332 THE AERONAUTICAL JOURNAL Uuiv.

im

subscribe to the final obsequies of the people who helped to stop the under-
carriage and the aeroplane after our friend had no further need of t h e m !
I have always regarded Major Taylor with reverent awe as the arch-figure-
juggler of aviation. How a man can design a floppy thing like a parachute
mathematically, and get it right, too, is quite beyond my limited comprehension.
But when such a man comes and does an ordinary parachute descent he seems
to come off his pedestal down to the earthly plane once more on a level with
myself.
Major Taylor's contribution is a very wonderful parachute which, for its
weight and area, has a much lower terminal velocity than any other parachute.
In it he has eliminated the inefficiency due to " fluting " at the periphery.
T. ORDE L E E S .

The CHAIRMAN said before calling upon Mr. J. W . W . Dyer to read his
Paper on " Airship Fabrics " he would like to tell those present a few things
about him. Before joining the Airship Service in 1915, Mr. Dyer was the author
of several works on the chemistry of cellulose. He joined the Airship Service in
1915 at Kingsnorth and became assistant chemist at that establishment, where he
worked mainly on the determination of the permeability of airship fabrics, and
the circumstances which governed the loss of gas-tightness and loss of strength
in those fabrics. Since the end of 1918 he had been chief chemist, and had
been engaged on important airship research work. During the last nine months
he had been directing experiments of exceptional importance in exposing various
gas-bag and outer cover fabrics to the tropical conditions of Egypt. A large
proportion of those g a s b a g s and outer cover fabrics had just come back from
Egypt, and Mr. Dyer was now studying them closely and reporting officially on
them. So he thought it would be seen that Mr. Dyer could speak with great
authority on his subject.
Mr. J. W . W . DYER then delivered the following lecture :—

AIRSHIP FABRICS.
A proper treatment of the construction and properties of airship fabrics
would require a book rather than an essay. In what follows, it is possible there-
fore to deal briefly only with the principal types of fabrics, classified according
to the main functions of each, describing their structure and behaviour and the
chief factors affecting their permanence when in service.
There are, then three main classes :—
I. Fabrics for the envelopes of non-rigid airships.
II. Fabrics for the g a s b a g s of rigid airships.
III. Fabrics for the outer covers of rigid airships.
In all cases low specific weight (expressed here in g r a m m e s per square metre*)
is clearly of the greatest importance.
I. Fabrics of this class must possess gas-holding properties, mechanical
strength and pliability. The internal pressure (equal to about 30111ms. of water)
which maintains the shape of the envelope sets up in its fabric considerable tension,
and this is increased by the distributed weight of the whole ship.
Non-rigid envelope fabrics are normally of rubber-proofed cotton. The
number and weights of the plies of cotton and the total weight and distribution
of the rubber depend on the precise purpose for which the fabric is intended.
Some typical fabrics may be illustrated by reference to standard British
practice." Cotton cloths are used having the weights and tensile strengths given
in the following t a b l e : —
t gms/m2-=-34 = ozs.yd2 correct to about 1%.
juitf.issi] THE AERONAUTICAL JOURNAL 333

Reference letter. W e ight g m s . / m 2 . Strength k g s . / m .

B no 1100
C 80 800
D 65 650
E 45 5io
It will be observed that all have nearly the same specific strength or a common
breaking l e n g t h ! of about 10,000 metres. Cotton is used in preference to linen
largely because its greater uniformity and smoothness make it more suitable
material for the proofer. Single-ply fabrics are not used for several reasons.
In the first place, local defects might cause serious w e a k n e s s ; secondly, they have
little resistance to tearing when wounded; thirdly, a sufficient weight of rubber
proofing of proper quality placed between two cotton plies gives greater gas-tight-
ness than the same weight and quality used a s facing of a single ply. In two
and three-ply fabrics the plies are held together by the rubber proofing which is
also the gas retaining member of the compound fabric. On the outer cotton face
is placed either camouflage colouring or a layer protective against the weather
or an attempt to combine the two. Typically the protecting layer is of a special
rubber proofing having a surface of aluminium powder printed on it. The
following schemes represent normal two and three-ply fabrics such as are used
respectively for small non-rigids and for the top lobes of large trilobe non-rigids.
The letters represent cottons of the above standard types.
TWO-PLY. THREE-PLY.
B C
Main gas holding rubber t Main gas holding rubber,
100 g m s . / m 2 . 100 g m s . / m 2 .
(C) . (C)
Outer protective rubber-aluminium Rubber layer, 30 g m s . / m 2 .
layer, 50 g m s . / m 2 . C
Outer protective rubber-aluminium
layer, 50 g m s . / m 2 .
Total Weight, 340 g m s . / t n 2 . Total W e i g h t , 420 g m s . / m 2 .
The letters in brackets ( ) indicate a ply laid with its yarns at 45" degrees to
those of the other ply or plies, i.e., diagonally or on the bias. This greatly
increases resistance to tearing and is discussed more fully below. It may be
stated here that rubber is chosen as proofing material not because, for a given
weight, it has the lowest permeability to gases of all readily accessible plastic
materials, but because other media which may be excellent for low permeability
have disadvantages of one sort .or another from which rubber is free. Gold-
beaters skin, oil proofing, gelatine proofing are all excellent for gas-tightness,
but expense, fragility, slow production of the first, defective strength and difficulty
of seaming of the second, and in the case of gelatine dependence for useful life
on retention of moisture not possible in hot dry climates, are reasons (among
others) why these substances a s at present applied are inferior to rubber for the
purpose of non-rigid envelope construction.
It will be convenient to treat the tensile, gas-holding and weathering proper-
ties of rubber-proofed fabrics in separate sections.
(a) Tensile properties.—The standard cotton fabrics of the table already given
will be regarded as the basic materials having the strengths there stated, and this
section will discuss'the strengths of balloon fabrics, built up from them, when
tested in various ways.
The simple tensile strength is normally measured on a piece of 5 cms. width
and about 20 cms. (effective) length. The ends of this test piece being secured in
suitable clamps, it is then treated so as either
t i.e., a length which would break under its own weight, often a useful basis of comparison-
 THE AERONAUTICAL JOURNAL [July, 1921
334

(a) T o stretch it at a constant rate, or

(b) T o increase the load on it at a constant rate till it breaks.
The second, the constant rate of loading test, is the one usually employed in
experimental work and now insisted on in contractual testing, and all tensile tests
discussed here have been made in that way. A figure purporting to give the
strength of a cotton fabric is highly arbitrary. This will appear more fully from
what follows, but it may be stated here that it depends on the kind of machine
used ((a) or (b) above), the humidity conditions of the fabric, and this, if equilibrium
has been reached, depends on the temperature 1 and humidity of the ambient air.
Further, the apparent strength depends on the rate of applying the load during
the test. The higher the humidity and the faster the rate of loading, the greater
is the apparent strength. 2
W h e n in use, the fabric is under compound stress and it is clearly desirable
to be able to predict its behaviour in such conditions from the simplest tests on
small pieces. The range of d a t a available is not such as to make this possible
for all two or three-ply fabric combinations that might be selected, but a good deal
is known with reference to certain much-used types. The points of interest are
obviously :—
(i.) The strength of the two and three-ply fabrics in simple tension and the
relation of this to the strengths of the component cloths.
(ii.) The strength of wounded fabrics.
(Hi.) The relation between simple tensile strength and the result of various
compound stress tests including bursting,
(iv.) The effect of a sustained load.
(i.) In tabular form below are given t h e strengths in simple tension of a
number of two-ply fabrics. Their structures are indicated by letters signifying
standard cottons from Table I. Letters in brackets signify bias ply. All results
are in kilos/metre.
T = strength of compound fabric.
*! and t 2 = strengths of component plies, t2 being used for the bias ply. The
t2 of columns 6 and 7 is the mean of the warp and weft values.
Figures in brackets denote strengths as specified in Table I. and not actually
found by testing the separated plies.
1. 2. 3. 4. 5. 6. 7.
Direction. Kilos. per Metre. T _il_
Structure of fabric. tested. T. t, t2 t , t2 t, t,
1
A (D) Warp 1560 (1250) (650) .82 .66
B (C) Warp 1306 (1100) (800) .69 •56
B (C) Warp 1183 980 626 •75 .62
Weft 1234 1080 580 •73 .64
C (C) Warp 1025 800 844 .64 •50
Weft 932 653 753 .64 •45
D (D) Warp 742 63S 553 .60 -.0
Weft 707 562 463 •58 •53
BX (BX) Warp 1000 625 852 .69 •43
Weft 1000 635 800 .69 •43
CC Parallel Warp 1705 (800) (800) 1.06 •5
Weft 1580 (800) (800) •99 •5

1
2
The distribution of moisture between the air and the cotton depends on the temperature.
It would appear that the different results given by different types of testing machine should
be explicable on this ground. The writer has, however, data which conflict with this view,
but they cannot be discussed here. It is doubtful whether comparisons under carefully controlled
or observed conditions have been made,
 July, 1921] HUE AERONAUTICAL JOURNAL 385

It will be observed t h a t there is a variable difference between the numbers

in columns 6 and 7, this difference being related t o the support in simple tension
given t o the straight ply by the biassed. A direct tension test on a diagonally
doubled fabric is of course somewhat unreal and unrelated t o the conditions
obtaining in actual service. I t is a rough guide only, and the above figures are
to be regarded as fairly typical while not serving for the deduction of any rule
applicable to all bias doubled fabrics.
(ii.) Resistance t o tearing, i.e., to the indefinite extension of a small wound,
is very important for non-rigid envelope fabric. It is a property conferred on the
fabric by the presence of the bias ply, whose marked influence in this respect is
shown by the figures given in Table I I I . Tearing or wound strength is given for
an initially wounded test piece by the expression
Load to produce tearing

W i d t h of test piece.
Provided the test piece is large enough to include the area of non-uniform stress
distribution 1 caused by the presence of the wound in the stressed specimen, the
value of this expression is independent of the width of the test piece and for a
given form and size of wound is a characteristic of the fabric. In Table I I I . it is
expressed a s a percentage of the strength of the unwounded material. T h e results
relate to test pieces 6in. wide by aoin. effective length having transverse central
cuts. T h e column headed b/s gives the ratio
Strength of bias ply (or plies)

Strength of straight ply,

and the percentage wound strength is seen to increase with increase in this ratio.
Further reference t o factors affecting tearing strength is m a d e in a later
section on weathering.

Tearing strength as percentage «

of full strength.
Fabric. i" cut. 1' cut. 1
r cut. b/s
i-ply proofed C 40 22 19 •—
CC parallel ... 33 24 18 —
Plain i-ply linen (130 gms./m 2 52 4i 33 • —

B (D) 57 46 41 •59
B (C) 62 45 — •73
C (C) 72 59 51 1.0
D (B) 79 69 63 i-7
C (C) C 5° 35 — •5
CC (C) 88 •— •— 2.0 •

(iii.) The behaviour of fabrics both unwounded and wounded under compound
stress conditions has been studied in three principal ways :—
(a) By loading the a r m s of a cross-shaped piece the central square is
placed under compound stress, the ratio of the two stresses being variable a t will.
There are, however, few results by this method, particularly on biassed fabrics.
It has the disadvantage that the stress on the test square is different by an unknown
amount from t h a t calculated as the quotient of the load by the width, on account
of enhanced stresses which exist a t the corners of the square.
1
The so-called " danger rectangle."
336 THE AERONAUTICAL JOURNAL (July, Ml

(b) Small cylinders of fabric sin. diameter and 30m. long having metal end
pieces have been used. They were subjected either to pressure above or
to pressure plus further longitudinal tension by loading them a s they were held
in position between the grips of a fabric testing machine. In other cases, using
similar cylinders bursting pressure only was applied with no additional longitudinal
tension. From test's of this nature with small cylinders the values for t h e ratio
Tensile strength by bursting "Tb
or —
Simple tensile strength Tt
•were found to be 1.4 t o 1.6 for various biassed fabrics, 0.6 t o 1.6 for parallel
doubled and about 0.7 for single ply.
(But see succeeding section.)
(c) Bursting tests on larger models.
L a r g e r cylinders ranging u p to about 3 ^ metres diameter and having hemi-
spherical ends have been used by Avorio in a number of bursting experiments.
H e found a scale effect with cylinders u p t o o n e metre diameter such that t|he
bursting stress decreased with increasing diameter. Beyond this u p to t h e largest
tested there was no further decrease. From the minimum results, his values lor
the ratio Tb/Tt were found to be 0.63 and 1.00 for parallel and diagonal two-ply
fabrics respectively.
In some tests carried out a year -or two a g o in this country, models 20 feet
long of t h e S S . type of non-rigid airship were used. T h e envelope except for one
gore w a s of strong three-ply fabric. Gores of t h e experimental two-ply fabrics
tilled the g a p , in turn being inserted by stuck seams. Pressure w a s supplied by
means of a blower, t h e circumference at b u r s t and b u r s t i n g pressure w a s
measured. Such tests were made on whole and variously wounded fabrics of
three types, the wounds being made a t t h e maximum diameter before the pressure
was applied. The weft yarns took the hoop tension and there were no transverse
seams. F o r comparison with the results of these experiments t h e weft strength
of the fabrics in simple tension, both unwounded and wounded, were also found.
Table, I V . gives the results (T„ for bursting tests, Tt for tensile tests) in
kilos/metre.
Table V. gives the ratio Tb/Tt for several of the conditions tried.
TABLE IV.

T„ Kilos/metre. Tt Kilos/metre.
Fabric. No.
wound.
r i"
hole.
3"
X
slit.
No.
wound.
V f"
hole. f
slit.
hole. hole.
C (C) 1110 ni5 1007 825 996 707 650 620

B (D). 1366 1004 920 — 1140 7i5 664 620

CC Parallel ... 1130 5r3 435 340 1250 604 500 4*3

T A B L E V.
Ratio Tb/Tt.
Fabric. No. wound. 1" hole. r hole. f slit.
. . B (D) • 1.20 1.40 1 •38 —
CC Parallel .90 .85 .87 .80
C (C) ... ... I.II 1.58 ' 1.5S 1.33
 tv.lv, 1921] THE AERONAUTICAL JOURNAL 337

The cottons in the three fabrics were not exactly of the weights specified in
Table I. Actually the total weight of cotton was, within a very few g r a m s per
square metre, the same in each fabric. The strength figures as they stand a r e
therefore proportional to specific strength. The results show the superiority of
diagonal over parallel fabric and also how misleading, for this comparison, a
simple tensile test is. The C (C) fabric has for small wounds a slight superiority
over the B (D), but there is an indication that with larger ones this might disap-
pear. T h e results of Table V. stand between Avorio's and those quoted in section
(b) above. Comparative bursting and tensile tests do not seem to have been made
under strictly comparable or observed conditions. Scale effect, seams, humidity,
temperature and rate of loading all affect the comparison. Rate of loading affects
the apparent strength of two-ply fabrics very considerably. Corresponding to
rates of 3olbs./in/min. and i5olbs./in/min. differences in apparent strength of
the order of i o per cent, and 20 per cent, for two-ply diagonal and parallel
respectively may be found, the higher rate giving the greater apparent strength.
This is for unwounded test pieces. The rate of rise of pressure was not taken in
the bursting tests described above but they probably correspond t o a rate of
loading. slower if anything than the slower of the two just quoted. T h e tensile
tests compared with them were made at the slow rate. H a d they been a t the
higher rate the tensile figures would have been higher and the ratio Th/Tt
correspondingly reduced.

Strength Under Sustained Load.

IV. This is of the greatest importance, but only comparatively lately has it
received attention. W o r k on the subject is at present in progress. The data now
available from earlier work are not numerous nor readily summarised in a general
statement. The figures in Tables V I . and V I I . will give an idea of the behaviour
of fabrics under conditions of sustained load.

TABLE VI.
Tests on warp direction of B cotton. Normal breaking load at 72 per cent,
humidity and i5olbs./in/min. loading = i29lbs./2".
Load applied. % of Normal Breaking Load. Time to break.
116 90 1 or 2 seconds,
no 85 A few seconds.
103 80 8 to 13 minutes.
95 75 1 to 2 hours.

TABLE VII.

Kilos/ netre. % of norma

Fabric. Normal Sustained breaking Time to
strength. load. load. break (days).
C (C) C warp ... 1900 9SO So 5
B D Parallel warp ... 1460 900 62 5
B (D) warp ... 1310 680 52 7
2
Single-ply linen 120 g m s . / m 1100 620 56 , 5

The fabric is not seriously weakened as a whole by sustained loading. If

after the break the remaining portions be tested in the ordinary way they are found
t o Have a strength little lower than that of new material. This behaviour of
338 THE AERONAUTICAL JOURNAL Uuiv.mi

fabrics under sustained load is clearly of g r e a t importance and introduces a hew

corrective to the estimates made of the stength of the fabric a s compared to the
stresses it is subject to in service.

The Fabric as a Gas Holder.

The two and three-ply fabrics, whose structure is described in an earlier
paragraph, each contain a layer of rubber proofiMg of i o o g m s . / m 2 . It is on this
principally that their gas-retaining properties depend, and in present-day fabrics
a proofing layer of ioo to 120 g m s . specific weight is adopted, the lower figure
being the normal standard. In the three-ply fabric the second layer between plies
of 3 o g m s . / m 2 is merely to hold the middle and outer plies of cotton together, though
it will undoubtedly help to reduce the permeability of the fabric, l h e outer layer
of 5 o g m s . / m 2 also helps when the fabric is new, but in use it probably ceases to
function as a g a s holder after a very short time while still remaining a more or
less efficient protective layer. Other weights and distributions of rubber have
been used, e.g., in earlier three-ply fabrics two 85gm. inter-ply layers were used.
In many fabrics an inner facing" h a s been used ot 15, 30, 40 or more g m s . / m 3 .
The composition of the rubber proofing varies somewhat with different makers.
Normally it is about 95 per cent, hard fine P a r a rubber, from 1 per cent, to 3 per
cent, of sulphur, often together with say 3 per cent, or 4 per cent, of inorganic
substances such a s lime, magnesia or litharge a s accelerators of vulcanisation.
Some proofers use no mineral accelerator. F o r standard British tabrics noi
vulcanisation is specified. Experience in this country with cold vulcanisation has
not been successful. There has been no attempt to set up a standard degree of
vulcanisation for the two excellent reasons that no one knows which is the best
one, and if they did, they would not in general be able to test a finished proofing
and say whether it had or had not been brought to that optimum degree. It may
be said here that the percentage of combined sulphur, a figure correlated with
degree of vulcanisation for the same type of mixing, varies in British proofing s
from about 0.3 to nearly 2.0, though more usually between 0.5 and 1.5. iNo
connection between this and permeability can be traced.
In testing proofed fabric, the permeability to hydrogen is measured with the
fabric a t normal humidity and 2o"C. The figures for two and three-ply fabrics
as described earlier may not exceed 12 and TO l i t r e s / m 2 / d a y respectively. Two-
ply fabrics, at any rate, are usually a litre or so below the specified figure. The
connection between weight of proofing and permeability is not very clear. W i t h
rubber films there is a regular decrease in permeability as the specific weight
increases. Such rubber film has, however, weight for weight, a very much higher
permeability than rubber proofing of the kind just described placed between plies
of cotton. The evidence on weight and permeability with regard to proofed fabrics,
as distinct from films of rubber, is by no means unequivocal. In one instance
the figures for a number of two and three-ply fabrics by the same maker were
compared. The former had almost exactly half the weight of proofing of the
latter, yet the mean permeability of the two-ply was very close to that of the
three-ply. On the other hand, there a r e numerous examples of pairs of fabrics,
each having the same between plies layer, and one having, in addition, a facing
of rubber where the faced fabric has a decidedly lower permeability. Two-p'y
fabrics of the type already described with an outer protective layer frequently
show on weathering a small immediate rise in permeability, and then if the
protective coat is a good one, no further rise for a long time. This change
corresponds to the early breakdown of the protective layer as a g a s retainer which
function is clearly performed to some extent when quite new. There are data
t i.e., next to the gas and therefore sometimes called the " gas layer," a term open to
objection. In the schemes for two and three-ply fabrics given earlier this layer would head.the
list.
Atfeittti THE AERONAUTICAL JOURNAL 339

of these kinds in abundance, but they do not settle the point. They all deal with
comparison between one and two and three layers of proofing. There is unfor-
tunately no set of permeability data for a series of fabrics having different weights
of inter-ply rubber of the same kind and applied in the same way. This matter
of permeability and weight has an obvious practical bearing, and it is of some
theoretical importance in connection with the mechanism of the passage of gases
through rubber m e m b r a n e s . ! In practice, for framing specifications, the rough
view is taken t h a t the more proofing there is the better the fabric should hold g a s .
This is demanded and obtained. The fact that this has no bearing either way
or any theory since the specification requirements -may leave the proofer, or at
any rate the ideal proofing with, so to speak, plenty to spare, leads to a further
observation.
The material used for the (statistical) investigations, by which it was sought
to discover what relationship, if any, existed between weight of proofing and
permeability, may have been very unsuitable. The method of spreading is very
important for the attainment of low permeability. It is necessary to spread a
ioogm. layer in a number of thin successive coats. T h e proofers who made the
fabrics from which the statistics were obtained were working to specification, and
had merely to keep the permeabilities of the different types below or not exceeding
certain values. The fact, therefore, that with 200gms./m 2 and with i o o g m s . / m 2
of proofing they produced fabrics with nearly identical permeability proves—well,
it proves that they did it, and not very much more unless it is quite certain that
in each case with the large majority of pieces proofed the very best possible
spreading was being done.
It may be stated here that it is possible too much stress has been laid on low
permeability to the neglect of other qualities, particularly endurance. Standards,
of course, have to be set up, and they imply measurement. Permeability is
readily measured, but up to now no endurance test that can be usefully
incorporated in a specification has been devised. " An English S u m m e r , " which
was once suggested as a trial period, is scarely a standard quantity. (The question
of endurance is discussed more fully in the sequel.)
Permeability is largely affected by temperature, to some extent by humidity,
and probably by tension. The first, the temperature effect, has been studied
fairly well in a preliminary manner. The temperature coefficient of permeability
is positive and large, viz., about 4 per cent, to 5 per cent, per i°C. Higher
humidity lowers permeability, but the effect for humidities short of saturation is
small for normal fabrics. T h e effect of tension does not seem to have been
examined. To test the permeability of fabric in tension would not be easy with
any of the existing forms of apparatus unless the tension were produced by high
excess gas pressure on one side and that would introduce undesirable complica-
tions. Small pressure excess has little effect on permeability. It was usual to
test with a hydrogen excess pressure of 3omms. of water, but it makes no practical
difference if the gas is at atmospheric pressure, at least with normal fabrics.
The permeability test to hydrogen is the one that is always made. Yet
permeability to air is in many ways the more important property of the fabric.
The loss of hydrogen through permeability is usually small and not costly to
replace. But the entrance of air lowers the purity of all the g a s in the envelope,
reducing lift and tending towards the formation of an explosive m i x t u r e . ' The
passage of nitrogen and oxygen through rubber proofing does not take place at
rates in accordance with Graham's Law. Diffusion proper may be operative, but
it is masked by some other effect, probably differences in the absorption of the
two g a s e s by rubber. Measurements on air, nitrogen, and oxygen permeability
have been made for a number of airship fabrics. Taking the hydrogen permeability
t It is not possible to discuss this here. Anything adequate would be lengthy.
340 THE AERONAUTICAL JOURNAL Uuly, 1921

as unity values for oxygen at i5.5°C. ranging from .25 to .32 were found, and
on one fabric for nitrogen the value .14.
These values may, of course, not be correct a t other temperatures. 1 It follows
from these figures that, from air, nitrogen enters the envelope about twice as
last as oxygen (assuming permeability proportional to partial pressure). In
(qualitative) agreement with this, it is found that the non-hydrogen fraction of
the g a s in the envelope is richer in- oxygen than ordinary air is. When the
amount of impurity is small the proportion of oxygen is greatest,- and as the
total impurity increases it approaches more and more to air in composition as
regards oxygen and nitrogen. There are two (or perhaps three) reasons for this.
The rate of entry of oxygen will fall off faster than that of nitrogen because the
difference in partial pressure of oxygen on the two sides of the fabric will diminish
more quickly than the corresponding value for nitrogen. Secondly, to make
good actual loss of g a s the envelope will be repeatedly replenished with fresh
hydrogen which usually contains o. 5 per cent, of nitrogen and no oxygen. Possibly
there will be a small amount of leakage of unchanged air inwards, e.g., from a
hole in a ballonet. The subject of air permeability needs further study both for
fabric and for seams in fabric.
Seams themselves are frequently responsible for a considerable fraction of
the total loss of gas from the envelope, the loss taking place through the stitch
holes. Unstitched seams, simply overlapped and taped on both faces have, if
properly made, a lower permeability over their a r e a than over the same area of
the plain fabric. Stitching adds nothing to the strength of the seam. It cannot
do so since a properly made stuck seam is stronger than the fabric which it unites.
Yet for a number of reasons that need not be discussed its use is persisted in
and the resultant bad effects on permeability have to be overcome as much as
possible. The passage of hydrogen laterally along the textile components of the
proofed fabric is the principal cause of seam leakage. How this may operate is
seen most easily by constructing diagrams of simple overlap, overlapped and taped,
overlapped, stitched and taped seams. The first is leaky; the second is
theoretically 1 gas-tight, not because the tape is gas-tight, but because the rubber
solution used seals the cut edges of the textile. T h e third is leaky, on paper and
in practice. An earlier reference was made to fabrics having an inner facing of
rubber of 15 to 50 or more g m s . / m 2 . Various functions have been attributed
to .this layer from time to time and they need not all be discussed here. One of
them is of importance in connection with seams, really only with stitched seams.
If two lengths of stitched and taped seam be made, one in two-ply fabric having
no inner face of rubber and the other in rubber-faced fabric, the seams in the latter
will, other things being equal, be found to be very much more gas-tight than those
of the former. The inner facing prevents the lateral passage of hydrogen along
the cotton into the shaft formed by the stitch hole. There is no doubt at all
about this effect. It is, however, not certain that the whole inner facing is
functioning. In one test, some seams were made with fabric a s described and
tested. From some of the pieces the rubber facing was removed 2 up to the edges
of the seam, but the seam permeability of these pieces was n o higher than that
of the others. It is quite possible, too, that the good effect of the extra facing
can be produced by thorough impregnation of the first coat or two of inter-ply
rubber. This experiment is being undertaken.

Seam permeability is stated in litres/lineal metre/day. From tests on pieces

1
It is stated that at 4odeg. C. the air permeability is equal to the hydrogen.
2
In the " air " impurity as much as 40% (instead of 21%) has been observed to be oxygen.
On the figures given for oxygen and nitrogen permeability so high a figure should not be possible,
but the discrepancy is not large.
1
And practically too in the sense explained earlier in this paragraph.
1
This may be done by rubbing gently with a piece of raw indiarubber
July, 19211 THE AERONAUTICAL JOURNAL 841

having seams across them the permeability in these terms can be calculated if the
fabric is sufficiently uniform for an estimate to be made of the permeability in
the non-seam region. W i t h rubber-proofed fabrics an approximation is possible,
but the better the seam the greater the uncertainty as to how much is real seam
permeability. The seam, of course, occupies a certain area, say 400 sq. cms.
per metre of seam. From this it can be seen that with normal fabrics of 10 to
12 i s / m 2 / d a y permeability there is no real excess of permeability due to the join
when the seam permeability is of the order of 0.5 i / m e t r e / d a y . In non-rigid
envelopes the total length of seams in fabric separating gas from air is very
roughly 2 metres for every square metre of fabric. Seam leakage need therefore
only be about 5 i s / m / d a y in order to double the all-over permeability of the
envelope. This figure is frequently exceeded and may indeed rise four or five-fold
with unfaced fabric even when new. Good workmanship can however do much
to keep down seam leakage even with fabrics theoretically bad for seaming, and it
means in this case plenty of rubber solution well rubbed in.
The above summary relates principally to new seams. There is not much
systematic information about different types of used seams. Another aspect of
seam leakage that needs attention is that of their air permeability under various
pressure conditions and its relation to loss of purity of g a s in the envelope.

Endurance and Deterioration.

Good rubber-proofed fabrics that are stored at a normal temperature and
shielded from light retain their properties unaltered for a long time to be measured
in years. In service they deteriorate more or less rapidly, the cotton losing
strength and extensibility and the rubber becoming more porous, usually conse-
quent upon discernible chemical change. The agents causing deterioration are
light, heat, and tension. Of these the principal one, either directly or indirectly,
is light. In the development of airship fabrics it is noteworthy how one point at
a time has received almost exclusive attention. In the early days it was insufficient
scouring, acidity in the cloth and means of overcoming it. 1 Then the importance
of protection from light was to the fore, the nature of the aluminium facing was
considered, and the necessity for a pigmented layer under the aluminium (still
imperfectly practised) insisted on. There is still more to be done on this subject,
but it is, in the writer's opinion, time that tension effects had their turn. This
view is based on experiments made during last summer which will be referred to
briefly later in this section.
The action of light on cellulose has been studied by Aston, who used linen
thread. No doubt the results apply also to cotton. He found that visible light
has scarcely any effect and that the specially active rays are in two groups in the
ultra violet. He further found that the rapidity of attack was greatly reduced by
the removal of oxygen. In agreement with this it has been shown that the
product of the action of ultra violet light on cotton in air has the properties of
the oxycelluloses. The textile cannot, of course, be protected from oxygen, but
it can be more or less completely shielded from light. T h e rubber is oxidised in
light and air, and apparently even more readily so is the free sulphur always in
the proofing. The sulphuric acid arising from this is a further cause of deteriora-
tion of the cotton. It may be observed that some of the free sulphur nominally
in the proofing is, in a new fabric, actually in the cotton textile, and on oxidation
is favourably situated for attacking the cotton. It is noteworthy, too, that light
stopping protective facings, judged to be good by the fact that the inter-ply rubber
deteriorates very slowly, do not prevent the relatively rapid oxidation of the free
1
It should have been mentioned earlier in describing the standard cottons that they are
specially scoured and afterwards treated with a dilute weakly alkaline substance, sodium acetate,
to counteract acidity. These processes were worked out about 1915 at the Manchester School
of Technology.
342 THE AEtlONAUTICAL JOURNAL Uuiy.mi

sulphur in the top textile and inter-ply rubber. The rate of oxidation of the
rubber may be judged by the acetone extract it yields. In a well-protected inter-
ply rubber whose acetone extract on six months' exposure (January to June) only
increased from 5.1 per cent, to about 7 per cent., the sulphur decreased from
1 per cent, to 0.25 p e r c e n t . Heat accelerates the decay of rubber in the light, and
also the attack on the cotton of the acid from free sulphur. The action of heat
alone on rubber in the absence of light is not comparable with the action of light.
No judgment of the probable weathering quality of a proofing can be based on
heat treatment tests. Experiments on the artificial weathering of proofed fabrics
using the mercury vapour lamp have not proved a satisfactory substitute for
exposure out of doors. Cotton is relatively more and rubber much less attacked
by the ultra violet lamp than in the' natural weathering test.
Methods of protection.—All these aim at keeping light away from the rubber,
cotton or both. They include :—
(i.) Reflecting facings of aluminium.
(ii.) Pigmented rubber facings,
(iii.) Dyes in the rubber,
(iv.) Dyes in the cotton.
Fabrics could be selected to typify each of these four methods used separately
as well as (i.) with (ii.) and (i.) with (iii.).
It is important to reflect as much light 1 as possible, and in British airship
practice the aluminium facing has been relied on for this, no systematic use of,
e.g., white pigments in addition have been made. 2 No aluminium facing alone
has ever been successful in giving good protection. The best facings are made
by printing aluminium powder on the rubber protective layer. Aluminium facings
are nearly always patently discontinuous even when new. In some cases large
amounts of the powder have been incorporated in the outer rubber layer, but
this weathers badly and soon falls off. A printed aluminium facing used in con-
junction with a suitably pigmented rubber layer immediately beneath has been
found the best of all methods yet tried in service if the protection of the fabric as
a whole be taken as criterion. The only pigmentation that has been much used
and found successful is customarily referred to as the " heavy litharge " layer.
It is simply the use of a high percentage, up to 20 per cent., of litharge in the
outer rubber layer. Its virtues have been attributed to opaque colloidal lead
sulphide formed by reaction between litharge and sulphur during vulcanisation.
Lead sulphide is undoubtedly present, but only some 10 per cent, of the lead
appears to be in this form.

Method (iii.), the use of suitable dyesi'in the rubber, has been found very
successful in protecting it from oxidation. The dyes have a marked absorption
in the ultra violet. Rubber so treated may be exposed for months, at any rate,
in this climate, and show scarcely any increase in either, acetone extract or in
permeability. A comprehensive series of weathering tests in Egypt on fabrics
protected in this way is now in progress in conjunction with the manufacturers.
As hitherto used, these dyes have failed.to protect the cotton. In fact, no, dye
protection for cotton has been successful© Generally, dyed outer cottons with
no outer rubber have been for ground camouflage. Such dyed cottons afford
slight transient protection to the rubber.

1
Total incident radiant energy would perhaps be a better term.
2
One is being tried in an exposure test at present.
\ y T h e method has been patented by the North British Rubber Co. An example^ is the use
of Toluene — azo — toluene — azo — B Naphthol.
© A n alizarin dye used in an experimental fabric weathered in Somaliland was favourably
reported on, but no use appears to have been made of it in practice. (Barr A.C.A., Report R
and M313.)
July. 1921] THE AERONAUTICAL JOURNAL 343

The part played by tension is causing deterioration on exposure is very

important. Until recently most weathering tests have consisted in the exposure
of pieces of fabric held in a frame but not under any tension. From comparative
weathering tests on pieces of fabric with and without tension it is found that the
rubber in the tcnsioned pieces has perished more than in the untensioned. In
certain fabrics a partial explanation at least is very simple. As a result of
stretching under tension the outer rubber film becomes roughened and the
aluminium no longer lies as a smooth layer but is easily rubbed away. Briefly,
tension has made the whole fabric more accessible to light. It seems probable,
however, from some other results that tension acts in other ways too. Three
different fabrics were made the subjects of comparative weathering tests of this
kind. Two of them had what proved to be relatively poor outer facings which
roughened and fell away to some extent as described above. The other outer
facing was excellent, being a " heavy litharge " rubber layer topped with a good
aluminium coat of considerable permanence. W i t h the first two there was greater
loss of strength of the top ply of cotton than in that ply of the comparison
untensioned pieces. W i t h the third fabric the loss of strength was negligible
both in the loaded and unloaded pieces. These strength data may be taken a s
some indication of the relative light penetrations, i.e., considerable in the first
two, little in the case of the third. Yet all these showed marked change in the
inter-ply rubber. The first two signified this change by increased acetone extract
and drop in permeability. The third showed a large rise in permeability and only
a small change in acetone extract.

The General Course of Deterioration.

The conclusions drawn with regard to this will not in general be the same
when based on exposure tests with untensioned pieces as when based on the
examination of samples from actual service. Exposure tests may sometimes
point to the rubber outlasting the textile. Experience with fabrics in service on
the whole points to the opposite conclusion. An exception must probably be made
in the case of the dye-protected rubbers. It is obviously better, since the fabric
cannot be everlasting, for it to become useless through loss of gas-tightness than
through loss of strength. Oxidative decay of the rubber (the inter-ply rubber
here is intended) may follow one of two courses in nature of products and con-
sequent effect on permeability. In the earlier stages there may be either a
steady and decided fall in permeability, sometimes to 2 or 3 l i t r e s / m 2 / d a y only,
or a very gradual, scarcely perceptible rise. 1 W h i c h of these two happens depends
immediately on whether a soft and tacky or a more or less dry and harsh oxidation
product is formed at the start. Following the fall of permeability will be a sudden
and large rise as the soft product hardens or crumbles.
It is not quite clear what determines the course that will be followed. Broadly
it appears to depend on rapidity or slowness of oxidative attack. If it is rapid,
the soft product and lowered permeability will be encountered. If it is slow, the
other product and condition. It is quite certain that in some cases, at any rate,
the same proofing can follow either course according to the nature of the exposure.
The actual seriousness of a given loss of strength of the outer 1 cotton will
depend mainly on whether it is the bias or straight ply. On this question of bias
or straight for outer ply there have been many alternations and different practice
has obtained in different branches of the service at the same time. The many
reasons, good or bad, for these changes cannot be discussed here. It is mow
agreed to put the bias ply outside. In this case a considerable weakening of that
ply means little loss in tensile strength of the whole fabric. And further, the
1
This may be preceded by an increase on the original value within a few days of first
exposure due to rapid loss of any gas-holding properties by the outer protective coat, if present.
1
It is clear, of course, that the attack is from without.
344. THE AERONAUTICAL JOURNAL [July, 1921

bias ply still contributes its anti-tearing qualities not seriously impaired when it
has lost considerably in strength individually. This is on condition that the rubber
layer between the plies is still substantially sound. If the rubber has decayed in
such a manner a s to become soft and infiltrate the yarns of the textiles, glueing
them into' what is, in effect, a sheet of continuous material, the fabric will have
a dangerously low tearing strength though it rnay have a good tensile strength
when unwounded. F o r the bias ply 'to d o its work there must be adhesion to
the straight ply, but not adhesion that has become amalgamation.
Finally, on the subject of endurance, there are several cases known of large
increases of permeability with no sensible chemical change in the inter-ply rubber
layers. Up to now it has received no explanation. 1 It does not seem at all certain
that tension on the fabric is a necessary prerequisite condition t for one or two
cases of it are known to have occurred in ordinary non-tension weathering tests.
It does seem probable t h a t it is confined to one type of proofing. The increases
in permeability are from, say, 10 or 12 i s / m 2 / d a y to 40 or 50 or even more. The
subject is still under investigation.

Classes II, and III., i.e., Fabrics for Rigid Airships.

From the foregoing short account of the fabrics of Class I. it will be seen
that though there are serious g a p s , a good deal of systematic work has been done
and a more or less coherent body of knowledge put together. This is not true
to the same extent of the fabrics in Classes II. and I I I . , though these are and
will be ahead of the others in importance. The properties and problems of these
fabrics are in general quite different from those of the fabrics in Class I. One
consideration of great importance for the latter does not arise in the same way
with fabrics for rigids, and that is the tensile strength. The strength of the
fabric of a non-rigid envelope is a principal measure of the strength of the ship
in a way that is not true of either kind of fabric in a rigid airship. The strengths
required for the latter fabrics are uncertain. Possibly the greatest stresses they
ever experience are those imposed when the g a s bags and outer covers a r e being
put in their places in the ship.
I I . Gas-bag fabrics.—The two requirements are low permeability and small
specific weight. The strength required is not great. Other necessities as flexi-
bility and (as far as possible) water repelling non-hygroscopic character are to
give permanence to the principal qualities.
In almost all cases gas-bags for rigid airships consist of goldbeaters skin
stuck to cotton fabrics. Silk and a mixture of silk and cotton have also been
used. No substitute for goldbeaters skin has yet been satisfactorily applied on
a large scale, though experiments on this point have been made and are being
continued with other animal membranes than the true goldbeaters skin as well as
with artificial films. The real goldbeaters skin is the membranous covering of
the caecum of the ox. In size it varies from about 27m. by 6in. to 40m. by ioin.,
and in weight from 14 to 28 g m s . / m 2 under ordinary humidity conditions. The
weight of skins when on the fabric will be rather greater because of overlaps,
say 2 5 g m s . / m 2 for an average single layer. For storage and transit the skins
are packed in barrels of salt crystals. Before use they are thoroughly washed,
scraped free from most of the adherent lumps of fat, and then soaked in a 5 per
cent, aqueous solution of glycerine which appears to be preferentially absorbed
from the solution by the skin to a noticeable extent. The glycerine is to maintain
flexibility. The cotton to which the skins are attached is D quality, i.e.,
6 5 g m s . / m 2 , and it is proofed with vulcanised rubber 2 0 g m s . / m 2 . T h e proofing
is to form a basis on which rubber solution can be spread. This is done thinly
with a pad, not a brush, and several coats are used. T h e skins, wet with glycerine

1
See a discyssion of the matter in T.1159 and R and M.584, both reports of the former A.C.A,
July, 1921] THE AERONAUTICAL JOURNAL 345

solutions, are applied to this 1 and good adhesion is obtained. There is an overlap
where each skin joins its neighbour. If two layers are applied, the second layer
is applied to the first with no adhesive between and so that the longitudinal
directions of the units of this layer are rectangular to that direction in the other
layer. The strength of goldbeaters skin in the transverse direction (i.e., the hoop
direction in the animal's intestine) is about twice that in the longitudinal
direction. W h e n the skins are laid and the fabric is " air d r y , " it is sometimes
varnished, in present practice an oil varnish being used. T h a t is a brief descrip-
tion of present British practice. A fabric is produced having a weight usually
below i 6 o g m s . / m 2 when double skinned and varnished. This refers to normal
humidity. The variation in weight with humidity is not large unless excessive
glycerine has been used. The permeability of skin-lined fabric is not very uniform.
Double-skinned fabric of the type described is frequently below o. q J / m f / d a y .
It is quite commonly lower; it is probable that most of the permeability observed
in normal pieces is not through the skins but through minute faults. In German
practice the adhesive used is a special glue and, at any rate recently, both skin
and cotton faces after make-up are treated with a water-repellent coating. The
German method and product both differ in several important particulars from the
British. Experiments with new adhesives and new methods of manipulation are
now in progress here * Differences in behaviour and properties between the two
types of fabric will now be briefly discussed in connection with the endurance and
causes of deterioration of skin-lined gas-bag fabrics.

Gas-bags are protected by the outer cover from the weather and from light,
and no question therefore arises of the deterioration of the textile or rubber (if
used) through the action of what is, with other fabrics, the most potent harmful
agency. The life of the gas-bag is the life of the skin lining, and methods of
protection have for their object the maintenance of the goldbeaters skin. This
is mainly a mechanical and not a chemical question. Considering first the use of
rubber adhesive, the skins are stuck by its means to cotton fabric. This is very
extensible even at low stresses. The skin also is extensible if it is kept in a moist
condition. To do so is the function of the glycerine, which is not a softener
per se, but which being hygroscopic, keeps in the skin an amount of moisture
greater than it would otherwise have. If a piece of new skin-lined fabric of this
type be broken in a testing machine, it is found that the skin remains intact up
to the point of rupture of the cotton even though extensions of, sav. 10 per cent,
and 17 per cent, in the warp and weft directions are obtained. If, however, such
a fabric be pulled on the bias and distorted with accompanying much greater
extension than in the direct pull, the skin will be cracked. Or, if the skin be dry,
as after exposure in warm air or low relative humidity, a small extension accom-
panying direct pull may be enough to break the skin. If the warmth and drying
be carried still further the contraction of the skin cannot be followed by the cotton,
with the result that at a region of weaker adhesion they part company and a
pucker is formed. If this is smoothed out by tension in this dry condition, once
more the skin will crack. If a piece of fabric which has been brought to this
puckered condition by heating in a dry oven at about 7o°C. be left for a time in
air at ordinary humidity, it will recover its extensibility and the puckers can be
pulled out without cracking the skin over them, though the separation of skin
from fabric in that region of course persists.

A similar puckering and relaxation is obtained when such fabric is exposed

in a climate with large variations in temperature and therefore in relative humidity
The latter and not absolute humidity is the moisture control. It appears that
1
Numerous important practical details and workshop methods for systematic and rapid
working are omitted.
1
It is unfortunate that this paper is being written at a time when interesting developments
ore in their early stages and by no means ripe for description or discussion,
 346 THE AERONAUTICAL JOURNAL tw«,i92i

after a number of alternations the skin loses its power of recovery even in a moist
atmosphere, and is in some way fundamentally altered. 1
The results suggest the trial of softeners that do not depend on moisture
for their action. This trial is being made.
T h e behaviour of glue-stuck skin-lined fabric is very different. There is no
puckering on dry heat treatment with these, and they can, even when dry, be
stressed considerably more than fabrics of the other type without any resultant
rupture of the skin. The glue alters the extensibility of the textile and under
moderate loads, at any rate, takes control. Further, the glued textile will tend
to expand in a moist atmosphere and contract in a dry one, behaving in this way
like the skin or, a t any rate, accommodating the one component to the other.
The whole question of the use of glue as adhesive including composition of glue,
mode of application, pre-treatment of textile and skins, as well a s after treatment
of the fabric with varnishes and water-repellents, is now e n g a g i n g attention, and
it would not be profitable to add more on the subject now.
It may be observed that neither glue-stuck nor rubber-stuck skin-lined fabric
will stand crumpling when warm and dry without serious permanent increase in
permeability.
The permeability of both types of skin-lined fabric is usually, but riot always,
lowered by increase of humidity. This lowering with the German glue-stuck
fabric is often very large. It is to be expected that this effect would be greater when
glue is the adhesive. T h e Germans appear not to aim a t as low permeability as is
expected with British fabric. This is in accordance with their lavish provision of
hydrogen. It is difficult to arrive at any reliable result for the all-over permea-
bility of either type of b a g . The only experimental method is to make loss of
purity curves. There are numerous sources of error to consider, and even
leaving these out of account the data do not lead to a hydrogen permeability result
without an assumption a s to the relative air and hydrogen permeabilities. It is
true that loss of purity itself is the important result, but if the calculations from
this to hydrogen permeability cannot be made, neither can the reverse one from
laboratory determinations. Qualitative comparative results must suffice unless a
fairly comprehensive investigation be undertaken of the relative air and hydrogen
permeabilities of a number of skin-lined fabrics.
The skin, as already stated, has appreciable strength and contributes to the
tensile strength of the complete fabric. Skins stuck with rubber also greatly
increase the resistance to tearing, but glue-stuck skins do not, probably because
of the impregnation of the textile by the glue. The following are typical figures
for the two kinds of skin-lined fabric.
Tensile strength Kilos/metre Tearing strength for "in. cut
warp direction. as percentage of unwonnded.
Complete Cotton ouly. Complete
Kind of Fabric. fabric. Skin removed. fabric. Cotton <nlv.

Rubber stuck (British) 825 670 . 65 40

Glue stuck (German) 860 608 40 40

Class III.—Outer Covers.

Ideally the outer cover would consist of a series of plane surfaces covering
1
It is possible that this is a result of other changes and not directly of the humidity .
alternation, e.g., of the development of acidity. There is no evidence on this point yet, but it
is hoped to obtain some from tests now in progress,
July.imi THE AERONAUTICAL JOURNAL Ml

the rigid framework and, with this framework, defining the shape of the ship.
This leads at once to the requirement that the outer cover shall be and remain
taut. Further, it must be weatherproof so as neither to admit rain to the inside
nor get waterlogged itself. It should also, as an extension of its waterproofness,
be water-repellent. It must, for retention of its own strength and for protection
of the gas-bags, be opaque to harmfully active l i g h t ; and, principally to avoid
superheating of the g a s , must reflect as much incident radiation as possible.
Added to this, it must have the lowest possible specific weight.
Linen, mercerised cotton and plain cotton have been used as the textile
components. The two latter are in use at present. Various doping schemes
used in the past have been only moderately successful and it would serve no useful
purpose to give an account of them here. Some single-ply cotton and linen fabrics
proofed with rubber having an aluminium finish were tried experimentally and
subjected to weathering and other tests before the importance of tautness in the
outer cover was fully realised. The dopes at first used were similar to aeroplane
dopes but of less contracting power. Now that the strength of the framework
is known to be sufficient, more powerfully contracting dopes, in all essentials the
same as aeroplane dopes, are used. These consist essentially of cellulose nitrate
or acetate, the latter having been used almost exclusively in the past, but the
former is introduced in the latest doping schemes for reasons to be mentioned
later. The function of the dope is first" and foremost to produce tautness. 1 It
is applied in a suitable solvent to a fabric already in slight tension. W h e n the
dope is dry, considerable tautening is found t o have taken place, the actual final
tautness being dependent on the nature and amount of the dope and the initial
tension of the fabric. Light absorbing pigments are introduced for the protection
of the textile. A red one, which has been found the best, consists of an oxide
of iron. The top coat contains aluminium powder to provide a reflecting surface.
This is only the briefest outline description of a doping scheme as now applied.
A few of the more important details, additions and alternatives to such a scheme
will now be briefly discussed.
The maintenance of tautness is a much more difficult matter than the attain-
ment of a good degree of initial tautness. In a moist atmosphere the tautness
is soon reduced through the absorption of moisture by the dope. It is stated
that by using very little dope an ever-taut fabric can be obtained because a balance
is struck between slackening of the dope and contraction of the textile, but under
such conditions probably a sufficient degree of waterproofness, protection for
textile and general robustness would not be attained. Experiment has shown that
pigmentation of the dope (initially done for quite another reason) is one of the
best means of minimising the slackening effect of exposure in moist a h V and the
best results are obtained with pigment in every coat of dope applied. Pigmenta-
tion of the right kind is thus excellent. It saves cellulose acetate, protects the
textile and maintains tautness under unfavourable conditions. The use of cellu-
lose nitrate dope instead of acetate has to be considered in this connection.,, l o o ,
for it is less affected as to tautness by changes in humidity. It is actually being1
employed now tor some p a r t s of outer covers, and is not used t h r o u g h o u t only
.because oj its g r e a t e r inflammability. The question of outer cover tautness, so
important on account of its b e a r i n g on speed, among other things, is a rather
1
Some apology is needed for the introduction of these elementary facts at this date in a
paper to an aeronautical society. It is done for the sake of a (limited) completeness and as
introductory to the further description. The great improvements in aeroplane dopes by which
airship outer covers are now benefiting are largely due to the chemical staff at the R.A.E., and
much of this section is taken from their reports where these matters are discussed in detail and
with the support of a wealth of experimental data.
1
It should be mentioned that earlier observations gave rise to the precisely opposite statement,
but fhe more recent experiments referred to above are very complete and probably their authors
had the earlier observations before them, and were satisfied that they were not generally true,
though no mention is mnde of this.
348 THE AERONAUTICAL JOURNAL [JuJv.mi

bigger problem than that of the tautness of aeroplane wing covers. T h e areas
involved are so much greater, and the view is held that dope cannot be expected
to do all t h a t is wanted in this way, 1 and in fact, the recent practice is to give
assistance by attaching the cover by inside ties to.diagonal wires.
Additional weatherproofness is sometimes attained by a finishing coat of
varnish (containing aluminium powder). Varnish here means a flexible nitro
• dope containing, e.g., castor oil.
In one acetate dope, linseed oil 2 is included as a water repellent, for which
purpose it appears t o be fairly efficient.
F o r the prevention of superheating there is the reflecting aluminium finish.
The efficiency of this varies considerably and depends largely on the nature of
the aluminium powder used. It seems possible that the reflecting power of
aluminium powder might be assisted by the use of zinc white which, out of a
large number of opaque substances tested, was found to possess the necessary
radiation characteristics in the highest degree, i.e., low absorptive power in the
region of maximum sunlight intensity and high emissive power with small rise
of temperature. It is probably not so effective in protecting the textile as the'
iron oxide pigment, but that might reinforce the zinc white in a layer beneath it.
Other problems that arise with outer covers touch the methods of applying
the dope. I n ' p r e s e n t practice pre-doping is used, i.e., a part of the dope is
mechanically spread before the cover is made up. I t is then laced in place and
the remaining dope (and varnish, if any) is applied in situ. It seems likely that
this will be abandoned as not making the most of the weight of dope it is desirable
to add. A light pre-doping may, however, be desirable t o give a slight stiffness
to the fabric to prevent distortion when putting in place prior to the doping proper.
Some of the foregoing, it will be seen, deals with proposals for methods yet
to be tried, and with regard to the new practice already in operation it is too early
yet to speak of its performance. It is known from many tests in connection with
aeroplane fabrics to give extraordinarily good protection to the textile. For outer
covers, as for gas-bag fabrics, this paper comes at a time when distinctly new
methods are being tried.
In conclusion, a word or two may be added on the general methods of chemical
and physical approach to the problems presented by these materials. I t is probably
true to say that applications of rigorous scientific methods and of well understood
principles are increasing, not merely in devising methods of tests but also in
working out new processes and employing new materials. Examples of this are
to be found in the studies of various solvents and solvent mixtures for dopes, in
preliminary researches on cellulose acetates and allied bodies, a n d in Aston's
work on the action of ultra violet light on cellulose, to name only a few examples.
On the other hand, there is a wide field where empirical methods are the only ones
available, and where the only good test is in an attempt to reproduce conditions
of service which the material will have to endure. All weathering" tests form the
main example of this kind of work. It is often, moreover, an empiricism where
it is not possible to select for variation and study a single factor, keeping all others
constant. In the interpretation of results, therefore, judgment and a weighing of
pros and cons may take the place of logical deduction. This perhaps is no draw-
back, for these qualities will certainly be required in devising any new material
as a result of the tests, for such material must always be a compromise.

1
It has been suggested that a slight positive pressure inside the cover should be maintained.
3
It is orthodox to object to the use of linseed or other drying oil in contact with textiles on
the score of tendering. It is possible that in conjunction with good light-absorbing pigments the
use of such oil might be free from objection.
July. 1921] THE AERONAUTICAL JOURNAL 349

DISCUSSION.
Mr. C. I. R. CAMPBELL said the paper was a comprehensive one and was
written by a gentleman who was, perhaps, the greatest authority on the subject.
A point which he (Mr. Campbell) thought required further consideration was the
value to be given to the bias ply in three-ply fabrics. In non-rigid design the
data available was very small indeed. In the non-rigid envelope the greatest
possible strength was required, particularly in a circumferential direction. A bias
ply was put in mainly for protection against accidental tearing, and no more
was really asked of it than that. One wanted to get the thinnest bias ply that
would do the work, and it was not known at present what was the minimum one
could g o to in that direction. He hoped Mr. Dyer would have particulars of
experiments on that matter before very long. The figures h e . g a v e for sustained
loads were very interesting and he (Mr. Campbell) had not seen them before. The.
fact appeared to be that a fabric could only sustain a tension of half its ordinarily
expected breaking load, and this had a marked bearing on what were customarily
believed To be fachtjrs of safety. It appeared that many of the factors of safety
which w e r e Talked about for non-rigid envelope fabrics should be halved, as the
loads were of a sustained character in most cases. Discussing outer covers,
which was one of the subjects upon which research, or rather perhaps practice,
was most incomplete. Mr. Dyer spoke of the importance of fireproofing. T h a t
was a point on which they all felt more strongly every day, and the evidence, he
thought, showed that if acetate dope were used with plenty of pigment in it,
appreciable fireproofness was secured. One did not get an incombustible outer
cover, but one that would not burn freely if a light were applied to it. At the
present time, when they were told airships were on the point of becoming extinct,
it was particularly valuable to have this information collected and put forward in
this authoritative manner, and they owed Mr. Dyer a debt of gratitude for having
taken the work in hand.
Dr. GUY BARR said with regard to the strength of two and three-ply fabrics,
the lecturer mentioned that jthe important things to consider were the relations
between the strength of the individual fabrics and that of the complete cover,
the strength of the wounded fabrics, and the effect of sustained loads. He (Dr.
Barr) would suggest a further important point—extensibility under compound
stress, which was important not only from the point of view of the adjustment
of local inequalities when stressed, but also from the point of view of the actual
volume of the ship. He referred to Table 7, bearing on the subject of strength
under sustained loads. It would be seen that the largest percentage of the normal
breaking load was retained in the second case dealt with (B.D., parallel warp).
It looked to him as if a considerable proportion of the loss—in the first and third
cases, at any rate, compared with the second—was due to the two plies coming
unstuck under the conditions of the experiment, and that this would not occur
to the same extent when the fabrics were in position with considerable pressure
inside. The difference between 50 and 62 was not large, but it was a considerable
proportion of the difference between 50 and 100. T h e remaining 38 per cent, was
probably to be ascribed simply to the fact that the effect of rate of loading appeared
10 be most serious at the lowest rates. Mr. Dyer referred to Dr. Aston's experi-
ments on the action of light on cellulose, and he (Dr. Barr) thought there was a
small mis-statement there. Dr. Aston found that the light from the mercury arc
produced the greatest deterioration in two particular portions of the spectrum,
but the fact that he found those two particular portions was presumably t o be
ascribed not to the fact t h a t those two parts of the spectrum of daylight were the
most active, but that they were the only two in the ultra-violet part of the mercury
arc spectrum studied in which there was any large amount of energy. H a d any
methods of protection been tried in which the reflecting'facing of aluminium on a
two-ply fabric had been combined with a dye in the cotton? For a long time
people had been satisfied with a yellow dyed cotton, which protected to some
350 THE AERONAUTICAL JOURNAL Uuly, 1921

extent the rubber underneath, and the dyed cotton itself did not deteriorate
extremely fast. The use of yellow dye showed a large improvement as compared
with undyed fabric. These reflecting facings had been tried with pigmented
rubber and with dyes in the rubber, but he had not seen any results of a combina-
tion of the reflecting pigments of aluminium with the dye in the cotton, and he
thought that combination should be tried. W i t h regard to the change in permea-
bility, which occurred occasionally on storage, when there was no obvious reason
why there should be a c h a n g e ; if one examined the fabric there was no evidence
of oxidation nor of anything else except change of permeability occurring. In
this connection there was an interesting observation made at the Bureau of
Standards that in one or two cases permeability did not increase, but decreased
to something like one-half of the original permeability instead of multiplying itself
by 5, which was happened in the case to which Mr. Dyer referred. He did not
know the explanation of it.
Colonel RICHMOND said he well understood Mr. Dyer's difficulty in compressing
his subject into the space allbwed for publication, and it was even more difficult
for those who had not received an advance copy to g r a s p the subject properly
from the precis into which the paper had been further compressed. It was a vast
subject, and even the written paper did not refer to a lot of things of which he
knew Mr. Dyer had intimate knowledge and on which they would all like to have
information and to be able to discuss. Amongst these the most important was,
perhaps, the mechanism of permeability. This was an extremely complicated
business. T a k e as an illustration the gold b e a t e r ' s sfeln membrane used in rigid
airship gasbags. This has a very Tow permeability to hydrogen, something like
1/200 of its permeability to water vapour. It had scarcely any permeability to
petrol. These were facts which it was very difficult to explain. ATso one might
refer to the fact that most colloidal substances seemed to have a limiting permea-
bility with weight. Some interesting work on this question had been done by
Mr. Ritchie, late of the Kingsnorth laboratory, but it did not g o far enough, and
he thought this question of the mechanism of permeability should be tackled as
soon as possible. He was sure it would have saved thousands of pounds during
the war in that more efficient means would have been found of restoring a large
number of envelopes which had to be scrapped. W i t h regard to rubber fabrics
in general, he would not prejudice the examination Mr. Dyer was going to make
of the Egyptian samples by saying much, but there was no doubt that in Egypt
all the dyes tried so far in the cloth were simply bleached in six weeks of ordinary
weather, unless they were protected by aluminium or some other pigment. The
dyes used in rubber undoubtedly preserved the nature of the rubber, but did not
preserve the cloth, and of all the samples received, those coated with a heavy
litharge coating outside seemed to have stood u p best. W i t h recard to skin-lined
fabrics for airship g a s b a g s , he would like to emphasise how necessary it was to
obtain perfect adhesion between the skin and the fabric. There was an idea some
time ago that the Germans thought it sufficient if the skin was lightly attached to
the fabric, the skin merely acting as a gas-holding medium. They knew now that
such a state of affairs would rapidly throw stresses on the skin which would tear
it in all directions when the relative humidity of the air was low. He had repeated
that experiment with synthetic films of various kinds. Provided the film was
elastic and the adhesion was good, the tensile strength of the film was not so
necessary as one might imagine. Mr. Campbell had emphasised the gloom of
the clouds which were supposed at present to be obscuring the sun from airships,
but he (Colonel Richmond) felt sufficiently optimistic about the future of them to
say that gold beater's skins as a means of lining airships would not do. There
were going to be so many airships that there would not be sufficient oxen in the
world to supply the skins. In spite of the financial gentlemen who said there was
not enough money in the country, the money would be found in other countries
for aircraft development, and he was sure we should not be long in competing:
J«i»,i92i] THE AERONAUTICAL JOURNAL 35i

successfully when once the public woke u p to the possibilities of airships. W i t h

regard to substitutes for gold beater's skin, one ingenious thing the Germans had
done was to g r o w the skins by means of bacteria, and also they found the, glue
used on their fabrics during the war so good that they were now making a fabric
which consisted of the cotton and glue alone, without any skins, and he thought
that promised very well. The aeroplane chemists had helped them enormously in
the matter of outer covers, but the chief point to airship pilots, which would not
strike the aeroplane chemists as being so important was the question of super-
heating, the question of getting the maximum reflection from the fabric. The
ideal fabric was either completely transparent—which was impossible—or a perfect
reflector. Of the three fabrics which had been examined for superheating in
Egypt, viz., one with an aluminium surface, one with a zinc-white surface, and
one with a red or gold-leaf surface, the best was the zinc-white, which had super-
heating, under normal conditions of about i6deg., the aluminium had 23deg. and
the red or gold 28deg. These were merely relative figures.
W i n g Commander T. R. CAVE-BROWNE-CAVE said he did not know if they
all realised that Mr. Dyer had really written two papers—the one he had read
and the printed one. The complete paper contained an enormous amount of
information on the very extensive research work which had been done on this
subject. It would be as well if Mr. Dyer would add to the paper when printed
in the Journal the detailed references t o other people's work to which he had
referred and also give a short bibliography. W i t h those additions he believed
that the paper would be for a very long time the classic on the question of
airship fabrics.
The paper was obviously more complete in its treatment of the non-rigid
fabric than where it dealt with the fabrics of the outer covers and g a s b a g s of
rigids. T h a t was because a great deal more work had been done on the rubber-
proofed fabric of the non-rigid. T h e conclusion generally drawn from the paper
was probably that the rubber-proofed fabric presented far greater difficulties than
the outer cover and gasbag fabrics of the rigids because of the deterioration of
the rubber. One was frequently tempted to say " let us abandon the rubber and
build u p a non-rigid envelope fabric with an outer surface like that of t h e rigid
and a skin-lined interior." T h a t scheme failed because one had to get the
necessary strength of the envelope by a compound fabric. Unless a substance
such as rubber were used it was extremely difficult to build up a compound fabric
which would have a reasonably good strength when wounded. Rubber was such
an elastic adhesive that it allowed the diagonal ply to reinforce the straight
plies to a much greater extent than was possible with anything else.
Mr. Campbell was very pessimistic about the deduction to be drawn from
the sustained load tests. T h e only way in which the envelope stress could be
increased to a dangerous value was by internal pressure, not by the rigging
tensions which were limited by the lift of the ship. Excessive internal pressure
was caused by a pilot's mistake and would only last at most a few minutes. It
was not fair to assume that the envelope was going to be blown u p to its bursting
pressure for long periods. An interesting point was the question of unsewn
seams. It had been established beyond all doubt that stitching in a seam was
detrimental. It contributed nothing to the strength, and was a most prejudicial
feature in regard to gastightness. But it did not look nice to fly about the
heavens with seams that were stuck and not sewn, and for that sentimental
reason non-rigid envelope seams were sewn. It was a curious state of affairs.
One difficulty in tackling the question of the deterioration of rubber-proofed
fabrics was the difficulty of reproducing the deterioration in an intensified form.
W h e n studying the deterioration of cellulose one could use mercury vapour lamps
with fair confidence, but there was no artificial means of producing the actual
deterioration of rubber in a more rapid form. He was rather disappointed t h a t
they had not had any remarks on that point from the rubber manufacturers. The
352 THE AERONAUTICAL JOURNAL Uuly, 1921

paper was, perhaps, rather pessimistic in that it drew attention to the various
difficulties that were experienced, but more particularly with regard to rubber-
proofed fabrics the development that had taken place was most creditable. In
1915 they started making envelopes of single-ply—a fabric designed to burst at
120 mm. pressure. They burst, however, when actually proof-tested at 25 mm.
That was the state of affairs when Adam, Ritchie and Dyer took over the question.
The last non-rigid deflated was the N.S.7. She was inflated in April, 1919,
deflated for a purely mechanical defect three months afterwards, inflated again
with the same envelope, and deflated three weeks ago because she was no longer
wanted for service. To keep an envelope inflated for three years and have it still
in an effective condition was a result to be proud of. Mr. Dyer was now free
to get on with the question of the rigid fabrics, and he thought they were all
confident that he could meet with corresponding success.
M. J. L. LAKE said that in going through an Elizabethan Grammar School
last summer he found in the Hall the words " Disce aut discede " — " Learn or
get out.'' The rubber manufacturers had done their best both a few years before
the war and during the war to learn the requirements of balloon and airship
work, and they had no desire to get out in spite of what had been said in the
House during the last few days. They felt that there was a great future for
airships.
He would like, as being perhaps the only representative of the manufacturers
present at the lecture and only a humble representative of the trade, to thank
Mr. Dyer for his learned paper which was full of most valuable information and
would be closely studied by the rubber manufacturers.
As one who had come into contact with the Government officials, he would
like to remark how very anxious they were to help the manufacturers now that
they had more time than they had during the war. The manufacturers appreciate
the ready help they were always willing to give them.
The CHAIRMAN invited other members to contribute remarks in writing,
announced that the Author would reply to the discussion in the Journal, and
proposed a vote of thanks for the paper.

SUMMARY OF POINTS RAISED IN A WRITTEN COMMUNICATION

FROM MR. A. D. RITCHIE, FORMERLY CHIEF AIRSHIP CHEMIST.

1. Avorio considers the dimensional effect observed by him to be due to

seams. Bursting tests on envelopes point to some, and possibly an important,
weakening due to seams.
2. The bursting tests at Kingsnorth were done with a rate of loading less
than 3olbs./in/min. ; at least 10 minutes were taken to reach bursting pressure.
3. Considerable weight, in my opinion, is to be. attached to the statistical
evidence as to the relation of weight and permeability.
(i.) It is shown clearly that results with one specification are very variable,
so that comparisons of specifications based on a few tests are misleading,
(ii.) As against the argument that the apparent similarity of permeability
for different weights is due to different degrees of care in manufacture,
I should suggest that a bad maker will always get many high permea-
bilities with light proofiings because they are harder to make, and it is
only good and uniform workmanship that will show the similarity. At
any rate, the following facts must be taken into consideration :—
(a) For some time permeability was not specified by the Admiralty;
there was only a general clause about workmanship in the
contract.
July, 1921] THE AERONAUTICAL JOURNAL 353

(b) The firm Worst for permeability was best for regularity of weight
(i.e., careful spreading).
(c) T h e minimum permeability (4 litres) was strikingly regular for
all makers and all specifications.
(d) Consideration of results for one specification showed no correla-
tion between variations of weight and variations of permeability.
(e) Some fabrics with as little as 50 g m s / m 2 between plies had
good permeability.
Owing to the nature of the fabrics made there were no
sufficient means to distinguish effects due to variations in
weights of a single layer of proofing from those due to variations
in number of layers or situation (i.e., between plies or on face).
4. Is loss of sulphur on weathering due to volatilisation?
5. Impregnation of the cotton with rubber would largely prevent seam
leakage, but the practical difficulties of impregnations are enormous.
6. Do the Germans soften their goldbeaters skin with glycerine? Do they
use Lecithin as I suspected ?
P. M. MATTHEW (communicated) : The three questions of chief interest to
proofers at the present time are :—
(1) The durability of proofing under tropical conditions.
(2) Its protective action on the cotton fabric.
(3) The efficiency of the seams in the various fabrics.
As the result of experience at hdme and in France during the past eight
years in connection with balloon and airship fabrics, and during the past forty
years in the Tropics with waterproof fabrics, I have no hesitation in coming to
the conclusion that durability is largely, if not entirely, a matter of vulcanisation,
the best method and " o p t i m u m d e g r e e " of which have yet to be determined.
Perfect vulvanisation, or " optimum cure " as it has been called, is generally
understood to mean the production in rubber of its maximum resiliency and tensile
strength. While it cannot be said t h a t this is in inverse ratio to its durability
under severe weathering conditions, I incline to the belief that something of the
kind is the case. In saying so, I have in mind the result of a series of
" w e a t h e r i n g " tests made during the summers of 1917 and 1918 with various
proofings made—
(1) W i t h pure unvulcanised rubber.
(2) W i t h rubber vulcanised to various degrees.
(3) W i t h unvulcanised (but vulcanisable) rubber.
As regards the best system of vulcanisation, which is perhaps still an open
question, I incline very strongly to dry heat for g a s proofing. I notice you
speak of finding a deposit of sulphur in the cotton of certain proofed fabrics. I
think you will find it generally present in those which have been vulcanised by
moist heat, in which the nature of the process prevents its escape.
As regards the various forms of protective coating on the outer surface of
the fabrics, it would appear that so far a suitable combination of rubber and
aluminium is the most effective. Here again the " optimum degree " of vulcanisa-
tion plays a highly important part in the matter of durability and incidentally,
of course, in fixing the aluminium. I may mention that in the case of our
samples (spec. No. 23 and 32) now being tested in Egypt, a red dye soluble in
rubber is incorporated in the aluminium mixing, no dye being used in the cotton.
Your remarks on the subject of seams are extremely interesting, and all
the observations which I made, both in this country and in F r a n c e during the
war, point to defective seams as the chief source of leakage in kite balloons.
354 THE AERONAUTICAL JOURNAL Uutv.mi

By the omission of stitching this would be greatly reduced, as also the risk of
d a m a g e to the fabric in making up. It has always appeared to me a s the result
of experience in cognate manufacture that with a gastight inner facing of rubber
on the fabric, and tape with a similar weight of facing, it should be possible
to make a perfect seam, and at the same time to reduce the width of the tape
now in use.
W i t h regard to g a s b a g s for rigids, it is interesting to note t h a t their life
is that of the skin lining, and that this is probably lengthened by the use of
rubber in place of glue as an "adhesive. Is it possible that the trouble caused
by the unequal shrinkage of the cotton fabric and t h e skins would be lessened
were the uniting solution vulcanised? It would almost seem that were this
possible it would reduce the tendency of the two to separate. It seems possible
that if the cotton fabric were well stretched in scouring and finishing (though
not so as to unduly reduce its strength) it might to some extent contract again
with the shrinkage of the skins. The drying-up and loss of pliability in the skins
is presumably due to decomposition or dessication of the animal matter. Could
this be retarded by applying some antiseptic or preservative along with the
glycerine ?
I trust you will pardon my troubling you with a rather long letter dictated
by interest in your work, and your very wise suggestion that t h e problems in
hand call for empirical as well as scientific methods of treatment.

REPLY TO POINTS RAISED IN T H E DISCUSSION.

I do not think there are any figures on the subject of three-ply fabrics
referred to by Mr. Campbell. If it is desired to reduce the weight of the
middle (bias) ply of a three-ply fabric it will be necessary t o consider what
degree of protection from tearing is to be provided.
I agree with Dr. Barr that extensibility is important, but so far as I know
it has scarcely been studied. Different pieces of fabric by the same maker have
different extensibilities for low loads such as are met with in practice. In
sustained load tests partial loss of adhesion of plies may be operative, but the
tests on single ply fabrics (linen in Table V I I . and cotton in Table VI.) show
that there is an effect independent of this ply separation. Dr. Barr's correction
as to the inferences from Aston's experiments is accepted. My statement needed
amplification. Dyes for protection of cotton with the aluminium facing as well
have not, I think, been tried. Of the much lowered permeability as reported
by Bureau of Standards I have met one case. Unfortunately I had very little
of the material. The rubber was unoxidised but very soft.
Major Richmond refers to a subject of great interest and greater difficulty,
viz., the mechanism of permeability. It is clear that the difference between
gases in respect of this property are not explicable on any of the recognised
characteristics of their molecules, but that a reaction, or perhaps better, a trans-
action, with the substance of the permeable film is also concerned in the total
effect. With regard to prevention of superheating and the use of zinc-white or
aluminium, I think a knowledge of the nature of the aluminium surface and of
the aluminium powder used is very important for the comparison. Figures have
been given for aluminium and zinc white reversing the order of efficiency stated
by Major Richmond, and further trials may be necessary.
W i n g Commander Cave-Browne-Cave refers to the desirability of a list of
references t o original papers. One that had been prepared is appended. It
is not so complete as one would wish. A large amount of the work on non-rigid
fabrics is contained only in departmental report principally from Kingsnorth and
Manchester and few of these were published by the A.C.A. The same is true
of work on skin-lined fabrics and outer covers. W i t h the exception of three or
J«i»,i92ti £HE AERONAUTICAL JOURNAL 355

four papers therefore the bibliography largely refers to work done for aero-
plane fabrics, but having either a general or particular bearing on some problem
of airship fabrics.
With regard to the points raised by Mr. Ritchie :—(2) If the comparison
tensile test had been done at an equivalent rate of loading the ratios T b /T t would
have higher than those given. ^
(3) Possibly I gave the statistical evidence on the weight-permeability
relation less than its deserts, but that is not to be regretted since it has called
forth such a clear statement of the case from Mr. Ritchie. These points have
to be taken seriously into account, though I still think the question not proven.
Comment must be limited, but the reasons for failure to carry conviction to my
mind are briefly :—
(1) There are few or no data on the weight-permeability relation for a
single layer of proofing covering a reasonably wide range of weights.
(a) There are such data (not very numerous it is true) for films and sheets of
rubber, and a roughly linear relation between weight and reciprocal of per-
meability has been stated, although various methods of preparation of film and
sheet were involved. Further data of this kind should be obtained to place
the matter beyond doubt.
(3) With regard to workmanship I should compare not so much one firm
with another, but the same firm (i.e., very likely the same workman).at work
on light and then on heavy proofings. Long practice in the art probably enable^
the spreader to boil it down to '' much rubber, less care; little rubber, more
care." I do not assert that this is the case, but only that it is a possible
explanation.
(4) I think sulphur may be, and probably is, lost by volatilisation on
exposure, but that oxidation takes place as well and may predominate. My
reasons are (i) that in a fabric exposed from January to June the rate of loss was
no higher in the hot weather than in the cold; and (ii) that from a fabric
kept during three months in a room always warm and for eight hours of the
day at 400 C. the loss was less than from a piece of the same fabric exposed to
the weather.
(5) I agree that impregnation of the cotton in the full sense of the term is,
impracticable. What was meant in the paper was the filling of interstices
between the yarns. It was suggested by proofers that this might be an improve-
ment. The full test has not yet been carried out, but a stitched seam made in
a piece of two-ply D fabric, where the interply rubber had for some reason
unknown been very thoroughly pressed into the cloth, gave a very much lower
seam permeability than is usual with stitched seams in unfaced fabrics, though
presumably this was not for the cause which might be expected to operate in a
genuinely impregnated cotton.
(6) Yes, glycerine is used in German skin-lined fabric. No further search
for lecithin has been made.
Mr. Matthew's remarks on the importance of correct vulcanisation are
interesting, and I should like to read a full account of the experiments to which
he refers, because it is a subject on which nothing seems to have been published
where maintenance of gas tightness is the principal property under consideration.
There appears to be no doubt that an inner facing of rubber is highly
beneficial in securing gas-tight seams, where stitched seams are concerned.
My remarks on rubber versus glue for skin adhesive were meant to favour
the latter, and experiments on most examples of the two types support my
argument. A vulcanised rubber as adhesive should certainly be tried.
The desiccation of the skins on exposure in hot climates seems to reach a
354 THE AERONAUTICAL JOURNAL Uuly.im

By the omission of stitching this would be greatly reduced, as also t h e risk of

c'amage to the fabric in making up. It has always appeared to me as the result
of experience in cognate manufacture that with a gastight inner facing of rubber
od the fabric, and tape with a similar weight of facing, it should be possible
to make a perfect seam, and at the same time to reduce the width of the tape
now in use.
W i t h regard to g a s b a g s for rigids, it is interesting to note t h a t their life
is that of the skin lining, and that this is probably lengthened by the use of
rubber in place of glue as an adhesive. Is it possible that the trouble caused
by the unequal shrinkage of the cotton fabric and the skins would be lessened
were the uniting solution vulcanised? It would almost seem that were this
possible it would reduce the tendency of the two to separate. It seems possible
that if the cotton fabric were well stretched in scouring and finishing (though
not so as to unduly reduce its strength) it might to some extent contract again
with the shrinkage of the skins. The drying-up and loss of pliability in the skins
is presumably due to decomposition or dessication of the animal matter. Could
this be retarded by applying some antiseptic or preservative along with the
glycerine ?
I trust you will pardon my troubling you with a rather long letter dictated
by interest in your work, and your very wise suggestion that the problems in
hand call for empirical as well as scientific methods of treatment.

REPLY TO POINTS RAISED IN T H E DISCUSSION.

I do not think there are any figures on the subject of three-ply fabrics
referred to by Mr. Campbell. If it is desired to reduce the weight of the
middle (bias) ply of a three-ply fabric it will be necessary t o consider what
degree of protection from tearing is to be provided.
I agree with Dr. Barr that extensibility is important, but so far as I know
it has scarcely been studied. Different pieces of fabric by the same maker have
different extensibilities for low loads such as are met with in practice. In
sustained load tests partial loss of adhesion of plies may be operative, but the
tests on single ply fabrics (linen in Table V I I . and cotton in Table VI.) show
that there is an effect independent of this ply separation. Dr. Barr's correction
as to the inferences from Aston's experiments is accepted. My statement needed
amplification. Dyes for protection of cotton with the aluminium facing as well
have not, I think, been tried. Of the much lowered permeability as reported
by Bureau of Standards I have met one case. Unfortunately I had very little
of the material. The rubber was unoxidised but very soft.
Major Richmond refers to a subject of great interest and greater difficulty,
viz., the mechanism of permeability. It is clear that the difference between
gases in respect of this property are not explicable on any of the recognised
characteristics of their molecules, but that a reaction, or perhaps better, a trans-
action, with the substance of the permeable film is also concerned in the total
effect. W i t h regard to prevention of superheating and the use of zinc-white or
aluminium, I think a knowledge of the nature of the aluminium surface and of
the aluminium powder used is very important for the comparison. Figures have
been given for aluminium and zinc white reversing the order of efficiency stated
by Major Richmond, and further trials may be necessary.
W i n g Commander Cave-Browne-Cave refers to the desirability of a list of
references to original papers. One that had been prepared is appended. It
is not so complete as one would wish. A large amount of the work on non-rigid
fabrics is contained only in departmental report principally from Kingsnorth and
Manchester and few of these were published by the A.C.A. The same is true
of work on skin-lined fabrics and outer covers. W i t h the exception of three or
July, mil j?HE AERONAUTICAL JOURNAL 355

four papers therefore the bibliography largely refers to work done for aero-
plane fabrics, but having either a general or particular bearing on some problem
of airship fabrics.
With regard to the points raised by Mr. Ritchie :—(2) If the comparison
tensile test had been done at an equivalent rate of loading the ratios T b /T t would
have higher than those given.
(3) Possibly I gave the statistical evidence on the weight-permeability
relation less than its deserts, but that is not to be regretted since it has called
forth such a clear statement of the case from Mr. Ritchie. These points have
to be taken seriously into account, though I still think the question not proven.
Comment must be limited, but the reasons for failure to carry conviction to my
mind are briefly :—
(1) There are few or no data on the weight-permeability relation for a
single layer of proofing covering a reasonably wide range of weights.
(2) There are such data (not very numerous it is true) for films and sheets of
rubber, and a roughly linear relation between weight and reciprocal of per-
meability has been stated, although various methods of preparation of film and
sheet were involved. Further data of this kind should be obtained to place
the matter beyond doubt.
(3) With regard to workmanship I should compare not so much one firm
with another, but the same firm (i.e., very likely the same workman).at work
on light and then on heavy proofings. Long practice in the art probably enable^
the spreader to boil it down to "much rubber, less care; little rubber, more
care." I do not assert that this is the case, but only that it is a possible
explanation.
(4) I think sulphur may be, and probably is, lost by volatilisation on
exposure, but that oxidation takes place as well and may predominate. My
reasons are (i) that in a fabric exposed from January to June the rate of loss was
no higher in the hot weather than in the cold; and (ii) that from a fabric
kept during three months in a room always warm and for eight hours of the
day at 400 C. the loss was less than from a piece of the same fabric exposed to
the weather.
(5) I agree that impregnation of the cotton in the full sense of the term is
impracticable. What was meant in the paper was the filling of interstices
between the yarns. It was suggested by proofers that this might be an improve-
ment. The full test has not yet been carried out, but a stitched seam made in
a piece of two-ply D fabric, where the interply rubber had for some reason
unknown been very thoroughly pressed into the cloth, gave a very much lower
seam permeability than is usual with stitched seams in unfaced fabrics, though
presumably this was not for the cause which might be expected to operate in a
genuinely impregnated cotton.
(6) Yes, glycerine is used in German skin-lined fabric. No further search
for lecithin has been made.
Mr. Matthew's remarks on the importance of correct vulcanisation are
interesting, and I should like to read a full account of the experiments to which
he refers, because it is a subject on which nothing seems to have been published
where maintenance of gas tightness is the principal property under consideration.
There appears to be no doubt that an inner facing of rubber is highly
beneficial in securing gas-tight seams, where stitched seams are concerned.
My remarks on rubber versus glue for skin adhesive were meant to favour
the latter, and experiments on most examples of the two types support my
argument. A vulcanised rubber as adhesive should certainly be tried.
The desiccation of the skins on exposure in hot climates seems to reach a
i3o"6 THE AERONAUTICAL JOURNAL lJuiv.im

point at which pliability cannot be restored by fresh treatment with glycerine.

Whether this is animal decay or attack by decomposition products of the rubber
adhesive and proofing- does not seem clear.

BIBLIOGRAPHY.
In an account, such as the foregoing, which is principally a description of
broad results based on much detailed experiment it merely confuses to refer to
individual original papers pertinent t o the subject under discussion a t the moment.
A list of the more important papers is therefore appended, roughly classified.
On some important matters there are no published reports available, the only
papers being confidential official ones.
The classification of the bibliography does not follow the divisions used in
the body of the report. It is sufficiently indicated by the sub-headings.
The letters T. and R. and M. are references to the well known A.C.A.
reports.
Strength (new fabrics).
T. 1204.
R. and M. 23, 29, 37, 180, 182.
Strength (effect of light, acidity).
T. 1019, 1381.
R. and M. 430, 585.
Permeability.
T. 998, 1124, 1197, 1224.
R. and M. 22, 232, 317, 360, 435, 447, 513, 584.
Weathering (non-rigid fabrics).
T. 602, 602a, 900, 1157.
R. and M. 313.
Dopes. Material (including solvents).
T. 816, 1269, 1412, 1502, 1550.
Tautness.
T. 1193.
R. and M. 569, 606.
Radiation and superheating.
T . 993-
R. and M. 329.
Bureau of Standards, technologic paper No. 128.