T H E INNERVATION O F T H E SHOULDER J O I N T ]

ERNEST GARDNER
Departnieiit of dnatom?y, Collegr. of ?&dicine, H7ny71e U i r i v e ~ i t y ,
Detroit, Micltigan

T W O FIGURES

INTRODUCTION

Certain features of the gross pattern of the nerve supply

of the shoulder joint have been known since the last century.
Nevertheless the innervation has been studied less than that
for the hip, knee and elbow, and there have been no investi-
gations of the distribution of nerves after they inter the
capsule.
Rudinger (1857) carried out the first systematic study of
the nerve supply of this joint. Ha found it to bc derived
from the axillary and suprascapular nerves. One branch of
thc former arose in the axilla and descended t o the antero-
inferior region of the capsule. Another branch arose from the
axillary nerve as it began its circumflex course and supplied
the posteroinferior region of the capsule. The latter distri-
bution was especially pronounced when t'he inferior branch of
the suprascapular nerve was lacking. He observed one case in
which a muscle bundle arose from the coracoid process and
anterior aspect of the capsule. A branch from the brachial
plexus supplied this muscle and also gave a twig to the
superior region of the capsule. He described superior and
inferior branches from the suprascapular nerve. The super-
ior one coursed laterally in the supraspinous fossa, gave a
twig to the coracoromial ligament aiitl tlieii split into 3
'Aided by a grant from the Division of Research Grants and Fellowships,
Piational Institute of Health, U. S. Public Health Service.

3
T H E ANATOMICAL HLSCORD. VOT,. 102. S O . 1
SEI'TEXBBR, 1948
2 RICHARD GENTRY JACKSON, M.D.

2. The presence of anterior or posterior plexuses on the

stomach.
3. Whether or not the posterior trunk is distributed to all
o r to only a part of the posterior surface of the stomach.
Table 1summarizes the statements of the different anatomists.
Agreement is shown in the following :
1. The anterior gastric vagal trunk and the posterior
trunk each contain fibers from both the right and left vagus
nerves.
2. Branches leave the anterior trunk just below the dia-
phragm and pass through the gastrohepatic omentum to the
liver.
3. Gastric branches of both the anterior and the posterior
trunks are distributed mainly along the lesser curvature of
the stomach.
METHODS

I n this study the vagus nerves were first dissected above

the diaphragm in 50 cadavers, starting just below the pul-
monary plexus and following them to the diaphragm. The
specimens used were those being dissected by freshman med-
ical and postgraduate students. I n every case the dissection
mas done before the student reached the area in his dissection,
Ihe nerves were charted and, after the heart had been re-
moved and the student was ready for the dissection of the
esophageal plexus, they were checked again to be sure that
n o branches had been missed. After the student had studied
the nerves, a number of dissections were photographed to
show the disposition of the main trunks and branches. In
some instances branches recorded in the original diagram
were broken by the students in their examinations and there-
fore do not appear in the photographs, but do appear in the
drawing beside the photograph. The photographs show the
size of the trunks and their approximate location on the esoph-
agus. Only the main trunks and divisions, the communicating
branches and the larger esophageal branches were charted ;
the pericardial, the diaphragmatic and the aortic branches
 NERYE SUPPLY O F SHOULDER JOIST 3

from the subscapular nerves are mentioned as an additional

source. I n Cunningham's Text-Book of Anatomy ('43), in
the section on joints, the nerve supply is said to arise from the
suprascapular, subscapular and axillary nerves, but in tlie
section on the nervous system there is nothing regarding
articular filaments from the subscapular nerves.
The most recent investigation appears to be that of Ark-
hangelsky ( '31) who dissected 40 shoulder joints. According
to him, the nerve supply is derived from the axillary, snpra-
scapular and anterior thoracic nerves. He states that the
axillary nerve always gives a branch about 3 em cranial to
the inferior border of the subscapularis muscle which descends
to the inferior border of this muscle and there diTrides into
twigs which are given to the anteroinferior region of the
capsule. I n about half the cases, he found thin branches which
approached the inferolateral region of the capsule near the
surgical neck of the humerus. A branch to the bicipital sulcus,
as described by Rauber, was much less common. Periosteal
and articular branches arise from the suprascapular nerves in
both supraspinous and infraspinous fossae. Arkhangelsky
dwelt at some length on the articular branches of tlie anterior
thoracic nerves. One of these supplies tlie acromial process
and probably gives fine twigs to the shoulder joint as well.
Another branch, which occurred in about 40% of the dissec-
tions, ascends in the bicipital sulcus.
So far as could be determined, there have been no complete
studies of the nerve supply of this joint in lower animals.
Ghetie ('39) mentions an articular branch of the axillary
nerve to the shoulder joint of the horse. Sasaoka ('38-'39)
mentions a similar branch in the cat.

MATERIALS A N D METHODS

Nerves were dissected in 7 adult shoulder joints so as to

establish the general pattern of distribution. Four other
joints were partially dissected to confirm points brought out
by study of fetal joints. Four fetal joints were sectioned
4 ERNEST GARDNER

serially at 10 p and stained by a modified llassoii’s method.2

Two were from a 11-week fetus and two from a 12-week fetus.
All were left attached to the trunk and sectioned in a plane
transverse to the trunk. The ages given are only approximate
since they were determined on the basis of crown-rump mea-
surements of fixed specimens. While the material was not
satisfactory for the study of nerve endings, it was excellent
for tracing nerve fibers, since even in very small bundles there
were usually a few large fibers which were easily distin-
guished.
Diagrammatic sketches were made of representative dis-
sections, and projection tracings were made of fetal sections
which showed characteristic features of the nerve supply.
OBSERVATIONS

The iiiajor features of existing descriptions were confirmed

and, in addition, branches hitherto undescribed were found.
The nerve supply is derived from the suprascapular, axillary
and anterior thoracic nerves, from the posterior cord of the
brachial plexus, directly from the stellate and perhaps other
sympathetic ganglia and possibly from the radial nerve.
I n the illustrations, both of gross and microscopic material,
branches are named according to their most specific origin.
Thus, “suprascapular” (fig. 2a) indicates articular twigs
from the suprascapular nerve.

The articular bralzches o f the alr;illary n c w e

Shortly after the ori’gin of the axillary nerve from the
posteyior cord of the brachial plexus a small branch arises
~vliich descends laterally across the subscapularis muscle,
turns around the lower edge of the muscle and divides into
twigs which enter the inferior region of the capsule (fig. 1,
ant.). Some of the terminal ramifications of these twigs mere
The first 6 dissections were done while the author was in the Department of
Anatomy, Stanford University, California, 1940-1941. The fetal material was
obtained through the courtesy of Prof. C. H. Danforth, Department of Anatomy,
Stanford Unirersity, California.
 NERTE SUPPLY O F SHOULDER JOIXT 5

,
:' ant. i b I : post.'...

C : post. '-\

From- ant. .- - - -.__

.*-

:' ant. \ d
Fig. 1, a-d Anterior aud posterior views of different types of distribution of
nerves to the shoulder joint. The articular capsule is represented by stippling.
(1, ant. S ~ O W Sthe articular branch of the lateral anterior thoracic nerre and
from the posterior cord.
6 ERNEST GARDNER

traced superiorly into the anterior and posterior regions of

the capsule. In one case a tiny filament was found wliich ac-
companied a branch of the anterior circumflex artery into the
bicipital sulcus (fig. l c , ant.).
In the fetal joints, the articular branch of the axillary nerve
has a similar course. It gives a twig which accompanies the
anterior circumflex humeral artery into the bicipital sulcus
and ascends until branches of this vessel enter the head of the
humerus (fig. 2c-h). Beyond this point the nerve fibers could
not be traced. During the initial part of its course, the bicip-
ital branch anastomoses with a bundle of fibers derived from
the posterior cord, and possibly with still another bundle de-
rived from the radial nerve (fig. 2j). Numerous twigs from
the main articular branch of the axillary nerve ramify in the
inferior region of the capsule (fig. 2i-k) and, in addition, give
small branches to the inferior edge of the subscapularis mns-
cle. Still other twigs from this articular nerve ascend in the
anterior and posterior regions of the capsule, adjacent to
their humeral attachments (fig. 2g-i) . These accompany blood
vessels and both vessels and nerves give branches which pene-
trate the fibrous layer of the capsule and enter the synovial
layer and adjacent portions of the humerus. The fibers wliich
ascend posteriorly reach the area of distribution of the lower
articular branch of the suprascapular nerve. The fibers which
ascend anteriorly overlap the area supplied by the branch
from the posterior cord.
AIaiiy of the small nerve bundles which enter the inferior
and anteroinferior regions of the capsule course in the fibrous
layer where they appear to terminate. These bundles are nu-
merous in comparison with those which accompany blood ves-
sels into synovial tissue. The arrangement is thus similar
to that found in the posterior region of the capsule of the knee
joint (Gardner, '48) and in the anteromedial region of the hip
joint (Gardner, '48). The articular twigs from the posterior
cord also contribute to this arrangement (see below).
 NERVE SUPPLY O F SHOULDER JOINT 7

Biceps tendon
,/ /
Ornohyoid
Ant. thoracic

Fig. 2, a-d Projection tracings of a fetal joint (sectioned in a plane trans-

verse to the trunk). Nerves are shown in solid black. The sections go'from cranial
t o caudal levels.
8 ERNEST GARDNER

Dorsal root qanq.

of axillary a
post. cord fibers
Fig. 2, e-h Coiitiiiuat,ion of the fetal sections.
BEBVE SUPPLE’ OF SEIOVLDEI: JOIST 9

Fig. 2, i-1 Continuation of t h e fetal sections.

10 ERNEST GdRDNER

T h e possible articular branches of the radial nerve

In one fetal joint a small twig arose from the radial nerve
near the origin of the profunda branch of the brachial artery.
I n another joint a similar branch arose somewhat more super-
iorly. I n both cases, the nerves ascended in the adventitia
of the brachial artery to the origin of the anterior circumflex
humeral artery (fig. 251). Here each seemed to anastornose
with the bicipital branch of the axillary nerve. It was difficult
to determine the ultimate distribution of the fibers from the
radial nerve and to be certain that they entered the bicipital
sulcus.
I n the same two fetal joints, a twig of the radial nerve
ascended in the long head of the triceps brachii, together with
an ascending branch of the profunda artery. It reached the
junction of the tendon of the long head and the capsule of the
shoulder joint and disappeared just as the capsule was
definitely reached (fig. 2k-1). Whether or not the fibers ac-
tually entered the capsule could not be determined in this
material.

The articular blraizch of the posterior cord of the

brachial plewits
This R ~ initially
S discovered in fetal material. I t s articular
distribution had been overlooked in the first dissections since
it appeared to supply the subscapularis muscle and its tendon.
In the fetal material it arises from the posterior cord, just
proximal to its termination and apparently from the portion
giving rise to the axillary nerve. It descends obliquely across
the ventral surface of the subscapularis muscle (fig. 2f).
When it reaches the tendon of this muscle it divides into a
number of twigs. One or two of the twigs anastornose with
the bicipital filament of the articular branch from the axillary
nerve (fig. 2g). A few others pierce the tendon of the sub-
scapularis muscle and enter the anterior region of the cap-
sule. This area is also reached by fibers ascending from the
articular branch of the axillary nerve. Fibers from both
 NERVE SUPPLY OF SHOULDER JOIST 11

nerves ramify mainly in the fibrous layer of the capsule, but

some of them penetrate deeply to the synovial layer in com-
pany with blood vessels. In the region of division of the
articular branch of the posterior cord, a small group of fibers
detaches itself and anastomoses with a group of fibers whose
origin is discussed below. The combined bundle of fibers
ascends in company with blood vessels to the joint capsule
superior to the tendon of the subscapularis muscle and pos-
terior to the coracoid process (fig. 2c-h). Here it ramifies in
an area characterized by a rather extensive vascular network.
The same area is also reached by fibers from the suprascap-
nlar and anterior thoracic nerves.
The articular distribution of this branch of the posterior
cord was later confirmed in dissections. It corresponds i n
most respects to the upper branch of the asillary nerve dc-
scribed by Rudinger (1857). It is not unlikely that a larger
series would show it arising from the axillary nerve as well
as from the posterior cord. Only figure Id, ant. shows the
articular branch from the posterior cord in its gross relation-
ships. The other dissections illustrated were done before the
existence of this branch was realized.

T h e articular bramhes of the aiaterior thoracic nerves

An articular branch of the lateral anterior thoracic nerve
was first noted in fetal joints and subseqnently confirmed in
dissections. I n its gross distribution it accompanies a branch
of the thoracoacrornial artery, crosses the coracoid process
superiorly and reaches the acromioclavicular joint (fig. Id,
ant.). During the terminal part of its course it gives a twig
to the anterosuperior region of the joint capsule. In one in-
stance, a very slender filamcnt arose from the main articular
branch soon after its origin and ran laterally, anterior to the
coracoid process, and entered the bicipital sulcus (fig. Id,
ant.). The other dissections illustrated in figure 1were done
before the existence of the articular branch of the lateral
anterior thoracic nerve was realized.
12 ERKEST 0.4RDNER

The distribution described above is similar to that fouiicl

in fetal joints in which it was observed that the fibers enter
the anterosuperior region of the capsule. Here they form
a number of small bundles, some of which ramify in the
fibrous layer where they overlap the area of distribution of
tlie upper articular branch of the suprascapular nerve. The
ramification is similar in arrangement to that formed by tlie
avillary nerve and posterior cord in the anteroinf erior region
of the capsule, although the numbers of nerve bundles are
not as great. Other fibers from the articular nerve penetrate
deeper into tlie capsule and become closely associated .with
the vascular network in the synovial tissue posterior to the
coracoid process. This region is also reached by fibers whose
origin is as follows.
I n one of the fetal joints an additional articular twig of the
lateral anterior thoracic nerve was found. This twig, which
was not found in any of the dissections, accompanied small
blood vessels across the subscapularis muscle and anasto-
mosed with sympathetic fibers descending in the adventitia of
the axillary artery. The combined trunk anastomosed with a
twig from the articular branch of the posterior cord and the
bundle of fibers resulting from this union ascended to the
vascular area posterior to the coracoid process (fig. 2c-h). The
contribution of the posterior cord was present in all the fetal
joints, while that of sympathetic fibers was found in two fetal
joints.
Syrnpatlzet ic fibers
I n two of the fetuses sympathetic fibers were traced into the
shoulder joint. In one, they arose from the stellate ganglion;
in the other from the sympathetic trunk just superior to the
stellate ganglion. I n both, the fibers descended in the adven-
titia of t h e subclavian and axillary arteries to the origin of
vessels ascending to the vascular area of the joint. Here
they entered into the formation of the ascending nerve bundle
described above. I n one joint the sympathetic fibers anasto-
inosed with a twig from the lateral anterior thoracic nerve.
 BERVE SUPPLY OF SHOTTLDER JOINT 13

Thus there is a direct path for sympathetic fibers to the

shoulder joint. Whether this is a fairly constant feature
cannot be stated at this time. I n the other two fetal joints,
fibers of similar origin mere found, but they were not as well
stained and were lost in the adventitia of the axillary artery.

The articular brafiches of the swprascapulm- Iierre

As this nerve enters the supraspinous fossa it gives a branch
which runs laterally, deep to the tendon of the supraspinatus
muscle. During its course it gives a twig to the periosteum of
the coracoid process and frequently one to the coracoacromial
ligament (fig. 1, post.). It ends as fine rami in the superior
region of the capsule, near the cartilaginous glenoid lip. In
the fetal joints, this nerve accompanies branches of the trans-
verse scapular artery into the fibrous layer of the capsule
near its fusion with the tendon of the supraspinatus muscle
(fig. 2a-b). It ramifies in this area and reaches anteriorly
into the distribution of the articular branch of the anterior
thoracic nerve, and probably into the vascular area as well.
The number of nerve fibers in the superior region of the
capsule is exceeded only by those in the anteroinferior area.
The articular branch also gives a twig to the acromioclavicular
joint. The fibers from the suprascapular nerve to the supra-
spinatus muscle reach as far as its tendinous insertion but
it was not possible to determine whether any of them con-
tinued into the capsule. Sympathetic fibers were observed ac-
companying the transverse scapular artery but it was not pos-
sible to trace them to the point of origin of articular vessels.
Just before entering the infraspinous fossa, the suprascap-
ular nerve gives a twig to the neighboring periosteum and one
or more filaments which run laterally and enter the capsule
deep to the tendons of the infraspinatus and teres minor mus-
cles (fig. 1, post.). I n the fetal joints, these fibers, together
with the blood vessels which they accompany, ramify in the
capsule adjacent to the glenoid lip (fig. 2c-f). Small bundles
14 ERNEST GARDNER

of fibers occasionally enter the synovial tissue. The lowermost

bundles overlap the area reached by fibers ascending from the
articular branch of the axillary nerve.
DISCUSSION

There are certain features of the nerve supply of the

shoulder joint which in many respects are similar to those
of the knee and hip. The nerve supply is characterized by
variation in gross distribution of nerves, by regional differ-
ences in intra-articular distribution and by a certain amount
of overlap of articular nerves.
Variation in distribution of articular nerves is evident
mainly in the manner in which they reach the joint. Once
within the joint there is a basic pattern of distribution in that
each nerve usually supplies a specific region of the joint.
The absence of certain articular branches, as determined by
dissection, is more apparent than real since many of these are
small and easily overlooked in dissection. For example, the
bicipital branch of the axillary nerve was found in but one of
the dissections, yet was present in all of the fetal joints.
The bicipital branch of the lateral anterior thoracic nerve is
less constant. According t o Arkhangelsky ('31) it is present
in 40% of the cases. I n this series it was found in but one of
the dissections. Also according to Arkhangelsky, at least. one
branch from the lateral anterior thoracic nerve supplies the
superior region of the shoulder joint. In the present study it
was found in both gross and microscopic material, so there is
little doubt that it is usually present. The other articular
branch of the lateral anterior thoracic nerve, that anastomo-
sing with sympathetic fibers and fibers from the posterior
cord, was present in but one of the fetal joints and is un-
doubtedly quite inconstant.
Riidinger (1857) described an upper branch from the axil-
lary nerve to the anterior region of the capsule. I n the
present study the branch which corresponds in distribution
to Riidinger 's description arose from the posterior cord, al-
though from that portion of the cord giving rise to the axillary
 NERVE SUPPLY O F SHOULDER JOIXT 15

nerve. Arkhangelsky did not mention such a branch, either

from the asillary nerve or from the posterior cord. I n gross
material, this branch appears to supply the subscapularis
muscle and its tendon. The fetal material, however, showed
that it either pierces the tendon or courses below it to reach
the capsule. It may be, therefore, that he assumed it to be a
muscular branch. It is quite likely that a series larger than
the present one would show it arising either from the postcr-
ior cord or the axillary nerve.
Each nerve entering the shoulder joint supplies a definite
region, but there is a certain amount of overlap with other
nerves, so that each region is supplied by more tliaii one nerve.
Thus, the antcroinferior region of the joint is supplied by the
asillary nerve, by the posterior cord and possibly by the radial
nerve. The anteroposterior region is supplied by the poster-
ior cord, the lateral anterior thoracic nerve and the suprascap-
ular nerve.
Many of the articular nerve fibers undoubtediy supply
blood vessels in the capsule and neighboring bone, since they
form bundles which are closely associated with the vessels
and decrease in size a s the vessels decrease. A number of the
fibers are undoubtedly vasomotor. This is borne out by the
presence of a direct path from sympathetic ganglia to ar-
ticular vessels. A similar direct path probably exists for the
hip joint (Gardner, '48). Fibers from lumbar ganglia reach
the femoral artery but their ultimate distribution is uncertain
since the artery i*eceives fibers directly from the femoral
nerve. The resulting mixture of fibers does, however, con-
tribute to articular vessels. In none of the fetuses could sym-
pathetic fibers descending in vascular adventitia be traced
more distally than the hip or shoulder.
A number of bundles of nerve fibers ramified mainly within
the fibrous layer of the capsule. This arrangement was par-
ticularly striking in the anteroinf erior and ,anterosuperior
regions of the capsule. These regions would be most subject
to compression or change of some sort during movement.
Thus, a situation similar to that in the knee and hip joints
16 ERNEST GARDNER

is present. I n these joints, the fibrous layer of the capsule

in the posterior region of the knee joint and the antero-
medial region of the hip joint contains many small bundles of
nerve fibers. I n the knee joint of the cat, the posterior region
of the capsule contains Ruffini-type nerve endings (Gardner,
'44).
The functional implications of articular nerve supply were
discussed in more detail in the study of the nerve supply of
the knee joint (Gardner, '48).
SUMMARY

The nerve supply of the human shoulder joint was studied

in dissections and in serial sections of fetal joints. The ar-
ticular nerves arise from the axillary, suprascapular and
lateral anterior thoracic nerves, from the posterior cord, from
sympathetic ganglia and possibly from the radial nerve.
The articular branch of the axillary nerve enters the infer-
ior region of the capsule where it ramifies and from which
fibers extend laterally into the bicipital sulcus and superiorly
in the anterior and posterior regions of the capsule. Recur-
rent branches from the radial nerve possibly reach both the
bicipital sulcus and the inferior region of the capsule.
The articular branch from the posterior cord supplies
mainly the anterior region of the capsule, extending both
superiorly and inferiorly. It anastomoses with the bicipital
branch of the axillary nerve, and also with sympathetic fibers
and occasionally a twig from the lateral anterior thoracic
nerve. The bundle of fibers resulting from union with the
latter two groups ascends to the anterosuperior region of
the capsule, especially to the vascular synovial tissue pos-
terior to the coracoid process.
A branch of the lateral anterior thoracic nerve supplies
the acromioclavicular joint and during its course gives a twig
t o the anterosuperior region of the capsule. In one dissection,
this nerve sent a branch to the bicipital sulcus.
I n two of the fetal joints, fibers from the stellate ganglion
and sympathetic trunk coursed in the adventitia of the suh-
 SENTI.: SUPPLY OF SHOULDER JOINT 17

clavian and asillary arteries, anastomosed with the branch

from tlie posterior cord, and eventually reached the vascular
area posterior to the coracoid process.
The suprascapular nerve gives a n upper branch to the
periosteum of the coracoid process, the coracoacromial liga-
ment and the superior region of the capsule. A lower branch
of tlie suprascapular nerve supplies the posteroinferior region
of the capsule.
Most of the articular fibers probably supply blood vessels
in the capsule and neighboring bone. The fibrous layer of the
anteroinferior and anterosuperior regions of tlic capsule con-
tain numerous small bundlcs of nerve fibers.
ACKNOWLEDGMENTS

I wish to express my thanks to Mrs. Gail Markham for her

aid in making the microscopic sections, and to Misses Evelyn
Erickson and Geraldine Chesney for their skillful preparation
of the illustrations.
LITERATURE CITED
ARKNANQELSKY,S. H. 1931 The innervation of the shoulder joint. Vestn. khir.,
29: (110. G7), 69-66.
DOCKAND VALENTIN.Cited by Q.uain and by Arkhangelsky.
CRUVFJLIIIER, J. 1844 The Anatomy of the Human Body. The first American
from the last Paris edition. By G. 8. Pattison. Harper and Bros.,
New York.
Cunningham’s Test-Hook of Ailatomy 1943 Eighth edition, edited by J. C.
Brash and E. R. Jamieson, Oxford University Press, New York.
FICK,R. 1904 Handbnch der Anatomic des Mensehen. Edited by K. ron
Bardeleben. Vol. 2, pt. 1, sect. 1. Gustav Fisrher, Jena.
G.~RDNFIR,E. 1944 The distribution and terniination of nerves in the knee
joint of the cat. J. Coinp. Neur., 60: 11-32.
1948 The innervation of the knee joint. Anat. Rec., 2 0 2 : 109-130.
1948 The innervation of the hip joint. Anat. Rec., 201: 353-372.
GHETIE, V. 1939 Die Innervation der Gelenkkapseln an den Gliedmassen des
Pferdes (fir Kliniker). Arch. f u r wks. n. prakt. Zierh., 75: 134-113.
Gray’s Anatomy of the Human Body 1942 Edited by W. H. Lewis, 24th
edition. Lea and Febiger, Philadelphia.
Norris’ Human Anatomy 1942 Tenth edition, edited by J. P. Srhacffcr. The
Blakiston Company, Philadelphia.
Piersol’s Human Anatomy 1930 Ninth edition, edited by G. Carl Huber. J. B.
Lippincott Company, Phi1:idelphia.
18 ERNEST GABDXER

Quain’s Elements of Anatomy 1909 Eleventh edition, edited by E. A. Schafer,

J. Symington, T. H. Bryce. Longmans, Green and Co., London.
MDINGER, N. 1837 Die Gelenknervrn des mcnscliliclien ICorpers. Erlangen.
Verlag roii Ferdinand Enke.
RAUBEX,A. Cited by Quain and by Arkhangelsky.
SASAOKA,S. 1938-1939 Ueber das Kaliber der niarklialtigen Nerveiif asern in1
Gelenkast. Jap. J. Med. Sc., 7 : Anat., 315-322.