Lecomber, “Flying the Sopwith Camel and Fokker Triplane,”
Flight International (April 8, 1978), quoted in Robert P. Harper, Jr.,
and George E. Cooper, “Handling Qualities and Pilot Evaluation,”
1984 Wright Brothers Lectureship in Aeronautics (New York:
American Institute of Aeronautics and Astronautics, 1984), p. 7.
3. Hallion and Gorn, On the Frontier, pp. 56–59.
4. Miller, The X-Planes X-1 to X-45, pp. 285–289. Years later, it should
be noted, Johnson himself sought to acquire “X-series” status for his
Model CL-1200 Lancer, a proposed derivative of the F-104, which
did in fact receive the designation X-27, though it was never actually
built and flown.
5. Robert A. Divine, The Sputnik Challenge: Eisenhower’s Response to
the Soviet Satellite (New York: Oxford University Press, 1993), pp.
7–9; Roger D. Launius, “An Unintended Consequence of the IGY:
Eisenhower, Sputnik, and the Founding of NASA,” AIAA Paper No.
2008-860 (2008).
6. J.R. Killian, Jr., “Memorandum on Organizational Alternatives for
Space Research,” accessed on July 5, 2011, http://history.nasa.gov/
sputnik/iv1.html.
7. Brian R. Page, “The Creation of NASA,” Journal of the British
Interplanetary Society 32 (October 1979): 449–450.
8. Killian, “Memorandum on Organizational Alternatives for Space
Research.”
9. Loyd S. Swenson, Jr., James M. Grimwood, and Charles C.
Alexander, This New Ocean: A History of Project Mercury, NASA
SP-4201 (Washington DC: NASA, 1966), p. 109.
10. William M. Bland, Jr., “Project Mercury,” in Emme, ed., The History
of Rocket Technology, pp. 212–240.
11. As related in Sylvia Doughty Fries, NASA Engineers and the Age of
Apollo, NASA SP-4104 (Washington, DC: NASA, 1992).
12. As noted in Hallion, Supersonic Flight, pp. 203–204.
13. The aircraft were the Bell X-1-3, X-1A, and X-1D; the Bell X-2 #1
and X-2 #2; the Bell X-5 #2; the Douglas D-558-1 #2; and two
Boeing EB-50 launch aircraft (one lost with the ground explosion
of the X-1-3 and one scrapped after the in-flight explosion and loss
277
Probing the Sky
of the X-2 #2). The aircrew were Jean Ziegler, Frank Wolko, and
Milburn Apt (both X-2s); Ray Popson (X-5 #2); and Tick Lilly
(D-558-1 #2). Additionally, Bell pilot Joseph Cannon was seri-
ously injured in the explosion of the X-1-3, fortunately surviving to
fly again.
278
APPENDIX
280
Technical Specifications for the Round One Aircraft
Probing the Sky
Douglas Douglas D-558-2

Aircraft Type: Bell XS-11 Douglas X-3 Northrop X-4 Bell X-5 Convair XF-92A
D-558-1
Jet only All rocket
Engine Number,
1 GE 1 West. 1 RMI 2 West. 2 West. 1 Allison 1 Allison
Manufacturer, 1 RMI XLR-11
TG-180 J34 XLR-8 J34 J30 J35 J35
Type 6,000
5,000 3,000 6,000 4,850 ea. 1,600 ea. 4,900 5,600
Static Thrust (lb)
Propellants (gallons and 311 LOX 293 260 345 LOX 240 340 560
230 kerosene 970 kerosene
type of propellant)2 DEA gasoline 378 DEA gasoline kerosene kerosene
T.O./Launch
12,250 10,105 10,572 15,787 21,900 7,820 9,960 15,560
(lb)
Weights
Landing (lb) 7,000 7,711 7,914 9,421 17,500 6,452 7,850 11,808
Height (ft) 10.85 12.14 12.67 12.8 12.52 14.8 12.2 17.75
Length (ft) 30.90 35.71 42.0 66.75 23.25 33.6 42.80

Douglas Douglas D-558-2
Aircraft Type: Bell XS-11 Douglas X-3 Northrop X-4 Bell X-5 Convair XF-92A
D-558-1
Jet only All rocket
Douglas
NACA 63-010 root NACA NACA 64(10)A011
Airfoil Section NACA 65-108 NACA 65-110 Modified NACA 65(06)-006.5
to NACA 63-012 tip 0010-64 to 64(08)A008.28
Hexagon3
31.9 max
Span (ft) 28.0 25.0 25 22.69 26.1 31.33
20.0 min
Wing Area 184.3 max
130 150.7 175.0 166.50 200 425
(sq ft) 166.9 min
Wing Root
6.18 6.21 9.04 10.58 10.25 5.61 min MAC 27.13
Chord (ft)
MAC4 10.05 max MAC
Tip Chord (ft) 3.09 5.1 4.11 4.67 0.0
Flap Area
11.6 n/a 12.58 17.22 16.7 15.9 n/a
(sq ft)
Aileron Area
6.30 7.94 9.8 8.08 17.2 [elev.] 7.24 76.19 [elev.]
(sq ft total)
Airfoil Section 65-006 65-008 63-010 hexagon none 65A-006 none
Horizontal Area (sq ft) 26 35.98 39.9 n/a 31.5 n/a
Tail n/a
Elevator Area 43.24
5.2 8.6 9.4 17.2 [elev.] 6.9 76.19 [elev.]
(sq ft total)
Area (sq ft) 25.6 25.68 36.6 23.73 16.0 29.5 75.35
Vertical
Tail Rudder Area
5.2 7.92 6.15 5.44 4.1 4.7 11.50
(sq ft)
Mach No. at 1.45 0.832 0.825 2.01 1.21 0.92 1.03 1.01
Perf. Altitude, ft 35,000 sea level 20,000 62,000 30,000 30,000 30,000 30,000
(mph) (960) (632) (585) (1,290) (822) (625) (700) (686)
1. Specification for XS-1 no. 1 (46-062) at time of first supersonic flight, October 14, 1947. 2. LOX: Liquid Oxygen; DEA: Diluted Ethyl Alcohol. 3. Douglas, not an NACA,
281
Technical Specifications for the Round One Aircraft
airfoil section. 4. MAC: Mean Aerodynamic Chord.

Acknowledgments
Writing a book as lengthy and complicated as a history of the early X-planes

could not have been done without the help of many people. They include
Tony Springer at NASA Headquarters, who provided funding for this and
earlier projects; Christian Gelzer, the project supervisor; and Peter W. Merlin,
who tracked down half-century-old weekly progress reports on the X-planes
in the Dryden archives. Thanks to Karl A. Bender and Freddy Lockarno, at
the Dryden Flight Research Center’s Technical Library, for tracking down
NACA Research Memorandums describing the data collected and problems
experienced with the early X-planes. Thanks as well to Albion Bowers for his
explanation of rocket science. Sarah Merlin receives my thanks for the first
round of copyediting. Thanks also to Dr. Richard P. Hallion, who added sig-
nificantly to the text, and to the copy editors at NASA Headquarters.
Finally, I wish to recognize the courage of the test and research pilots, be
they NACA, contractor, or military, who flew the X-planes. They were the
explorers of an undiscovered territory. Like the Wright brothers, they had to
find their way without guideposts. The body of experience that had been built
up over the previous decades was of no use in understanding the demands of
transonic and supersonic flight. The engineers, wind tunnel researchers, and
pilots were on their own. But in their efforts to overcome the unknowns, they
were unknowingly setting the stage not simply for supersonic flight, but for a
far greater leap. The procedures, safety rules, and experience gained in breaking
the sound barrier were later used when the NACA became NASA, the agency
whose new role was to explore space.
Curtis Peebles
282
Selected Bibliography
I. Reports and Memorandums, by Type

General
Beeler, De E., and George Gerard. “Wake Measurements Behind a Wing
Section of a Fighter Airplane in Fast Dives.” NACA TN No. 1190, 1947.
Briggs, L.J., and H.L. Dryden. “Pressure Distribution over Air Foils at High
Speeds.” NACA Report No. 255, 1926.
Briggs, L.J., and H.L. Dryden. “Aerodynamic Characteristics of Twenty-Four

Airfoils at High Speeds.” NACA Report No. 319, 1929.
Briggs, L.J., and H.L. Dryden. “Aerodynamic Characteristics of Circular-Arc

Briggs, L.J., G.F. Hull, and H.L. Dryden. “Aerodynamic Characteristics of

Buckingham, Edgar. “Jet Propulsion for Airplanes.” NACA Report No. 159,
1923.
Caldwell, F.W., and E.N. Fales. “Wind Tunnel Studies in Aerodynamics

Phenomena at High Speeds.” NACA Report No. 83, 1920.
Chief, Bureau of Aeronautics, United States Navy. “Transfer of Aircraft to the

NACA Under Public Law 266 of the 81st Congress.” December 27, 1950.
Chief of Naval Operations, United States Navy. “Transfer of Aircraft to the

NACA Under Public Law 266 of the 81st Congress.” January 5, 1951.
Day, Richard E. Coupling Dynamics in Aircraft: A Historical Perspective. NASA

SP-532. Edwards AFB, CA: NASA Dryden Flight Research Center, 1997.
283
Probing the Sky
Harper, Robert P., Jr., and George E. Cooper. “Handling Qualities and Pilot
Evaluation.” Wright Brothers Lectureship in Aeronautics. New York:
American Institute of Aeronautics and Astronautics, 1984.
Jones, Robert T. “Wing Plan Form for High-Speed Flight.” Collected Works of
Robert T. Jones. NACA Report No. 863, in NASA TM X-3334. Washington,
DC: NASA, 1976.
Loftin, Laurence K., Jr. Quest for Performance: The Evolution of Modern Aircraft.
NASA SP-468. Washington, DC: NASA, 1985.
Marschak, Thomas A. The Role of Project Histories in the Study of R&D, Report
P-2850. Santa Monica: The RAND Corporation, January 1964.
McAvoy, William H., Oscar W. Schey, and Alfred W. Young. “The Effect on
Airplane Performance of the Factors That Must Be Considered in Applying
Low-Drag Cowling to Radial Engines.” NACA Report No. 414, 1931.
Moss, Sanford A. Superchargers for Aviation. New York: National Aeronautics

Council, Inc., 1942.
National Advisory Committee for Aeronautics, High-Speed Flight Station,

“Research Aircraft Projects Panel 1952 Annual Report.”
National Advisory Committee for Aeronautics, High-Speed Flight Station,

“Research Aircraft Projects Panel 1953 Annual Report.”
Parks, James H. “Experimental Evidence of Sustained Coupled Longitudinal

and Lateral Oscillations from a Rocket-Propelled Model of a 35° Swept
Wing Airplane Configuration.” NACA RM L54D15, 1954.
Phillips, William H. “Effects of Steady Rolling on Longitudinal and Directional

Stability.” NACA TN No. 1627, 1948.
Polhamus, Edward C., and Thomas A. Toll. “Research Related to Variable

Sweep Aircraft Development.” NASA TM 83121, 1981.
Saltzman, Edwin J., and Theodore G. Ayers. Selected Examples of NACA/

NASA Supersonic Flight Research. NASA SP-513. Edwards AFB, CA: NASA
Dryden Flight Research Center, 1995.
284
Shortal, Joseph Adams. A New Dimension—Wallops Island Flight Test Range:

The First Fifteen Years. NASA RP-1028. Washington, DC: NASA, 1978.
Stack, John, W.F. Lindsey, and R.E. Littell, “The Compressibility Burble and
the Effect of Compressibility on Pressures and Forces Acting on an Airfoil.”
NACA Report No. 646, 1938.
Stillwell, Wendell H. “Results of Measurements Made During the Approach

and Landing of Seven High-Speed Research Airplanes.” NACA RM
H54K24, 1955.
Strass, H. Kurt, and Edison M. Fields. “Flight Investigation of the Effect

of Thickening the Aileron Trailing Edges on Control Effectiveness for
Sweptback Tapered Wings Having Sharp- and Round-Nose Sections.”
NACA RM L9L19, 1950.
Thompson, Milton O., with J.D. Hunley. Flight Research: Problems Encountered
and What They Should Teach Us. NASA SP-2000-4522. Washington, DC:
NASA, 2000.
Weick, Fred E. “Drag and Cooling with Various Forms of Cowling for a
‘Whirlwind’ Engine in a Cabin Fuselage.” NACA TN No. 301, 1928.
Whitcomb, Richard T. “A Study of the Zero-Lift Drag-Rise Characteristics

of Wing-Body Combinations Near the Speed of Sound.” NACA RM
L52H08, 1952.
Williams, Walter C., and A. Scott Crossfield. “Handling Qualities of High-

Speed Airplanes.” NACA RM L52A08, 1952.
Wood, Donald H. “Tests of Nacelle-Propeller Combinations in Various

Positions with Reference to Wings. Part I. Thick Wing–NACA Cowled
Nacelle–Tractor Propeller.” NASA Technical Report No. 415, 1933.
Williams, Walter C. Letter to NACA Headquarters. “Choice of Color for

Research Aircraft at Edwards.” NACA HSFRS, December 3, 1951.
Bell XS-1 (X-1)

Angle, Ellwyn E., and Euclid C. Holleman. “Determination of Longitudinal
Stability of the Bell X-1 Airplane from Transient Responses at Mach
285
Probing the Sky
Numbers up to 1.12 at Lift Coefficients of 0.3 and 0.6.” NACA RM

L50I06a, 1950.
Beeler, De E., and John P. Mayer. “Measurements of the Wing and Tail Loads
During the Acceptance Tests of Bell XS-1 Research Airplane.” NACA RM
L7L12, 1948.
Carman, L. Robert, and John R. Carden. “Lift and Drag Coefficients for the
Bell X-1 Airplane (8-Percent-Thick Wing) in Power-Off Transonic Flight.”
NACA RM L51E08, 1951.
Drake, Hubert M., and John R. Carden. “Elevator-Stabilizer Effectiveness

and Trim of the X-1 Airplane to a Mach Number of 1.06.” NACA RM
L50G20, 1950.
Drake, Hubert M., Harold R. Goodman, and Herbert H. Hoover. “Preliminary

Results of NACA Transonic Flights of the XS-1 Airplane with 10-Percent-
Thick Wing and 8-Percent-Thick Horizontal Tail.” NACA RM L8I29,
1948.
Drake, Hubert M., Milton D. McLaughlin, and Harold R. Goodman. “Results

Obtained During Accelerated Transonic Flight Tests of the Bell XS-1
Airplane in Flights to a Mach Number of 0.92.” NACA RM L8A05a, 1948.
Goodman, Harold R., and Hubert M. Drake. “Results Obtained During

Extension of U.S. Air Force Transonic-Flight Tests of XS-1 Airplane.”
NACA RM L8I28, 1948.
Goodman, Harold R., and Roxanah B. Yancey. “The Static-Pressure Error of

Wing and Fuselage Airspeed Installations of the X-1 Airplanes in Transonic
Flight.” NACA RM L9G22, 1949.
Hamlin, Benson. The Design Conception of Supersonic Flight. Amherst, NY: The
Amherst Museum/Niagara Frontier Aviation and Space Museum, 1996.
Mattson, Axel T. “Force and Longitudinal Control Characteristics of a 1/16-

Scale Model of the Bell XS-1 Transonic Research Airplane at High Mach
Numbers.” NACA RM L7A03, 1947.
United States Air Force. Air Force Supersonic Research Airplane XS-1 Report
No. 1. Wright Field, OH: USAF Air Materiel Command, January 9, 1948.
286
Williams, Walter C., and De E. Beeler. “Results of Preliminary Flight Tests of

the XS-1 Airplane (8-Percent Wing) to a Mach Number of 1.25.” NACA
RM L8A23a, 1948.
Williams, Walter C., Charles M. Forsyth, and Beverly P. Brown. “General

Handling Qualities Results Obtained During Acceptance Flight Tests of
the Bell X-1 Airplane.” NACA RM L8A09, 1948.
Young, James O. Meeting the Challenge of Supersonic Flight. Edwards AFB, CA:
Air Force Flight Test Center, 1997.
Bell X-1A
Drake, Hubert M., and Wendell H. Stillwell. “Behavior of the Bell X-1A
Research Airplane During Exploratory Flights at Mach Numbers Near 2.0
and at Extreme Altitude.” NACA RM H55G25, 1955.
Powell, J.L. “X-1A Airplane Contract W33-038-ac-20062, Flight Test Progress

Report No. 15, Period from 9 December through 20 December 1953.” Bell
Aircraft Corporation Report No. 58-980-019, February 3, 1954.
Bell X-2
Stiffler, Ronald. The Bell X-2 Rocket Research Aircraft: The Flight Test Program.
Edwards AFB, CA: Air Force Flight Test Center, August 12, 1957.
Douglas X-3 Stiletto

Day, Richard E., and Jack Fischel. “Stability and Control Characteristics
Obtained During Demonstration of the Douglas X-3 Research Airplane.”
NACA RM H55E16, 1955.
Douglas Aircraft Company, Flight Testing Division. X-3 test reports:
“Flight Report No. 7,” 1953.
Douglas Aircraft Company, “Model X-3 Mock-up Conference Airplane

Descriptions and Illustrations.” Santa Monica, CA: Douglas Aircraft
Company, December 6, 1948.
287
Probing the Sky
Fischel, Jack, Euclid C. Holleman, and Robert A. Tremant. “Flight Investigation

of the Transonic Longitudinal and Lateral Handling Qualities of the
Douglas X-3 Research Airplane.” NACA RM H57I05, 1957.
Fleming, Frank N. “Evolution of the Configuration of the X-3 Supersonic

Research Aircraft.” Santa Monica, CA: Douglas Aircraft Company, August
1, 1949.
Keener, Earl R., Norman J. McLeod, and Norman V. Taillon. “Effect of

Leading-Edge-Flap Deflections on the Wing Loads, Load Distribution, and
Flap Hinge Moments of the Douglas X-3 Research Airplane at Transonic
Speeds.” NACA RM H58D29, 1958.
Marcy, William L. “High-Speed Loads Measured on the Douglas X-3 Research

Airplane.” NACA RM H57L08, 1958.
Marcy, William L., Harriet J. Stephenson, and Thomas V. Cooney. “Analysis of

the Vertical-Tail Loads Measured During a Flight Investigation at Transonic
Speeds of the Douglas X-3 Airplane.” NACA RM H56H08, 1956.
NACA HSFRS, X-3 Progress Report Memorandums for Research Airplane

Projects Leader:
“Progress Report for the X-3 Research Airplane for the Period November
1 to November 30, 1953.”
“Progress Report for the X-3 Research Airplane for the Period July 1 to
July 31, 1954.”
“Progress Report for the X-3 Research Airplane for the Period August 1 to
August 31, 1954.”
“Progress Report for the X-3 Research Airplane for the Period September
1 to September 30, 1954.”
“Progress Report for the X-3 Research Airplane for the Period October 1
to October 31, 1954.”
NACA HSFRS, X-3 Status Reports for Chief of Research:
“X-3 Status Report for Period 14 September to 18 October 1952.”
288
“X-3 Status Report for Week Ending 25 October 1952.”
“X-3 Weekly Status Report Week Ending 25 April 1953.”
“X-3 Weekly Status Report Week Ending 2 May 1953.”
“X-3 Weekly Status Report Week Ending 20 June 1953.”
“X-3 Weekly Status Report Week Ending 18 July 1953.”
“X-3 Weekly Status Report Week Ending 25 July 1953.”
“X-3 Weekly Status Report Week Ending 1 August 1953.”
“X-3 Weekly Status Report for Week Ending 1 November 1953.”
“X-3 Weekly Status Report for Week Ending 15 November 1953.”
NACA HSFRS Memorandum for Chief of Research. “X-3 Airplane Visit of

Harold F. Kleckner of the Douglas Company to Langley on October 18,
1951.”
NACA HSFRS Memorandum for Chief of Research. “Proposed Modifications

of X-3 Airplane to Rocket Power and Air Launch.” October 24, 1952.
NACA HSFS. “Flight Experience with Two High-Speed Airplanes Having

Violent Lateral-Longitudinal Coupling in Aileron Rolls,” NACA RM
H55A13, 1955.
Stephenson, Harriet J. “Flight Measurements of Horizontal-Tail Loads on the

Douglas X-3 Research Airplane.” NACA RM H56A23, 1956.
Northrop X-4 Bantam

Drake, Hubert M. “Stability and Control Data Obtained from First Flight of
X-4 Airplane.” NACA RM L9A31, 1949.
Matthews, James T., Jr. “Results Obtained During Flights 1 to 6 of the Northrop
X-4 Airplane (A.F. no. 46-677).” NACA RM L9K22, 1950.
289
Probing the Sky
NACA HSFRS. X-4 Progress Memorandums for NACA Research Airplanes

Project Leader:
“Acceptance of the X-4 Airplane.” n.d.
“Progress Report on Acceptance Tests of X-4 Airplanes from January 1 to

January 14, 1949.”
“X-4, January 14 to June 10, 1949.”
“X-4, June 17 through July 1, 1949.”
“X-4, June 20, 1949.”
“X-4, July 16 through July 29, 1949.”
“X-4 Airplanes July 30 through August 12, 1949.”
“X-4 Airplanes July 31 through August 12, 1949.”
“X-4, August 27 to September 14, 1949.”
“X-4 Airplanes October 8 to October 21, 1949.”
“X-4, November 19 to December 2, 1949.”
“X-4, December 3 to December 16, 1949.”
“X-4, December 31, 1949, to January 1, 1950.”
“X-4, January 27 to February 24, 1950.”
“X-4, February 24 to March 24, 1950.”
“X-4, May 5 to May 19, 1950.”
“X-4, July 15 to July 28, 1950.”
“X-4, August 26 to September 8, 1950.”
“X-4, September 8 to September 22, 1950.”
290
“X-4, September 22 to October 6, 1950.”
“X-4, October 6 to October 20, 1950.”
“X-4, November 3 to November 17, 1950.”
“X-4, December 15 to December 29, 1950.”
“X-4, May 18 to June 1, 1951.”
“X-4, June 16 to June 29, 1951.”
“X-4, June 30 to July 13, 1951.”
“X-4, July 14 to July 27, 1951.”
“X-4, July 28 to August 10, 1951.”
“X-4, August 11 to August 25, 1951.”
“X-4, October 6 to October 19, 1951.”
“X-4, October 20 to November 2, 1951.”
“X-4, November 3 to November 16, 1951.”
“X-4, January 26 to February 8, 1952.”
“X-4, February 9 to February 22, 1952.”
“X-4, March 8 to March 21, 1952.”
“X-4, March 22 to April 4, 1952.”
“X-4, April 5 to April 18, 1952.”
“X-4, April 19 to May 2, 1952.”
291
Probing the Sky
“X-4, May 17 to May 30, 1952.”
“X-4, May 31 to June 13, 1952.”
“X-4, June 28 to July 11, 1952.”
“X-4, July 12 to July 25, 1952.”
“X-4, October 1 to November 1, 1952.”
“X-4, January 1 to January 31, 1953.”
“X-4, February 1 to February 28, 1953.”
“X-4, March 1 to March 31, 1953.”
“X-4, April 1 to April 30, 1953.”
“X-4, May 1 to May 31, 1953.”
“X-4, July 1 to July 31, 1953.”
“X-4, September 1 to September 30, 1953.”
Sadoff, Melvin, and A. Scott Crossfield. “A Flight Evaluation of the Stability

and Control of the X-4 Swept-Wing Semitailless Airplane.” NACA RM
H54G16, 1954.
Sadoff, Melvin, and Thomas R. Sisk. “Summary Report of Results Obtained

During Demonstration Tests of the Northrop X-4 Airplanes.” NACA RM
A50I01, 1950.
292
Valentine, George M. “Stability and Control Data Obtained from Fourth and
Fifth Flights of the Northrop X-4 Airplane (A.F. No. 46-676).” NACA RM
L9G25a, 1949.
Williams, Walter C. “Results Obtained from Second Flight of X-4 Airplane

(A.F. No. 46-676).” NACA RM L9F21, 1949.
Williams, Walter C. “Results Obtained from Third Flight of Northrop X-4

Airplane (A.F. No. 46-676).” NACA RM L9G20a, 1949.
Bell X-5
Banner, Richard D., Robert D. Reed, and William L. Marcy. “Wing-Load
Measurements of the Bell X-5 Research Airplanes at a Sweep Angle of
58.7°.” NACA RM H55A11, 1955.
Bell Aircraft Corporation. “Bell X-5 Ready for Flight Tests.” June 14, 1951.
Bellman, Donald R. “Lift and Drag Characteristics of the Bell X-5 Research
Airplane at 59° Sweepback for Mach Numbers from 0.60 to 1.03.” NACA
RM L53A09c, 1953.
Briggs, Donald W. “Flight Determination of the Buffeting Characteristics of the

Bell X-5 Research Airplane at 58.7° Sweepback.” NACA RM L54C17, 1954.
Campbell, John P., and Hubert M. Drake. “Investigation of Stability and

Control Characteristics of an Airplane Model with Skewed Wing in the
Langley Free-Flight Laboratory.” NACA TN No. 1208, 1947.
Childs, Joan M. “Flight Measurements of the Stability Characteristics of the

Bell X-5 Research Airplane in Sideslips at 59° Sweepback.” NACA RM
L52K13b, 1953.
Donlan, Charles J., and William C. Sleeman. “Low-Speed Wind-Tunnel

Investigation of the Longitudinal Stability Characteristics of a Model
Equipped with a Variable-Sweep Wing.” NACA RM L9B18, 1949.
Finch, Thomas W., and Donald W. Biggs. “Preliminary Results of Stability

and Control Investigation of the Bell X-5 Research Airplane.” NACA RM
L52K18b, 1953.
293
Probing the Sky
Green, Warren E. The Bell X-5 Research Airplane. Wright-Patterson AFB, OH:
Wright Air Development Center, March 1954.
Kress, Robert W. “Variable-Sweep Wing Design.” AIAA Paper No. 83-1051,

1983.
Martin, James A. “Longitudinal Flight Characteristics of the Bell X-5 Research

Airplane at 59° Sweepback with Modified Wing Root.” NACA RM L53E28,
1953.
NACA HSFRS, X-5 Flight Research Summaries:
“Investigation of Flight at Speeds Below Minimum Drag, Flight No. 60.”

January 27, 1953.

January 29, 1953.

February 6, 1953.
NACA HSFRS, X-5 Progress Memorandums for Chief of Research:
“Progress Report for the X-5 Research Airplane for the Period May 3 to
May 16, 1952.”
“Progress Report for the X-5 Research Airplane for the Period July 12 to
July 25, 1952.”
“Progress Report for the X-5-1 Research Airplane for the Period July 26 to
August 8, 1952.”
“Progress Report for the X-5-1 Research Airplane for the Period October
1 to November 1, 1952.”
“Progress Report for the X-5-1 Research Airplane for the Period July 1 to
July 31, 1955.”
“Progress Report of Engineering and Instrumentation for the X-5 Airplane

During the Period August 25 to September 7, 1951.”
294

During the Period December 1 to December 14, 1951.”

During the Period December 15 to December 28, 1951.”

During the Period December 29, 1951, to January 11, 1952.”
NASA Langley Research Center. “Summary of NACA/NASA Variable-Sweep

Research and Development Leading to the F-111 (TFX).” NASA LRC
Working Paper LWP-285, 1966.
Nugent, Jack. “Lift and Drag of the Bell X-5 Research Airplane in the 45°
Sweptback Configuration at Transonic Speeds.” NACA RM H56E02,
1956.
Perry, Robert L. Innovation and Military Requirements: A Comparative Study.

RAND Report RM-5182PR. Santa Monica, CA: The RAND Corporation,
1967.
Reed, Robert D. “Flight Measurements of Horizontal-Tail Loads on the Bell

X-5 Research Airplane at a Sweep Angle of 58.7°.” NACA RM H55E20a,
1955.
Rogers, John T., and Angel H. Dunn. “Preliminary Results of Horizontal-

Tail Load Measurements of the Bell X-5 Research Airplane.” NACA RM
L52G14, 1952.
United States Air Force. “Report of Major Aircraft Accident—X-5, 50-1839A.”

Edwards AFB: AFFTC, October 27, 1953.
Videan, Edward N. “Flight Measurements of the Dynamic Lateral and

Longitudinal Stability of the Bell X-5 Research Airplane at 58.7° Sweepback.”
NACA RM H55H10, 1955.
Douglas D-558-1 Skystreak

Angle, Ellwyn E., and Euclid C. Holleman. “Longitudinal Frequency-Response
Characteristics of the D-558-I Airplane as Determined from Experimental
Transient-Response Histories to a Mach Number of 0.90.” NACA RM
L51K28, 1952.
295
Probing the Sky
Barlow, William H., and Howard C. Lilly. “Stability Results Obtained with
Douglas D-558-1 Airplane BuAero No. 37971 in Flights up to a Mach
Number of 0.89.” NACA RM L8K03, 1949.
Beeler, De E., Donald R. Bellman, and John H. Griffith. “Flight Determination

of the Effects of Wing Vortex Generators on the Aerodynamic Characteristics
of the Douglas D-558-I Airplane.” NACA RM L51A23, 1951.
Bellman, Donald R. Letter to NACA Headquarters. “Looseness of Stabilizer

Mountings for the D-558-I Airplane.” November 5, 1952.
Bellman, Donald R. Letter to NACA Headquarters. “Reply to Request for

Information on D-558-I Vertical-Tail Load Investigation.” January 12, 1953.
Douglas Aircraft Company. “Skystreak World’s Speed Records.” El Segundo,

CA: DAC, September 19, 1947.
Heinemann, Edward H. “The Development of the Navy-Douglas Model

D-558 Research Project.” El Segundo, CA: Douglas Aircraft Company,
November 17, 1947.
Hyatt, 1st Lt. Abraham, United States Marine Corps Reserve. “Proposed High-
Speed Research Airplane.” USN Bureau of Aeronautics Memorandum
AER-E-225-AH. September 22, 1944.
Keener, Earl R., and Mary Pierce. “Tabulated Pressure Coefficients and
Aerodynamic Characteristics in Flight on the Wing of the Douglas D-558-I
Airplane for a 1g Stall, a Speed Run to a Mach Number of 0.90, and a
Wind-Up Turn at a Mach Number of 0.86.” NACA RM L50J10, 1950.
Keener, Earl R., James R. Peel, and Julia B. Woodbridge. “Tabulated Pressure
Coefficients and Aerodynamic Characteristics Measured in Flight on the
Wing of the Douglas D-558-I Airplane Throughout the Normal-Force
Range at Mach Numbers of 0.67, 0.74, 0.78, and 0.82.” NACA RM
L50L12a, 1951.
NACA, “NACA Aircraft Accident Investigation Report Douglas D-558-1

Airplane, BuNo 37971.” Muroc, CA: NACA Muroc Flight Test Unit. May
3, 1948.
296
Sadoff, Melvin, William S. Roden, and John M. Eggleston. “Flight Investigation

of the Longitudinal Stability and Control Characteristics of the Douglas
D-558-I Airplane BuAero No. 37972 at Mach Numbers up to 0.89.”
NACA RM L51D18, 1951.
Soulé, Hartley A. Letter to NACA Headquarters. “Request for Comments

on Abandonment of D-558-I Vertical-Tail Load Investigation.” December
19, 1952.
Thompson, Jim Rogers, William S. Roden, and John M. Eggleston. “Flight

Investigation of the Aileron Characteristics of the Douglas D-558-I Airplane
BuAero No. 37972 at Mach Numbers Between 0.6 and 0.89.” NACA RM
L50D20, 1950.
Williams, Walter C. “Flight Measurement of the Stability Characteristics of the

D-558-1 Airplane BuAero No. 37971 in Sideslips.” NACA RM L8E14a, 1949.
Williams, Walter C. “Limited Measurements of Static Longitudinal Stability

in Flight of Douglas D-558-1 Airplane, BuAero No. 37971.” NACA RM
L8E14, 1948.
Douglas D-558-2 Skyrocket

Ankenbruck, Herman O., and Chester H. Wolowicz. “Lateral Motions
Encountered with the Douglas D-558-II All-Rocket Research Airplane During
Exploratory Flights to a Mach Number of 2.0.” NACA RM 54I27, 1954.
Ankenbruck, Herman O., and Theodore E. Dahlen. “Some Measurements of

Flying Qualities of a Douglas D-558-II Research Airplane During Flights
to Supersonic Speeds.” NACA RM L53A06, 1953.
Brunn, Cyril D., and Wendell H. Stillwell. “Mach Number Measurements and
Calibrations During Flights at High Speeds and at High Altitudes Including
Data for the D-558-II Research Airplane.” NACA RM H55J18, 1956.
Dahlen, Theodore E. “Maximum Altitude and Maximum Mach Number

Obtained with the Modified Douglas D-558-II Research Airplane During
Demonstration Flights.” NACA RM L53B24, 1953.
Fischel, Jack. “Effects of Wing Slats and Inboard Fences on the Longitudinal
Stability Characteristics of the Douglas D-558-II Research Airplane in
297
Probing the Sky
Accelerated Maneuvers at Subsonic and Transonic Speeds.” NACA RM

L53L16, 1954.
Fischel, Jack, and Cyril D. Brunn. “Longitudinal Stability Characteristics in

Accelerated Maneuvers at Subsonic and Transonic Speeds of the Douglas
D-558-II Research Airplane Equipped with a Leading Edge Wing Chord-
Extension.” NACA RM H54H16, 1954.
Fischel, Jack, and Donald Reisert. “Effects of Several Wing Modifications

on the Subsonic and Transonic Longitudinal Handling Qualities of the
Douglas D-558-II Research Airplane.” NACA RM H56C30, 1956.
Fischel, Jack, and Jack Nugent. “Flight Determination of the Longitudinal

Stability in Accelerated Maneuvers at Transonic Speeds for the Douglas
D-558-II Research Aircraft Including the Effects of an Outboard Wing
Fence.” NACA RM L53A16, 1953.
Fischel, Jack, Robert W. Darville, and Donald Reisert. “Effects of Wing-

Mounted External Stores on the Longitudinal and Lateral Handling
Qualities of the Douglas D-558-II Research Airplane.” NACA RM
H57H12, 1957.
Sjoberg, S.A., James R. Peele, and John H. Griffith. “Flight Measurements

with the Douglas D-558-II (BuAero No. 37974) Research Airplane: Static
Longitudinal Stability and Control Characteristics at Mach Numbers up
to 0.87.” NACA RM L50K13, 1951.
Sjoberg, S.A. “Preliminary Measurements of the Dynamic Lateral Stability

Characteristics of the Douglas D-558-II (BuAero No. 37974) Airplane.”
NACA RM L9G18, 1949.
Consolidated Vultee XF-92A Dart

Holleman, Euclid C. “Flight Measurements of the Lateral Response
Characteristics of the Convair XF-92A Delta-Wing Airplane.” NACA RM
H55E26, 1955.
Holleman, Euclid C., John H. Evans, and William C. Triplett. “Preliminary

Flight Measurements of the Dynamic Longitudinal Stability Characteristics
of the Convair XF-92A Delta-Wing Airplane.” NACA RM L53E14, 1953.
298
Redd, Joseph W., Jr., Maj. Charles E. Yeager, and Lt. Col. Frank K. Everest.
“Performance Flight Tests of the XF-92A Airplane, USAF S/N 46-682, with
a J33-A-29 Power Plant.” AFFTC Technical Report No. 53-11. Edwards
AFB, CA: Air Force Flight Test Center, 1953.
Sisk, Thomas R., and Duane O. Muhleman. “Lateral Stability and Control
Characteristics of the Convair XF-92A Delta-Wing Airplane as Measured
in Flight.” NACA RM H55A17, 1955.
Sisk, Thomas R., and John M. Mooney. “Preliminary Measurements of Static

Longitudinal Stability and Trim for the XF-92A Delta-Wing Research
Airplane in Subsonic and Transonic Flight.” NACA RM L53B06, 1953.
II. Autobiographies, Biographies, and Memoirs

Bridgeman, William, and Jacqueline Hazard. The Lonely Sky. New York: Henry
Holt and Company, Inc., 1955.
Cebeci, Tuncer. Legacy of a Gentle Genius: The Life of A.M.O. Smith. Long
Beach, CA: Horizons Publishing, Inc., 1999.
Connor, Margaret. Hans von Ohain: Elegance in Flight. Reston, VA: American
Institute of Aeronautics and Astronautics, 2001.
Crossfield, A. Scott, and Clay Blair, Jr. Always Another Dawn: The Story of a
Rocket Test Pilot. New York: Arno Press, 1972.
Everest, Lt. Col. Frank K., as told to John Guenther. The Fastest Man Alive.
New York: E.P. Dutton, 1958.
Glennan, T. Keith, edited by J.D. Hunley and in association with Roger D.

Launius. The Birth of NASA: The Diary of T. Keith Glennan. NASA SP-4105.
Washington, DC: NASA, 1993.
Golley, John (in association with Frank Whittle and with technical assistance
from Bill Gunston). Whittle: The True Story. Washington, DC: Smithsonian
Institution Press, 1987.
Grierson, John. Jet Flight. London: Samson Low, Marston & Co., Ltd., 1944.
299
Probing the Sky
Heinemann, Edward H., and Rosario Rausa. Ed Heinemann: Combat Aircraft

Designer. Annapolis, MD: Naval Institute Press, 1980.
Hoover, Robert A. “Bob,” with Mark Shaw. Forever Flying. New York: Orion, 1996.
Johnston, A.M. “Tex,” and Charles Barton. Tex Johnston: Jet-Age Test Pilot.
Washington, DC: Smithsonian Institution Press, 1991.
Lundgren, William R. Across the High Frontier: The Story of a Test Pilot—Major
Charles E. Yeager, USAF. New York: William Morrow, 1955.
Meyer, Corwin H. Corky Meyer’s Flight Journal: A Test Pilot’s Tales of Dodging
Disasters—Just in Time. North Branch, MN: Specialty Press, 2006.
Nahum, Andrew. Frank Whittle: Invention of the Jet. Duxford, U.K.: Icon
Books, 2004.
Peebles, Curtis, ed. The Spoken Word: Recollections of Dryden History: The Early
Years. SP-2003-4530. Washington, DC: NASA, 2003.
Phillips, W. Hewitt. Journey in Aeronautical Research: A Career at NASA Langley

Research Center. Washington, DC: NASA, 1998.
Späte, Wolfgang. Top Secret Bird: The Luftwaffe’s Me 163 Comet [sic]. Missoula,
MT: Pictorial Histories Publishing Company, 1989.
Stoliker, Fred, Bob Hoey, and Johnny Armstrong. Flight Testing at Edwards:
Flight Test Engineers’ Stories 1946–1975. Edwards AFB, CA: Flight Test
Historical Foundation, 1996.
Whittle, Sir Frank. Jet: The Story of a Pioneer. New York: Philosophical Library,
1954.
Yeager, Chuck, and Leo Janos. Yeager: An Autobiography. New York: Bantam,
1985.
Young, James O., ed. The Men of Mach One: A Supersonic Symposium. Edwards
AFB, CA: Air Force Flight Test Center, 1990.
Ziegler, Mano. Rocket Fighter. Garden City, NY: Doubleday and Co., Inc.,
1963.
300
III. Miscellaneous Sources

Abzug, Malcolm J., and E. Eugene Larrabee. Airplane Stability and Control:
A History of the Technology That Made Aviation Possible. Cambridge, U.K.:
Cambridge University Press, 1997.
Alegre, Gregory. Campini Caproni, no. 5 of the Ali d’Italia series. Turin: La
Bancarella Aeronautica, n.d.
Aronstein, David C., and Albert C. Piccirillo. Have Blue and the F-117: Evolution
of the “Stealth Fighter.” Reston, VA: American Institute of Aeronautics and
Astronautics, 1997.
Ball, John, Jr. Edwards: Flight Test Center of the USAF. New York: Duell, Sloan,
and Pearce, 1962.
Becker, John V. The High-Speed Frontier: Case Histories of Four NACA Programs,
1920–1950. NASA SP-445. Washington, DC: NASA, 1980.
Boyne, Walter J., and Donald S. Lopez, eds. The Jet Age: Forty Years of Jet
Aviation. Washington, DC: National Air and Space Museum in association
with the Smithsonian Institution Press, 1979.
Bradley, Robert E. “The Birth of the Delta Wing.” American Aviation Historical
Society (winter 2003).
Brown, Capt. Eric, RN. “An Ill-Fated ‘Swallow’…but a Harbinger of Summer.”

Air Enthusiast 10 (July 1979).
Burnet, Charles. Three Centuries to Concorde. London: Mechanical Engineering

Publications Limited, 1979.
Constant II, Edward W. The Origins of the Turbojet Revolution. Baltimore: The
Johns Hopkins University Press, 1980.
Divine, Robert A. The Sputnik Challenge: Eisenhower’s Response to the Soviet

Satellite. New York: Oxford University Press, 1993.
Emme, Eugene M., ed. The History of Rocket Technology: Essays on Research,
Development, and Utility. Detroit, MI: Wayne State University Press, 1964.
301
Probing the Sky
Fries, Sylvia Doughty. NASA Engineers and the Age of Apollo. NASA SP-4104.
Gray, George W. Frontiers of Flight: The Story of NACA Research. New York:
Alfred A. Knopf, 1948.
Gorn, Michael H. Expanding the Envelope: Flight Research at NACA and NASA.
Lexington, KY: The University Press of Kentucky, 2001.
Hallion, Richard P. “Before B-2: The History of the Flying Wing,” Part 1: “to
1945”; Part 2: “since 1945.” Air Power History 41, nos. 3–4 (fall–winter
1994).
Hallion, Richard P. “Convair’s Delta Alpha.” Air Enthusiast Quarterly 2

(February 1977).
Hallion, Richard P. “Lippisch, Gluhareff, and Jones: The Emergence of the

Delta Planform and the Origins of the Sweptwing in the United States.”
Aerospace Historian 26, no. 1 (spring 1979).
Hallion, Richard P. On the Frontier: Flight Research at Dryden, 1946–1981.

SP-4303. Washington, DC: NASA, 1985.
Hallion, Richard P. “Richard Whitcomb’s Triple Play.” Air Force Magazine 93,
no. 2 (February 2010).
Hallion, Richard P. “Rocket Dreams.” World War II 23, no. 4 (October–

November 2008).
Hallion, Richard P. “Serendipity at Santa Monica: The Story of the Douglas

X-3.” Air Enthusiast Quarterly 4 (October 1977).
Hallion, Richard P. Supersonic Flight: Breaking the Sound Barrier and Beyond—
The Story of the Bell X-1 and Douglas D-558. London: Brassey’s, 1997 ed.
Hallion, Richard P. “X-4: The Bantam Explorer.” Air Enthusiast Quarterly 3

(June 1977).
Hallion, Richard P., and Michael H. Gorn. On the Frontier: Experimental Flight
at NASA Dryden. Washington, DC: Smithsonian Books, 2003.
302
Hallion, Richard P., ed. NASA’s Contributions to Aeronautics, vol. 1: Aerodynamics,

Structures, Propulsion, Controls. NASA SP-2010-570-Vol 1. Washington,
DC: NASA, 2010.
Hansen, James R. Engineer in Charge: A History of the Langley Aeronautical

Laboratory, 1917–1958. NASA SP-4305. Washington, DC: NASA, 1987.
Hansen, James R. The Bird Is on the Wing: Aerodynamics and the Progress of
the American Airplane. College Station, TX: Texas A&M University Press,
2004.
Hunley, J.D., ed. Toward Mach 2: The Douglas D-558 Program. NASA SP-445.
Washington: DC: NASA SP-4222, 1999.
Jenkins, Dennis R. X-15: Extending the Frontiers of Flight. NASA SP-2007-562,

Knaack, Marcelle Size. Post–World War II Fighters. Washington, DC: Office

of Air Force History, 1986.
Launius, Roger D. “An Unintended Consequence of the IGY: Eisenhower,

Sputnik, and the Founding of NASA.” AIAA Paper No. 2008-860, 2008.
Libis, Scott. Skystreak, Skyrocket, & Stiletto: Douglas High-Speed X-Planes.

North Branch, MN: Specialty Press, 2005.
Matranga, Gene J., C. Wayne Ottinger, Calvin R. Jarvis, and D. Christian

Gelzer. Unconventional, Contrary, and Ugly: The Lunar Landing Research
Vehicle. NASA SP-2004-4535. Washington, DC: NASA, 2004.
Miller, Jay. The X-Planes X-1 to X-45. Hinckley, U.K.: Midland Publishing,
2001.
Page, Brian R. “The Creation of NASA.” Journal of the British Interplanetary

Society 32 (1979).
Pavelec, Sterling Michael. The Jet Race and the Second World War. Annapolis,
MD: Naval Institute Press, 2007.
Peebles, Curtis. “Risk Management in the X-Planes Era: D-558-II vs. X-1A at
Mach 2.” Quest 11, no. 4 (2004).
303
Probing the Sky
Peebles, Curtis. Road to Mach 10: Lessons Learned from the X-43A Flight Research
Program. Reston, VA: AIAA, 2008.
Rivas, Brian. A Very British Sound Barrier: DH 108—A Story of Courage,

Triumph, and Tragedy. Newark, U.K.: Red Kite/Wing Leader Publishers,
2012.
Rotundo, Louis. Into the Unknown: The X-1 Story. Washington, DC:
Smithsonian Institution Press, 1994.
Schatzberg, Eric. Wings of Wood, Wings of Metal. Princeton, NJ: Princeton

University Press, 1998.
Schlaifer, Robert, and S.D. Heron. Development of Aircraft Engines and Fuels.
Boston: Harvard Business School, 1950.
St. Peter, James. The History of Gas Turbine Development in the United States…A
Tradition of Excellence. Atlanta, GA: International Gas Turbine Institute of
the American Society of Mechanical Engineers, 1999.
Smith, G. Geoffrey. Gas Turbines and Jet Propulsion for Aircraft. New York:
Aircraft Books Inc., 1946.
Swenson, Loyd S., Jr., James M. Grimwood, and Charles C. Alexander. This
New Ocean: A History of Project Mercury. NASA SP-4201. Washington,
DC: NASA, 1966.
Taylor, John R. Jane’s Pocket Book of Research and Experimental Aircraft. New
York: Collier Books, 1976.
Thompson, Milton O., and Curtis Peebles. Flying Without Wings: NASA Lifting
Bodies and the Birth of the Space Shuttle. Washington, DC: Smithsonian
Institution Press, 1999.
Winter, Frank H. “‘Black Betsy’: The 6000C4 Rocket Engine, 1945–-1989.”

Paper presented at the 23rd Symposium on the History of Astronautics,
40th International Astronautical Congress of the International Astronautical
Federation. Malaga, Spain: October 1989.
Witkin, Richard. The Challenge of the Sputniks. New York: Doubleday Headlines
Publications, 1958.
304
Young, James O. Meeting the Challenge of Supersonic Flight. Edwards AFB, CA:
Air Force Flight Test Center History Office, 1997.
305
Index
Numbers in bold indicate pages with illustrations and figures
A See also research aircraft/X-series research

A-12 program, 140, 264–66n31 aircraft
AD-1 research aircraft, 208 aircraft carriers, 207–8, 235
aerial refueling and X-5 formation test flights, airflow
221–24, 239n29, 239n33 calculation of as inviscid fluid, 1
aerodynamics compressibility effects and swept-wing
flow separation and, 2–3 configurations, 17, 39–40n43
speed and rules for, v compressibility effects at transonic and
supersonic flight and, 34 supersonic speeds, 3, 6–7
transformation of technology, tools, and flow separation, concept of, 2
procedures for supersonic era, v, 3–4 flow separation, discovery of, 2
Aerojet rocket engine, 10 flow separation, research on, 2–3
ailerons potential flows, 17
D-558-1 data collection and research research on supersonic flight, 1–2
results, 54–56, 59–61 X-3 design and compressibility effects,
D-558-1 lateral stability and aileron- 137–38
effectiveness studies, 69–70, 73 airfoils
D-558-2 lateral stability and aileron- circular-arc airfoils, 3
effectiveness studies, 112 D-558-2 wing design, 87
X-5 aileron overbalance and effectiveness, drag and design of, 2–3
234 high-speed flight and design of, 14–15
aircraft research on high-speed aerodynamics,
carrier landings and design of, 207–8, 235 2–3, 4
changes in design of, 198 Air Force Academy, U.S., 195
designation system for numbering, Air Force/Army Air Forces (AAF)/Army Air
37–38n32, 38n34 Corps, U.S.
design of for jet propulsion, 8, 37n24 aerial refueling operations, 239n33
evolutionary process of development of, v, aircraft designation system, 37–38n32,
3–4 38n34
first flight, v, 1 airfoil research by, 2–3
military role in aircraft design, 35–36n12 communication between NACA and, 230
307
Probing the Sky
fighter aircraft design competition, 243–44 value of air launching, 53

German rocket plane development, X-3 all-rocket/air-launch configuration,
concerns about, 6 140–41
Mach 2 sustained speeds, aircraft for, 135, X-15 air launching, 140
160n2 XS-1 launching from, 15, 25, 27, 32, 33,
research aircraft designation system, 53, 124
37–38n32 See also B-29 Superfortress
rocket-powered research aircraft study, Allison J33 engines, 247, 248, 250–51, 254,
10–12 262–63n10
transonic research aircraft, role in Allison J35 engines, 210, 211–12
development of, 9–12 altitude
turbojet-powered research aircraft study, D-558-2 altitude-record flight, 101, 102
10–12 Mach numbers, air temperature, and, 78n12
variable-sweep aircraft used by, 236 stability and control and, 56
X-3 flights, 146–49 American F-100A aircraft, 38n35
X-5 flights, 213–16, 226–30 American Rocket Society, 273
X-5 program role, 212, 238n10 Ames-Dryden AD-1 research aircraft, 208
XF-92A flights, 246, 249–54, 262–63n10 Ames Research Center/Ames Aeronautical
XF-92A program role, 243–47 Laboratory
XS-1 program publicity, 26–27 creation of, 7
XS-1 program role, 12, 14–15, 16, 25–31, mission and purpose of, 7
34, 72 Model 7002 wind tunnel testing, 248
XS-1 wing design, 20 research culture at, 274
Air Force Flight Test Museum, 195 space research by, 274
Air Force Museum, U.S., 157, 195, 232, 257 XS-1 test site tour, 27
Air Force Test Center, vi Ankenbruck, Herman O., 113–14, 116
air launching and mother ships Antelope Valley College, 133n63
advantages of, 124 Apollo 11 Command Module, 81–82n51
aircraft available for air launching, 15 Apollo program, 275
chase planes, 100, 116, 227 Apt, Milburn, 77n7, 276, 277–78n13
D-558-1 air-launching recommendation, 53 area rule design, 87, 259, 261, 264–
D-558-2 air launch and P2B-1S engine 66n31, 272
failure, 118 Armstrong, Neil, 118, 232, 275
D-558-2 air launching, 95–101, 109, 110, Armstrong Flight Research Center (Dryden
115–17, 124 Flight Research Center), vi, 74, 197
D-558-2 air launching, loading into launch Armstrong Whitworth AW 52 aircraft, 196–97
aircraft for, 117 Army Ballistic Missile Agency (Marshall Space
D-558-2 air launching, modifications for, 84 Flight Center), 274
drop-body testing, 19 Arnold, H.H. “Hap,” 6, 7
rocket-powered aircraft launching from, 11 Ascani, Fred, 29
308
Index
Ashkenas, Irving, 169 XS-1 test site considerations, 16,

Atomic Energy Commission, 274 38–39n41
Avro Type 707 aircraft, 264–66n31 Bell L-39 aircraft, 87
Bell P-39 Airacobra aircraft, 7, 9
B Bell P-63 aircraft, 247
B-1/B-1B strategic bomber, 236, 272 Bell X-1/XS-1 aircraft. See X-1/XS-1
B-2A Spirit stealth bomber, 196–97 (Experimental Sonic-1) aircraft
B-24 aircraft, 86 Bell X-2/XS-2 aircraft. See X-2/XS-2 aircraft
B-29 Superfortress Bell X-5 aircraft. See X-5 aircraft
aerial refueling formation test flights, 221, Bell XP-59A jet, 7–8, 9, 11, 22, 37n24, 243
223–24 Bellman, Donald, 70–71
D-558 launch plane role of, iv, 13, 95, Big Bear Lake, 116
96–101, 110 Biggs, Lyman J., 2–3
drop-body testing launch plane role, 19 Boeing B-29 Superfortress. See B-29
loading aircraft for air launching, 117 Superfortress
P2B-1S engine failure and near tragedy, 118 Boeing EB-50 aircraft, 277–78n13
P2B-1S Superfortress, 96–101, 110, Boeing supersonic transport (SST), 236
115–17, 130n30 Boeing XB-47 jet bomber, 87
XS-1 launch plane role, 15, 25, 27, 32, 33, Boulton Paul P.111 aircraft, 264–66n31
53, 96, 124 Boulton Paul P.120 aircraft, 264–66n31
XS-1 transport to Muroc attached to, 24 Boyd, Albert, 28, 29, 30, 31, 49, 214, 252–53
B-36 bombers, 140, 198 Bridgeman, William “Bill”
B-52 bombers, 140 D-558-2 all-rocket/air-launch flights,
B-58 bombers, 264–66n31 96–101
Barnes, Florence “Pancho,” 33 D-558-2 altitude-record flight, 101
Becker, John V., 11–12, 74 D-558-2 flights, 94–95, 109, 112–13,
Beech Bonanza aircraft, 198 117, 126
Beeler, De E., 21 D-558-2 speed-record flights, 99–100
Bell, Lawrence D. “Larry” X-3 flights, 142, 143–44, 145–46
XS-1 flight test program direction and goals, Brown, Clinton, 20
27, 29 BT-13 aircraft, 247
XS-1 project and Collier Trophy, 42n83 Buckingham, Edgar, 4, 5
Bell Aircraft Corporation buffeting
“black” aircraft program, involvement in D-558-1 data collection and research
first, 7 results, 56–57, 60, 62–64, 66, 73
Niagara Falls facility, 213 D-558-2 pitch-up characteristics and, 93
XS-1 program publicity, 26–27 external stores research flights, 120, 123
XS-1 program role, 12, 14–15, 22–23, SB2C Helldiver design, 7
25–31 X-4 data collection and research results, 180
XS-1 project and Collier Trophy, 42n83 X-5 handling characteristics, 225–26
309
Probing the Sky
XS-1 data collection and research results, Carl, Marion E.

27–28, 29, 33 D-558-1 canopy design and crash helmets,
Burstein, Adolph, 244 79n18
Busemann, Adolf, 39–40n43, 75–76 D-558-2 flights, 114
Bush, George H.W., 80n43 Mach 2 flight, 114
Bush, Vannevar, 6, 37n21 speed record flights, 49–50, 78n12
Butchart, Stanley P. Carolina Aviation Museum, 81–82n51
D-558-1 flights, 66–67, 70, 71 Champine, Robert A. “Bob”
D-558-2 air launch and P2B-1S engine D-558-1 flights, 59, 59, 62
failure, 118 D-558-2 flights, 90
education of, 66–67 D-558-2 pitch-up characteristics, 91–92
experience of, 66–67, 80n43 transfer of, 130n38
external stores research flights, 122 Chance Vought F7U/F7U-3 Cutlass program,
flight tests, 69 64, 160n7, 195
P2B-1S pilot role, 115–16 chase planes, 100, 116, 227
pilot-escape systems, opinion about, 77n7 circular-arc airfoils, 3
X-4 flights, 191 Clousing, Lawrence, 27
X-5 flights, 221, 231–32 cockpit design, D-558-1 design, 53, 70
Collier Trophy, 42n83, 76
C Comet I jetliner, 200n4
C-74 transport aircraft, 140 Concorde SST, 264–66n31
C-119 cargo aircraft, 213, 226 Consolidated-Vultee (Convair) F-106 aircraft,
C-124 transport aircraft, 140 264–66n31
Caldwell, Frank W. Consolidated-Vultee (Convair) XF-92A aircraft.
D-558-1 canopy design and crash helmets, See XF-92A aircraft
79n18 Consolidated-Vultee (Convair) YF-102/F-102
propeller aerodynamics research, 2 aircraft, 259–61, 259, 260, 261, 264–66n31
speed record flights, 49–50, 78n12 control surfaces
California Institute of Technology, Jet Propulsion painting of D-558-1 control surfaces, 58
Laboratory, 274 supersonic flight and effectiveness of, 1
California State Polytechnic College, 81–82n51 XS-1 horizontal stabilizer, elevator, and tail
Campini engine and Secondo Campini, 36n13 design, 38n35
Cannon, Joseph, 277–78n13 See also specific surfaces
canopy Cooper, George, 194
D-558-1 design, 48, 53, 79n18 cowling design and location, 4
D-558-2 design, 88 Cox, L.K., 180, 181
X-4 design and lock, 173, 183 Cox, R.B., 58
XF-92A design, 247, 260 Crawford, Alden R., 20
XS-1 design, 88 Crimson Test Tube, 48. See also D-558-1
Skystreak
310
Index
Crossfield, A. Scott engines for aircraft, 12

D-558-1 flights, 64, 65, 70, 71 flight-test program direction, plan, and
D-558-2 flights, 92, 97, 102–8, 109, goals, 27, 85
111–18, 115, 121, 125–26, 130n38 funding for, 44
D-558-2 pitch-up characteristics, 92, models of, 9
102–8, 125 personnel involved in program, iv, 13
dirt road named after, 254, 255, 263n18 phase 1 aircraft and goals, 12–13, 27, 85
education of, 64 phase 2 aircraft and goals, 12–13, 27, 85
experience of, 64, 67, 126 phase 3 aircraft and goals, 12, 27, 85
external stores research flights, 120 research flight preparations, iv, 13
first flight jinx, 254, 255 specifications and preliminary design
hydrogen peroxide accident, 131–32n54 development, 12
inertial coupling research missions, 155 D-558-1 Skystreak
Mach 2 flight, 112–17 accident and death of Lilly, 51–53, 57, 74,
North American Aviation job, 117–18, 126, 277–78n13
132n55 air-launching recommendation, 53
pilot-escape systems, opinion about, 77n7 canopy design, 48, 53, 79n18
X-4 contributions to aviation research, 195 cockpit design, 53, 70
X-4 flight data reports, 186–88 contributions to aviation research, 71–75,
X-4 flights, 185–88, 189, 190, 193, 194, 124, 268, 269, 271
205n70 control surfaces, painting of, 58
X-5 aircraft, value of, 235 data collection and research results, 13,
X-5 flights, 217, 218, 231 50–51, 53–57, 67–71, 73–74, 81n44
X-5 handling characteristics, 230, 231 data collection at level flight, requirement
X-15 flights, 118, 126 for, 48
XF-92A flights, 254–56, 255, 258 design of, 44, 45–48, 68, 72–73, 85–86,
XF-92A handling characteristics, 258 139, 199, 268
XS-1 flights, 126 design of, approval for, 13–14
Crossfield Pike, 254, 255, 263n18 design team, 45
Crowley, John designation and numbering of, 77n2
swept-wing testing, approval for, 18–19 display of aircraft at end of program,
XS-1 specifications for research 81–82n51
requirements, 21 dive tests, 57, 67–68, 73
XS-1 test site considerations, 16 elevator vibrations and replacement of
Curtiss SB2C Helldiver aircraft, 7, 9 elevators, 64–65
fates of aircraft and perceived importance,
D 81–82n51
D-21 aircraft, 264–66n31 final flights and end of program, 71
D-558 program flight safety recommendations, 53
development of, decisions leading to, 8–12 flight-test program direction, plan, and
development program for, 12–13 goals, 27, 71–74
311
Probing the Sky
flight tests, 48–51, 53–71, 66 vortex generator research flights, 62–63,

frequency-response data, collection of, 73, 80n37
68–69 wing design, 46–47, 281
fuel system and capacity, 47, 58, 280 XS-1 compared to, 45
funding for, 44 D-558-2 Skyrocket
fuselage design and materials, 45–46, air launching, loading into launch aircraft
86, 87 for, 117
instrumentation for data collection, 50–51, air launching and P2B-1S engine failure, 118
61–62 air launching of, 95–101, 109, 110,
landing gear configuration and landing gear 115–17, 124
problems, 47, 48, 51 air launching of, modifications for, 84,
modifications to, 57–58 94–96, 110
NACA, transfer to, 59, 130n30 all-rocket/air-launch configuration,
NACA research flights, 58–69, 59 95–101, 123
nickname for, 48 altitude-record flight, 101, 102
paint schemes, 47–48, 51, 58, 72, canopy design, 88
78–79n15, 250 contributions to aviation research,
personnel involved in program, 44 124–27, 271
phase 1 aircraft designation, 12–13 data collection and research results,
pilot-escape system, 47, 77n7 121, 127
pilots, dangers faced by, 64 data collection from D-558-1 aircraft and
pressure-distribution research, 62–63, 73 design of, 13
shortcomings of, 270 design of, 73, 85–88, 139, 199
smallest aircraft around largest jet display of aircraft at end of program, 119,
engine, 46 133n63
specifications, technical, 280–81 dive tests, 89
speed brakes, 78–79n15, 211 elevator design, 88
speed capabilities of, 45, 68, 71–74 engine test firings, 116
speed record flights, 48–50, 78n12 external stores research flights, 119–23,
stability, control, and handling 119, 127, 132n60, 271
characteristics, 270 final flights and end of program, 118–19,
stabilizer and tail design and materials, 123–24, 133n63
38n35, 46, 67–69, 70–71, 138, 281 first flights, 88–91
stall/spin behavior, 62 flight-test program direction, plan, and
structural design and materials, 46–47 goals, 27, 85–86, 111, 124–27
tail-loads studies, 69, 70–71, 73–74 flight tests, 88–124, 100, 121, 129n16,
TG-180 engines issues, 50, 51–53, 58–59, 132n59
62, 74 fuel systems, turbopump, and capacities,
third aircraft, 44, 68, 72, 75 86, 87, 88, 89, 95, 97, 128n4, 280
tip-tank study, 71, 73 fuselage design, 86, 87
transonic flight with, 27 inertial coupling behavior of model, 152
312
Index
instrumentation for data collection, 87, 126 takeoff runs, length of, 88–89, 94–95
JATO-boosted flights, 89, 89, 94–95 wing fences, 88, 97, 102–6, 103, 108, 256
jet power for and configuration wing-load data collection flights, 122
specifications, 85–86, 87, 88, 90, 95, 96, wing modification to add sawtooth leading
98, 101–2, 280–81 edge, 106–8, 107
jet-rocket system research flights, 101–9, wing-strain-gauge calibration, 91
102, 124–27 Dassault MD 550 Mirage I aircraft, 264–66n31
landing gear, 87, 136 Dassault MD 550 Mirage II aircraft, 264–66n31
launch areas for, 115–16 Dassault MD 550 Mirage III aircraft,
Mach 2 flight, 110, 112–17 264–66n31
NACA, transfer to, 89, 96, 97, 101, 109, Dassault MD 550 Mirage IV aircraft,
130n30 264–66n31
NACA research flights, 89–94, 101–24 data collection systems and methodologies
number of research flights, 157 automatic computing equipment for data
performance potential, 95, 99–100, 124, analysis, 204–5n69
128n4 backlogs of data processing, 175
personnel involved in program, 121 “computers” for data analysis, 24, 91
phase 2 aircraft designation, 12–13 D-558-1 data collection at level flight,
pilot-escape system, 47, 77n7 requirement for, 48
pitch-up characteristics, 91–94, 97, D-558-1 data collection and research
102–9, 125–26, 125, 127, 133n64, 157 results, 13, 50–51, 53–57, 67–71,
research flight preparations, 13 73–74, 81n44
rocket power for and configuration D-558-1 instrumentation, 50–51, 61–62
specifications, 12, 85–86, 87, 89–90, 95, D-558-2 data collection flights and
96, 97, 101–2, 114–15, 114, 280–81 research results, 121, 127
shortcomings of, 270 D-558-2 instrumentation, 87, 126
specifications, technical, 280–81 D-558-2 wing-strain-gauge calibration, 91
speed capabilities and speed records, 84, frequency-response data, collection of,
99–100, 102, 109, 111, 112–17 68–69
stability, control, and handling instrumentation for data collection,
characteristics, 90–94, 97, 99–101, standard, 171
102–12, 125–26, 127, 270 Reeves Electronic Analog Computer
stabilizer and tail design, 88, 108–9, (REAC), 252
128n5, 138, 281 stability data collection methods, 252
stall/spin behavior, 62, 92 transformation of technology, tools, and
supersonic flight with, 27 procedures for supersonic era, v
survival of aircraft at end of program, women “computers” for data analysis, 24,
126–27 175, 204–5n69
swept-wing design, 73, 85–88, 124–26, X-3 data collection and research results,
125, 137–38, 281 147–50
tail-loads studies, 73–74, 127
313
Probing the Sky
X-5 data collection and research results, X-4 dives and dive brakes, 176, 179–80,
214, 215–26, 232–34 181, 188–89, 191, 193, 196, 203n53
XF-92A data collection and research X-5 dive tests, 219–20
results, 251–57 DM-1 glider, 244–45
XS-1 data collection and research results, Donaldson, E.M., 49
54, 56, 57, 73–74 Donlan, Charles J., 208, 236
Davidson, Milton, 8–9 Donovan, Robert C., 45
Davis, Frank W., 244 Doolittle, James H. “Jimmy,” 274
Day, Richard E. “Dick,” 154, 155–56 Douglas Aircraft Corporation
Daytona Beach, 16 D-558 program role, 12–14, 44
DC-3 aircraft, 4 Model 558 High-Speed Test Aircraft. See
de Havilland, Geoffrey, Jr., 168–69 D-558-1 Skystreak; D-558-2 Skyrocket;
de Havilland DH 100 Vampire jet fighter, 168 D-558 program
de Havilland DH 106 jet airliner, 168, 200n4 Douglas C-54 Skymaster transport, 15
de Havilland DH 108 Swallow aircraft, 168–69, Douglas C-74 transport aircraft, 140
170, 196–97 Douglas C-124 transport aircraft, 140
Defense, Department of (DOD), 273–74 Douglas X-3 aircraft. See X-3 Stiletto
delta-wing aircraft drag
advantages and value of, 247, 253, airfoil design and, 2–3
258–61 area rule and fuselage design, 87, 259,
birth of, 243 261, 264–66n31, 272
German designs, 244–45 external stores research flights, 119–23,
improvement recommendations, 253–54 119, 127, 132n60, 271
model testing of concept, 248 flow separation and, 2–3
popularity of design, 243 high fineness ratio and, 45–46, 138
wing fences, 255, 256, 257 Mach 2 flight preparations, 114–15
XF-92A design, 244–45, 247, 281 research on, 72
Derry, John, 168 swept wings and, 17–20, 207
design theory and transformation of technology, X-3 design and, 136, 137–38
tools, and procedures for supersonic era, X-4 lift-over-drag studies, 190–92, 195
v, 3–4 X-5 handling characteristics, 225
Diehl, Walter Stuart, 9, 11 Drake, Hubert, 156
digital fly-by-wire flight control systems, drop-body testing, 18–20
196–98, 272 Dryden, Hugh L.
dive maneuvers airfoil research by, 2–3
D-558-1 dive tests, 57, 67–68, 73 D-558-2 modifications, support for, 95–96
D-558-2 dive tests, 89 Mach 2 flight, approval for, 113
loss of aircraft during high-speed dives, photo of, 2
6–7, 9 wing design and drag, Volta Congress
X-3 dive tests, 145–46 presentation about, 39–40n43
XS-1 Mach 1 flight, confirmation of, 42n83
314
Index
Dryden Flight Research Center (Armstrong endplates on wings, 57

Flight Research Center), vi, 74, 197 engines
ducted-fan engine designs, 5, 8. See also Jeep cowling design and location, 4
piston-engine/combustion-chamber engine installation of and engine efficiency, 4
Durand, William F., 6, 7–8, 9 nacelle design and location, 4
See also jet and turbojet engines
E Everest, Frank K. “Pete,” 98–99
EB-50 aircraft, 277–78n13 F-100A aircraft, concerns about
Edwards, Glen, 172 performance of, 153
Edwards Air Force Base/Muroc Army Air Field X-3 flights, 147, 148–49
aeronautical research and development X-4 flights, 183
at, vi X-5 flights, 215
Air Force Flight Test Museum, 195 XF-92A flights, 249–51, 254
D-558-1 test site, 88–89, 89, 97 Explorer II balloon flight, 101
facilities at, 24 external stores research flights, 119–23, 119,
housing near, 24 127, 132n60, 271
landscape and terrain around, 24
renaming of, 96 F
space research at, 274 F3D-2 Skyknight aircraft, 160n7
street names, 263n18 F3H-1 Demon aircraft, 160n7
weather conditions at, 24 F-5/T-38 aircraft, 159, 160n7, 272
XP-59A program role, 7–8 F7U/F7U-3 Cutlass program, 64, 160n7, 195
XS-1 test site, 22–24 F-8 aircraft, 197
XS-1 test site considerations, 16, 27, F-14/F-14A Tomcat aircraft, 156, 236, 272
38–39n41 F-16 aircraft, 198
Eisenhower, Dwight D., 113, 273–74 F-80 Shooting Star, 113, 221–23
ejection seats. See pilot-escape systems/ F-86 aircraft
ejection seats chase plane role of, 100, 227
elevators F-86 all-moving tail, 68, 81n46
D-558-1 data collection and research F-86A deployment for Korean War, 81n44,
results, 54–55, 64–65, 67–69 81n46
D-558-2 design, 88 F-86A speed record, 78n12
movable and flight control, 38n35 F-86E stabilizer design, 38n35
XS-1 stabilizer, elevator, and tail design, handling characteristics, 249
32–33, 38n35, 138 high-speed research with, 74
elevons performance of, 113
concept behind, 167 pitch-up characteristics and swept-wing
X-4 use of, 167, 172, 183, 193, 194, 196 design, 92
XF-92A use of, 247 speed capabilities, 146
Elizabeth, Lake, 115 XP-86 aircraft, 28, 87
Ellis, Macon C., 20 F-94 aircraft, 249
315
Probing the Sky
F-100/F-100A aircraft, 38n35, 113, 153, Fokker trimotor aircraft, 4

155–56, 158 formation test flights and aerial refueling,
F-101 Voodoo aircraft, 156 221–24, 239n29, 239n33
F-102/YF-102 aircraft, 259–61, 259, 260, frequency-response data, collection of, 68–69
261, 264–66n31 Frost, Richard, 25, 30, 32
F-104 Starfighter aircraft fuel systems
design and development of, 113, 158–59, armor shielding for hoses and tanks, 53,
272–73, 277n4 58, 138
inertial coupling behavior, 156 D-558-1 fuel system and capacity, 47, 58,
success of design and use of by air forces, 280
159 D-558-1 tip-tank study, 71, 73
X-3 development and, 141, 158–59, D-558-2 fuel systems, turbopump, and
161n10, 195 capacities, 86, 87, 89, 95, 97, 128n4,
F-106 aircraft, 264–66n31 280
F-111/F-111A aircraft, 208, 236, 272 fuel hoses, 58
F-117A stealth fighter, 198 wet wings and fuel tanks, 47
Fairchild C-119 cargo aircraft, 213, 226 X-3 fuel system and capacity, 138, 280
Fairey Delta 2 aircraft, 264–66n31 X-4 fuel system and capacity, 170, 200n8,
Fales, Elisha N., 2 280
Fédération Aéronautique Internationale (FAI), 49 X-5 fuel-cell pressure gauge problems,
Ferri, Antonio, 75–76 213–14
Fields, Edison M., 203n50 X-5 fuel system and capacity, 212, 280
fighter aircraft XF-92A fuel system and capacity, 246, 280
design and development of, 37n24 XS-1 fuel system, turbopump, and fuel
design competition, 243–44 capacity, 15–16, 21, 25–26, 41n63,
X-5 as basis for, 210, 235 128n4, 280
Finch, Thomas W., 215–17 Full-Scale Tunnel, 245
Fischel, Jack, 103–7, 154, 155–56 fuselage
Fisk, James B., 273 area rule design, 87, 259, 261, 264–
FJ-1 aircraft, 247 66n31, 272
Fleming, Frank N., 135, 136, 151 D-558-1 design and materials, 45–46,
flight control systems 86, 87
armor shielding for cables, 53, 138, 170 D-558-2 design, 86, 87
fly-by-wire systems, 196–98, 272 high fineness ratio, 45–46, 138
transformation of technology, tools, and X-3 design, 136, 137, 138
procedures for supersonic era, v XF-92A design, 242, 245, 247
XF-92A system, 245, 247, 249, 253, 257,
259 G
fly-by-wire flight control systems, 196–98, 272 General Electric “black” aircraft program, 7
flying-wing aircraft, 167–68, 170, 171–74, General Electric I-A Whittle turbojet engines,
196–97 7–8
316
Index
General Electric J85 engines, 160n7 D-558-2 pitch-up characteristics, 92, 93

General Electric TG-180 axial flow turbojet X-4 flights, 183, 185
engine, 10, 46, 48, 50, 52–53, 58, 74, 85 X-4 operational concerns, 178
Germany Grumman F-14/F-14A Tomcat aircraft, 156,
aircraft development in, 6, 168 236, 272
delta-wing aircraft, 244–45 Grumman XF10F-1 Jaguar fighter, 235, 236
jet engine research and development in, 5
rocket propulsion development in, 6, 10 H
swept-wing research and development in, Hallion, Richard P., 161n10, 236
20, 39–40n43, 85 Heinemann, Edward Henry “Ed”
Gilruth, Robert R. D-558 program plan and goals, 12, 85
contributions to X-plane and space D-558-1 design role, 45, 46, 48
programs, 275 D-558-1 landing gear and tires, 47
photo of, 14 Heinkel, Ernst, 5
wing-flow technique, invention of by, 14, 18 Heinkel He 178 aircraft, 5
XS-1 horizontal stabilizer, elevator, and tail helmets and D-558-1 canopy design, 53,
design, 38n35 79n18
XS-1 wing design, 14–15 Hemphill, Thomas, 244
Glennan, T. Keith, 274 High-Speed Flight Station/High-Speed Flight
Gloster Meteor aircraft, 49 Research Station
Goodlin, Chalmers H. “Slick” dress code at, 121
experience of, 23 establishment of, 23–24, 23
XS-1 flight tests, 23, 24, 27, 32, 42n83 facilities at, 24
XS-1 program publicity, 26–27 housing near, 24
XS-1 supersonic flight by, 27, 28 Mercury program role, 274–75
Goodrich Corporation, 47 Muroc Flight Test Unit, 23–24, 23, 34, 50,
Gough, Melvin N. 275
photo of, 26 name change of, 132n55
XS-1 flight-test program direction and research culture at, 274–75
goals, 25, 26 X-planes at, 199
XS-1 test site tour, 27 High Speed Tunnel, 8-Foot, 76
Great Britain High-Speed Wind Tunnel, 7-by-10-Foot, 138,
jet engine research and development in, 5, 206, 208
7, 8, 9 Hilton, W.F., 1
sound barrier, first British aircraft to exceed, Hoover, Herbert, 42n83
168 Hoover, Robert A. “Bob,” 29, 32
Greene, Carl, 11 Horton Ho 229 aircraft, 196–97
Griffith, John H. Hunsaker, Jerome, 20
Chance Vought job, 130n38 Hyatt, Abraham, 12
D-558-1 flights, 59, 62–63, 64 hydrogen peroxide accident, 131–32n54
317
Probing the Sky
hypersonic flight jet-assisted-takeoff (JATO) solid rocket

knowledge, dedication, expertise, and boosters, 89, 89
courage for development of, vi jet engines
transformation of technology, tools, and aircraft design for, 8, 37n24
procedures for, v–vi, 269 applications for, complacency about, 4, 5,
hypoxia, 228–29 17
applications for and benefits of, reevaluation
I of, 5–6
inertial coupling (roll coupling), 147, 150–56, axial-flow turbojet engines, 5
157–59, 163n38, 233–34, 272 centrifugal-flow jet engine, 5
Irvine, Jack, 244 development of and aircraft development,
Italy, jet engine research and development in, 5 139, 160n7
ducted-fan designs, 5, 8
J early concepts of, 4–8, 36n13
J33 engines, 247, 248, 250–51, 254, gas turbine engines, importance of
262–63n10 development of, 4
J34 engines hybrid piston engine/jet concepts, 4–5
D-558-2 engines, 88, 95, 96, 98 Jeep piston-engine/combustion-chamber
P.1101 engine test program proposal, 210 engine, 6–7, 7, 8, 37n21
performance of, 160n7 limitations of early engines, vi
X-3 engines, 135, 139–40 limitations of research with, 11
J35 engines, 210, 211–12 materials for, 6
J46 engine, 135, 139–41, 157, 160n7, 210 ramjets, 244, 246, 247
J47 engines, 210 research aircraft with turbojet propulsion,
J57 engines, 153 study of, 9–12
J58 engines, 140 transformation of aircraft development
J75 engines, 140 and, v–vi
JA-37 Viggen aircraft, 264–66n31 See also specific engines
Jacobs, Eastman N. Jet Propulsion Laboratory, California Institute of
dispute between Theodorsen and, 19 Technology, 274
Jeep engine–powered aircraft research, Johnson, Clarence L. “Kelly,” 272–73, 277n4
6, 7 Johnson, Richard L., 185, 253
turbojet engine and aircraft research, Johnson Space Center (Manned Spacecraft
support for, 8, 37n24 Center), 275
wing design and drag, Volta Congress Jones, Robert T.
presentation about, 39–40n43 delta-wing research, 244
Jansen, George, 98 dispute between Theodorsen and, 19
JAS-39 Gripen aircraft, 264–66n31 swept-wing research paper, distribution of,
Jeep piston-engine/combustion-chamber 19, 20
engine, 6–7, 7, 8, 37n21
318
Index
swept-wing theoretical analysis and Langley Memorial Aeronautical Laboratory

calculations, 17–18, 19–20, 39–40n43, adjustment for personnel transferring to
72, 85 Muroc for XS-1 program, 24
XS-1, swept-wing configuration for, 20–21 Full-Scale Tunnel, 245
Jones, Walter P. High Speed Tunnel, 8-Foot, 76
D-558-1 flights, 64, 65 High-Speed Wind Tunnel, 7-by-10-Foot,
D-558-2 flights, 97, 102 138, 206, 208
death of, 130n38, 192 jet engine research by, 5–7
education of, 64 research culture at, 274
Northrop test pilot role, 130n38, 192 space research by, 274
X-4 flights, 186, 188, 190–91, 192 wind tunnel testing breakthroughs, 74–76
X-5 flights, 217, 218 XS-1 flight-test program direction and
goals, 30
K XS-1 program control, 16
Kent, Marion, 23 XS-1 test site considerations, 16, 21, 22,
Killian, James R., Jr., 273–74 25–26, 27, 40n52
Kinkaid (mechanic), 131–32n54 Lewis, George W.
Kleckner, Harold F., 152 Langley jet engine research, support for, 6
Klein, P.B., 30 transonic research aircraft, role in
Korean War, 81n44, 81n46, 146 development of, 8, 9–10
Kotcher, Ezra XS-1 program role, 21
Mach 0.999 study, 10–12, 42n83 Lewis Research Center, 274
turbojet-powered aircraft, limitations of life-support systems, v
research with, 11 lift
XS-1 program role, 12, 42n83 swept wings and, 17–20, 207
Kuhn, Thomas, v X-4 lift-over-drag studies, 190–92, 195
X-5 handling characteristics, 225
L Lilly, Howard C. “Tick,” 42n83, 49, 50–53, 56,
Land, Edwin, 274 57, 74, 170, 276, 277–78n13
landing gear Lippisch, Alexander, 244–45
D-558-1 landing gear configuration and Lippisch DM-1 glider, 244–45
problems related to, 47, 48, 51 Lockheed A-12 program, 140, 264–66n31
D-558-2 landing gear, 87, 136 Lockheed Blackbird aircraft, 264–66n31
wheel and tire size, 47 Lockheed F-104 Starfighter. See F-104
X-3 landing gear, 136 Starfighter aircraft
X-3 tire failures, 143, 145 Lockheed F-117A stealth fighter, 198
X-4 landing gear lock and landing gear Lockheed P-38 Lightning aircraft, 6–7, 9, 57,
door issues, 170, 176, 177, 178, 179, 86
180, 183 Lockheed P-80/P-80R aircraft, 16, 27, 37n24,
X-5 landing gear, 212 49, 247
XF-92A landing gear, 247, 256 Lockheed XF-90 aircraft, 160n7
319
Probing the Sky
Lockheed XST Have Blue stealth aircraft, 198 May, Eugene F. “Gene”
LR-8/XLR-8/XLR-11/RMI 6000C4 rocket D-558-1 canopy design and crash helmets,
engines 79n18
D-558-2 use of, 12, 85–86, 87, 89–90, D-558-1 flights, 48–49, 49, 57, 58–59
95, 96, 97, 114–15, 114 D-558-2 flights, 89, 95
nozzle extensions, 114–15 Mayer, John, 33
X-3 use of, 140, 157 McCormack, James, 273
X-15 use of, 140 McDonnell F3H-1 Demon aircraft, 160n7
XS-1 use of, 22–23 McDonnell F-101 Voodoo aircraft, 156
Lusk, Arthur, 169 McKay, John B. “Jack”
Lyndon B. Johnson Space Center, 275 astronaut wings, qualification for, 126
D-558-1 flights, 71
M D-558-2 air launch and P2B-1S engine
M-21 aircraft, 264–66n31 failure, 118
Mach 0.999 study, 10–12, 42n83 D-558-2 flights, 118, 123, 126
Mach 1 flight, 27–31, 32–34, 42n83, education of, 71
81–82n51, 88, 114 experience of, 71, 126
Mach 2 flight, 110, 112–17 external stores research flights, 120, 122
Mach 2 sustained speed performance, 135, X-4 flights, 194, 205n70
160n7 X-5 flights, 231–32
Mach numbers, altitude, and air temperatures, X-15 flights, 126
78n12 XS-1 flights, 126
Manned Spacecraft Center (Johnson Space McTigue, John, 35n2, 157
Center), 275 MD 550 Mirage I aircraft, 264–66n31
Marietta Field, 16 MD 550 Mirage II aircraft, 264–66n31
Marine Corps Air-Ground Museum, 81–82n51 MD 550 Mirage III aircraft, 264–66n31
Marschak, Thomas A., 139 MD 550 Mirage IV aircraft, 264–66n31
Marshall Space Flight Center (Army Ballistic Mercury program, 274–75
Missile Agency), 274 Messerschmitt Me 163 Komet rocket fighter,
Martin, Bill, 249 39–40n43, 166, 167–68, 169, 209, 244
Martin, John F., 88, 89, 128n6 Messerschmitt Me 262 jet fighter, 39–40n43
materials Messerschmitt P.1101 aircraft, 209–10, 235
evolutionary process of aircraft development Meyer, Corwin H. “Corky,” 235
and changes in, v, 3–4 Meyer, J.C., 216
transformation of technology, tools, and MiG-15 aircraft, 81n44, 81n46
procedures for supersonic era, v, 3–4 MiG-19 aircraft, 38n35
wood-and-fabric to all-metal aircraft, MiG-23/27 aircraft, 236, 272
transition to, 4, 35–36n12 MiG-25 aircraft, 156
Mattson, Axel T., 33 military services
Maxwell Air Force Base, 195 aircraft design role of, 35–36n12
320
Index
aircraft designation system, 37–38n32, National Advisory Committee for Aeronautics

38n34 (NACA)
research aircraft, funding for, 8 airfoil research, support for, 3, 4
transonic research aircraft, opinions about communication between Air Force and, 230
need for, 11 creation of, 3
transonic research aircraft, role in duration of existence, 241n60
development of, 9–12 evolutionary process of aircraft
missiles development, role in, 3–4
nuclear weapon–armed rockets, attitudes focus of research on practical applications,
toward practicality of, 37n21 3–4
surface-to-air missiles (SAMs), 262n7 human contributions of X-plane programs,
V-2 missiles, 136 275
Model CL-1200 Lancer aircraft (X-27 aircraft), laboratory/research center structure, 274
277n4 mission and purpose of, 3
model testing name change of, 241n60
delta-wing concept testing, 248 propulsion research by, 4
drop-body testing, 18–20 Research Aircraft Projects Panel, 140–41
inertial coupling behavior, 151–52 space research by, 273–74
size of models for wind-tunnel testing, Special Committee on Jet Propulsion, 6
74–75 swept-wing concepts, failure of NACA to
sting for mounting models, 75 develop, 20–21, 72
XS-1 model, 32–33 variable-sweep wing studies, 208, 235–36,
See also wing-flow technique 237n3
Moise, Jack, 131–32n54 X-planes research plan, 71–72
motorjet, 36n13 National Aeronautics and Space Administration
Munk, Max, 17, 39–40n43 (NASA)
Muroc Army Air Field. See Edwards Air Force laboratory/research center structure, 274
Base/Muroc Army Air Field mission and purpose of, 3, 241n60
Muroc Flight Test Unit, 23–24, 23, 34, 50, 275 NACA merger and creation of, 3, 241n60,
Murray, Arthur A. “Kit,” 221, 227–29, 231, 253 274
museum-display aircraft, 81–82n51, 119, space research by, 273–74
133n63, 157, 166, 195, 232, 257 National Air and Space Museum, 81–82n51,
MX-324 glider, 171 119, 133n63
MX-656 project, 135, 160n2. See also X-3 National Museum of Naval Aviation, 81–82n51
Stiletto National Museum of the Air Force, 166
MX-810 project, 169. See also X-4 Bantam National Science Foundation, 274
Navy, U.S.
N aerial refueling operations, 239n33
N-9M aircraft, 196–97 aircraft designation system, 38n34
nacelle design and location, 4 D-558 program funding from, 44
321
Probing the Sky
D-558 program role, 12–14, 72–73

transonic research aircraft, role in P
development of, 9–12 P-38 Lightning aircraft, 6–7, 9, 57, 86
variable-sweep aircraft used by, 236 P-39 Airacobra aircraft, 7, 9
Nord 1500-02 Griffon II fighter, 244, P-47 Thunderbolt aircraft, 7, 9
264–66n31 P-51 Mustang, viii, 18, 19, 20
North American Aviation, Crossfield role at, P-63 aircraft, 247
117–18, 126 P-80/P-80R aircraft, 16, 27, 37n24, 49, 247
North American Aviation XB-70 Valkyrie aircraft, P-84 aircraft, 27–28
264–66n31 P.13 fighter, 244
North American FJ-1 aircraft, 247 P.111 aircraft, 264–66n31
North American X-15 aircraft. See X-15 aircraft P.120 aircraft, 264–66n31
Northrop, John Knudsen “Jack,” 167 P.1101 aircraft, 209–10, 235
Northrop B-2A Spirit stealth bomber, 196 Panavia Tornado strike aircraft, 236
Northrop F-5/T-38 aircraft, 159, 160n7, 272 paradigm shift, v–vi
Northrop N-9M aircraft, 196–97 Phillips, William Hewitt “Bill,” 150–52, 156,
Northrop SM-62 Snark aircraft, 197 163n38
Northrop X-4 aircraft. See X-4 Bantam pilot-escape systems/ejection seats
Northrop XB-35 flying-wing bomber, 167, 171 capsule systems, 47, 77n7
Northrop XP-79B flying-wing fighter, 170, 171, D-558-1 Skystreak, 47, 77n7
196–97 D-558-2 Skyrocket, 47, 77n7
Northrop YB-49 flying-wing bomber, 167, 172, human limits and design of, 47, 77n7
196–97 transformation of technology, tools, and
Northrop YF-89D Scorpion aircraft, 192 procedures for supersonic era, v
nuclear weapons X-2 aircraft, 47, 77n7
A-bomb, Soviet development of, 37n21 X-3 Stiletto, 153
nuclear weapon–armed rockets, attitudes X-4 Bantam, 173
toward practicality of, 37n21 X-5 aircraft, 212
Nugent, Jack, 103–4 XF-92A aircraft, 245–46, 247
Pilotless Aircraft Research Division (PARD), 152
O pilots/test pilots
oscillations accidents and loss of, 77n7, 227–31,
D-558-2 dynamic-lateral-stability 275–76, 277–78n13
characteristics, 90–91, 99, 101, 126 age of when breaking the sound barrier, 2,
Dutch roll lateral-directional oscillation, 88 35n2
pilot-induced oscillations, 174 altitudes and speeds experienced by,
supersonic yaw, 126 274–75
X-4 characteristics, 180 changes in D-558-2 program, 130n38
courage of, vi
dangers faced by, 64
322
Index
death of and naming of street after, 263n18

demands of test programs and operations R
on, 74 R-1535 radial piston engine, 6
education of, 64, 65, 66–67, 71 ramjets, 244, 246, 247
experience and skills of, vi, 64, 66–67, 71, Reaction Motors Incorporated 6000C4 engine.
80n43, 88, 126 See LR-8/XLR-8/XLR-11/RMI 6000C4 rocket
reluctance to fly aircraft, 128n6 engines
voice of when breaking the sound barrier, Reeves Electronic Analog Computer (REAC),
2, 35n2 252
Pinecastle Army Air Field, 21, 22, 24, 26, refueling formation test flights and aerial
40n52 refueling, 221–24, 239n29, 239n33
pitch and pitch control Reid, Henry
D-558-2 pitch-up characteristics, Jones swept-wing research paper,
91–94, 97, 102–9, 125–26, 125, 127, distribution of, 20
133n64, 157 Langley jet engine research, support for,
X-4 characteristics, 167, 170, 171–72, 5–6
179, 186–87, 202n46 XS-1 flight-test program direction and
XS-1 aircraft, 32–33 goals, 25–26, 30
Planes of Fame Museum, 133n63 Republic P-47 Thunderbolt aircraft, 7, 9
point-defense mission, 246–47, 262n7 research aircraft/X-series research aircraft
Popson, Raymond A. “Ray,” 227–31, 276, accidents with, 77n7, 227–31, 275–76,
277–78n13 277–78n13
Pratt & Whitney engines “black” aircraft program, first, 7–8
J58 engines, 140 contributions to aviation research, vi,
J75 engines, 140 269–76
R-1535 radial piston engine, 6 cost of research activities, 113
President’s Scientific Advisory Committee criticism of X-planes, 272–73
(PSAC), 273–74 delays in research programs, 175
pressure-distribution research demands of test programs and operations,
D-558-1 flights, 62–63, 73 74
X-3 flights, 156 designation system for numbering,
propellers 37–38n32, 38n34
airfoil design and efficiency of, 2–3, 4 engine test firings, 116
research on high-speed aerodynamics, 2–3 fates of and perceived importance,
propulsion 81–82n51
NACA research on, 4 funding for, 8
supersonic flight and, 34 human contributions of X-plane
transformation of technology, tools, and programs, 275
procedures for supersonic era, v–vi Jeep engine–powered aircraft, 6–7, 7, 8
Purcell, Edward, 274 jet engine development and aircraft
development, 139, 160n7
323
Probing the Sky
museum-display aircraft, 81–82n51, 119, rudder

133n63, 157, 166, 195, 232, 257 D-558-1 data collection and research
NACA research plan, 71–72 results, 54–55
opinions of military services about need D-558-2 rudder lock, 99
for, 11 X-4 rudder cable changes, 181, 182
photo of, 199 X-4 rudder control, 171–74, 197
public opinion about, 119 XF-92A rudder-pulse tests, 255
Round One aircraft, vi, 269–73, 274–76
Round Two aircraft, vi, 273–74 S
war and development of, 8 Saab 35 Draken aircraft, 264–66n31
See also specific aircraft Saab JA-37 Viggen aircraft, 264–66n31
Richardson, E.W., 180, 181 Saab JAS-39 Gripen aircraft, 264–66n31
Ridley, Jackie L. “Jack,” 21, 29, 32, 33–34, 253 Sadoff, Melvin, 177–78, 179, 180, 187
Ritland, Osmond J., 28 safety procedures and transformation of
Robert J. Collier Trophy, 42n83, 76 technology, tools, and procedures for
rocket-powered aircraft supersonic era, v, 3–4
air-launching of, 11 Salina Field, 16
contributions to aviation research, v–vi Samet, Arthur, 23
German development of, 6, 10 SB2C Helldiver aircraft, 7, 9
limitations of research with, 11, 12 Schwartz, A.M., 169
research aircraft using rocket motors, study semi-tailless aircraft, 167, 168, 195–98,
of, 10–12 200n4, 235. See also X-4 Bantam
speed capabilities for research aircraft, 11 Shannon, Ellis D. “Sam,” 248, 253
Stack, Robert, opposition to, 9, 11, 12, Sherman, Albert E., 5
42n83 Shick, Ralph H., 244
supersonic regime research with, vi shock-stall, 14
Rogers Dry Lake shock waves
conditions at, 22 fuselage design to avoid shock waves near
flooding of, 51, 69, 102, 130n31, wings, 45–46
142–43, 175 XS-1 horizontal stabilizer, elevator, and tail
XS-1 test site, 22–24 design, 38n35
rolls and roll behavior Sleeman, William C., 208
D-558-1 data collection and research SM-62 Snark aircraft, 197
results, 59–61 Smith, A.M.O., 45, 46, 85
inertial coupling (roll coupling), 147, Smith, George F., 21–22, 25, 28–29
150–56, 157–59, 163n38, 233–34, 272 Smith, R.G., 45
Root, L. Eugene “Gene,” 12, 45 Smith, Stan, 16, 38–39n41
Rosamond, 116 Sopwith Camel, 270
Round One aircraft, vi, 269–73, 274–76 Soulé, Hartley, 23, 27, 30–31, 70, 73, 152, 275
Round Two aircraft, vi, 273–74 sound barrier
coining of phrase, 1
324
Index
conditions for pilots during flights to Spirit of St. Louis, 81–82n51

break, 32 Sputnik, 241n60, 273
first British aircraft to exceed, 168 SR-71 aircraft, 264–66n31
mythology surrounding, 1–2, 11, 34, 35n2 stability and control
opinions about breaking, 32, 135 altitude and, 56
research aircraft to collect data on, 9 D-558-1 data collection and research
XS-1 flight to break, 32–34, 42n83 results, 53–57, 59–71, 73
South Base D-558-1 handling characteristics, 270
D-558-1 aircraft at, 44 D-558-1 lateral stability and aileron-
D-558-2 aircraft at, 84, 86, 110 effectiveness studies, 69–70, 73
hangar at, iv, 13 D-558-2 dynamic-lateral-stability
training facility role, 86 characteristics, 90–91, 99, 101, 126
Spaatz, Carl, 29 D-558-2 handling characteristics, 90–94,
space program, start of U.S., 273–75 97, 99–101, 102–12, 270
Space Shuttle Program, 275 D-558-2 lateral stability and aileron-
Space Task Group, 275 effectiveness studies, 112
Sparks, Ralph, 181 D-558-2 pitch-up characteristics,
Special Assistant to the President for Science 91–94, 97, 102–9, 125–26, 125, 127,
and Technology, 273 133n64, 157
speed data collection methods, 252
aerodynamic rules and increases in, v external stores research flights, 119–23,
compressibility effects during high-speed 119, 127, 132n60, 271
flight, 3, 6–7 inertial coupling (roll coupling), 147,
D-558-1 speed capabilities, 45, 68, 71–74 150–56, 157–59, 163n38, 233–34, 272
D-558-2 speed capabilities and speed semi-tailless aircraft, 167, 168, 195–98
records, 84, 99–100, 102, 109, 111, Sopwith Camel, 270
112–17 tailless aircraft, 167–68
evolutionary process of aircraft development variable-sweep-wing aircraft, 208
and increase in, v, 3–4 X-3 handling characteristics, 142–46,
F-86 speed capabilities, 146 148–50, 157–59, 270
structural heating and, 139 X-3 inertial coupling research, 147, 151,
X-3 speed capabilities, 140, 146, 271 152, 153–56
X-3 tire failures and, 143 X-4 handling characteristics, 167, 171–74,
XF-92A speed capabilities, 250, 251 176, 177–78, 179–80, 183–92, 193,
speed records 194, 196–98, 270
D-558-1 flights, 48–50, 78n12 X-5 handling characteristics, 215–26,
Donaldson record, 49 235, 271
F-86A flights, 78n12 X-5 inertial coupling behavior, 233–34
requirements for official record, 49 XF-92A handling characteristics, 248–49,
speed-record flights, 99–100, 113 251–57, 258–59, 271
spin and spin control. See stall/spin behavior XS-1 data collection, 25, 26, 30–31
325
Probing the Sky
stabilizers X-4 characteristics, 171–72, 177, 178, 183

adjustable and flight control, 38n35 X-5 characteristics, 217–18, 220–21,
all-moving one-piece design, 138 227–31, 229, 235
D-558-1 data collection and research XF-92A characteristics, 249
results, 54–57, 64, 67–69 Standards, U.S. Bureau of
D-558-1 stabilizer and tail design, 38n35, airfoil research by, 2–3
46, 67–69, 70–71, 138 jet engine research by, 4
D-558-2 stabilizer and tail design, 88, Stanley, Robert, 25, 27–28
108–9, 128n5, 138 Strass, H. Kurt, 203n50
F-86 all-moving tail, 68, 81n46 Strategic Air Command, 239n33
X-3 stabilizer and tail design, 138 structures
XS-1 stabilizer, elevator, and tail design, NACA research on, 4
32–33, 38n35, 138 speed and structural heating, 139
Stack, John transformation of technology, tools, and
Collier Trophy award for slotted tunnel, 76 procedures for supersonic era, v
D-558 program role, 8–9, 9, 42n83, 71–72 Su-17/22 aircraft, 236
D-558-1 swept-wing design, 85 Su-24 aircraft, 236
photos of, 9, 10 subsonic flight
rocket-powered aircraft, opposition to, 9, evolutionary process of aircraft development
11, 12, 42n83 for, v, 3–4
slotted tunnel development, 75–76 knowledge, dedication, expertise, and
transonic research aircraft, development of, courage for development of, vi
8–9, 10 validity of information and concepts in
turbojet-powered aircraft, commitment to supersonic era, v, vi
acquire, 12 supersonic flight
Vickers Swallow supersonic airliner role, aircraft design for, 34, 191
236 dangers of and mystery surrounding, 1–2
XS-1 horizontal stabilizer, elevator, and tail reliability and validity of information and
design, 38n35 concepts for, vi
XS-1 program control, 16, 21 research aircraft development to study,
XS-1 project and Collier Trophy, 42n83 71–72
XS-1 specifications for research routine nature of, 269–70
requirements, 15–16, 21–22 transformation of technology, tools, and
XS-1 test site considerations, 16 procedures for supersonic era, v–vi, 3–4,
XS-1 wing design, 14–15 269–70
stall/spin behavior wing-flow technique for research, viii
D-558-1 characteristics, 62 XS-1 flights, 26–27, 28–31
D-558-2 characteristics, 62, 92 supersonic transport (SST), 236, 264–66n31
flying-wing aircraft, 171–72 surface-to-air missiles (SAMs), 262n7
sweeping wings during, 228 Swallow aircraft (de Havilland DH 108),
X-3 characteristics, 145 168–69, 170, 196–97
326
Index
Swallow supersonic airliner, 235–36 D-558-1 data collection and research

Sweeney, William “Bill,” 12 results, 54–57, 64–65, 66, 67–69
swept wings D-558-1 stabilizer and tail design, 38n35,
advantages of, 17–18, 207 46, 67–69, 70–71, 138, 281
carrier landings and, 207 D-558-1 tail-loads studies, 69, 70–71,
concepts using, failure of NACA to develop, 73–74
20–21, 72 D-558-2 stabilizer and tail design, 88,
D-558-2 design, 73, 85–88, 124–26, 125, 108–9, 128n5, 138, 281
137–38, 281 D-558-2 tail-loads studies, 73–74, 127
drop-body testing of concept, 18–20 F-86 all-moving tail, 68, 81n46
German research and development of, failure of during high-speed dives, 6–7
20, 85 inertial coupling behavior and design of,
lack of compressibility effects with, 17, 156
39–40n43 semi-tailless aircraft, 167, 168, 195–98,
lift, drag, and, 17–20, 207 200n4
limitations of, 207 strengthening of tail structures, 7
pitch-up characteristics, 91–94, 125–26, tailless aircraft, 39–40n43, 167–68
125, 127, 133n64, 157 wind tunnel testing and X-3 tail design, 138
sawtooth leading-edge extension, X-3 stabilizer and tail design, 138, 281
106–8, 107 X-4 semi-tailless design, 166, 167, 235, 281
stability and control and, 39–40n43 X-4 tail, structural integrity of, 181, 182,
theoretical analysis and calculations, 185
17–18, 19–20, 39–40n43, 72, 85 X-5 tail design, 281
wing fences, 88, 97, 102–6, 103, 108, 256 X-5 tail-loads studies, 73, 232–33
wing-flow testing of concept, 18–20, 19, XF-92A design, 244, 247
20 XF-92A tail design, 281
wing slats, 104–6, 107, 108 XS-1 stabilizer, elevator, and tail design,
X-2 design, 27, 137–38 32–33, 38n35, 138, 281
X-3 design, 136, 137–38, 137, 158, 281 technology
X-3 design and compressibility effects, changes in, 198
137–38 conservative attitudes toward, 37n21
XS-1, swept-wing configuration for, 20–21, process for technological and scientific
27 breakthroughs, 39–40n43
Symington, Stuart, 34 transformation of for supersonic era, v–vi,
3–4, 269–70
T TG-180 axial flow turbojet engine, 10, 46, 48,
T-38/F-5 aircraft, 159, 160n7, 272 50, 52–53, 58, 74, 85
tails Theodorsen, Theodore, 19, 20
all-moving slab tail, 138 Thiokol XLR-99 engines, 140
all-moving tail and flight control, 38n35, Thompson, Floyd, 14–15
68, 81n46 Thompson Trophy air races, 23
327
Probing the Sky
time, breaking the sound barrier and reversal Swallow supersonic airliner, 235–36
of, 2, 35n2 symmetrical wing sweeping, 208
transonic flight wing-pivot system, 208, 272
loitering in transonic regime, 45 X-1 model testing of concept, 206, 208
research aircraft development to study, X-5 wing design, pivot point, and sweep
8–12, 71–74 angle, 210–11, 211, 222, 236,
research on, onset of, 2–4 237–38n7, 281
thrust capacity for, 85–86 See also X-5 aircraft
wing-flow technique for research, viii Vensel, Joseph “Joe,” 21, 23, 115
Truszynski, Gerald “Gerry,” 21, 23, 50, 275 Vickers Swallow supersonic airliner, 235–36
Tsien, Hsue-Shen, 17, 39–40n43 Victorville, 116
Tu-22M aircraft, 236 Voigt, Waldemar, 209
Tu-144 aircraft, 264–66n31 von Kármán, Theodore, 39–40n43
Tu-160 aircraft, 236, 272 von Ohain, Hans, 5
Tucker, Charles, 170–71, 172, 173–75, 176, vortex generator research
177, 178, 179, 180 D-558-1 flights, 62–63, 73, 80n37
Turcat, Andre, 264–66n31 XS-1 flight, 80n37
Turner, Harold, Jr., 9 Vought XF8U-1 Crusader aircraft, 235
Voyles, James, 27
U Vulcan strategic bomber, 264–66n31
Ulmer leather gaskets, 126–27 Vultee BT-13 aircraft, 247
University of the South, 257
unmanned aerial vehicles (UAVs), 264–66n31 W
Walker, Joseph A. “Joe”
V astronaut wings, qualification for, 126
V-2 missiles, 136 D-558-1 flights, 65–66
Van Every, Kermit D-558-2 flights, 118, 126
D-558-1 design role, 45 education of, 65
D-558-2 design, 125 experience of, 65, 126
D-558-2 design role, 86, 87–88 formation test flights, 222–24
Vandenberg, Hoyt S., 42n83 inertial coupling research missions, 153–55
variable-sweep wings X-3 flights, 149–50, 153–55, 156, 157
advantages and value of, 208, 235, 236 X-4 flights, 189, 190
aircraft built using concept, 236 X-5 accident panel testimony, 230
asymmetrical oblique wing sweeping, 208 X-5 flights, 216, 217–20, 231
carrier landings and, 207–8, 235 X-5 handling characteristics, 230
limitations of, 208 X-15 flights, 126
NACA/NASA studies of, 208, 235–36, XS-1 flights, 126
237n3 Wallis, Bruce, 235–36
stability, control, and handling Weick, Fred, 4
characteristics, 208 Weil, Joe, 156
328
Index
Welch, George “Wheaties,” 153 slotted tunnel development, 75–76

Wendover Field, 16 speed and aircraft design research in, v
Westinghouse engines sting for mounting models, 75
24C turbojet engines, 12, 85, 86, 87, 95 transformation of technology, tools, and
J30 engines, 170 procedures for supersonic era, v
J34 engines. See J34 engines X-3 tail design testing, 138
J46 engine, 135, 139–41, 157, 160n7, 210 wing fences
jet engine development by, 160n7 delta-wing aircraft, 255, 256, 257
Whitcomb, Richard T., 79–80n27, 259, 261, swept-wing aircraft, 88, 97, 102–6,
264–66n31, 272 103, 108
Whittle, Frank, 5, 7 wing-flow technique
Whittle engine, 5, 7 concept of, 18
Williams, Walter C. development of, 14, 18
aerial refueling formation test flights, 221 model use in adaptation of, viii
contributions to X-plane and space swept-wing concept testing, 18–20, 19, 20
programs, 275 transonic and supersonic airflow research
D-558-1 paint schemes, 51 with, viii
D-558-1 research results, 53–57 winglet, 57, 79–80n27
D-558-2 program role, 115 wings
housing for near Muroc, 24 D-558-1 data collection and research
inertial coupling research, 156 results, 54–57, 59–64, 67, 81n44
Mach 2 flight, approval for, 113 D-558-1 design and materials, 46–47, 281
Muroc Flight Test Unit role, 21, 23–24, 23 D-558-2 wing-strain-gauge calibration, 91
painting of D-558-1 control surfaces, 58 endplates, 57
X-4 equipment concerns, 181 failure of during high-speed flight, 3
X-4 flight data reports, 173–74, 186–87 flaps on, 88
XF-92A program role, 254 fuselage design to avoid shock waves near
XS-1 flight-test program direction and wings, 45–46
goals, 25, 26, 27, 28, 41n63 leading-edge slats, 88
XS-1 program publicity, 27 reverse taper design, 88
wind tunnels strengthening of wing structures, 7
choking in, 74–76, 124–25 swept-wing aircraft, 256
Collier Trophy award for slotted tunnel, 76 trailing-edge thickness, 188–89, 203n50
Full-Scale Tunnel, 245 wet wings and fuel tanks, 47
High Speed Tunnel, 8-Foot, 76 X-4 design and material, 170, 188–89,
High-Speed Wind Tunnel, 7-by-10-Foot, 203n50, 281
138, 206, 208 X-4 wing flaps on trailing edge, 170
limitations of research with, 1, 71–72, XS-1 wing design, 14–15, 20–21, 24, 27,
74–75, 269 46, 137–38, 281
Model 7002 wind tunnel testing, 248 See also delta-wing aircraft; swept wings;
size of models for testing in, 74–75 variable-sweep wings
329
Probing the Sky
wingtip vortices, 221–24 flight tests, 21–24, 32–34, 42n83

Wolf, Robert W., 9–10 flight test site for, 16, 21–24, 25–26, 27,
Wolfe, J.E., 253 40n52
Wolko, Frank, 276, 277–78n13 fuel system, turbopump, and fuel capacity,
Woods, Robert J,, 12, 209, 210 15–16, 21, 25–26, 41n63, 128n4, 280
Woolams, Jack fueling and handling facilities for, 26
death of, 23 funding for, 28
XS-1 flight tests, 22, 32 inertial coupling behavior during falling
XS-1 test site considerations, 16 bodies test, 151–52
Wright, Ray H., 75 Mach 1 flight, 32–34, 42n83, 81–82n51,
Wright brothers, v, vi, 1, 269 88, 114
Wright Field, 10–12 Mach 1 flight goal, 27–31
Wright Flyer, 81–82n51 model testing, 32–33
Wright-Patterson Air Force Base, 195, 232, 257 NACA, transfer to, 50
negotiations to build, 12
X paint schemes, 250
X-1/XS-1 (Experimental Sonic-1) aircraft performance potential, 128n4
accidents with, 22, 40n52, 277–78n13 personnel involved in program, 21
air-launching from B-29 aircraft, 15, 25, pitch control, 32–33
27, 32, 33, 53, 124 publicity about program, 26–27, 34, 42n83
altitude for flight tests, 15, 16, 27–28, 29, reports about program, 34, 42–43nn84–85
30–31 specifications, technical, 280–81
canopy design, 88 specifications for research requirements,
conference about results of program, 34 15–16, 21–22
contributions to aviation research, 269, 271 speed requirements for research, 15–16
control of program, 16, 28–30, 34 stability and control data collection, 25, 26,
D-558-1 compared to, 45 30–31
data collection and research results, 54, 56, stabilizer, elevator, and tail design, 32–33,
57, 73–74 38n35, 138, 281
design of, 14–15, 45, 139, 199 supersonic flight capabilities of, 26–27,
designation system for numbering, 28–31
37–38n32 survival of aircraft at end of program,
development of, decisions leading to, 8–12 126–27
display of aircraft at end of program, takeoff and landing of, 15–16, 21, 25, 27,
81–82n51 41n63
fate of and perceived importance, transport to Muroc attached to B-29
81–82n51 aircraft, 24
flight-test program direction, plan, and turbojet propulsion for, vi
goals, 25–31, 41n63, 72, 74 vortex generator research flight, 80n37
weather conditions for testing, 22
330
Index
wing design, 14–15, 20–21, 24, 27, 46, Mach 2 sustained speed performance, 135,
137–38, 281 160n7
X-1A inertial coupling flight, 152, 157 mockup of, 134
X-2/XS-2 aircraft NACA research flights, 135–36, 147–50
loss of aircraft, 127, 277–78n13 number of research flights, 157
pilot-escape system, 47, 77n7 paint schemes, 141, 150
wing design, 27, 137–38 performance potential, 134, 136, 139,
X-3 Stiletto 141, 146
Air Force flights, 146–49 pilot-escape system, 153
all-rocket/air-launch configuration, 140–41 pressure-distribution research, 156
contract to develop, 135, 160n2 safety modifications, 74
contributions to aviation research, 135, shortcomings of, 270
157–59, 271–72 specifications, technical, 280–81
data collection and research results, speed and structural heating, 139
147–50 speed capabilities, 140, 146, 271
design of, 136–39, 137, 147, 151, 153, stability, control, and handling
157–59, 199 characteristics, 142–46, 148–50,
display of aircraft at end of program, 157–59, 270
157, 257 stabilizer and tail design, 138, 281
dive tests, 145–46 stall/spin behavior, 145
drag and design of, 136, 137–38 taxi tests, 142
engine-cooling arrangement, 136–37 thrust requirements, 136, 139–41, 147
engine failure during flight, 146 tire failures, 143, 145
engines for, 136–37, 139–41, 157 transport of to lakebed for takeoff, 143, 158
F-104 Starfighter development and, 141, wing design, 136, 137–38, 137, 158, 281
158–59, 161n10, 195, 272–73 X-4 Bantam
final flights and end of program, 156–57 aircraft for NACA, preference for, 181
flight-test program direction, plan, and canopy design and lock, 173, 183
goals, 135–36, 141, 157 contributions to aviation research,
flight tests, 135–36, 142–50, 153–56 195–98, 272
fuel system and capacity, 138, 280 design of, 166, 169–70, 169, 175, 182,
fuselage design, 136, 137, 138 198, 199, 209
inertial coupling research, 147, 151, 152, display of aircraft at end of program, 166,
153–56, 157–59, 272 195, 257
instrumentation for data collection, dive brakes, 176, 179–80, 181, 188–89,
147–48, 156 191, 193, 196, 203n53
J34 engines for, 135, 139–40 elevons, 167, 172, 183, 193, 194, 196
J46 engines for, 135, 139–41, 157, 160n7 engine malfunctions and replacements,
landing gear, 136 175, 176, 180, 182, 188, 189–90,
load feel control settings, 144 192–93
331
Probing the Sky
engines, access to, 177 contract to develop and build, 212

engines for, 170 contributions to aviation research, 213,
engines for, concerns about, 181 235–36, 272
final flights and end of program, 192–95 data collection and research results, 214,
flight status of both aircraft, 182 215–26, 232–34
flight-test program direction, plan, and design of, 199, 209–13, 209, 211, 227,
goals, 167, 169, 180–82 229, 231
flight tests, 170–80, 182–94, 205n70 display of aircraft at end of program, 232,
fuel-leak problems, 175–76, 182, 185, 189 257
fuel system and capacity, 170, 200n8, 280 dive tests, 219–20
instrumentation for data collection, 169–71, engines for, 211–12
174–75, 176–77, 181, 183, 185, 193 final flights and end of program, 231–34
landing gear lock and landing gear door flight characteristic differences between
issues, 170, 176, 177, 178, 179, 180, 183 aircraft, 229
lift-over-drag studies, 190–92, 195 flight-test program direction, plan, and
NACA research flights, 185–94 goals, 212–13
pilot-escape systems, 173 flight tests, 213–24, 226–34
rudder cable changes, 181, 182 formation test flights, 221–24, 239n29
rudder control, 171–74, 197 fuel-cell pressure gauge problems, 213–14
safety modifications, 74 fuel system and capacity, 212, 280
semi-tailless design, 166, 167, 195–98, inertial coupling (roll coupling), 233–34
235, 281 instrumentation for data collection, 214
shortcomings of, 270 jet fighter design based on, 210, 235
specifications, technical, 280–81 landing gear, 212
stability, control, and handling NACA, transfer to, 213
characteristics, 167, 171–74, 176, NACA research flights, 216–26, 231–34
177–78, 179–80, 183–92, 193, 194, number of research flights, 231
196–98, 270 pilot-escape systems, 212
stability, control, and handling safety modifications, 74
shortcomings, 186–88, 202n46 shortcomings of, 271
stall/spin behavior, 171–72, 177, 178, 183 specifications, technical, 212, 280–81
tail, structural integrity of, 181, 182, 185 speed brakes, 211, 226
wing damage, 193–94 stability, control, and handling
wing design and material, 170, 188–89, characteristics, 215–26, 235, 271
203n50, 281 stall/spin behavior, 217–18, 220–21,
wing flaps on trailing edge, 170 227–31, 229, 235
X-5 aircraft symmetrical wing sweeping design, 208
accident and death of Popson, 227–31, tail design, 281
277–78n13 tail-loads studies, 73, 232–33
aileron overbalance and effectiveness, 234 wing design, pivot point, and sweep angle,
Air Force flights, 213–16, 226–30 210–11, 211, 222, 236, 237–38n7, 281
332
Index
wing-root fillets, 219, 226 final flights and end of program, 256–57
wing slats, 211, 223, 226 flight control system, 245, 247, 249, 253,
X-15 aircraft 257, 259
air launching of, 140 flight-test program direction, plan, and
contributions to aviation research, vi, 273 goals, 254–55
design and development of, 272 flight tests, 248–57
display of aircraft at end of program, flying mockup for proposed interceptor, 243
81–82n51 fuel system and capacity, 246, 280
engines for, 140 funding for, 247
fate of and perceived importance, fuselage design, 242, 245, 247
81–82n51 improvement recommendations, 253–54
flight-test program direction, plan, and interceptor mission of, 246–47
goals, 124 jet power for, 247, 248, 250–51, 254,
X-27 aircraft (Model CL-1200 Lancer aircraft), 262–63n10
277n4 landing gear, 247, 256
X-43A aircraft, 139 Model 7002 designation, 244, 247–48
XB-35 flying-wing bomber, 167, 171 NACA, transfer to, 254
XB-47 jet bomber, 87 NACA research flights, 254–57, 255, 258
XB-52 bomber prototype, 198 paint schemes and bare-metal finish, 242,
XB-70 Valkyrie aircraft, 264–66n31 246, 250, 250, 258
XF8U-1 Crusader aircraft, 235 pilot-escape systems, 245–46, 247
XF10F-1 Jaguar fighter, 235, 236 propulsion system for, 244
XF-86 aircraft, 30 ramjet power for XP-92, 244, 246, 247
XF-90 aircraft, 160n7 refueling formation test flights, 221
XF-92A aircraft rocket power for XP-92, 244, 246, 247
Air Force flights, 246, 249–54, 262–63n10 rudder-pulse tests, 255
birth of, 243–44 shortcomings of, 271
canopy design, 247, 260 specifications, technical, 280–81
cart for takeoff run, 246, 247 speed capabilities, 250, 251
contract to develop, 247 stability, control, and handling
contributions to aviation research, 256, characteristics, 248–49, 251–57,
258–61, 264–66n31, 272 258–59, 271
data collection and research results, stall/spin behavior, 249
251–57 tail design, 244, 247, 281
delta-wing design, 244–45, 247, 281 taxi tests, 248–49
design of, 199, 242, 243–48, 245, 246, weapons aboard, 246
258, 260 wing fences, 255, 256, 257
display of aircraft at end of program, 257 XP-92 design and prototypes, 243–48
elevons, 247 XLR-8/XLR-11 rocket engines. See LR-8/
engine malfunctions and replacements, XLR-8/XLR-11/RMI 6000C4 rocket engines
250–51, 256, 262–63n10 XLR-99 engines, 140
333
Probing the Sky
XP-56 tailless fighter, 171

XP-59A jet, 7–8, 9, 11, 22, 37n24, 243
XP-79B flying-wing fighter, 170, 171, 196–97
XP-86 aircraft, 28, 87
XS-1 aircraft. See X-1/XS-1 (Experimental
Sonic-1) aircraft
XS-3 aircraft
jet-powered aircraft design, 27
Mach 1 flight goal, 27
X-series research aircraft. See research
aircraft/X-series research aircraft
XST Have Blue stealth aircraft, 198
Y
YB-49 flying-wing bomber, 167, 172, 196–97
YB-60 bombers, 140
Yeager, Charles E. “Chuck”
injury to falling from a horse, 33
X-1A inertial coupling flight, 152, 157
X-3 flights, 147, 147, 148–49
X-4 flights, 182
XF-92A flights, 249–51, 258
XF-92A handling characteristics, 258
XS-1 flight tests, 21, 29, 32–34
XS-1 Mach 1 flight, 32–34, 42n83,
81–82n51, 88, 114
XS-1 project and Collier Trophy, 42n83
YF-12 aircraft, 264–66n31
YF-89D aircraft, 130n38
YF-102/F-102 aircraft, 259–61, 259, 260,
261, 264–66n31
York, Herbert, 274
Z
Ziegler, Jean L. “Skip,” 213–14, 226, 276,
277–78n13
Zisfein, Melvin B., 161n10
334

Probing The Sky

Uploaded by

Copyright:

Available Formats

Probing The Sky

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Probing The Sky

Uploaded by

Copyright:

Available Formats

Curtis Peebles

Copyright © 2014 by the National Aeronautics and Space Administration.

This publication is available as a free download at

National Aeronautics and Space Administration

Introduction: Toward—and Into—the Unknown v

Chapter 1: Confronting the “Sound Barrier”: The Bell XS-1................................ 1

Chapter 2: Flying Test Tube: The Douglas D-558-1 Skystreak........................... 45

Chapter 3: Proving the Swept Wing: The Douglas D-558-2 Skyrocket.............. 85

Chapter 6: Transformative Pioneer: The Bell X-5............................................... 207

Chapter 7: Progenitor of the Delta: The Convair XF-92A................................... 243

Round One: A Reflection 269

Toward—and Into—the Unknown

In 1935, when British aerodynamicist W.F. Hilton inadvertently coined the

The Onset of Transonic Research

Wood-and-fabric biplanes were giving way to all-metal monoplanes, aircraft

The Turbojet Revolution

P-38 Lightnings, Republic P-47

Two Roads Toward the Transonic

aide Harold Turner, Jr. (Turner had

It appears quite possible to construct a single engine aircraft based

D-558-2. The Bureau of Aeronautics also outlined a development program that

correct what NACA representatives saw as shortcomings, the D-558-1 design

• Flexibility to permit changes of all principal components such as

Why Not Swept Wings?

The debate was settled by the end of

Flight Tests and Unresolved Issues

this concluded Bell’s obligations under the contract, company management

Settling Differences, Building a Program

“will not undertake to supply a pilot and

Langley’s personnel proved less than enthusiastic in the change of program

Toward—and Beyond—the Unknown

opposed to engineering requirements, behind the switch from

the Langley library. Jones incorporated a citation listing it in the

56. Hallion, Supersonic Flight, pp. 89–91.

82. From flight transcript, reprinted in Young, Meeting the Challenge of

1950); De E. Beeler and John P. Mayer, “Measurements of the

First Flights and Speed Records

Instrumentation Data Collected

Airspeed-altitude recorder Indicated airspeed and pressure altitude

Three-component accelerometer Normal, longitudinal, and transverse acceleration

Angular-velocity recorder Rolling velocity

Sideslip-angle recorder Sideslip measurement

Wheel-force recorder Aileron and elevator forces

Pedal-force recorder Rudder-pedal force

Control-surface position recorder Aileron, elevator, rudder, and stabilizer position

Wing-bending moment/sheer load and horizontal

Common timing circuit Synchronize collected data

It was found…that very little photographic contrast was being

Disaster: The NACA Loses Its First Pilot

First Research Results

Skystreak’s airspeed calibration flights, up to Mach 0.85, were published less

The stability measurements used in the report were primarily derived

Modifications and Flights Resume

When the aircraft was sprayed with the white undercoat…the

NACA Research Flights Continue

As the speed of the airplane is increased a right-wing heaviness

recording instruments and were not noticeable to the pilot. The

1952–1953: The Skystreak’s Twilight Years

least contribute to an operational naval fighter. The Skystreak represented a