The American journal of anatomy

eos rete

Spcaaeen ar ee
Sheeencrerren ery

age
Mas CRISS

er,

pay’ yet ip 4
a i, a 7 m4"

the, =

5 i elatnd
7 pert, ¥

i v

"i ee
1"

Au ity") F
Z i i
I big 3- ‘i
eas
Ey

Heh

THE AMERICAN JOURNAL

OF

ANATOMY

EDITORIAL BOARD

CHARLES R. BARDEEN GeHorGcE 8. HUNTINGTON
University of Wisconsin Columbia University
Henry H. DoNaLpson FRANKLIN P. Maui
The Wistar Institute Johns Hopkins University
Sruon H. Gace J. PuayrarrR McMourricw
Cornell University University of Toronto
G. Cart HUBER CHARLES 8S. MINoT
University of Michigan Harvard University

Groraet A. PIERSOL
University of Pennsylvania

Henry McE. Knower, Secretary
University of Cincinnati

VOLUME 13
1912

THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY
PHILADELPHIA, PA.

ee

COMPOSED AND PRINTED AT THE
WAVERLY PRESS

By rae Witiiams & WILKINS CoMPANY
Bautimore, U.S. A.

CONTENTS

1912

No. 1. MARCH

J. PARSONS SCHAEFFER. The genesis and development of the nasolacrimal
passacesmmenane ehintycome: Mouneseme see cer so: eee acer we seer

CHARLES R. Essicx. The development of the nuclei pontis and the nucleus
arcuatus in man. Twelve figures

Ruskin M. Luamon. The sheath of the sinoventricular bundle. Five figures

ALBERT Kuntz. The development of the adrenals in the turtle. Nine
ALU OS PY tor S0 coh ict RO Ec cave eh POOR Ee ET eee ae ENO 2 eres SEE

No. 2. MAY

Darmon A. RutNenart. The nerves of the thyroid and parathyroid bodies.
JEWS EF SDSS foe OP ese a IE hha rhs tlh o Renee Ala cistc eee rea (osc! a EO ©

ALEXANDER 8S. Becca. The anomalous persistence in embryos of parts of the
peri-intestinal rings formed by the vitelline veins. Five figures.........

Joun Lewis Bremer. The development of the aorta and aortic arches in
ie] OL OMS MUN UO) an "qbidss he ANE Rees Atco on coh gcc ae ma oomaans Mee sre comes

JEREMIAH S. Feracuson. The behavior and relation of living connective
tissue cells in the fins of fish embryos with especial reference to the
histogenesis of the collaginous or white fibers. Ten figures..............

H. E. Jorpan anp K. B. Steete. A comparative microscopic study of the
intercalated dises of vertebrate heart muscle. Twenty-three figures... ..

Cuas. W. GREEN. A new type of fat storing muscles in the salmon. Onco-
PANG ITS Wl Aol, AD yo) MMEUERE Sa ascoone 2 yoo doo gnc occcouesae ncee

il

bo

103

ie

129

175

iV CONTENTS

J. PARSONS SCHAEFFER. Types of ostia nasolacrimalia in man and their genetic
SigMInCAMces- HinbeeMAIGULES=) +m -.2. eetaers ccs stam ste e's ee te eh eet ste eccues 183

FREDERICK Trnney. The development of the veins and ‘ymphatics in
Tragulus meminna, erxleben.. Fourteen figures ........................ 193

Noi=3; JULY

FraNKLIN P. Matyi. On the development of the human heart. Thirty-
SO Vie Mitr Uses ee ri cny eee rel aot oe kee cit ake eR Scots 2h Ie ea ae ee eg Ske 249

Oswatp 8. Lowstey. The development o the human prostate gland with
reference to the development of other structures at the neck of the

lininany: bladder mlevenmytl CUTCS a= seg: ele irse ee rm cate ere rene eer: 299

Exror R. Ciark. Further observations on living growing lymphatics: their
relation to the mesenchyme cells. Eighteen figures..................... 347

No. 4. SEPTEMBER

Hat Downey. The attachment of muscles to the exoskeleton in the cray-
fish, and the structure of the crayfish epiderm Five figures............. 381

Orro F. Kampmerer. The development of the thoracic duct in the pig.
Thirty-ive figures (ive ,colored plates)... ec.cnec- aoe ase tet ee ee OL

Freperic T. Lewis. The form of the stomach in human embryos with notes
upon the nomenclature of the stomach. Twelve figures................. 477

THE GENESIS AND DEVELOPMENT OF THE
NASOLACRIMAL PASSAGES IN MAN

J. PARSONS SCHAEFFER

Yale University

From the Anatomical Laboratories of Cornell and Yale
THIRTY-ONE FIGURES

A brief review of the literature on the nasolacrimal passages
[lacrimal ducts (lachrymal canaliculi), lacrimal sac, nasolacrimal
duct] shows that diverse views were from time to time advanced
on the genesis and development of these passages. Before con-
sidering the material studied in this investigation I want to refer
to some of the theories held by earlier writers. I do not wish to
give a complete résumé on the history of the development, but
rather in a brief manner indicate the evolution in our knowledge
concerning the genesis and development of these passages.

v. Baer (28-37) thought that the nasolacrimal passages had
their origin in a diverticulum from the‘ Rachenhohle.’ His theory!
presumably was based upon hypothetical conclusions, since it is
entirely unsupported.

Burdach (737) in his ‘Die Physiologie als Erfahrungswissen-
schaft’ writes briefly concerning the genesis of the nasolacrimal

' Die Bildung des Thranenkanals (in birds) glaubte ich in einer Ausstiilpung
der Rachenhohle, die zuerst nur wenig vor der Eustachischen Roéhre hegt und
sehr bald nach dieser sichtbar wird, zu erkennen, doch habe ich bisher noch nicht
den gesammten Vorgang verfolgt. Uber Entwickelungsgeschichte der Thiere,
Thiel 2, S. 116.

Der Thrinengang stiilpt sich auch hier (in mammals) aus der Rachenhdhle
gegen das Auge hervor und liegt Anfangs hinter den Muscheln, die nur, indem sie
sich verlingern, sich iiber ihn ziehen. Uber Entwickelungsgeschichte der Thiere,
Phelk2S2219,

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 1
MARCH, 1912

2 J. PARSONS SCHAEFFER

passages, but does not state his meaning clearly.2. He apparently
thought that the nasolacrimal passages had their origin in a
diverticulum or skin-fold (‘Hautfalte’) in the region of the medial
palpebral commissure (internal canthus) and, since the naso-
optic fissure is obliterated by the eighth week of embryonal life,
he must have thought that the ‘Hautfalte’ grew into the sub-
stance of the maxilla, ultimately reaching the nasal cavity. Bur-
dach may have had the right conception of the development of
the nasal end of the nasolacrimal duct, but in the genesis of the
nasolacrimal passages from a skin-fold (‘Hautfalte’) in the region
of the medial palpebral commissure he erred (providing the writer
interprets his statement correctly). It is difficult to say what
Burdach meant by his ‘Hautfalte.’ That the anlage of these
passages comes to lie in the body of the maxilla is true, but it
comes about in an entirely different way as will be seen subse-
quently. His theory would not explain the origin of the paired
lacrimal ducts.

Erdl (45) and Coste (’47—’59), according to Ewetzky, believed
that the furrow ‘‘weleche am Naseneingange beginnt und am
inneren Augenwinkel miindet, auf ihrer ganzen Linge iiber-
briickt und dergestalt in einem Canal verwandelt.’? That the
naso-optic fissure becomes constricted or shut off from the sur-
face by its lips closing in and coalescing with each other, thus
establishing the nasolacrimal connections, was indeed the accepted
theory for some time. The theory is, of course, erroneous because
the anlage of the passages is for some time represented by a solid
plug or strand of epithelial cells which early becomes detached from
the surface. The strand of cells becomes cord-like and acquires
a lumen secondarily (see subsequent paragraphs). Neither would
this theory explain the pairing of the lacrimal ducts.

So far as my review of the literature would prove, Born (’76)
was the first investigator to properly interpret the earliest stages

> Der innere Augenwinkel ist mehr verlingert als bei Erwachsenen und steht
tiefer als der fiussere; schon in der achten Woche erscheint in ihm die Karunkel
und eine zur Mudnasenhohle sich sekende Hautfalte als Anfang des Thrinenkanals.
Die Thrainenpunkte ragen im fiinften Monat sehr stark hervor und im siebenten
etwas mehr zuriick.

GENESIS OF NASOLACRIMAL PASSAGES IN MAN 33

of the nasolacrimal passages. He investigated this field in am-
phibia and found a structure homologous with that described
by Coste for mammals, but he found that its genesis did not agree
with Coste’s hypothesis.

Since Born’s conception of the genesis of the nasolacrimal pas-
sages In amphibia applies also, broadly speaking, in other forms,
it may not be amiss to briefly quote his own words:

Der Thranencanal der Amphibien bildet sich durch Einwachsung
und Abschntirung eines Epithelstreifens von der Nase bis zum Auge
hin der dann ein Lumen bek6émmt und sich mit der Nasenhéhle in
Verbindung setzt.

While the above did not clear up the origin of the lacrimal
ducts in mammals, it nevertheless proved to be the correct inter-
pretation of the genesis of the main portion of the nasolacrimal
duct in all investigated forms up to the present. According to
Born, in amphibia, a solid strand of epithelial cells, extending
from the eye to the nose, becomes detached from the surface
epithelium and this strand of cells later acquires a lumen. The
strand of cells retains connections with the surface epithelium
at both the ocular and nasal ends. This strand of cells becomes
both the lacrimal ducts and the whole of the nasolacrimal duct.

Born later (78) investigated lizards and birds and found that
the basic principles concerning the anlage of the nasolacrimal
passages in these forms agreed with what he found in amphibia.
He, however, found that the cord of cells differentiated along
the course of the oculo-nasal furrow; also that it differed somewhat
in its further development. In both forms (lizards and _ birds)
a solid cord of cells became isolated from the surface. In lizards
the isolation was complete, i.e., there remained no connection
with the surface epithelium at any point; both lacrimal ducts
and the nasal end of the nasolacrimal duct developed as sprouts
from the mother cord of cells. As in amphibia the lumina of
these several channels were established later. In birds the cord
of cells retained connection with the surface epithelium at both
the ocular and nasal ends. One of the lacrimal ducts, however,
developed as a sprout from the ocular end of the mother cord of
cells.

4 J. PARSONS SCHAEFFER

Ewetzky (’79) studied the embryology of the nasolacrimal
connections in ‘Rindsembryonen’ and in the main agreed with
Born’s findings.

Born (’82) investigated reptiles and found that the solid cord
or strand of epithelial cells retained connection with the surface
epithelium at the ocular end, but that the nasal end of the naso-
lacrimal duct grew as a sprout from the mother cord of cells.
He also found that there is no doubling or pairing at the ocular
end; that is, he found but one lacrimal duct.

Legal (’83) investigated the pig and came to the same general
conclusion as did Born, and claimed (for pig) that the superior
lacrimal duct was wholly a part of the original mother cord of
cells. He further claimed that the mother cord of cells retained
superiorly and dorsally a connection with the epidermis in the
region of the palpebral fissure.* He concluded that the inferior
lacrimal duct grew as a sprout from the mother cord, but he
found that the sprout did not reach the free border of the inferior
eyelid; therefore remaining ‘funktionell unbrauchbar.’!

Kolliker (’84) believed that both lacrimal ducts developed as
sprouts from the mother cord.

Ewetzky (88) thought that the ocular end of the mother cord
divided into two forks, and that these forks in turn developed into
the lacrimal ducts (superior and inferior).

Jouves (’97) studied the sheep and man, and found in a 19 mm.
human embryo both lacrimal ducts present but without any con-
nection with the surface epithelium at this time.

Cosmettatos (98) investigated the rabbit, and Stanculeanu
(00) the bird, the sheep, and man. These investigators, accord-

3. . . . ganz hinten endlich bleibt besténdig eine Verbindung mit der
Lidfurche erhalten.

4 Bei Schweinsembryonen ist die Thriinenkanalanlage eine solide, von der
tiefen Epidermisschicht des Thrinenfurchengrundes ins Bindegewebe einwuch-
ernde Leiste, die sich bis auf das hinterste Ende am innern Augenwinkel von der
Epidermis abschniirt, und mit dem vordern, stark auswachsenden Ende mit der
Nasenhohle verbindet: der abgeléste, solide Epithelstrang stellt den spitern
einfachen Thriinennasengang und das obere Thranenrérchen dar, das untere
sprosst aus demselben hervor, bleibt aber, da es die freie Lidfliche nicht erreicht,
funktionell unbrauchbar.

GENESIS OF NASOLACRIMAL PASSAGES IN MAN 2)

ing to Fleisher, depending upon the form studied, came to the
conclusion that one of the two lacrimal ducts was wholly or partly
a portion of the original mother cord of cells, and that the other
lacrimal duct and the remaining portion of the nasolacrimal duct
developed as sprouts from the mother cord.

Hammar (’02) shows a model® of a human embryo 18.5 mm.
long in which both the nasal and ocular ends of the anlage of the
nasolacrimal passages are free from the surface epithelium. The
anlages of the lacrimal ducts are well illustrated in the model.

Fleisher (06) published his researches on the pig, the guinea
pig, the mouse, the rabbit, and man. He arrived at the follow-
ing general conclusion:

Aus diesen Priparaten geht hervor, dass beim Menschen die Ent-
wicklung der Tréinenréhrchen in derselben Weise vor sich geht, wie
bei den anderen, von mir untersuchten Saéugetieren, durch selbsténdige
Sprossung jedes der Roérehen aus dem Augenende der Triinenleiste,
die sich vollstandig vom Epithel abschniirt.

Fleisher, therefore, disagrees with Legal and some others on the
genesis of the lacrimal ducts (lachrymal canaliculi), and conforms
with Ko6lliker and more nearly to Ewetzky. He is also in accord
with Matys who came to similar conclusions for Spermophillus
eitillus.

Lang (11) reports his findings in a human embryo, aged from
seven to eight weeks. He finds that the left side of his embryo
agrees with the conclusions of Fleisher and Matys. On the right
side he, however, finds the superior lacrimal-duct anlage wanting.

In subsequent portions of this paper I wish to record my pre-
liminary observations on the genesis and development of the
nasolacrimal passages in man. I now hope to carry this study
farther and, if an investigation of a larger number of human
embryos warrants, will report my later observations and con-
clusions in a subsequent paper.

In looking over material for the substance of another paper
I noticed variations in the development of some portions of the

° Studien iber die Entwicklung des Vorderdarms und einiger angrenzenden
Organe, Archiv f. mikrosk. Anat., Bd. 59, taf. 26, fig. 14.

6 J. PARSONS SCHAEFFER

nasolacrimal passages. I, therefore, felt that there was need of
an investigation of the genesis of these passages In man, based
upon an examination of a larger number of human embryos than
was formerly done. Fortunately there were available for this
study good series of appropriately aged embryos showing the
genesis and early stages of the nasolacrimal passages. The
embryos ranged in age from thirty days to ‘term.’ A certain
amount of material of the early extra-uterine period was also
studied; together with a large number of adult specimens.

It is well known that at one stage of the embryo there is a
furrow or fissure—the naso-optic fissure—extending from the eye
to the nasal pit. This fissure is bounded superiorly by the lateral
nasal process and inferiorly by the maxillary process. The naso-
optic fissure gradually disappears by a growth and coalescence
of the structures bordering it. In this manner the fissure is
‘out-folded’ as it were and thus becomes shallower and shallower
until its ultimate obliteration. The epidermis along the course
of the floor of the now very rudimentary fissure concerns us for
some time longer with reference to the anlage of the nasolacrimal
passages.

Before the naso-optic fissure is entirely obliterated we have in
frontal sections a thickening of the deeper layers of the epidermis
along the floor of the very rudimentary fissure (fig. 2). This
initial thickening is the anlage of the nasolacrimal passages. It
is at first, and remains so for some time, a solid cord-like struc-
ture of epidermal cells, at all points a part of the epidermis, along
the floor of the remains of the naso-optie furrow, extending from
the neighborhood of the eye towards the nose.

My observations began on embryos aged approximately from
thirty to thirty-two days. In these embryos I could find no
evidence whatever of an anlage of the nasolacrimal passages.
In fig. 1 we have a frontal section through the remains of the
naso-optic fissure (human embryo aged approximately thirty-
three days). Note that there is no evidence of the anlage of the
nasolacrimal passages along the floor of the rudimenatry fissure.
The epidermis appears uniform in thickness at all points, 1.e.,
the epidermis is not thickened along the floor of the fissure.

“NI

GENESIS OF NASOLACRIMAL PASSAGES IN MAN

Fig. 1 Frontal section through the nasal fossa (f/) and the now-rudimentary
naso-optie furrow (no), from a human embryo aged approximately thirty-three
days. Note that there is no evidence of thickening of the epidermis along the
floor of the rudimentary naso-optic furrow (no), to establish the anlage of the
nasolacrimal passages. > 62.4

Fig. 2 Asimilar section to that illustrated in fig. 1, from a human embryo aged
thirty-four days. Note the plug-like thickening of the epidermis in the position
of the rudimentary naso-optic furrow. This is the first evidence of the anlage
of the nasolacrimal passages. > 62.4.

Fig.3 A corresponding section to those illustrated in figs. 1 and 2, from a
human embryo aged approximately thirty-five days. Note the marked in-growth
of the epidermal plug in comparison to the plug represented in fig. 2. X 62.4.

Fig. 4 A frontal section through the anlage of the nasolacrimal passages,
from a human embryo aged forty-three days. Note that the cord of epidermal
cells is now entirely separated from the surface. It is wholly surrounded by
mesenchymal cells. In the sections preceding and following this the separation
from the surface was just as complete as in that shown here. 62.4.

The first evidences of the anlage of these passages I found in a
12 mm. embryo, aged about thirty-four days. The anlage is
well illustrated in figs. 2 and 7 as a plug-like thickening (in frontal
section) of the deeper layers of the epidermis along the floor of
the remains of the naso-optic fissure. Note especially that the
surface layer of flat epidermal cells is not included in the anlage

8 J. PARSONS SCHAEFFER

of the nasolacrimal passages. In this respect I am in accord with
Born, Legal and others and at variance with Ewetzky’s first
paper (79). As to age for the first evidences of the anlage of the
nasolacrimal passages I agree rather closely with Ewetzky, who

anterior

-Meatus nasi
inferior

Con, nas.
inferior

N

Sep. nasi :

Fig. 5 Selected sections from a series through the developing nasolacrimal
passages (human embryo aged from forty-three to forty-five days). Note that
nowhere are the anlages of the nasolacrimal passages in connection with the sur-
face. The lacrimal ducts are already in evidence. All of the ‘passages’ are yet
solid cords of epithelial cells, and are indicated in deep black. 14.

found that ‘“‘die Entwickelung des Thrinencanals beginnt um
das Ende der 5. oder im Anfang der 6. Woche des Foétallebens.”’
The anlage of the nasolacrimal passages soon becomes suffi-
ciently developed to sink into the corium along the course of
the rudimentary naso-optie fissure. The anlage grows rapidly

GENESIS OF NASOLACRIMAL PASSAGES IN MAN g

and in a brief time it has grown to such an extent that it reaches
into the underlying mesenchymal tissue. Witness for example
figs. 3 and 8, which represent frontal sections through the left
nasal fossa of an embryo aged approximately thirty-five days.
In these figures the ridge of epithelial cells has grown into the
underlying mesenchymal tissue. Note especially, however, that
the anlage still retains its connection with the rete mucosum
of the epidermis along the course of the naso-optic groove, in
which it has its genesis. Note further that the anlage is as yet
solid and that there is no evidence of lumen formation.

Fig. 9 represents a semi-frontal section through a later stage of
the anlage (embryo aged about forty-three days). Note that
now the cord of epithelial cells is entirely detached from the sur-
face, 1.e., it has entirely lost its connection with the rete mucosum
of the epidermis from which it arose. The anlage of the naso-
lacrimal passages now lies well embedded in the mesenchymal
tissue. It will be further noticed that the central cells of the
cord-like anlage (fig. 9) have taken the stain less deeply, and
apparently there is already an attempt at lumen formation. Some
of the central cells seem to be in a state of ‘necrobiosis.’ The
cells of the cord are apparently becoming re-arranged to form a
wall in anticipation of a later lumen.

In the serial frontal sections through the nasal cavity of a
forty-three day embryo represented in figs. 10 and 11, the com-
plete isolation from the surface of the nasolacrimal passages at
this stage of development is well illustrated. The embryo from
which these photomicrographs were made is in a splendid state
of preservation. It is human embryo no. 3 (Hess Embryo) of
the Cornell University Series. It belongs to the research col-
lection of Professor and Mrs. Gage.

This embryo shows several very important points in connec-
tion with the development of the nasolacrimal passages: In the
first place we find at this stage that the anlages of the nasolacri-
mal passages are nowhere connected with the epidermis, but that
they are entirely encompassed by mesenchymal tissue. In the
second place it will be noticed that the cords of cells are solid,
with here and there evidences of lumen formation. The series

10 J. PARSONS SCHAEFFER

Figs. 6, 7, 8 and 9 Photomicrographs of frontal sections showing several
stages in the development of the nasolacrimal passages; a, 6, c = remains of the
naso-optic furrow; f=nasal fossa; e=eye; d=different stages of the anlage of the
nasolacrimal passages.

Fig. 6 Note that there is no evidence of the anlage of the nasolacrimal pas-
sages in the region of the naso-optic furrow (a). Human embryo aged thirty-
three days. X 68.

Fig. 7 Note the plug-like anlage of the nasolacrimal passages (d) from the
deep layers of the epidermis along the floor of the naso-optie furrow (b). Human
embryo aged from thirty-four to thirty-five days. X 68.

Fig. 8 In this section the anlage of the nasolacrimal passages is considerably
advanced over that shown in fig. 7. Note, however, notwithstanding that it has
pushed its way into the underlying mesenchymal tissue, that the anlage still
retains its connection with the deeper layers of the epidermis along the floor of
the naso-optic furrow (c). Human embryo aged about thirty-six days. X68.

Fig. 9 In this section the anlage of the nasolacrimal passages has lost its con-
nection with the surface. Human embryo aged forty-three days. 68.

GENESIS OF NASOLACRIMAL PASSAGES IN MAN 1a

Figs. 10 and 11 Photomicrographs of frontal sections in the region of the
developing nasolacrimal passages, from a human embryo aged forty-three days.
Note the anlages of the nasolacrimal passages and that they are entirely separated
from the surface. The lacrimal-duct anlages are well advanced and show as
sprouts from the mother cord. The nasal end of the cord has not developed suffi-
ciently to come in contact with the mucous membrane of the inferior nasal meatus.
The lacrimal ducts are also some distance from the free borders of the eyelids at
this time. The section represented in fig. 10 is the most ventral of the series
and that represented in fig. 11 the most dorsal. Some of the intervening sections
of the series are, of course, omitted. Embryo no. 3—Hess, Cornell University
Series. n= anlage of nasolacrimal duct; inf = anlage of inferiorlacrimal duct;
sup = anlage of superior lacrimal duct; e=eye. X 27.

ib J. PARSONS SCHAEFFER

also shows some irregularities and lateral buds from the main
cords. These may account for the very common diverticula
of the adult nasolacrimal duets (fig. 30). Finally the series shows
that the lacrimal ducts (lachrymal canaliculi) begin as sprouts
from the ocular end of the mother cord of cells. Both the supe-
rior and inferior lacrimal ducts are about equally advanced in
development, but neither of them have progressed far enough to
reach the free borders of the eyelids and thus establish the defini-
tive connections between the anlages of the nasolacrimal passages
and the epidermis.

The lacrimal ducts are also solid cords of cells and show no
evidence of lumen formation. In this series (figs. 10 and 11)
it would indeed be difficult to say which of the lacrimal ducts
(superior or inferior) was the first to begin its budding from the
mother cord. Presumably they started budding approximately
at the same time.

So far as my observations would prove there is considerable
variation in the development of the lacrimal ducts, notwithstand-
ing that both ducts begin, I believe, as buds from the mother
cords of cells. The two ducts do not always begin their devel-
opment at the same time. If they do begin at the same time
then one or the other is often relatively tardy in its growth.

In figs. 12 and 13 are represented frontal sections through the
nasal cavity of a forty-two to forty-five day embryo. On the
left side of this embryo (figs. 12 and 13) the superior lacrimal
duct is well advanced, almost reaching to the free border of the
eyelid. The inferior lacrimal duct on the other hand is extremely
backward in its development. The only structure present—a
small lateral bud from the mother cord of cells, that may be the
beginning of the inferior lacrimal duct, is shown in fig. 12. I,
however, am not at all sure that this is the anlage of the inferior
duct. It is a well known fact that one or the other lacrimal-duct
anlage may fail to reach the free border of the eyelid. It is,
therefore, possible, had this embryo (figs. 12 and 13) continued
its development to ‘term,’ that it would have been born without
a drainage duct for the inferior eyelid. On the other hand, on

GENESIS OF NASOLACRIMAL PASSAGES IN MAN 13

the right side of the same embryo both lacrimal-duct anlages are
equally developed.

To say, from the condition found in the embryo represented
in figs. 12 and 13, that the superior lacrimal-duct anlage is a por-
tion of the original mother cord of cells is I believe erroneous.
I rather hold, in such cases, that the inferior-duct anlage is tardy
in its development, and that both ducts have their anlages in
buds from the mother cord. Of course as stated above one or
the other duct may at times, for unknown reasons, fail to develop
far enough to gain coalescence with the free border of the eyelid;

ede i

‘12 WSS

Figs. 12and13 Photomicrographs of frontal sections of a human embryo aged
forty-two to forty-five days, in the region of the early lacrimal passages of the
left side. Note that the superior lacrimal duct (sld) is well advanced. The only
evidence of an inferior duct is seen in fig. 12, at point marked X. Whether this
early condition would have led to an absence of the inferior lacrimal duct is of
course not known. s/d = superior lacrimal duct; « = anlage (?) of inferior lacrimal
duGt >< 32:

or the duct may reach the border of the lid but fail to establish
a lumen at this point. Cases have also been reported in which
supernumerary lacrimal puncta and lacrimal ducts were present
(Weber and others). There are at times small buds arising from
the lacrimal-duct anlages, and in all probability these at times
continue to develop independently until they reach and gain
coalescence with the eyelids. The lumina for these super-
numerary lacrimal ducts are, of course, established just as they
are in the regular ducts.

14 J. PARSONS SCHAEFFER

By the beginning of the third month (in some cases before)
the lacrimal-duct cords have developed sufficiently to come in
contact with the epithelium on the free borders of the eyelids.
We will, however, find that the ducts are as yet in places solid
cords. Portions of the mother cord, especially the portion that
is destined to become a portion of the lacrimal sac, are active in
lumen formation at this time. The fundus of the lacrimal sac
apparently develops as a sprout from the mother cord. The
nasal end of the mother cord has not developed sufficiently to
come in contact with the mucous membrane of the inferior
nasal meatus. It is, however, not far removed, and in a later-
stage embryo it will be found coalesced with the nasal mucous
membrane. Lumen formation at the point of coalescence of the
mother cord with the nasal mucous membrane is delayed approxi-
mately until ‘term’ (figs. 22, 23:and 24).

Frontal sections of embryos aged approximately one hundred
days will show that the nasal end of the mother cord of cells has
developed to the nasal mucous membrane and has coalesced with
it. Both lacrimal-duct cords have grown to the free borders of
the eyelids and have coalesced with the epithelium at these points.
The mother cord of cells, or the portion destined to become the
ereater portions of the lacrimal sac and the nasolacrimal duct,
has by this time established irregular lumina at various points.
The latter are best developed at the ocular end of the primary
cord and towards the nasal end of the cord. The lacrimal-duct
cords have also established lumina at various points, especially
in the regions of the knees, or what will later become the ampullae
of the lacrimal ducts.

In figs. 14, 15 and 16 (embryo aged one hundred and seven days)
is well illustrated the irregular manner in which the cords of epi-

Figs. 14,15,and 16 Photomicrographs of frontal sections through the naso-
lacrimal passages of a human embryo aged one hundred and seven days. Note
both lacrimal ducts (fig. 14) in contact and fused with the epidermis in the region
of the free borders of the eyelids. The lacrimal ducts have not yet established
lumina in the regions of the eyelids (fig. 14). The remaining portions of the ducts
aremore or less patent throughout. Note the irregularity of lumen formation in
the nasolacrimal duct (fig. 16). sld = superior lacrimal duct; ild = inferior lacri-
mal duct; cld = common lacrimal duct; /s = lacrimal sac; nld = nasolacrimal duct.
2 10)-9),

15

S IN MAN

GENESIS OF NASOLACRIMAL PASSAG

16 J. PARSONS SCHAEFFER

thelial cells establish lumina. Note that the lacrimal ducts are
more or less patent throughout, save at the free borders of the
eyelids where solid cords still persist. The horizontal and verti-
cal portions of the lacrimal ducts are well shown (fig. 14). The
lacrimal sae (fig. 15) is well advanced but the remainder of the
nasolacrimal duct is not wholly patent. Even at this early stage
there is some evidence of beginning diverticula from the naso-
lacrimal duct (fig. 16). The connection with the inferior nasal
meatus is, of course, not yet established (fig. 16).

According to my studies, the ocular end of the mother cord
is the first to establish a lumen. The horizontal portions of the
lacrimal ducts become patent before the vertical portions (figs.
14 and 17). The last parts of the vertical portions of the lacri-
mal ducts to become patent are the junction points between the
lacrimal-duct cords and the epithelium of the free borders of the
eyelids (fig. 17). The nasal end of the mother cord establishes
a lumen before the middle portion of the cord (figs. 19 and 20).
The middle portion remains solid, according to my series of
embryos, for some time longer (fig. 19). The last portion of the
nasolacrimal passages to become patent is at the point of coal-
escence between the nasal sprout of the mother cord and the
nasal mucous membrane. This is usually deferred, as stated
before, until ‘term’ or even later (fig. 24).

In the adult we find varying positions on the lateral wall of
the inferior nasal meatus for the ostium of the nasolacrimal
duct. The ostium also varies as to shape, and it is occasionally
duplicated. Rarely we find a triplicity of the ostium.

The position of the ostium, i.e., whether at the highest point
of the inferior nasal meatus, or at varying distances below the
above point on the lateral nasal wall, depends, of course, largely
upon the original point of coalescence of the mother cord of cells
with the nasal mucous membrane (fig. 28). At times the area of
coalescence between the cord of cells and the nasal mucous mem-
brane is rather extensive (fig. 21). In such cases, due to the irre-
gular mode of lumen formation, two or more ostia may readily
be formed (instead of the usual single ostium) with a bridge of
intervening tissue remaining permanently. The different shapes

Dy

Figs. 17-21 Photomicrographs of frontal sections through the nasolacrimal
passages of a human embryo aged one hundred and twenty days. Note the solid
portions of the lacrimal ducts in fig. 17. In fig. 18 we have a patent section (/s)
of the ocular end of the nasolacrimal duct, and in fig. 19 a section of the mid-por-
tion of the nasolacrimal duct, still solid (mld). Note the well established lumen
(nld) at the nasal end of the nasolacrimal duct in figs. 20 and 21. Note how exten-
sive the contact point between the nasolacrimal duct and the inferior nasal meatus
will be (fig. 21). sld = superior lacrimal duct; ‘ld = inferior lacrimal duct; ls =
lacrimal sac: n/d = nasolacrimal duct; inc = inferior nasal concha; inm = inferior
nasal meatus. 19.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 1

17

1s J. PARSONS SCHAEFFER

Figs. 22-27 Photomicrographs of sections through the nasolacrimal duct.

Fig. 22 Fromaterm child. Note that the connection between the nasolacri-
mal duct and the inferior nasal meatus is not yet established. X 3.4.

Fig. 23. From a seven-month fetus. The connection between the nasolacri-
mal duct and the inferior nasal meatus is not established. 6.1.

GENESIS OF NASOLACRIMAL PASSAGES IN MAN

Meatus nasv

infertor

~
>
8
=
=
Oo
2
=
~
S
SL
=

Concha nasalts
wfertor

Fig. 24 From aterm child. Barrier between the nasolacrimal duct and the
inferior nasal meatus still present. Note the irregularity of the nasolacrimal
duct. Compare with fig. 30. > 2.9.

Fig. 25 From an adult aged sixty years. Note the circular and regular con-
dition of the nasolacrimal duct. Reconstruction seen in fig. 29. 2.6.

Fig. 26 From an adult aged sixty-five years.

Note the marked diverticulum
(X) from the nasolacrimal duct.

The nasolacrimal passages of this individual
are seen in reconstruction in figs. 30 and 31. 2.6.
Fig. 27 From an adult aged seventy years.

In the region of the ostium of
the common lacrimal duct. X 2.8.

Fig. 28 Photographs of transections of the nasolacrimal duct at the point of
entrance (ostium of nasolacrimal duct) into the inferior nasal meatus. Note

the different types of ostia (0). The sections are from adults and are magni-
fied from two to four times.

20 J. PARSONS SCHAEFFER

Fig. 29 Reconstruction of the nasolacrimal passages of an adult aged sixty
years. Note the regularity of the nasolacrimal duct and compare with figs. 30
and 31. X 3.2.

GENESIS OF NASOLACRIMAL PASSAGES IN MAN Pall

51

Figs. 30 and 31. Reconstruction of the nasolacrimal passages of an adult aged
sixty-five years. Fig. 30 represents a medial view and fig. 31a lateral view of the
reconstruction. Especially note the irregularity, due to diverticula, of the naso-
lacrimal duct. The portions indicated in black at the inferior extremity of the
nasolacrimal duct is a portion of the inferior nasal meatus. XX 3-2.

De, J. PARSONS SCHAEFFER

of the ostia are doubtless due to the angle at which the original
eord of cells meets the nasal mucous membrane. The position of
contact also makes a difference. If the ostium is at the highest
point of the inferior meatus, 1.e., Just caudal to the attachment
of the inferior nasal concha to the lateral nasal wall, the opening
of the nasolacrimal duct is usually a large, wide, open-mouthed
ostium, unguarded by folds of mucous membrane (fig. 28).
If, on the other hand, the ostium is farther caudal on the lateral
wall it is usually slit-like and more or less guarded by folds of
mucous membrane (figs. 28).

Even at term the embryo presents very irregular nasolac-
rimal ducts (fig. 24). After birth the walls of the ducts become
more and more regular In the adult we very frequently find
large diverticula from the nasolacrimal duct, and these added to
other irregularities give us at times extremely irregular lumina
(figs. 30 and 31). On the other hand we find adult ducts in
which the lumina are very simple and regular (fig. 29). The
lumina of the adult lacrimal ducts (lachrymal canaliculi) are gen-
erally more or less irregular.

SUMMARY

1. The strand of thickened epithelium—the anlage of the naso-
lacrimal passages—along the floor of the rudimentary naso-optic
fissure becomes entirely separated from the surface, and for some
time is wholly surrounded by mesenchymal tissue.

2. The strand or cord of epithelial cells thus isolated from the
surface is for some time without a lumen.

3. From the mother cord of cells both lacrimal ducts and the
nasal end of the nasolacrimal duct grow as sprouts. The cephalic
portion of the lacrimal sac also grows as a sprout from the mother
cord.

4 Considerable variation occurs in the development of the
lacrimal ducts, i.e., as to number, time, and degree of develop-
ment.

5. The lumina of the several portions of the nasolacrimal pas-
sages are established in an irregular manner. The ocular end

GENESIS OF NASOLACRIMAL PASSAGES IN MAN 23

of the mother cord is the first to establish a lumen. The point
of coalescence between the nasal end of the cord and the mucous
membrane of the inferior nasal meatus is the last to become patent
—the lumen here is established approximately at ‘term’ or even
later. The horizontal portions of the lacrimal ducts establish
lumina before the vertical portions.

I wish to take this opportunity for expressing grateful acknowl-
edgment to Professors Gage and Kerr for the material placed
at my disposal in this investigation and for other courtesies
extended. Iam also indebted to Professor Ferris for reading the
manuscript.

BIBLIOGRAPHY

Ask, Fritz 1907 Uber die Entwickelung der Caruncula lacrimalis beim Men-
schen, nebst Bemerkungen tiber die Entwickelung der Triinenrérchen
und der Meibomschen Driisen. Anatom. Anz., Bd. 30.

v. Barer, Karu Ernst 1828-1837 Uber Entwickelungschichte der Thiere. K6nigs-
berg.

Born, G. 1876 Uber die Nasenhodhlen und den Thrinennasengang der Amphi-
bien. Morphol. Jahrbuch, Bd. 2.

1879 Die Nasenhohlen und der Thrinennasengang der amnioten Wir-
belthiere. Morphol. Jahrbuch, Bd. 5.

1883. Die Nasenhodhlen und der Thrinennasengang der amnioten Wir-
belthiere. Morphol. Jahrbuch, Bd. 8.

Burpacu, KartuG. 1837 Die Physiologie als Erfahrungswissenschaft. Leipzig.
nt = oD

CosmMetratos, F. 1898 Recherches sur le développement des voies lacrymales.
Thése de doctorat. Paris.

CostE, JEAN J. 1847-1859 Historie générale et particuliére du développement
des corps orgninsés. Paris.

Emmert, E. 1876 Angeborenes Fehlen aller 4 Thrinenpunkte und Thrinen-
rorchen. Archiv. f. Augen- und Ohrenheilk, Bd. 5.

Erpt 1845 Die Entwickelung des Menschen und des Hithnchensim Ei. Leipzig.

Ewerzky, TH. 1879 Beitrige zur Entwickelungsgeschichte des Auges. Arch.
f. Augenheilk., Bd. 8.
1888 Zur Entwickelungsgeschichte des Thrinennasenganges beim
Menschen. Archiv. f. Opthal., Bd. 34, Ab. 1.

FuLeisHer, B. 1906 Die Entwicklung der Thrainenrérchen bei den Siugetieren.
Archiv. f. Opthal., Bd. 62, H. 3.

24 J. PARSONS SCHAEFFER

Hammar, J. Ava. 1902 Studien iiber die Entwicklung des Vorderdarms und
einiger angrenzenden Organe. Archiv. f. mikroskop. Anat. und Ent-
wicklungsch., Bd. 59.

Jouves, A. 1897 Recherches sur le développement des voies lacrymales chez
Vembryon de mouton et l’embryon humain. Thése de doctorat.
Toulouse.

K6uurker, A. 1879 Entwicklungsgeschichte des Menschen und der héhern
Tiere. Leipzig.
1884 Grundriss der Entwicklungsgeschichte des Menschen und der
hohern Tiere. Leipzig.

Lecat, E. 1883 Die Nasenhdhlen und der Thrinennasengang der amnioten
Wirbelthiere. Morphol. Jahrbuch, Bd. 8.

Lane, Paun 1911 Zur Entwickelung des Thrinenausfithrapparates beim Men-
schen. Anat. Anz., Bd. 38.

Maanus, H. 1880 Mangel der unteren Thrinenpunkte und Warzchen auf
beiden Augen. Centralbl. f. prakt. Augenheilk.

Marys, 1905-1096 Die Entwickelung der Triinenableitungswege. Zeitsch. f.
Augenheilk, Bd., T4 u. 16.

ScHAEFFER, J. Parsons 1911 The lateral wall of the cavum nasi in man, with
especial reference to the various developmental stages. Annals of
Otol. Rhinol. and Largynol., June.
1911 Variations in the anatomy of the naso-lachrymal passages. An-
nals of Surgery, August.

SrANCULEANU, G. 1900 Recherches sur le développement des voies lacrymales
chez homme et chez les animaux. Archiv d’ophtalm., Tom. 20.

Wesepr, A. 1861 Opthalmologische Miscellen und Aphorismen, II Zwei Faille

von iberzihhgen Canaliculi lacrymales. Archiv f.Opthalmologie,
IBGE Sa Absale

THE DEVELOPMENT OF THE NUCLEI PONTIS AND
THE NUCLEUS ARCUATUS IN MAN

CHARLES R. ESSICK
From the Anatomical Laboratory of The Johns Hopkins University

TWELVE FIGURES

The following paper shows the origin of the gray matter found
on the ventral surface of the adult rhombencephalon from the
‘Rautenlippe’ or rhombic lip, along its attachment to the medulla,
and the path of migration of the cells formed here to the position
characteristic of the fully developed brain. Almost all of the
cells of the arcuate and pontine nuclei arise by karyokinetic divi-
sion around the attachment of the roof of the fourth ventricle
and then wander to their proper places. The pathway of this
cellular migration proves to be a very superficial one, remarkable
both for its constancy and its definite limits. The arcuate nucleus
forms by a migration over the surface toward the ventral median
fissure; the pontine nuclei choose a path which corresponds in
every detail to the fibro-nuclear mass which I described for the
adult as the corpus ponto-bulbare.

It is well known that all of the nuclear material in the central
nervous system is derived from that portion of the ectoderm which
closes in to form the neural tube; and our knowledge of the exact
manner of this cellular distribution is due mainly to the re-
searches of His. This author has pointed out that all of the nerve
cells in the central nervous system first passed through the stage
of neuroblasts and in their development are usually wont to leave
the place of their origin so as to enter into the formation of the
gray matter at a distance. The newly formed gray masses may:
(1) remain in the neighborhood of the matrix, e.g., motor and
arcuate cells of the spinal cord, or, (2) pierce the substance of
the medullary tube in a radial direction and collect on the surface
into an independent layer. Such wandering of cells from the

25

26 CHARLES R. ESSICK

matrix of the original ventricular gray matter takes place in the
formation of the cerebral cortex. (3) Subsequent to a definite
bending in the medullary wall, newly formed gray masses may be
transferred from the dorsal to the ventral portion of the brain by
a migration of cells in a tangential direction, e.g., arcuate forma-
tions, olive and accessory olives, and a part of the nuclei lying in
the pons. The third of these processes as described by His is
responsible for the development of the gray matter constituting
the arcuate nuclei of the medulla and the basilar nuclei of the pons.

The compact manner in which the neuroblasts arrange them-
selves in their migration to the pontine flexure, has attracted the
attention of many observers both in macroscopic and microscopic
preparations, nevertheless, with the exception of Streeter (’12)
only a casual mention of it has been made by them. Blake (’00)
in his description of the roof of the fourth ventricle noted cells
which were transferred to the ectal surface of the oblongata by
the formation of the secondary rhombic lip and he could trace
them in many embryos as far cephalad as the trigeminal nerve.
He ventured the suggestion that they might be connected with the
ganglia of some of the cranial nerves. His (’04) has given a good
illustration (fig. 103) of the rhombencephalon of a 5 em. fetus
and has shown the outlines of the thick mass of cells passing from
the rhombic lip to the pontine formation. From its appearance
with the naked eye, as well as in serial sections, he identifies it
with the Corpus Trapezoides, which thus occupies a superficial
position at this time (p. 163). Streeter (07) in dissections of the
Seventh nerve in pig embryos called attention to the presence of a
ganglion mass connected with the pons ganglia which could be
traced backward as a surface ridge between the facial and acoustic
nerves, to end on the dorso-lateral surface of the restiform body.
Since then he has suggested (712) two possible origins for the
pontine nuclei: the corpus ponto-bulbare and the mantle zone of
the pontine region. Neuroblasts from the latter source emerge
through the marginal zone as happens with the cortical cells of
the cerebellum. Orzechowski (’08) in human foetus measuring
17 and 23 cm. has described ganglion masses connecting the
rhombic lip, lateral recess wall and pons, which he considers the

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN Bib

embryonic corpus ponto-bulbare. He believes that the adult
structure may contain portions of undeveloped embryonic tissue
and be responsible for the frequent tumors in the cerebello-pon-
tine angle. With the exception of the mention made by His
of a migration of cells to form the arcuate nuclei nothing has
been contributed to their development.

TABLE 1
(GiswieeniD) COLLECTION NUMBER PLANE OF SECTION THICKNESS
mm. be
20 128 Coronal 50
20 368 Sagittal 20
20 22 Transverse 50
23 382 Sagittal 50
24 405 Sagittal 40
30 227 Sagittal 50
30 5) Sagittal 50
30 S6 Coronal 50
33 211 Sagittal 50-100
33 145 Sagittal 50-100
35 199 Sagittal 50
46 95 Sagittal 100
50 96 Sagittal 100
50 S4 Transverse 50
50 1S4 Sagittal 50-100
SO 172 Transverse 100
96 484 Transverse 40
113 490 Transverse 30
115 219 Sagittal 50-100
1438 508 Transverse 40
188 509 Transverse 50
295 491 Transverse 40

In carrying out this investigation the writer was given abundant
opportunity to examine the large collection of human embryos
brought together in this laboratory by Professor Mall. Table
1 gives a list of the embryos used in this study.

Wax plate reconstructions were made after the method of Born.
Human, pig and rabbit embryos, stained in toto with alum-
cochineal, were prepared for dissection as described by Streeter
(04, p. 87). Whole brains stained in iron-haematoxylin also
gave brilliant differentiation. The most instructive specimens of

28 CHARLES R. ESSICK

the migrating strands of nuclear material were obtained in em-
bryos stained in methylene blue. After previous hardening
(10 per cent formalin is excellent), the brain is carefully taken out
of the skull and all of the pia mater dissected off; very great care
must be exercised in removing this vascular membrane in order
that the tiny penetrating vessels do not tear the surface of the
brain which then stains very deeply along the ruptured edges.
The specimen is placed into an aqueous solution of methylene
blue (saturate aqueous methylene blue and water equal parts)
for thirty to sixty seconds, rinsed in water, and transferred to water
for study. The whole brain is tinted blue but the most prominent
parts take a more intense stain so that all of the fine surface
irregularities are outlined in great detail. This brings out with
remarkable clearness the anastomosing strands of cells converging
into the pontine formation.

Confusion might arise out of the terms employed here so that
a word may not be out of place concerning their meaning. Inas-
much as the flexures of the brain as well as the position of the head
are not fixed, | have used the words ‘cephalad’ (forward, front,
anteriorly, cerebrally), ‘eaudad’ (backward, behind, spinalward),
‘dorsal,’ ‘ventral,’ ‘lateral’ and ‘mesial’ just as if the central nerv-
ous system were a simple straight tube placed in the head as the
spinal cord is in the body. This, it seems to me, will facilitate
the description of relations of parts which are constantly shifting
their positions in relation to the body. In addition, I might state
that I have used the term ‘neuroblast’ loosely, so as to include all
undifferentiated cells which have not taken on a definite form.

In considering the development of these basilar masses it may
be of advantage to review briefly some of the relations which exist
in the rhombencephalon just before the cells, destined for the
pontine and arcuate nuclei, set out from their germ centers. His
(91) has carefully reconstructed some of the intramedullary
nuclei and nerve roots with their relations to the surface and brain
flexures in an embryo of 22 mm. (figs. 5 and 17). He has called
attention to the fact that at this time, towards the end of the
second month, the formation of new neuroblasts has ceased in the
medulla and it is only with difficulty that a mitotic figure is dis-

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN 29

covered around the ventricular cavity where great numbers were
present in the earlier stages. With the cessation of its activity,
the epithelium lining the cavity of the fourth ventricle becomes
very sharply marked off from the underlying nervous tissue and
it would be expected that the various nuclear masses in the me-
dulla have received their allotment of cells, further growth consist-
ing of increase in size of individual elements and the addition of
nerve fibers.

Fig. 1 Profile reconstruction of 22 mm. embryo. ™X 4.5. Taken from His—
Entw. d. mensch. Rautenhirns, fig. 5. I have drawn in the abducens nerve (N.
vr) and its nucleus (Nu. vt).

If now one looks at fig. 1 (a profile drawing taken from His,
(91), to which I have added the sixth nucleus and its nerve) many
differences from the adult are evident, the most striking perhaps
being the great, flexure in the pontine region—the cerebellar thick-
ening almost touching the medulla. Just under the floor of the
fourth ventricle appears the nucleus n. abducentis usually an
elongated mass of cells lying immediately behind the ventricular
furrow formed by the bend in the neural tube. From this nucleus
the axones pass obliquely ventrally through the tegmentum to
emerge just behind the summit of the pontine flexure in a series
of rootlets which behave much as the hypoglossal nerve roots.
They are quickly gathered together to form a single nerve trunk.

30 CHARLES R. ESSICK

The facial nucleus has a remarkably constant form, the outline of
which is similar both in sagittal and coronal sections. It might
be compared to a pear, the smaller cephalic extremity tapering
off bluntly. The nucleus preserves this constricted end in the
adult as has been brought out in a model made by Mr. Weed in
this laboratory. Cephalad through half of its extent it lies dorso-
lateral and parallel to the superior olive and extending far in front
of the outline of the nucleus n. abducentis. It will be observed
that the relation to the superior olive is that of the adult yet one
would miss the familiar appearance of the facial nucleus seen in
transverse sections through the cephalic pole of the inferior olive.
In other words, the caudal tip of the facial nucleus is distant a
considerable interval from the cephalic tip of the inferior olive.
The olivary complex, still very incompletely developed, is made
up of an elongated mass of cells situated near the raphe. It shows
a marked bend conforming to the flexure of the medulla in the
neck region. Its cephalic pole, as projected on the lateral sur-
face, falls behind the transverse level of the seventh nucleus. Of
the greatest importance is the histological appearance of the
rhombencephalon, the ventral surface of which is made up of the
marginal veil (‘Randschleier’ of His) and in its nuclear free net-
work run the fibers comprising the long association tracts. This
layer, striking in sections on account of the dearth of nuclear
material, forms a brilliant background which permits one to readily
outline the nuclet wandering over its surface later.

From the embryological series of this laboratory definite evi-
dences of the migration leading up to the formation of pontine
nuclei appear in an embryo of 28 mm. (Mall Collection No. 382).
Fig. 2 was drawn from a wax-plate reconstruction of the rhomb-
encephalon of this embryo. Here the degree of medullary
development corresponds pretty accurately to that of the 22 mm,
embryo just described. The behavior of the cells ning the ven-
tricular cavity deserves particular attention, inasmuch as they fur-
nish the neuroblasts for the future pontine nuclei. The ependyma
covering the floor of the fourth ventricle over the basal and alar
plates has lost all signs of the great activity which it showed
during the formation of the tegmental structures. The cells

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN 31

eee

Caudal Wall of Lateral Recess
Roof Atfachment of Fourth Ven}.
he ‘

Fig. 2 Ventro-lateral view of a wax-plate model of the rhombencephalon of a
23 mm.embryo. X 18. (No. 382).

€

32 CHARLES R. ESSICK

composing it are sharply marked out into a definite lamina and
only after searching through many microscopic fields is one able
to detect evidences of cell division. In marked contrast to this
inactive region, the lip-plate which makes up the caudal wall of
the lateral recess and the roof of the fourth ventricle just behind
it, is found busily engaged in producing new elements. The fur-
row formed by the attachment of the roof plate to the medulla,
contains great numbers of karyokinetic figures in every high power
field of the microscope and in this neighborhood the ependymal
zone is not so sharply differentiated into such a thin layer as covers
the medullary floor nearer the midline. Its cells are more closely
packed, its nuclei take on a deeper stain, and the line of demarca-
tion from the subjacent tissue is partly destroyed by the proto-
plasmic processes of the new neuroblasts which are beginning to
push toward the surface of the brain. The exact manner of arriv-
ing at the surface is illustrated by a more fortunate section (fig.
3) through a slightly older embryo. Here the deeply staining
cells, poor in protoplasmic envelope, may be seen to leave their
position near the ventricular cavity, and to come together at the
surface where they form a tin sheet of closely arranged cellular
material. When once they have gained the surface of the brain
they migrate toward the pontine flexure, always preserving their
superficial position.

By referring to fig. 2 a very good idea can be obtained of the
zone of proliferating cells and the area covered by the migrating
neuroblasts that have gained the surface of the rhombencephalon.
I shall omit the description of the arcuate formation for the pres-
ent and consider only that narrow elongated column of cells which
is seen to turn toward the pontine flexure. It is very easy to
identify the densely-staining closely-arranged nerve cells in sec-
tions and I have imitated the appearance one gets specimens
stained in toto by shading this column. The cells, that have left
the ventricle, converge into a well-defined band which, as it curves
around the restiform body, embraces the more anterior of the root-
lets of the glossopharyngeal nerve and passes between the facial
and acoustic nerves as far forward as the trigeminal nerve. At
this stage the cellular sheet is very thin, being but 4-5 cells deep

DEVELOPMENT Of THE NUCLEI PONTUS IN MAN 33

between the seventh and eighth nerves and where it ends behind
the fifth nerve being but a single cell in depth.

We have then a narrow well-defined band of neuroblasts derived
from germinal centers situated along the attachment of the roof
of fourth ventricle to the medulla, and moving over the surface of
the brain toward the pontine flexure. The histological character-
istics make it possible to trace them as far as the trigeminal
nerve as a sheet which gradually thins out toward its advancing
edge. One might well compare the process to ice growing over a
pond, yet unlike the latter the new material is formed at the shores
only and the whole sheet moves out over the surface its thin
advancing edge to meet its fellow from the opposite side.

It should be noted that the degree of development of the
rhonibencephalon does not always correspond absolutely with the
measurements of the human embryos given in this table. A
priori, it would not be expected that at any given stage each organ
would always be found to correspond to those of another embryo
of like measurement, but in addition to the personal elements in
measuring, the fluid in which they are measured often accounts for
the difference of a few millimeters more or less.. Embryo No.
405 measuring 24 mm. shows a younger stage from the standpoint
of pontine development than No. 382 just described. In the
former embryo we can see the same active participation of the
ventricular epithelium in the production of new elements and the
same distribution of karyokinetic figures, yet the front ranks
of advancing neuroblasts have only reached the level of the facial
and acoustic nerves. This would give us a stage slightly younger
than No. 382 where the advancing edge has gained the transverse
level of the trigeminal nerve. Furthermore another possible error
is introduced by the measurements which I have made to show
the growth of the pons and they must be interpreted freely since
the plane of section is rarely perfect. Obliqueness of section
therefore precludes accurate comparison yet the differences are
great enough to draw general’ conclusions.

Nos. 227, 75, and 86 (measuring 30 mm.) furnish valuable
steps leading up to the fusion of the columns of the advancing
neuroblasts derived from the two halves of the brain. The fre-

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 1

34 CHARLES R. ESSICK

quency and wide distribution of the karyokinetic figures occurring
around the attachment of the ventricular roof as well as the caudal
wall of the lateral recess, speak for the active participation that
the rhombic lip is taking in the production of pontine nuclei and
as a result the roof and recess wall are thickened perceptibly. The
last fetus has been sectioned transversely through the medulla
and fig. 3. shows well the large production of cells along the roof
attachment. It is impossible to figure dividing cells at this mag-
nification, yet in this one section which I have illustrated, I was
able to count as many as 75 karyokinetic figures immediately
beneath the membrana limitans interna. A very few could be

Fig. 3 Germinal centers for the basilar nuclei at the roof attachment of the
fourth ventricle in a 30 mm. fetus. X 6.0. (No. 86, slide 35, section 1).

made out where the cells converge at the surface, while only an
occasional mitotic figure is met with among the cells turning
around the restiform body. No evidences of indirect division
could be found in the layer when it has arrived in front of the
pontine flexure. Wesee, then, that the mitosis is confined sharply
to the central canal. The increased production of new elements
is also brought to one’s attention by the increase in depth of the
column passing between the facial and acoustic nerves to 10-15
superimposed neuroblasts; moreover those which had gained the
trigeminal nerve in the 23 mm. embryo have moved forward and
mesially in front of the pontine flexure.

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN 35

The three specimens measuring 30 mm. give us the final steps
in the completion of the first anlagen of the nuclei pontis. In
No. 227 the cell lamina has not reached the ventral median fis-
sure, it being possible to trace it from the border of the fourth
ventricle forward to the fifth nerve where it curves mesially at
almost a right angle toward its counterpart from the other side.
The sagittal sections containing the sixth nerves mark the thin
advancing edges of the cellular sheets approaching each other from
the two sides of the brain. In No. 75 the most advanced cells
have almost succeeded in gaining the midline, while in No. 86
the two columns have fused across the raphe. During their
entire course from the rhombic lip on the dorsal surface to the
raphe on the ventral surface, the wandering neuroblasts have
kept a superficial position, only occasionally is there any tendency
for any of the cells to penetrate into the clear, almost nuclear-free
marginal veil. As they pass between the seventh and eighth
nerves the cells are constricted into a narrow band 0.2 mm. wide,
but on reaching the pontine flexure they spread out into a fan-
shaped layer 0.6 mm. in caudocephalic extent. The lemnisci
medialis and lateralis, which up to this stage had occupied a super-
ficial position, are now covered over by a thin bridge of tissue and
we can begin to speak of a tegmental and basilar part of the pons.

Nos. 211 and 145 (33 mm.) are cut sagitally and give us an |
opportunity to study the earliest pontine nuclei in their relation
to the emergent nervus abducens. Fig. 4 is a camera lucida out-
line of the section through this nerve. The axones after leaving
their nucleus take a ventro-cephalic course through the tegmen-
tum and emerge from the neural tube just behind the most
prominent part of the pontine flexure. The young pontine neuro-
blasts, on the other hand, lie wholly in front of this flexure,
spread out into a sheet whose caudo-cephalic extent is 1.25 mm.
and whose depth is 0.057 mm. at its thickest part, tapering down
to the thickness of a single cell both caudally and cephalically.
Between the most cephalic rootlets of the sixth nerve and the most
caudal cells of the pontine nuclei is an appreciable interval
(almost 0.5 mm.) so that one is at once reminded of the condition

36 CHARLES R. ESSICK

Olivo- arcuate
migration

Olivo- Arcuate
migration.

~f Olivo- Arcuate *
migration

6 oe

Figs. 4to7 Camera lucida tracings of sagittal sections through the rhomben-
cephalon of human embryos from Prof. Mall’s collection. The nucleus facialis
(Nu. viz) has been dotted in by profile reconstruction.

Fig. 4 83mm. fetus. xX 9.5. (No. 145, sl. 19, sect. 3).

Fig. 5 35mm. fetus. xX 9.5. (No. 199, sl. 37, sect. 1).

Fig. 6 50mm. fetus. 7.2. (No. 96, slide 48).

Fig. 7 115mm. fetus. X 7.7. (No. 219, sl. 40, sect. 4).

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN 30

seen in the adult lower animals where the abducens nerve emerges
from the brain some distance behind the pontine formation.

At 35 mm. (No. 199) the increased cellular activity around the
wall of the fourth ventricle is shown by the great numbers of
dividing cells and the twofold increase in depth of the migrating
column passing between the facial and acoustic nerves. Already
the neuroblasts which are crowding from both sides toward the
midline have piled up over the ventral surface of the brain, so
that, near the raphe, they are now four times (0.22 mm.) as deep
as the stage preceding. The free interval between the emergent
abducens and the caudal edge of the pons is decreased to half
what it was in the 33 mm. embryos. The important contribu-
tion which this embryo adds to the development of the basilar
part of the pons, consists in a few strands of longitudinal fibers
lying near the midline within the thin sheet of superficial neuro-
blasts newly descended from the lateral walls of the ventricle.
Extending in a direction parallel to the axis of the central nervous
system, these inconspicuous fiber bundles separate from the well
developed bundle of axones comprising the lemniscus medialis at
the level of the cephalic edge of the pontine sheet and plunge into
the latter where they take up a middle position as far as its caudal
edge. Here again they leave the nuclei pontis and join the fiber
mass constituting the medial lemniscus. It is impossible to
trace these isolated fibers except where they lie among the pon-
tine nuclei, but, as we shall see when, by continual addition to
their number, more of their course can be determined, these few
strands represent the first beginnings of the longitudinal fibers
which are recognized in the basilar part of the adult pons as the
cortical projection system. They are represented in fig. 5 by two
dotted lines traversing the pontine nuclei. Concerning the first
appearance of the pyramidal tract there has been a general unan-
imity of opinion, the most important work being that of Flech-
sig’s work on myelinization time. Tiedemann (16) thought he
saw pyramids in fetus of the third month but he was evidently
looking at the eminences formed by the inferior olives which at
this time lie adjacent to the ventral median fissure and cause an
elevation in the position occupied by the future pyramidal tract.

38 CHARLES R. ESSICK

The latter subsequently forces Hts way between the olives and
gradually displaces them laterally. Certainly at this time the
number of fibers making up the pyramidal tract is insufficient
to cause the surface markings on the medulla which we know as
pyramids. Reasoning back from the interval of time—four
months—between the appearance and mayelinization of other
systems, Flechsig came to the conclusion that the pyramids must
first be laid down between the middle and the end of the fifth
month. In cross sections through the olive of an 80 mm.! (erown-
rump) fetus he is unable to recognize any tissue which may be
regarded as matrix for the pyramids but thinks they arise from
fibers growing down from the cerebral cortex with remarkable
rapidity in the second half of the fifth month. W. His (04) has
given a valuable table (p. 155) showing the various fiber systems
which he was able to identify at each stage of embryonic growth.
In a fetus of 83 mm. he was unable to find the pyramidal tract,
but at a length of 120 mm. he saw evidences of its appearance
together with cross pontine fibers. He gives us no statement
as to the part of the brain in which he observed the pyramidal
tract, merely noting its presence or absence in the various embryos
in his collection. With this statement of the present knowledge
of the cortical projection system, I shall omit until later the
reasons for believing that the few axones isolated by the early
pontine nuclei represent the anlage of the cortico-spinal tract in
this 35 mm. fetus.

No. 95 (46 mm.) is the youngest stage in which I could determine
cross fibers among the pontine nuclei. They are most conspic-
uous at the lateral borders where they gather together into com-
pact strands to form the brachia pontis. Here the fibers have
a superficial position, embracing laterally the corpora restiformia
as the latter turn sharply into the cerebellar hemispheres. It is
possible to trace the axones coming from the pontine nuclei for
some distance into the cerebellum until their course parallels the

1 The crown heel measurement which Flechsig used was 11 centimeters. For
the sake of ready comparison with my study I have put his measurements into this
form from the table given by Mall in Handbuch der Entwicklungsgeschichte des
Menschen-Keibel und Mall, Leipzig, 1910, p. 205.

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN 39

large mass of fibers coming up in the inferior peduncle. The
trigeminal nerve in this stage sends its rootlets through these
cross fibers in an oblique direction to reach its intramedullary
nuclei. The cerebro-spinal neurones which were seen among the
relatively thin sheet of nuclei covering the ventral surface of the
pontine flexure in the stage just described (85 mm.), are increased
enormously during the interval left in this series, and we have a
striking similarity to the picture one gets in sagittal sections
through the pontine region of the adult brain. The thickened
layer of cells (now 0.642 mm.) are invaded by large anastomosing
strands of fibers which collect at the cephalic and caudal border
of the nuclear sheet into a solid bundle. This behavior is one of
the pecularities of the cortical projection system as it lies among
the pontine nuclei where, as is well known, the otherwise compact
fiber tract is broken up into smaller fasciculi by the cross fibers
and nuclei of the pons. Caudad the reunited fibers fuse insepar-
ably with the median lemniscus soon after leaving the nuclei
pontis; cephalad I am unable to trace them beyond the cephalic
flexure.

His has suggested that this interweaving of cross fibers of pons
with pyramidal tract points to an alternating time of deposition
of the component parts of the two systems—the development
proceeding in a direction away from the central canal. This for
the most part is true. We find the new cells which have migrated
from the ventricular walls, spreading themselves over the surface
of those already descended to the pontine flexure, and as the new
axones come from the cortex they tend to grow among the younger
nuclei, i.e., to grow nearer the surface. Thus each fasciculus
when it enters the pontine nuclei, pushes along near the surface
but it is soon deeply buried by new cells which are continually
streaming down from the ventricle. As a result there is a separa-
tion of the pyramidal tract into a series of fasciculi which unite
again at their exit from the caudal border of the pons. Some of
the cells, however, after passing between the seventh and eighth
nerves forsake their superficial position and plunge between the
cross fibers of the pons. This is well illustrated in fig. 9 the more
deeply staining young cells are seen forcing their way between

40 CHARLES R. ESSICK

the transverse pontine fibers going into the brachium pontis.
Moreover the fasciculi of the pyramidal tract keep on growing so
that one must infer that some of the axones coming down from the
cortex add themselves to the bundles more deeply placed in the
pons.

During the period of growth between 35 mm. and 46 mm.
enormous numbers of neuroblasts have come down from the lateral
boundaries of the fourth ventricle. ‘These new cells together with
their processes sent out transversely and the cortical axones
threading their way among them have increased the thickness of
the basilar portion to 0.642 mm. There is a tendency for the cells
from both sides to crowd toward the midline, thus giving rise to
the typical crescentic shape of the pontine nuclei which one ob-
tains in transverse sections through the pons. The increase in
thickness is also accompanied by an increase in caudo-cephalic
extent. The latter, however, does not proceed with the same
proportional rate as the former, so that the sagittal sections of
pontine nuclei are becoming more and more oval. In spreading
caudad the interval between the nuclei pontis and the abducens

‘rootlets has been gradually reduced until in this fetus the more
cephalic axones are surrounded by pontine nerve cells.

In fetus of 50 mm. (Nos. 84, 96, 184) the number of cerebro-
spinal neurones have increased to such an extent that it is now
possible to follow them with sufficient accuracy to be certain that
we are dealing with the axones of the cortical projection system.
The fibers splitting up among the pontine nuclei already form
comparatively large bundles (fig. 6) which are collected together
into a solid fasciculus at the cephalic end of the pons. Here they
come into close relationship with the lemniscus medialis, but it
is not impossible to trace the large fiber mass into the internal cap-
sule. Traced cerebrally the crura gradually diverge from the
midline and turning around the cephalic flexure they lie ventral
and lateral to the nucleus hypothalmicus, while the medial lem-
niscus has a more dorso-lateral position with regard to this
nucleus. The fibers making up the pyramidal tract can be traced
definitely into the internal capsule. Spinal-ward I have been
unable to differentiate the projection system from the medial lem-

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN 4]

niscus soon after it has left the caudal border of the pons. There
is then no question but that we have been dealing with the begin-
nings of the pyramidal tract as early as 35 mm.; its behavior
among the pontine nuclei making identification certain. Former
observers have confined most of their attention to the medulla
oblongata where it is hopeless to try to pick out the few strands of
fibers when they first grow down from the cortex. The increase
in number of these axones is so gradual that it is only in the older
fetus where enough fibers are collected to form the surface mark-
ing on the medulla which we can recognize as pyramids. Flechsig
is sure that there are no pyramids at 80 mm. and probably “‘the
pyramidal tract is completely lacking.” To harmonize the sys-
tem with other observed facts he assumes that they must grow
down rather rapidly from the cortex when once they start, since
their myelinization occurs after birth and the usual interval be-
tween the formation of a nerve fiber and its acquirement of a mye-
lin sheath is about four months. This of necessity would have the
pyramidal tract appear about the middle to the end of the fifth
month or 14 to 16 em. To this one must answer that a myelin
sheath does not appear on every axone of this system simultane-
ously; it begins rather on isolated fibers and is first complete at the
age of two years.

In these fetus of the eleventh week the basilar part of the pons
has reached a thickness of 0.7 mm. (fig. 6). The abducens nerve
rootlets are almost entirely surrounded by nuclear material after
they leave the tegmentum, only the caudal two or three fasciculi
being free. Great numbers of neuroblasts are encountered pass-
ing between the seventh and eighth nerves, forming a stream
0.16 mm. deep, while the germ centers at the ventricular margin
are busily producing new cells. A fortunate sagittal section
through No. 96 has been illustrated to show the participation
which the greatly thickened lateral recess wall takes in contribut-
ing cells to the pons. For purposes of orientation a wax-plate
reconstruction was made with the section drawn on its cut sur-
face (fig. 8). As the cerebellum in its growth crowds against the
medulla, this caudal wall is flattened out and becomes part of the
mesial wall of the recess. A separation of the cells coming from

42

CHARLES R. ESSICK

got of Fourth Ventricle

Fig. 8 Sagittal section through the cerebellum and lateral recess

of a 50 mm.
fetus from a wax-plate reconstruction. 15. (No. 96).

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN 43

the recess wall from those of the ventricular roof is purely arbi-
trary, since the two origins are really continuous with each other.
This association in the adult was first pointed out by Orzechow-
ski (08, p. 41). For embryological reasons the nuclear thicken-
ings of the lateral recess wall with their accompanying fibers
occurring in the adult should be included in the structure which I
have termed the corpus ponto-bulbare. At this stage the cavity
of the fourth ventricle shows a peculiar tendency to form small
outpouchings along the attachment of the roof at the place where
the pontine nuclei are being formed. From two to four such
recesses can be made out extending laterally for a considerable
distance from the main ventricular cavity and causing the exter-
nal surface to be thrown up into ridges. In section they may be
round or slit-like and are lined with deeply staining cells, great
numbers of which are found in process of karyokinetic division.
The production of neuroblasts at this stage is enormous and these
lateral extensions from the ventricle furnish a greater expansion
of ependymal surface and thus increase the germ layer where cell
division can take place.

In the older fetus the system of ventricular outpouchings be-
comes more complicated and secondary processes are formed
which are distinctly tubular. The size of the lumen varies, being
sometimes less than the width of a single nucleus. It is always
lined with a simple layer of cells which are definitely ependymal
and as long as pontine nerve cells are being formed these tubules
can be made out with a little difficulty among the closely packed
neuroblasts but always the center of mitotic activity. With the
emigration of the last of the new elements the ventricular pro-
longations stand out with much greater clearness. This is partic-
ularly well shown in the five and eight months fetus—in the latter,
one is struck by the greater number of such tubules both in the
roof attachment and the caudal wall of the lateral recess.

At the beginning of the fourth month, as shown by No. 172
(80 mm.), the evidences of marked cellular activity, i.e., extensive
mitosis and deeper staining are still present around the roof
attachment of the fourth ventricle and the caudal wall of the
lateral recess. This fetus does not illustrate any new principle

44 CHARLES R. ESSICK

@ Ventriculus EF:
Quartus

s :\Tractus
Nuc *{ Solitarius

Hypoglossus 35

Migrating

Pontine Nuclei

Corpus
Restitorme

Corpus. ey
Ponto- Bulbare

Fig. 9 Oblique section through the rhombencephalon of an 80 mm. fetus.
x 7.5 (No. 172, sl. 200, sect. 2).

but furnishes a beautiful single section which includes all of the
relations of the cells wandering to the pontine nuclei. Owing to
the cervical flexure a cut which sections the spinal cord trans-
versely passes tangentially through the medulla and pons (fig.
9). The great thickness at which these sections were cut gives
the migrating column of nuclei a very deep red stain, almost
black, and I have represented it as black; the plane of section falls
behind the roof thickening which marks the true germinal centers,
but it does show with remarkable clearness the whole path of the
neuroblasts starting from the ventricular edge, encircling the
corpus restiforme, passing between the facial and acoustic nerves,
to take their place among the cross fibers of the pons.

During the period extending through fetus 508 (143 mm.)
there is a continued addition of cells to the basilar part of the pons.
In a fetus of the thirteenth week, 96 mm. (No. 484), the maximum
production of new cells has been reached. At this time the ven-
tricular edges on both sides of the brain are full of karyokinetic
figures and extending from these places are two thick columns of

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN 45

closely packed young cells which pour into the pons between
the seventh and eighth nerves. The cells which had already
descended from the rhombic lip show no tendency to assume the
ganglionic form. They are still rather closely arranged, the cyto-
plasm scanty and clear, the nucleus small and containing one or
two chromatin condensations but no real nucleolus. ‘Two differ-
ent reactions toward the haematoxylin can be made out among
the nuclei of these cells—the one quite densely staining and
usually smaller nucleus, the other slightly larger and more vesicu-
lar. The latter form the larger proportion of nuclei. From the
extensive cross-fiber system already present one must assume that
great numbers of these pontine cells have sent out nerve processes
and have taken up their final position. The newly added cells
tend to remain superficially yet a considerable portion push in
between the transverse fibers and cells already fixed. Although
the stain of No. 490 (113 mm.) does not permit of good histo-
logical study, yet it may be readily seen that the pontine nuclei
are still receiving great numbers of new elements from the rhom-
bie lip. At 143 mm. (No. 508) the production of neuroblasts
destined for the basilar parts of the hind brain has diminished very
appreciably. The mitoses around the ventricular margin are
fewer and the ependymal lining has begun to be separated quite
sharply from the underlying nervous tissue. Many of the cells
in the path of migration have larger and clearer protoplasmic
bodies, apparently unwilling to complete their journey to the
pontine region. Others appear as all of those in earlier stages
with elongated almost naked nucleus pushing on toward the pons
before assuming the ganglionic form. Of the cells which have
long since gained their permanent position in the pons, many can
now be recognized as ganglion cells. The nucleus is very much
larger, although rarely showing a distinct nucleolus—the cyto-
plasm, paler than the framework in which it is embedded, is also
increased in amount. The greater number of cells have grown
slightly—possessing a smaller, more densely staining nucleus .
surrounded by a clear protoplasmic envelope.

During the interval between 143 mm. (No. 508) and 188 mm.
(No. 509) the formation of neuroblasts ceases entirely and the

46 CHARLES R. ESSICK

further development of the pons consists of an addition of axones
(with their later myelinization) and the maturity of the nerve
‘cells. The rhombic lip has given up all signs of activity in the
next stage of which I had access to serial sections. In No. 509
(188 mm.) the lining of the central canal is uniformly at rest in
the medullary region and is now as sharply demarcated around
the roof attachment as that covering the basal and alar plates.
The roof thickening now reminds one of those sections through the
adult medulla which pass through the ponticulus of Henle. Com-
paratively few of the cells have not descended into the pontine
region but are taking on the characters of adult ganglion cells
along the path where the pontine cells migrated at an earlier period.
Just how many cells fail to move into the pontine region but take
up their position around the restiform body varies in the differ-
ent brains. This helps us to understand the wide differences
which were noted in the size of the fully developed corpus ponto-
bulbare. With the disappearance of the closely packed nuclei
around the attachment of the rhombic lip and the consequent
elearing up of the roof thickening, the ventricular outpouchings
stand out with great clearness. They are not very unlike tubular
glands with a single cell lining them and show a tendency to
branch frequently.

The increase in the number and size of the bundles of cross
fibers in the pons gives the basilar part a greater thickness in
the adult where one finds the fiber material outweighing the
nuclear material. In the first beginnings on the contrary just the
reverse holds true, and the nuclei pontis alone form the protu-
berances in the pontine region. The individual ganglion cells,
although larger than the preceding stages, are still immature.
The protoplasmic bodies do not accept the counter stain and few
of the nuclei possess a well formed nucleolus.

By the eighth month (No. 491) the protoplasm of the ganglionic
cell is no longer clear but takes up the counter stain. Many of
them now look like the ganglion cells of the adult except for their
smaller size. The entire migratory path is strewn with cells.
Passing between the facial and acoustic nerves the column can be
traced around the restiform body into the roof of the fourth ven-

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN 47.

tricle and the mesial wall of the lateral recess. A great many of
the cells can be recognized as ganglion cells of the corpus ponto-
bulbare, but as a whole the appearance is one of immaturity.
Here and there can be found a cell whose protoplasm stains but as
a rule the large vesicular nuclei are surrounded by a colorless
zone. The migration of nuclear material in the medulla has
ceased entirely in this stage and one has to expect only the matur-
ity of those elements already present.

Throughout the description I have disregarded a very impor-
tant factor in development which transforms the hind brain of
an embryo of the second month into the adult form: I refer to the
obliteration of the pontine flexure. The maximum flexure in the
neural tube occurs about the time of the appearance of the first
nerve cells on the ventral surface and then diminishes gradually
so that at birth there is still an indication of it on the ventricular
floor in front of the emmentia abducentis by a furrow running
transversely. The relations of the tegmental nuclei to one
another as well as to the olivary complex are distorted by the
extreme flexion of the brain in the pontine region. The reduction
of this may be regarded as taking place around the nucleus n.
abducentis as an axis, since it is situated immediately beneath
the ventricular floor just behind the bend in the brain. The other
nuclei, superior olive, facial nucleus and olivary complex are dis-
tributed around its circumference and are consequently separated
from one another. A glance at fig. 1 will show this arrangement.
Figs. 4, 5, 6, and 7, are camera lucida drawings of sections through
the nucleus n. abducentis with its emergent root bundles, which
were selected from such sagittal series as illustrated the change in
position of the nuclear masses during the obliteration of the pon-
tine flexure. The nucleus facialis is projected into the section as
indicated by the broken lines. As the neural tube unbends,
the olivary complex and pontine nuclei are gradually pushed
toward one another until the cephalic tip of the former comes to
be covered by the latter. The abducens nerve which in younger
stages (figs. 4,5 and 6) pursues a straight course within the me-
dulla is bent by this process, so that it takes a caudal direction in
order to reach the surface of the brain (fig. 7). The facial nu-

48 CHARLES R. ESSICK

cleus at first separated from the olivary body by a considerable
interval comes to lie in the same transverse section as the latter,
while the pontine nuclei cover up the cephalic two-thirds of this
nucleus. We have, then, in addition to the mere increase in size
of the pontine and olivary nuclei an actual alteration of their posi-
tions as a result of the straightening out of the neural tube. As
a consequence nuclear masses which were separated from one
another, are crowded together and the course of the sixth cranial
nerve altered.

Having considered the origin of the main mass of nuclei pontis,
the possibility of cells from other sources must not be overlooked.
In the region of the pontine flexure near the raphe one can make
out at an early period collections of cells extending from the ven-
tricular floor to the pontine nuclei with which they are connected.
They occupy the position which is held by the nuclei reticularis
tegmenti pontis (Flechsig) in the adult. Long before any cells ap-
peared superficially on the pontine flexure the karyokinetic figures
had disappeared in the ependymal sheet near the raphe, so that it
is highly improbable that the nuclei pontis depends on this portion
of the neural tube for many of its elements. In addition these
cells of the nuclei reticularis tegmenti pontis are evident long
before the pontine nuclei appear and never have the characteristic
appearance of young wandering neuroblasts during pontine de-
velopment. In some of the older embryos a thin sheet (one to two
cells deep) are migrating from the wall of the lateral recess in front
of the dorsal cochlear nucleis but the layer is narrow and composed
of comparatively few cells. These cells join the pontine nuclei
behind the trigeminal nerve. It is hardly necessary to exclude
other sources if one considers seriously the great production of
new cells around the rhombic lip. This begins at 23 mm. and con-
tinues incessantly until the fetus has passed 143 mm. in crown-
rump measurement. Couple with this extensive period the short
time in which any mitotic division is complete and the great
numbers met with in every section and it will not take a great
stretch of imagination to account for all of the cells in the nuclei
pontis.

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN 49

ARCUATE NUCLEI

Examination of different adult brains in microscopical sections
reveals a great variation in the amount of nuclear material which
goes to make up the basilar portions of the brain stem. This
is especially true of the arcuate nuclei where small, more or less
isolated patches of nuclear material may often be scattered along
the ventral and lateral surfaces of the medulla as far as the resti-
form body. The arcuate nucleus proper, the most constant of
these masses, lies near the ventral median fissure superficial to
the pyramidal tract, extending from a point caudal to the olive
up to and fusing with the pontine nuclei. At its caudal extrem-
ity, under the olive, this mass is always of greater dimensions and
tapers off somewhat as it is followed toward the pons—in some
brains disappearing here and there for a,few sections, in others
forming a continuous narrow strip under the whole length of the
medulla. The arcuate nuclei proper, as well as these superficial
isolated masses lying more laterally, will be shown to have a com-
mon origin and at one time to be actually continuous with one
another. The principles governing their development are iden-
tical with those which we have studied in connection with the
pontine formation. The same germ centers around the attach-
ment of the rhombic lip contribute cells which migrate superfi-
cially over the medulla in front of the cervical flexure.

The formation of the arcuate nucleus, unfortunately, is not so
simple as that of the pontine nuclei, but is complicated by the
simultaneous development of the olivary complex. As His has
shown, the latter begins as a migration from the alar plate of the
rhombencephalon early in the second month. Toward the end
of this month the olive can be outlined readily although it has
only a small fraction of the cells which it contains in the adult.
At this time one can make out in embryos of about 20 mm. large
elongated nuclei, almost devoid of a protoplasmic body leaving
the ventricular margin along the attachment of the rhombic
lip. They are arranged in strands of a single cell in depth and two
or three in width, streaming over the surface of the medulla just
under the external limiting membrane. This migration is directed

50 CHARLES R. ESSICK

toward the portion of the medulla which is under the partially
formed olivary complex and recalls the undifferentiated wandering
cells seen in connection with the nuclei pontis. A great many
leave the surface at various points and plunge into the depth to
join the neuroblasts already massed up in the olivary nuclei. A
broad sheet, however, remains superficially and can be traced
from the roof attachment ventrally. In some brains (Nos. 368
and 453) this sheet has moved among the vagus rootlets and
advanced almost to the emerging hypoglossal roots. In another
(No. 22) the migrating cells cover the entire ventral surface of
the medulla just in front of the cervical flexure having met, across
the raphe, those moving down from the opposite side. In other
words, there exists at this period a band of superficial undifferen-
tiated cells uniting the roof attachment on both sides which is
not unlike the early pontine bridge in fetus of 30 mm. The
former begins just in front of the cervical flexure and subtends
one-half to two-thirds of the olivary complex (fig. 2), but unlike
the latter many cells leave it everywhere and make their way into
the substance of the medulla to form gray matter in the interior.

It is striking (1) that the neuroblasts of the developing arcuate
nuclei imitating the pontine formation, pay no attention to the
raphe but cross it in an uninterrupted sheet; (2) that they appear
before the anlage of the nuclei pontis, and if one turns to mamma-
lian embryos (I have studied pig and rabbit) which are slightly
larger than 20 mm. in crown-rump measurement, (3) that a well-
developed arcuate formation exists just as in the human material.

Turning to the adult brain we find each arcuate nucleus a dis-
crete mass which is separated from its counterpart by the raphe.
Moreover the arcuate nucleus is peculiar to man so that from a
phylogenetic standpoint we should expect to find it developing
later than the pontine nuclei inasmuch as it is last to be acquired.
Furthermore, very soon after its formation in pigs and rabbits one
looks for it in vain. At 51mm. only comparatively few cells can
be found, while the superficial layer of migrating cells has dis-
appeared completely from the subolivary region of a fetal pig
of 60mm. In man, on the other hand, when once there is a col-
lection of cells over the medulla in the subolivary region (as in

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN ai

No. 22) all of the later stages invariably show nuclear material
in the position which we know will be occupied by arcuate nuclei.
There appears but one rational explanation which will harmonize
all of these apparently jarring facts which we have determined.
In human embryos at the beginning of the second month there is
an intramedullary migration of cells from the rhomboid lip to
make up the olive, toward the end of the month the path of migra-
tion becomes more and more superficial until many of the cells
actually cross the raphe before plunging into the medulla. In
the lower mammals the comparatively simple olivary complex
soon acquires its allotment of cells and when production of oli-
vary neuroblasts ceases in the roof attachment, those on the
surface soon find their way into the interior. In man, on the
contrary, before the olive has received all of its cells and while the
migration from the rhombic lip is still proceeding actively, neuro-
blasts which cannot be differentiated from those destined for the
olive, begin to wander over the surface among the vagus roots.
These elements stop on the ventral surface near the raphe and
constitute the anlage of the arcuate nucleus. Stated differently,
we are probably not dealing with arcuate formation in human
embryos of 20 mm. where a cell lamina lies on the surface of the
medulla in the place where we know the arcuate nucleus ought to
be.

Just when the arcuate neuroblasts begin to descend from the
rhombic lip can only be conjectured; this uncertainty has led me
to call it ‘olivo-arcuate migration.’ Probably at 30 mm., as
exemplified by No. 86, most of the thick superficial sheet of cells
in the arcuate region represents a migration of olivary elements
(fig. 10). Here the deeply staining nuclei form a continuous
lamina over the ventral surface, the caudo-cephalic extent of
which corresponds to the spinal one-half of the olive. Even older
embryos present this pons-like structure as figs. 4, 5, and 6 illus-
trate. First in a fetus of 80 mm. (No. 172) does one meet with
any large number of superficial neuroblasts under the cerebral
one-half of the olive. Here almost the entire surface of the
medulla is the seat of cellular migration. From the cervical
flexure almost to the pontine nuclei the pyramidal tract is covered

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, No. 1

5 ye CHARLES R. ESSICK

by a superficial sheet of cells many of which are pushing their
way into the medulla near the raphe. The wandering of cells
to the region of the cerebral half of the olive fills up, in this em-
bryo, the gap between the pons and the band (olivo arcuate
migration, (fig. 2) which was first completed in embryos of 20 mm.

The addition of new elements is even more marked in the 96
mm. fetus (No. 484) where great numbers of moving cells are
directed toward the ventral portion of the medulla immediately
behind the pons. Here the cells are leaving not only the rhombic
lip to pursue a course similar to the earliest olivo-arcuate migra-
tion but also from the ventral edge of the thick column of migrat-
ing pontine nuclei. All along the corpus ponto-bulbare of this
fetus neuroblasts can be seen to leave its ventral edge and migrate
directly toward the ventral median fissure. In the adult it is
well known that the arcuate nuclei fuse across the midline as
one nears the pons, although at the caudal end of the olive they
present two discrete swellings which lie some distance from the
midline. Two mechanical factors are concerned in breaking up
this uninterrupted sheet of nuclear material which is so striking
in the younger fetus (fig. 10). These are the formation of the
external arcuate fibers and the growth of the pyramidal tract.
Already in this fetus a considerable number of arcuate axones are
crossing in the raphe, the main mass of nuclei, however, still lie
on either side of the midline (fig. 11). It remains for the con-
stant interstitial addition of pyramidal axones to bring about the
further separation of the arcuate nuclei. It is apparent that the
cortical projection system must occupy a very inconspicuous
part of the cross sectional area of the medulla of this embryo,
when one considers the superficial position of the olivary complex
and their proximity to the ventral medial fissure. Compare with
this the fig. 12 which is a camera lucida tracing of No. 508 (148
mm). The level of this section was made to correspond with that
of No. 484 by choosing both about one-tenth of the distance from
the caudal to the cephalic pole of the olive. The extensive addi-
tion to the pyramidal tract has pushed the olives apart as well as
the arcuates. The latter, remaining superficial to the rapidly
growing nerve system, have been drawn away from one another.

DEVELOPMENT OF THE NUCLEI PONTUS IN MAN 53

Along the raphe can be found a few cells, the remains of the con-
necting bridge, and these persist in this position even in the adult.
Farther laterally one may often find small isolated masses at
almost any point along the periphery of the medulla, the number

Roof Attachment
/ a

4 Corpus
\ Restiforme

Substanhia ;

Gelatinosa \'

10

12

Fig. 10 Cross section through the lower olivary region of a 30 mm. fetus. X
16.5. (No. 86, sl. 28, sect. 13).

Fig. 11 Camera lucida tracing of a cross section through the lower olivary
region of a96mm. fetus. X 8.2. (No. 484, sl. 3, row 2, sect. 7).

Fig. 12 Camera lucida tracing of a cross section through the lower olivary
region of al43 mm. fetus. > 7.3. (No. 508, sl. 3, row 4, sect. 2).

and amount varying with different brains. These represent
portions of the basilar nuclei which have not descended to the
position of the arcuate nuclei proper. It will be remembered that
in No. 508 many of the pontine nuclei are assuming ganglionic

54 CHARLES R. ESSICK

form. This is not true in the arcuate formation where the nuclei
are still very densely staining and there is very little protoplasm
in the bodies. In No. 509 (188 mm.), however, the protoplasmic
body is represented by a clear unstained area around a pale
vesicular nucleus. Here it is possible to speak of young ganglion
cells with certainty although most of the elements are still undif-
ferentiated.

To conclude, then, we have in the rhombic lip or ‘“Rautenlippe’
of His a common ancestor for the olive, pontine nuclei and arcuate
nuclei—the nuclei pontis being formed by a migration through a
restricted pathway, the corpus ponto-bulbare; the nuclei arcuati
along with part of the olive by a superficial migration over the
ventral surface of the medulla.

BIBLIOGRAPHY

Buake, J. A. 1900 The roof and lateral recesses of the fourth ventricle con-
sidered morphologically and embryologically. Jour. Comp. Neur.,
vol. 10.

Essicx, C. R. 1907 The corpus ponto-bulbare—a hitherto undescribed nuclear

; mass in the human hind brain. Amer. Jour. Anat., vol. 7.
1909 On the embryology of the corpus ponto-bulbare and its relation
to the development of the pons. Anat. Rec., vol. 3.

Fuecusie, P. 1876 Die Leitungsbahnen in Gehirn und Riickenmark des Men-

schen. Leipzig.

His, W. 1891 Die Entwicklung des menschlichen Rautenhirns vom Ende des
ersten bis zum Beginn des dritten Monats. Leipzig.
1904 Die Entwicklung des menschlichen Gehirns wahrend der ersten
Monate. Leipzig.

OrzEecHowskI, K. 1908 Ein Fall von Missbildung des Lateralrecessus. Ein
Beitrag zur Onkologie des Kleinhirnbriickenwinkels. Arbeiten aus
dem Neurol. Instit. Wien, Bd. 14.

STREETER, G. L. 1904 The development of the cranial and spinal nerves in the
occipital region of the human embryo. Amer. Jour. Anat., vol. 4.
1907 On the development of the membranous labyrinth and the acous-
tic and facial nerves in the humanembryo. Amer. Jour. Anat., vol. 6.
1912 The development of the nervous system. Manual of Human
Embryology, Keibel and Mall, vol. 2.

TIEDEMANN, F. 1816 Anatomie und Bildungsgeschichte des Gehirns in Foetus
des Menschen. Niirnberg.

THE SHEATH OF THE SINO-VENTRICULAR BUNDLE!
RUSKIN M. LHAMON

Assistant Professor of Anatomy, Philippine Medical School

From the Anatomical Laboratory of Stanford University

FIVE FIGURES

The present investigation, undertaken at the suggestion of
Professor Meyer, is concerned chiefly with certain aspects of the
connective tissue sheath around the atrio-ventricular bundle,
or, as suggested by Retzer, the sino-ventricular bundle. As
this sheath has recently been mentioned and described to a
greater or less extent by many investigators, it is hoped that the
present communication will be of some interest.

A review of the literature upon this subject will be given only
so far as the connective tissue sheath is concerned. Kent in
1893, noted a considerable development in the region of the auri-
culo-ventricular groove. In 1906, Tawara while recognizing
the Purkinje fibers as the branchings of the bundle, stated that
the bundle and its branches were everywhere isolated from the
cardiac muscle by connective tissue up to the point where the
transition into muscle cells takes place, and that at this point
also a transition of the connective tissue sheath into perimysium
occurs. Keith and Flack in 1906, Fahr in 1908, and Méncke-
berg in 1908, were impressed with the fibrous sheath and con-
firmed Tawara’s findings. De Witt in 1909, while making dis-
sections of the bundle and its branches preparatory to reconstruct-
ing it, noted that she could not remove the sheath because of the
danger of breaking the strands of the bundle in many places
along the finer branches and at the region of the nodal points.

1Tt is believed that this is the first application of the injection method to the
demonstration of the ramifications of the conductive system. These results are
so easily obtained and give such a complete picture of the system, especially in
the left ventricle, that after testing it the editor cannot refrain from thus calling

attention to the value of the method.

55

56 RUSKIN M. LHAMON

Her models are, therefore, representations of the bundle and con-
nective tissue sheath. Curran in 1909, stated that a mucous
bursa, was constantly present in relation to the bundle and its
branches in the human, calf, beef, and sheep hearts which he
examined. According to Curran, the so-called bursa is best
demonstrated by gross dissection, but he also succeeded in dis-
tending it by a blow pipe, the character of which was however
not stated, and in many cases, saw bubbles of air in the region
of his ‘reticulum’ upon the interauricular septum, and upon the
sides of the interventricular septum, along the courses of the right
and left branches. What Curran means by ‘reticulum’ is not
quite clear, but presumably he refers to the gross structure of the
bundle in the region where the ‘Knoten’ (Tawara) is found. He
also stated that the bursa contains a fluid which is of greater con-
sistency and more tenacious than ordinary lymph. In some
cases, it was said to be so plentiful as to exude in the form of a
droplet when the bursa was punctured, in others the overlying
tissues could be made to puff out by making pressure on both
ends of the branch. The place where this bursa was said to reach
its greatest dimensions was just under the cartilaginous part of
the septum, where the impact or friction with surrounding parts
would presumably be greatest. From his microscopic work
Curran concluded that the nature of this space varies from “very
loose areolar tissue with fluid in the cellular spaces, which always
connect with each other along the line of the bundle, to distinct
cavities, filled or lubricated with fluid and no trabeculae crossing
the intervening space. In the usual form, there are one or two
large spaces or one continuous space with very fine trabeculae
crossing from the walls of the canal to the auriculo-ventricular
muscle.’’ Curran, however, does not state whether these spaces
are lined with a special endothelium or by a membrane similar
to synovial membranes.

Dogiel in 1910, sought to throw doubt upon the existence of the
auriculo-ventricular system as a muscular connecting link between
the heart chambers and stated that the origin and the termina-
tion of such a nerveless bundle, its relation to the nerves, to the
connective tissue, and to the rest of the heart muscle, are anatom-
ically and physiologically unknown, and according to him, no

THE SHEATH OF THE BUNDLE OF HIS Bik

evidence has been shown to prove that the fasciculi of the bundle
are separated from the heart muscle by connective tissue. Dogiel
also states that the endocardium is not sufficiently transparent
so that the fasciculi of the bundle and connective tissue sheath
can be seen with the unaided eye!

The subject of the ‘Sinus-knoten,’ upon which Keith and Flack,
Keith and Mackenzie, Fahr, Ménckeberg, Koch, Thorel and
Wenkebach have done recent investigations, and the relations of
the ‘Sinus-knoten’ to a connective tissue sheath, are not con-
sidered in this article, for the writer has sought to limit himself
to the relations of the connective tissue sheath of the ‘Knoten’
and the parts of the bundle peripheral to this.

In the present investigation eighty-seven hearts were used;
of these thirty-eight were beef hearts, thirty sheep, fifteen calf,
and one was from a lamb six weeks old. About three-fourths of
these hearts were used for injections, and many of them were
afterwards dissected. Others were used in part for microscopic
work. The rest were used for both gross dissection and micro-
scopic work. Pieces of tissue including parts of the bundle with
some of the cardiac muscle were removed from regions (where
the bundle begins its course from the place) of the ‘Knoten,’
where the bundle passes under the cartilaginous septum and the
main branches pass down upon the latter, and also farther down
from regions where the branches of the bundle were very small.
Serial sections of the complete system were not made, however,
for it was thought that an examination of the bundle with its
connective tissue envelope in different regions in a number of
hearts would be entirely adequate for the purpose of this investi-
gation. The tissues were fixed in formalin or Zenker’s fluid,
dehydrated, embedded in paraffin, and cut 5 to 7} and 10 micra
in thickness. The ordinary haematoxylin and eosin stain was
used, although some sections were stained with Van Gieson. In
some instances where tissue was removed for miscropic study after
injection of the sheath, the celloidin method was used.

Since a description of the atrio-ventricular bundle as it is seen
by gross dissection would be a useless repetition of what has
already been well done by others, it is omitted here. The bundle
is easily located and easily dissected from its origin to the terminal

58 RUSKIN M. LHAMON

fasciculi and in all essential points I found it as described in the
literature.

Although twenty uninjected hearts of beef, calf, and sheep,
and others which were injected, were carefully dissected to deter-
mine the presence of a bursa, no evidence whatever of the exist-
ence of such a structure was obtained. The sheath and bundle
with its larger and smaller branches could not be freed from its
bed until the strand of connective tissue uniting it with the under-
lying musculature had been broken. Between the bundle and the
heart muscle there seemed to be a line of cleavage in places where
the connecting trabeculae were very fine and consequently broke
easily, yet, these fine connections were always present. Under
the cartilaginous or the bony septum of the beef heart, where
the main bursa was located by Curran, a considerable amount of
loose connective tissue was constantly present in the hearts which
I examined, but a definite and preformed bursal space was never
found. At the point where the left branch of the main bundle
passes out from under a layer of muscle to le immediately
beneath the endocardium, a slight depression was noted usually
in the endocardial layer. When the endocardium was carefully
dissected off from this place and the edge of the layer of muscle
lifted up with a forceps, very fine trabeculae were constantly seen
connecting the muscle with the sheath of the bundle. Since
this was the place where Curran located the main bursa, and
where he stated the greatest friction existed between heart muscle
and bundle we should expect some evidences of a bursa here. If,
as Curran seems to think, a protective mechanism be necessary
this loose connective tissue sheath could perhaps serve for pro-
tection of the bundle against the forcible impact of the surround-
ing structures.

A careful examination of the stained section from uninjected
specimens failed likewise to reveal any space. On the contrary,
the sections showed a thick layer of connective tissue enclosing the
bundle. On the one hand, this layer passed over into the con-
nective tissue of the cardiac muscle; on the other, it formed the
envelope and framework within which the bundle lay embedded
so that each strand had its sheath. The stained sections further

THE SHEATH OF THE BUNDLE OF HIS 59

showed that the attachment of the sheath to the individual
fasciculi was relatively loose, and whether the latter were regular
or irregular, the sheath was usually closely applied to them,
although a slight separation was seen in some cases. In the main
mass of the bundle, small blood vessels, nerve bundles and some
fat were noticed.

Repeated attempts to demonstrate the presence of a fluid
as found by Curran, were made by pressing on the superficial
parts of the bundle. For this purpose, hearts were chosen where
the main branches of the bundle were superficial enough to form
a plainly visible and palpable strand under the endocardium.
Since the right branch of the bundle is more cylindrical in form
and lies upon the convex side of the septum, it seemed better
adapted for this purpose. If a bursa containing fluid exists, it
should be found most easily along this part of the bundle. How-
ever, when pressure was made at distant points and the fingers
were made to approach each other, no bulging or pouching of the
endocardium from fluid pressure underneath occurred. What
happened was a gathering of the endocardium and the tissues
superficial to the bundle into a number of ridges or undulations,
running transversely to the long axis of the bundle. These
results certainly speak against the existence of fluid in a bursal
space and can have resulted only from the tissues being bound
down tightly to the underlying structures. Attempts to with-
draw fluid by means of a very fine capillary pipette were also
unsuccessful.

The injection of various fluids into the tissues directly beneath
the endocardium surrounding the bundle and its branches, was
also used as a means of demonstrating the presence or the absence
of a bursa. For this purpose sixty hearts were used. ‘Twenty-
nine of these were beef hearts, twenty-four sheep, six calf and one
from a six weeks lamb. An ordinary hypodermic syringe with a
relatively fine needle was used for these injections. Watery
suspensions of India ink, of finely precipitated Prussian-blue,
of Prussian-blue gelatine, and at times air, were used as injection
media. These were injected into the tissues between the bundle
and the endocardium, and between the bundle and the heart

60 RUSKIN M. LHAMON

muscle along both the right and the left branches. At times the
needle was pointed downward, at others, upward. In order to
reach the point just under the cartilage in the ox heart where Cur-
ran described a bursa a centimeter in diameter, the needle was
inserted from high up on the left side of the septum under the
right cusp of the aortic valve, and directed downward and to the
right a distance of 8 to 10 mm., this procedure being employed
as a result of examination of dissected specimens. The amount of
pressure used in making the injections was not gauged, but it -
was very slight, firm pressure never seeming necessary.

The results of the first injections were good pictures of extrava-
sation, and subsequent dissections showed this to be the case. No
confining limits could be made out, for the injection material
extended as readily into the connective tissues between the car-
diae muscle fibers and out under the endocardium as into the
tissues immediately around the bundle. This extravasation was
found in all regions, whether along the course of the main bundle
under the cartilaginous septum, or along the larger or the smaller
branches of the system. Hence, unless the walls of a preformed
_space had been ruptured by pressure, which was unlikely, we
must conclude that a bursa, if it exists, must extend out under the
endocardium and in between the cardiac musculature. Micro-
scopic examination of stained sections also showed the diffuse
character of the injection mass typical of extravasation. In the
case of beef heart no. 12, however, a different result was appar-
ently obtained. In this specimen, the attempt had been made to
inject the ‘main bursa,’ but accidentally the sheath of the bundle
had been pierced and the injection solution delivered within
it. On dissecting the main bundle from the ‘Knoten’ to the divi-
sion into the septal branches, it was seen that, the injection mass
was confined entirely within the sheath except at the point of
puncture, where some fluid had extravasated. This result sug-
gested the idea that it might be possible constantly to inject:the
sheath, and by using air in the syringe it was also found that the
sheath could be inflated. If, for example, air was injected from
the right side into the sheath of the right branch at the base of
the moderator band or trabecula supraventricularis, it ran up

THE SHEATH OF THE BUNDLE OF HIS 61

towards the main branch and, in some instances, at once appeared
on the left side within the sheath of the left branch. Following
out this same idea, a complete injection of the system was
obtained from its point of departure from Tawara’s ‘Knoten’
to the very fine ramifying branches which lie under the endocar-
dium, and even to the finest terminal fasciculi which end in
the myocardium. The chief difficulty was to insert the needle
at just the proper depth to pierce the sheath. The point selected
for the purpose of injection was immaterial as long as the fascicu-
lus chosen was somewhat thicker than the needle.

Contrary to Dogiel’s view, the endocardium is always trans-
parent enough to permit accurate localization of the fasciculi.
These are readily seen as they pass along under the endocardium,
and after some trials few mistakes were made in inserting the
needle. .

Twenty-five hearts were used for the injection of the sheath
in this manner, eighteen of these were fresh beef hearts, and one
a lamb’s heart, the rest being formalin-preserved hearts of cattle.
In all of these a practically uniform result was obtained. The
sheath of the atrio-ventricular bundle could be filled with the
colored fluids without difficulty. The lamb’s heart gave an
especially delicate and fine picture of the injection. However,
as a rule, it was difficult to force the fluid upward to the main
bundle and farther upward to the region of the ‘ Knoten,’ although
this was done in some eases.

Fig. 1 is a photograph of beef heart no. 36 of this series, and
although the figure is a retouched photograph it does not repre-
sent adequately the extent nor the completeness of the injection.
This heart was opened by an incision which extended up from the
apex parallel to, and about three centimeters from the anterior
longitudinal suleus. The view shows the left ventricle with its
outer wall reflected to the left. The injection mass used was a
suspension of India-ink. Points of injection were at 1, where
the left branch appears superficially and begins its downward
course, and at 2, where some of the ramifying divisions of the
left branch have come to lie upon the posterior papillary
muscle. The fluid, injected at 7, ran down within the sheath of

62 RUSKIN M. LHAMON

the left branch to the region where this divides at about the junc-
tion of the upper and middle thirds of the septum. From here,
three main paths lead off; an anterior which traversed a false
tendon to reach a papillary muscle, a middle which passed down
to branch out upon the septal wall as far as the apex, and a pos-
terior which crossed a false tendon to reach the base of the pos-

Fig. 1 Photograph of ox heart, retouched, showing injected sheath of atrio-
ventricular bundle. One-half actual size. 1 and 2, points from which entire left
side of sheath was injected; 3 and 4, a cut false-tendon, connecting the septal
with the outer wall of the chamber; 5, marks the region above apex, where a very
fine anastomosis is not well represented in the photograph.

terior papillary muscle. On the septal wall, the whole sheath
was injected from the one point 7. In the same way the system
of branches on the outer wall of the ventricle was filled from
point 2. The false tendon connecting points 3 and 4 was after-
ward cut to facilitate photographing the specimen. At 5 two
tendon threads are shown which carry Purkinje fasciculi, but the
sheaths of these were not completely injected.

THE SHEATH OF THE BUNDLE OF HIS 63

A detailed description of the coursing of the finer branches as
shown in fig. 1 is unnecessary, but a short summary seems Justi-
fied. On the septal wall the fibers unite, branch and reunite
to form a network, the meshes of which constantly grow smaller
until in the apex of the heart a mantle of fine anastomosing fibers
occurs. This lies at a region under 5 and is not well shown in the
figure. Upon the outer wall of the ventricle, the injected mesh-
work extends from the apex almost to the attachment of the leaf-
let of the mitral valve. Here, the fasciculi appear relatively
smaller, unite at more frequent intervals than those upon the
septal wall, and form a finer network. Moreover, the points of
union are more characteristically node-like. Often, fine fascicull,
unnoticed before injection, became apparent and could be seen
to branch and to unite with others until they were finally lost to
view because the sheath was uninjected.

With the sheath system in the right ventricle similar results were
obtained. ‘The fluid ran down in the sheath of the right branch
and passed to the outer wall of the chamber by way of the mod-
erator band, where anastomoses occurred between the branches,
as in case of the left ventricle. From the outer wall other
strands travelled back to form a septal network. The right divi-
sion, however, appeared somewhat harder to inject than the other,
possibly because of a closer attachment of the sheath. Upwards,
at the beginning of the interventricular septum, it was found that
the injection had not extended to include the main branch. This,
also, may have resulted from a too firm union between sheath and
bundle, or perhaps from the use of too little pressure. Dissec-
tion of beef heart no. 37, the next in the series, showed that the
sheath of the main bundle had been filled with fluid as far as the
interauricular region—the ‘ Knoten’—and other specimens showed
the same condition.

In the heart of the young lamb only the left side of the system
was injected. Here, too, a very fine and delicate system of anas-
tomosing fasciculi, similar in all respects to those in the larger
beef hearts, was brought into view. With the formalin-preserved
hearts of beef and calf, a like result was obtained but the extent to
which the fluid would penetrate within the sheaths of these was

64 RUSKIN M. LHAMON

naturally limited because of the hardening and loss of elasticity
of the tissues.

That the sheath does not simulate a bursa in containing fluid
was well shown by injecting a fasciculus from two points, when the
fluids would run together perfectly without hindrance from any
other contained fluid. Attempts to demonstrate a lining by
means of a silver stain were also unsuccessful nor was it possible
to demonstrate microscopically the existence of definitely ar-
ranged nucleii which the connective tissue cells of the inner layer
of a bursa should show.

un Gna tan ae endocardium

i “onsen += —- myocardium

------- == anastomosing fascicull
eons under endocardium

~----~------- injected fasciculi
---°" penetrating myocardium

QU pericardium

/
/

Fig. 2. Ox heart; cross-section; diagram of left ventricular wall, acual size,
showing depths within myocardium at which the injected branches of the sino-
ventricular system could be traced with the naked eye.

Tawara, who traced the Purkinje fasciculi and their sheaths
through the myocardium to their terminations or transitions into
heart muscle and perimysium respectively, states that the sheaths
everywhere form a closed system and isolate completely the fas-
ciculi from the heart muscle. From an examination of my speci-
mens these statements are wholly confirmed, for even with the
naked eye, it was possible to find injection masses in some in-
stances as far as fifteen millimeters in the myocardium. Fig. 2
is a diagram of part of the left ventricular wall of beef heart no.
26, and-illustrates this fact. Furthermore, it was possible, by

THE SHEATH OF THE BUNDLE OF HIS 65

securing serial sections of pieces of myocardium containing parts
of the injected system, to trace on the one hand, the sheath with
the enclosed injection mass as far upward in the region of the
‘Knoten’ as the connective tissue of the auricular muscle, and on
the other as far into the ventricular wall as the point of termina-
tion of the Purkinje fibers, where the sheath becomes the perimy-
sium.

=
= ee en
moe ESS
== == = —— —— B
Sye=se 6
ee S
ea TD
~\ SRR
WK MOY
SSK CSc (63
Te = SN:
SS SS se-£
NSS
Wa
WS

Fig. 3 Section showing two Purkinje fasciculi under the endocardium, A,
endocardium; B, injection mass within connective tissue sheath; C, Purkinje
fasciculus; D, connective tissues; 2, cardiac muscle. X 275.

For this microscopical work, pieces of tissue including parts of
the injected system were taken from the different regions, 1.e.,
from the region of the ‘Knoten,’ from that of the main bundle,
of the main branches, from the sub-endocardial anastomosis, and
from the terminations within the myocardium. Paraffin and
celloidin were employed for embedding and the sections were
stained with haematoxylin-eosin and van Gieson. Where the
fasciculus had appeared as a definite strand in the gross, the micro-

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 1

66 RUSKIN M. LHAMON

scopic examination showed that the injection mass was entirely
held within the connective tissue envelope. At points, this
sheath had been dissected and separated around the entire cir-
cumference of the fasciculus. At other points the attachments of
the sheath were still partly intact and limited the injection mate-
rial to a part of the circumference. Where the strands anasto-
mose, or where a side branch is given off, as Tawara has stated,
the connective tissue follows each part, giving it a complete
investment. This was well shown by studying serial sections of a
part of the bundle where numbers of strands go to form one large
branch. The India-ink, or the Prussian-blue mass made the
tracing of the sheath easy, for by looking backward and forward
in the series the anastomosis of the near-by fasciculi enveloped by
a continuous connective tissue covering was marked by the con-
tinuity of the enclosed black or blue mass.

Fig. 4 shows a camera lucida drawing of a section of tissue from
beef heart no. 26, the section being cut somewhat obliquely to
the long axis of the cardiac muscle cells. The figure shows the
transition point of a Purkinje fiber into the heart muscle, the
fiber showing some of the India-ink injection mass within its
sheath. In the series from which this section was taken the fiber
A is found to be a branch from B, the sheaths and the enclosed
injection material of the two strands, therefore, are continuous.
The fasciculus B, traced through the cardiac tissue for a relatively
great distance, gives off the short fiber A which passes at once to
the wedge-shaped fasciculus of cardiac muscle cells /, and becomes
lost in the latter at the apex. As shown in the figure, the injec-
tion material had travelled along in the sheath-like covering of the
fiber A to the point where this change takes place and the sheath
is seen to be directly continuous with the perimysium of the heart
muscle.

This mode of transition of a fiber into a group of muscle cells,
is apparently one of the ways in which the sino-ventricular sys-
tem may terminate within the myocardium of the ventricles.
Tawara has discussed other forms. While an inquiry into these
terminal mechanisms and the histology of the same, has not been
the special object of this investigation, nevertheless, in tracing
out the ramifications of injections, into the sheath, the above

THE SHEATH OF THE BUNDLE OF HIS 67

transition was observed. Such a transition could, perhaps, be
considered a means to offset the great numerical disproportion
which must exist between the fibers of the atrio-ventricular system
and the heart muscle fibers.

Fig. 5 shows a camera lucida drawing from the auricular region
—the ‘Knoten’—in a section of tissue from beef heart no. 12.

Fig. 4 Camera lucida drawing; section of ventricular myocardium, beef heart.
A and B, Purkinje fascicul; C, injection material within connective tissue sheath;
D, cardiac muscle fibers; #, fasciculus of heart muscle fibers in which fiber A ter-
minates. XX 275.

Point A shows the auricular muscle fibers cut in cross section.
From these, bundles of fibrillae pass off to enter a point of union
for many such fibril fasciculi, which place Tawara has called a
node. In this nodal point the fibrils are seen as confused strands
coursing in many different directions. The clear spaces G repre-
sent the indentations of the node which have been cut in the sec-
tioning, and within which the connective tissue nuclei can be

68 RUSKIN M. LHAMON

Fig. 5 Cameralucida drawing; section from region of the ‘Knoten,’ showing a
large nodal point with connections. A, cardiac muscle fibers (auricular); B,
fibril strands from muscle fibers to the node; C, cross striations of fibril bundles
within the node; D, connective tissue; E, injection mass between connective tissue
and node; F, spaces produced by injection mass; G, indentation in node, showing
connective tissue within.

seen. Cross striations of the fibril strands are plainly seen at
many points C. Surrounding these structures is the connective
tissue D which forms the sheath of the atrio-ventricular system
in this region, and which encloses the injection mass / within
it. Since uninjected specimens, do not show such a great separa-
tion between sheath and system, it seems probable that the
large spaces F were made by the injection and later fixation and
dehydration of the tissues kept them in the distended position
while the injection material settled to a small mass.

There exists, then, in beef, calf and sheep hearts, a complete
connective tissue envelope applied to the sino-ventricular sys-
tem in all its ramifications. It clothes the system in its entirety
beginning above at the point where the main bundle takes its
origin and continues downward along the branching ventricular
fasciculi to their terminations. In the sub-endocardial region,
it is in relation with the sub-endocardial connective tissue,
while within the myocardium it is related to the interstitial myo-

THE SHEATH OF THE BUNDLE OF HIS 69

cardial connective tissue. It is remarkable that the sheath can
be so easily dissected loose from the bundle which it covers and
that it can nevertheless withstand the pressure necessary to inject
it so extensively. Although a rapid diminution of pressure
undoubtedly occurs as the fasciculi branch and anastomose,
the pressure at the point of injection is relatively great; never-
theless the sheath withstands it, if ordinary care be used. Fur-
thermore, in view of the reduction of pressure which must result
from the progressive division and anastomosis of the fasciculi the
extent to which the sheath can be injected is very great.

The question might well be asked whether or not this system,
which can be injected so easily, is in relation with the lymphatic
system. A consideration of the origin and the termination of the
sheath will furnish the answer. It has been shown that the tissue
forming the sheath is continuous, on the one hand, with the peri-
mysium of the auricular muscle, on the other with the perimy-
sium of the ventricular fibers. Consequently, if we look upon the
sheath as composing the walls of a lymph space we must neces-
sarily look upon the perimysium in a similar way, and accept the
fact that this lymph space extends out around the fasciculus of
heart muscle cells and around the single cells. Such a concep-
tion would of course seem untenable.

The functions of the sheath lie outside the scope of this paper,
excepting so far as the idea of a lubricating mechanism is con-
cerned. If, by a bursa, we understand a structure distinct from
the sheath, the existence of such a bursal space was not confirmed
by this investigation. On the other hand, if the inflated connec-
tive tissue sheath was mistaken for a bursa, the interpretation of
its function as a lubricating apparatus seems impossible. Other
writers, notably Tawara, Keith, Fahr, and Ménckeberg, have
spoken of its function as one of isolation and insulation. How-
ever, Dogiel, whose opinion is worthy of great consideration, has
recently stated that the existence of such a sheath has never been
satisfactorily proven, and hence feels under no necessity to ascribe
to it a function. The writer offers the facts herein reported as
further proof of the indisputable existence of the sheath as a very
definite structure which isolates the atrio-ventricular system from
the rest of the heart, even to its terminations.

70 RUSKIN M. LHAMON

CONCLUSIONS

1. In hearts of beef, calf and sheep there is a definite connective
tissue sheath surrounding the sino-ventricular bundle, which com-
pletely invests and isolates that bundle from the heart muscle.

2. The existence of this sheath can be demonstrated in its
entirety by injections, for the sheath is capable of maintaining
its integrity against considerable pressure from within.

3. This sheath begins above where the fasciculi of the bundle
are continuous with the musculature of the interauricular septum
(Tawara’s ‘Knoten’) and clothes the system up to its ventricular
termination, i.e., to its transition into cardiac muscle fibers.

4. The sheath does not simulate a bursa, save very remotely,
perhaps, nor is it part of the lymphatic system of the heart.

5. Hence this sheath is not a bursa and no synovial bursa
exists, for if a bursa be present in any portion, these injections
show that it is coextensive with the bundle itself.

BIBLIOGRAPHY

Curran, E. J. 1909 A constant bursa in relation with the bundle of His; with
studies of the auricular connections of the bundle. Anat. Rec., vol.
3, no. 12.

Der Wirt, Lyp1a M. 1909 Observations on the sino-ventricular connecting sys-
tem of the mammalian heart. Anat. Rec., vol. 3, no. 9

DoargEt, J. 1910 Die Bedingungen der automatischen-rhythmischen Herzkon-
traktion. Pfliigers Archiv., vol. 135.

Faur 1907 Uber die Moskulése Verbindung zwischen Vorhof und Ventrikel
(das Hissche Biindel) im normalen Herzen und beim Adams—Stokes
Symptom-komplex. Virch, Archi. f. Path. Anat., Bd. 188.

1908 Zur Frage der Atrioventrikuliren Muskelverbindung in Herzen.
Verh. deutsch pathol Ges. 12.

KeritH AND Fiack 1907 Form and nature of the muscular connections between
the primary divisions of the vertebrate heart. Jour. Anat. and Phys.
London, vol. 41.

Kent, STANLEY 1893 ° Researches on the structure and function of the mamma-
lian heart. Journal of Physiology, vol. 14.

Moncxesera, J. G. 1908a Uber den sogennanten abnormen Sehnenfiiden in
linken Ventrikel des menschlichen Herzens und ihre Beziehungen zum
Atrio-ventrikulirbiindel. Verh. deutsch pathol. Ges. 12.
1908b Untersuchungen iiber das Atrioventrikulir-biindel im mensch-
lichen Herzen. Gustav Fischer. Jena.

ReEtTzER, Ropert 1908 Some results of recent investigation on the mammalian
heart. Anat. Rec., vol. 2, no. 4. ;

Tawara, §S. 1906 Das Reizleitimgssystem des Siingetierherzens. Gustav
Fischer. Jena.

THE DEVELOPMENT OF THE ADRENALS IN THE
TURTLE

ALBERT KUNTZ

From the Laboratories of Animal Biology of the State University of Iowa

NINE FIGURES

CONTENTS
MNO CUI HOT Sarre ae ci ons oo costes eas) ORT IE OR ee oeseens Lene tata pest ones? 71
(USE VALLONS AAG Se seed cole date oe ER EN ee ieee tout Mastek 76
Bick igade viel opirientt. ©5275 das .'.00 < «<5 % Storey Rm ae Sk achat a Sr ah Pee ae 76
ConbicalisubsGameecnras.ccn.c8 ws 5618 Eee casein e tere Boe ne 76
Chromaiimisulbstamces. 4.4.05: 0.1509 oreo Oa aiAraciae oe a 79
HhebeTACeEVCLOPINEM Ui Uesnis ce seh cio bcd 42 kee Oe aor als Shy siateesicbipoess & 81
Genetic relationship and differentiation of chromaffin cells.................. 85
SORITIILD A Soe ORe Mice, Foe nO EE EEE TRC Ce Daa a ihre ereries mae Roee 87
ESky Lito erarseayo lanypec seceeete ee g ne Bice be scan, 3.15 ea Aa RO Rata ah ee Ue ea 88
INTRODUCTION

A review of the literature bearing on the development of the
adrenals in the several classes of vertebrates reveals several oppos-
ing theories of development each of which has been supported
by investigators of recognized ability. The adrenal system is
a compound system. In the lower vertebrates it includes (accord-
ing to the nomenclature proposed by Balfour) the interrenals
and the suprarenals. In the higher vertebrates these two sys-
tems are consolidated more or less completely into a single pair
of adrenal organs composed of the cortical substance which cor-
responds to the interrenals of the lower vertebrates, and the med-
ullary, or chromaffin,! substance which corresponds to the supra-
renals. These organs have been described by some investigators

1 The term, chromaffin substance, will be used in this paper in preference to the
term, medullary substance. A true medulla occurs onlyin the adrenals of mam-
mals.

71

i2 ALBERT KUNTZ

as having a homogeneous origin; i.e., both the cortical and the
chromaffin substance arising from the same source, and by others
as having a heterogeneous origin; i.e., the cortical and the chro-
maffin substance arising from separate sources. The advocates
of the theory of the homogeneous origin of the adrenals have in
turn derived them from the mesenchyme, the peritoneal epithe-
lium, the germinal epithelium, the epithelium of the pronephros,
the epithelium of the mesonephros and from the peripheral part of
the sympathetic nervous system. The advocates of the theory of
the heterogeneous origin of the adrenals have in turn derived the
cortical substance from the mesenchyme, the peritoneal epithe-
lium, the epithelium of the pronephros and the epithelium of the
mesonephros, while they have in turn derived the chromaffin sub-
stance from the mesenchyme and from the sympathetic nervous
system.

A comprehensive review of the literature bearing on the devel-
opment of the adrenals would be foreign to the purpose of this
paper. An attempt will be made only to indicate the trend of the
progress of our knowledge in this field by brief reference to some
of the more important papers. For a more comprehensive review
of the literature and for a complete bibliography, the reader is
referred to the work of Poll? (06).

The most important investigations on the adrenal system be-
fore the classical work of Balfour are those of Leydig (’53). This
investigator described both the interrenals and the suprarenals
in the fishes and concluded that the latter are derived from the
sympathetic nervous system. Indeed, he says, “‘ As the pituitary
body is an integral part of the brain, so are the suprarenal bodies
part of the sympathetic system.”

The work of Balfour (78) shows conclusively that in the elas-
mobranch fishes the interrenals are of mesodermal origin, while
the suprarenals are derived from the sympathetic ganglia located
along the abdominal aorta. The findings of Balfour were cor-
roborated by the work of some of his pupils, notably that of
Mitsukuri (’82).

2 Hertwig’s Handbuch, 1906, vol. 3, part 1, chap. 2, pp. 443-616.

DEVELOPMENT OF THE ADRENALS IN THE TURTLE 73

Although the theory that the suprarenals, or the chromaffin
substance, are derived from the sympathetic nervous system was
so well supported by the work of these early investigators, later
observers have repeatedly failed to recognize the nervous origin
of these bodies and have attempted to derive them from some
other source. This condition is probably due in part to a failure
to recognize the homology of the chromaffin substance of the
adrenals of the higher vertebrates with the suprarenals of the
fishes and in part to an erroneous interpretation of observations
coupled with an unwillingness to believe that any part of the
adrenals is derived from the sympathetic nervous system.

Gottschau (’83), working with mammalian embryos, noted the
intimate association of the chromaffin substance with the corti-
cal substance and concluded that both arise from the mesenchyme.
Janosik (’83) observed that in mammalian embryos cells advance
from the peritoneal epithelium into the mesenchyme in the region
in which the adrenals arise. He concluded, therefore, that these
bodies are derived from the germinal epithelium. Like Gottschau
he believed that the cortical and the chromaffin substance are
derived from the same source. O Schultze (97) concluded, from
observations made on embryos of Vespertillio murinus, that the
entireadrenal anlage is derived from sympathetic ganglia and
that it is later differentiated into a cortical and a medullary
portion. Minot (’97) found no evidence of the sympathetic
origin of any part of the adrenals in human embryos, but sup-
ported the older view of Gottschau that both the cortical and the
chromaffin substance are derived from the mesenchyme.

The above citations are sufficient to indicate that there has
been no general agreement in the conclusions to which the earlier
investigators were led. A similar lack of agreement, though less
marked, is also prevalent among the later investigators.

Flint (00), in his paper on the blood vessels of the adrenals,
does not attempt to determine the ultimate source of either the
cortical or the chromaffin substance. He shows, however, that
in embryos of the pig the cortical substance arises first and that
the chromaffin substance arises from cells which wander in from
the outside. That the cells which give rise to the chromaffin

74 ALBERT KUNTZ

substance are derived from the sympathetic nervous system, in
his opinion, requires further proof.

Aichel (’00) advanced a new theory of the origin of the adrenals.
According to this author, the interrenals alone in the selachians
are homologous with the adrenals of the higher vertebrates.
These, he believes, arise from the peritoneal invaginations of the
pronephros. The so-called suprarenals in the selachians, he con-
cludes, are derived from retrograding canals of the same body.
In embryos of both the rabbit and the mole, according to Aichel,
both the cortical and the chromaffin substance are derived from
the epithelium of the pronephros. This view finds no support
in the work of other investigators.

Wiesel (’01), although his observations were made on embryos
of the pig which were too far advanced to reveal the earliest traces
of the adrenal anlagen, concluded that the cortical substance is
derived from the peritoneal epithelium, while the chromaffin sub-
stance is derived from the prevertebral sympathetic plexuses.

Whitehead (’03) studied the development of the adrenals in
embryos of the pig from their earliest anlagen. His conclusions
agree in general with the conclusions of Wiesel above cited.

By an exhaustive review of the literature and by extensive ob-
servations, Poll (06) has shown that the weight of evidence is
in favor of the view that in all the classes of vertebrates the cor-
tical substance of the adrenals arises from the peritoneal epithe-
lium, while the chromaffin substance is derived from cells which
become separated from the anlagen of the sympathetic nervous
system. The genetic relationships of the cells which give rise
to the chromaffin substance, however, as well as the processes
by which these cells become associated with and approximated to
the cortical substance and by which they become transformed into
typical chromaffin cells have not been understood.

During my studies of .the development of the sympathetic
nervous system in embryos of the Loggerhead turtle (Thalas-
sochelys caretta), my attention was attracted by the phenomena
involved in the development of the adrenals. Embryos of this
species afford excellent material for the study of these phenomena
because the chromaffin material is comparatively abundant, the

DEVELOPMENT OF THE ADRENALS IN THE TURTLE 19

embryos are comparatively large, develop comparatively slowly
and afford excellent histological preparations.

The development of the adrenals has been investigated less
extensively in the Reptilia than in the other classes of vertebrates.
Our limited knowledge of the development of these organsinthis
class of vertebrates is the more appreciable because the adrenals
in this, the most primitive class of the Amniota represent a tran-
sition stage in the evolutionof the highly specialized adrenal organs
in the higher vertebrates from the more primitive adrenal system
in the Anamnia.

Observations which have been recorded on the development of
the adrenals in the Crocodilia and the Ophidia are only frag-
mentary and inconclusive. More or less extensive observations
on the development of these organs in the Sauria, primarily in
certain species of the genus Lacerta, have been recorded by
Braun (’79, ’82), von Mihalcovies (’85), Weldon (’85), Hoffmann
(89) and Soulie (03). von Mihalcovies, being an advocate of
the theory of the homogeneous origin of the adrenals, derived
both the cortical and the chromaffin substance from the germinal
epithelium. The other investigators above mentioned agree in
deriving the chromaffin substance from the anlagen of the pre-
vertebral sympathetic plexuses. They do not agree, however,
as to the origin of the cortical substance. According to Braun,
the cortical substance arises directly from the mesenchyme. <Ac-
cording to Weldon and Hoffmann, it arises from the epithelium
of the pronephros. According to Soulie, it arises from the per-
itoneal epithelium.

The only extended observations on the development of the
adrenals in the Chelonia which have been recorded, as far as the
writer is aware, are those of Poll (’04, 706) on embryos of Emys
europaea. This author describes the origin of the cortical sub-
stance from the peritoneal epithelium in detail. He has little
to say, however, concerning the chromaffin substance except
that it is derived from the anlagen of the prevertebral sympathetic
plexuses.

The following observations are based almost exclusively on
embryos of the Loggerhead turtle (Thalassochelys caretta).

76 ALBERT KUNTZ

I take pleasure in expressing my indebtedness to Professor F. A.
Stromsten for the use of a large number of embryos of this species
which were collected by him at the Dry Tortugas, Florida, during
the summer of 1910. It is a real pleasure also to express my deep
sense of obligation to Professor G. L. Houser for helpful sugges-
tions during the progress of this investigation and for reading the
manuscript.

OBSERVATIONS
Early development

Cortical substance. The anlagen of the adrenals arise in
embryos of Thalassochelys caretta during the eighth day of incu-
bation as buds of cells which proliferate from the peritoneal
epithelhum just laterad to the root of the mesentery and approx-
imately at the middle level of the mesonephros (fig. 1, ad). At
the close of the eighth day of incubation a short series of buds
may be observed on either side of the aorta rising into the mesen-
chyme between the aorta and the mesonephros. These buds
are at first more or less wedge-shaped with a broad base resting
on the peritoneal epithelium from which they arise (fig. 1,ad).
There is no well marked differentiation apparent at this stage
between the cells composing the adrenal buds and the cells of the
adjacent mesenchyme.’ The nuclei of the former often appear
somewhat clearer than the nuclei of the latter and their cytoplasm
stains somewhat more intensely. These buds may be readily
recognized, however, by the relatively compact and more or less
regular arrangement of the cells composing them. The exact
number of buds taking part in the development of the cortical
substance of each adrenal organ is not easily determined. These
buds do not all arise simultaneously. During the ninth and the
tenth day of incubation, after the earliest buds have become nearly
or completely separated from the peritoneal epithelium, buds
may still be observed in the anterior region of the zone of prolifera-
tion which exhibit the earliest phases of their development from
the peritoneal epithelium.

DEVELOPMENT OF THE ADRENALS IN THE TURTLE Th

At the close of the eighth day of incubation, the buds in the
middle region of the adrenal zone rise in the mesenchyme as high
as the middle level of the aorta and in sections occupy the greater
part of the area between the aorta and the mesonephros (fig. 2,
ad). The cells in these buds are now arranged in more or less
regular rows rising from the base of the wedge-shaped aggregate
toward its apex which is usually slightly recurved toward the
mesial surface of the mesonephros. Mitotic figures may occasion-
ally be observed in these growing buds. It is obvious, therefore,
that these cells retain the power of further propagation by division
after they have advanced into the adrenal anlagen, thus giving
the latter the capacity of independent growth after they are no
longer connected with the peritoneal epithelium.

As development advances, the adrenal buds become larger and
advance farther into the mesenchyme, rarely, however, rising but
slightly higher than the middle level of the aorta. They no
longer appear wedge-shaped in transverse sections, but have be-
come rounded or oval in outline, retaining connection with the
peritoneal epithelium only by a slender stalk (fig. 3, ad). During
the tenth day of incubation, some of the buds become almost or
completely separated from the peritoneal epithelium. They now
appear in transverse sections as condensations in the mesenchyme
between the aorta and the mesonephros and are often so closely
approximated to the mesial surface of the latter that except for
a slight difference in the staining properties of the cells compos-
ing them it becomes difficult to distinguish between these two
complexes. Such conditions, doubtless, are responsible for the
conclusions of some of the earlier investigators that the cortical
substance of the adrenals is derived either directly from the
mesenchyme or from the epithelium of the mesonephros.

At the close of the eleventh day of incubation, the adrenal an-
lagen have become separated from the peritoneal epithelium
throughout nearly the entire extent of the adrenal zone. In
transverse sections, they appear more or less circular in outline,
being modified in form somewhat by the limitations of the area
between the aorta and the mesonephros (fig. 4, ad). In the ante-
rior and the middle region of the adrenal zone, the adrenal anlagen

78 ALBERT KUNTZ

to NES
4

Fig. 1 Transverse section through the anterior region of the adrenal zone in
an eight-day embryo of Thalassochelys caretta. XX 150. ad, adrenal buds; ao,
aorta; m, mesonephros; mes, mesentery; sy, anlagen of sympathetic trunks.

Fig. 2 Transverse section through the middle region of the adrenal zone in an
eight-day embryo of Thalassochelys caretta. X 150. ad, adrenal buds; ao,
aorta; m, mesonephros; mes, mesentery; sy, anlagen of sympathetic trunks.

Fig. 3 Transverse section through the middle region of the adrenal zone in
a nine-day embryo of Thalassochelyscaretta. > 150. ad, adrenal buds; ao, aorta;
m, mesentery; sy, anlage of sympathetic trunk.

Fig. 4 Transverse section through the adrenal zone in an eleven-day embryo:

of Thalassochelys caretta. X 150. ad, adrenal anlagen; ao, aorta; mes, mesen-
tery; sy, anlagen of sympathetic trunks; syc, sympathetic cells.

DEVELOPMENT OF THE ADRENALS IN THE TURTLE 79

are now located below the middle level of the aorta, while in the
posterior region they lie close to the ventro-lateral aspects of the
latter and are not far removed from the peritoneal epithelium in
the angle between the mesentery and the mesonephros. The
primary arrangement of the cells in irregular rows is no longer
apparent. At this stage, however, the cells seem to be arranged
in more or less regular concentric layers.

At the close of the thirteenth day of incubation, the aorta has
become relatively smaller than in the preceding stages and the
internal organs are located farther ventrally with respect to the
former. In the anterior region the adrenal anlagen now lie along
the ventro-lateral aspects of the aorta, while farther posteriorly
they lie distinctly below its ventral level (fig. 5, ad). In the pos-
terior region the anlagen of the adrenals have become somewhat
compressed dorso-ventrally and appear more or less oval in trans-
verse sections. ‘The mesial surfaces of the right and the left
adrenal have now become closely approximated to each other in
this region. Denser aggregates of cells may now be observed
scattered irregularly throughout the section of the adrenal anlage.
This condition, doubtless, represents the initial stage in the dif-
ferentiation leading to the characteristic arrangement of the cells
in the cortical substance of the adrenal gland.

After the close of the thirteenth day of incubation, the adrenal
anlagen become somewhat farther removed from the aorta. They
increase in size materially and the cells assume a more definite
arrangement. At the close of the thirteenth day, the denser
aggregates of cells noted above have become more conspicuous
and are variously connected with each other by strands of closely
aggregated cells (fig. 6, ad). The meshes within this complex
stain less intensely than the denser areas and contain relatively
few cells.

The observations thus far recorded agree in all essentials with
the observations of Poll (’04, 706) on the development of the
cortical substance of the adrenals in Emys europaea.

Chromaffin substance. The above observations pertain only
to those parts of the adrenal anlagen which arise from the peri-:
toneal epithelium and give rise to what is commonly known as

SO ALBERT KUNTZ

the cortical substance. The chromaffin substance arises in an
entirely different manner. As has been repeatedly maintained
by certain investigators and denied by others, for the several
classes of vertebrates, and as has been shown by Poll also for the
turtle (Emys europaea), this part of the adrenal system is derived
from the anlagen of the prevertebral sympathetic plexuses. It
is, therefore, primarily of ectodermal origin.

During the eighth day of incubation, when the adrenal buds
appear as small cell-ageregates arising from the peritoneal epi-
thelium, the anlagen of the ganglia of the sympathetic trunks are
already present as irregular cell-groups lying in the mesenchyme
approximately at the dorsal level of the aorta (fig. 1, sy). As
development advances, the anlagen of the sympathetic trunks
become larger and, as the writer has shown in an earlier paper
(11 a), cells become separated from them and migrating ventrally
give rise to the anlagen of the prevertebral sympathetic plexuses
which are located along the ventro-lateral aspects of the aorta.

During the eleventh day of incubation, sympathetic elements
may be traced ventrally from the anlagen of the sympathetic
trunks along the lateral surfaces of the aorta. In the anterior
and the middle region of the adrenal zone where the adrenal an-
lage is still located between the mesonephros and the aorta, sym-
pathetic cells may in some sections be observed in contact with
the anlage of the cortical substance, but no such elements could
be observed, at this stage, among the cortical cells (fig. 4, syc).

At the close of the thirteenth day of incubation, sympathetic
cells may be observed as far ventrally as the ventral level of the
aorta and the anlagen of the prevertebral sympathetic plexuses
have become well established. The adrenal anlagen now lie in
close proximity with the anlagen of the prevertebral sympathetic
plexuses and in some sections small groups of sympathetic cells
may be observed in contact with the cortical mass (fig. 5).

As development advances, the cells composing the anlagen of
the prevertebral sympathetic plexuses become more numerous
until at the close of the nineteenth day of incubation the aorta is
completely encircled ventrally by conspicuous aggregates of sym-
pathetic cells (fig. 6, pv). The adrenal anlagen are now somewhat

DEVELOPMENT OF THE ADRENALS IN THE TURTLE 81

farther removed from the aorta and numerous small cell-groups
may be traced from the larger cell-aggregates constituting the
anlagen of the prevertebral sympathetic plexuses toward the ad-
renal anlagen (fig. 6, cc). Many of these cell-groups may be
observed closely approximated to the dorsal and the mesial surfaces
of the adrenal anlagen, while not infrequently such cells may be
observed just beneath the surface of the adrenal anlage among the
cortical cells (fig. 6, cc’).

There is no observable difference at this stage between the cells
of sympathetic origin which advance toward the anlagen of the
adrenals and the cells which remain within the sympathetic
plexuses. That these are the cells, however, which give rise to
the chromaffin substance associated with the cortical substance
in the adrenal organs can not be doubted. These cells not only
become aggregated at the dorsal and the mesial surfaces of the
adrenal anlagen, but, as development advances, groups of these
cells may be found completely encircling the cortical substance of
the adrenals and penetrating into the spaces which appear between
the aggregates of the cortical cells as the latter approach more and
more closely the arrangement which is typical of the adrenal
glands in the mature state. Furthermore, sympathetic cells
do not advance farther ventrally into the mesentery at this level,
and, as the writer has shown in the earlier paper referred to above,
the cells which give rise to the sympathetic plexuses in the walls
of the digestive tube are derived from other sources; viz., the
hind-brain and the vagus ganglia.

Later development

The record of the later development of the adrenals is essentially
a record of the further growth of the glands, the rearrangement
of the cells, the adjustment of the chromaffin substance to the
cortical substance and the differentiation of the elements derived
from the sympathetic nervous system into typical chromaffin
cells.

During the later stages of incubation, the development of the
adrenals advances comparatively slowly. The adrenal zone be-

82 ALBERT KUNTZ

Fig. 5 Transverse section through the adrenal zone in a thirteen-day embryo
of Thalassochelys caretta. X 150. ad, adrenal anlagen; ao, aorta, pv, anlage of
prevertebral sympathetic plexus; sy, anlagen of sympathetic trunks.

Fig. 6 Transverse section through the adrenal zone in a nineteen-day embryo
of Thalassochelys caretta. > 150. ad, adrenal anlagen; ao, aorta; cc, cells des-
tined to give rise to chromaffin substance; cc’, cells of sympathetic origin among
cortical cells; pv, anlage of prevertebral plexus.

Fig. 7 Transverse section through the adrenals in a twenty-five day embryo
of Thalassochelys caretta. x 100. ad, adrenals; ao, aorta; cc, chromaffin
cells; pv, anlage of prevertebral sympathetic plexus.

Fig. 8 Section of adrenal in a thirty-six -day embryo of Thalassochelys caretta.
x 300. cc, chromaffin substance; co, cortical substance; vs, vascular space.

DEVELOPMENT OF THE ADRENALS IN THE TURTLE 83

comes comparatively shorter and the glands gradually assume the
general form and location which they maintain during adult life.
In embryos twenty-five days old, the adrenals are still closely
associated with the anlagen of the prevertebral sympathetic
plexuses and a few cells apparently continue to advance from the
latter into the masses of cells of sympathetic origin which are
becoming approximated more and more closely to the cortical
substance. In the anterior region, the right and the left adrenal
are distinct and are located approximately at the level of the
coeliac plexus which lies between them. Farther posteriorly the
right and the left adrenal lie in contact with each other and are
apparently connected by a broad bridge of cortical cells; the
principal mass of the prevertebral plexuses being embraced in the
angle between them (fig. 7, ad). Numerous aggregates of cells
of sympathetic origin lie in close contact with the cortical sub-
stance (fig. 7, cc). These aggregates are most conspicuous at the
dorsal and the mesial aspects of the adrenals, but aggregates of
considerable size may be observed surrounding the entire mass of
the cortical substance. The cells of the cortical substance are
becoming arranged into more or less distinct aggregates with
numerous vascular spaces appearing between them. Not infre-
quently the aggregates of cells of sympathetic origin penetrate
deeply into these vascular spaces, while in sections occasionally
such cell-groups occur completely surrounded by the cortical
substance.

After the twenty-fifth day of incubation, the association of the
cortical and the chromaffin substance becomes more intimate.
In embryos from thirty to thirty-six days old, the masses of cells
of sympathetic origin have become more conspicuous than in the
earlier stages and penetrate more deeply into the spaces between
the aggregates of the cortical cells (fig. 8, cc). Groups of cells of
sympathetic origin which in sections appear to be completely sur-
rounded by the cortical substance are now scattered irregularly
throughout the entire section. Such areas in the section probably
do not represent groups of cells which have become completely
separated from the masses of cells of sympathetic origin aggrega-
ted at the surface of the cortical substance, but are, doubtless,

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, No. 1

84 ALBERT. KUNTZ

sections of columns of such cells which have penetrated deeply”
into the vascular spaces between the aggregates of the cortical
cells.

In embryos forty-three days old, the adrenals have become more
intimately associated with the renal organs and are located on
the ventro-mesial surfaces of the latter in approximately the same
position which is maintained by them throughout adult life. A
larger proportion of the cells of sympathetic origin have pene-
trated into the vascular spaces between the aggregates of the
cortical cells. The distribution of the cortical and the chromaffin

Fig.9 Transverse section through adrenal in a forty-three-day embryo of
Thalassochelys caretta. > 75. cc, chromaffin substance; co, cortical substance;
mc, mesenchyme cells; vs, vascular spaces.

substance, therefore, approaches closely the distribution of these
substances in the mature gland. Fig. 9 is introduced to illustrate
the distribution of the cortical and the chromaffin substance and
their relation to each other in an embryo forty-three days old.
Neither during embryonic development nor in the adult con-
dition do the vascular spaces between the aggregates of cortical
cells, into which the chromaffin cells penetrate, show well defined
walls which are characteristic of arteries or veins, but are lined
by a single layer of flattened epithelial cells (fig. 8, vs). That these
spaces are blood vessels, however, is indicated by the presence in

DEVELOPMENT OF THE ADRENALS IN THE TURTLE 85

‘them of numerous blood corpuscles. Wherever aggregates of
chromaffin cells are found in the cortical substance they occur in
or adjacent to such spaces. It is probable, therefore, that chro-
maffin cells penetrate into the cortical substance only in these
vascular spaces.
It may be noted at this point that not all the elements which
- become differentiated into chromaffin cells ever become incorpor-
ated into the adrenal organs. During the later stages of develop-
ment in embryos of Thalassochelys caretta aggregates of these
cells may be found entirely apart from the adrenals. In Chryse-
mys marginata, in the adult condition, I have also observed aggre-
gates of chromaffin cells associated with the sympathetic ganglia
in proximity with the adrenals. Similar conditions were observed
by Poll in Emys europaea and by some of the earlier investigators
in various species of other classes of vertebrates.

GENETIC RELATIONSHIP AND DIFFERENTIATION OF CHROMAFFIN
CELLS

In a series of earlier papers,’? the writer has shown, for the
several classes of vertebrates, that the cells which become sepa-
rated from the cerebro-spinal nervous system and advance periph-
erally to give rise to the sympathetic nervous system are the
descendants of the ‘germinal’ cells (Keimzellen) of His; viz., the
indifferent’ cells and the ‘neuroblasts’ of Schaper. The vast
majority of these cells (in the lower vertebrates perhaps all of
them) advance peripherally as cells of the ‘indifferent’ type.
These cells have the capacity of becoming differentiated into
neuroblasts or into embryonic supporting cells. Some of the
‘indifferent’ cells, however, retain the capacity for further propa-
gation by division and give rise to new generations of ‘indifferent’
cells after they have become separated from the cerebro-spinal
nervous system.

As indicated in an earlier section of this paper, the cells which
become separated from the sympathetic anlagen and enter the
~ anlagen of the adrenals there to give rise to the chromaffin sub-

3 See bibliography.

S6 ALBERT KUNTZ

stance are, during the earlier stages of development, identical,
as far as may be determined by microscopic observation, with the
cells which remain in the sympathetic plexuses. We are forced
to the conclusion, therefore, that they are cells of the ‘indifferent’
type.

The question now arises; why do these cells which become asso-
ciated with the cortical cells of the adrenals, as well as certain.
other aggregates which remain apart from the cortical substance,
become differentiated into chromaffin cells while other cells en-
dowed with the same initial capacity do not? This question we
can not answer conclusively at present. It is suggestive, however,
that the differentiation of these cells into chromaffin cells takes
place comparatively late in the course of development.

The cells destined to become transformed into chromaffin cells
show very little evidence of differentiation before the thirty-sixth
day of incubation. Even at the forty-three day stage which was
the latest embryonic stage of Thalassochelys caretta at my dis-
posal and which falls within a week of the time of hatching, many
of the cells of sympathetic origin associated with the cortical
substance of the adrenals still remain apparently in their undif-
ferentiated condition. Some of the cells of sympathetic origin
in the adrenals, however, have assumed a polyhedral form which
is the typical form of the chromaffin cells. The cytoplasm of
these cells also stains somewhat more intensely than in the earlier
stages, but does not as yet present the coarsely granular appear-
ance, when stained by the iron-haematoxylin method, which is
characteristic of mature chromaffin cells.

In several earlier papers (11 a, 711 b, 711 d), evidence was pre-
sented in support of the theory that the processes involved in the
peripheral displacement of the cells giving rise to the anlagen of
the sympathetic nervous system are stimulated and controlled
by the influence of hormones which are produced in the regions
toward which the cells advance. The displacement of the cells
of sympathetic origin which give rise to the chromaffin cells from
the sympathetic plexuses into the adrenals must, doubtless, be
accounted for in a similar manner. This displacement can not
be accounted for by the mechanical processes involved in growth.

DEVELOPMENT OF THE ADRENALS IN THE TURTLE 87

These cells do not advance toward the adrenals in continuous
cell-columns, nor do they advance along a single path. Appar-
ently they advance in small aggregates along various paths
through a relatively compact mesenchyme. The processes here
involved are probably initiated and controlled by the influence of
hormones which are produced by the cells of the cortical sub-
stance. The differentiation of ‘indifferent’ cells into chromaffin
cells in the adrenals occurs after they have been associated with
the cortical substance for a considerable interval. By this time
the cortical cells have probably already assumed a secretory func-
tion. It is probable, therefore, that the differentiation of ‘in-
different’ cells into chromaffin cells is also stimulated by the influ-
ence of hormones which are produced by the cortical cells. The
masses of chromaffin cells which remain apart from the adrenals
probably arise from aggregates of ‘indifferent’ cells whose differ-
entiation into neuroblasts or into supporting elements was delayed,
but which were located sufficiently near to the adrenals to fall
within the sphere of influence of the hormones produced by the
cortical cells.

SUMMARY

1. The anlagen which give rise to the cortical substance of
the adrenals arise in embryos of the Loggerhead turtle (Thalas-
sochelys caretta) as buds of cells which proliferate from the peri-
toneal epithelium. The chromaffin substance is derived from the
anlagen of the prevertebral sympathetic plexuses. These findings
agree with the findings of Pollin embryos of Emys europaea and
with those of some of the earlier investigators in other types of
vertebrates.

2. The cells which become differentiated into the chromaffin
cells, like the majority of the cells which advance peripherally
from the cerebro-spinal nervous system into the anlagen of the
sympathetic nervous system, are cells of an indifferent type;
viz., the ‘indifferent’ cells of Schaper.

3. The processes involved in the displacement of the ‘indiffer-
ent’ cells from the anlagen of the prevertebral sympathetic plex-
uses toward the adrenals and the differentiation of these cells into

88 ALBERT KUNTZ

chromaffin cells are probably stimulated and controlled by the
influence of hormones which are produced by the cells of the cor-
tical substance.

BIBLIOGRAPHY

(The following list contains only those papers to which reference is made in this
paper)

AIcHEL, O. 1900 Vergleichende Entwickelungsgeschichte und Stammesge-
schichte der Nebennieren. Arch. mikr. Anat., vol. 56, pp. 1-80.

Baurour, F. M. 1878 Development of elasmobranch fishes. London, pp. 237-
247.

Braun, M. 1879 Ueber Bau und Entwickelung der Nebenniere bei Reptilien.
Zool. Anz., Jahrg. 2, pp. 238-239.

1882 Bau und Entwickelung der Nebennieren bei Reptilien. Arb. a.
d. zool-zoot. Inst., Wurzburg, vol. 5, pp. 1-30.

Fut, J. M. 1900 The blood-vessels, angiogenesis, organogenesis, reticulum,
and histology of the adrenal. The Johns Hopkins Hospital Reports,
vol. 9, pp. 153-230.

Gorrscuau, M. 1883 Structur und embryonale Entwickelung der Nebennieren
bei Séugetieren. Arch. Anat. u. Phys. Anat. Abt., Jahrg., pp. 412-448.

HorrmMann, C. K. 1889 Zur Entwickelungsgeschichte der Urogenitalorgane
beiden Reptilien. Zeitschr. f. wiss. Zool., vol. 48, pp. 261-800.

JANosik, J. 1883 Bemerkungen itber die Entwickelung der Nebenniere. Arch.
mikr. Anat., vol. 22, pp. 738-745.

Kunz, A. 1909a A contribution to the histogenesis of the Sympathetic nervous
system. Anat. Rec., vol. 3, pp. 158-165.

1909 b The role of the vagi in the development of the sympathetic
nervous system. Anat. Anz., vol. 35, pp. 381-390.

1910 a The development of the sympathetic nervous system in mam-
mals. Jour. Comp. Neur., vol. 20, pp. 211-258.

1910 b The development of the sympathetic nervous system in birds.
Jour. Comp. Neur., vol. 20, pp.’ 284-308.

1911 a The development of the sympathetic nervous system in cer-
tain fishes. Jour. Comp. Neur., vol. 21, pp. 177-214.

1911 b The development of the sympathetic nervous system in turtles.
Amer. Jour. Anat., vol. 11, pp. 279-312.

1911 ¢ The evolution of the sympathetic nervous system in verte-
brates. Jour. Comp. Neur., vol. 21, pp. 215-236.

1911 d The development of the sympathetic nervous system in the
Amphibia. Jour. Comp. Neur., vol. 21, pp. 397-416.

DEVELOPMENT OF THE ADRENALS IN THE TURTLE 89

Leypie, F. 1853 Anatomisch-histologische Untersuchungen iiber Fische und
Reptilien. Berlin.

Minor, C. 8S. 1897 Human embryology, New York.

von Minatcovics, V. 1885 Untersuchungen iiber die Entwickelung des Harn-
und Geschlechtsapparates.der Amnioten. Internat. Monatschr. Anat.
Hist., vol. 2, pp. 387-414.

Pout, H. 1906 Die vergleichende Entwickelungsgeschichte der Nebennieren-
systeme der Wirbeltiere.. Hertwig’s Handbuch d. vergl. u. exper.
Entwg. d. Wirbeltiere, vol. 3, pp. 448-616.

ScuHuttze, O. 1897 Grundriss der Entwickelungsgeschichte des Menschen und
der Siugetiere, Leipzig.

Soup, A. 1903 Recherches sur le développement des capsules surrénales chez
les vertébrés supérieurs. Jour. del’anat. et phys. Par Année 39, pp. -
197-293.

Wetpon, W.F.R. 1885 Onthe suprarenal bodies of vertebrates. Quart. Jour.
micr. Soc., vol. 25, pp. 137-150.

WHITEHEAD, R. H. 1903 The histogenesis of the adrenal in the pig. Amer.
Jour. Anat., vol. 2, pp. 349-360.

Wiese, J. 1901 Ueber die Entwickelung der Nebenniere des Schweines, ins-
besondere der Marksubstanz. Anat. Hefte, Weisb., vol. 16, pp. 115-
150.

THe NERVES Of THE THYROLD AND PARATHYROID
BODIES

DARMON A. RHINEHART

From the Laboratory of Anatomy, Indiana University
FIVE FIGURES

Prior to the year 1867 the nerves of the thyroid were usually
described as vaso-motor for the supply of the numerous blood
vessels.

Peremeschko (’67) examined thin teased preparations of thy-
roid that had been macerated in acetic acid and water, and found
many more nerves than the gland was thought to contain. He
described some of these as following the arteries, and others as
leaving the vessels, dividing again and again, and finally losing
themselves as fine varicosed branches in the interfollicular con-
nective tissue.

Poinearé (’75), after macerating the gland in dilute acetic
acid colored with fuchsin, found abundant nerves and nerve
plexuses, which he considered to be a separate nerve supply for
this organ, connected with the central system by the nerves enter-
ing the gland. He also described ganglion cells lying in groups
or clumps either in the substance of the larger nerves, in the nerves
at their places of branching, or alongside the nerve stems.

Anderson (’94) described the formation of very elaborate peri-
vascular plexuses from which fine fibers penetrate between the
follicles and form perifollicular plexuses in which the follicles
seem to be imbedded. These nerves are very irregular in their
course, do not anastomose, and, after repeated divisions, the ter-
minal fibrillae end in knobs on the bases of the cells. At no time
do they enter the cells nor do they penetrate between them.

Berkeley (95) described perivascular plexuses similar to those
of Anderson, and a primary and secondary plexus surrounding

91

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 2
MAY, 1912

92 DARMON A. RHINEHART

each follicle, the former lying some distance from the bases of the
cells, the latter directly on the cells. Small curved branches
from the secondary plexuses end in knobs and look as if they pen-
etrate between the cells while others are straight indicat ng that
their endings he on the basal ends of the cells.

Crissafulli’s (92) description of these nerves agrees in almost
all particulars with that of Anderson, with the exception that he
found ganglion cells here and there throughout the g and.

TECHNIQUE

The only technical methods for staining nerve fibers that were
found of value in this work were the intra-vitam methylene blue
method of Ehrlich and the Golgi method.

The methylene blue method was carefully tried in all its forms,
but was only valuable in staining the larger nerves accompanying
the blood vessels into the gland.

The original Golgi method in the rapid and mixed form to-
gether with the modification of Berkeley gave poor results. The
most successful and complete staining of the nerves was secured
by the use of the ‘double impregnation’ method of Cajal, modified
as to the length of time the tissues were allowed to remain in the
different fluids.

This method consists in fixing small pieces of tissue for from six
to eight days in a mixture made up of four parts of a 3.5 per cent
solution of potassium bichromate and one part of a 1 per cent
solution of osmic acid. After fixation and a careful washing in
distilled water the tissues are put into a | per cent solution of
silver nitrate for three days. This is followed by a second fixa-
tion for an additional three days, and a second impregnation
with silver. After the first day in the second silver solution free
hand sections are examined at intervals until the nerves are found
to be satisfactorily stained. The tissues are then rapidly dehy-
drated in several changes of 95 per cent and absolute alcohol, and
are quickly imbedded in celloidin, the blocks being hardened in
chloroform. Sections, 25 to 100 microns thick, are cut under oil,
cleared on the slides in earbol-xylol and xylol, and mounted

NERVES OF THYROID AND PARATHYROID BODIES 93

in thick xylol-balsam. The slides should then be warmed until
enough of the xylol has evaporated to leave the balsam brittle
on cooling, and while still warm, covered by warmed cover slips.
Preparations prepared in this way are permanent.

THE NERVES OF THE THYROID

The nerves which go to the thyroid are entirely of the non-
medullated variety and reach it from the neighboring cervical
sympathetic ganglia by following the perivascular connective
tissue and the tunica adventitia of the thyroid arteries. It is
often very difficult to examine these nerves, for they are obscured
in sections stained by the chrome silver method by large amounts
of black and brown precipitates, and by the staining of the fat.
By carefully examining sections where there is only a small amount
of fat, or better, by studying sections stained by the methylene
blue method, the nerves accompanying the blood vessels can be
followed. They appear as large, wavy, irregular strands, com-
posed of individual varicosed axis-cylinders collected together
into bundles similar to the wires in a cable.

These nerves do not branch elaborately outside the gland; the
only branches coming from them are small ones which form the
perivascular plexuses in the arterial walls, and probably furnish
the vaso-motor supply. As soon as the arteries penetrate into
the gland substance they branch profusely, this branching being
accompanied by a corresponding branching of the nerves, so that
as the arteries decrease in size there is a similar decrease in the
size of the accompanying nerves. The nerves supplying these
smaller arteries (fig. 2) form an elaborate perivascular plexus, and
give off the branches which penetrate between the follicles, and
form the perifollicular plexuses. The latter are the only ones
that can be said to be the true glandular or secretory nerves.

The arterial or perivascular plexuses are formed by the branches
of a few relatively large nerve trunks (fig. 2, A), lying in the con-
nective tissue immediately surrounding the vessels, and having
a course parallel to that of the arteries. The branching from these
nerves is very irregular; the branches taking any direction after

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 2

94 DARMON A. RHINEHART

leaving the main trunk. They extend for a longer or shorter
distance around the vessels, divide again and again, and finally
become resolved into an intricate network of fine end fibrillae.
This elaborate nervous plexus is located in the perivascular
connective tissue and in the tunica adventitia of the arterial walls,
while some of the finer fibrillae penetrate into the tunica media.

The larger nerves are composed of bundles of axis-cylinders,
and the branching of these consists merely in some of these axis-
cylinders leaving the main bundle and taking a different direc-
tion. As this method of branching continues the axis-cylinders
finally come to lie singly. From these single fibers are given off
true branches, which do not decrease in size but are usually more
irregular, more varicosed, and have a more wavy course. The
final end fibrillae are often very short, are beset with many vari-
cosities, and soon end in irregular enlargements or end-knobs.

The many crossings and recrossings of the nerves, together with
their elaborate branching, is very misleading, and often gives the
appearance, especially under low magnification, of true anasto-
moses. In the thicker sections where the nerves are great in
number it is often impossible to distinguish the separate fibers.
But if relatively thin sections are examined under high magni-
fication the fibers can all be traced individually. Therefore,
contrary to the statement of Berkekey, true anastomoses do not
occur. In examining these nerves it is seen that the origin and
ending of the most of the branches cannot be traced in a single
section for they are only fragments, the remainder having been
cut away.

All the nerves that go to make up the perivascular plexuses are
more or less varicosed. Even the axis-cylinders that form the
large nerve bundles are beset with many irregular enlargements.
There may also be precipitations of chrome-silver between the
axis-cylinders in places, so as to give the nerves the appearance
of solid cords. At the points of branching of all the nerves there
are irregular, triangular, black masses, which may be either a
portion of the nervous structure or a black precipitate. The
smaller nerves, where the axis-cylinders lie singly, are more vari-
cosed than the corresponding parts of the larger nerves; this

NERVES OF THYROID AND PARATHYROID BODIES 95

being especially true of the final end fibrillae. In fact these
nerves appear as if consisting of many irregular enlargments con-
nected by fine threads (fig. 4). The varicosities may be only
slightly larger than the fibers or they may be twice as large; the
larger ones are cylindrical, spherical, or irregularly triangular, the
majority, however, are either oval or spindle-shaped.

The final endings of the nerves consist of fine end branches.
The smallest fibers usually divide into two, sometimes more, end
fibrillae. The course of these fibrillae is very irregular, they are
very varicosed and soon end. The tip consists of a spindle-
shaped enlargement or end-knob. This ending sometimes occurs
in the perivascular connective tissue, sometimes in the tunica
adventitia, but most often in the tunica media in relation to the
smooth muscle fibers of the vessel walls. Occasionally endings
from the perivascular plexuses can be seen penetrating the peri-
vascular connective tissue and terminating in relation with the
bases of the epithelial cells of the immediately adjacent follicles.

The plexus surrounding the arteries is more dense and com-
plex than that surrounding the accompanying veins. The venous
plexus resembles the arterial in formation and architecture but
it is not nearly as elaborate. The peri-capillary plexus is usually
formed from the few end branches of a single fiber, running par-
allel to the course of the vessel.

The most important nerves of the thyroid from a physiological
standpoint are those that have their endings in intimate relation
to the gland cells, for they are the true glandular or secretory
nerves. These nerves pass into the gland substance alongside
of, or in the walls of the arteries, together with those that form
the perivascular plexuses and furnish the vaso-motor supply.
They assist in the formation of the perivascular plexuses but there
is no way of distinguishing them from the vaso-motor nerves.
There may be, however, a physiological difference in the two
kinds of nerves, but if such exists, it cannot be determined histo-
logically.

As has been said before, the nerves do not leave the large arte-
ries and pass directly into the gland substance, for this does not
occur until the vessels, by branching, have decreased very much in

96 DARMON A. RHINEHART

size. From the perivascular plexuses of these smaller arteries
are given off here and there nerves which penetrate into the inter-
follicular connective tissue for a greater or lesser distance and
form the perifolliicular plexuses. These branches are almost
always single fibers, for none of the larger nerves leave the ves-
sels as such.

The perifollicular plexuses completely surround all of the fol-
licles of the thyroid, or, as Anderson has said, there is present a
diffuse plexus of nerves in which the follicles seem to be imbedded
(fig. 3) The nerves coming from the perivascular plexuses may
supply a follicle near the artery, they may supply one at a con-
siderable distance, or their branches may enter into the formation
of the plexuses of several follicles. These nerves divide again and
again into a large number of fine, varicosed fibers which completely
surround a single follicle or adjacent follicles in a dense nervous
network. There is absolutely no regularity, arrangement or
method in the place or manner of branching, nor in the direction
the branches may take after leaving the main stem. Neither is
there any regularity in the distribution of the glandular nerves,
nor anything that can be said of their distribution, except that
they enter into the formation of the plexuses around one or more
of the follicles of the thyroid. There are no primary or secondary
plexuses, as described by Berkeley, but a single one surrounding
each follicle.

In regard to the anastomoses of the nerves of these plexuses
the same statement applies that was made in regard to the peri-
vascular nerves. In the thinner sections the nerves can all be
traced as individuals, and the branches can be traced to the parent
stem if it is included in the section.

The perifollicular nerves are also very varicosed, the varicosi-
ties corresponding in size and shape to those found on the nerves
of the perivascular plexuses.

Perhaps the most important fact to be determined concerning
the nerves of the thyroid is the position and manner of the final
endings of the follicular nerves, for the effect of nervous impulses
on the secretory activity of the gland cells would depend, to a
great extent, on the intimacy of the relations of the nerves to them.

NERVES OF THYROID AND PARATHYROID BODIES 97

The presence or absence of a definite basement membrane around
the follicles is an important point in regard to nerve and cell
relations. The majority of investigators are of the option that
a definitely formed basement membrane is absent, but that there
is a condensation of the connective tissue immediately surround-
ing the follicles, and this gives support to the bases of the epithelial
cells (Baber).

The final endings of the nerves are short, fine, very varicosed
fibrillae which lie in this condensed connective tissue (figs. 1 and
4). Some of these end fibrillae are curved, as described by Berke-
ley, while others are straight, and still others, and perhaps a
majority, are irregular. They always end in a knob-like enlarge-
ment on the basal ends of the cells; each cell, however, does not
come into relation with the end of a nerve, but only a few scat-
tered here and there throughout the follicles. I have been unable,
by most carefully examining a large number of sections with the
best magnification obtainable, to see any of these endings either
entering the cells or penetrating into their intercellular substance.

In some of the literature concerning the nerve supply of the
thyroid the statement is made that ganglion cells were found in
different locations within the gland. Anderson and Berkeley,
however, with whom I agree, state that these structures are not
present. Thereare,in many of the sections stained by the chrome-
silver method, numbers of black precipitations which resemble
ganglion cells very closely in size and shape and could easily be
mistaken for them. A careful examination of many of these does
not reveal anything resembling a nucleus or nucleolus, while even
if nerve fibers enter them, the arrangement is not characteristic
of ganglion cell processes.

THE NERVES OF THE PARATHYROID

A careful examination of all the literature accessible concern-
ing the parathyroid has failed to reveal any reference whatever
to its nerve supply, and one of the latest histologies makes the
statement that this subject needs further investigation.

The technique used in investigating the nerves of the parathy-
roid was the double impregnation method of Cajal as given in

98 DARMON A. RHINEHART

detail elsewhere in this paper. Several of these small bodies
were carefully dissected out and carried through separately, but
with poor results. Those sections of the parathyroid in which
the nerves are the most abundant are attached to some of the
more satisfactorily stained sections of thyroid. The treatment
in both eases is, therefore, identical.

The first and most striking thing observed on examination of
these sections of parathyroid is the scarcity of the nerves as com-
pared to the great numbers found in the thyroid. It.would at
first seem that this might be due to incomplete staining, which
is not at all probable, for in the adjacent thyroid tissue partially
or completely surrounding the parathyroid, the nerves are
beautifully stained. The similar density of the two bodies leads
me to believe that a method giving a satisfactory result in one
would lead to a like result in the other.

The arteries supplying the parathyroid bodies are branches of
the thyroid arteries, and may take either or both of two routes
in reaching their places of distribution. When the bodies are
not closely connected there is one relatively large branch that
passes directly into the parathyroid, but when the thyroid par-
tially or completely surrounds the parathyroid, more numerous
but smaller branches take origin from the thyroid vessels within
the thyroid gland and pass through the intervening small amount
of loose connective tissue into the parathyroid. Inasmuch as
the nerves follow these arteries, they arise from the large nerve
bundles around the thyroid vessels and accompany their branches
into the parathyroid. It is very probable, therefore, that the
nerves supplying both glands constitute a single set of sympa-
thetic fibers.

Around the parathyroid arteries there are formed nerve plexuses
resembling the perivascular plexuses of the thyroid vessels, dif-
fering, however, in not being nearly so elaborate, and in consist-
ing of single fibers. The branching of the nerves accompanies
the branching of the arteries, so that the smaller arterial twigs
usually carry with them a single nerve fiber. There are no nerves
around the veins or capillaries.

NERVES OF THYROID AND PARATHYROID BODIES 99

In most instances the nerves can be traced along the vessel
wall, within which are also found their few terminal branches.
A few other fibers are present which cannot be followed along a
vessel, but which seem to run in the connective tissue between the
groups and cords of gland cells. It is probable, however, that
these accompany smaller vessels which are too poorly stained to
be visible. I am of the opinion that the nerves are entirely vaso-
motor for the supply of the blood vessels, and that there are no
special glandular or secretory nerves in the parathyroid, because
none of the fibers leave the supporting connective tissue and pene-
trate into the cell groups.

CONCLUSIONS

1. The nerves of the thyroid are entirely non-medullated and
reach it from the cervical sympathetic ganglia by following the
thyroid arteries.

2. In the thyroid there are formed elaborate nervous plexuses
around all the blood vessels and all the follicles, the nerves form-
ing the latter coming from the plexuses surrounding the smaller
arteries.

3. The perivascular nerves end in the walls of the blood vessels
and furnish the vaso-motor supply, while those of the perifollicu-
lar plexuses end on the bases of the epithelial cells and probably
carry impulses influencing secretion.

4, All the nerves are varicosed but do not anastomose.

5. The nerves of the parathyroid come from the same set that
supplies the thyroid and pass into it along with the branches from
the thyroid arteries. These nerves probably all end in the vessel
walls and are vaso-motor in function.

6. No ganglion cells are found in either the thyroid or para-
thyroid bodies.

100 DARMON A. RHINEHART

BIBLIOGRAPHY

ANDERSON, OskaR A. 1894 Zur Kenntniss der Morphologie der Schilddriise.
Archiv fiir Anatomie und Physiologie, Anat. Abt.

BaBER, E.C. 1876 and 1881 Contributions to the minute anatomy of the thyroid
gland of the dog. Phil. Trans. of the Royal Society of London; vol.
166, part 2; vol. 172, part 3.

BERKELEY, Henry J. 1895 Nerves of the thyroid gland of the dog. Johns
Hopkins Hospital Reports, vol. 4.

CrissaFuLui, E. 1892 I Nervi della Glandola Tiroide. Bulletino mens. della
Acad. Gioenia si scienz. nat. in Catania, Nuova Serie, Fasc. 25.

DoateL, A. 8. 1910 Methylenblau zur Nervenfairbung. Enzyklopadie der
mikroskopischen Technik, Band 2, Berlin.

Harpesty, Irvine 1902 Neurological technique. Chicago.

Huser, G. Cart 1892 Zur Technik der Golgi’schen Methode. Anatomischer
Anzeiger., Band 7.

Kauuius 1910 Golgische Methode. Enzyklopadie der mikroskopischen Tech-
nik, Band 1. ;

Masor, RaupH H. 1909 Studies on the vascular system of the thyroid gland,
Amer. Jour. Anat., vol. 9.

PEREMESCHKO 1867 Ein Beitrag zum Bau der Schilddriise. Zeitschrift fiir
wissenschaftliche Zoologie, Band 8.

Porncarh 1875 Notesurl’innervation dela glandethyroide. Journal del’anat.
et de la phys., Tome 11.

Witson, J. Gorpon 1910 Intra vitam staining with methylene blue. Anat.
Rec., vol. 4, no. 7, July.

NERVES OF THYROID AND PARATHYROID BODIES 101

All drawings were made with an Edinger drawing apparatus using Leitz compens
oculars nos. 4 and 6, and 4 and 16 mm. apochromatic objectives.

Fig. 1 Drawing showing the relation of the nerve endings to the gland cells.
x 650.

Fig. 2. A small artery in the interior of the thyroid showing the elaborate peri-
vascular plexus. A, large nerve bundle breaking up to form the plexus; 6, nerve
leaving the perivascular plexus to form the perifollicular plexus. > 630.

Fig. 3. Several adjacent follicles showing the intimate connections of the adja-
cent perifollicular plexuses. > 250.

Fig. 4 One of the follicles shown in fig. 3. .N, endings of the follicular nerves;
V. varicosities on the nerves; H, the epithelial cells, the intercellular substance
stained a light brown. X 650.

Fig. 5 Section of the parathyroid. A, small artery accompanied by nerves
entering the parathyroid from the thyroid; B, supporting connective tissue; C,
nerves in this connective tissue; D, nerve endings. X 150.

102

THE ANOMALOUS PERSISTENCE IN EMBRYOS OF
PARTS OF THE PERI-INTESTINAL RINGS
FORMED BY THE VITELLINE VEINS

ALEXANDER S. BEGG

Professor of Histology and Embryology in Drake University, Des Moines, Iowa
FIVE FIGURES

While studying the development of the pancreas, Dr. Fred-
eric T. Lewis found two embryos which present anomalies of
the intra-embryonic portion of the vitelline veins. He has
referred to one of these, a human embryo of 11.5 mm., in Keibel
and Mall’s Human Embryology (German ed., p. 421); the other,
a pig embryo of 10 mm., has not been previously recorded.
These specimens, which prove to be of considerable embryological
interest, were placed at my disposal, and I have made a careful
study of them at the Harvard Medical School, in coéperation
with Dr. Lewis, to whom I am indebted for many valuable sug-
gestions. Wax reconstructions have been prepared, both of the
abnormal and of normal specimens, which show the course of the
veins and the close correlation which exists between their arrange-
ment and the form of the pancreas. The models have been de-
posited in the collection at the Harvard Embryological Labora-
tory, where they may be examined at any time.

The peri-intestinal rings formed by the vitelline veins were
first made known by His. In a familiar figure, here reproduced
as fig. 1, he showed that the right and left vitelline (or omphalo-
mesenteric) veins anastomose with one another at three places,
namely (1) ventral to the intestine within the liver; (2) dorsal
to the intestine below the dorsal pancreas; and (3) ventral to the
intestine above the yolk-stalk. Thus two venous rings are pro-
duced, each of which encircles the intestine. He showed, more-
over, that the left half of the upper ring and the right half of the

103

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 2

104 ALEXANDER S. BEGG

lower ring degenerate, and that the remaining portions of the two
vitelline veins form a single vein winding about the intestine.

These relations are correctly represented in fig. 1, but in one
respect the drawing of His is subject to criticism. ‘Two vessels
are seen ascending along the intestine to fuse at the lower ventral
anastomosis. Are these the right and left vitelline veins as Evans
has labelled them in his copy of this figure (Keibel-Mall, German
ed., p. 653) and as His designated them in the younger ‘Embryo R’
(5 mm.)? Or is the left vessel, V.p. in the figure, the superior
mesenteric vein and the right vessel the fused pair of vitelline
veins? Ifthe modelof His’s ‘Embryo A’ (7.5 mm.), as reproduced
by Ziegler, is examined, it will be found that both vessels shown in
fig. 1, are continued beyond the loop of intestine along the yolk-
stalk, thus representing the right and left vitelline veins respec-
tively. This, however, is an error. The left vessel in an embryo
of the stage in question does not extend beyond the intestinal
loop. It is the superior mesenteric vein, and the right vessel
represents the original pair of vitelline veins, which have fused.

The development of the single stem formed by the vitelline
veins may be observed in rabbit embryos. At the time when the
lower ventral anastomosis is formed, the yolk-sac is close to the
intestine. The veins coming from the right and left halves of the
sac meet and anastomose ventral to the intestine and imme-
diately separate to encircle it. With the formation of the yolk-
stalk the ventral anastomosis appears to be drawn out in a single
stem, which increases in length with the formation of the primary
intestinal loop. The prolonged ventral anastomosis becomes
separated from the mesentery, so that it appears as a single vein
which swings across the abdominal cavity with a peritoneal invest-
ment of its own. In this condition, in human embryos of the
third month, it was observed by Luschka (’63).

The superior mesenteric vein apparently arises in human em-
bryos of about 5 mm. Thus, in a 4.9 mm. specimen, Ingalls has
found several small veins ascending behind the intestine to join
the dorsal anastomosis of the vitelline veins, which, it should be
noted, is plexiform. These ascending veins probably give rise
to the superior mesenteric vein. In a 7 mm. human embryo,

ANOMALIES OF THE VITELLINE VEINS 105

Elze has shown that the superior mesenteric vein is a well-defined
stem which empties into the spiral vessel formed from the peri-
intestinal rings. The place of junction comes to lie on the left
side of the intestine, both in human embryos (Elze) and in pig
embryos (Lewis, Thyng). In other words it has shifted ventrally,
and the mesenteric vein appears to join the left half of the lower
peri-intestinal ring. Thus at the stage shown in His’s figure,
when the spiral vein has been formed from the peri-intestinal
rings, the veins which unite near the intestine are the superior
mesenteric vein and the fused vitelline trunk. On the other hand,

Fig. 1 His’s diagram showing the formation of the portal vein, V.p. V.w. and
V.u.d., parts of the right umbilical vein. V.w.s., left umbilical vein. V. Ar.,
ductus venosus. !

the place where the right and left vitelline veins unite is near the
yolk-sac, as shown in the reconstructions by Lewis and Thyng;
and this is far removed from the area included in His’s figure.

The true relations of these vessels, as here described, have
doubtless been well understood by investigators of the venous
system, but it is difficult to find an explicit account of them.
Luschka recognized a vitelline vein coming from the yolk-sae
and a mesenteric vein coming from the mesentery, but apparently
he did not consider the possibility that the mesenteric vein might
be derived from a left vitelline vein. This possibility, suggested

1 For the use of the electrotype of this figure, and for many facilities for study

and investigation during my stay at the Harvard Laboratory, I am deeply in-
debted to Professor Charles S. Minot.

106 ALEXANDER S. BEGG

by His’s models and figures, was rejected by Dexter and Lewis,
both of whom figured the elongated ventral anastomosis of the
vitelline veins, and portions of the right and left veins of the yolk-
sae which unite to produce it. Hochstetter, in his admirable
résumé in Hertwig’s Handbuch, neither figures nor describes
the notable elongation of the ventral anastomosis, and Elze fails
to recognize it, since he describes the vitelline trunk which crosses
the abdomen as the left vitelline vein.

With the explanation which we have made, His’s diagram (fig. 1)
will make clear the nature of the anomaly shown in fig. 2. This
figure represents a model of the veins of a pig embryo of 10 mm.;
viewed from the left side. In addition to the veins, it shows
portions of the stomach and liver, including the gall-bladder, and
also the dorsal and ventral pancreases and a large portion of the
primary intestinal loop. The distal part of this loop and the
yolk-stalk had been cut away before the embryo was sectioned.
In reconstructing the organs, only the epithelial portion was
included.

In this specimen the fused vitelline veins form a rather narrow

vessel showing evidence of atrophy at several points. Within
the umbilical cord it occupies a distinct fold of the mesentery.
Upon reaching the abdominal cavity the vein leaves the intestinal
mesentery and crosses, free from it, to the connective tissue about
the duodenum. Ventral to the duodenum it suddenly enlarges
and is joined by the superior mesenteric vein. The latter, through-
out most of its course, forms part of a net-like system of channels
lying in the mesentery. It is a largé vein which passes backward
and upward in a sweeping curve to join the vitelline vein. In
joining the vitelline vein it passes ventral to the intestine instead
of dorsal to it, and the main trunk formed by the union of these
vessels is on the right side of the intestine instead of on the left.
The embryo presents, therefore, a persistence of the right half of
the lower peri-intestinal ring, which forms a portion of the main
channel to the liver. In the 7.8 mm. embryo described by Thyng,
the right half of the lower ring was not found, and it presumably
atrophies normally in still younger embryos.

pe

VF umb—

Paned

©

Vines.sup

& Pane. d.

S

V. Li.

Pane.v.

Vv nLES.S Up

Pane. dN li.

ID mes sup. 4 \

4

Figs. 2,3 and 4 Wax reconstructions of parts of the liver, intestine, and ad-
jacent veins. 30 diam. Fig.2 Pig embryo: 10mm. Harvard Embryological
Collection, Series 1698. Fig. 3 Human embryo: 10mm. H. E. C., Ser. 1000.
Fig. 4. Human embryo: 11.5 mm. H. E. C., Ser. 189. Pane.d., Panc.v., dor-
sal and ventral pancreases. Ves.fel., gall bladder. V.li., splenic vein. V. mes.
sup., Superior mesenteric vein. V.v., trunk formed by the fusion of right and left
vitelline veins. V.wmb., umbilical vein.

107

108 ALEXANDER S. BEGG

The left half of the lower ring normally forms a large vessel
which winds around the dorsal wall of the intestine just posterior
to the duct of the dorsal pancreas, and then ascends to the liver.
The glandular mass of the dorsal pancreas, in growing forward on
the right side of the intestine, encounters this vein and becomes
molded about it. It sends ‘ventral processes’ forward, usually
on the right side of the vein, but sometimes on its medial side.
In the abnormal embryo there are two ventral processes of the
dorsal pancreas, both of which are shown in the figure. The nor-
mal course of the superior mesenteric vein, after being joined by
the vitelline vessel, would be under the duct of the dorsal pancreas
and upward on the medial side of these processes, and the shape
of the pancreas in the abnormal embryo indicates that such a
vessel was present at an earlier stage. It has, however, disap-
peared and the left half of the lower peri-intestinal ring, together
with the dorsal anastomosis of the vitelline veins, is represented
by a slender vessel which passes under the dorsal pancreas near
its distal extremity. There it is joined by the splenic vein. Be-
fore the left limb of the lower ring receives the splenic vein, it
presents a small branch directed toward another short branch
across the top of the pancreas. These vessels may formerly have
connected with one another. The unusual course of these rep-
resentatives of the dorsal anastomosis of the vitelline veins may
be explained by the plexiform nature of the original connection.
The upper peri-intestinal ring has developed normally. — Its left
half has disappeared, and its right half persists as the portal
vein.

Finally it should be noted that the ventral pancreas in this
embryo is bi-lobed, and that it bifurcates over the upper edge of
the abnormal vein. If its lobes correspond with those usually
found (ef. Lewis, ’11) it is evident that the entire ventral pancreas
has been displaced to the right, since the ventral process of the
dorsal pancreas approaches its left lobe. Its relation to the vein
suggests that such a displacement has occurred.

The abnormal human embryo (11.5 mm.), which has been
modelled in the same way as the 10 mm. pig, is shown in fig. 4.
Above it, in fig. 8, a normal specimen of 10 mm. is presented for

ANOMALIES OF THE VITELLINE VEINS 109

comparison. ‘The smaller embryo is somewhat younger and fails
to show the rotation of the intestinal loop, but in regard to the
veins the specimens are quite comparable. In the normal em-
bryo the left half of the upper ring and the right half of the lower
ring have disappeared. In the abnormal embryo the left half
of the upper ring is absent, but the right half of the lower ring
remains as the direct continuation of the fused vitelline veins.
The ventral portion of the left half of the lower ring has disap-
peared, but its dorsal portion remains as the continuation of the
superior mesenteric vein. Although this anomaly differs from
that in the pig in many ways, there is a striking resemblance in
the dorsal displacement of the mesenteric vein, which passes
beneath the pancreas near its extremity. The explanation of this
feature is not apparent.

In the normal human embryo the duct of the dorsal pancreas
opens nearer the stomach than the common bile duct. The
distance between the two outlets, calculated from the wax recon-
struction, is 0.16 mm. In the abnormal embryo, however, the
relative position of these outlets is reversed (as already recorded
by Lewis) and I find that the duct of the dorsal pancreas opens
0.12 mm. below or caudal to the orifice of the common bile duct.
It is possible that the abnormal arrangement of the adjacent
veins led to this anomaly, but this cannot be affirmed. The small
and rather rudimentary ventral pancreas in the 11.5 mm. speci-
men extends downward and forward in close relation with the
left side of the abnormal vein.

As a summary of the observations which we have recorded, a
diagram (fig. 5) is presented, in which the normally persistent
portions of the peri-intestinal rings may be compared with the
parts found in the pigandin man. In these figures the term por-
tal vein is applied to the vessel formed by the union of the superior
mesenteric and splenic veins, in accordance with anatomical
usage, and is not extended to include the vessel made by the
junction of the superior mesenteric and fused vitelline veins. It
would be interesting to find adult specimens which had passed
through the abnormal stages figured, but apparently such cases
have not been recorded. In the human embryo which we have

110 ALEXANDER 8S. BEGG

described, after the obliteration of the vitelline trunk, essentially
normal relations would be restored. But in the pig the superior
mesenteric vein would cross in front of the duodenum, and it is
probable that this condition will some time be found in adult
animals.

NN

Liver

“V, lien.

~V. mes. sup.

Fig. 5 Diagrams showing, in ventral view, the variations observed in the peri-
intestinal venous rings. The probable position of the obliterated portions is
indicated by stippled vessels. A, normal human embryo. #&, abnormal pig em-
bryo. C, abnormal human embryo. V.lien., splenic vein. V.mes.swp., superior
mesenteric vein. V.p., portal vein. V.v., trunk formed by the fusion of the vitel-
line veins.

BIBLIOGRAPHY

Dexter, F. 1902 On the vitelline vein of the cat. Amer. Jour. Anat., vol. 1,
pp. 261-267.

Euze, C. 1907 Beschreibung eines menschichen Embryo von zirka 7 mm.
grosster Linge. Anat. Hefte, Abth. 1, Bd. 35, pp. 409-492.

Evans, H. M. 1911 Die Entwicklung des Blutgefiisssystems. Handbuch der
Entw. des Menschen, herausgegeben von F. Keibel und F. P. Mall,
Bd. 2, pp. 551-688.

His, W. 1885 Anatomie menschlicher Embryonen. III. Zur Geschichte der
Organe. Pp. 1-260.

InGatits, N. W. 1908 <A contribution to the embryology of the liver and vascular
system inman. Anat. Rec., vol. 2, pp. 338-344.

Hocustertrer, F. 1906 Die Entwickelung des Blutgefisssystems. Handbuch
der verg. und exp. Entw. der Wirbeltiere, herausgegeben von O. Hert-
wig. Bd. 3, Teil 2, pp. 21-166. (Published as an ‘Abdruck’ in 1902.)

Lewis, F. T. 1903 The gross anatomy of al2mm. pig. Amer. Jour. Anat., vol.
2, pp. 211225.
1911 Die Entwicklung des pancreas. Handbuch der Entw. des Mens-
chen, herausgegeben von F. Keibel und F. P. Mall, Bd. 2, pp. 418-436.
1911 The bi-lobed form of the ventral pancreas in mammals. Amer.
Jour. Anat., vol. 12, pp. 3889-400.

LuscuKxa, H. 1863 Die Anatomie des menschlichen Bauches. Die Anat. des
Menschen, Bd. 2, Abth. 1, pp. 1-377.

Tuyne, F.W. 1911 The anatomy of a7.8 mm. pigembryo. Anat. Rec., vol. 5,
pp. 17-45.

THE DEVELOPMENT OF THE AORTA AND AORTIC
ARCHES IN RABBITS

JOHN LEWIS BREMER
From the Harvard Medical School, Boston

NINE FIGURES

The development of the primary blood-vessels in the body of
the embryo has for many years been a matter of dispute. Evans,
in the German edition of the second volume of the Keibel-Mall
Embryology, sums up the matter as follows:!

Whether the first blood-vessels of the embryonic body arise by in-
growth from the yolk-sac capillaries, or whether the embryonic vessel-
stems, or at least a part of them, originate in situ from the mesoderm
of the body, is still an open question. Both views have found their
supporters; the name of His is connected with the first mentioned idea,
the names of Riickert and Mollier especially with the second.

In birds it is possible to prove that the greater part of the de-
scending aortae develop from the mesial border of the capillary
plexus which has extended in from the yolk-sac, and this is very
probably true of mammals also; but (to quote again):

For the cranial part of the aorta, on the other hand, the results are
contradictory. His describes it as arising from a further ingrowth of
the same extra-embryonic capillaries which form the aortain its more
caudal portion; the capillary chain grows finally over the blind end of
the pharynx, turns ventral-ward, and joins the cranial part of the heart
cavity. In rebuttal, Riickert and Mollier have stated in numerous
articles that the aortae arise in loco from cells of the visceral layer of
the mesoderm. It is impossible at present to insist that the anlagen
found on the yolk-sac are the only ones for the endothelium of the body
vessels. (Keibel-Mall Entwickelungsgeschichte, vol. 2, p. 552, etc.)

1 Tn the American edition of this work some of the results of the present paper
have been added.

111

Wee JOHN LEWIS BREMER

In this paper and by means of the following reconstructions
made from serial sections of embryos in the Harvard Embryo-
logical Collection, I hope to show clearly that the view of His and
his supporters is in the main correct, that the cranial part of the
aorta arises as an extension of the capillary network of the yolk-
sac; and also to throw more light on the development of the ven-
tral aorta, the aortic arches, and the pulmonary artery. Forthe
study of this question I have chosen to work primarily with the
rabbit, partly because of the excellence of this material in this
laboratory, and partly because the presence of the ‘lateral hearts,’
described by Rathke, and easily recognizable in this species even
in early stages, readily marks the position of this part of the blood-
vessel net, and makes interpretation of the secondary foldings
much simpler.

To some extent this same material was used by Dr. F. T. Lewis
in a paper on “The Intra-embryonic Blood-vessels of Rabbits
from 84 to 13 days,” which, accompanied by a demonstration of
sections and graphic reconstructions, was read at the meeting
of the American Association of Anatomists in 1903, but never
-published in full. In the report of the Proceedings in The Amer-
ican Journal of Anatomy, vol. 3, a résumé is given as follows:
‘From the network of vessels in the splanchnopleure of the yolk-
sac, all intra-embryonic vessels are apparently derived as off-
shoots. The network ends mesially in the two aortae. With
the formation of the pharynx, this net is so folded as to produce
dorsal and ventral aortae with the connecting first arch.” It
will be seen that Lewis agrees with His as to the origin of the
dorsal aorta, but discards the idea that this vessel grows forward
around the tip of the pharynx to join ventrally with the anterior
end of the heart. A glance at figs. 1 and 3 will show that this
is correct; dorsal aorta, first arch, ventral aorta and heart anlage
are all laid down almost simultaneously.

Here a few words are needed on the character of the early blood-
vessels. The most recent investigations in this field have been
carried onalmost exclusively by careful injections of fresh embryos,
which are then studied as transparent objects or are converted
into sections from which reconstructions are made. This method

AORTA AND AORTIC ARCHES IN RABBITS 113

presupposes that all vessels are injectable, and in fact the claim
is made that many collapsed vessels cannot be distinguished in
sections until opened and marked by the injection mass. While
in no way wishing to belittle the value of this method of research,
or to discourage the increase of the many beautiful and valuable
preparations obtained by its use, I still feel that its limitations,
as they are shown in this paper, should be pointed out. As
insisted on by His in 1900, and by many other authors (His’ name
is used as that of the champion of this idea), the first blood-vessels
on the yolk-sac and elsewhere are solid cords or strands of cells,
without lumen: or to use other words, the actual vessels are always
preceded by solid growths, which secondarily become hollow
and form vessels. These solid growths, for which I wish to pro-
pose the term ‘angiob ast cords,’ usually take the form of nets,
which may persist until the separate strands are hollow, as shown
by the injections of Evans and others, or may, as I hope to show,
disappear in part without ever becoming injectable. The two
views are clearly shown by a comparison of two figures, one from
His (00, 2, fig. 91), the other from Evans (’09, figs. 1, 2, 3) both
showing the caudal end of the aorta of a chick embryo; by the
injection method the capillary network is revealed, while His
represents a network of solid sprouts preceding the hollow vessels.
In this case, since an injection of these so-called solid sprouts would
give practically the same picture as would be obtained if they
were not seen and so left out of the drawing (the network being
similarly placed throughout), we have no direct proof that the
sprouts are not potentially hollow, or in other words merely col-
lapsed; but in the development of the anterior part of the aorta
there are nets of solid angioblast cords present at an early stage,
parts of which have certainly never been shown by injections,
and may therefore, for the present at least, be considered solid.
Here and there in this solid network there are hollow spaces, or
true vessels, unconnected at first with one another and with the
lateral capillary net except by the solid angioblast cords, and
therefore not to be reached by any injection mass from this
lateral net; for such unconnected hollow spaces I suggest the
term ‘angiocysts.’ Thus the angioblast cords retain certain char-

114 JOHN LEWIS BREMER

acteristics of the blood-islands, in that they also change from
solid to hollow independently; but in the angioblast cords there
is no sign of the formation of blood-cells. It was the observance
of these isolated spaces, which later fuse to form large vessels,
that lead to the often repeated statements of Riickert and Mollier
and others that the dorsal aortae arise in situ from the cells of
the mesoderm; and in truth the connection with the lateral
capillary net is short lasting and sometimes extremely tenuous.
Tiirstig (84, 1) recognized the presence of solid cords leading a
short distance from these hollow spaces, but did not trace their
connections; others of this school have missed them entirely.
The story of the development of the primary arterial system
can best be told with the aid of the figures. Fig. 1, a reconstruc-
tion of the angioblast cords of one side of a rabbit embryo of
five segments, shows these cords, streaming in from the network
over the yolk-sac, the cut ends where the reconstruction was dis-
continued showing at the right border of the figure. They have
grown from right to left of the figure, occasionally anastomosing,
until near the median line, which lies at the left of the figure.
- The shape of the meshes of this net indicates, it seems to me, the
direction of this growth, and the rapidity with which it has oc-
curred. Cephalad the net is limited by the proamnion, where no
mesoderm exists; caudad the net continues beyond the portion
drawn. In a few places the cords have become hollow; here and
there near the median line, and especially at the right of the figure,
where a considerable chain of hollow spaces extends longitudin-
ally, near the lateral border of the embryo proper. This chain,
the future lateral heart, lies beneath the coelom, and like the
coelom at this stage is situated only in the anterior third of the
embryonic body. At the left of the figure the net ends rather

Fig. 1 Rabbit, 8} days, 5 segments, 3.4 mm., H.E.C. no. 650. Reconstruction
of the angioblast of the left side, anterior two-thirds of the embryo, seen from the
entodermal cavity. The median line of the body is to the left of the drawing, solid
cords of angioblast come from the yolk sac on the right. Hollow vessels are indi-
cated by the shading. Brackets show position of somites; arrows mark regions
where the aortic network is as yet incompleté; x — y indicates plane of section of
figeZen oe dol:

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 2

115

116 JOHN LEWIS BREMER

regularly, and the meshes are slightly smaller, especially toward
the head. Closer examination will show that each cord extends
from the lateral heart almost directly toward the median line,
then suddenly spreads longitudinally, as though its further direct
course were blocked by some obstruction. In all but two places,
which are marked by the arrows, the longitudinal strands of this
mesial net have anastomosed with others; at these two places
the longitudinal network is interrupted, and we may see clearly
that this part of the net, the future dorsal aorta, is an ingrowth
from the lateral vessels.

WQS
Md, Gx LEE geN
Ge», | Se Rion Pe,
6) @) OLES
RG, og Ses”
; OO
Som Of 2 5 SC 5 SOG bs te
a “ae ZERO Go SSROO
QOS Bwagy.. a
ae gC) AGO
Sion ae od ih oS OBS, OG ge
iS,
Ent. SHES a b

Fig.2 Rabbit, 83 days, H.E.C. no. 650, section 317. Section through the plane

x«-—yinfig. 1. Md.Gr. = medullary groove; Som.=5th somite; Hnt.=entoderm.
a, b, c, ete. = sections of the angioblast network. » 175.

If we examine fig. 2, a section of the same embryo, we can see
these cords in their relation to the germ layers. The embryo has
become shrunken in preservation, so that the layers are separated
slightly from one another. The angioblast cords, indicated by
the small letters, are seen to adhere now to the entoderm, now to
the mesoderm, a fact which made them hard to follow, but which
is obviously the result of this shrinkage. Let us imagine that
the layers are close together. It will be seen that the somite will
then touch both the ectoderm and the entoderm; it is this that
forms the obstruction to further growth of the angioblast cords
toward the median line. Where the somite does not exist, as in
the head region, the cords extend further, and are then halted
by the closely approximated medullary groove and entoderm;

AORTA AND AORTIC ARCHES IN RABBITS Nal 7s

this can be seen by noting the position of the somites in fig. 1.
Just lateral to the somite is the thin neck of mesoderm, the nephro-
tome, which coming between two thicker portions, will be held
above the entoderm, to form the roof of a small longitudinal
‘anal; it is in this canal that the angioblast cords have room to
expand and become hollow (fig. 2, a). In the head region, a
similar though broader canal is provided by the uplifted lateral
edge of the medullary groove; and in this region the reconstruc-
tion shows that the hollow parts of the net are found more widely
distributed; in fact their chain follows forward the line of the
spreading edges of the medullary groove. Thus, as in the case
of the vertebral arteries, a longitudinal trunk is made by the anas-
tomosis of branches from transverse vessels.

Another place where the blood-vessels have an opportunity
to develop freely is beneath the coelom. At this early stage the
coelom is almost exclusively represented by the amnio-cardiac
vesicles, which le cephalad to the level of the somites. Here is
seen the development of the chain of spaces which is to become
the lateral heart; the extension caudally of this is limited by the
absence of the coelom.

Here then, in this early stage of the embryo before the pharynx
has begun to be formed, we see a flat sheet of angioblast cords,
forming a network, lying between entoderm and mesoderm, de-
rived as an extension from the lateral extra-embryonic area. Its
growth is hmited cephalad by the absence of mesoderm in the
proamnion; mesially by neural groove or somite. It occupies
two areas especially where further growth is possible, namely
under the nephrotome or the angle of the medullary groove, and
under the amnio-cardiace vesicle.

Turning now to the next older embryo, one of six or seven seg-
ments, figs. 3 and 4, we can see that two changes have taken place;
a portion of the net has vanished, and the remaining portions
include more hollow vessels, which are still, however, connected
by solid angioblast cords. The persistent parts of the net are
exactly those indicated in the last paragraph as occupying favor-
able positions; beneath the coelom lies the lateral heart, and
beneath the edge of the neural groove hes the dorsal aorta. This

Fig. 3 Rabbit, 83 days, 6 to 7 segments, 3.4 mm., H.E.C. no. 624. Reconstruction of the
angioblast of the right side, anterior third of the embryo, dorsal view. Orientation asin fig. 1.
The coelom is represented as opened to show the lateral heart within. Ao.d. = network to
form the dorsal aorta; Ar./ = first aortic arch; Con. Art.=conus arteriosus; Ve.=venous end
of lateral heart, with vitelline vein entering it. 2x-y indicates plane of section of fig. 4.
Broken line marks limit of medullary groove. X 200.

118

AORTA AND AORTIC ARCHES IN RABBITS 119

reconstruction does not include the region of the segments.
Cephalad, this raised edge, the top of which is indicated by the
dotted line, makes a wide sweep laterally to form the future
optic vesicles, and extends over the region of the coelom; so that
a wide portion of the net may remain here, and may join with
the vessels under the coelom. Thus the first aortic arch is formed,
a net of vessels and cords connecting the dorsal aortic net with
the anterior end of the lateral heart.

agers at cela Oo IEG

— 2 Sn
‘WPOoSTee

Fig. 4 Rabbit, 8} days, H.E.C. no. 624, section 101. Section through plane
x-— yin fig. 8. Md.Gr., medullary groove; Coe., coelom; Ao.d., three strands of
network for dorsal aorta; a, isolated portion of angioblast; b, part of network; Ve.,
lateral heart. > 175.

Of the original net of angioblast cords between the lateral heart
and the aortic net almost nothing remains. Here and there a
few isolated cells can be found, not connected with the rest of
the network; one of these is shown in fig. 4, others, though recog-
nized, were left out of the reconstruction for the sake of clearness.
Other cords have also been lost; of those connecting the lateral
heart with the extra-embryonic angioblast only the most caudal
remains, the others have either entirely disappeared or have lost
their lateral connections. The cause of this latter destruction
of the net is shown in fig. 4; the lateral heart, covered by a reflec-
tion of the mesoderm, has expanded so far into the coelom that
vessels from it to the lateral net must take a curving course, and
would probably be compressed between mesoderm and entoderm.
This has occurred progressively from before backward, until the
lateral heart gradually leaves the region of the coelom; here no

120 JOHN LEWIS BREMER

such influence is brought to bear on the angioblast cords, and
here they enlarge and become the vitelline veins.

Another agency at work in the further development of these
blood-vessels is shown in fig. 4. The shape of the layer of ento-
derm indicates a longitudinal folding of this layer to form the
pharynx, of which a point near ‘a’ is to be the lateral edge, and a
union of the layer near ‘6’ with a similar point on the opposite
side of the embryo is to complete the floor. This fold ends ante-
riorly by curving to the median line under the network of the first
aortic arch. Thus the pharynx is due, in the rabbit at least, not
so much to the usually described pouching forward of the ento-
derm as to a lateral folding of the layer, and the floor of the phar-
ynx is completed by a union of the entoderm of the two sides,
which soon fuses and forms two continuous sheets, one for the
floor of the pharynx, the other for the upper wall of the archen-
teron.

This folding in a more advanced stage is shown by the shape
of the blood-vessels in figs. 5 and 6, since the vessels always lie
close to the entoderm. The dorsal aortae, still showing, by their
frequent subdivision, signs of their origin from a network of
vessels, are in the same relative position as before, dorsal to the
entoderm of the pharynx. The lateral part of the network has
been rolled in underneath the pharynx, whose crescentic outline
is marked by the plexus of vessels which forms the first arch. By
this folding or rolling in process the lateral edge of the network
now lies beneath the pharynx and near the median line, and as the
mesoderm makes its way between the floor of the pharynx and the
roof of the archenteron, new shoots from these vessels pass toward
the median line, and may even anastomose with others from the
opposite side. This ventral plexus of vessels, many of which are
at first solid cords, is the first indication of the ventral aorta.
It is connected with the lateral heart, as can best be seen in
fig. 6, by means of a slender vessel, the conus arteriosus, or bulb;
and the lateral heart has, so to speak, lagged behind in the fold-
ing, so that the curve of the blood-vessels and of the entoderm in
transverse sections of the embryo makes an S, the upper curve
of which comprises the dorsal aorta, first arch, and ventral aortic

Ao. d.

os Co nArt.
g}

Ve.

Fig. 5 Rabbit,9 days, 8 segments, 3.2 mm., H.E.C.no.621. Reconstruction of
the blood-vessels of the head end of the embryo; seen as though looking forward
from the anterior end of the lateral hearts, to show network for first arch. <Ao.d.,
dorsal aortae; Con. art., conus arteriosus; Ve., lateral hearts; Ao.v., network for
ventral aorta; y, extension toward median line; x, blind ends of obliterated ves-
sels from yolk-sac. > 125.

Fig.6 Rabbit, 9 days, 8 segments, 3.2 mm., H.E.C. no. 709. Reconstruction
and lettering similar to those in fig. 5. At y the extension of the ventral aortic
network has fused with that of the other side. X 125.

Fig. 7 Rabbit, 9 days, 11 segments, 3.8 mm., H.E.C. no. 619. Reconstruction
and lettering similar to those of fig. 5. Ar.1, Ar.2, first and second aortic arches.
125:

121

2? JOHN LEWIS BREMER

plexus, while the lower curve includes conus arteriosus and lateral
heart. The upper curve corresponds with the pharynx, the lower
is below the pharynx and associated with the coelom.

If we notice the position of the lateral heart in figs. 4 to 7 we
can see it is gradually rolled over and inverted, at its anterior end,
so that whereas in fig. 4 it projected dorsally into the coelom and
therefore was connected with the other vessels ventrally only,
in fig. 6 the projection is distinctly lateral. In fig. 7, from a rab-
bit of eleven segments, the lagging lateral hearts have ultimately
met in the median line and partly fused, but not until the pharynx
floor has been completed and the ventral aortic network established,
so that the inverted and fused lateral hearts le at a distinctly
lower level, connected with the ventral aortae by the ascending
conus arteriosus, which itself is composed of two fused halves.
The blood-vessels of this part of the embryo now lhe in three
tiers or levels; dorsally, the dorsal aortae, ventrally the heart,
and between these two the ventral aortic plexus, joined to the
former by the first arch and to the latter by the conus arteriosus.

On the recognition of this middle tier, the plexus of the ventral
aortae, depends the proper understanding of the development
of the rest of the aortic arches and of the pulmonary arteries.
If we turn back to fig. 3, we shall see that the three levels are
already indicated in the angioblast net; the dorsal aorta, in plexus
form, occupies a distinct region toward the median line, the lateral
heart lies toward the right of the figure, while, connected with this
by the conus arteriosus and with the dorsal aorta by the first
arch, a portion of the net remains, expanded longitudinally, and
lying beneath the mesial border of the coelom. ‘This is to form
the ventral aorta, and is to lie in the floor of the pharynx, while
the conus is to lead from this level to the heart, which occupies a
more ventral position.

It will be noticed that in fig. 3 there is an extension of the ven-
tral aorta caudad from the conus arteriosus. This may be pre-
cocious in this case, for in the two embryos of eight segments,
figs. 5 and 6, no such extensions were found, while in the embryo
of eleven segments it is again present. This caudad growth of the
ventral aortic net lies in the thin sheet of mesoderm between the

AORTA AND AORTIC ARCHES IN RABBITS 123

floor of the pharynx and the pericardial cavity, and ultimately
reaches the level of the pulmonary anlagen.

THE AORTIC ARCHES

The first aortic arch has already been described as a persistent
portion of the original angioblast net, folded around the anterior
end of the pharynx. The cords of this net become hollow, as is
shown by Evans in the drawing of the injected vessels of a duck
embryo (Keibel-Mall, vol. 2, fig. 398). Later, as is seen in figs.
7 and 8, the net becomes reduced to two vessels on each side, and
in the rabbit frequently breaks up into capillaries before being
reduced to a single trunk.

The second aortic arch (figs. 7 and 8) arises as a lateral exten-
sion from the plexus of the ventral aorta, frequently double on
one or both sides, consisting at first of solid cords and hollow
spaces, and met by much shorter outgrowths from the dorsal
aorta. The potentially double character of this arch, even
after it has attained a considerable development, is shown in fig.
8, x and y. While the second arch is becoming established the
plexus of the ventral aorta is extending still further caudad, and
again giving off lateral branches, which run between the entoder-
mal pouches to the dorsal side of the pharynx, and again are met
by shorter growths from the dorsal aorta. Thus the third and
fourth arches are formed and, as a glance at fig. 8 will show, they
also, from their plexiform origin, are potentially multiple on each
side.

With the growth of the pharynx the conus arteriosus, which
joined the ventral aortae originally cephalad to the second arch,
has moved caudad, and later opens almost directly into the third
and fourth arches. This condition is represented in fig. 9. Here
still we see the plexus of the ventral aorta extending caudad from
the ventral part of the fourth arch, as it did earlier from the sec-
ond arch. Its plexus formation is easily recognized, as it lies
in the thin sheet of mesoderm between the pharynx and the peri-
cardial cavity, but it no longer crosses the median line on account
of the presence of the keel-like growth of the pulmonary anlage
from the ventral wall of the pharynx. From the dorsal part of

124 JOHN LEWIS BREMER

Con. Art.

Fig. 8 Rabbit, 10 days, 23 segments, 3.2mm., H.E.C.; no. 940. Reconstruc-
tion of the blood-vessels of the head end of the embryo, seen from the ventral side ;
the heart removed by cutting through the conus arteriosus. Lettering same as in
fir 5, andealso Ag, Arz2, Air.3,.An-4; aortic arches; #.C., external carotid arte-
ries, cut; « and y, remains of second channel for second aortic arch. 125.

the fourth arch and from the dorsal aorta sprouts arise, at first
in part solid, which curve around the pharynx and, entering the
same sheet of mesoderm in which the ventral aorta lies, form there,

AORTA AND AORTIC ARCHES IN RABBITS PAS

on the left side of the embryo, an anastomosing plexus. On the
other side of this embryo these sprouts from the dorsal aorta are
much simpler, though still double, while a lateral extension of the
plexus of the ventral aorta replaces that derived, on the left, from
the dorsal sprouts. In other embryos there is great variation in
these vessels. There can be no doubt that these vessels from the
dorsal aorta are about to join the ventral plexus and form another
arch; nor can there be any question that in this case there are two

Fig. 9 Rabbit, 103 days, 3.2 mm., H.E.C., no. 560. Reconstruction and let-
tering similar to those of fig. 8. P.art., extension of ventral aortic network as
pulmonary arteries. >< 125.

channels on each side. While not wishing to go deeply into the
controversy over the presence or absence of a sixth aortic arch, I
may say that it seems to me that the final solution should come
from a further study of the entodermal pouches, of the branches
of the nerves, and of the cartilages in this region. We have seen
that the four first arches are potentially double on one or both
sides, and we now find that the last arch is so also, with, as is

126 JOHN LEWIS BREMER

known, great variation in the points of origin and forms of
anastomosis. We know that the fourth pouch is forked at the
end, but so is also the third pouch. Branches of the vagus have,
in a few instances, been found which seem to indicate the pres-
ence of an extra arch; but aberrant nerve branches are not
unknown elsewhere. As far as the early development of these
vessels is concerned, there is nothing certainly to prove the pres-
ence of an interpolated arch.

THE PULMONARY ARTERY

It will be noted that the sprouts for this last arch arise chiefly
from the dorsal vessels, instead of from the ventral net. I also
wish to point out that the net grows beyond the arch, before
the arch has become complete. In other words this extension of
the ventral aortic net forms well defined pulmonary arteries, one
on each side, before the pulmonary arch exists; the pulmonary
artery is in no sense a branch of the pulmonary arch, and more-
over, in the strictest sense, the arch extends only from the dorsal
aorta to the pulmonary artery, the ventral part of the vessel
usually called the arch is really the ventral aorta. The persist-
ent pulmonary arteries are entirely ventral; they have been
joined during embryonic life by branches from the dorsal aorta,
but such branches are only temporary.

Here I must add a few words in regard to some recent state-
ments on the development of the pulmonary arteries. Evans
(09, fig. 21) gives a figure of the injected pulmonary arteries of a
pig embryo of 12 mm., and in the German edition of the Keibel-
Mall Embryology he copies (fig. 396) a figure of Fedorow showing
the pulmonary arteries of a guinea-pig embryo of twenty-one
days. In both cases the vessels form a narrow plexus in front of
the trachea, and Evans? concludes that here, asin other growing
vessels, the main trunks are preceded by a capillary net. That
he is correct in the main, that the pulmonary arteries do arise as
the extension of the net of the ventral aorta, we have just seen;

2 Dr. Evans hasvery readily and kindly acknowledged his error to me; Fedorow
did not make the misinterpretation.

AORTA AND AORTIC ARCHES IN RABBITS 127

but these plexuses of the pulmonary arteries across the median
line are of later occurrence, after much mesoderm has grown in
between the trachea and the pericardial cavity. It is unfor-
tunate that the pig and the guinea-pig were chosen as illustrations,
for, as I have shown in previous papers (’02, ’09), in these two
species only, so far as I am aware, is such a secondary net found.
After the formation of this net one of the connections with the
pulmonary arches is lost, so that in these two embryos the trunk
of the pulmonary arteries seems very long before the right and
left branches are given off. In those two papers I spoke of the
pulmonary arteries as branches of the pulmonary arches, a
statement which I am now very glad to correct.

SUMMARY

In the rabbit, the dorsal aorta, the first aortic arch, the conus
arteriosus and the lateral heart are all parts of an original net-
work of angioblast cords derived from the extraembryonic plexus
of blood-vessels. Those portions of this network which are me-
chanically favored in their position persist; the other portions dis-
appear. The favored portions lie (1) under the coelom, (2) under
the nephrotome, or (3) under the raised edge of the medullary
eroove. The connection between first arch and lateral heart
is permitted by the lateral expansion of the medullary groove
which extends over the coelom.

This net of angioblast cords is folded in the formation of the
pharynx, so that its lateral edge, anterior to the lateral heart,
becomes the ventral aorta. In this folding the lateral heart is
retarded, and thus comes to lie on a lower plane, more ventral.
The connection between the two planes is by means of the conus
arteriosus.

From the net of the ventral aorta a plexus grows mesially and
caudally, still forming the ventral aorta. Lateral growths from
this pass around the pharynx, often in plexus form, and make
the second, third and fourth aortic arches. The conus arteriosus
is moved caudad to a position opposite the fourth arch.

A further extension of the plexus of the ventral aorta, situated
between the floor of the pharynx and the dorsal wall of the peri-

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 2

128 JOHN LEWIS BREMER

cardial cavity, but prevented from crossing the median line by
the presence of the median pharyngeal outgrowth to form the
trachea, extends to the lungs as the pulmonary arteries, which
are later joined by vessels springing from the dorsal aortae.
These vessels, which may be double and plexiform, constitute
the fifth (and sixth) arch.

Although dealing in this paper with the development as found
in rabbit embryos, I have examined various other species, as
chick, pig, sheep, etc., and feel satisfied that in all essential points
the story of the development of these primary vessels in other
vertebrates will be found similar to that here described.

LITERATURE CITED

BREMER, JOHN L. 1902 On the origin of the pulmonary arteries in mammals.
Am. Jour. Anat.,.vol. 1, p. 137, and Anat. Rec., vol. 3, p. 334.

Evans, H. M. 1909 In Keibel-Mall Entwickelungsgeschichte, vol. 2, p. 551,
etc. Anat. Rec., vol. 3, p. 498.

Frporow, V. 1910 Ueber die Entwickelung der Lungenvene. Anat. Hefte.
Bd. 40, 8. 529.

j His, W. 1900 Lecithoblast und Angioblast der Wirbelthiere. Abh. d. math.-

phys. Klasse d. Kgl. Siichs. Ges. d. Wiss. Bd. 26, no. 4. Leipzig.

Rickert u. Mouuier 1906 Die Entstehung der Gefiisse und des Blutes bei
Wirbelthieren. Handb. d. vergl. u. expt. Entw. d. Wirbelthiere, her-
ausg. von O. Hertwig. Bd. 1, S. 1019.

Tirstiac, J. 1884 Untersuchungen iiber die Entwickelung der primitiven Aorten.
Schrift., herausg. v. d. Naturf.-Ges. bei. d. Univ. Dorpat. Bd. 1.

THE BEHAVIOR AND RELATIONS OF LIVING CON-
NECTIVE TISSUE CELLS IN THE FINS OF FISH
EMBRYOS WITH SPECIAL REFERENCE TO THE
HISTOGENESIS OF THE COLLAGINOUS OR WHITE
FIBERS

JEREMIAH 8. FERGUSON

Assistant Professor of Histology
Cornell University Medical College, New York City

TEN FIGURES

In the process of connective tissue development the cells first
arise, the fibers later appear. This sequence is established be-
yond controversy. The ontogenetic relation of cell and fiber is
not, however, so thoroughly established. The theories advanced
may be grouped under three heads: (1) Intra-cellular origin, the
cells may transform into fibers (Schwann, Valentin, Boll, Flem-
ming, Spuler, Livini). (2) Extra-cellular origin, the fibers arise in
the intercellular substance by its fibrillation, or possibly as a
secretion from the cells (Henle, Merkel, Virchow, KoOlliker).
(3) Epicellular origin, the fibers form in an ectoplasm at the sur-
face of the cell (Schultze, Hansen, Golowinski, Mall).

These theories have all been primarily founded upon the results
of examination of ‘fixed’ or ‘killed’ tissue, or upon the study of
fresh teased tissue. Living connective tissue has been studied in
the mesentery of the frog and other animals under conditions
which are accompanied by marked inflammatory reaction and
certain stages of the formation of exudates and of scar tissue
have been thus investigated, and more recently movements of
connective tissue cells have been observed in tissue cultures but
so far as I know the theories of the histogenesis of connective
tissue have not been examined with reference to the behavior of
living cells under normal conditions.

129

130 JEREMIAH S. FERGUSON

Living connective tissue cells have been seen in tissue cultures
by Harrison, Burrows, Carrel and Burrows, Margaret R. and W.
H. Lewis and others, to exhibit a certain motility, and Harrison
has recently emphasized the stereotropic tendency of connective
tissue cells in cultures when in contact with foreign surfaces,
glass, spider-web, ete. But so far as I know, the histogenesis of
connective tissue fibers has not been so studied, and at best the
culture method is open to some criticism on the ground that while
the connective tissue cells are undoubtedly alive and active, yet
they exist under very unusual, if not abnormal, conditions whose
effects have not yet been subjected to complete analysis. Under
these conditions the behavior of the connective tissue elements
while probably similar, is not certainly in exact conformity with
that of the tissue within the embryo.

In order that deductions based upon these several methods of
examination be adequately controlled it appeared desirable that
developing connective tissue be studied in the living animal under
conditions which were in every respect normal, or which, at least,
resulted in no inflammatory reaction. In mammals this endeavor
is fraught with considerable difficulty owing to the size of the
mammalian embryo and the depth beneath other tissues, often
not transparent, at which the connective tissue lies.

During the past summer I had the opportunity, through the
courtesy of the Marine Biological Laboratory at Woods Hole, of
studying connective tissue in the fins of living fish embryos under
conditions which were wholly normal and unaccompanied by
any evidence of inflammatory reaction.

If a free swimming Fundulus embryo is placed on a hollow
ground slide it will continue to swim, often actively, and its heart
beat and circulation are maintained. It may be observed for
some minutes and at the end of observation may be returned to
the aquarium to continue an uneventful existence for hours or
days thereafter. If a drop of chlorotone is added, or frequently
without its addition, the fish will remain quiet for some minutes,
thus permitting continued observation of connective tissue cells
in his semitransparent fins. Certainly cells studied under these
conditions are open to no criticism of abnormality.

LIVING CONNECTIVE TISSUE CELLS 131

The viability of the animals is unaltered for I have kept them
for several days after such observation without any indication of
decreased activity on their part. Even embryos which have been
quieted by chlorotone, as well as those immersed for hours in a
solution of Bismark brown in sea water, I have resuscitated and
kept alive and in an apparently normal and usual condition for
two or three days; they could easily have been kept longer had it
seemed advisable.

The tissue selected for observation was in the fins of free swim-
ming pelagic and Fundulus embryos. The embryos studied were
chiefly of Fundulus and varied from 5 te 20 mm. in total length.
The most favorable subjects were from the time of hatching, 5
to 6 mm., up to 12 mm. in length. The pectoral and caudal fins
were usually selected as most available for observation. In
such embryos the fin consists of a central frame work formed by
the jointed rays, lepidotrichia, with their attached muscles, and
a superficial integument of pavement epithelium with its sub-
jacent basement membrane. ‘The finer fin rays, actinotrichia,
continue the jointed rays to the margin of the fin. The fin at
this stage is very thin and the epidermis lies almost in contact
with the fin rays. But between adjacent rays is an interval which
lodges on either side the afferent and efferent blood vessels, bor-
dered by chromatophores, and between them a loose mass of mes-
enchymal connective tissue in which the cells may be readily
observed.

In embryos 5 to 6 mm. long the connective tissue in the pectoral -
fins consists chiefly of a mass of round cells confined to the proxi-
mal portion, and beyond this mass a distal fringe or ‘skirmish
line’ of scattered stellate cells. In the unpaired fins, which are
less advanced in their development only the scattered stellate
cells are represented, the invasion of the round cell mass having
not yet occurred. In later stages, as in the caudal fin of the same
embryo, the zone of round cells has advanced distalward among
the actinotrichia nearly to the fin margin, leaving behind between
the lepidotrichia an area of more mature cells, stellate and spindle,
and a few fine fibers well separated by broad spaces occupied by
tissue fluids. The spindle cells and fibers preponderate in the

Laz JEREMIAH S. FERGUSON

proximal, the round cells in the distal zone of the fin. Hence,
one follows the sequence of development in passing from the dis-
tal toward the proximal portion of such a, fin. Older embryos
show the same zones of transition but in them the formation of
fibers in the proximal region is more advanced.

In mammalian tissue one finds three stages in the histogenesis
of connective tissue, a primitive cellular stage, a syncytial stage,
and a fibrous stage. The first is characterized by the predomi-
nance of round cells, the second by stellate, the third by spindle
and lamellar cells. The same succession of cell types is present
in the finsof embryo fish and there is a corresponding succession of
histogenic stages. Fibers do not appear prior to the appearance
of cellular processes. Fine fibers appear coincidently with stel-
late cells, coarse fibers and fiber bundles develop later.

In the distal portion of the fin fine fibers first appear in the round
cell area coincidently with the transition from round to early
stellate forms. At exactly this period I have observed the first
indication of motion, the throwing out of pseudopods by the round
cells, in the connective tissue cells of the living embryo. Fig. 1
-shows such changes in two cells on the border of the round cell
area near the posterior end of the ventral fin. There is at this
time relatively little locomotion, as is shown in the figure by
referring the position of the cells a and b to the relatively fixed
point, a prominence on the margin of an adjacent chromatophore
(ch).

The first appearance of fibers in the distal portions of the fins
has been very properly connected by Harrison, and by Goodrich
with the origin of the dermal fin rays from the ‘scleroblast’ cells
which closely resemble the connective tissue cells and like them
are of mesodermal origin. In the region of the actinotrichia
in the distal portion of the fin, it is difficult to distinguish between
the early forms of these coarse fibers and the true connective tis-
sue fibers, but the actinotrichia are confined to the region of the
last one or two joints of the jointed fin rays, and there they pro-
ject, as Goodrich has shown, from between the two opposed
dermal plates which form the distal section of the jointed fin ray.
If therefore one studies a region proximal to the last section and

LIVING CONNECTIVE TISSUE CELLS 135

selects the interval between the jointed rays the primitive actino-
trichia are thereby excluded.

In such portions of the caudal fins of 6 mm. embryos, and in
equivalent places in later stages, are typical connective tissue
fibers mostly occurring as coarse longitudinal bundles with fine
oblique anastomoses. Single fibers occur in the intervals of the
coarser bundles. It is along these fibers and fiber bundles that
the stellate and spindle cells are disposed. These cells are readily
seen in the living fish, though the ease of observation is subject
to much variation in different individuals and to a less extent in
different portions of the same embryo.

My observations were made on living embryos immersed in sea
water, some with, some without the addition of chlorotone. In
some cases a few drops of a saturated solution of Bismark brown
were added to the sea water in which the fish was kept, the effect
of which after a time was to slightly increase the color contrast
between the connective tissue cells and surrounding structures.
The stain seemed almost inocuous, for fish could be kept in it
for several days without apparent effect on their vitality. Many
of the fish thus examined were later killed, and the fins stained
and mounted in toto, or sectioned. The various cell types seen
in life were readily recognizable in corresponding locations in the
stained preparations.

It is in life difficult or impossible to distinguish between the
spindle and lamellar types, though in ‘fixed’ tissue they may be
morphologically distinct. In the living animal one can see a
stellate or a spindle cell elongate, approach and flatten itself
against a connective tissue fiber or fiber bundle, becoming some-
times so attenuated as to be scarcely distinguishable from the
fiber against which it lies; it may at any time acquire increased
thickness. Such a relation to a connective tissue fiber is shown
by the cell 6 in fig. 2. The relationship is again exhibited by the
two cells shown in fig. 3, one of which a, approached a small
fiber bundle, became flattened against it, then rotated to the oppo-
site side of the fiber at 9.30 a.m., and later freed itself from the
contact. Its locomotion can be observed in relation to the chro-
matophore (ch) which served as a fairly fixed point. Similar

134 JEREMIAH S. FERGUSON

cells are frequently seen flattened against the surface of fiber
bundles, blood-vessels, or fin-rays, and exhibiting a slow stereo-
tropic locomotion. Many of these cells would seem to be identi-
cal with those which in stained preparations we are accustomed
to call lamellar cells.

That connective tissue cells exhibit a certain amount of motion
is no new observation. It has been well known since the inflam-
matory reaction to injury or infection was studied in the mesen-
tery by Arnold and others. I have observed that the extent and
rapidity of the motion varies with the cell type. The round cell,
or primitive type, presents relatively little motion, it being limited,
so far as I have observed, to the very slow projection and retrac-
tion of minute pseudopods. Even this evidence of activity seems
rather to be limited to those later phases of the cellular stage which
foreshadow the transformation of the round cells to the stellate
type of the succeeding stage. This transformation is indicated
by the fact that the motion is more noticeable near the border
of the round cell area than in its interior, and also because at the
extreme margin of such a cellular area one may by careful scru-
tiny observe an extensive alteration from round to stellate types,
some cells passing rapidly to an approximate spindle form. The
type of motion exhibited by the round cells, when observable,
is well shown by fig. 4, cells a—c being observed at the extreme
margin of the round cell area, cells d-e just within the margin,
and cells f—g well in the interior of the area.

While the general trend of cell change is from round to stellate
to spindle cells, a change may often be observed to occur in the
reverse direction, as occurred to the cell shown in fig. 5, and that
in fig. 6. Such retrograde changes are less frequently observed,
and the transformation is less extensive than are the progressive
changes from the round to the stellate forms. The retrograde
stellate phase is also more frequently of a transient character
(fig. 5). Thus, a stellate cell may by retraction of its processes
temporarily assume a spheroidal form but it soon again projects
pseudopods and regains its stellate character. Or a typical,
bipolar, spindle shaped cell may extend a third process, or even
several additional processes (figs. 2, 5 and 6), but, so far as I have

LIVING CONNECTIVE TISSUE CELLS 135

observed, such processes are limited in size and usually of short
duration. This reverse transformation may be likened to an elas-
tic rebound brought about by an inherent resistance to change of
form reacting against an impelling foree which directs the trans-
formation from the round to the spindle type. The cell frequently
balks at the change, but the general trend from round to stellate
and from stellate to spindle form is inevitable.

Motion resulting in change of form is perhaps most active in the
stellate type of connective tissue cell. The general trend of this
motion seems to be indicated in fig. 5 J, in which a typical round
cell selected for observation at the margin of the round cell mass
in a pectoral fin of a 6 mm. embryo was seen within a period of
six minutes to elongate and then to pass through successive stel-
late shapes to a typical spindle form. But the succession is not
always so rapid. Stellate cells exhibit all sorts of morphological
transformations in rapid sequence (fig. 7) and this stage of con-
nective tissue development is of relatively more transient duration
than either the preceding or the succeeding stage. Moreover, the
shape of the cell is undoubtedly influenced to some extent by
its surroundings and the duration of a particular stellate, spindle
or lamellar shape may in some cases be thus determined.

Likewise, spindle cells undergo considerable transformations
in form, the most frequent of which undoubtedly result in the
lamellar shapes on the one and in the stellate on the other hand.
Because of the limitations of the microscope in the delineation of
the ‘third dimension’ it is most difficult in the colorless living
tissues to differentiate between the lamellar and spindle types of
cell but the evidence of fixed and stained tissues shows the lamel-
lar to be the more mature, the spindle the earlier type, and I
have observed nothing in the living tissues to indicate the con-
trary unless indeed it be that both types appear to be somewhat
dependent on their surroundings, for as already stated these forms,
in the same cell, seem to be more or less interchangeable. That
spindle cells frequently and freely revert to the stellate type there
is abundant evidence. There is also evidence that these cells may
be capable of still further transformations than those of mere form.
A syncytial stage in the development of connective tissue has

136 JEREMIAH S. FERGUSON

long been assumed. That this stage in its most typical form
presents those cell pictures which we are accustomed to regard
as stellate cells is well known. It is generally recognized that
this syncytial stage passes into one in which the fibers appear and
the syncytium is replaced by a tissue of cells and fibers. The
syncytial stage has been presumed to be preceded by a cellular
stage and to those who have traced the origin of the mesoderm
from the time of egg fertilization it would appear logical, even
necessary, that at a sufficiently early period a cellular character
must obtain, though Mall has questioned the preéxistence of this
cellular condition. The transformation from the cellular to the
syncytial condition has been ascribed on the basis of stained sec-
tions, to either of two processes: either the syncytium arises by
incomplete division of preéxisting cells or the syncytium results
from the fusion of the preéxisting cells. That some syncytia arise
by incomplete cell division is very probably true. This appears
specially obvious in such placental tissues as the superficial cells
of the chorionic villi. I know of no convincing evidence that it
does occur in the connective tissues.

Since I have been unable to observe mitotic figures in the living
connective tissue cells of the fish which are under discussion I
cannot offer any evidence pro or con the origin of a connective
tissue syncytium by incomplete cell division. I have, however,
frequently observed a phenomenon which simulates the fusion of
processes of adjacent stellate cells after the manner of a typical
connective tissue syncytium. In figs. 2 and 8 JI the neighboring
cells, which were at first entirely distinct and separate, were within
a brief period seen to send out processes which on contact appar-
ently fused. But of course one cannot say without subsequent
fixation and staining of the identical cells, a process presenting the
greatest difficulties, that the fusion was actual and complete.
Even in stained sections the question is often difficult to deter-
mine. While the fusion was apparent I am not at all sure that it
was actual. Not, however, in every case when cell came into con-
tact with cell did such apparent fusion occur. ‘This is shown in
fig. 8 I, in which processes from the cells a and 6 came into contact
tip to tip, yet though fusion seemed imminent it did not occur

LIVING CONNECTIVE TISSUE CELLS Tae

and the contour of each cell at the point of contact remained clear
and distinct. Moreover it-would seem that since connective tis-
sue cells move extensively along the surfaces of the syncytium
that syncytium could scarcely arise by fusion of its cells.

The spindle cells exhibit a certain stereotropism. They are
prone to take their position alongside a connective tissue fiber or
fiber bundle or against the surface of a blood-vessel or dermal
fin-ray. When in contact with a broad surface, such as that of a
blood-vessel or one of the lepidotrichia, the cells frequently assume
a flattened, lamellar form. This is shown by stained sections, in
which that type of cell predominates in these locations, and by the
observation of living spindle cells which frequently move up to
a blood-vessel or a fin-ray and then become so thinned out against
the surface that they finally vanish, being in the living tissue in-
distinguishable from the refraction lines which surround the larger
bodies. Again, the spindle cells very frequently move up to a
connective tissue fiber or bundle and then elongate along the nar-
row filament until, as before, the cell finally appears to vanish by
its extreme attenuation. Such a result was observed a moment
later than the recorded observation in the case of the cell shown in
eo:

It frequently happens that the spindle cells after such elongation
again thicken to a typical spindle form, and may even throw out
other processes, but in so doing, if the cell is observed in relation
to some relatively fixed point, e.g., a joint of a dermal fin-ray, a -
chromatophore, or a blood-vessel, it will be seen that the cell has
changed its relative position; it has exhibited locomotion. Loco-
motion is not a distinguishing character of the spindle cell; it is
exhibited by the stellate cells, possibly also to a very limited
extent by those round cells which are only just beginning to pre-
sent pseudopod formation. But the character of the locomotion
in the several types of cells differs decidedly. In the stellate
type locomotion may take any direction and resembles a very
active amoeboid motion, processes being extended along the sur-
faces of fine fibers, then either retracted or increased in size until
the whole cell has come to occupy the place of the former process.
Though locomotion in the stellate cells is not entirely confined

138 JEREMIAH S. FERGUSON

to the direction of visible fiber lines, yet a projecting process of
such a cell often appears to envelope or to become coincident with
a fiber. In the spindle cells locomotion is always so far as I have
observed, in the direction of the fiber lines: usually these cells
merely slide along the surface of fibers, blood-vessels and similar
structures.

I have observed that the stellate cells are more prone to le in
relation with the finer, the spindle cells with the coarser fibers;
the coarser fibers in most cases, because of their size, being pre-
sumably fiber bundles rather than single fibers. This relation-
ship is to be expected in as much as in stained preparations one
finds the stellate cells present with those finer fibers which repre-
sent the earlier stages in fiber formation.

That fibers do lie without the cell in both embryonic and mature
connective tissue is generally conceded. That they he within the
cell in reticular tissue, which in a way is comparable to an early
or embryonic type of connective tissue, I have recently demon-
trated by means of the Bielschowsky stain.t. The types of fiber
development by fusion of intracellular granules described by
Spuler and by Lavini though perhaps not conclusively demon-
strated, at least show that certain granules which are in relation
with the first appearance of fibrils do lie within the substance of
the stellate, mesodermal, connective tissue cells. Moreover, I
have found in embryonic tissues (fig. 9) Just such appearances as
I have described for reticular tissue.2. By means of the Biel-
schowsky method such appearances can be shown throughout
embryonic connective tissue. I have observed them in pig
embryos, of various ages, in the limb buds, the head, the cervical
region, and in the back throughout the whole length of the embryo
from the occiput to the caudal tip, also in the umbilical cord. In
many of these locations Ihave made similar observations on human
embryos of older stages but in which the connective tissue was
still actively developing. One is at a loss to explain the method
by which fibers arising within the cells arrive at a location out-
side the cell body when these cells are in active motion. The

1 Am. Jour. Anat., vol. 13, page 277, 1911.
2 Loe. cit., in which see especially fig. 4 and fig. 8, pages 285 and 289.

LIVING CONNECTIVE TISSUE CELLS 139

ectoplasmic theory of Hansen does not satisfactorily account for
it and its elaboration by Mall is not as specific in this particular
as one might wish. These theories do not appear to fully har-
monize with the relatively active motion and locomotion of the
connective tissue cells which I have observed in the fins of living
fish and which Harrison, Burrows and others have also to some
extent recognized in tissue cultures. The cells are not suffi-
ciently quiescent to permit of endoplasmic retraction with deposit
of ectoplasmic fibers unless this retraction is rapidly performed,
in which case it should be observable in the living embryo. I
have in one or two cases suspected such a method of deposit but
have notas yet been able to convince myself that it actually occurs;
in fact I now doubt if it occurs at all.

The ectoplasmic theory presupposes that the fibers arise at the
surface of the cell. This I have found to be not always the case.
The clear delineation of fibers by the Bielschowsky method makes
it possible to follow their course within the cell more carefully
than ever before and I find that the blackened fibrils within the
cell both in pig embryes (fig. 9)and in the fish’s fin very frequently
pass close to the nucleus, sometimes ending almost in contact with
this structure, but more frequently passing by so closely as to be
in actual contact with the nuclear membrane. I am aware that
Golowinski using the iron haematoxylin method, demonstrated
the presence of fibers at the surface of the connective tissue cells
of the umbilical cord and that the apparent relation to the nu-
cleus was explained by him as due to obliquity of section. But I
have not in my preparations been able to convince myself of the
adequacy of this explanation. I have found fibers to be not
always at the surface of the cell, they may and frequently do pene-
trate entirely through the cytoplasm of the cell, as I have previ-
ously described for mature reticular tissue. In the developing
connective tissue, as well as in reticular tissue, such penetration of
cells by the fibers is so frequent as to appear quite characteristic.
It seems to me that the intimate relation of connective tissue cells
and fibers in embryonic tissues can only be accounted for by ‘tak-

3 Loe. cit., see fig. 10, page 293.

140 JEREMIAH S. FERGUSON

ing cognizance of the plasticity of the connective tissue cellular
cytoplasm, and also of the active motion of connective tissue cells
during the period in which the fibers are being formed, so that the
finer connective tissue fibers become, by the cellular activity,
embedded in the plastic cytoplasm of the cells during their stereo-
tropic locomotion. The plastic character of the cellular cyto-
plasm is admirably shown by the rapid changes in form of the
connective tissue cells in the fins of living fish embryos.

I have already stated that the spindle cells of connective tissue
in the fins of living fish undergo active locomotion. In fact this
seems to be a most prominent function of the spindle cell type.
Most frequently the cell glides along connective tissue fibers which
often appear to be thus partially enveloped by the cytoplasm.
In recording this stereotropism I am able to corroborate, for the
living cells of embryo fish, the observations of Harrison on tissue
cultures in which he finds that the connective tissue cells are spe-
cially prone to follow along the surface of fixed objects. Such ob-
jects in normal living subjects are most frequently the connective
tissue fibers and fiber-bundles already deposited, though as pre-
viously stated, I have also observed connective tissue cells moving
along the surface of the dermal fin rays and of blood-vessels.
In this form of activity the cells adapt themselves more or less to
the shape of the surface along which they are moving. They
wrap themselves about or rotate around the finer fibers (fig. 3)
and they flatten themselves against the larger objects (figs. 2,
3and 5). In this attenuated condition they still move along the
surface of fibers, often at a considerable rate of speed. One such
cell I have recorded in a preliminary communication! was found in
ten minutes to have covered a distance of 50u, a rate of Ip in
every twelve seconds.

The striking similarity of the living, spindle, connective tissue
cells to those of fixed tissue is indicated in fig. 10 which shows
several such cells from a 100 mm. pig embryo. The magnifica-
tion is the same as that used for the observation of the living cells.
The-similarity in the form of the cell and the relation of cells to
fibers is apparent on comparison with the preceding figures. One

4 Biol. Bull., vol. 21, page 272, fig. 2, 19J1.

LIVING CONNECTIVE TISSUE CELLS 141

ean scarcely avoid the interpretation that the cells shown in fig.
10 were at the moment of ‘fixation’ moving along the surface of
the fiber bundles against which they le.

The pronounced morphological relation between the connective
tissue cells and fibers cannot but have an equally close functional
relation. What those functional relations may be we are not
now possessed of the data to fully determine.

Further studies will be necessary to fully understand the part
played by the active moving connective tissue cells in the pro-
duction and growth of the collaginous fibers. So far as they are
now determined the essential phases of the process in which the
connective tissue cells are concerned appear to be: (1) the connec-
tive tissue arises from a primitive mass of round mesenchymal
cells; (2) there is a change of form from round to stellate, to
spindle, and eventually to lamellar cells; (8) certain fibers seem
first to appear within the cells, possibly at their surface (Hansen,
Golowinski) ; (4) there is formed a reticulum pervading the inter- —
cellular ground substance whose fibers may be, though they not
necessarily are, identical with those first arising within the cell;
(5) coincident with the origin of fibers there begin amoeboid move-
ments in the stellate and spindle cells; (6) there is an increase in
size and number of fibers in the reticulum and they aggregate
into bundles, synchronously with which first the stellate and later
the spindle cells move ae the surface of the fiber and fiber
bundles.

In conclusion I desire to express my sincere thanks to the
Marine Biological Laboratory at Woods Hole, Massachusetts,
for the opportunities so kindly placed at my disposal.

BIBLIOGRAPHY

Arnotp 1893 Arch. f. path. Anat., vol. 132, p. 502.

Bout, Franz 1872 Arch. f. mikr. Anat., vol. 8, p. 28

Burrows, M. T. 1911 Jour. Exp. Zool., vol. 10, p. 63.

CARREL, ALEXIS, AND Burrows, M.T. 1910 Jour. Am. Med. Assoc., vol. 55, 1379.
1911 Jour. Expr. Med., vol. 18, p. 416.

erauson, J.S. 1911 Am. Pa Anat., vol. 12, p. 277.
1911 Biol. Bull. vol. 21, p. 972.

FLEMMING 1891 Festschr. f. R. Virchow.
1897 Arch. f. Anat., p. 171.
1906 O. Hertwig’s Handbuch. d. vergleich. u. exper. Entwickl. d.
Wirbeltiere, vol. 2, p. 1.
GOLowInsk1 1907 Anat. Hefte, vol. 33, p. 205.
Goopricu E.S8. 1904 Quart. Jour. Mic. Sc., vol. 47, p. 465.
Hansen, F.C. C. 1899 Anat. Anz., vol. 16, p. 417.
Harrison; R.G. 1893 Arch. f. mik. Anat., vol. 43.
1895 Anat. Anz., vol. 10, p. 138.
1910 Jour. Exp. Zool., vol. 9, p. 787.
1911 ‘Science, vol. 34, p. 279.
HENLE, J. 1841 Allgemeine Anat., S. 197, u. 397.
v. Korrr 1907 Ergebnisse d. Anat. u. Entwickel., vol. 17, p. 247.
Lewis, MarGaret R. ann W.H. 1911 Anat. Record, vol. 5, p. 377.
Lirvint, F. 1909 Monitore zool. Ital., vol. 20, p. 225.
Matt, F. P. 1902 Amer. Jour. Anat., vol. 1, p. 329.
MERKEL, Fr. 1895 Verhandl. d. Anat. Gesellsch., vol. 9, p. 41.
Scuwann, THEODOR 1839 Morphological Researches, trnsl., 1847.
Sputer 1897 Anat. Hefte, vol.7, p. 115.
VALENTIN R. Wagner’s Handworterbuch der. Physiologie, vol. 1, p. 670.
VircHow, Rupotr 1852 Verhandl. d. Phys. Med. Ges., vol. 2, pp. 150, 314.
1860 Cellular Pathology, transl. from 2d German ed., p. 43.
ZIEGLER 1903 General Pathology, transl. from 10th German ed., p. 353.

PA

EXPLANATION OF FIGURES

1 From a Fundulus embryo, of 5.5 mm. total length, showing beginning amoe-
boid movements of two cells, a and b, on the border of the round cell area at
the posterior extremity of the ventral median fin. The observation extends over
a period of fifteen minutes. The last seven drawings were made without change of
focus for the purpose of eliminating variation in form due to the examination of
different levels. a, b, two connective tissue cells. Ch, prominence on the sur-
face of a chromatophore, the body of the cell is not represented. The numerals
in this and succeeding figures indicate the exact time at which each recorded
drawing was completed.

2 Connective tissue cells, one of which, a, exhibits transformation from a
spindle to a stellate type, and another, b, becomes flattened against a connective
tissue fiber. There was an apparent anastomosis between cell a and the proto-
plasmic process p of an adjacent cell. f, connective tissue fiber; 7, margin of a
joint of a fin ray, giving a fixed point inrelation to which locomotion may be deter-
mined. Other letters and numerals as in the preceding figures.

3 Two connective tissue cells exhibiting some locomotion. One of these, a,
assumed a lamellar like relation to a fiber bundle while rotating about it. At
9.25 a.m. this cell became momentarily so thin as to almost escape observation.
Ch, Ch’, chromatophores. Other letters and numerals as in the preceding figures.

142

LIVING CONNECTIVE TISSUE CELLS PLATE 1
JEREMIAH 8. FERGUSON

ah RON Ch ates os
aS) Oe: =, “arab, ee
423 vas 4| 26 4.27 4.29

3 kL
Gy
ee ae, ane, Oe aoe , Paz?
4.3] 4.34 ee 4.38

A 918 ( fe I23 92
gr
Zh of Vm yp ae
pe {A -
thos “ffck Lae
Q 2. 2 7 9 3 | Y) 93 8) 3 4] f 937 93 SS) 9.4

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, No. 2

PLATE 2
EXPLANATION OF FIGURES

4 Amoeboid motion resulting in change of form exhibited by connective tissue
cells of the primitive or ‘round’ type. J, cells a—-c, from the extreme margin; JJ,
cells d-e, from just within the margin; and J//J, cells f-g, from the interior of a
round cell area. J and JI from the pectoral, 7/7 from the caudal fin. Numerals
as in the preceding figures.

5 I, transformation of a round to a spindle cell in the pectoral fin of a Fun-
dulus embryo 6 mm. long, 20 days after fertilization, 11 days after hatching.
From 4.20 to 4.22 p.m. there was in this cell an apparently retrograde change from
spindle to stellate form but at 4.24 p.m. this had been proven temporary. JJ,
transition of a stellate cell to a temporary spindle form. Letters and numerals as
in preceding figures.

6 Apparent retrograde change from spindle to a stellate form in a cell under-
going rather slow locomotion. The stellate phase of such cells is nearly always
temporary. From the same embryo as fig. 5; cap., blood-capillary. Other letters
and numerals as in the preceding figures.

144

LIVING CONNECTIVE TISSUE CELLS PLATE 2

JEREMIAH S, FERGUSON

d 4
ie A se ote ob Hy

9.59

914

1 9p fs hy/Bfu

8592 893 854 855 857 859

0
1 G2 go fe F? fe
ai 42

418 x fe 20 421 42K 422 423 4234424

0 eb ghd
Ly f. a
by be Lo to,

i> eo 6
= ion We tar
O15 é ae 920 922 f

eee
4 ERRAND

457 459 5.00 S01 S502 S01 306
14

_

PLATE 3

EXPLANATION OF FIGURE

¢ Stellate connective tissue cells from the fins of four embryo fish, I-IV.
exhibiting rapid change of form. Owing to difficulties of observation it is not
possible to make drawings oftener than at 1 to 3 minute intervals; hence, the
actual changes of form were much more frequent than the record shows. A-K.
_ nine connective tissue cells. Numerals as in the preceding figures.

146

LIVING CONNECTIVE TISSUE CELLS PLATE 3
JEREMIAH S. FERGUSON

9h Ls5"

WAC ee

1228 8612300 1232—tit«éD;«

a

PLATE 4

EXPLANATION OF FIGURES

8 Stellate connective tissue cells exhibiting locomotion and, on contact, appar-"
ent fusion. J, from the pectoral fin of a 10 mm. Fundulusembryo. The fusion
between cellsaand 6 is apparent only. bandcappearto form part of the anas-
tomosing syncytium. J/,from the caudal fin of a6 mm. Fundulusembryo. Cells
e and g on coming into contact at 4.44 p.m. apparently fused after the manner of the
cells which form the delicate early connective tissue syncytium. Letters and
numerals as in the preceding figures.

9 Stellate connective tissue cells in the subectodermal mesenchymal syney-
tium of a 25 mm. embryo pig. Fibrils pass through the cells very close to the
nucleus. Bielschowsky stain.

10 Connective tissue cells from the praevertebral (J) and intermuscular (JZ)
connective tissue of a 100 mm. pig embryo. The form of the cells and their
contact relations to adjacent connective tissue and muscle fibers is strikingly simi-
lar to the amoeboid connective tissue cells of the living fish embryo. The appear-
ance suggests that such cells were quite probably moving along the surfaces with
which they are in contact.

148

LIVING CONNECTIVE TISSUE CELLS PLATE 4
JEREMIAH S, FERGUSON

de! .
I ok’ dss
4.37

——
SS
—~

a 438 4.40

EN

149

A COMPARATIVE MICROSCOPIC STUDY OF THE
INTERCALATED DISCS OF VERTEBRATE
HEART MUSCLE
H. E. JORDAN anv K. B. STEELE

From the Anatomical Laboratory of the University of Virginia

° TWENTY-THREE FIGURES
CONTENTS
IL lb atingoye Icio) ah RR hed SO OA SEs a tho poole Goce b ote Go mmamaaocacemoar 151
ieeViaternialkan dime thod Sessa ne ae eee eer crenata eter 152
HED Eeseriptives .... stew aa he ee Ce ee SRR Re Peay Meret crete ia aa altyeks crs oles 153
ACCA Ghalttimateriall eee re ie oon oe mek en ee re crarcr sn Minus hss ceca seater auda 153
it. Monk yeeaarncetere’ 36 BA ePrints sch eaie uaueier 153
O° “Bate pueweeies ce . Ska Jas or eer eminem pone eS Ln arts ie ee 157
3: GUN hap eee ss ls Fake ee AAI ee lenges sot pate agers 159
AL © hitgornnita appar teie e ce es eee, sos gay Sey eet ee teers 160
5s OPOSSII eee ek Akio 5. Si 4e ee I ae ae) State aro 162
(ceed s Rp hoahaahh a¥ec| Ove le atte WEARS Reel Sic, oot eo BONES eer Eers Cee tee 8 162
SAT fed Bubb eto) (sje ee oa ala eh ne DE Soca ich rs Gch pear cc ORR CR OCs 162
Cooks BY 16 We pan te fs aah anaes A ANA tees 1c, cece Saline Aen aaa a ee 162
OR Erg ee een ih. 2 ctathae Strap sets cen wetdaienchocuseyenioe 163
LOS MET OUtee ery eer eee icin over a eon arvaicua teak radereth acs Se 164
Bee Woungvangefoetalematentallpens 4 esee See ic as Sarre 165
LEV Guime area e ascites A Rese ee eee cael ce Sees ee 165
a. Second, third and fourth weeks.................. 165
[ oj aal SHS ekIS So eS Pio i Sere cic Did OG SEC eo 165
Cz oastiweels Of gestatlOn geen tec cms ors cee aneas © 165
2h Cate OlehoOuly days. «sc cc eeaee tence aeae oe sisi usre or ogeior 165
SU Childwottoursyeans tac... ee ere Oise pey ae osteo 165
TV EMDISCUSSIOM GH err ener Mire inte Ay seciige. tdi co as er ence Petaisic Llahaiatsy tishellenes 167
Mice SUTRA YAM COMCIMSIONS 5.0.0.5 cs b:c.s,5.5 bc o eactene seelere se heal tee see srerae & 172
THEM eG UTE TC LGE Ot teehee a te a hero chs. ove =o! ESPON ave) tov) 9 Sc Stolen ehe care 173

I. INTRODUCTORY

On the basis of histologic findings in the heart of the humming-
bird, Jordan (’11) believes he has demonstrated that intercalated
discs are not correctly interpreted as intercellular elements mark-
ing cell boundaries, as recently again urged in the papers by

151

152 H. E. JORDAN AND K. B. STEELE

Zimmermann (710) and his students, Palezewska (10) and Wer-
ner (10). It is the purpose, in part, of this investigation to
present further facts in support of the non-cellular interpreta-
tion of vertebrate heart muscle.

Marceau (’04) states that he was unable to demonstrate discs
in the heart muscle of vertebrates lower than birds;! also in goose
and duck. Moreover, he says that discs develop only some time
after birth. The further object of our research is to test the
factual basis of these assertions and to determine, if possible,
from structural marks, the probable function of the intercalated
dises.

II. MATERIAL AND METHODS

The material studied twncludes heart muscle of adult human,
monkey, sheep, bat, cat, guinea-pig, mouse, rabbit, chipmunk,
opossum, humming-bird, lizard, turtle, toad, frog and trout; also
of guinea-pig of last week of gestation, and of first, second, third
and fourth weeks of post-natal life; of a cat embryo of four days,
and of a four-year-old child; and of Limulus.?

Zimmermann’s technic* was employed except for adult human,
cat, rabbit, sheep and lizard hearts. Mammalian cardiac muscle
of the latter group has been fully described and illustrated by
Werner. We shall confine our descriptions for the most part
to the heart muscle of forms not previously studied. The
descriptions in every instance apply to the ventricular tissue.

1J—n the second volume (1911) of ‘Plasma und Zelle,’ Heidenhain says
‘Schaltstiicke wurden bisher in keinen Falle bei niederen Wirbeltiere beobachtet;
sie finden sich by Végeln zum ersten Male.’

2 We are indebted to Messrs. H. F. Jackson and E. L. Powers for kindly col-
lecting the material of toad, frog, trout and Limulus at Cold Spring Harbor, L. I.;
and to Dr. W. H. Schultz, of the Hygienic Laboratory, Washington, for the young
and foetal guinea-pig hearts.

3 Following this technic, small pieces of tissue were treated for twenty-four
hours with a solution of 90 parts of absolute alcohol plus 10 parts of 25 per cent
HNO;. The tissue was then washed in several changes of 94 per cent alcohol or
until the latter remained neutral to litmus paper. It was then passed into dis-
tilled water, from which it was transferred to a solution of 1 gram of Gribler’s
haemalum in 10 cc. of water. Here it remained for from eight to ten days when it
was washed in distilled water and then carried through the ordinary steps of the
paraffin technic. In our own experience we have obtained equally successful
preparations by staining sections for twenty-four hours on the slide.

INTERCALATED DISCS OF HEART MUSCLE 153

III. DESCRIPTIVE

A. ADULT MATERIAL

In every case the myocardium consists of a close-meshed net-
work of coarser and finer branching striped muscle trabeculae.

1. Monkey

When a large expanse of tissue is under view, the discs are
seen to lie more or less closely aggregated in definite regions.
There is evidently roughly an alternation of disc-containing and
dise-free areas (figs. 1 and 2). On close examination the disc-
containing areas for the most part correspond to the axial regions
of the mesh. In physical terms these regions correspond to
areas of greatest stress during contraction (figs. 1 and 2). The
discs appear commonly in three distinct forms: (1) they may be
compact and wide,‘ i.e., as wide as a single fiber; (2) they may
be narrow, i.e., not much wider than a single fibril; (3) they may
consist of rows of spherical granules, connected severally by a
delicate deep-staining membrane. These three main types of
dises are all represented in fig. 1. The ‘compact’ dises under
higher magnification are seen to be composed of rows of longer
or shorter rod-like granules. The rows of spherical granules
usually span the intervals between adjacent discs of more com-
pact forms (forms 1 and 2). The myofibrillae are clearly evi-
dent, singly and in bundles, giving the fiber a distinct longitu-
dinal striation. The striations (fibrillae) passing through the
smaller dises of form 2 are much coarser than the adjacent fibrils
(fig. 1). These fibrils are evidently modified, being probably in
a State of greater contraction; and the narrower discs would seem
to be foci of still greater contraction in the fibrils. If this in-
terpretation is valid—as many facts about to be detailed very
strongly indicate—then the wider discs are reasonably conceived
as the result of a fusion or association of narrower discs; and the

4 ‘Wide’ and ‘narrow’ refer to the transverse axis of longitudinal sections,
‘coarse’ or ‘robust’ and ‘delicate’ to the longitudinal axis, and ‘deep’ and ‘super-
ficial’ to the diameter of the uncut fiber.

154 H. E. JORDAN AND K. B. STEELE

All figures unless otherwise specified are magnified 1800 diameters.

Fig. 1 Expanse of cardiac muscle of monkey in longitudinal section, showing
several types of intercalated discs, their general disposition in the axis of the mesh;
and their relation to each other, the dark bands, and the myofibrillae.

Fig. 2 Longitudinal section of two medially fused muscle trabecular of mon-

key heart, showing the axial disposition of the discs, and the super-nuclear loca-
tion of several.

INTERCALATED DISCS OF HEART MUSCLE P55

rows of granules as areas of local contraction in otherwise relaxed
and distinct fibrils, the inter-granular membrane possibly repre-
senting a condensed membrane of Heidenhain (M line; meso-
phragma, Heidenhain).

The dises are almost invariably at the level of, and, where
present, displace the dark (anisotropic) bands.* However, they
are frequently wider than these bands. Occasionally, when ro-
bust, they may extend approximately halfway or even entirely,
through the lighter band (fig. 6). But they are certainly not
generally bounded on both sides by Krause’s membrane (telo-
phragina, Heidenhain), an observation urged in favor of an inter-
cellular interpretation, as has been stated by some investigators,
e.g., Heidenhain.

In fig. 2 all of the discs appear as wider or narrower more or
less compact granular bands. The two separate collections are
localized at the points of stress. One disc appears to span the
line of junction. Several lie partially over the nucleus. All are
superficial and of comparatively little depth.* These structural
variations, superficial and occasional super-nuclear position, ren-
der untenable interpretation as cement lines. Fig. 3 illustrates
four successive superficial discs unconnected by ‘risers’ to form
‘steps.’ Moreover, they gradually shade laterally into the ani-
sotropic bands which they in part displace. In fig. 4, three discs
are shown overlying the deeper nucleus. In fig. 5 are illustrated
two wide dises of finely granular character. Such discs are fairly
common. In this particular fiber the two discs lie at slightly dif-
ferent depths. The important point is that in passing from a

5 The term anisotropic is not here employed in a definitely physical sense. Itis
used simply to denote the darker-staining substance in the Q and Z band of striped
muscle, in accordance with common custom. Careful study of cardiac muscle
with the micropolariscope, under the same physical conditions that very strik-
ingly revealed anisotropic vegetable fibers and inorganic crystals, failed, however,
to disclose any definitely anisotropic substance in the myofibrillae. The point
urged is the similarity between the darker transverse and the intercalated discs.
It would seem, moreover, that not all striped muscle can be resolved into sharply
defined isotropic and anisotropic discs. On the other hand, as far as the micro-
polariscope gives evidence, ‘stripes’ appear in the absence of anisotropic granules.

6 Hence ‘discs’ strictly defined is a misnomer; they are more properly desig-
nated ‘bands.’ The Q and J ‘bands,’ moreover, are more correctly denominated
‘discs.’

156 H. E. JORDAN AND K. B. STEELE

higher to a lower level of focus, when one disc fades out of view
as the other passes in, no differentiated connecting substance
(‘riser’) appears, as one would expect if these two dises repre-
sented portions of an irregular cement line cut tangentially, as
interpreted by Zimmermann and his students.

The most plausible interpretation, it would seem, must regard
the dark-staining granules or rodlets of the ‘discs’ as local modi-
fications or contractions at anisotropic levels of the myofibrillae.

3 eae 5 6

Fig. 3. Branching fiber of monkey heart with four plate-like discs, showing
their relation to the dark bands.

Fig. 4 Short portion of fiber of monkey heart showing super-nuclear position
of three discs.

Fig. 5 Two granular dises of monkey heart at different levels of focus. In
passing from the higher to the lower level of focus no connecting membrane or
‘riser’ appears.

Fig. 6 Atwo-step, comb-like, dise of monkey heart, in longitudinal width equal
to that of the isotropic and two anisotropic dises. The two portions are connected
by a coarse deep-staining membrane. The ‘teeth’ of the ‘comb’ are interpreted
as locally contracted portions of adjacent fibrillae.

Fig. 6 shows a rare variation. Here the disc may extend the
longitudinal distance of half or even an entire lighter band. In
form it has a ‘comb’ structure. When at different levels, as here,
the several sections are connected by a robust deeply-staining
membrane. The appearance simulates protoplasmic processes or
‘intercellular bridges.’ However, if this were a correct interpre-
tation, they would undoubtedly be very much more numerous.

INTERCALATED DISCS OF HEART MUSCLE 157

The ‘teeth’ of the ‘comb,’ we believe, must be interpreted as
local, elongate contractions of the myofibrillae. Occasionally
such dises consist of only one or several ‘teeth;’ again three or
four ‘combs’ may appear in longitudinal series, giving the appear-
ance of a contracted striped insect muscle. Frequently successive
dises limit a region of darker sarcoplasm; frequently also dises
le in regions of darker sarcoplasm. The significance of these
observations is that they indicate a relationship between discs
and a condition of contraction. Possibly the dises outline lim-
its of special physiological states. The loose granular dises are
perhaps simply ‘comb’-dises with shorter ‘teeth.’

2” Bat

In the heart of the bat, in addition to the types of dises de-
seribed for monkey, there appear discs in the form of a line of
dark-staining, connected, spherical granules extending across
as many as eight separate fibers; and directly over underlying
nuclei (fig. 7). Such a structure does not even vaguely suggest
a cell membrane or cement line. It becomes intelligible only
in terms of local modifications (contractions?) at approximately
the same level in many adjacent fibers. It would seem to rep-
resent the morphologic picture of a physiologic state sharply
localized longitudinally, widespread laterally. In fig. 8 are shown
two series of narrow dises connected by more or less delicate
granular inter-disc membranes (‘risers’). These inter-connec-
tions, and the localization of the dises in the axial region of the
muscular mesh, here again suggest the effect, or phase, of a defi-
nite physiologic state. The inter-dise membrane may represent
less extreme local contractions in the intervening fibrillae, or a
dislocated modified membrane of Heidenhain, or possibly simply
‘anisotropic’ granules obliquely aligned. Careful focussing reveals
the fact that these complex step-like discs lie superficially through-
out and encircle the fiber in ring-fashion. This single observation
disposes definitely of a cell-border interpretation.

158 H. E. JORDAN AND K. B. STEELE

Fig. 7 Longitudinal section of cardiac muscle of bat, showing a granular
intercalated ‘disc’ running across eight fibers.

Fig. 8 Heart muscle of bat showing two long step-like discs winding super-
ficially around their fibers. The ‘risers’ probably represent anisotropic granules
aligned longitudinally, or perhaps a distorted membrane of Heidenhain.

INTERCALATED DISCS OF HEART MUSCLE 159

3. Guinea-pig

Fig. 9 illustrates the typical condition in adult guinea-pig.
The discs are again aggregated in more or less definite areas;
they may be wide or narrow, delicate or robust, compact or loosely
granular. At the left, two discs are shown connected by a row
of granules; a similar disc, of mono-serial coarse granules, ex-
tends uninterruptedly across the two middle fibers. The same
condition, producing the disc-structure, prevails at approxi-

Fig. 9 Section of guinea-pig heart muscle illustrating the several types of
discs, and their relation to each other, the fibrillae and the darker bands.

mately the same level in the four adjacent fibers; the appearance
is incompatible with a cement line. Fig. 10 shows a disc of
three ‘steps’ and two ‘risers’ extending across a nucleus. Fig.
11 illustrates a patch of compact and pale granular discs. Fig.
12 shows a variation occasionally met with in the guinea-pig and
other mammals. Such a disc has not the remotest resemblance
to a cement line or to an intercellular bridge. The thickenings

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, No. 2

160 H. E. JORDAN AND K. B. STEELE

(blocks). of the dises are most likely local contractions in the fibrils,
the connecting deeply-staining ‘membrane’ being formed of the
lateral coalescence of less extreme locally contracted portions
of intervening fibrils; or it represents perhaps a thickened, dis-
torted membrane of Heidenhain; or it may be the product of
the combination of both of these elements. This is a more com-
mon type of dise in young hearts.

10 11 . 12

Fig. 10 Guinea-pig heart muscle fiber, showing the super-nuclear position of
a three-step disc.

Fig. 11 ,Guinea-pig heart-muscle fiber, showing compact and granular varie-

ties of plate-like discs. The loosely granular discs are the less highly developed
type. '
Fig. 12 Guinea-pig cardiac fiber, with block-like discs at different aniso-
tropic levels and connected by delicate granular membranes giving a zig-zag
appearance to the complete structure. The ‘blocks’ are locally contracted por-
tions of the fibrillae, the connecting membranes represent probably ‘anisotropic’
granules in linear arrangement, or one of the ‘membranes’, i. e., M or Z.

4. Chipmunk

Heart muscle of the chipmunk exhibits all the types of dises_
above described. It shows also especially clearly and abundantly
a type illustrated in fig. 13. Here the discs appear as oval thick-
enings on the fibrils. In other locations three or four (still more
elongated) may appear at successive anisotropic levels, giving
the appearance of deeply-stained striped insect muscle. Fig. 14

INTERCALATED DISCS OF HEART MUSCLE 161

shows a step-like dise with six ‘steps’ and five ‘risers.’ The
‘steps’ consist of deep-staining granules; the ‘risers’ are more
delicate granular structures. Several of the discs shade off later-
ally beyond the ‘risers,’ into more compact granular anisotropic
bands. It would be impossible, we believe, to interpret this
structure in terms of a cement line or cell boundary. The ‘steps’

14

13

Fig. 13 Branching cardiac fiber of chipmunk, with numerous isolated larger
and smaller oval discs, for the most part at the darker levels of the fibrillae.

Fig. 14 Cardiac fiber of chipmunk with step-like discs, composed of coarse
granular thick ‘steps’ and delicate granular ‘risers.’

Fig. 15 Fiber of humming-bird heart, showing the short, compact deeply-
staining dises characteristic of this muscle.

of the dise here are clearly at the levels, and displace portions, of
the darker bands, the entire structure lying superficially.

162 H. E. JORDAN AND K. B. STEELE

5. Opossum

Conditions in the heart of the opossum are almost identical
with those in guinea-pig. As concerns the intercalated discs,
they differ only from their homologues in guinea-pig in being
somewhat paler, more delicate and apparently less numerous.
The deeper-staining character of the disc-containing portion of
the muscle trabeculae is especially conspicuous in the opossum
heart.

6. Humming-bird

Fig. 15, added chiefly for the purpose of completing the series
of illustrations, shows a fiber of humming-bird heart muscle to
illustrate the relative abundance, comparative form, and typical
appearance of this tissue. Humming-bird heart muscle has been
fully described in a former paper’ by one of the authors.

7. Turtle

In turtle heart the discs are fairly abundant (fig. 16). The
discs here all appear as narrower or wider plates. They are again
superficial and at the anisotropic levels, displacing the darker
bands, and in many instances shading gradually into them at one
or both ends. None of the rare varieties above described for
mammalian muscle appear in turtle nor in lower forms, even
step-like dises being very rare. The localization of the discs in
definite transverse areas is apparently absent. The discs here
are stouter than in frog and toad.

8. Toad

Fig. 17 illustrates conditions in the heart muscle of toad.
The majority of dises are very narrow, though occasionally dises
the width of an entire fiber appear. The discs are situated super-

7In this paper the darker band was regarded as the Q band without con-
sideration of the possibility that it may represent the Z line modified in con-
traction. This possibility is discussed below.

INTERCALATED DISCS OF HEART MUSCLE 163

ficially and are of little depth, and apparently promiscuously
placed. On close examination, under higher powers, the appar-
ently compact delicate discs are seen to be granular. They are
abundant only in certain areas, and not generally as plentiful
as this particular illustration would indicate.

16 ae 17

Fig. 16 Four adjacent fibers of turtle heart. The discs are superficial, numer-
ous, and very like those of bird heart; always at anisotropic levels, usually com-
pact and only very rarely in ‘steps.’

Fig. 17 Cardiac muscle of toad.

9. Frog

Practically the same description holds for frog (fig. 18) as for
toad. The important point is the presence of discs in tissue
below birds, where they have been denied. There appears abso-
lutely no evidence here that they mark cell boundaries.

164 H. E. JORDAN AND K. B. STEELE

10. Trout

In heart muscle of trout (fig. 19, representing several levels
of focus), the discs are distinctly fewer in number than in higher
groups. Occasionally also the oval type appears as thickenings

18 abe inert

Fig. 18 Cardiac muscle of frog.
Fig. 19 Cardiac muscle of trout. The intercalated discs are similar to the
simpler types of hearts of higher forms, but distinctly less numerous.

on the more distinct stouter (contracted) fibrils. The relation
with respect to the nuclei and the anisotropic bands is the same
as described for higher vertebrates. The discs are more abundant
in contracted (darker) regions, and near branches. Occasionally

INTERCALATED DISCS OF HEART MUSCLE 165

they extend across two fibers, and very frequently the discs are
at the same level in adjacent fibers. They are always superficial,
never appearing in the mid-line of a fiber when the nucleus is in
focus. They appear in patches, and without regard to the posi-
tion of the nuclei, and are only very rarely in ‘steps.’

B. YOUNG AND FOETAL MATERIAL
1. Guinea-pig

a. Second, third and fourth weeks. In young guinea-pigs of the
second to fourth week intercalated discs are already present
(figs. 20 and 21). The majority, however, are in the shape of
very narrow bands. These may be arranged in ‘steps’ connected
by ‘risers.’ Many of the apparently compact discs are resolved
under higher magnification into a series of blocks (local thicken-
ings of fibrils), and the discs in general are less compact than at
older stages. Again there is absolutely no evidence of cells and
boundaries.

b. First week. During the first week discs are already present,
and at the usual levels, but all have a light-staining character.
In this tissue the granular discs shade gradually laterally into
the darker bands.

c. Last week of gestation. During the last week of gestation,
coincident with the appearance of striations, dises first make
their appearance (fig. 23). No indication of discs could be seen at
any earlier period. The discs here are narrow and consist of
faintly-staining granules.

2. Cat of four days

In a cat embryo of four days the dises are already sparsely
present. They are evidently just making their appearance.
They are narrow, pale and granular.

3. Child of four years

In this tissue the discs are abundant, and mostly of the type of
narrow plates of slight depth. Relative to their number in the

166 H. E. JORDAN AND K. B STEELE

adult heart, however, they are meager in amount. Whether
this fact is due exclusively to the young, or, in part also, to the
diseased condition (tubercular meningitis) is not at present clear.
In the case of guinea-pig it is certain that the discs are relatively

20 21

Fig. 20 Cardiac muscle of guinea-pig of four weeks of post-natal life. The
intercalated dises appear aggregated in the axis of the mesh, are of simple charac-
ter and mostly of paler or darker granular composition.

Fig. 21 From same preparation as fig. 20, showing the beginning of the for-
mation of step-like discs.

less abundant and simpler in young than in the adults. Possibly
the same relation obtains throughout mammals. However, there
may also be a relationship between the discs and the condition of
health and disease, a point which is now being investigated.

INTERCALATED DISCS OF HEART MUSCLE 167

Fig. 22 Cardiac muscle of guinea-pig of first week of post-natal life. The
dises are abundant, but exclusively of the granular (or less compact) narrow dise
type.

Fig. 23 Cardiac muscle of guinea-pig of last week of gestation. Cross stria-
tions and dises are just beginning to make their appearance.

IV. DISCUSSION

An attempt to establish a relationship between rate of heart-
beat and the number of discs was unsuccessful. If a series of
animals, ranged according to the reported and observed relative
abundance of intercalated dises, be compared with the same series
ranged according to the rate of heart-beat, a correspondence
appears at certain points between as many as three successive

168 H. E. JORDAN AND .K. B. STEELE

members (e.g., dog, man, sheep). But at other points there is
absolutely no correspondence (e.g., sheep, cat, rabbit). The
relationship, therefore, if it exists at all, cannot be a simple one.
Factors, besides rate of beat, must affect the relative abundance
of the dises. Such factors may be the force of the beat, or the
instant or total amount of work done. That the rate, simply, does
not determine the number of the discs appears furthermore from
the fact that the discs are more abundant in the adult than in
the young (e.g., guinea-pig and man), whereas the rate of heart-
beat relative to age varies in the reverse ratio.

Moreover, the number of the discs varies in different portions
of the same heart and in different individuals of the same species.
This may mean, however, that they vary according to the phase
or state of function, or perhaps according to the total amount of
function (i.e., age of the individual). The observations regarding
the relative abundance of the discs above stated may thus have
no absolute (final) significance. Sufficient observations under
uniform conditions have not yet been made for an accurate seri-
ation of heart muscles from the standpoint of the abundance of
dises. The physiologic significance (normal and abnormal) of
these structures will appear in full only after a careful compara-
tive study of the same individual under varying internal and
external states, both normal and morbid; of animals of the same
species at different ages; and of animals from the various groups
under relatively uniform conditions of age, health and function.
When all the factors helping to determine the presence and abun-
dance of intercalated discs are thus known and accounted for, it
may become possible to arrange animals in identical series from
the standpoint both of the rate of the heart-beat, and from the
number of the intercalated discs. It seems clear that a relation-
ship of some degree exists between the presence and abundance of
these dises and function (rhythmic contraction), but in detail the
relationship remains obscure.

In a histologic study of the lung of the white mouse one of us
(Jordan) recently discovered that the tunica media of the proxi-
mal end of the pulmonary arteries consists of striped (cardiac)
muscle for a considerable distance. This seemed to offer, there-

INTERCALATED DISCS OF HEART MUSCLE 169

fore, an excellent material for testing Heidenhain’s interpreta-
tion of the intercalated discs as regions where new sarcomeres
(‘inokommata’) are being added to the growing cardiac fibers.
According to Heidenhain, the heart. can enlarge only by inter-
stitial growth, 1.e., by terminal additions to the cardiac elements
(trabeculae). In the case of the pulmonary arteries, however,
there appears no reason for postulating growth by this method;
the media here, developed from truncus arteriosus to be sure, can
nevertheless, undoubtedly increase in amount in the same way
as elsewhere in arteries. Moreover, striated muscle elsewhere does
not increase by means, nor show evidence of, intercalated dises.
But intercalated discs are present in the pulmonary media;
furthermore the greater abundance of the discs here coincides
with the time of less rapid growth, and less close developmental
relationship with the heart. The presence of intercalated dises
in the media of the cardiac end of the pulmonary arteries in the
mouse would seem to be correlated with the ‘beat’ (strain?)
here occurring in common with the heart. Still other facts
controverting Heidenhain’s interpretation that the intercalated
discs provide for the ‘interkalare Laingenwachstum’ of the
cardiac fibers are: (1) the absence of transition stages between
the discs and fully formed sarcomeres; (2) their absence during
stages of most rapid (foetal) growth; (38) their numerical increase
even after the heart has attained its normal bulk; (4) their pres-
ence in aged and diseased hearts; and (5) their considerable struc-
tural variation—every type capable of resolution, however, into
very similar elementary units.

Militating most strongly against Zimmermann’s interpretation
of the discs in terms of intercellular elements, is our observation
of the superficial location of the complex step-like forms. The
more complex step-like types appear only where the entire fiber
is included within the plane of section. Under such circumstances
the successive ‘steps’ can be traced completely around a fiber
by lowering and again raising the level of focus. Many such are
then seen to form rings or even short spirals. The discs are of
course not complete in the step forms, but are interrupted, con-

170 H. E. JORDAN AND K. B. STEELE

sisting of ‘steps’ at different anisotropic levels connected by deli-
cate membranes spanning the intervening ‘isotropic’ bands.

Attempts to alter the number of the discs experimentally by
stimulation with varying strengths of an electric current have
proved unsuccessful. Nor are they appreciably affected in tissue
fixed in a state of rigor mortis. Material is now being collected
for a study of these discs in various pathological conditions of the
heart. Discs could not be demonstrated in Limulus heart.

A comparison of the illustrations accompanying the articles
by Zimmermann’s students, Werner and Palezewska, shows that
in the human heart the discs are commonly bounded on both
sides by the so-called ‘Krause’s Z-lines,’ whereas in lower mammals
the dises are narrower than the space between two Z-lines, and
consequently bounded on only one side by this line. Heidenhain
likewise illustrates the discs in human heart muscle as bounded
on both sides by a Z-line. Granting that this interpretation
of the striped condition is correct, especially then in man do the
dises correspond to the Q or reputed anisotropic levels—as we
have urged on the basis of a different interpretation—and in so
far support our contention that they represent modifications of
the fibrillae at anisotropic levels. But all of these illustrations
differ from the far more widely prevalent condition of our ma-
terial, in that the so-called ‘Z-line’ or ‘Krause’s membrane’ is
represented much too delicate. The darker stripe (seen both in
fresh and stained material) is usually stout, and frequently almost
half as wide as the alternate lighter segments; this is more par-
ticularly the case in human material. Since the sarcolemma is
only occasionally, and then only imperfectly, festooned between
these lines no definite suggestion is given of a ‘Krause’s mem-
brane.’ Especially in the regions where the discs appear abun-
dantly are the dark stripes robust. Having naturally directed
our attention chiefly to these regions, we interpreted appearances
as indicating a condition of semi-contraction, according to the
illustration of Tourneux (see Traité d’Histologie; par Prenant,
Bouin et Maillard; tome 1, p. 442. Paris).

It seems possible, however, in the light of this illustration and
the theoretical interpretation involved, that the fibers are in

INTERCALATED DISCS OF HEART MUSCLE 7a

condition of full, or nearly full, contraction. During contrac-
tion the substances of the anisotropic and isotropic bands are
supposed to intermingle and ultimately change their relative
locations. Such a transition condition may account for the
indistinct, or absence of, stratification of the anisotropic and iso-
tropic substances under the micropolariscope. Moreover, in the
contracted condition (according to Tourneux’ diagram; see also
M. Heidenhain, ‘Plasma und Zelle,’ ’11, p. 677) the darker stripe
is at the level of the Z-line, itself supposed to consist of aniso-
tropic substance. The Z-line seems to have thickened by reason
of the accumulation of ‘anisotropic’ substance about it, forming
the ‘contraction band’ of Rollet. Thus the darker stripe may
indeed represent the Z-line, plus considerable additional aniso-
tropic substance. The ‘Z-lines’ of Heidenhain and Zimmermann
correspond apparently to the darker lines in our specimens, repre-
senting more likely a ‘contraction band.’ The illustrations of
these investigators are faulty in that they show the darker stripes
too delicate, always single, continuous, and too uniform. If the
darker stripes are indeed the Z-lines, now grown robust in con-
traction, the regions of the fibers containing the intercalated
dises are in a state of more pronounced contraction, according to
the theory of Rollet and Tourneux. This deduction, then, is in
complete accord with our position that the discs are somehow
a concomitant of contraction; and further that they represent
modifications (irreversible contractions?) of the fibrils at the dark
(anisotropic) levels, the anisotropic substance having shifted in
contraction to the Z-line. It seems clear that a complete eluci-
dation of the question of the structure and function of the inter-
calated discs awaits fuller knowledge of the physical and chemical
changes undergone by the cardiac myofibrillae during contraction.
Our interpretation of the robust, sometimes double, dark stripes
(the only stripes visible in non-human material) as the aniso-
tropic bands seems in closer accord with our knowledge of skeletal
and striped muscle generally.

V2 H. E. JORDAN AND K. B. STEELE

V. SUMMARY AND CONCLUSIONS

1. By the use of Zimmermann’s technic it was possible to dem-
onstrate intercalated discs in all heart muscle examined, except
that of Limulus. Of lower vertebrates the material included that
of turtle, lizard, frog, toad and trout, in which forms the presence
of discs has been denied.

2. In guinea-pig, in which form only the matter was tested,
intercalated dises appear early during the last week of gestation,
coincidently with the appearance of striations. A progressive
increase in number, complexity and density was noted during
the first, second, third and fourth weeks of post-natal life. <A
similar more pronounced difference obtains between the heart of
the young and that of the adult guinea-pig. In a cat embryo of
four days the discs are already present but few in number, pale,
and loosely granular in structure.

3. Compared with mammals (e.g., monkey, bat, chipmunk),
in lower vertebrates the discs become progressively less numerous
(except in birds, e.g., humming-bird), narrower, less compact
(more granular) and less complex. Conditions with respect to
the dises in young mammalian hearts are very similar to those in
the hearts of lower vertebrates. With increase of age, there is a
progressive increase in number, complexity, density and width
of dises (e.g., guinea-pig).

4. The comparative study of vertebrate heart muscle gives no
evidence favoring the interpretation of the discs as structures
marking cell boundaries, e.g., cement lines or intercellular bridges.

5. Specific points in the evidence against an intercellular inter-
pretation are: (a) their superficial location; (b) their relationship
to the dark (‘anisotropic’) band,i.e., they displace these bands and
shade laterally into them; (c) their position frequently over a
nucleus; (d) their relation to the myofibrillae; (e) their random
arrangement with respect to the nuclei; (f) structurally and
tinctorially they seem to be of the same nature as the so-called
anisotropic substance; and (g) their absence before the appear-
ance of striations. .

INTERCALATED DISCS OF HEART MUSCLE iis

6. The dises are interpreted in terms of local contractions (or

ageregations of ‘anisotropic’ granules) in the muscle fibrils. The
different modes of association of such single contraction foci
give rise to all the various types of dises described, 1.e., granules,
blocks, ovals, plates (composed of closely apposed longer and
shorter rodlets), ‘combs,’ ‘steps’ and saw-teeth forms.
. 7. The presence of discs would seem to be correlated with the
function of rhythmic contraction characteristic of cardiac muscle,
and may represent a fixed phase of a contraction wave (local or
general), or more probably is the result (of the nature of an irre-
versible strain condition) of the total amount of function. The
latter idea is supported by the fact of (a) their absence in the
mammalian foetus, and their increasing abundance and coarse-
ness with age; (b) their general location in lines corresponding
roughly with the axes of the heart muscle mesh; (c) in general,
their greater abundance in hearts of more rapid beat; and (d)
their presence also in the striated muscle of the media in the proxi-
mal (beating) end of the pulmonary arteries (e.g., mouse).

LITERATURE CITED
HemenHAIn, M. 1901 Uber die Struktur des menschlichen Herzmuskels-
Anat. Anz., Bd. 20:
1911 Plasma und Zelle. Jena.

Jorpan, H. E. 1911 The structure of heart muscle of the humming bird, with
special reference to the intercalated dises. Anat. Rec., vol. 5, no. 11.

Marceau, F. 1904 Recherches sur la structure et le developpement compare
des fibres cardiaque. Ann. des Se. Nat. Zool., vol. 19.

PauczEwska, IRENEVON 1910 UberdieStruktur der menschlichen Herzmuskel-
fasern. Arch. f. mikr. Anat. u. Entwickl., Bd. 75.

Werner, Marie 1910 Besteht die Herzmuskulatur der Siugetiere aus allseits
scharf begrenzten Zellen oder nicht? Arch. f. mikr. Anat. u. Entwickl.,
Bd. 75:

ZIMMERMANN, K. W. 1910 Uber den Bau der Herzmuskulatur. Arch. f. mikr.
Anat. u. Entwickl., Bd. 75.

2

ss
oh
J hs
.
4 : re
;
4
ar
Pid) ibs
ai) \8 =
ra> F
+
' a

se a are ne
Te ‘ i
tA IGA ia
oa 1]
=i, a7) hie iy Oe.
eT Eo mee
at aes
. oes |
yiethe aan "
A ’ . ah PAs oa
&¥ { AY a hee
wf MD
a> Ed
, ef wa Ll
+ er Ph eo eee
Por aera *.
coh bs art ae
iis ‘
nid Oe We
ni. i
ae |
lara oa
| Fi
Kadtele AG Pee 4s... (0
ve of d
prek oe :

SSTaRea ust:

F
7 ;
: t
} ,
: See
—
-
A
ae { .
- . 4
’
4

ba

bd eT LARS vb)
- ge A -
A wy AY.
ie ait) c 7 ’
= t a
* “
>
. = A
btn > i
' ¥
P é .
¥ * x
4 q . pate
A 4 a ¥
is \

¥

a FS sth topes ee, Pat eran s

Be AL We i
;

“hae?

.
4 ‘ ; a,
sa iy 2
Ficoll:
j . a t
i
iol
y
,
f ” ; ‘
o/ i \
e > a
* %
sd
:
«4
5
a ‘
~
} ane
anes
. a nm af
~ 1 - : ‘
: i
‘ P by : a
2 $4 4
i
7 ",
:
+
i :
} ) ad ;
, “7, 4 :
( ie ee 94 *
che ;
s A , .
~ t : tr
ia
n ay
i =.¥ Fp . aed ~
a. Wis
ey
iF
iB yt
. - p. r
> Mie Gah)
f eee
r Pas N

ic ee

Se)
2 ~
os i we =“
. .
cate IN Tose
a
1 *

A ae 9a a

Soa a il

ee GUA, | feo.

, “ i p
7s, F ‘ .
eee E
~ ° j he i | ‘ :
; Lan F4 hia PO 4 re ye | oe
1) ee 1
' ae i { ' ’ .
ny a
afia
* ' ne -
' eh) | pf
Wal ak,
ct ri j
: . ' te Vay
: ha
‘ + q rT «&
; : OMNI
- J i
: P Aaa
“dl th Sy isi
i ar ah Wy
‘ s he : . ay i
. . ,
2 , ; 5 “Ray & Aiteh /
ho i ee” Oo et Oe eee

A NEW TYPE OF FAT STORING MUSCLE IN
THE SALMON, ONCORHYNCHUS
TSCHAWYTSCHA!

CHAS. W. GREENE

Department of Physiology and Pharmacology, Laboratory of Physiology,
University of Missouri

TWO FIGURES (ONE PLATE)

The king salmon, Oncorynchus tschawytscha, like a number of
other fishes, possesses a thin and superficial muscle lying along
the side of the body just under the lateral line. This muscle
extends the full length of the body of the fish from the pectoral
girdle to the caudal peduncle.? It is thickest at the lateral line
and becomes gradually thinner as it spreads out in a sheath over
both the dorsal and the ventral surfaces of the deep portion of the
great lateral muscle. In transverse section the shape of the
muscle on each side is that of two scythes placed together base
to base. When the muscle tissue is coagulated, as in heating, its
dark appearance makes it stand out with prominence. In so far
as the king salmon is concerned this dark muscle presents charac-
teristics of peculiar interest which seem to me to justify a special
characterization and report.

The dark muscle is a differentiated portion of the lateral muscle
mass. Its fibers run in a longitudinal direction. The muscle is
broken into short segments, myomeres, by transverse connective
tissue bands, the myocommata. In comparison with the remain-
der of the great lateral muscle, the profundus, the dark muscle is
characterized by its relatively small fibers. These fibers are very
compactly arranged having a minimum of interstitial connective

1 Published by permission of the U.S. Commissioner of Fisheries.

2 In an anatomical description of the salmon muscles I have given this muscle

the name, ‘musculus superficialis lateralis.’ The deep portion of the great lateral
muscle is called the ‘musculus profundus lateralis.’ Paper now in manuscript.

175

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 2

176 CHAS. W. GREENE

tissue. ‘The muscle as a whole is marked off from the great lateral
muscle by a pretty definite fibrous partition. This separation is
not loose enough to be easy of dissection, but when the connective
tissue is softened, as in the process of cooking, the muscle very
readily separates from the deep portion of the lateral muscle.

I have examined the histological structure of this dark muscle,
however, chiefly with special reference to the normal loading of fat.
The muscle possesses the usual gross arrangement of sarcolemma,
sarcoplasm, and fibrillae. The fibrillae are band shaped and
appear in cross section as doubly refractive lines not unlike the
appearance of bacilli arranged side by side. ‘These fibrillae vary
in size, but are in the neighborhood of from 1 to 1.2u in cross
section in their long diameter. At the surface they are radially
arranged with reference to the axis of the fiber, forming a rather
definite layer around the circumference of the fiber. But they
are irregularly placed throughout the central portion of the fiber,
as shown in figs. 1 and 2.

The most striking characteristic of the normal dark muscle is
the relatively large amount of sarcoplasm. The sarcoplasm forms
_ thin layers between the doubly refractive lines shown in cross
section of the fibrillae, and fills up the spaces where groups of
fibrillae are brought in contact, i.e., the angles of Cohnheim’s
areas. In paraffin sections of adult muscle, both from the young
salmon and from mature adults, there is always present in the
sarcoplasm a number of clear globules or vacuoles. ‘The number
and size of these vacuoles is so great as to obscure the usual rela-
tions of the sarcoplasm, fig. 1. The vacuoles are largely in the
angles of Cohnheim’s areas, but may be present in spaces between
the fibrillae in the individual rows. These vacuoles represent
spaces from which fat has been extracted during the imbedding
process. On the whole the muscle fibers are characterized by
resemblance to the more generalized type of striated fiber. Espe-
cially in the younger fibers do we find an excessive amount of
sarcoplasm both around and between the fibrillae. Other types
of striated muscle of the salmon have smaller fibrillae and rela-
tively less sarcoplasm.

A NEW TYPE OF FAT STORING MUSCLE ae

There is nothing particularly characteristic of the sarcolemma:
of the normal dark fiber except that it is strikingly widely sepa-
rated from the surface of the fiber. Where the sarcolemma is so
separated the space between it and the fiber is filled in with
definite spherical vacuoles from which fat has been dissolved in
the preparation.

There is abundant evidence of the cleavage of these dark muscle
fibers in material obtained from the relatively young salmon.
This evidence consists in the arrangement of fibers in groups and
in the presence of various stages of the separation of fibers, a
point chiefly identified by the arrangement and the formation of
sarcolemmal partitions and by the disposal of fat in the fibers.

The normal loading of fat

The dark muscle is heavily loaded with fat. The fat is present
both between the muscle fibers and within the muscle fibers as
shown in fig. 1.

The amount of fat between the fibers is relatively small. It
is present in drops from 5 to 20u in diameter, chiefly in areas
which contain blood vessels. The fact peculiar to this muscle
and on which I wish to lay emphasis is the presence of enormous
quantities of intramuscular fat. I have followed this fat by the
special method of staining with scarlet red, Bell’s modification of
the Herxheimer method, and bave confirmed the observations
by paraffin sections of both young and adult tissues.

The fat is distributed within the muscle fiber in two regions.
First, within the sarcoplasm throughout the substance of the
muscle, and second, under the sarcolemma but outside the sar-
coplasm. The sarcoplasmic fat is in droplets of extreme varia-
tion in size. In the normal mature muscle these fat drops run
from a fraction of a micron to as much as 6, or in rarer instances
even 10u in diameter. In one typical young fish the average of
the large intramuscular droplets is from 4 to 6 win diameter. The
fat droplets are located at the point corresponding with the angles
formed by Cohnheim’s areas. Great variety exists as to the size

178 CHAS. W. GREENE

of the droplets and among these larger drops, sometimes in close
approximation to them, will be found numerous smaller droplets.
The smallest droplets present are visible only under the oil im-
mersion lens. In numerous instances the smallest droplets are
found between the adjacent fibrillae.

In the normal tissue great quantities of fat are also found over
the surface of the fiber but under the sarcolemma. That is to
say, the fat is stored between the sarcolemma and the muscle
substance. There is great variation shown by the individual
fibers. Some of them present almost continuous rings of fat
droplets, especially in the younger fish, while others show fat
around only a small portion of the surface. In the scarlet red
stained material it is not always easy to demonstrate this relation-
ship of the fat droplets, but in paraffin material stained by Mal-
lory’s aniline blue stain it is easy to confirm the fact that the fat
is within the sarcolemma.

The appearance of the fat and its relations within the dark
muscle fibers, espec_ally within the sarcoplasm of the fibers, is
best shown by the two figures, one of which (fig. 1) gives the posi-
tive picture obtained by scarlet red staining of frozen sect ons of
formalin-fixed tissue; the other (fig. 2) the negative obtained by
the imbedding method. The presence of fat within muscle fibers
is well enough known, but I have thus far found no reference in
the literature to any such concentrated loading as is represented 1n
this dark muscle from the king salmon. My general report on
this work, together with comparisons with other muscular tissues
in the king salmon will give evidence for making the assumption
expressed in the title, namely, that we are dealing here with a new
and special type of fat-storing muscle.

The functional significance of this type of muscle seems to me
to be found largely in the fact of its ability to store and to liber-
ate again such unusual quantities of fat. The muscle has the
function of lipogenesis strongly developed. It would seem that
it is illustrative of one specific tissue in which Loevenhart’s sug-
gestion of a lipogenesis is a specific physiological function. To
what extent this function supersedes or displaces the general
contractile power, if it does either, remains to be determined.

PLATE 1
EXPLANATION OF FIGURES

1 <A cross section of the superficial lateral or dark muscle of salmon no.
97, a fish from the McCloud River at Baird, California. This preparation was
made as afrozen section of material after a fixation of four months in 10 per cent
formalin. The section was stained 'n an alkaline-alcoholic solution of scarlet red
fat stain and counterstained with Delafield’s haematoxylin. It was mounted in
glycerine. The figure shows a striking amount of fat present throughout the
fibers with a few large globules between the fibers. Magnification, Leitz ocular
3, objective 7, camera lucida outlines.

2 A cross section of the superficial lateral or dark muscle of fish no. 97,
from the McCloud River at Baird, California. The material is stained with Mal-
lory’s aniline blue connective tissue stain. The figure is characterized by the
large number of clear spaces which represent vacuoles produced by the extracting
of the fat in the imbedding process. Magnification, Leitz ocular 3, objective
1/12, camera lucida outlines.

180

A NEW TYPE OF FAT STORING MUSCLE
CHAS. W. GREENE

PLATE 1

‘oA
Zs af, ay
MY

Wz

NN
ZZ
AU

mi

tt
y)\\\

A
Ip

a He
C ISKEINS

,

181

TYPES OF OSTIA NASOLACRIMALIA IN MAN AND
THEIR GENETIC SIGNIFICANCE

J. PARSONS SCHAEFFER

Yale University

From the Anatomical Laboratory of the Yale Medical School

FIFTEEN FIGURES

During a recent investigation on the genesis and development
of the nasolacrimal passages in man, my attention was frequently
directed to the marked variations that exist, in the adult, in the
manner of communication between the ductus nasolacrimalis
and the meatus nasi inferior. I, therefore, wish in this note to
refer to the principal types of ostia nasolacrimalia found, and to
eall attention to their probable genetic significance.

LOCATION

At the outset I may say, according to the material studied,
that the ostium of the ductus nasolacrimalis is invariably located
somewhere on the ventral portion of the lateral wall of the meatus
nasi inferior. The variations encountered are due to differences
in type, position within the above limits, and to duplication.
Notwithstanding that the large series of specimens examined for
the substance of this communication invariably presented the
ostium of the ductus nasolacrimalis on the lateral wall of the
meatus nasi inferior, Geddes! reports an unusual and apparently
unique abnormality in an Irish male subject of the age of twenty-
eight years, in which the ductus nasolacrimalis communicated
with the meatus nasi medius. I will refer to this unusual abnor-
mality in a subsequent paragraph.

Within the limits of the ventral portion of the lateral wall of
the meatus nasi inferior there is considerable variation as to the

1 An abnormal nasal duct. Anatom. Anz., Bd. 37, no. 1, 1910.
183

184 J. PARSONS SCHAEFFER

position of the ostium nasolacrimale. It is located from 15 to
20 mm. dorsal to the limen nasi, and from 30 to 40 mim. dorsal to
the naris (anterior naris). Considerable variation also exists in the
cephalo-caudal plane: It is frequently found in the most cephahe
portion of the meatus nasi inferior, immediately caudal to the
attachment of the concha nasalis inferior to the lateral nasal wall.
Again we see specimens in which the ostium is 10 mm. (rarely
more) caudal to the above point. Between these two extremes
we, of course, encounter ostia at various distances caudal to the
attached border of the concha nasalis inferior (figs. 4, 6, 9 and 11).

DUPLICATION OF THE OSTIUM

The ostium nasolacrimale is usually a single opening. We,
however, encounter specimens in which the opening is duplicated.
A triplicity of the ostium was encountered in the series studied.
Fig. 8 represents a specimen in which the ostium was duplheated.

TYPES OF OSTIA

The unqualified statements found in some of our text-books,
that the ductus nasolacrimalis at the point of communication with
the meatus nasi inferior, is provided with a valve (plica lacrimalis
or the so-called valve of Hasner) is certainly at variance, in many
instances, with the real anatomic condition. A study of this
ostium or communication in a large series of cadavers at once
demonstrates that there is no unvarying typical form but that,
on the other hand, we are dealing with several normal anatomic
lypes of ostia nasolacrimalia. In order to make more compre-
hensible the several types of ostia encountered in this study it
may be well to refer to the illustrations, figs. | to 11, in the aecom-
panying plate. The drawings were made from actual dissections

Figs. | to 11 Drawings of actual dissections illustrating the various types of
ostia nasolacrimalia encountered in this study. The reader is referred to the text
for a further consideration of them. The concha nasalis inferior is represented as
partly cut away so as to expose for study the manner of communication between
the ductus nasolacrimalis and the meatus nasi inferior. ?.

OSTIA NASOLACRIMALIA IN MAN PLATE 1

J. PARSONS SCHAEFFER

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 2

IS5

OSTIA NASOLACRIMALIA IN MAN 187

of the region. The concha nasalis inferior is represented as partly
dissected away so as to expose for study the ostium of the ductus
nasolacrimalis.

In fig. 5 is represented a fairly common type of ostium. The
ductus nasolacrimalis passes through the nasal mucous membrane
rather obliquely. The ostium of the duct is rather indefinite and
slit-like, and is essentially a potential space. It is guarded by a
fold of mucous membrane (plica lacrimalis). The ostium con-
tinues caudally towards the floor of the nose in a very shallow,
gutter-like depression in the nasal mucous membrane—the depres-
sion or gutter becoming shallower and shallower until its ulti-
mate disappearance.

Another very common type of ostium nasolacrimale is repre-
sented in figs. 3 and 6. The ostium is located immediately caudal
to the attached border of the concha nasalis inferior, i.e., at the
most cephalic portion of the meatus nasi inferior. It passes more
or less directly through the mucous membrane of the nose and in
this respect contrasts strongly with the types represented in figs.
5and 10. The ostium of this type (figs. 3 and 6) is unguarded by
folds of mucous membrane, and always presents as a wide, un-
guarded, open-mouthed and more or less circular opening. This
important type of ostium nasolacrimale, of very common occur-
rence, is not even mentioned in many of our texts. It is easily
located and probed, and in all respects stands in marked contrast
with the slit-like types. This type of ostium is not provided with
a ‘valve.’

In fig. 9 is shown a frequent type of ostium. The ostium proper
is usually more or less open and somewhat guarded by a plica
lacrimalis. Extending from the ostium caudally is a rather deep,
gutter-like groove which tends to become deeper as it approaches
the floor of the nose. The gutter does not disappear by becoming
shallower and shallower as in type fig. 5, but it ultimately termi-
nates in a blindly ending pouch in the nasal mucous membrane.

In fig. 10 is represented the extremely narrow, slit-like type of
ostium. It is essentially a potential opening and is well guarded
by a plica lacrimalis. It passes very obliquely through the nasal
mucous membrane. This type of ostium is usually located with

188 J. PARSONS SCHAEFFER

difficulty. The narrow slit extends in a cephalo-caudal direction
and it contrasts somewhat with the wide slit-like type represented
in fig. 5.

Rarely we find an anomalous type of ostium represented in
fig. 7. The ostium is located on a raised or nipple-like projec-
tion of the nasal mucous membrane.

The ostium represented in fig. 4 is also unusual. It is an ex-
tremely small, circular opening located immediately caudal to
the attached border of the concha nasalis inferior. The common
type of ostium for this position is the wide, unguarded, open-
mouthed type.

The other figures represent variations of the principal types
already referred to. The normal anatomic types, since they occur
so frequently in the series of specimens studied, are represented in
figs. 2,5, 6,9and 10. Figs. 4 and 7 are illustrations of anomalous
ostia, occurring very infrequently in the series. Figs. 1, 3, 8 and
11 are variations of the normal anatomic types.

GENETIC SIGNIFICANCE OF THE SEVERAL TYPES OF OSTIA
NASOLACRIMALIA

It is well known from the work of many observers that the naso-
lacrimal passages have their anlage in a thickening of the rete
mucosum of the epidermis, along the floor of the rudimentary
naso-optic furrow. This solid strand of epidermal cells ultimately
becomes entirely isolated from the surface (in man) and for some
time is wholly encompassed by mesenchymal cells.2. From the
ocular end of the isolated epidermal strand of cells develop two
sprouts which become the ductus lacrimales (superior and inferior).
There is also a sprouting from the nasal end of the strand of cells
which in time grows sufficiently to establish coalescence with the
mucous membrane of the lateral wall of the meatus nasi inferior.
The anlages of the nasolacrimal passages later become canalized.?

The manner of coalescence of the strand of epidermal cells
with the nasal mucous membrane especially concerns us with

2 J. Parsons Schaeffer, The genesis and development of the nasolacrimal passages

inman. Amer. Jour. Anat., vol. 13, no. 1, 1912.
3 Loc. cit.

OSTIA NASOLACRIMALIA IN MAN 189

reference to the adult types of ostia nasolacrimalia. The embry-
ology of the nasolacrimal passages demonstrates that the point
of coalescence of the strand of epidermal cells with the mucous
membrane of the lateral wall of the meatus nasi inferior is incon-
stant. The place of coalescence may be at the most cephalic
point of the meatus nasi inferior (fig. 12). On the other hand,

Concha nasalis
media

Ductus naso-
lacrimalis

Meatus nasi
Concha nasalis inferior
r

inferio

Meatus nasi
inferior
Ductus naso-
lacrimalis

Concha nasalis
inferior

Figs. 12 and 13 Outline drawings of frontal sections (from human embryos)
through the region of the developing nasolacrimal passages. Note in fig. 12 that
the strand of epidermal cells, the anlage of the nasal end of the ductus nasolacri-
malis, has coalesced with the mucous membrane of the nose at the most cephalic
portion of the meatus nasi inferior. In fig. 13 the point of coalescence is farther
caudal on the lateral wall of the meatus nasi inferior. The ductus nasolacrimalis
is represented in the drawing as solid (see previous paper as to the time and manner
of canalization of the duct).!

the point of coalescence may be much farther caudal on the lateral
wall of the meatus nasi inferior (fig. 13). There is also consider-
able variation in the ventro-dorsal plane.

The point of coalescence of the strand of epidermal cells with
the nasal mucous membrane, of course, determines the position
of the adult ostium nasolacrimale (figs. 6 and 11). The manner
of this early coalescence also materially influences the type of

4 Loc. cit.

190 J. PARSONS SCHAEFFER

adult ostium: If at the most cephalic point of the meatus nasi
inferior, the ‘penetration’ of the nasal mucous membrane is more
or less direct, and the large, unguarded, open-mouthed ostium
of the adult would very likely result (compare figs. 6 and 12).
On the other hand, if the strand of cells strikes the lateral wall of
the meatus nasi inferior obliquely, the resulting ostium is likely
to be slit-like and more or less guarded by a plica lacrimalis (com-
pare figs. 5 and 10 with fig. 13).

The area of coalescence between the strand of epidermal cells—
the anlage of the nasolacrimal passages—and the nasal mucous
membrane is, at times, quite extensive (fig. 14). In the process of
lumen formation, in such eases, it is reasonable to believe that
several ostia, communicating between the ductus nasolacrimalis
and the meatus nasi inferior, may be formed. I believe that these
extensive areas of coalescence account for most cases of dupli-
cation of the ostium nasolacrimale. Intervening bridges of mu-
cous membrane would, of course, remain as the division planes
between the several ostia (figs. 8 and 14 should be compared).

In the development of the nasolacrimal passages we frequently
see bud-like projections extending from the main strand of epi-
dermal cells (fig. 15). These buds are frequently seen near the
meatus nasi inferior. It is reasonable to believe, in some instances,
that several of these nasal sprouts establish communication with
the meatus nasi inferior. Such a development would account
for a duplication of the ostium of the ductus nasolacrimalis. It
would also account for the rarer condition in which the several
ostia open into independent, short canals which in turn open into
the ductus nasolacrimalis. Usually however the bud-like pro-
jections from the main strand of cells do not develop sufficiently
to establish coalescence with the nasal mucous membrane, but
end blindly. The latter would account for the very common
diverticula from the adult ductus nasolacrimalis.’ Doubtless
many of these buds disappear entirely. It is well known that we
may also have a duplication of one or both of the ductus lacrimales.
This duplication must be explained along similar lines.

5 J. Parsons Schaeffer, Variations in the anatomy of the nasolacrimal passages.
Annals of Surgery, August, 1911.

OSTIA NASOLACRIMALIA IN MAN 191

Concha nasalis.

media
Meatus nasi
inferior
Concha nasalis Ductus naso-
inferior lacrimalis

Meatus nasi ~—"

communis Anlage of the naso-
lacrimal passages
entirely isolated

Concha nasalis
inferior from the surface

Fig. 14 Outline drawing of a frontal section through the developing ductus
nasolacrimalis (human embryo). The duct is represented as solid. Note the
very extensive coalescence of the duct with the nasal mucous membrane. Com-
pare with figs. 12 and 13.

Fig. 15 Frontal section through the nasal fossa and the anlage of the nasolac-
rimal passages, from a human embryo aged forty-three days. Note that the anlage
is entirely isolated from the surface. The ductus lacrimales have started to sprout
from the ocular end of the strand of epidermal cells. The nasal end of the strand
of cells has not developed sufficiently to come in contact with the nasal mucous
membrane. Especially note the lateral buds from the main strand of cells.
xX 16.5.

The case of Geddes, in which the ductus nasolacrimalis com-
municated with the meatus nasi medius, can be explained by a
lateral bud from the main strand (fig. 15). In such a case the
accessory or lateral bud, instead of ending blindly or developing
sufficiently to establish communication with the meatus nasi
inferior, established connections with the meatus nasi medius.
For some reason or other, the nasal end of the main strand of cells
did not establish connections with the meatus nasi inferior. If,
on the other hand, the connection was established, the cord not
becoming canalized, would in all likelihood undergo resorption
caudal to the sprout which established the definitive connections
with the meatus nasi medius. That Geddes was dealing with a
true portion of the ductus nasolacrimalis and not with a false

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 2

192 J. PARSONS SCHAEFFER

passageway is proven by the fact that the lumen of the duct lead-
ing from the meatus nasi medius was lined with similar epithelium
to that lining the remainder of the ductus nasolacrimalis.

If in the cases, where we have an extensive coalescence between
the strand of epidermal cells and the nasal mucous membrane
(fig. 14), the whole area of coalescence should become patent,
we would readily duplicate some of the other adult types of ostia
represented in the plate, figs. 1 to 11, making due allowance for
changes in the further development of the ostium.

I wish to take this opportunity for expressing grateful acknowl-
edgment to Professcr Kerr for sending me a series of specimens
for study from the Cornell anatomical collection. The other
material studied was from the Yale anatomical series. Fully
two hundred ostia nasolacrimalia were included in this investi-
gation.

THE DEVELOPMENT OF THE AXIAL VEINS
AND LYMPHATICS IN TRAGULUS
MEMINNA, ERXLEBEN

FREDERICK TILNEY

From the Anatomical Laboratory, Columbia University
FOURTEEN FIGURES!
MATERIAL

The material on which this investigation is based consists
of five embryos of the ruminant ungulate, Tragulus meminna
(Indian Chevrotain) measuring 5mm., 6 mm., 13 mm., 20 mm.
and 23 mm. respectively. These were obtained by exchange
from the Smithsonian Institute through Dr. A. Hrdlicka. The
series represents the entire embryological material of this form
collected during a number of years in Borneo and adjacent
islands of the Malay Archipelago by Dr. Abbott of the Insti-
tution.

In view of the very great importance of Tragulus to the phylo-
genetic interpretation of the marsupial and placental vascular
system, and of the probability that further embryos of this
animal will not soon again become available for study, it seems
advisable to publish the main results of the investigation based
on the material now in hand. This becomes all the more im-
perative since we possess the complete account of the develop-
ment of the posteava in embryos of Didelphis marsupialis fur-
nished by McClure (’03), and hence have the opportunity of
directly comparing the available series of Tragulus with the
corresponding marsupial stages.

The two smaller (6 mm. and 6 ,mm) and the two larger (20
mm. and 23 mm.) embryos of Tragulus here described proved
to be admirably fixed and preserved and well adapted for minute
and detailed study. The intermediate embryo of 13 mm. is
in a less satisfactory condition and has hence only been used

‘Expense of illustrations borne by author.

193

194 FREDERICK TILNEY

to bridge the gaps between: the earlier and the later stages in
‘certain regions, in which its preservation permitted definite
conclusions.

In a single adult specimen of Tragulus, McClure (06) found
that the post-renal division of the post-cava was placed directly
in front of the aorta and formed by the union of the common
iliac veins, this union taking place ventral to the aorta. These
relations correspond to one of the three types of post-cava de-
scribed by him (’03) in Didelphis marsupialis and figured in
his plate 2 as no. 8. McClure concludes that this condition
allies the venous organization of Tragulus more closely to that
of the marsupials than to any of the known ruminants. In
a verbal communication to the writer, Dr. McClure states that his
earlier observations have been confirmed by the subsequent exami-
nation of three additional adult specimens of the animal.

Beddard (’07) observed the same relations of the aorta to the
post-renal segment of the post-cava in three adult specimens of
Tragulus. One male adult and three male fetuses in the Col-
umbia collection show the post-renal segment of the post-cava
directly in front of the aorta. The conditions in the adult speci-
men are illustrated in fig. 1. In Tragulus the axial venous chan-
nel from the confluence of the internal and external iliac veins
(fig. 1, 36 and 37) to the renal level consists of two parts; a,
the common iliac veins (fig 1, 38) (paired portion of the post-
cava, this latter term being used to keep the account in accord
with MecClure’s description) and b, the unpaired portion of the
post-cava, ventral to the aorta (fig. 1, 39). The unpaired por-
tion is 6.5em. in length and extends from the confluence of the
common iliac veins to the renal anastomosis. The common
iliacs are 4.9 em. in length and extend from the confluence of
the internal and external iliac veins to the paired portion of the
post-cava. Each common iliac vein receives the sex vein of
the corresponding side (fig. 1, 40). Tragulus differs, in this
respect, from the majority of marsupials in which the sex veins,
as has already been shown by Schulte (07) and Schulte and
Tilney (’09), empty into the unpaired portion of the post-cava
or into the post-cava and left renal vein.

VEINS AND LYMPHATICS IN TRAGULUS 195

With reference to the development of the systemic lymphatics
the material does not furnish conclusive evidence in all particu-
lars. The organization of the lymph sac, however, is identical
with that observed in the domestic cat by Huntington and
McClure (710). The younger embryos present all the characteris-
tics of the primary venous stage, while the older specimens show
the several details of structure typical of the late veno-lymphatic,
pre-lymphatic and definite lymphatic stages. The systemic
lymphatics, at first, are entirely independent of the lymph sacs
and only secondarily acquire connections with them in completing
the thoracic duct formation. In the specimens examined they
showed no genetic relation to the veins, although their exact
histogenesis could not be determined on account of the limited
material.

The Born method of reconstruction was employed with a
magnification of 100 diameters for purposes of topographical
study and 200 diameters for obtaining regional detail. All meas-
urements in the embryo were made from sections after fixation
upon the slide and, in most cases, computed from figures obtained
at a given magnification. None of the embryos was injected.

Venous organization of the 5 mm. embryo (fig. 2)

This embryo presents an imperfectly closed neural tube in
the entire spinal region. Its venous channels have a bilateral,
symmetrical arrangement, there being four sets of paired trunks,
all parallel with the long axis of the body. In addition to these
drainage lines there are two plexuses, one, the perimescnephroic
plexus surrounding the mesonephros, the other, the umbilico-
cardinal plexus which is still active in draining the territory of
the body wall into the umbilical and post-cardinal veins.

1. The umbilical veins (fig. 2, 6). These are the largest chan-
nels present; they are as yet equal in size and show no tendency:
to lose their bilateral, symmetrical arrangement. No com-
munication has been established with the hepatic sinusoids, so
that the two umbilical veins are still independent vessels. They
are, to a considerable extent, concerned in drainage of the body

196 FREDERICK TILNEY

as well as in placental circulation. This is shown by the rich
umbilico-cardinal plexus which permeates the lateral somatic
wall. Each vein opens into the sinus venosus of its own side
in common with the vitelline entrance and forward of the Cuvier-
ian approach (fig. 2, 10).

2. The omphalomesenteric veins (fig. 2, 9). These veins, also
paired and symmetrical, are smaller than the umbilicals. Al-
though their point of entrance into the sinus venosus is the
most mesial and ventral of all vessels entering the heart, the
plane of the main trunk passes about midway between the um-
bilical and post-cardinal veins, so that the course of the vessel
as it approaches the sinus is deflected sharply forward and inward.
Two divisions of this vessel may be recognized, an intra-hepatic
portion which has entered into the formation of the sinusoids
and is situated at the dorso-mesial angle of the liver. It thus
comes into close proximity to the coelomic angle. The second
or omphalic portion passes out upon the yolk sac, leaving the
body through the external umbilicus. In this way, each om-
phalomesenteric vessel, in its intrahepatic portion, has its long
axis parallel both with the umbilical and post-cardinal veins,
but interposed between it and each of these channels are the
mesial branches of the perimesonephroic plexus. No evidence
of any distinctly mesenteric branch was found.

3. The post-cardinal veins (fig. 2, 5). These are present as
a pair of vessels of medium size which have become definite
channels throughout their entire length. They hold their typi-
eal position dorso-lateral to the mesonephros and still retain
part of the plexiform connection with the umbilical veins (fig. 2,
7). This connection becomes greatly reduced as the promontory
is reached. The size of the vessel increases as it ascends to
become confluent with the pre-cardinal, but before doing so it
turns forward alinost at right angles to its axial course and by
this horizontal arm enters the duct of Cuvier. Immediately
below this. sharp forward bend, the post-cardinal receives a
large tributary from the cephalic extremity of the perimeson-
nephroic plexus. In the greater part of its course many branches
from the plexus join this vessel. At some points the vein has

VEINS AND LYMPHATICS IN TRAGULUS 197

a tendency to reduplicate its channel; the redundant element
then lies mesial to it. In the region of the promontory several
small dorsal tributaries enter the vessel.

4. The pre-cardinal veins (fig. 2, 2). These vessels consti-
tute the fourth pair of symmetrical channels and appear in the
typical position of the early stages. The main trunk is parallel
to the neural tube and descends toward the Cuvierian duct,
into which it enters, after curving sharply forward and inward.
It follows the direction of the neural axis cephalad, adapts itself
to the cervical flexure, and turns forward nearly at right angles
to its descending portion. In this way the vessel may be
regarded as presenting two limbs, a, an arched or horizontal limb
which is assuming the position of the vena capitis lateralis, and
b, a straight or descending limb in the jugular line. The straight
limb receives several small, dorsal tributaries whose develop-
ment plays an important rdle in the genesis of the veno-lym-
phatics.

5. The perimesonephroic plexus (fig. 2, 8). This forms a
dense network around the mesonephros throughout its entire
length. It has its greatest density along the ventral surface of
the organ. At many points the large mesonephroic sinuses com-
municate with the plexiform channels. The plexus becomes much
reduced in size on the mesial and lateral surfaces of the meso-
nephros, where an irregular series of anastomosing channels
establish communication with the post-cardinal vein. During
the course of development, the ventral portion of the peri-
mesonephroic plexus takes part in the evolution of certain
axial vessels; the mesial anastomosing branches give rise to
the lateral portion of periaortic plexus (fig. 3, 46 and fig. 4,
46).

6. The umbilco-post cardinal plexus fig. 2, 7). This connects
the umbilical and post-cardinal veins. It represents the inter-
mediate phase during which the umbilical vein has preserved
its capacity as a drainage line for the body wall and is about to
give this function over to the post-cardinal. In many places
the plexus has already broken down and appears to be selecting
the post-cardinal as its ultimate channel.

198 FREDERICK TILNEY

Such evidence of lymphatic organization as this embryo affords
confines itself to the primitive venous fundaments of the jugu-
lar lymph saes. If these elements did not so nearly duplicate
the conditions of the early ground plan of the sac in the eat,
it would be difficult, without further material, to recognize their
true significance. They appear as several dorsal tributaries in
connection with the straight portion of the pre-cardinal vein,
with the promontory and -with the cephalic extremity of the
post-cardinal vein. Three of these branches are related to the
straight limb of the pre-cardinal, one to the promontory and
one to the post-cardinal The tributaries in relation with the
pre-cardinal vein are irregular and have the appearance of redun-
dancies in the vein line. One in particular, the third and most
caudal of the series, is larger and more irregular than the rest.
Another characteristic of early lymph sac formation is the marked
increase in size of the pre-cardinal vein and promontory as they
draw together toward the Cuvierian duct. This augmentation
does not appear to be in response to the entrance of new tributary
lines, but, as in the cat, seems to be a redundant growth of the
venous network prior to its differentiation into distinct veno-
lymphatic channels. Upon these grounds, and especially because
of their striking correspondence to the conditions observed in
the cat, these structures may fairly be taken to represent the
elements which determine the early or primary venous stage of
lymph sac organization.

The venous organization of the 6 mm. embryo (fig. 5 and fig. 6)

The advance observed in this embryo is characterized by the
emergence of a definite sub-cardinal drainage line out of the
perimesonephroic plexus and the disturbance of the bilateral
symmetry in the umbilical veins. There are five sets of paired
venous channels, the general disposition of whose course is paral-
lel to the long axis of the body. The perimesonephroic and
umbilico-cardinal plexuses have lost their definite outlines.

1. The omphalomesenteric veins (fig. 6, 9). Notwithstanding
the fact that the cephalic portion of each vitelline vein has become

VEINS AND LYMPHATICS IN TRAGULUS 199

involved in the formation of the hepatic sinusoids, it is still possi-
ble to recognize two well defined venous channels passing through
the liver, parallel with the body axis and situated in the angle
formed by the junction‘of the dorsal and mesial hepatic surfaces.
The anterior walls of these channels are subject to great irregu-
larities occasioned by their relations to the hepatic sinusoids;
their dorsal and mesial walls are well defined. Each channel
is situated opposite the coelomic angle, dorso-lateral to the gut
tube, directly ventral to the sub-cardinal vein and ventro-mesial
to the post-cardinal vein. The two intra-hepatic vitelline vessels
not only drain the sinusoids of their respective sides but commu-
municate with each other by several transverse anastomoses which
follow a semi-circular course in front of the gut tube. The most
cephalic of these anastomotic channels is the largest and lies at
a level a little below the vitelline entrance into the sinus venosus.
As it approaches this sinus, each vitelline vein undergoes a change
in course, curving outward and forward, then upward and in-
ward to a point slightly mesial to the entrance of the umbilical
vein. The intra-hepatic vitelline channels may be traced cau-
dad to a large sinus-like blood space situated immediately below
the liver. This large sinus has been greatly augmented by the
confluence of the two umbilical veins and the omphalic portions
of the vitellines. From its position it may be conveniently
referred to as the sub-hepatic sinus (fig. 6, 17), which, in this
embryo, receives blood from the large umbilical veins and a
single omphalic vessel of medium size. It delivers blood to
the hepatic sinusoids, and the two intra-hepatic vitelline chan-
nels, having a more ample communication with the channel of
the right side. ‘Two veins of small size appear in the gut wall
in regions below the liver; their relations identify them as the
mesenteric veins. The left mesenteric vein is the larger of the
two; it establishes connection with the left intra-hepatic vitel-
line channel at a distance of 100 cephalad of that vessel’s de-
parture from the sub-hepatic sinus. The right and smaller
mesenteric vein joins the left vein by a semicircular anastomosis
behind the gut at a point 300 below the entrance of the latter
vein into the left intra-hepatic vitelline channel. The follow-

200 FREDERICK TILNEY

ing components:-may be distinguished in the omphalomesenteric
drainage line.

a. The omphalic or vitelline element, a single vein which
drains the blood from the yolk sac into the sub-hepatic sinus.

b. The intra-hepatic vitelline element, appearing as a right
and left channel, of which the right is to persist as the hepatic
portion of the post-cava, receiving the hepatic revehent veins
and, during intra-uterine life, the ductus venosus; the left chan-
nel takes part in the development of the hepatic revehent veins
and sinusoids. By its early association with the mesenteric
veins, this vessel aids in the formation of the hepatic portal
system of post-natal life.

c. The mesenteric elements, appearing in the gut wall as
right and left veins, the left communicating with the left intra-
hepatic channel, while the right vessel gives its drainage to the
left mesenteric by a semicircular anastomosis behind the gut.
These latter elements give rise to the definitive portal system.

2. The umbilical veins (fig. 5, 6), These vessels, in greater
part, still retain their symmetrical arrangement, but the general
venous symmetry of the earlier stage is now to some degree
disturbed in the region in which the omphalic portion of the
vitelline veins become confluent with the umbilical veins to
form the sub-hepatic sinus. The significance of the umbilical
veins below the sinus appears to be so different from that above
it as to justify the distinction of infra-sinal and supra-sinal por-
tions in either vein. The capacity of the infra-sinal portion
is several times that of the supra-sinal. The left infra-sinal
segment has increased greatly in size and, in places, is a double
channel. The right infra-sinal segment is also double, but for
a shorter distance. The supra-sinal segment of the umbilical
vein is a relatively slender vessel, extending from the outer side
of the sub-hepatic sinus to the sinus venosus. In its entire
course cephalad it receives numerous anastomotic branches from
the umbilico-cardinal plexus and is thus still largely concerned
in drainage of the body wall. It aids in the return of blood
from the sub-hepatic sinus and so affords two direct passages

VEINS AND LYMPHATICS IN TRAGULUS 201

from the placenta to the heart, pending the establishment of
the well defined ductus venosus.

3. The post-cardinal veins (fig. 5, 5). These vessels occupy
the typical post-cardinal position. They may be traced caudad
as far as the cloaca and there lose their identity in an indefinite
plexus. Followed cephalad, each vessel is seen to attain its
maximum diameter at the level of the middle of the mesonephros.
They still give evidence of the early influence of the perimeso-
nephroic plexus; in several regions a channel is observed lying
parallel to the main vessel and continuous with it above and
below. The redundant element may be ental or ectal in position.
The genesis of this intermediate cardinal element is not yet
entirely clear. Like the sub-cardinal vein, it seems to represent
the general process by which definite channels are evolved from
more primitive plexus formations along the lines of axial growth
and drainage. The cephalic portion of the post-cardinal vein
is still intimately related to the supra-sinal segment of the umbili-
cal vein by branches of the umbilico-cardinal plexus. In the
region of the promontory, each vessel increases in size and re-
ceives a number of dorsal tributaries.

4. The sub-cardinal veins and the perimesonephroic plexus (fig.
5, 12). One of the characteristic changes in this stage is the
development of a definite sub-cardinal vein out of the perimeso-
nephroic plexus. This vein is a slender vessel with here and
there a remnant of the former plexus draining into it. This
is particularly the case about the lower pole of the mesonephros
where the plexus is still rich and shows a tendency to give rise
to other axial lines besides the sub-cardinal channel. Here it
is possible to recognize a longitudinal drainage line which is being
selected along the inner aspect of the plexus, mesial to the sub-
cardinal itself and ventro-lateral to the aorta. This element
is in the position of the cardinal collateral channel of McClure.
Because of its course and relations, three portions of the sub-
cardinal vein may be recognized. 1. The caudal vertical por-
tions 2, the middle or horizontal portion and 3, ‘the cephalic
vertical portion. The caudal vertical portion begins as a slender
vessel from the caudal extremity of the peri-mesonephroic plexus;

202 FREDERICK TILNEY

it passes cephalad along the ventro-mesial surface of the meso-
nephros, having its long axis parallel with the aorta. Upon
reaching the level of the sub-hepatic sinus the vessel swings
dorsad and slightly lateral so that its axis is now turned almost
at right angles to its caudal portion. The position which it
now occupies distinguishes it as the middle or horizontal portion
of the vein. In this portion of its course the vessel closely fol-
lows the direction of the intra-hepatic segment of the vitelline
vein as the latter is sweeping dorso-cephalad away from the
sub-hepatic sinus. The two vessels are still completely sepa-
rated by the coelom with a measured distance of 10u between
them. The angle determined by the junction of the caudal
and middle portions of the vein lies in a direct line with the intra-
hepatic portion of the vitelline vein, so that the projection cau-
dad of the axis of the latter vessel would coincide with the axis
of the caudal portion of the sub-cardinal vein, and would thus
foreshadow the axis of the future post-cava (on the right side).
It is not possible to state that the region of closest approach
is the region of the final confluence of these two vessels, but it
seems probable from the relations of the sub-hepatic sinus, that
the ductus venosus, intrahepatic vitelline and sub-cardinal veins
of the right side unite at this, rather than at some more cephalic
point. The cephalic vertical portion continues the vein cephalad
and is the smallest as well as the shortest of the three divisions.
It taps the post-cardinal vein 200 above the cephalic pole of
the mesonephros. It is parallel to the intra-hepatic portion of
the omphalomesensenteric vein, from which it is separated by
a mean distance of 20un.

5. The pre-cardinal veins (fig. 5, 2) and veno-lymphatics. The
general arrangement of the pre-cardinal veins has undergone
no marked change. The vessels still present an arched and a
a straight limb. There has been, however, considerable modifi-
cation in the character of the dorsal tributaries. These have
expanded and become confluent in several places. The expan-
sion has continued after confluence has taken place and the
channels have been converted into irregular spaces. The four
dorsal tributaries to the pre-cardinal vein have all united, while

VEINS AND LYMPHATICS IN TRAGULUS 203

the tributaries coming into the promontory and cephalic portion
of the post-cardinal have joined to form a large blood space.
The tendency of the pre-cardinal and post-cardinal veins to
increase in size as they approach the duct of Cuvier is still evi-
dent. On the right side there are clear signs of a beginning
fenestration in the base of the pre-cardinal, thus indicating the
inception of the para-pre-cardinal line.

Marked changes are met with in passing from the primitive
organization of the 6 mm. embryo to the condition of the 20 mm.
embryo. As touching upon the stages intermediate between
the two, reference will be made to the 13 mm. specimen. A
noteworthy acquisition in the 20 mm. embryo is found in the
now fully developed lymph sacs, and also in the formation of
the main segments of the systemic lymph channel. These two
elements are as yet entirely distinct and separate. The plan
of venous drainage foreshadowed in the 6 mm. embryo has been
carried well on towards completion. In this process a single
post-cava has been acquired as far as the inter-renal anastomosis
while below that level, two large symmetrical channels, repre-
senting the cardinal collateral veins, codperate in the formation
of a double post-cava.

The lymphatic organization in the 20 mm. embryo (fig. 7)

In the 20 mm. embryo, the jugular lymph sacs (fig. 7, 25)
are situated in the neck region, one on either side, on the lateral
aspect of the great vessels and nerves. Each sac has a general
wedge shape, with its base looking inward and forward, while
its edge lies between the vertebral column and the dorso-lateral
surface of the body. Its greatest diameters are attained about
midway between its cephalic and caudal poles. The 3rd, 4th,
5th and 6th cervical ganglia lie dorsal to the sac. The third
cervical nerve traverses its cephalic pole, the sixth nerve passes
beneath its caudal pole, while the fourth and fifth nerves go.
directly through it. The right sac is 1.61 mm. in length, the
left sac 1.73 mm. The maximum ventro-dorsal diameter of
the right sac is 1.68 mm., that of the left sac being 1.6mm. The

204 FREDERICK TILNEY

maximum transverse diameter of the right sac is .54 mm., that
of the left sac is .48 mm. It will be convenient to describe the
structure as having a body with caudal and cephalic processes.

The body of the lymph sac (fig. 7, 25). This portion of the
sac appears as a prominent feature of cross sections in the neck
region. The walls of the sacs are thinner than those of the veins.
The mesial and ventral walls are smooth and regular, the lateral
wall presents many irregular projections. The broad ventral
portion of the sac lies directly back of the jugular vein (fig. 7, 35).
Mesially it is in relation with the carotid artery, the vagus and
sympathetic nerves, but separated from them by an extensive
plexus of venous channels draining into the jugular system.
In its subsequent development, this plexus allies itself with the
lymphatic system.

Processes of the sac. The contour of the sac becomes irregu-
lar in several regions by prolongations from its walls. These
prolongations serve as the processes by which the sac acquires
its ultimate connections with the venous system and the systemic
lymphatics. From the cephalic pole a large number of processes
reach upward into the head region and end blindly. Some of
these prolongations serve to connect with the systemic lymphatic
trunks from above. The caudal processes are larger and show
a more definite arrangement. Their number and disposition
differ on the two sides. This difference depends upon the rela-
tions which each sac bears to the adjacent veins. The left sac
is still freely connected with the jugular vein; the right sac is
entirely cut off all venous connection. The left sac is typical
of the conditions in the late veno-lymphatie period. The right
sac affords a good example of the pre-lymphatic stage. Due
to its free communication with the left jugular vein, the left
sac is filled with blood. Caudally, it is divided into two main
processes, one of which turns mesad and ventrad in the direction
of the jugular vein, the ventro-mesial process; the other continues
caudad in line with the general axis of the sac, the dorso-lateral
process. The ventro-mesial process enters directly into the vein,
On the right side the sac has already been emptied of blood and
has lost connection with the jugular vein. The dorso-lateral
process (fig. 7, 26) is present on each side. It does not connect

VEINS AND LYMPHATICS IN TRAGULUS 205

directly with the veins but proceeds caudad to form three other
processes, namely, the lateral process or subcutaneous duct,
the dorsal descending process or thoracic duct approach (fig.
7, 26a) and the ventral descending or broncho-mediastinal ap-
proach. The ventral and dorsal descending processes become
divergent immediately above the cephalic vein; the ventral proc-
ess passes caudad on the ventro-lateral aspect of the jugular
vein to join the broncho-mediatinal systemic channels on the
right side. On the left this connection is not completed. The
dorsal descending process extends a short distance caudad, dorso-
lateral to the thyreo-cervical artery to end blindly.

Organization of the systemic lymphatic drainage lines. The
evidence concerning the development of the lymph sac in Tra-
gulus shows that this structure is derived from the venous system.
In its early stages the sac is wholly independent of the systemic
lymphatics. Subsequently it joins with the systemic lymphatic
channels, and thus serves as the connecting link between these
channels and the venous system. In the material studied it
was impossible to discern any genetic relation between the devel-
oping systemic lymphatics and the veins. Such regions as gave
the earliest pictures of the organization of the systemic lym-
phatics revealed these elements as independent mesenchymal
spaces, at first presenting the form of a plexus. By a process
of expansion and confluence this plexus comes to form definite
channels. In the 20 mm. embryo development has advanced
too far to offer anything that is conclusive as to the actual histo-
genesis of these lymph spaces. The process referred to as con-
fluence and expansion is not carried on with the same degree
of rapidity or effectiveness in all regions of the embryo; it confines
its activities to certain selected districts which remain, for some
time, separate. The systemic lymphatic channels arise from
three major segments which, because of their relations to the
venous system, may be termed the azygos, pre-azygos and post-
azygos segments, corresponding to similar segments in the cat
as recently described by Huntington (’10).

1. The azygos segment (fig. 7, 28). It is in this segment that
the systemic lymphatic channels have attained their greatest
development. Longitudinally the segment reaches caudad from

206 FREDERICK TILNEY

the azygos-Cuvierian junction to a level slightly above the inter-
renal anastomosis. It is notable for its unusual size as well
as for the striking resemblance it bears to the reptilian type of
axial lymph channel (fig. 9, 28). On approaching its cephalic
extremity the channel breaks up into a rich plexus (fig. 8, 50).
The same is true of its caudal extremity. The vessel on the
left side is the larger. It is situated ventro-mesial to the azygos
vein and dorso-lateral of the aorta. The right channel holds
the same relative position. After proceeding a short distance
caudad these two parallel channels rapidly expand and become
confluent across the median line behind the aorta. The azygos
segment interposes itself between that vessel and the two azygos
veins, which latter are connected with each other by anastomos-
ing vessels passing behind the aorta. The relations of the azygos
segment to the other large channels are illustrated in cross section
in fig. 9.

2. The post-azygos segment (fig. 7, 29). This portion of the
systemic lymphatic channel pertains to the abdominal region
where it appears as an irregular vessel, at times double and
again fused, behind the aorta. It continues downward in this
condition to the iliac bifurcation where it becomes a considerably
dilated, single vessel again bifurcating at its caudal extremity.
At the point of this bifurcation it alters its relations in such a
manner as to lie lateral to the iliac vein, taking up the ultimate
position of the ilio-lumbar lymphatic trunks. At its cephalic
extremity it ends in a plexus which has already established sev-
eral connections with the plexus at the caudal extremity of the
azygos segment.

3. The pre-azygos segmént (fig. 7, 30 and 31). In this portion
of the systemic lymphatic channel the process of expansion is
apparently less active than elsewhere. Two general lines of
development may be traced, one on the right in relation to rem-
nant of the right pre-cardinal vein and aorta, the other on the
left in relation to the large brachio-cephalic arterial trunk. Each
line presents a cephalic and caudal element. On the right side the
caudal element communicates with the plexus at the cephalic end
of the azygos segment (fig. 8, 50), and from here extends cephalad

VEINS AND LYMPHATICS IN TRAGULUS 207

between the aorta and pre-cardinal vein as far cephalad as the
arch of the aorta. The cephalic element begins as a plexus at
the bifurcation of the brachio-cephalic artery and continues cepha-
lad dorso-lateral of the jugular vein as far as the jugulo-sub-
clavian junction. At this point it is separated by an interval
of approximately 0.1 mm. from the blind end of the dorsal descend-
ing process of the lymph sac ( thoracic duct approach). The
pre-azygos segment on the left side is even more divided, for
although it presents a cephalic and caudal division, each of
which appears as a clear cut channel, it was impossible to detect
any connection between these two divisions on the one hand
and the azygos segment on the other. It ends blindly above
and appears to have no connection with the thoracic duct approach
of the sac, from which it is separated by an interval of about
0.1 mm. Its caudal end is also independent of any connection.
The caudal division begins as a distinct channel at the derivation
of the internal mammary artery and extends almost as far caudad
as the junction of the brachio-cephalic trunk with the aortic
arch. So that, above the azygos segment, the line along which
the thoracic duct is destined to develop its connection with the
duct approach of the lymph sac consists of two as yet indepen-
dent elements, the cephalic and caudal divisions of the left pre-
azygos segment.

A summary of the systemic lymphatic organization shows
that the following elements must be recognized in the future
line of the thoracic duct:

(a). The azygos segment (figs. 7, 28 and 9, 28). A spacious
sinus-like channel, in places entirely surrounding the aorta and
ending at either extremity in a plexus, in many regions resembling
the peri-aortic lymphatic sinus of reptiles.

(b. The post-azygos segment,(fig. 7, 29), a series of alter-
nating plexiform and sinus-like channels, connected cephalad
with the caudal plexus of the azygos segment.

(c). The pre-azygos segment,(fig. 7, 30 and 31), consisting
on the right side of two independent divisions, the cephalic and
caudal, of which the latter is connected with the cephalic plexus
of the azygos segment while the former has not yet established

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 2

208 FREDERICK TILNEY

communication with the thoracic duct approach of the lymph sac.
On the left side, the two divisions are also present, but neither
has made connection with other parts of the lymphatic system.

Lymphatic organization in the 23 mm. embryo

Axial lymphatic organization has been carried to its consum-
mation in the 23 mm. embryo. The imperfectly crystallized
conditions of the next younger specimen have already marked
out the line along which this development would proceed.

Body of the lymph sac (fig. 10). The body of the sac on both
sides lies in front of the 3rd, 4th, and 5th cervical ganglia, the
6th ganglion and the interspace above it lying entirely below
the caudal limit of the vesicle. The long body axis now bears
the ratio of 17 to 1 to the long axis of the lymph sac, whereas
in the 20 mm. embryo this ratio was 10 to 1. Measurements
of the sac as computed from the mounted sections show that
this decrease in size is not merely relative but absolute, as the
following values of the long axes show:

23-MM. EMBRYO 20-MM. EMBRYO
mm. mm.
IRF OG 3G 3. cin as eh ame nara 1.49 | AGT

1GAISEVC80 45 hae ek eee: 1.61 Le

The ventro-dorsal and transverse diameters also show a decrease-
That this attenuation has chiefly affected the more ventral por-
tion of the sac is shown by the fact that the two nerves (4th
and 5th cervicals) which in the 20mm. embryo passed through
the sac have now freed themselves, while it also appears that
the process which released them has at the same time produced
a neck in the sac itself. By this neck the sac approaches the
jugular vein and attains its ultimate systenal lymphatic and
venous connections. It will be obvicus, therefore, that this cervi-
cal portion corresponds in general to the caudal extremity in
the sac of the next younger specimen.

Processes of the sac. Several of the cephalic processes may be
traced for a considerable distance into the head region and un-

VEINS AND LYMPHATICS IN TRAGULUS 209

doubtedly denote the connections of the sac with lymphatic
trunk lines of the head. The main interest with reference to
these processes centers about those which are derived from the
caudal extremity or what may be termed the cervix of the sac
Upon reaching the level marking the entrance of the cephalic
vein into the jugular, the cervix of the sac breaks up into four
processes, namely:

1. The dorsal process which extends dorsad accompanying
the cephalic vein and receives the dorsal somatic tributary of
the sac, the subcutaneous duct.

2. The mesial process, the now much reduced primary veno-
lymphatic connection which extends mesad dorsal to the jugular
vein but does not tap into it.

3. The dorsal descending process which in the 20 mm. embryo
was designated the thoracic duct approach and which has now
acquired its full connection with the preazygos segment of the
duct, especially on the left side. This process descends along
the dorso-lateral surface of the thyreo-cervical artery.

4. The ventral descending process which affords communi-
cation with the broncho-mediastinal channels. It appears in
a position ventral to the thyreo-cervical artery. As this process
enters the sac it forms an acute angle with the dorsal descending
process which lodges the cephalic vein as it is opening into the
jugulo-subclavian junction. In addition to the broncho-medias-
tinal approach, the ventral descending process has developed
still a third, the jugulo-subclavian approach. This is a slender
prolongation from the mesial side of the process which, upon
reaching the jugulo-subclavian angle, forms a direct communi-
cation with the venous system. This is the so-called secondary
venous tap of the lymphatic into the venous system and by its
establishment determines the transition from the prelymphatic
to the lymphatic stage. The secondary tap is made in a charac-
teristic manner. From the wall of the vein at its jugulo-sub-
elavian junction a process which has the appearance of a tubular
redundancy pushes its way cephalad between the lateral vein
wall and the mesial wall of the sac. After a short distance it
meets and opens into the jugulo-subelavian approach of the
lymph sac thus producing a channel between the latter and the

210 FREDERICK TILNEY

venous system. The venous orifice of this duct-lke structure
is wide; the saccular orifice is elongated and narrow, while the
duct itself is placed between the walls of the sac and the vein.
This arrangement makes it appear that distension of the vein
would act on the jugulo-subclavian approach in such a way as
to produce a valvular effect. By this means the sac has acquired
its ultimate connection with the venous system (fig. 11, A, B,
ang ©).

The systemic lymphatic drainage line. The transition from
the conditions of the systemic lymphatic organization of the 20
mm. embryo to those of the 23 mm. is characterized by a con-
fluence of the several previously established segments, with the
result that the thoracic duct is now a continuous channel and
at the same time has upon the left side, acquired connection with
the lymph sac.

The azygos segment. This portion of the systemic lymphatic
line gives evidence of the least change of character. It is still
a capacious channel situated dorsal to the aorta, in places sur-
rounding the vessel. If changed at all, it has somewhat lost
in capacity, especially because its periaortic plexuses are less
rich and numerous. In relation to the azygos veins, it still is
interposed between them and the aorta. The plexuses which,
in the early conditions, were observed at its cephalic and caudal
extremities have given place to definite channels (fig. 12). From
the cephalic plexus there has arisen a single large trunk which
is paralleled by a second small lymph vessel, both of which pass
over into the pre-azygos segment. The caudal plexus develops
two channels which communicate with the post-azygos segment.

Pre-azygos segment (fig. 12). Here the greatest change has
occurred, for the plexiform and irregular channels of the earlier
pre-azygos segment have become defined as two parallel trunks.
The larger of these is the more constant and apparently repre-
sents the main line of drainage. Upon the leftside the confluence
of the several divisions of the pre-azygos segment has been
carried so far as to form a complete connecting vessel in com-
munication with the dorsal descending process of the lymph
sac of that side. On the right side this confluence is not as com-

VEINS AND LYMPHATICS IN TRAGULUS PAL

plete. The cephalic division of the pre-azygos segment has met
and fused with the thoracic duct approach of the sac. The
caudal division is still independent.

Post-azygos segment (fig. 10, 29). This segment has also at-
tained more definite outline. Its caudal dilatation is larger and
its ilio-lumbar appendages more extensive, so that upon the left
side there is an uninterrupted thoracic duct line, which has
resulted from the fusion of the pre-azygos, azygos and post-
azygos segments. The right duct line is still incomplete. The
ventral descending process of the sac on both sides has already
established connection with the truncus broncho-mediastinalis.
In the main, this trunk is still a dense plexus situated ventral
to the thymus; in several places, however, it loses its plexiform
character to become a distinct channel. Retaining these general
relations to the thymus; it passes to the caudal extremity of
that organ, where it undergoes considerable reduction, but may
be traced across the aortic arch to the root of the lung.

THE VENOUS ORGANIZATION IN THE LATE EMBRYONIC STAGES

The advance in the venous organization in the 20 mm. embryo
depends on a modification in the relations between the umbilical
and omphalomesenteric veins. Coalition of the right sub-cardi-
nal with the right emphalomosenteric has produced a definite
post-caval system as far caudad as the renal anastomosis. Below
this level the cava arises in a manner somewhat different from
that observed in the majority of mammals already studied.

Renal anastomosis and cardinal collateral veins

Upon reaching the level of the kidneys, the post-caval drainage
line becomes greatly expanded to form a large, irregularly quad-
rilateral channel situated in front of the aorta. This large chan-
nel establishes the renal anastomosis and, from its position,
may be termed the inter-renal segment of the post-cava. It
presents two cephalic and two caudal angles. From its right
cephalic angle the post-caval drainage line is continued toward
the heart by means of the right sub-cardinal vein. Its left
cephalic angle receives the left suprarenal vein. The renal vein

212 FREDERICK TILNEY

enters the inter-renal segment immediately below the cephalic
angle (fig. 13, 57).

Two vessels of equal size enter the inter-renal segment, one
at either caudal angle. They are practically parallel to each
other, are placed ventro-lateral to the aorta and separated from
each other by a mean distance of 60u. Caudally they arise
from the junction of the internal and external iliac veins. The
proximity of these two parallel channels to the median line,
their position ventral to the aorta and mesial to the ureters
makes it probable that they are derivatives of the ventro-mesial
element of the peri-mesonephroic plexus and hence should be
considered the cardinal collateral channels (fig. 18, 27).

Immediately above the entrance of each cardinal collateral
vessel, a vein from the mesonephros empties into the inter-renal
segment. This mesonephroic vein subsequently becomes the sex
vein and will hereafter be referred to by that term (fig. 13, 40).

Caudad of its sub-cardinal portion, the axial drainage line is
therefore made up as follows, beginning from the confluence of
the internal and external iliac veins.

1. The two cardinal collateral veins, constituting the so-called
paired portion of the post-cava (fig. 13, 2).

2. The inter-renal segment, situated directly ventral to the
aorta and between the kidneys. It receives the cardinal col-
lateral and sex veins at its caudal angles while the renal veins
empty into it by its cephalic angles. At its right cephalic angle
it passes over into the sub-cardinal portion of the post-cava
(e159):

The relative dimensions of the two portions change consider-
ably in the further development of the post-cava; the following
tabulation gives the length of the two segments in the older
embryos as compared with the adult conditions.

20-MM. EMBRYO 23-MM. EMBRYO) ADULT
mm. | mm. | mm.
Length of inter-renal segment............ 0.71 0.75 65
Length of paired portion (cardinal col- |
Ls tera) heroes sce: ree ole ooh sn iene eto 0.14 1.25 49
Distance between renal and sex veins...... 6.26 | 0.30 | 54

VEINS AND LYMPHATICS IN TRAGULUS 213

From these figures it appears that the ratio of the inter-renal
segment to the paired portion of the cava (cardinal collaterals)
is 1 to 1.6 in the 20 mm. embryo as against 1 to 0.75 in the adult
animal. In other words, if the increment of growth has remained
constant in the inter-renal segment, it has been reduced one
half in the cardinal collateral vessels while passing from the 20
mm. stage to adult. These facts make it clear that longitudinal
expansion of the inter-renal segment plays an important role in
determining the position and relations of the post-renal segment
of the cava. This assumption is further borne out by the fact
that the distance between the sex veins and the renal veins pro-
gressively increases in passing from the embryonic stages to the
adult.

The pre- and post-cardinal veins

The pre-cardinal veins may still be identified almost in their
entirety even in these late stages. In the greater part of the
neck they constitute a pair of channels, the internal jugular
veins. High up in their cephalic portion, they receive a short
trunk which has resulted from the confluence of the temporo-
facial and internal maxillary veins. This trunk represents the
external jugular vein. As the two internal jugular veins approach
the thorax they become confluent and form a single large vessel.
It is into this confluent element that the sub-clavian, cephalic,
internal mammary and vertebral veins enter. On passing into
the superior mediastinum the single vessel again becomes a double
channel. The vessel on the right is much the larger. This por-
tion of the jugular system represents the pre-cava which, after
becoming much reduced in size, enters the heart. The left chan-
nel proceeds further caudad and finally joins the left post-cardi-
nal vein to form the duct of Cuvier.

The left post-cardinal vein participates in the formation of
the left azygos vein. During this process that redundant car-
dinal channel, earlier observed in the perimesonephroic plexus
and referred to as the intermediate element, seems to take a
prominent part. In the cephalic portion of the azygos major
vessel, the original post-cardinal channel determines the ultimate

214 FREDERICK TILNEY

vein line, but in those regions corresponding to the perimeso-
nephroic plexus of the younger stages, the intermediate element
is obviously the selected channel. The relations of this element
to the aorta in the 6 mm. specimen indicate the future line of
the azygos vessel. In the 13 mm. embryo the intermediate ele-
ment has become a more prominent channel than the post-cardi-
nal. <A right azygos ‘vein, corresponding in position to the left
vessel, is formed in like manner, except that it establishes its
ultimate drainage connections by a series of cross anastomoses
with the left vein. These anastomoses, at first, are diffuse and,
in the 20 mm. embryo, appear as irregular channels between
the two azygos lines. The most cephalic channel is well defined
and larger than the rest. In the 23 mm. embryo the formation
of distinet cross vessels has been carried much further so that
these inter-azygos connecting lines are now arranged in seg-
mental series, are twelve in number, and occur at the junction
of each dorsal segmental vein with the azygos vessel of its respec-
tive side (fig. 10, 62). Another set of anastomosing vessels serves
to connect the two azygos channels with the sub-cardinal and
cardinal collateral lines. These anastomotic vessels are placed
along the sides of the aorta and thus correspond in their relations
to the mesial branches of the perimesenephroic plexus observed
in the younger stages. Two sets of these vessels may be dis-
tinguished, a cephalic anastomosis which traverses the anlage of
the supra-renal body and undergoes a gradual reduction in pass-
ing from the 13 mm. to the 23 mm. stage. This plexus forms
a connection between the azygos vein and the corresponding
side of the post-cava in its inter-renal segment. The second
plexus is much more extensive. It establishes a communication
between the azygos veins and cardinal collateral portions of
the post-cava (fig. 14, 61). The significance of this connection
in its bearing upon the possibilities of post-caval formation will
subsequently be discussed. That it gradually diminishes in sig-
nificance as growth proceeds is witnessed by the fact of its actual
reduction in passing from the 13 mm. to the 23 mm. embryo.
The post-cardinal veins thus concern themselves with the azygos
system, allowing the selection of the caval drainage line to fall

VEINS AND LYMPHATICS IN TRAGULUS 215

upon the other channels. The venous plexus between the pre-
aortic and post-aortic veins becomes so extensive as to completely
invest the aorta except for a small interval immediately in front,
on either side of the median line. As it approaches the levels
marking the derivation of the hypogastric arteries the plexus
becomes more voluminous and finally communicates with the
iliac trunks.

From the conditions of these embryonic stages it is obvious
that the venous return from the internal and external iliac veins
depends upon a large peri-aortic plexus rather than upon discrete
venous channels. The increasing prominence of certain axial
channels in this plexus clearly indicates the process by which
the plexus itself is to be replaced by veins in the line of the lon-
gitudinal growth of the body. The course of these axial channels
is parallel to the long body-axis; two are placed ventro-lateral
to the aorta, the cardinal collateral veins; two are situated dorso-
lateral to the aorta, the post-cardinals. At the indeterminate
stages represented by the 20 mm. and the 23 mm. embryos the
internal and external iliac veins may select one of several possi-
bilities for the continuance of their drainage lines toward the
heart. They may choose either or both of the post-cardinal
veins to the exclusion of the cardinal collaterals, so that the
venous blood would reach the inter-renal segment of the cava
by the plexiform channels of the inter-renal post-cardinal plexus.
They may select both cardinal collaterals to the exclusion of the
post-cardinals and so establish a communication with the inter-
renal segment of the cava. The adult conditions, showing that
the post-renal segment of the cava is pre-aortic in position,
clearly demonstrate that the selection has fallen upon the cardi-
nal collateral veins and that the post-cardinals play no part in
the formation of this portion of the cava.

These possibilities of selection which must be reckoned with
in discussing the post-renal cava in Tragulus emphasize again
the statement of Schulte (’09) that
homonymous venous channels are not necessarily morphological equiva-

lents but are rather homodynomous, agreeing in function because they
drain similar areas:—and it thus appears that the anatomicalname of

216 FREDERICK TILNEY

veins designate not morphological but physiological units. The term
post cava only indicates a hydro-dynamic line. The variously named
cardinals are merely dilated portions of the reticulum along the major
hydro-dynamic lines, which, responding to the large volume of blood
they transmit, dominate the picture.

Although the facts cited above account for the acquisition of
a pre-aortic post-renal segment of the cava in the adult Tragulus,
they do not furnish a complete explanation of the process by
which the ultimate relations of this vessel are attained. In the
20 mm. and 23 mm. embryos the post-renal segment of the cava
presents two portions, each of which is pre-aortic in position,
namely, the paired portion and the unpaired portion. The un-
paired portion in these stages, constitutes about one-third of
the entire post-renal segment of the cava (fig. 18, A and B): In
the adult, while the paired and unpaired elements still enter into
the formation of the post-renal segment, their proportions have
greatly changed. The unpaired portion instead of being one-
third as long as the paired portion is now six times longer. In
other words five-sixths of the post-renal segment of the cava is
represented in the adult by a single unpaired pre-aortic channel,
the remaining one-sixth being represented by the paired portion
(fig. 1). This marked change in proportion may be due to one
of three possibilities; 1, the fusion of the two cardinal collateral
veins across the median line in the cephalic two-thirds of their
course; 2, a caudal migration of the angle of confluence of the
cardinal collaterals; and 3, the longitudinal expansion of the
inter-renal segment alone. The position and relations of the
sex veins considerably lessen the difficulties in deciding which
of these processes is the active one. It has already been shown,
in the older embryos, that the renal veins mark the cephalic
limits of the inter-renal segment while its caudal limits are in-
dicated by the sex veins. The portion of the cava between these
limits, therefore, must be considered the inter-renal segment in
all stages. Upon this basis, the change in proportion of the
two elements of the post-renal cava, observed in passing from
the embryonic to the adult stages, may be explained by the longi-
tudinal expansion of the inter-renal segment. ‘The ultimate wide

VEINS AND LYMPHATICS IN TRAGULUS PA Bo

separation between the renal and sex veins seems to indicate
that this was the process by which the post-renal cava has been
changed from an embryonic channel in greater part paired to
an adult vessel in greater part unpaired. The probability of
this explanation is further sustained by the fact that the incre-
ment of growth from the early stages has favored the inter-renal
segment (see page 213). The relations between the length of
the post-renal segment of the aorta and that of the inter-renal
segment of the cava also show changes which are significant in
this connection. The measurements of the aorta were taken
from the point of derivation of the right renal vein to the iliac
bifurcation; those of the inter-renal segment from the entrance
of the right renal vein to the point of confluence of the two cardi-
nal collateral veins.

In the 20 mm. embryo the length of the inter-renal segment
of the cava was 0.33 that of the post-renal segment of the aorta.
In the 23 mm. embryo this value has increased to 0.40 and in
the adult to 0.87. Thus there has been a relative increase in
the rate of growth in the inter-renal segment of the cava as com-
pared with the post-renal segment of the aorta.

The change in the relations of the sex veins and arteries is
further evidence of this relative increase in the inter-renal seg-
ment of the cava. Both of the older embryos show the sex veins
and arteries in close relation to each other (fig. 13, A and B).
The veins enter the inter-renal segment practically in common
with the entrance of the cardinal collateral veins. The sex arter-
ies arise separately from the aorta at a level only slightly caudal
to that of the veins. The adult specimen in the Columbia col-
lection shows the sex arteries arising from a short common trunk
given off from the aorta at a point 1.5 cm. above the iliac bifur-
cation, while the sex veins enter the common iliac veins (paired
portion of the post-cava) 4 mm. below the point of entrance of
these latter channels into the unpaired portion of the post-cava
(fig. 1). Thus the sex arteries, which in the embryo arise from
the aorta caudad of the sex veins, in the adult arise from a level
distinctly cephalad of these veins. This marked change in rela-
tions appears to have its explanation in the relatively more rapid

218 FREDERICK TILNEY

longitudinal growth of the inter-renal segment. of the cava as
compared with the post-renal segment of the aorta. The shift-
ing of the sex veins from their more primitive point of inoscula-
tion may be due to a caudal migration of the angle of confluence
of these vessels with the post-renal segment of the cava or it
may be the result of certain mechanical changes due to the caudal
migration and descent of the testis.

The intra-hepatic portion of the right omphalo-mesenteric vein
has gained ascendency over all the venous spaces of the liver
and appears as a definite channel situated in the right dorso-
mesial angle of that organ. It constitutes the hepatic portion
of the post-cava. At the cephalic pole of the liver the vessel
is large, receiving, in this region, the two major revehent trunks
which drain the hepatic sinusoids. Immediately below the inos-
culation of these revehent vessels the cava diminishes in size,
taking up as it proceeds caudad, several lesser, hepatic revehent
tributaries. When the caudal pole of the liver is reached the
vessel swings slightly mesad and dorsad, to pass over into the
sub-cardinal portion of the cava. The mesenteric portion of the
omphalomesenteric vessel has now become the portal vein and
drains into one of the largest advehent branches of the um-
bilical channels. The post-caval drainage line thus utilizes the
right intra-hepatic portion of the omphalomesenteric vein in pass-
ing through the liver, and the right sub-cardinal vein as far
caudad as the inter-renal segment.

The umbilical drainage system presents itself as the typical
single channel of foetal life. It makes its way through the um-
bilical fissure of the liver and then enters that organ. In the
liver it breaks up into the rich plexus of the umbilical portal
system. ‘The ductus venosus is given off from one of the main
stems of this plexus and passes obliquely upward to enter the
post-cava in common with the confluence of the major hepatic
revehent veins. These observations apply equally to the 20 mm.
and 23 mm. embryos.

VEINS AND LYMPHATICS IN TRAGULUS 219
SUMMARY

The development of the axial lymphatics in Tragulus presents
the following characteristics:

1. Two distinct anlagen, one for the lymph sac and the other
for the systemic lymphatics.

a. The lymph sac is derived from the venous system; passes
from the primary venous into the veno-lymphatic stage, at the
end of which period it loses all connection with the veins and
so enters upon its pre-lymphatic stage. Ultimately it establishes
a secondary connection with the venous system and so becomes
definitely lymphatic.

b. The axial systemic lymphatics develop in three distinct por-
tions, namely the azygos, pre-azygos and post-azygos segments.
The exact histogenesis of these segments could not be determined,
but they convey the impression of plexiform channels arising
independently in the mesenchyme and then rapidly expanding
into discrete axial vessels. No connection between these seg-
ments and the veins was observed at any point. By confluence
the segments became integrated to form the axial systemic lym-
phatics.

2. The final union of the two distinct anlagen determines the
completed axial lymphatic line. In this manner the lymph sac
becomes intermediary in establishing a communication between
the venous and peripheral lymphatic system, a mode of organi-
zation which resembles that of the cat and so, no doubt, the
general ground plan in mammals.

Advancing from the stage of symmetrical channels, the venous
development is much concerned with modifications in the peri-
mesonephroic plexus. This vascular network, bearing intimate
relation to the mesonephros, is subsequently converted into such
distinct channels as the sub-cardinal and cardinal collateral veins
which have adapted themselves to the general line of axial growth.

These facts seem to sustain the proposition that all definite
venous channels have their inception in a plexus and emerge
from this plexus as definite veins, under the influence of certain
hydro-dynamic factors, which are in the interest of most efficient

220 FREDERICK TILNEY

venous return. Lest it be argued that the perimesonephroic
plexus from which the above named vessels arise is a special
case in Tragulus, it may be stated that this plexus has already
been described by Brown (11) in cat embryos, and has been
observed by the writer in several sauropsid forms (chick and
scleroporus embryos). The organization of the post-renal seg-
ment of the cava also appears to be an instance of the emergence
from a plexus of selected axial channels; in this case the channels
happen to be the cardinal collaterals. The plexus itself surrounds
the aorta; the cardinal collateral vessels mark its ventral limits;
the dorsal limits are formed by the post cardinals. A pre-aortic
post-renal segment of the cava is established in part by the
selection of the cardinal collateral veins as axial channels and
in part by a marked longitudinal expansion of the inter-renal
segment. These conditions observed in Tragulus definitely ally
its post-cava with the marsupial type. The similarity thus
established between the venous organization of this aberrant
ungulate and that of the marsupials has a clear phylogenetic
significance. Itdoes not, however, shed much light on the more
fundamental problems involved in the development of the post-
cava. In fact, it merely serves to open the question as to what
hydro-dynamic and other mechanical factors must control the
selection of the ultimate venous drainage channels in the axial
line of the body.

In conelusion, the writer desires to acknowledge his indebted-
ness and express his appreciation to Professor Huntington for
his direction and assistance in preparing this paper.

VEINS AND LYMPHATICS IN TRAGULUS 221

LITERATURE CITED

Brepparp, F. E. 1907 A preliminary note upon some characteristics of the
venous system of Tragulus meminna and allied genera. Anat. Rec..,
no. 5 of Amer. Jour. Anat., vol. 7, no. 1, p. 111.

Brown ALFRED JEROME 1911 A note on postcardinal omphalomesenteric com-
munications In the adult mammal. Anat. Rec., vol. 4, no. 12, p. 425.

Huntineton, Gro. §. 1910 The genetic principles of the development of the
systemic lymphetics vessels in the mammalian embryo. Anat. Rec.,
vol. 4, no. 11, p. 399.

HuntinetTon, Geo. §., and McCuvurz, C. F. W. 1910 The anatomy and devel-
opment of the jugular lymph sac in the domestic cat (Felis domestica).
Amer. Jour. Anat., vol. 10, no. 2, p. 177.

McCuurg, C. F. W. 1903 A contribution to the anatomy and development of
the venous system of Didelphis marsupialis. Part 1. Amer. Jour.Anat.,
ViOle 2, NOstos eomle

1906 The post-cava of an adult Indian Chevrotain (Tragulus meminna,
Erxleben). Anat. Anzeig., Bd. 29, No. 13.

ScHuLTE, HERMANN von W. 1907 The range of variations in Monotremes and
Australian Marsupials. Amer. Jour. Anat., vol. 6, no.3, Proceedings,
Ds a0.

ScHULTE, HmRMANN von W., and TitNry, FrepERIcK. 1909 A note on the
organization of the venous return with especial reference to the iliac
veins. Anat. Rec., vol. 3, no. 11, p. 555.

PLATE 1

EXPLANATION OF FIGURE

1 Adult Tragulus meminna from a dissection in the study collection of the
Department of Anatomy, Columbia University, showing the pre-aortic position
of the post-cava. 24, aorta; 34, ureter; 36, internal iliac vein; 37, external iliac
vein; 38, common iliac vein (paired portion of post-cava); 39, unpaired portion
of post-cava; 40, sex vein; 41, spermatic artery; 42, caudal vein; 43, common iliac
artery; 44, external iliac artery; 45, internal iliac artery; 57, renal vein; 48, renal
artery; 61, caudal artery.

222

VEINS AND LYMPHATICS IN TRAGULUS
FREDERICK TILNEY

PLATE 1

223
THE AMERICAN JOURNAL OF ANATOMY, VOL. 13

3, NO. 2

PLATE 2

EXPLANATION OF FIGURE
2 From a reconstruction of a 5 mm. Tragulus embryo. Collection No.
204. X 100. Showing the symmetrical arrangement of the axial venous channels
and the plexuses in connection with them. 1, aortic arches; 2, pre-cardinal vein;
3, dorsal pre-cardinal tributaries; 4, dorsal aorta; 5, post-cardinal vein; 6, um-
bilical vein; 7, umbilico-post-cardinal plexus; 8, perimesonephroic plexus, 9, om-
phalomesenteric vein; 10, duct of Cuvier.

224

PLATE

VEINS AND LYMPHATICS IN TRAGULUS

FREDERICK TILNEY

PLATE 3

EXPLANATION OF FIGURES

3 Reconstruction of a portion of the perimesonephroic plexus in a 5 mm.
Tragulus embryo. Collection No. 204. X 200. 45, post-cardinal vein; 24, aorta;
46, mesial portion of perimesonephroic plexus; 47, ventral portion of perimeso-
nephroic plexus.

4 Cross section showing the relations of the perimesonephroic plexus
ina5mm. Tragulusembryo. Collection No. 204. * 200. 4, post-cardinal vein;
8, perimesonephroic plexus; 13, Wolffian duct; 46, mesial portion of the peri-
mesonephroic plexus; 48, mesonephros.

PLATE 3

AND LYMPHATICS IN TRAGULUS

VEIN

FREDERICK TILNEY

PLATE 4

EXPLANATION OF FIGURE

5 Reconstruction of a 6 mm. Tragulus embryo. Collection No. 205.
< 100. Showing the arrangement of the axial venous channels. 1, aortic arches;
2, pre-cardinal vein, showing its horizontal and vertical limbs; 3, dorsal pre-car-
dinal tributaries; 5, post-cardinal vein ;6, umbilical vein; 7, umbilical-post-cardinal
plexus; 9, omphalomesenteric vein; 10, duet of Cuvier; 11, sub-hepatic sinus;
12, sub-cardinal vein; 13, Wolffian duct; 14, Cloaca; 24, aorta.

228

VEINS AND LYMPHATICS IN TRAGULUS PLATE 4
FREDERICK TILNEY

PLATE 5

EXPLANATION OF FIGURE

6 Reconstruction of a 6 mm. Tragulus embryo. Collection No. 205. > 100.
Showing the region of the sub-hepatic sinus in ventral view. 4, post-cardinal
vein; 6, umbilical vein; 7, umbilico-post-cardinal plexus; 9, omphalo-mesenteric
vein; 11, sub-hepatic sinus; 12, sub-cardinal vein; 13, Wolffian duct; 14, Sinus
venosus; 16, intestine; 24, aorta.

230

VEINS AND LYMPHATICS IN TRAGULUS PLATE 5
FREDERICK TILNEY

231

PLATE 6

EXPLANATION OF FIGURE

7 Reconstruction of a 20mm. Tragulusembryo. Collection No. 202. Xx 100.
Showing the axial veins and lymphatics. 6, umbilical vein; 17, pre-cava; 18,
hepatic portion of post-cava; 19, sub-hepatie portion of post-cava; 20, post-
cardino-cardinal collateral anastomosis; 2/, cardinal collateral vein; 22, confluence
of iliae veins; 23, arch of aorta; 24, aorta; 28, jugula lymph sac; 26, dorso-lateral
process of the lymph sac; 26a, dorsal descending process of the lymph sac (tho-
racic duct approach) ; 26b, ventral descending process of the lymph sac (broncho-
mediastinal approach); 28, azygos segment of thoracic duct; 29, post-azygos
segment of thoracic duct; 30, cephalic division of pre-azygos segment of thoracic
duct; 31, caudal division of pre-azygos segment of thoracic duct; 32, sympathetic
nerve; 33, vagus nerve; 34, ureter; 35, jugular vein; 51, azygos vein; 62, dorsal
segmental vein; 68, pancreas.

VEINS AND LYMPHATICS IN TRAGULUS PLATE 6
FREDERICK TILNEY

PLATE 7

EXPLANATION OF FIGURE

8 Reconstruction showing the junction of the azygos and pre-azygos segments
of the thoracic duct in a 20mm. Tragulusembryo. Collection No. 202. X 100.
24, aorta; 28, azygos segment of the thoracic duct; 31, caudal division of the pre-
azygos segment of the thoracic duct; 50, junction of the azygos and pre-azygos
segments of the thoracic duct.

VEINS AND LYMPHATICS IN TRAGULUS PLATE 7
FREDERICK TILNEY

PLATE 8

EXPLANATION OF FIGURE

9 Cross section through the azygos segment of the thoracic duct showing
its resemblance to the reptilian type of thoracic duct. 20mm. Tragulus embryo.
Collection No. 202. X 100. 24, aorta; 28, azygos segment of thoracic duct com-
pletely surrounding aorta: 51, azygos vein.

VEINS AND LYMPHATICS IN TRAGULU
FREDERICK TILN

PLATE 8

PLATE 9

EXPLANATION OF FIGURE

10 Reconstruction showing the axial veins and lymphatics in a 23 mm.
Tragulus embryo. Collection No. 228. X 100. 2, pre-cardinal vein; 23, arch
of aorta; 25, jugular lymph sac; 26, dorso-lateral process of the lymph sac; 28,
azygos segment of thoracic duct; 29, post-azygos segment of thoracic duct;
$1, caudal division of the pre-azygos segment of the thoracic duct; 32, sympa-
thetic nerve; 33, vagus nerve; 35, jugular vein; 51, azygos vein; 62, dorsal seg-
mental vein; 65, sub-hepatic portion of post-cava.

VEINS AND LYMPHATICS IN TRAGULUS PLATE 9

FREDERICK TILNEY

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 2

PLATE 10

EXVLANATION OF FIGURE

11 A, B, and C. Three serial sections showing the manner in which the left
jugulo-sub-clavian tap of the lymph sac is made in a 23 mm. Tragulus embryo.
Collection No. 228. » 100. 26a, dorsal descending process of the lymph sac
(thoracic duct approach) ; 27, ventral descending process of the lymph sac (broncho-
mediastina approach); 52, jugulo-sub-clavian approach; 53, confluence of the
jugular veins; 54, thymus; 54, thyro-cervical artery; 56, sub-clavian vein; 63,
ventral prolongation of the jugulo-sub-clavian approach. The relations of the
structure marked 52 in the three figures indicate the manner in which the second-
ary connection between the lymph sac and the venous system is made.

240

PLATE 10

ULUS

I
x

VEINS AND LYMPHATICS IN TRAG

FREDERICK TILNEY

241

PHATE 11

EXPLANATION OF FIGURE

12 Reconstruction showing the junction of the azygos and pre-azygos
segments of the thoracic duct in a 23 mm. Tragulus embryo. Collection No.
228. 100. 238, arch of aorta; 24, aorta; 28, azygos segment of the thoracic
duct; 30, cephalic division of the pre-azygos segment of the thoracic duct; 31,
caudal division of the pre-azygos segment of the thoracic duct; 30-31, confluence
of the two divisions of the pre-azygos segment of the thoracic duct; 50, junction
of the azygos and pre-azygos segments of the thoracic duct.

242

VEINS AND LYMPHATICS IN TRAGULUS
FREDERICK TILNEY

PLATE 11

PLATE 12

EXPLANATION OF FIGURES

13a and 13b Schemata giving a ventral view of the relations of the post-
cava to the aorta in a 20 mm. and a 23 mm. Tragulus embryo respectively, as
shown by reconstructions of these stages. 21, cardinal collateral vein; 24, aorta;
40, sex vein; 41, spermatic artery; 57, renal vein; 59, inter-renal segment of the
cava; 60, iliac bifurcation of the aorta.

VEINS AND LYMPHATICS IN TRAGULUS PLATE 12
FREDERICK TILNEY

13A 1B

PLATE 13
EXPLANATION OF FIGURE

14 Cross section showing the anastomosis between the cardinal collateral
and azygos veins in a23 mm. Tragulusembryo. Collection No. 228. 21, cardinal
collateral vein; 24, aorta; 28, azygos segment of the thoracic duct; 51, azygos
vein; 61, anastonosis between cardinal collateral and azygos veins.

VEINS AND LYMPHATICS IN TRAGULUS
FREDERICK TILNEY

PLATE 13

247

~ :
e

. Bd . Lae ae
| | Ea 7,

»
q

ON THE DEVELOPMENT OF THE HUMAN HEART

FRANKLIN P. MALL
From the Anatomical Laboratory of the Johns Hopkins University

THIRTY-SEVEN FIGURES

In my recent study! on the musculature of the adult human
heart it was necessary to refer constantly to the development of
this organ, and in general my description of the course of the mus-
cle bundles was also based upon their development. This made
it necessary to study numerous serial sections of embryo hearts,
as well as whole hearts which had been removed from the embryo
and dissected under the binocular microscope.

It is my purpose now to give as accurate a description as possible
of several points which were obscure to me at the beginning of
my study, so this report is to be viewed as supplementary to the
excellent study by His as well as the recent one by Tandler.
First of all an attempt was made to study the course of the muscle
bundles and the formation of the vortex in the smallest hearts by
means of direct observation upon whole hearts, stained and
unstained, under the binocular microscope. This study was
controlled by that of serial sections of other hearts, an abundance
of material of both kinds being available. It soon became
apparent that the muscle wall of the entire heart had to be
included in this study, which soon showed that the critical point
lay in the wall of the atrial canal, that is the common canal be-
tween the atria and ventricles. The study led back to the study
of the valves at this point, an understanding of which is really
the key to the whole situation. This resulted in locating definitely
the atrio-ventricular muscle bundle (bundle of His) in all stages
of development.

1 Mall, F. P., Onthe muscular architecture of the ventricles of the human heart.
Amer. Jour. Anat., vol. 11, 1911.

249

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 3
JuLy, 1912

250 FRANKLIN P. MALL

The points to be discussed will be considered in the following
order: A, Subdivisions of the early heart; B, Formation of the
septum and atrio-ventricular valves; C, The atrio-ventricular
bundle; D, Musculature of the left ventricle.

A. SUBDIVISIONS OF THE PRIMARY HEART TUBE

In an embryo about 2 mm. long (No. 391), which was modeled
in wax by Dr. Dandy, the heart is shown as a relatively straight
tube with its arterial end directed towards the head (fig. 1).
Its muscle wall is of even thickness and communicates throughout
its whole length along the dorsal midline with the rest of the meso-
derm, that is it has a complete dorsal mesentery.? At its anterior
end the heart tube shows a slight dilatation just before the arte-
ries arise from it. Midway between the two ends of the muscle
tube there is an indentation on the left side which marks the begin-
ning of the bulbo-ventricular groove, that is, it separates the
atria and left ventricle on the one hand from the right ventricle
and the bulb on the other. Within the muscle tube there is
suspended by means of numerous fine fibrils the collapsed endo-
cardial lining. These fibrils will be considered later when the
development of the valves is discussed.

The heart now separates rapidly from the rest of the mesoderm
in subsequent stages and is soon suspended in the pericardial
coelom, remaining attached to the body only at its venous and
arterial ends. The indentation on the left side of the heart,
mentioned above, becomes more pronounced, the heart rolls
upon itself quite rapidly as succeeding stages of development
show. In an embryo 2.5 mm. long (No. 3) the heart is separated
from the dorsal midline in its middle, while in another of the same
length (318) the separation is more pronounced. In one 3.5 mm.
long (164) the separation is complete and the various subdivisions
of the final heart tube can be outlined with precision (fig. 2).

2 A complete description of this embryo isgivenby Dandy. Amer. Jour. Anat.,
vol. 10,1910. Sections through the heart are shown in his figs. 1 and 2. Evans
(Keibel-Mall, Manual of human embryology, vol. 2, fig. 409).also pictures sections
through the heart of thisembryo. Also by Mall, Ibid., vol. 1, figs. 382-386.

LEGENDS FOR ALL OF THE FIGURES

A, atrium

A.C., atrial canal

A.Cu., anterior endocardial cushion
A.F., annulus fibrosus

Ao., aorta

A.P., anterior papillary muscle
B.S., bulbo-spiral band

A.V.B., atrio-ventricular bundle
B., bulb

C.S., coronary sinus

T.V.C., interventricular canal
F.O. foramen ovale I

F.O.2 foramen ovale II

L.A., left atrium

L.Cu., left endocardial cushion
L.O., left venous ostium

L.P., large papillary muscle of the
right ventricle

L.R.V., longitudinal bundle
right ventricle

L.V., left ventricle

M.p.m., medial papillary muscle

P., pulmonary artery

P.Cu., posterior endocardial cushion

P.}., posterior papillary muscle

R.A., right atrium

R.Cu., right endocardial cushion

R.O., right venous ostium

R.V., right ventricle

of the

S.A., septum of the atria
S.S., sino-spiral band

S.V., septum of the ventricle
V., ventricle

Fig. 1

Ventral view of the model of an embryo 2 mm. long (No.391).

x 100.

The pericardium has been removed from in front of the heart.

251

252 FRANKLIN P. MALL

I mentioned above in describing the heart of the embryo 2 mm.
long that its endothelial lining is collapsed and suspended by a mass
of fine fibrils within the muscular tube. In slightly older stages
the arrangement of the endothelial lining is changing, becoming
dilated on the venous side of the heart. This change is beginning
in No. 3, is more advanced in No. 318 and is complete in No. 164
(fig. 2). His? has pointed out that the endothelial lining hugs
the muscle wall closely in the embryonic atrium, while it remains
suspended for a time in the rest of the heart. This arrangement
is so pronounced in the early heart that it affords a way by which

Fig. 2. From the reconstruction of a heart of an embryo 3.5 mm. long (No. 164).
< 66. View from the left side.

we may determine with precision the exact portion of the heart
tube from which the atrium arises. My specimens show con-
clusively that the atrium arises exclusively from the free heart
tube and that the sinus venosus does not contribute to its forma-
tion. This being established it follows that as soon as the heart
tube is fully separated from the body walls that the anlage of the
entire adult heart is to be found between its arterial and venous
attachments.

In the embryo 3.5 mm. long (No. 164) the completed heart
tube is seen, which is S-shaped and twisted upon itself so that the

3 His, W., Anatomie mensch. Embryonen. Theil 38, Leipzig, 1885.

DEVELOPMENT OF THE HUMAN HEART 253

arterial and venousendsare brought close together. At the venous
end the muscle wall is shghtly dilated which marks the atrium;
this is lined closely with endothelium which encircles the cavity
within. No delicate fibrils are here seen between the muscle wall
and its endothelial lining. Then foilows an upper bend to the
heart after which there is a dilatation projecting towards the left
side, the former marking the atrial canal and the latter the left
ventricle. The lower connecting piece unites the left ventricle
with the bulb which Jater on gives rise to the right ventricle. In
the atrial canal (Haller’s auricular canal) the endothelial tube is
seen as a solid strand of cells suspended freely in the muscle wall

Fig. 3 Section of the heart of the embryo 3.5mm. long. X 66.

by the delicate fibrils already mentioned. In the left ventricle
the tube shows a distinct cavity, while throughout the rest of the
heart tube the cavity is irregular but not pronounced. The form
of the endothelial tube is shown in fig. 2 and again in a semidia-
erammatic figure of a transverse section through the atrium, ven-
tricle and bulb in fig. 8. The delicate fibrils, which no doubt
belong to the endothelial cells are present in large number through-
out the whole heart tube, excepting in that which forms the atrium.
In another embryo (No. 384, 2 mm. long), considerably smaller
than the one Just described and probably pathological, the degree
of development of the heart is practically identical with the one

Q1

35 mm. long (No. 164).

254 FRANKLIN P. MALL

In my collection there are two other embryos slightly more
advanced in development than the one just described which bear
upon the exact origin of the atria from the heart tube. They are
Nos. 486 (4 mm. long) and 470 (4 mm. long). Neither of these
specimens has been studied carefully as a whole, so the number of
myotomes in each can not be given. Nor have the measurements
been corrected by the drawings and the sections.

In No. 486 the single atrium as described above (No. 164) is more
pronounced, is dilated and filled with blood,,while the form of the
endothelial tube is much the same. However, the atrium is
sharply separated from the sinus venosus and there are a few
fibrillae between the endothelial and muscular walls. In the left
ventricle the endothelial and the muscular layers are just begin-
ning to interlock to form the first trabeculae.

Embryo No. 470 shows the heart more advanced than in No.
486. Theatrium has become double, that is there are two atria.
The endothelial tube in it is distended and its separation from the
sinus venosus is still pronounced. The trabecular formation in
the left ventricle is somewhat more pronounced than before.

From now on the changes in the heart take place very rapidly,
as the general form of the embryo also changesrapidly. Thehead
is bent upon the body which is well curved upon itself with pro-
nounced limb buds. In the next embryo, No. 239 (4 mm. long),
the subdivisions of the heart are sharply defined and in the follow-
ing stage, No. 463 (3.9 mm.), the preliminary subdivisions are
complete (fig. 4). During this time the embryo curls upon itself
and the limb buds are formed. In embryo 239 the two atria are
very pronounced, the right communicating with the sinus venosus.
The atrial canal is sharply defined, first as a constriction and
secondly by a great increase of the fibrillar mass between the
endothelial and muscular walls. The trabecular system is well
formed in the left ventricle and has extended into the bulbus,
that is into the right ventricle. In this specimen it is clear that the
course of the circulation is from the right to the left atrium, then
first to the left ventricle after which it enters the right. The
right atrium still lies in the notch between the left ventricle and
the bulb on the posterior side of the heart. Only a little later

DEVELOPMENT OF THE HUMAN HEART PASS)

does it project to the right side of the bulb which becomes its per-
manent position. This is seen in No. 463 (fig. 4).

Although the two embryos just mentioned are practically of
the same stage of development, the difference of the degree of
development of their hearts is most pronounced. No. 463, which
is perfectly preserved, has a heart with larger atria, a more con-
stricted atrial canal, a large left ventricle and a pronounced but
contracted bulb. The muscular wall of the whole heart is con-
tinuous without a single break init. That surrounding the atrial
canal is sharply defined forming a continuous ring connecting at

Fig. 4 Section of the heart of an embryo 3.9 mm. long (No. 463). X 66.

all points with the atria above and the left ventricle below. This
is mentioned especially because a share of this connecting ring
disappears while the remaining portion becomes the atrio-ventric-
ular bundle.‘

B. FORMATION OF THE ATRIO-VENTRICULAR VALVES

In the earliest stages described (No. 391), while the muscle
wall of the heart is still in the form of a straight tube and is con-
nected throughout its length with the body wall, the endothelial
tube is separated from the muscular tube by a marked layer of
delicate fibrils. In their papers upon the heart both His and

4 A few data regarding all of the embryos described in this paper are given in
my catalogue of 500 specimens. Anat. Rec.. vol. 5, 1911.

256 FRANKLIN P. MALL

Tandler speak of these fibrils but they give no very definite infor-
mation regarding their nature. In his study of the chick His®
describes the space between the endothelial tube and muscular
wall of the heart, which later in development fills with connective
tissue arising from the inner tube. In the atria, where this space
is never pronounced, the secondary thickening is also insignifi-
eant. Later His® also observes this space in young human em-
bryos. Many fibrils extend from the endothelial tube, which
when they are pronounced, draw out the side of the tube in a char-
acteristic way. This he pictures. He is uncertain whether these
fibrils are natural or produced by the hardening reagents used.
In the atrium the inner tube hugs the muscle wall closely. In the
atrial canal the space is filled with two pronounced cushions of
connective tissue, while in the ventricle the muscle forms trabec-
ulae which are soon covered with endothelium. In the bulb
this space is very marked and filled with a delicate connective
tissue framework. He does not show conclusively the meaning
of this tissue.

Tandler? describes and figures this substance well in a human
embryo with fifteen somites. Although his figure shows beauti-
fully the inwandering of nuclei from the endothelium, and al-
though he speaks of a reduction and enlargement of these filbrils,
he is unwilling to decide whether or not they are due to the method
of preservation and of staining of the sections. He states ex-
pressly that the tissue resembles very much Wharton’s jelly. It
seems to me, however, that the evidence of His and Tandler is
sufficient to show that this tissue is not due to coagulation but
a constant normal constituent of the developing heart. That it
is distributed in a definite way in different portions of the heart
and in different stages of development speaks almost conclusively
for this opinion. Its origin and meaning is however a different
question.

> His, Untersuch ueber die erste Anlage des Wirbelthierleibes. Leipzig, 1868,
S. 141.

6 His, Anat. mensch. Embryonen. Th. 5, 1885, 8. 141.

7 Tandler, Entwicklungsgeschichte des Herzens. Keibel-Mall Handbuch d.
Entwicklg. d. Menschen, Leipzig, 1911, Bd. 2, 8. 524.

DEVELOPMENT OF THE HUMAN HEART 257
In the youngest embryos studied the reticular mass between
the endothelial cells and the muscle wall appears either homogene-
ous, or as composed of most delicate fibrils, or of coarser fibers,
according to the method of preservation. In general it appears
like the most delicate reticulum of the mesenchyme and under all
circumstances any stain which tinges the fibrils tinges also the
endothelial cells. So intimate is this connection that it forces
the conclusion that the fibrils together with the endothelial cells
form a syneytium. In the younger hearts the endothelial nuclei
lie altogether on the inner side of the fibrils as has been repeatedly
observed, but as soon as the trabecular system begins to form in
the left ventricle some of the endothelial nuclei invade the fibrillar
layer. This is first seen in embryo No. 239 (44 mm. long). Here
the trabecular system is quite completely formed in the left
ventricle by an interlocking of processes from the muscular and
endothelial layers. In the right ventricle the process is not so far
advanced, while in the atrial canal and the bulb the reticular
layer is invaded by endothelial nuclei but not by muscle cells.
A similar arrangement is found in embryo No. 3, which as No.
239, is intensely stained. Fig. 5 is from the posterior endocardial
cushion of an embryo 4.3 mm. long, (No. 148), showing that the
nuclei of the cushion are invading it from its endothelial side. All
this is more pronounced in No. 463 (3.9 mm.) which is more ad-
vanced in development and is perfectly preserved. In this speci-
men it is quite easy to demonstrate that the nuclei of the endothe-
lium and the reticular mass belong together, for they are distinctly
intermingled and yet are separated from the muscular layer (fig.
6). Since the nuclei and fibrils belong together and since it has
been demonstrated that the reticulum of the liver is developed
from the endothelial cells, [ shall speak of the reticulum between
the endothelium and muscle layer of the heart as endothelial
fibrils. The great importance of this distinction is at once appar-
ent for it shows that connective tissue arises also from endothelial
cells and that the intima of the entire vascular system including the
the valves of the heart has a like origin.
I think that I have now shown that the endothelial fibrils
are constant in the heart and that we must holdthe endothelial

258 FRANKLIN P. MALL

cells responsible for the production of the connective tissue of the
endocardium as well as of the valves. Further study will prob-
ably show that endothelial connective tissue is by no means of
rare occurrence. At any rate it has been definitely settled that
the endothelial cells of the liver give rise to the connective tissue
of the liver lobule.

Fig. 5 Section of the posterior endocardial cushion of an embryo 4.3 mm. long
(No. 148). X 360.

In my study on the development of the connective tissue I
was astonished to find in macerated and digested frozen sections
that the endothelial tube with its surrounding reticulum can be
isolated.’ Insuch specimens it is impossible to separate the nuclei
from this continuous mass of reticulum; together they form a syn-
cytium. This connection was demonstrated in pig embryos 20
mm. long. Although this was entirely out of harmony with the
results obtained for other connective tissues, which always arise
from the mesenchyme, it had to be accepted and so it was recorded.

® Mall, Development of the connective tissues from the connective syncytium
Amer. Jour. Anat., vol. 1, 1902, p. 354.

DEVELOPMENT OF THE HUMAN HEART 259

Ina measure this was confirmed by Kon® who observed the devel-
opment of the reticulum in the liver of a foetus in the middle of
pregnancy. Mollier!? in his beautiful study on the development
of the blood shows conclusively the connection between the endo-

Fig. 6 Section of the anterior endocardial cushion in the atrium of an embryo
3.9 mm. long (No. 463). > 360.

thelial cells of the liver and the surrounding reticulum. This
he follows back to a human embryo 10 mm. long and in subsequent
stages the connection of the endothelial with the reticulum is

9 Kon, Die Gitterfasergeriist der Leber, etc. Archiv. fiir Entwickl.-Mechanik,
Bd. 25, 1908.

10 Mollier, Die Blutbildung in der embryonalen Leber des Menschen. Archiv
fiir mik. Anat., Bd. 74, 1909.

260 FRANKLIN P. MALL

complete, that is, it forms a syneytium. Although Mollier
believes that the capillaries of the liver arise from the mesenchyme
of the capsule, which is impossible, it answers our purpose to
state that he shows that the connective tissue of the liver develops
from the endothelial cells and not from other mesenchyme cells.
Those of us who see the primary vascular tree of the liver arising
from the endothelhal wall of the omphalomesenteric veins by a
process of reduction of this large vessel to form sinusoids, recog-
nize Mollier’s “‘ origin” of capillaries of the liver as only a second-
ary contact between the sprouting capillaries when they reach
the mesenchyme of the capsule of the anlage of the liver. For
the present purpose it is clear that Mollier demonstrates also that
the connective tissue of the liver arises from endothelial cells.
This relationship has been amply confirmed by F. T. Lewis in
the liver of a human embryo 7.5 mm. long.!! So for the liver the
chain is complete; throughout development the connective tissue
of the lobule is in direct continuity with the endothelial cells of
the blood capillaries and therefore they give rise to them. Within
the lobule there are only endothelial cells and epithelial cells and
‘no one finds the reticulum arising from the latter.”

In embryo No. 239 the endothelial fibrils are quite unequally
distributed throughout the heart. In the atria, as mentioned
above, they form but a very thin layer. In the atrial canal the

1 Lewis, F. T., Entwicklung der Leber. Keibel-Mall, Handbuch der Entwickl.,
Bd.2,1911. §.397, fig. 291.

Quite recently Mollier finds the endothelial cells and reticulum of the spleen
as one continuous reticulum from which he concludes that the endothelial cells
arise directly from the mesenchyme (Arch. f. mik. Anat., Bd. 76,1911). The oppo-
site conclusion may be drawn equally well. I have published figures which corre-
spond with Mollier’s, giving at the same time a conclusive experiment to prove that
the circulation of the spleen is entirely through the pulp spaces; these have the
value of blood capillaries (Mall, Amer, Jour. Anat., vol. 2, p. 315). Pathologists
have been of the opinion that in endarteritis the intima thickens by a proliferation
of endothelial cells, and that connective tissue may arise from these cells. Mar-
chand has questioned the truth of this statement, but recently the subject has been
reinvestigated by Baumgarten (Arbeiten auf dem Gebeite der Pathologischen
Anatomie, Leipzig, 1904, Bd. 4) who showed that proliferation of endothelial cells
may form a thickening of the intima without any rupture of the elastica interna,
thus excluding entirely any participation of connective tissue cells in the process.
See also von Szily, Anatom. Hefte 35, 1908, Taf. 45-47.

DEVELOPMENT OF THE HUMAN HEART 261

fibrils are heaped up into two mounds to form the well known
endocardial cushions, which have between them a transverse
slit. The posterior cushion extends upward into the left atrium
and then along its posterior surface into the right atrium and ends
at the opening of the sinus venosus. The anterior cushion also
extends into the left atrium along its anterior border and reaches
to the septum primum which is just beginning to form. Below,
in the left ventricle, the endocardial cushions blend with the endo-
thelial reticulum covering the trabeculae. The interlacement of
the endothelial and muscular layers to form the trabeculae extends
into the right ventricle, but in the bulb the two layers are quite
sharply defined and separated.

The nuclei of the endothelial syncytium form first of all the
inner layer of the heart, but in the endocardial cushions of the
atrial canal as well as in the bulb the nuclei gradually extend
towards the muscular coat. In other words the nuclei of the inner
coat are gradually invading their reticular layer.

In the heart of embryo 463 the differentiation of the endothe-
lial syneytium is more pronounced than in the specimen just
described. The heart is now well formed with two pronounced
atria, a much constricted atrial canal and a marked constriction |
of the interventricular canal (fig. 4). The bulbo-ventricular and
the interventricular grooves are well formed. The septum of
the atria and that of the ventricles are well marked. The endo-
thelial syncytium is most pronounced in the endocardial cushions
and in the bulbus. The cells are quite equally distributed through-
out the syneytium but they are somewhat more numerous imime-
diately under the endothelial covering than near the muscle layer
of the heart (fig. 6). The posterior cushion does not reach as
far into the left ventricle-as the anterior and is also less extensive
in the atria; it reaches nearly to the sinus venosus. The anterior
endocardial cushion is a large sickle-shaped affair, encircles the
heart in front as the border of the atrial septum (septum primum)
which is now forming. The anterior cushion ends on the medial
side of the opening from the sinus venosus. The space in the
atria between the cushions marksthe primary foramen ovale (fora-
men ovale I).

262 FRANKLIN P. MALL

In an embryo 4.8 mm. long (No. 148) practically the same con-
ditions are seen as in the embryo just described. If anything
it is a little more advanced in development. A section of the
endocardial cushion is shown in fig. 5. From now on new con-
ditions arise which when concluded separate the heart into its
right and left halves.

The anterior and posterior cushions are now well formed, the
superior septum (primum) and the septum of the ventricles (sep-
tum inferior) are beginning but the septum aorto-pulmonale
(aortic septum) is still absent. While these are forming, up to
the next stage, when the muscle wall of the atrial canal begins to
break down, the heart gradually enlarges without changing very
markedly its external form. The steps which I am about to de-
seribe are well established in various mammals, but I shall repeat
them hastily in order to confirm them allin the human heart. In
doing this I shall include in the descriptions the following speci-
mens:

Embryo No. 80, C. R. length 5 mm.

Embryo No. 1386, C. R. length 4 mm.
Embryo No. 116, C. R. length 5 mm.
Embryo No. 241, C. R. length 6 mm.
Embryo No. 2, C.R. length 7 mm.
Embryo No. 383, C. R. length 7 mm.
Embryo No. 380; C. R. length 7.5 mm.
Embryo No. 113, C. R. length 8 mm.
Embryo No. 397, C. R. length 8 mm.
Embryo No. 422, C. R. length 9 mm.
Embryo No. 163, C. R. length 9 mm.

In Embryo No. 80 the anterior and posterior cushions are con-
siderably thicker than before but hold practically the same rela-
tion to the heart as in Nos. 463 and 148. In the left ventricle the
cushions are spread out and have attached themselves to the
trabecular system. As there are two attachments which corre-
spond in position to the anterior and posterior papillary muscles,
it is proper to speak of them as such. The septum aorto-pul-

DEVELOPMENT OF THE HUMAN HEART 263

monale is well formed and its two ridges reach well into the bulbus
to the interventricular foramen. Much the same arrangement is
found in Embryos No. 136, 116 and 880 which are of the same
stage of development as No. 80.

Fig. 7 Transverse section of the atrial canal and bulb of the heart of anembryo
9 mm. long (No. 422). x 40.

Fig. 8 Transverse section of the heart of an embryo 7 mm. long (No.2). & 40.

In Nos. 241, 422 and 2 the lumen of the bulb is »-shaped, the
endocardial cushions are much more pronounced than before;
they are ready to fuse as is indicated in figs. 7 and 8. The sep-
tum primum and the interventricular septum are well marked.
In 397 the septum primum is very thin above so that it is uncer-
tain whether or not it has broken through to form the foramen
ovale II.

264 FRANKLIN P. MALL

In 388, 118 and 168 the foramen ovale IT has just formed, being
smallest in the first and largest in the last. The cushions are well
developed in No. 383; the anterior reaches to the septum primum
and the posterior to the opening of the sinus into the right atrium.
There is a large space between them connecting the two atria.
High up in the atria the septum primum is broken though forming
the foramen ovale II, as shown by Born. The arrangement of
the endocardial cushions with the space between them and the

Fig.9 Sagittal section. Embryo8 mm. long (No.113). x 40.

foramen ovale II is well shown in No. 113 (fig. 9). Both cushions
now course to the sinus venosus and are blended with the con-
nective tissue above it. Behind the cushions there is a muscle
strand from the sinus to the ventricle which marks the position
of the atrio-ventricular bundle. The thin septum primum reaches
to the two cushions, and high up it is perforated by an opening
with sharply defined walls. Between the anterior and posterior
cushions well within the atrial canal the two atria still communi-
cate with each other; in this region the cushions are as yet not

DEVELOPMENT OF THE HUMAN HEART 265
blended. In No. 163 the blending of the cushions is complete
(fig. 10) and the permanent foramen ovale is fully established.
Together the united endocardial cushions form a cubical plug

which blocks the center of the atrial canal leaving a channel on

It also projects into the left ventricle and is attached

either side.
The atrial canal

to its walls forming the two papillary muscles.
is divided into two canals which now form the right and left ostia.

The two cushions give rise to the medial cusps of the bicuspid
and tricuspid valves, that is, the medial cusp of the tricuspid
valve and the anterior cusp (B.N.A.) of the mitral valve; the
right halves of each cushion make the former and the left halves

the latter.

i

(| HK at ul

Fig. 10 Coronal section. Embryo 9 mm. long (No. 163). X 40.

The septum aorto pulmonale is still incomplete in the speci-
mens described, and the interventricular foramen is wide open,
but the septum of the ventricle is well formed, extends upward
and blends with the posterior endocardial cushion behind. Be-
hind the posterior endocardial cushion a marked band of muscle
extends from the wall of the sinus venosus to the interventricular
foramen where it spreads over the inner walls of the two ventricles
as shown in fig. 9. This is the sino- or atrio-ventricular bundle
(bundle of His). These structures I shall describe in the heart of
embryo No. 353 which is an unusually well preserved specimen,
perfect in every respect and of the right stage of development

for this purpose.

266 FRANKLIN P. MALL

Embryo No. 353 is 11 mm. long with a well formed arm and
hand plate. A profile outline of the embryo may be seen in the
figure by Evans." The sections are 10u thick in a coronal direc-
tion and slightly oblique, that is, they strike the heart transversely.
The heart which has been modeled in wax is 1.7 mm. wide and 2.2
mm. long. The apex is cleft as is so often the case in this stage of
development.'* The septum aorto pulmonale is complete and the
anterior and posterior cushions are fully united into a single mass

Fig. 11 View from below of a model of the heart of an embryoll mm.long (No.
353). > 50. The ventricles have been cut off. The connective tissue septa are
colored yellow.

of connective tissue. This mass extends to a point up in the atrial
septum on its dorsal side, to the left valve of the opening into
the sinus venosus. The complete union of the two cushions has
obliterated the foramen ovale I and the foramen ovale ITI is well
above the common fibrous process of the united cushions. A
view from below, that is after the apex of the heart is cut off,

13 Wyans, Keibel-Mall Handbuch, Bd. 2, fig. 469.
14 Mall, Bifid apex in the human heart. Anat. Rec., vol. 6, 1912.

DEVELOPMENT OF THE HUMAN HEART | 267

shows that the cushions mark the borders of the medial sides of
the right and left ostia, these being notched to indicate the extent
of the anterior and posterior cushions. This is well shown in
Fig. 11 which was drawn from the model. On the dorsal side,
the posterior cushion extends well down the posterior border of
the interventricular foramen, that is it makes part of the border
of the septum of the ventricle. These are the primary connec-
tions of the two endocardial cushions. There are also secondary
connections which in a measure involve the valves lateral to the
ostia.

On the lateral side of either ostium there is a rounded endo-
cardial thickening which marks the beginning of the lateral valves.
These are already observed in Embryo 422, 9 mm. long (fig. 7).
To anticipate the description I may state that the right cushion
marks the center of the anlage of the anterior and posterior cusps
of the tricuspid valve, and the left the anlage of the posterior
cusp of the mitral valve as seen in fig. 11. The septum aorto
pulmonale soon blends with the cushions through a dorso-lateral
wing which is divided into two branches to encircle in part the right
venous ostium, one of which blends with the right lateral endo-
cardial thickening, and the other, the medial, blends with the
anterior process of the medial valves now represented by the right
lower wing of the anterior endocardial cushion. It is thus seen
that through the blending of the septum aorto pulmonale with the
right side of the anterior endocardial cushion, the right venous
ostium is nearly encircled by endothelial connective tissue. This
connection may still be seen in the adult heart where the septum
aorto-pulmonale (the tendon of the conus) is found to blend with
the fibrous ring of the right ostium at the anterior border of the
attachment of the medial cusp of the tricuspid valve.

The large space marked by the interventricular foramen at
the root of the aorta remains constant in all subsequent stages of
development and is termed by Quain" the vestibule of the aorta.
This name, which is appropriate, I shall adopt and use in my de-
scription. The vestibule in fig. 11 is common to both ventricles,

1 Quain’s Anatomy, 10th Edition, vol. 2, fig. 317.

THE AMERICAN JOURNAL OF ANATOMY, -VOL. 13, NO. 3

268 FRANKLIN P. MALL

but as the septum of the ventricles and the septum aorto pul-
monale approach each other more and more to form the perma-
nent membranous septum, the vestibule becomes transferred to the
left ventricle as may be seen in figs. 16 and 17. An open inter-
ventricular foramen in the adult always communicates between
the aortic vestibule and the space below the medial cusp of the
tricuspid valve, as is clear by observing Spalteholz’s figure.

The topography in the wall of the left ventricle is much easier
to define. 'The common endocardial mass borders the left venous
ostium and each of its two horns are continuous with pro-
nounced muscular bands, the papillary muscles, which extend to
the more solid muscular wall of the heart. In their course from
the valve to the outer wall of the heart muscle the papillary mus-
cles communicate continuously with the trabecular system. Both
the anterior and posterior papillary muscles connect with the
lateral valve which is being extended around the left ostium by
an ‘‘undermining”’ process, already well described by His. So in
this early stage of development the anterior papillary muscle
unites the anterior tip of the medial and lateral valves (anterior
and posterior B.N.A.) with the anterior wall of the left ventricle,
and the posterior muscle unites the posterior tip with the posterior
wall of the heart as seen in fig. 11. The vestibule of the aorta
connects the aorta with the left ventricle; it is encircled by the
border of the ventricular septum. In the course of time the
border of the ventricular septum unites with the septum aorto
pulmonale and thus finally separates the two ventricles of the
heart. When viewed through the aorta the muscular interven-
tricular septum usually makes in the adult the right border of the
vestibule but often it projects into the vestibule as isnormally
the case in the pig and the ox. In such specimens, as well as in
the pig and the ox, the right semilunar valve arises directly from
the interventricular muscular septum. This shows to what ex-
tent the valves must ‘sink into’ the bulbus in passing from the
stage represented in No. 353 to the adult form.!7

16 Spalteholz, Hand atlas, vol. 2, fig. 420.
17 The normal position of the membranous septum is described in an article on
aneurysms arising from it, in the Anatomical Record, vol. 6, 1912.

DEVELOPMENT OF THE HUMAN HEART 269.

The anatomy of the heart of embryo No. 353 may serve as a
basis to describe the final closure of the interventricular opening
and the formation of the membranous septum. In this specimen
the septum aorto pulmonale has grown down to the interventric-
ular foramen and blends with the right tip of the anterior endo-
cardial cushion which is lodged in the foramen. The aortic sep-
tum. also extends to the lateral side of the right venous ostium
and the posterior cushion extends down the anterior border of
the ventricular septum. So the interventricular foramen is

Fig. 12 Coronal section. Embryo 13 mm. long (No. 175). X 40.

bounded above by the union of the septum aorto pulmonale and
the anterior cushion, in front by the septum aorto pulmonale,
behind by the extended portion of the posterior cushion and
below by the muscle of the ventricular septum.

In embryos 109 (10.5) and 317 (12.5) the ventricular septum
is much more developed than in the stage just described. In
these two specimens the fibrous tissue forming the septa and
valves is much as in No. 353, but the muscle wall has grown up
and nearly closes the interventricular foramen. However, there

270 FRANKLIN P. MALL

is still a free communication between the right ostium, with the
right ventricle and the vestibule of the aorta. The interventricu-
lar foramen is somewhat smaller in No. 175 (18 mm.) which is
cut in a more fortunate plane to illustrate this point than the

~ Tc

Fig. 13 Sagittalsection. Embryo 14mm. long (No. 144). X 40.

Fig. 14 Coronalsection. Embryo 17.2 mm. long (No. 424). X 24.

other two embryos (fig. 12). The opening is gradually becoming
smaller in the following embryos in the order of their enumera-
tion, No. 423 (15.2), 144 (14, fig. 18), 424 (17, fig. 14). In 390
(15.5) it is uncertain whether it is closed or not, in 409 (16) it is

DEVELOPMENT OF THE HUMAN HEART PT

just closed, and in 482 (18) it is closed but its connection with the
right ventricle is still indicated. That the order of development
does not correspond with the length of the embryo is due to the
method of measuring; No. 144 was measured on the glass slide.
But a comparison with the profile drawings of these specimens
shows that the order of closure of the foramen corresponds with
the degree of development of the external form. In No. 423 the
interventricular foramen (0.1 mm. in diameter) is situated well
anterior, at the point of junction between the septum aorto pulmo-
nale and the right wing of the anterior endocardial cushion. It is
under the medial cusp of the tricuspid valve in exactly the position
taken by the atrio-ventricular bundle. In No. 424 the foramen
is barely 0.02 mm. in diameter, and were not the vascular system
injected with india ink the opening would probably be overlooked.
It is present in but a single section. Here it is again located with
the right limb of the atrio-ventricular bundle below the medial
cusp of the tricuspid valve well anterior. On the left side it
communicates with the vestibule of the aorta exactly in the posi-
tion the atrio-ventricular bundle lies in the adult heart. That
this is of significance will be considered when the atrio-ventricular
bundle is discussed. In this stage the posterior cusp of the aortic
valve still lies somewhat distant, but as the valves sink deeper
and deeper into the vestibule of the aorta the position of the inter-
ventricular opening comes to he behind the posterior cusp adja-
cent to the left limb of the atrio-ventricular bundle. In an adult
heart with a patent interventricular foramen I have found the
atrio-ventricular system streaming through this opening, thus
showing that there is an association between them.

The atrio-ventricular valves are not as difficult to trace in their
development in the successive stages as has been that of the for-
mation of the membranous septum. In the youngest embryos,
that is those under 3.5 mm. long, the endocardial connective tissue
which was quite equally distributed in the earliest stages has
gradually rearranged itself. first becoming less pronounced in the
atria and then becoming well dove-tailed with the trabecular
system in the ventricle and bulb. As soon as the atrial canal
is well formed the endocardial connective tissue arranges itself

272 FRANKLIN P. MALL

there in the form of two cushions. the anterior and posterior endo-
cardial cushions. In the bulb two ridges are also formed which
ultimately give rise to the septum aorto pulmonale, but as this
latter structure has been considered by Greil and by Tandler, I
have taken it up only in so far as it bears upon the formation of
the membranous septum.

In the embryo 3.9 mm. long (No. 463) the two endocardial cush-
ions of the atrial canal are wellformed (fig.4). The posterior is short
and reaches from the sinus to the ventricle, while the anterior is
much more extensive for it reaches from the sinus also around the
upper and anterior part of the atrium through the atrial canal to
the bottom of the ventricle. In general they repeat that which
is shown in Greil’s fig. 3 taken from the heart of Lacerta.’ The
two cushions are confined almost wholly to the left ventricle.
However, the side of the lower tip of the anterior cushion passes
through the interventricular foramen and continues as the anterior
medial endocardial thickening of the bulb. A little later at 4.8
mm. (No. 148) the same arrangement is still seen except that the
left lateral tip of each of the two cushions is more intimately
attached to the trabecular system of the ventricle. These attach-
ments mark the beginning of the anterior and posterior papillary
muscles. The cushions gradually become more and more pro-
nounced until the embryo is 7 mm. long (No. 2) when their lower
right tips begin to enter the interventricular canal. With the
formation of the septum primum the anterior cushion gradually
approaches the posterior with which it ultimately blends (fig. 9).
Before this takes place the septum primum is perforated forming
the foramen ovale II; somewhat later the foramen ovale I is com-
pletely obliterated.!® By this time the two cushions have blended
into a solid mass which obstructs the atrial canal, leaving on either
side an ostium. The common cushion or valve mass is now
wider than it is thick (fig. 15), hangs well into the left ventricle
where its two corners or tips are well supported by the two papil-
lary muscles (figs. 11 and 12). The right half rests upon the

18 Greil, Beitrage zur vergleich. Anat. u. Entwicklungsg. d. Herzens u. d. Trun-
cus arteriosus d. Wirbelthiere. Morph. Jahrbuch, Bd. 31, 1903.
19 See also Greil, l.c., figs. 5 and 12.

DEVELOPMENT OF THE HUMAN HEART 2

inferior septum behind (the right posterior tip, projects well
into the interventricular foramen, while the right anterior tip
blends with the septum aorto pulmonale. It becomes clear, by
comparing this stage with older ones, as well as with the adult
(figs. 15 to 18), that this common central mass is destined to pro-
duce the medial valves of the two venous ostia, namely anterior
cusp of the mitral valve and the medial ‘cusp of the tricuspid valve.

But in addition to the anterior and posterior endocardial cush-
ions there are also two lateral cushions well demonstrated by
His and by Greil. One of these, the left lateral, is first seen in an
embryo 9 mm. (No. 422) long, (fig. 7), but both are not well pro-
nounced until the embryo is 11 mm. long (No. 353, figs. 11 and
16). At this time a lateral’ process from the septum aorto pul-
monale reaches to and blends with the right lateral cushion. The
left lateral is still sharply defined and isolated. The line of con-
nection between the right lateral cushion and the septum is always
marked in the adult by a tendon which passes into the aortic sep-
tum or by a small papillary muscle which holds the same position.
This muscle is called medial papillary muscle by Henle.2° Between
the tendon or muscle and below the anterior tip of the medial
cusp there is alway a clear field, the bottom of which is formed by
the membranous septum.

Semi-diagrammatic reconstructions of the tendinous masses
at the base of the heart are shown in figs. 15 to 18. Fig. 15 is
from the heart of embryo No. 163 (9 mm.) with the left lateral
cushion from No. 422 (9 mm.) added. The muscular wall of the
atrial canal is indicated and the connective tissue is stippled.
Fig. 16 is from embryo No. 353 (11 mm.) which shows that the
origin of the aorta has shifted well into the left ventricle. In
fig. 17, No. 296 (17 mm.) the sinuses marking the three semi-
lunar valves are indicated upon the two wings of the septum aorto
pulmonale, and upon the anterior endocardial cushion. Fig.
18 shows the arrangement of the connective tissue at the base of
the adult heart. The parts of the valves which are formed by

20 Henle, Gefasslehre, II Auflage, 1876, fig. 33. This tendon is constant, in fact
the most constant of all of the attachments of the valves in the right ventricle.
It is pictured in all anatomies, but is not recognized in the B.N.A.

\

274 FRANKLIN P. MALL

the undermining process are indicated by cross hatching. The
course of the atrio-ventricular muscle bundle dividing within the
interventricular foramen and crossing the ventricular septum,

Fig. 15 Semidiagrammatic reconstruction of the heart of an embryo 9 mm.
long (No. 163). The left lateral endocardial cushion has been added from another
embryo of 9 mm. (No. 422).

Fig. 16 Semidiagrammatic reconstruction of the heart of an embryo 11 mm.
long (No. 353). The muscle encircling the atrial canal is reduced and the atrio-
ventricular bundle is seen passing behind the endocardial cushions to spread over
the ventricular septum below.

is also shown. These figures indicate clearly the fate of the pri-
-mary dividing masses of connective tissue at the base of the heart -
in the embryo.

DEVELOPMENT OF THE HUMAN HEART (5)

In young hearts the septum aorto pulmonale encircles the right
ostium as shown in fig. 17 and from it the tricuspid valve arises.
In the adult the septum aorto pulmonale blends with the anterior

A.Y.B.

Fig.17 The same as fig. 16 from the heart of an embryo 17 mm. long (No. 296).
Most of the muscle wall of the atrial canalis absent. The atrio-ventricular bundle
is shown and the semilunar valves of the aorta are indicated.

Fig. 18 Sketch of the base of an adult heart showing the valves. The atrio-
ventricular bundle is shown. The figure may be compared with figs. 15, 16 and
We

end of the medial and anterior cusp of the tricuspid valve, while
the same tip of the valve is tied to the conus by the medial tendon
which frequently enlarges into a papillary muscle. In following
its development it is found that this tendon marks the primary

276 FRANKLIN P. MALL

union of the septum aorto pulmonale to the valve and in its sub-
sequent development is separated from the septum as the valve
enlarges. The beginning of this process is shown in fig. 19 which
is from a section of embryo 86 (80 mm.). It is here seen that the
valve mass is enlarging and that its lower border is being sepa-
rated from the muscle of the right ventricle as the tendinous cords
are forming. In so doing the medial papillary muscle is forming
and at the same time the trabeculae which encircle the base of the

co
o
fu

co
Ny

08 &

Muy
LF,

Miny
Dy

——=

Fig. 19 Section through the conus and right ostium of the heart of an embryo
30 mm. long (No. 86). X 20. The attachment of the tricuspid valve to the sep-
tum aorto pulmonale and the ventricular septum is shown.

right ventricle as shown in fig. 11, become fused to produce the
crista supraventricularis.2!_ This at first ends in the septum aorto
pulmonale but in the adult it passes around and over the medial
tendon and passes down the anterior inner wall of the right ven-
tricle. In the embryo the moderator band arises just below the
medial cusp of the tricuspid valve, but in its further development
is shifted towards the apex where it binds the large papillary mus-

21 The ‘crista’ is correctly figured in Toldt’s Atlas, fig. 946, while in Spalteholz,
fig. 424, it is pictured as extending down to the base of the large papilary muscle.
It is this extension which contains the right limb of the atrio-ventricular bundle
and therefore must represent the moderator band. A study of the development of
this portion of the heart shows that this is the case.

DEVELOPMENT OF THE HUMAN HEART Bd.

cle with the extension of the crista supraventricularis.2 In the
ox, pig and sheep, however, it retains its original position and
contains the right limb of the atrio-ventricular bundle.

It is thus seen that the medial cusp of the tricuspid valve is
attached in front by the medial tendon (or muscle), and behind
by the large papillary muscle, and the inequality of these two
structures accounts for the double appearance of the lateral valve.
In reality no true tricuspid valve is present and correctly speaking
there is no tricuspid valve. Both are bicuspid with medial and
lateral cusps. Both are tied down by two muscles, the two papil-
lary muscles on the left side and the large papillary muscle and
the median tendon on the right side.

The atrio-ventricular cushions expand not only by their own
growth but to a greater extent by a process of undermining the
ventricle wall all around the venous ostia.. To what extent this
burrowing has taken place is marked by the attachment of the
valves to the muscle walls of the heart. The tendons nearest the
tips of the valves were the first to form, while those nearer the
bases of the valves were formed subsequently. This has all been
demonstrated by His. By this process of undermining the attach-
ment of the base of the valve to the wall of the ventricle, the atrial
portion is telescoped into the ostia. The muscle of the atrium

22 'Toldt’s ‘crista’ (l.c., fig. 946) ends in the medial papillary muscle (anterior),
which according to its development is correct. Spalteholz (l.c., fig. 424) extends
the crista past the anterior tendon down to the base of the large papillary muscle
(posterior). This connecting band contains the right limb of the atrio-ventricular
band and is pictured by Tawara on his plate 7. Retzer (J. H. H. Bull., vol. 20,
1909, and Anat. Rec., vol. 6, 1912) associates the moderator band with the crista,
in fact he says that when absent it is represented by the crista. He recognizes
fully the meaning of the anterior (medial) papillary muscle, but when it is recalled
that the right limb of the atrio-ventricular band passes on the posterior side of
this muscle and the crista on its anterior, the identity of these two structures is
disproved. In fact in the embryo the crista reaches to the medial muscle and the
moderator band is considerably below it. In the adult heart this band is pushed
to the base of the large (posterior) papillary muscle as pictured in Spalteholz and
it contains the right limb of the atrio-ventricular bundle as shown by Tawara;
this I have been able to confirm. His, who introduces this term ‘crista’ (Beitrage
zur Anat. d. Mensch. Herzens, Leipzig, 1886, 8. 9) is not clear in his description of
its attachment to the septum. According to its development it should end in the
septum aorto pulmonale, that is, at the point of origin of the medial papillary
muscle.

278 FRANKLIN P. MALL

extends into the atrial part of the valves, which at first is continu-
ous through the tendinous cords with the ventricular muscle. At
first the atrio-ventricular muscle connection is through the main
wall of the ventricle, but as this is resolved into the trabecular
system with the growth of the valves and the formation of the
papillary muscles the connection between atrium and ventricle
is through the tendinous cords which are at first muscular. Later
with the degeneration of the muscle in the cords the muscular
connection between the atria and ventricles was believed to have
been broken down completely. In lower vertebrates the mus-
cular connection between the atria and ventricles through the
trabecular system remained throughout life, and the significance
of this has been fully demonstrated by Gaskell”? and His, Jr.*4
In man, however, all of the cords are converted into connective
tissue, but the muscle of the atria extends well into the valves or it
may reach to its free border (Kirchner).” It has also been
stated that in some rare instances the tendinous cords of the mitral
valve remain muscular in the adult (Oehl).2°. But the muscular
connection through the valves is first interrupted at the free thin
edges and later the muscular fibers in the tendinous cords dis-
appear. The break in the muscular connection at the free edges
of the valves will be considered in discussing the atrio-ventricular
bundle.

Returning to the description of embryo No. 353 (11 mm.),
it is possible to determine with considerable precision the various
adult connections of the valves. The most definite valve is the
anterior cusp of the mitral which is formed by the union of the
left lateral tips of the anterior and posterior endocardial cushions.
Each tip is here bound to the trabecular system by well formed
muscle strands, one of which passes to the anterior wall of the
heart and the other to the posterior as well as to the septum. It
is clear that these two strands, which appear much earlier than in

23 Gaskell, On the innervation of the heart with special reference to the heart of
the tortoise. Jour. Phys., vol. 4, 1884.

24 W. His, Jr., Die Thitikeit desembryonalen Herzens. Arbeiten aus der med,
Klinik zu Leipzig, F. C. W. Vogel, 1893.

25 Kiirchner, Wagner’s Handworterbuch, vol. 2, 1844.

2€ Oehl, Henle’s Anatomie, Gefiisslehre, Braunschweig, 1876.

DEVELOPMENT OF THE HUMAN HEART 279

this specimen and continue throughout development, are the
anlages of the anterior and posterior papillary muscles respectively.
In addition a heavy strand of tissue encircles the lateral side of
the left ostium and unites the apices of the two papillary muscles.
In the middle of this, protruding into the left ostium, is seen the
conspicuous left lateral cushion already noted in an embryo
9 mm. long (No. 422, fig. 7). This arrangement corresponds with
what is found in the adult heart. Each papillary muscle not only
attaches itself to the tips of the anterior cusp but also to the tips
of the posterior cusp. (It is confusing to use the B.N.A. terms
anterior and posterior in naming the cusps when the embryological
terms should be medial and lateral.) Between the two papillary
muscles there is often a third or lateral papillary muscle, or nu-
merous small muscles or the two muscles may be widened to fill
this area. At any rate the structures found around the left ostium
in embryo 353 represent fully what is present in the adult. We
have in all cases two papillary muscles, each of which communi-
cates freely with the tips of the two cusps of the mitral valves.
On the dorsal side and in front of the left ostium the muscle of the
atrium and ventricle is continuous, as is shown in fig. 16. ©

As the heart grows larger more chordae tendinea are formed
necessarily nearer to the base of the heart. It follows that the
primary condition found in No. 353 represents only the tips of the
valves and that their subsequent enlargement is due partly to
stretching of the anlage and partly to the undermining of the wall
of the ventricle.

The united anterior and posterior endocardial cushions were
‘projected at first into the left ventricle but subsequently its left
tip becomes lodged in the interventricular foramen (fig. 11).
To this smaller portion the septum aorto pulmonale attaches
itself anteriorly and the inferior muscular septum posteriorly.
The left half hangs freely within the left ventricle throughout life,
forming a loose flap or sail which is suspended between the left
venous ostium and the vestibule of the aorta. While the semi-
lunar valves of the aorta are somewhat distant from this flap in
embryo No. 353, later on they are pushed down to it so that the
aorta becomes finally attached to its base (fig. 17).

280 FRANKLIN P. MALL

On the right side of the heart the septum aorto pulmonale
soon blends with the lateral tip of the anterior cushion and by
the time the septum is complete, as in No. 3538, it divides, one
portion of which is attached directly to the left anterior tip of
the common cushion and the other encircles the right venous
ostium and blends with the right lateral cushion. The posterior
side of the ostium is formed by trabeculae which pass from the
lateral cushion mainly to the muscular septum and others which
course symmetrically into the trabecular system of the ventricle.
It is by no means easy to determine three systems to correspond
with the three cusps of this valve. Neither is it easy to recognize
the three cusps in the adult unless we associate the medial papil-
lary muscle, which is constant, with the anterior cusp and the
large papillary muscle with the posterior cusp: If this is done it
is possible to name the valves in the heart of embryo No. 356 as
follows (figs. 15 to 18):

The median cusp is attached in this embryo to the septum aorto
pulmonale in front and tothe muscular septum behind. It hascor-
responding attachments in the adult. The anterior cusp is attached
partly ‘to the septum aorto pulmonale and to the large papillary
muscle. The posterior cusp is attached to the large papillary
muscle. If this division is correct the lateral cushion belongs to
the anterior cusp while the posterior cusp is developed entirely
from the lateral side of the ventricle. The medial papillary mus-
cle is constant and always extends as a muscle, or as a tendon,
from the region of the septum aorto pulmonale (tendon of the
conus) to the anterior cusp which is marked by a similar strand
of tissue from the septum to the lateral cushion in the embryo.
That this is the case is shown in sections of older hearts, one of
which is given in fig. 19. The posterior cusp, over the large papil-
lary muscle, is of irregular formation and by a process of exclu-
sion the muscle strands in the embryo which encircle the right
ostium behind the right cushion, should give rise to it. At any
rate the right and left ostia of numerous hearts of embryos less
than 20 mm. long, are slit-like and similar, so that it is impossible
to state that one is encircled by a tricuspid valve. Each is
bordered by medial and lateral valves, and each medial and lateral

DEVELOPMENT OF THE HUMAN HEART 281

valve is bound to the ventricle by anterior and posterior papil-
lary muscles. In the course of time the valves and papillary
muscles will be renamed to correspond with their development,
which will also accord with the adult condition.?’

C. THE ATRIO-VENTRICULAR BUNDLE

No complete deseription can be given to the formation of the
tricuspid and mitral valves without considering the fate of the
muscular wall of the atrial canal. This at once defines definitely
the atrio-ventricular bundle which is embryologically the rem-
nant of the auricular canal after the greater portion of its muscular
wall has been destroyed by the formation of the valves.

His, Sr., described the breaking down of the muscular wall of
the atrial canal in small embryos and states that it is nearly broken
down in an embryo 13.8 (?) mm. long. Later His, Jr., showed
the presence of a muscle bundle connecting atrium with ventricle
in the new-born child and discusses the connections between
atrium and ventricle in lower animals as well as in the embryo.
He noted that the character of the heart beat changed in the chick
at the time the atrial canal is forming, at the time the muscular
connection between the atria and ventricles is reducing.?® It is
easy to read into this paper that His thought that the muscular
connection between the atria and ventricles never does break down
completely, although he never states it. Were there acomplete
separation of the muscle wall of the atria and ventricles before the
formation of the atrio-ventricular bundle is formed heart block
should occur in the embryo. That the atrio-ventricular bundle
is aremnant of the primary connection in the embryo is mentioned
by Tandler as a possibility but he believes it improbable. He is

°7 Direct dissection of the heart of an embryo 30 mm. long (No. 90, a) shows that
the medial valves of the two ostia form a saddle hanging over the ventricular
septum. Instead of single lateral cushions double lateral cushions are present on
both sides, the posterior being larger than the anterior on the right side. A simi-
lar condition is sometimes seen in serial sections. In case the ventricles are cut
through transversely the valve system can be beautifully seen from below, and this
method of investigation will probably settle this question definitely.

SVBNSE dines Ga yo} 10)

282 FRANKLIN P. MALL

rather inclined to think that it is a new formation, as has been
asserted by Retzer.?9

In following the atrio-ventricular bundle in the foetus Fahr®®
found it in one of 160 mm. long, Tawara*! in one 100 mm. long,
Monckeberg® at 75 mm., Tandler in human embryos 28.5 and 19
mm. long, and Retzer in the pig from 15 to 20 mm. long. His,
Sr., states that the muscle wall between the atria and ventricles
is nearly broken down in an embryo 13.8 (?) mm. long. It is
evident that if the bundle is a new formation it must appear in
embryos about 15 mm. long. However, in order to anticipate my
own report upon this subject it may be stated that the bundle is
not a new formation but is the remnant of the wall of the atrial
canal after its anterior and lateral sides have been broken in the
formation of the tricuspid and mitral valves. The portion on
the posterior wall which connects the sinus-with the ventricle never
breaks down; early in development it shows changes in structure
which differentiate it from the rest of the heart muscle. It holds
a constant position just behind the posterior endocardial cushion
immediately over the septum of the ventricle, a position which
is marked at first by the interventricular foramen and later by the
membranous septum.

The atrial canal is first well formed in an embryo 3.9 mm.
long (No. 463); it is now markedly constricted and is pretty well
filled up by the two large endocardial cushions (fig.4). Its
outer wall is uniformly muscular throughout. There is practically
no changes in subsequent stages until the lateral endocardial
cushions make their appearance (fig. 15), both of which are pres-
ent in an embryo 11 mm. long (No. 353, fig. 16). At this time
there is a break in the region of the left cushion, and possibly over
the right cushion, as well as in the bottom of the bulbo-ventricular
groove, that is between the aorta and the left ventricle. At
13 mm. (No. 406) the break is more marked over the left endocar-

29 Retzer, Anat. Rec., vol. 2, 1908.

30 Fahr, Virchow’s Archiv, Bd. 188, 1907.

31 Tawara, Das Reizleitungssystem des Saugethierherzens, Jena, 1906.

32 Ménckeberg, Untersuchungen iiber des Atrioventriklarbiindel in mensch-
lichen Herzens. Jena, 1908.

DEVELOPMENT OF THE HUMAN HEART 283)

dial cushion, there are several breaks on the right lateral side, a
large one just behind the aorta and one on the medial dorsal side
of the left ostium. In other words the muscular connections
between the atria and ventricles are as follows: A large one in
front of and a large one behind the left ostium, a number of small
ones lateral to the right ostium and a large one from the sinus to
the ventricle dorsally just over its inferior septum. In another
embryo of the same stage of development (No. 317, 12.5 mm.)
there are two marked muscular connections, one in front of each
of the ostia. There is a very marked one from the sinus to
the inferior septum below. This is composed of muscle fibers
which arise as two horns to which the dorsal muscle fibers of the
atria stream. This is clearly the atrio-ventricular muscle, as
shown in fig. 17. From now on it is the chief muscular connec-
tion but for some time additional connections are often seen.
For instance in embryos 14 mm. (No. 144) and 15.2 mm. (No.
423) there is an additional connection just in front of the right
ostium, in embryos 16 mm. (409), 18.5 mm. (481) and 20 mm.
(368) there is an additional connection just in front of the left
ostium. In a number of specimens the atrio-ventricular bundle
is not confined to the region immediately over the inferior septum
but is spread out over the dorsal side of the left ostium (432, 18.
mm.; 431, 18.5 mm.; 368, 20 mm.), while in one specimen 21 mm.
long (460) there is a single additional strip on the left lateral side.
In general, however, the main bundle which is at first broad and
associated more with the left ventricle than the right, gradually
becomes constricted so that it is well formed and rounded by the
time the embryo is 20 mm. long. It is beginning to be isolated
at 11 mm. and well separated at 13 mm.

The position of the bundle is shown in sagittal sections in figs.
-9 (No. 113, 8 mm. long), 13 (144, 14 mm.), 20 (890, 15.5 mm.),
21 (482, 18 mm.), 22 (431, 18.5 mm.) and 23 (368, 20mm.). Itis
clear in all of these hearts that the muscle connection is from the
sinus between the posterior cushion and the annulus fibrosus
into the septum of the ventricle. This is its position in the adult
heart. In transverse sections the bundle is shown in figs. 8 (No.
2,7mm.), 24 (353,11mm.) and 26 (317, 16 mm.). Here again it

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, No. 3

284 FRANKLIN P. MALL

is clear that the bundle lies just back of the endocardial cushion.
Coronal sections of the heart giving the position of the bundle are
shown in figs. 27 (46, 21 mm.), 29 (409, 16 mm.) and 30 (423,
15.2 mm.). In these figures it is clear that the bundle passes
below the medial cusp of the tricuspid valve and between it and
the interventricular septum. The extreme dorsal connection is
shown in fig. 30. In earlier stages while the interventricular fora-
men is still present, it is easily seen that the bundle forms its wall
as indicated in figs. 12 (175, 13mm.) and, 14 (424, 17.2 mm.).

Fig. 20 Sagittal section. Embryo 15.5 mm. long (No. 390). x 18.
Fig. 21 Sagittal section. Embryo 18 mm. (No. 4382). 40.

These sections show definitely that the bundle is present in the
youngest hearts after the atrial wall has begun to break down, and
that its course is exactly the same as it is in the adult. After the
embryo is over 20 mm. long the bundle is easily seen, as has been
demonstrated by other observers. Figs. 15 to 18 gives a summary
of what I have stated.

The additional strips above connecting atria with ventricles
in various stages in embryos less than 20 mm. long may be of sig-
nificance in view of Romberg’s and Kent’s observations. Rom-
berg®* found muscular connections between the anterior and pos-

33 Krehl and Romberg, Ueber die Bedeutung des Heramuskels, ete., Arbeiten aus
der med. Klinik zu Leipzig, 1893, p. 72; and His, Jr., Die Thitigkeit des embryo-
nalen Herzens, Ibid., p. 25.

DEVELOPMENT OF THE HUMAN HEART Dap

terior cusps of the tricuspid valves. Kent*! found such bundles
in the left wall of the heart as well as in the right wall in young
rats and young rabbits. It is possible that some of these strips

Fig. 22 Sagittal section. Embryo 18.5 mm. long (No. 431). X 40.
Fig. 23 Sagittal section. Embryo 20 mm. long (No. 368). 40.

may be constant or they may be variations. At any rate their
presence has not been established as has been the atrio-ventricular
muscle of His.

34 Kent, Researches on the development of the pee heart. Journal of
Physiology, vol. 14, 1893.

286 FRANKLIN P. MALL

So far I have shown that there is a band of tissue apparently
muscular, which connects the sinus, or atria, with the ventricle.
At first it is broad but it gradually becomes constricted. In an
embryo 12.5 mm. long (317) the muscle fibers of the atria stream
towards the connecting muscle as two horns, somewhat later it
arises from a thickened mass in the neighborhood of the sinus and
finally from the mass which has become converted into a nodule.
About this time according to His, Jr., the nerves begin to invade
this region. However, I have been able to trace the ganglion
cells through the septum to the valves in but a single specimen 34
mm. long (No. 249) which shows that the nerve fibers really do
enter the bundle on its atrial side. ‘To what extent they penetrate
the ventricular portion is still undecided.

It is not easy to determine with certainty the destruction of the
lateral and anterior walls of the atrial canal, and my statements
rest upon repeated studies of this region in all of my embryos.
Not only does the endocardial thickening play a rdéle in this proc-
ess but the connective tissue of the epicardium also plays a part,
as His, Sr., has shown. To what extent the outer connective tis-
- sue plays a part is well shown in the reconstruction of an embryo
11 mm. long (853). The model shows that it forms a collar
encircling entirely the heart between the atria and ventricles and
also extends into the anterior and posterior longitudinal sulci.
On the two lateral sides the encircling connective tissue is drawn
into the valves as they become larger and this process completes
the separation of the atrial and ventricular musculature. Be-
tween the atria and the aorta it forms a large plug which soon
comes into apposition and blends with the anterior endocardial
cushion. Behind it encircles the atrio-ventricular bundle so that
' this bundle becomes lodged between the posterior cushion and the
outer connective tissue. The relation here found corresponds
with the position of the bundle in the adult, for it then lies behind
the medial cusp of the tricuspid valve which arises from the endo-
cardial cushion, and the right annulus fibrosus, which arises from
the outer connective tissue.

That the bundle differs in structure from the rest of the muscle
of the heart of the foetus has been shown by Ménckeberg, who

DEVELOPMENT OF THE HUMAN HEART

ae es,

ae

Pig. 24 Transverse section of the heart of an embryo 11 mm. lonen(Non 353) < 33.
Mig. 25 Enlarged drawing of the square shown in Fig. 24 including the atrio-ventricular muscle.
x 360.

288 FRANKLIN P. MALL

Fig. 26 Transverse section. Embryo 16 mm. (No. 317). xX 40.

Fig. 27 Coronal section through the heart of an embryo 21 mm. long (No. 460).
A square is around the atrio-ventricular muscle.

289

DEVELOPMENT OF THE HUMAN HEART

Enlarged drawing of the atrio-ventricular muscle shown in Fig. 27.

Fig. 28
x 360.

290 FRANKLIN P. MALL

deseribes and pictures it in a specimen of the fifth month. Fahr
states that the bundle is less pronounced in the foetus than in the
adult but that its position is the same. Tandler found it in an
embryo 20 mm. long as ‘‘distinguishable even under low powers
by its staining properties.”’ The nuclei are dark and the cell
border stains faintly with eosin. At 20.5 mm. the cells of the
bundle have become larger both in the atrium and the ventricle.

In suitable sagittal sections of the heart the atrio-ventricular
bundle may be recognized in an embryo 8 mm. long (fig. 9) as
a strand of tissue extending from the region of the sinus to the
ventricle. This tissue is somewhat separated from the rest by
delicate spaces and its cells stain somewhat more intensely than
the rest. In embryo pigs 7 mm. long the bundle appears more
like epithelium with a tendency towards vesicle formation in the
immediate region of the sinus. In larger foetuses of the sheep the
epithelial nature is most pronounced. In an embryo 11 mm.
long (figs. 24 and 25) the numerous small spaces around the bundle
encircle it from the region of the sinus just behind the posterior
endocardial cushion, along its course behind the medial cusp of
the tricuspid valve to the upper border of the ventricular septum.
This same arrangement is found in another embryo (12.5 mm. long,
fig. 6) cut in the same plane and stained in the same way. In
this specimen the muscle fibrils are not present in the bundle but
they are found in the muscle cells of the atria and the ventricles.
In a sagittal section of an embryo 14 mm. long (fig. 13) the bundle
is recognizable throughout its extent by its surrounding spaces
and not by its staining properties. In 175 (fig. 12) the bundle
‘ap. be followed into the two ventricles. The same structure and
arrangement 1s made out in embryos 15.2 mm. (423), 15.5 mm.
(390), 16 mm. (409, fig. 29) and 17.2 mm. (424, fig. 14). In
older specimens the interventricular foramen is Just closed and
the bundle reaches to it from the region of the sinus. Two limbs
are easily outlined one of which reaches to the moderator band of
the right ventricle, which is pronounced in the human embryo.

The atrio-ventricular muscle is well lodged between the valve
and the annulus fibrosus in an embryo 18 mm. long (482, fig. 21);
it again extends into the moderator band. The same is observed

DEVELOPMENT OF THE HUMAN HEART 291

in another specimen of the same size and prepared in the same way
(No. 43, 18.5 mm. long). At this stage the nerves have reached
the region of the sinus as is seen in specimen (460, 21 mm. long).
In this specimen the bundle ean be followed by the spaces around
it, figs. 27 and 28. Another specimen (368, 20 mm.), which is
unusually well preserved to show the fibrillae of the heart muscle,
shows that the degree of development of the atrio-ventricular
bundle is the same as that of the rest of the musculature of the
heart.

Fig. 29 Coronal section to show the position of the atrio-ventricular muscle.
Embryo 16 mm. long (No. 409). x 18.

Fig. 30 Coronal section, well dorsalwards, to show the extension of the atrio-
ventricular muscle to the coronary sinus. Embryo 15.2 mm. long (No. 423).
<5:

The spaces around the bundle in embryos from 10 to 20 mm.
long appear to continue in the adult; possibly they form the bursa
of the bundle described by Curran** and the spaces which encircle
the Purkinje fibers. Brilliant injections of these have been made
by Lhamon.*

In an embryo 34 mm. long, No. 249, the atrio-ventricular bun-
dle is chiefly muscular as it passes to the ventricle. Near the sinus

35 Curran, A constant bursa in relation with the bundle of His. Anat. Rec.,
vol. 3, 1909, -

35 Lhamon, The sheath of the sino-ventricular bundle. .Amer. Jour. Anat.,
vol. 13, 1912.

292 FRANKLIN P. MALL

it is invaded by nerve cells. From now on there is a change in the
structure which is well shown in two foetuses 50 mm. long (Nos.
84 and 96). In both the cells are more ‘epithelial’ in appearance
and in the former this differentiated band can be followed into
both of its divisions, one of which reaches to the moderator band.
The limb of the bundle which passes to the left ventricle is most
sharply defined. In a foetus 80 mm. long (No. 172) this bundle
is easily made out and appears much as it does in other foetuses,
or as it is in the adult.

It is thus seen that the atrio-ventricular bundle is present in an
embryo 8 mm. long and that as soon as the muscle of the rest of
the atrial canal is broken down it becomes outlined by encircling
spaces which are now formed. Later it is composed of muscle
which resembles much the rest of the heart musculature. When
the nerves invade the septum of the atria the atrio-ventricular
muscle showsmarked histological changes which remind one much
of the Purkinje fibers. These I have followed into both ventricles
to the moderator band on the right side in the human heart, and
through it in the pig. In the foetal sheep up to 15 em. long much
the same conditions prevail and by using a great variety of stains
I was able to follow the atrio-ventricular bundle beyond the mod-
erator band; no Purkinje fibers were found. As we now know the
continuity of the bundle with the Purkinje system it is not remark-
able that in early stages the structure of the main bundle simulates
that of this system. In the adult heart Purkinje fibers are found
throughout the ventricular portion of the atrio-ventricular bundle
and this is all in harmony with what Ihavefound. I cannot leave
the subject without expressing the suspicion that the differentia-
tion of the Purkinje system is in some way due to the influence of
nerves when they appear in the wall of the sinus.

D. MUSCULATURE OF THE LEFT VENTRICLE

It is practically impossible to unravel the musculature of ven-
tricles from serial sections alone. However, it is possible to gain
quite a clear picture of the muscle bundles by direct observation
of the whole heart with the aid of the dissecting microscope.
About two dozen hearts were removed from human embryos

DEVELOPMENT OF THE HUMAN HEART 293

ranging from 10 to 40 mm. in length and these were first carefully
studied with various enlarging lenses including the binocular. It
was found that the hearts from embryos over 15 mm. long could
be dissected under the binocular to great advantage, especially
after they had been stained in alum cochineal. So all of the
hearts were stained and preserved in alcohol. No dissection was
made hastily and during the preparation numerous sketches made.
In the course of time after many attempts had been made I finally
satisfied myself regarding the various stages of the development
of the chief muscle bundles of the left ventricle.

The following specimens were dissected; the smallest embryo
was an unusually good one from a tubal pregnancy. The heart

Number and length of embryos whose hearts were removed entire, studied and dissected

NO. LENGTH NO. LENGTH
426 10 293 | 19
12 263 | 23
360 14 | | 25
434 15 | 118 25
90¢ 17 | Davis Dif
17, | 33 27
90d 18 | 90a 30
90b 18 227 40

283b 19 |

was removed and examined many times in various fluids by trans-
mitted and reflected light. The best drawings were obtained
from the stained heart while it was in 60 per cent alcohol. These
are shown as diagrams in figs. 31, 32 and 33. It is noticed at once
that most of the striations on the surface of the heart are circular,
which is probably the arrangement of all the fibers in younger
specimens. However, at the apex certain changes have taken
place which fully account for the arrangement of the muscle
bundles in the adult heart. In front of the heart, at the bulbo-
ventricular furrow, the fibers leave the bulb and penetrate the
medial side of the left ventricle, as indicated in fig. 31. Behind
the fibers leave the surface of the left ventricle and enter the

294 FRANKLIN P. MALL

inferior septum which is just beginning to form from below and
behind (fig. 33). Thus it is seen that the vortex of the apex is
laid down in the heart of an embryo 10 mm. long (fig. 32).

In a set of perfect serial sections of the heart of an embryo

32
Fig. 31 Outline drawing of the front of the heart of an embryo 10 mm. long (No-
426). The course of the fibers is indicated.
Fig. 32 The same heart, viewed from below.
Fig. 33 The same heart, viewed from behind.

11 mm. long (353) no such interlacing of the fibers can be seen,
excepting well up near the base of the heart where the fibers from
the right side cross over the anterior longitudinal furrow, while

DEVELOPMENT OF THE HUMAN HEART 295

in the posterior longitudinal furrow the fibers from the left side
enter the septum. ‘There is every indication from the study of
older specimens that the ‘vortex’ formation is at first well up in
the middle of the heart and that only later it is pushed down to
the apex of the left ventricle.87 I was unable to determine this
point with certainty from my sections. However, it is clear that
in an embryo 3.9 mm. (463) long (fig. 1) the posterior bundle,
that is the bulbo-spiral, enters the septum and forms its upper
border just below the inter-ventricular foramen. Below this the
sino-spiral should pass into the left ventricle in front, but sections
do not show that the cells in this region of the heart are passing in
any special direction. However, it is seen that the circular muscle
of the left ventricle appears at the base of the heart and not at its
apex. In this embryo the interventricular foramen is 0.15 mm.
in diameter; in one 8 mm. long (113) it is 0.8 mm. in diameter and
in one 11 mm. long (353) it is 0.4 mm. in diameter. This indicates
that the ventricles grow longer by a down growth of their walls
and not as an upgrowth of the inferior septum.?® With this
in view it is clear that the bulbo-spiral should be located near the
base of the heart at the time of its earliest appearance.

In an embryo 12 mm. long the superficial fibers on the dorsal
side of the heart, that is the bulbo-spiral, could be followed into
the ventricular septum and the sino-spiral around the heart into
the left ventricle. The structures showing the direction of the
muscle fibers could only be seen after the whole heart had been
stained with cochineal and methylene blue. In a specimen 14
mm. long (3860) the heart was dissected from behind under the
binocular microscope and the sino-spiral bundle was followed.
These fibers passed on both sides of the ventricular septum in
front to the septum aorto pulmonale, that is they formed the longi-
tudinal bundles of the right ventricle; the spreading out of the sino-
spiral in the left ventricle as well as the interpapillary muscle

37 Mall, Bifid apex in the human heart. Anat. Rec., vol 6, 1912.

38 Flack (Further advances in physiology, New York, 1909) says that the ventric-
ular septum grows downward as the ventricles expand and unite. The atrio-ven-
tricular bundle is thus associated from the first with the oldest part of the septum,
that is, its upper border which encircles the interventricular foramen.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 3

296 FRANKLIN P. MALL

bundles were also seen. A large band of fibers crossing the right
ventricle was incorporated with this system; no doubt it is the
moderator band, which is very pronounced in the embryo.

The two main bundles could be followed with precision in
the heart of an embryo 17 mm. long, the bulbo-spiral passing into
the upper part of the septum and the sino-spiral to the anterior
wall of the left ventricle.

By the time the embryo reaches 25 mm. in length it is easy to
demonstrate that the strands of muscle which were observed in
younger specimens are destined to become the main muscle bun-
dles of the heart in the adult. First of all there are fibers on the
dorsal side of the heart which cross diagonally the posterior longi-
tudinal suleus to reach the apex of the right ventricle. From
here they pass to the left ventricle and pierce it near the apex
anteriorly to form the anterior horn of the vortex.3? From this
horn fibers pass on the right side of the ventricular septum to the
septum aorto pulmonale, that is they are the longitudinal fibers
of the right ventricle. Under the sino-spiral band loops of fibers ~
are seen which encircle the left ventricle and end in the posterior
‘triangular field. All of these loops enter the septum from behind
and form the bulbo-spiral fibers. This arrangement was clearly
demonstrated in a heart from an embryo 17 mm. long, as well as
in specimens 25 and 27 mm. long (figs. 34 to 37). _In one of these
(No. 33) the sino-spiral turns upward at the apex of the right
ventricle along the lower part of the anterior longitudinal sulcus
and then enters the left ventricle. At this point the fibers blend
with the trabeculae of the right ventricle and then give rise to
the longitudinal fibers of the septum of the right ventricle. The
bulbo-spiral bundles form loops reaching towards the apex ofthe
left ventricle, that is, they are only an extension of the posterior
triangular field. This, as has been shown in the previous study,
remains in the adult as the circular fiber of the left venous ostium.
At the extreme posterior border of the left ventricle there is a
marked raphé to correspond with the base of the posterior papil-
lary muscle much as in the adult pig’s heart. The sino-spiral

39 Mall, Amer. Jour. Anat., vol. 11, 1911.

DEVELOPMENT OF THE HUMAN HEART 297

loops now form the anterior horn of the vortex and the lower end
of the loops form the posterior horn.?°

To sum up: The simple heart tube is made up of circular fibers
which in a general way remain circular throughout development.
After the heart is well kinked upon itself the fiber bundles around

Fig. 34 Course of the fibersin the heart of an embryo 17 mm. long. The view is
from behind.

Fig. 35 Thesame as fig.34 from the heart of an embryo 25 mm.long. The fibers
which pass to the base on the right ventricle are present.

Fig. 36 The same as fig. 35, from an embryo 27 mm. long.

Fig. 37 The same as fig. 35, from another embryo 27 mm. long (No. 33).

the venous and arterial ends remain circular but there are marked
deviations at the apex which correspond with the kinking at this
point. Those fibers that arise from the bulbo-ventricular groove
in front encircle the left ventricle and enter the ventricular sep-

40 His, Anat., mensch. Embryonen, Bd. 3, S. 177, makes a statement which is
just the opposite of mine. -

298 FRANKLIN P. MALL

tum behind quite high up so that they form its border just below
the interventricular foramen. Those that arise behind encircle
the right ventricle and enter the heart in front of through the an-
terior longitudinal suleus. Nowas the heart grows by an extension
of the ventricles downward these two encircling bands also grow
and ultimately become interlocked at the apex of the left ventricle,
the fibers from the bulb forming the bulbo-spiral band and the
posterior horn of the vortex; the fibers from the sinus side, the
sino-spiral band form the anterior horn of the vortex. As the
sino-spiral band becomes larger and larger it is shifted over the
greater part of the bulbo-spiral band on the posterior side of the
heart. On the anterior side of the heart with the growth of the
anterior longitudinal sulcus, the sino-spiral band gradually extends
sending bundles to both sides of the septum, that is it constantly
straddles the septum from its lower and under side. By this
extension it binds together the inner walls, especially the papil-
lary muscles, of the two ventricles with each other. These are
the two strands shown in figs. 35 to 37.

The bulbo-spiral band first extends from the bulb over the left
- ventricle, and with its growth penetrates its wall, ultimately
including its apex to end in the posterior horn of the vortex (figs.
31to33). Insodoingit produces loops which constantly turn upon
themselves to return to the base so that when viewed together
the bulbo-spiral band appears as a series of telescoped loops of
sheet fibers as is now well known. At the base of the left ven-
tricle the fibers are more circular forming a thick fleshy mass on
its dorsal side, the triangular field, which in the adult still marks
the embryonic conditions; all loops of the bulbo-spiral band are
but an extension of this field as is shown by the development of
the fibers as well as by the architecture of this muscle in the adult.

THE DEVELOPMENT OF THE HUMAN PROSTATE
GLAND WITH REFERENCE TO THE DEVELOP-
MENT OF OTHER STRUCTURES AT THE NECK
OF THE URINARY BLADDER

OSWALD 8S. LOWSLEY

From the Anatomical Laboratory, Johns Hopkins University

ELEVEN FIGURES (THREE COLOR PLATES)

A review of the literature on the embryology of the prostate
gland and other structures at the neck of the human bladder
discloses a great diversity of findings. Much has been written
about the middle lobe of the prostate and there seem to be two
views very firmly held with regard to its development, one best
expressed by Griffiths and utilized by Tandler and Zuckerkandl!
to the effect that the middle lobe is an independent structure
which may sometimes be lacking; the other supported by Pallin,
Jores and others who believe that the middle lobe is always formed
by ingrowths from the lateral lobes.

Griffiths? concludes from his studies: (1) That the middle :obe
may be either present or absent at the time of puberty and in
adult life before enlargement takes place. . (2) That this lobe is
independent, having glands of its own which open on parts of the
hinder wall of the prostatic urethra. (3) That this region devel-
ops separately from the part of the urethra just mentioned in the
same way as the lateral lobes do from the part of the urethra on
each side of the verum montanum, and it is not the result of an
extension back of gland tissue from the lateral lobes into the inter-
val between the vasa deferentia beneath the neck of the bladder.

! Tandler and Zuckerhandl, Folia Urologica, Bd. 5, 1911, p. 587, in their discussion
of prostatic hypertrophy state that the middle lobe is anatomically constant and
believe that it is morphologically and embryologically independent.

> Griffiths, Jour of Anatomy and Physiology, vol. 23, p. 374, 1889.

299

300 OSWALD S. LOWSLEY

Griffiths’ states in another paper that there can be no enlargement
of the third lobe unless there were gland tubules originally there.
Any one of the lobes may enlarge without involvement of the others.
This author states that enlargement does not take place in that
part of the prostate behind the urethra and anterior to the verum
montanum (posterior lobe).

Mansell Moullin‘ states that the prostate is not an urinary
organ but that the point of origin of the prostatic glands has
simply been displaced in the course of racial development from
the lining of the Wolffian ducts to that structure into which they
open. The question of the median lobe, according to this author,
depends upon the extent to which this displacement has occurred
in each individual. So long as the glands are restricted to the
prostatic sinus there is no medium portion. In some instances a
greater or smaller number are displaced towards the bladder and
they not infrequently occupy the middle line. Usually they remain
on the posterior wall of the urethra and form a more or less con-
spicuous median lobe. Exceptionally they make their appearance
upon the anterior wall as well.

Keibel does not agree with this opinion and considers the pros-
tate to be an urinary organ because its glands arise from the ure-
thra above the openings of the Wolffian and Miillerian ducts.

Evatt® in a study of a 12 em. (crown-heel measurement) foetus
which he considered to be three and one-half months of age, by
means of a wax model described the middle lobe to be made up of
branches of the two largest prostatic ducts which come together
in the midline behind and these, with two smaller ducts immedi-
ately above them, form a centrally placed lobe above the level of
the point of entrance of the genital cord. This he considers to
be the middle lobe and consists of ducts derived from both sides
of the prostate and cannot therefore be regarded as an azygos
structure.

3 Griffiths, Jour. of Anatomy and Phy iology, London, vol. 24, p. 236, 1890.
4 Mansell Moullin, Jour. of Anatomy and Physiology, vol. 29, 1895.

> Keibel, Archiv fiir Anatomic und Physiologie, 1896.

6 Evatt, Jour. of Anatomy and Physiology, vol. 43, p. 314, 1908.

THE HUMAN PROSTATE GLAND 301

Gustaf Pallin? summarizes the results of his study of human
embryos aged three and four months by means of wax models as
follows:

The prostate gland is deposited in male embryos in the third month
by separation of the solid longitudinal folds on the outer side of the epi-
thelial wall of the urethra. Three groups of prostatic tissue are dis-
tinguished, (1) Cranialwards from the genital cords lying dorsally, (2)
Caudalwards from the genital cords lying dorsally,.(3) Ventral.

Both of the first groups go out from the prostatic furrows. From the
cranial the main mass of the base of the prostate will be composed; the
third lobe seems not to be composed of independent glands but ramifi-
cations of the cranial glands can grow into the midline and then these
become gland parts. The caudal dorsal structure forms the lateral and
hind part -of the side lobes. The ventral group at first occupies the
greater part of the forward urethral wall. The number of its glands
becomes reduced in the fourth month and it appears then in the midline
as a forward lobe. In certain cases the reduction of this lobe amounts
to complete atrophy.

This article is very extensively quoted and accompanied as it
is by drawings from wax models and very accurate descriptions
has influenced many workers.

Jores’ states that the middle lobe can be considered only as a
glandular commissure connecting the two lateral lobes and not as
an independent structure.

The posterior lobe is generally referred to in the literature as
originating from ingrowths from the lateral lobes.

The number of tubules of the various parts of the prostate
emptying their secretions into the prostatic urethra is usually
stated in text-books to be between twenty and thirty, while the
proportion of glandular tissue compared to the interglandular
stroma is quoted by different authors to be from one-third or one-
half (Kasuyoshi Nakasima) to five-sixths (Walker). The later
writers seem to be agreed that the prostate begins to develop at
about the third month of intra-uterine life. K6lliker thought
that it was not present until the fourth month and Mihalkovies
placed the fifth month as its‘beginning. The smooth muscle of
the gland, according to the latter and also Tourneux, began to

7 Gustaf Pallin, Archiv fiir Anatomie und Physiologie, 1901.
8 Jores, Virchow’s Archiv fiir pathologische Anatomie, Bd. 135, 1894, p. 224.

302 OSWALD S. LOWSLEY

develop at the middle of the fifth month. Pallin, on the other
hand, found the musculature to be developing at the fourth
month.

Albarran® describes a group of glands under the neck of the
bladder which open into the urethra and whose tubules lie between
the mucosa and the musculature. He calls this the subcervical
glandular group and states that it varies greatly in different imdi-
viduals and may be entirely lacking in some.

The bladder, trigonum vesicae, and internal sphincter have
received a great deal of attention by writers on the anatomy of
this region.

J. Griffiths!® believed the trigonum vesicae to be composed
only of the innermost bands of muscular bundles of the bladder
wall, while the outermost longitudinal bundles pass on to the
neck of the bladder.

W. Waldeyer" called attention to the following facts in regard
to the trigonum vesicae: (1) There is a separate development of
its musculature which is continuous with the musculature of the
ureters and the prostatic urethra. (2) There is an absence of a
submucosa over the trigone. (3) It has a smooth, firm, thick
layered mucous membrane.

Versari” concludes from his studies that normally the muscula-
ture of the trigonum vesicae is made up of (a) the trigonal portion
of the internal sphincter, (b) part of the muscular layer of the ure-
ters, and (ce) the muscle bundles of their sheaths. In adults there
are present in the trigonal region bundles which come from the
muscular layer of the bladder.

Walker!® agrees with the above findings in part. He observes
that from the ureter on each side a thick band of muscle passes
down towards the urethra. These bands converge and unite
so that this longitudinal muscle flows over the margin of the ure-
thral opening in a continuous sheet. In the center of the triangle

9 Albarran, Maladies de la prostate, p. 526, 1902.

10 J. Griffiths, Jour. of Anatomy and Physiology, 1891, p. 535.

11 W. Waldeyer, Das Trigonum Vesicae, Sitzungsberichte der akadamie der Wis-
senschaften in Berlin, 1897, p. 732.

12 Versari, Ric. d. Lab. d. Roma, vol. 18, 1907.

13 Walker, Jour. of Anatomy and Physiology, vol. 40, p. 190, 1906.

THE HUMAN PROSTATE GLAND 303

formed by these bands of muscle the fibers appear to interlace
indiscriminately.

Delbet declares the trigonum vesicae to be an appendage of
the urethral walls. Congenital absence of a ureter shows the
trigonum to be lacking on that side. Passavant has described a
case in which the trigone was entirely separate from the bladder
wall.

In regard to the internal sphincter of the human bladder Ver-
sar1® concludes from his investigations that (1) The smooth muscle
sphincter of the urinary bladder of man constitutes a structure
by itself, which develops independently of the middle (circular)
layer of the bladder, the circular muscle layer of the urethra and
the musculature of the ureters. (2) The sphincter is made up of an
urethral anda trigonal portion, and it is the urethral portion only
which assumes the form of a ring surrounding the initial part of the
urethra. The first groups of the fibers of the sphincter arranged
in bundles correspond to the anterior arch of the urethral por-
tion; from there immediately follow those of the urethral portion
of the posterior arch, and these last are apparently those of the
trigonal portion. The posterior arch of muscle extends little by
little, with new bundles either upwards to occupy part of the tri-
gonal area or downwards along the posterior wall of the urethra,
so that it comes to have an extent much greater than the anterior.
On the other hand, the older view held by Krause, Hyrtl, Gegen-
bauer and others is that the sphincter is a continuation downward
of the circular musculature of the bladder.

The seminal vesicles begin to develop at about the third month
(MeMurrich).!® Pallin found that the ejaculatory ducts show no
suggestion of smooth muscle at the sixth month. The colliculus
seminalis of man is, according to D. Berry Hart,!7 the analogue of
the hymen and lower one-third of the vagina, the Miillerian ducts
not being represented in the adult human male except by the hyda-

1 Poirier and Charpy: Traite d’anatomie humaine, vol. 5, p. 110.

15 Versari, Ric. d. Lab. d. Roma, vol. 13, 1907.

16 MeMurrich, The development of the human body, Philadelphia, 1902.

17 Hart, A contribution to the morphology of the human urino-genital tract,
Jour. of Anatomy and Physiology, 1901, p. 330.

304° OSWALD S. LOWSLEY

tid of Morgagni and some rudiments near the testes. Primrose!’
believes that the uterus masculinus must be looked upon as the
homologue of the series of structures formed in the female by the
fused portions of the Miillerian ducts. Minot! states that in
the male the Miillerian ducts remain rudimentary and their middle
‘portions usually abort leaving the upper fimbriated ends to
develop into so-called hydatids of Morgagni, and the lower or
caudal ends to unite within the genital cord to form the so-called
uterus masculinus, a rudimentary representation of the female
uterus and vagina.

The prostates used in this investigation were obtained from
Dr. Mall’s collection of human fetuses which were preserved in
alcohol or in formalin (4 per cent). They were imbedded in
paraffine and cut,in series, being stained with haematoxylin and
eosin or Mallory’s stain. The measurements taken are crown-
rump, and the ages of embryos are estimated according tothe
table in Keibel-Mall’s Manual of Embryology.”

Before taking up the discussion of the various specimens it
seems best to state the terminology to be used. The various
parts of the prostate gland will be referred to as follows: (1) The
middle lobe, or that part of the gland which is situated between
the bladder and the ejaculatory ducts under the floor of the ure-
thra (prespermatic and post urethral). (2) Lateral lobes, or
those parts of the gland which arise from the prostatic furrows
and the lateral walls of the urethra and extend laterally and pos-
teriorly from that structure. (3) Posterior lobe, or that part of
the prostate which lies dorsal to the ejaculatory ducts above their
entrance into the urethra and dorsal to the urethra below this
point (post spermatic and post urethral). This is the part of
the prostate which is felt per rectum. (4) Ventral lobe, or that
part of the organ formed by glands arising from the anterior or
ventral wall of the prostatic urethra.

18 Primrose, Jour. of Anatomy and Physiology, 1899, p. 64.
19 Minot, Human embryology, New York, p. 490.
2 Keibel-Mall, Human embryology, Philadelphia, vol. 1, 1910, pp. 180-200.

THE HUMAN PROSTATE GLAND 305

Fetus 5 cm. long:?! (ten weeks)

A study of the bladder and prostatic portion of the urethra
of a human male fetus two and one-half months of age shows
several interesting features. The bladder at the trigonal region
has about the same circumference as it has over the rest of its
area, being at this stage a tubular structure which narrows down
gradually as it approaches its orifice and nowhere is there a sharply
outlined portion which will later become the vesical orifice. The
prostatic portion of this tubular structure is marked only by the
change in its shape.

The organs all seem to be composed of embryonic connective
tissue which is similar throughout. The walls of the bladder are
uniform in size everywhere except in the region of the trigone
which is nearly twice as thick as any other portion. The con-
nective tissue strands can be traced from the ureters out into the
trigone and the latter structure is quite definitely superimposed
upon the bladder wall. The increased thickness of the base of
the vesical wall extends throughout the trigonal region but is
most marked at the beginning or interureteral region.

By following this gradually narrowing tube down there is °
noticed a change in shape so that there is a slight notch formed in
the ventral wall of the urethra and a projection into the lumen
of the posterior wall so that this structure takes on the shape of
a very widely spread inverted V. This marks the beginning of the
verum montanum. Further down there are noticed two little
notches on the floor of the urethra, one on each side of the verum
montanum. ‘The one on the right is more pronounced than that
on the left. The outer or lateral portions of the lumen which
will later become the prostatic furrows point in a horizontal direc-
tion and at this period of development show no tendency to be
directed downward (fig. 1).

In no portion of the prostatic urethra is there any thickening of
tissue or outgrowth of epithelial cells indicating the development
of prostatic gland tissue. There is no specific arrangement of
tissue planes and it is not possible to pick out the exact site of the

21 These measurements are all crown-rump and not crown-heel.

306 OSWALD S. LOWSLEY

internal sphincter. Below the verum montanum the urethra
becomes more or less star-shaped, indicating a collapsed circular
tube.

Fetus 7.5 cm. long: (thirteen weeks)

There is considerable change noted in the appearance of the
bladder at this stage. Throughout its entire area the wall com-
posing the base is thicker than at any other portion of its cireum-
ference, and the nearer one approaches the trigonal region the
greater is the thickness. The musculature of the bladder is dis-

Fig. 1 5 em. human fetus two and three-fourths months. Prostatic portion of
urethra.

tinctly made out as deeply staining tissue composed of circular,
interlacing and longitudinal strands which are easily differentiated
from the connective tissue elements forming the major portion of
the bladder wall. The strands of muscular tissue are much
larger and more abundant at the base or inferior portion than at
any other. In the region just superior to the trigone where the
bas-fond will later develop, the inferior wall is three times as thick
as it is at any other part of the circumference. ‘The mucous mem-
brane is gathered in folds on the inferior interior surface of the
organ throughout its length, while elsewhere it is smooth.

The trigonum vesicae is about five times as thick as the remain-
ing portion of the vesical wall. The muscular strands composing

THE HUMAN PROSTATE GLAND 307

it are much finer in texture than those found elsewhere and many
of them can be traced between the two-ureters and out into other
portions of the trigone.

There is a very sudden narrowing of the vesical walls at the
site of the developing internal sphincter and the lumen of the
bladder changes from an oval to a triangular shape and then into
a rather narrow horizontal slit with a vertical slit connecting with
its anterior wall as shown in fig. 2. Considerably below this it
becomes triangular in shape again.

An examination of the urethra of the thirteen weeks old fetus
from the bladder outwards reveals the fact that two large solid

/Ventral

‘ \ Middle Lobe
i) Tubules

Urethra \: Lateral
Lobe
Middle
Lobe.
Sem

Fig. 2 7.5 em. human fetus. Three months. X 20.

evaginations extend posteriorly from its floor. Outward or cau-
dally from these buds other larger evaginations have developed
forming tubules, in some of which, lumina are present. In others
the lumina are poorly developed, and in still others are solid (fig.
2). These structures, 12 in number, which are without question
developing prostatic tubules, are separated by a considerable
space from the main mass of prostatic tissue and occur directly
on the floor of the urethra in a position that is universally accorded
to the middle lobe, i.e., between the bladder and the entrance of
the ejaculatory ducts, and extend posteriorly to occupy a posi-
tion under the floor of the urethra and between the bladder and
ejaculatory ducts.

308 OSWALD 8. LOWSLEY

The tubules of the lateral lobes arise from the sides of the ure-
thra and from the bottom and in some eases a little to the inside
of the depressions at the sides of the urethra, which are commonly
called prostatic furrows. These structures are in nearly every
case larger than the ones making up the middle lobe. They are
thirty-nine in number, twenty-six of which are arranged definitely
in pairs and three of which are unpaired send branches forward.

In this series no glandular tissue is noted in the region dorsal
to the ejaculatory ducts. At the lower or caudal end of the pros-
tate the tubules become centrally located near the midline and
here we have structures which later grow back dorsally to the
ejaculatory ducts and become the posterior lobe of the prostate.
Beginning evaginations and tubules occur throughout the pros-
tatic urethra on its roof or ventral wall. The general direction of
the growth in this region as well as in the others is bladderwards
with the exception of the outermost of the lateral lobes and the
posterior lobe which send a few branches caudalward. The
tubules of the ventral lobe are twelve in number, eight of which
are paired, the other four being directly in the middle line. One
of the latter tubules is quite long and presents a very definite
lumen as all of the larger ones do; those smaller in size are nearly
all solid epithelial outgrowths.

There are no glandular outgrowths from the urethra below the
apex of the prostate and there are no signs of glandular growth in
the subcervical region where Albarran’s tubules are found in
later stages.

The vasa deferentia appear under the bladder at the entrance
of the ureters as two small tubes surrounded by a thick layer of
developing muscle and connective tissue. As they descend they
approach each other and behind the middle of the trigonum vesi-
cae they are enveloped in the same tissue, being bound very firmly
together, and between them is noticed a very small lumen sur-
rounded by rather delicate but distinct connective tissue layers,
which is taken to be the unobliterated upper portion of the fused
Miillerian ducts. The vasa deferentia and their enveloping
tissue increase enormously in size as they descend, so that under
the internal sphincter this structure is larger than the beginning

THE HUMAN PROSTATE GLAND 309

of the urethra and its surrounding tissue. Under the main por-
tion of the middlelobe thelumina of the vasa deferentia spread out
laterally as shown in fig. 2 and form a lateral branch or bending
which marks the first appearance of the seminal vesicles. Im-
mediately below this point at the beginning of the ejaculatory
ducts, the lumina constrict, and the surrounding tissue is less
abundant. The structures surrounding the ejaculatory ducts
which have to this point remained separate now intermingle with
the muscular walls of the urethra and firmly bind the two struc-
tures together.

At this point the utriculus prostaticus which has been very
small begins to enlarge and finally becomes larger than the two
ejaculatory ducts which decrease 1 in size as they epEroact the ure-
thral lumen.

In the progress of the two ejaculatory ducts and utriculus pros-
taticus toward the lumen of the urethra, the tissues surrounding
the two structures are closely bound together, but the mass of
tissue around the ejaculatory ducts and utriculus prostaticus
maintains its identity and by its further progress pushes the flat-
tened floor of the urethra up into a mound-like projection trans- |
forming its triangular lumen into an inverted semilunar shaped
structure and forming the verum montanum, the main tissue of
which is thus derived from the walls of the ejaculatory ducts.

Fetus 8 cm. long: (thirteen weeks. No. 54 in Dr. Mall’s collection)

This fetus shows exactly the same distribution of elements as
the one just described. The number of tubules is reduced, prob-
ably due to the fact that the sections were three times as thick as
the ones discussed above and this might have obscured some of
them.

The middle lobe region gives rise to seven glands. Twenty-
seven tubules are found in the two prostatic furrows which will
form the lateral lobes and six others are located on the floor of the
urethra caudalward from the entrance of the ejaculatory ducts
which represent the posterior lobe of the prostate. These latter
send branches caudalward or anteriorly as well as cranialward or

310 OSWALD S. LOWSLEY

posteriorly, and none of them extend as far back as the ejaculatory
ducts.

On the ventral wall of the urethra are located thirteen struc-
tures composing the anterior lobe.

The subeervical glands of Albarran were not found.

Fetus 12.5 cm. long: (sixteen weeks)

The bladder of the fetus studied here was contracted so that a
direct comparison with the younger stages which were distended,
is not possible. However, it is very evident that there has been
an enormous development of smooth muscle between the last
specimen described and this one. Throughout the whole vesical
wall the muscle bundles making up the three coats of the bladder
stand out very clearly. The mucous membrane and the connec-
tive tissue underlying it are thrown into thick folds extending into
the lumen of the bladder and almost obliterating it. In this
specimen as in those previously described the trigonum vesicae
is very much thicker than any other part of the bladder wall.
~ Muscle fibers are seen in large number extending from the ure-
teral walls and forming a thick layer which seems to be superim-
posed upon the true vesical wall as described by Poirier and
Charpy.

The folds in the bladder mucosa disappear from the trigonal
region, although they persist in other portions of the vesical lumen,
until the region located at about the middle of the trigonum vesi-
cae is reached where it is smooth and regular. In this same
region the trigonal portion of the vesical wall is no thicker than
any other portion, although lower down just before the commence-
ment of the internal sphincter the trigonal portion is about twice
as thick as any other part of the bladder wall.

Just below the internal sphincter which is quite well developed
at this age there are noted eight very slight finger-like evagina-
tions from the floor of the urethra. These structures extend only
a short distance into the submucous tissue and will form the sub-
cervical group of glands decribed by Albarran.

THE HUMAN PROSTATE GLAND BL |

A very short distance below the lower part of the internal
sphincter there are seen numerous gland tubules which are situated
near the periphery of the greatly thickened urethral wall. These
ends of prostatic tubules are arranged in four groups which are
rather widely separated from one another by the stroma of the
urethral wall. These groups of gland tubules are located one in
each lateral wall, one between the floor of the urethra and ejacu-
latory ducts, and one in the anterior wall. Surrounding each
tubule is noted a slight differentiation of tissue from that com-
posing the wall of the urethra, which in this stage shows equal
thickening on all sides, and there is notas yet any bulging, although
the posterior wall is enlarged as it was in the 13 weeks fetus due
to the présence of the ejaculatory ducts and the utriculus Dees:
taticus with their surrounding tissue layers.

Careful identification of the various groups of tubules mentioned
above and tracing their course from section to section reveals
the fact that even at this early stage these structures have many
branches which communicate with the urethra by means of one
rather small duct. In one case there are three extensively
branching tubules whose ducts join.and empty into the urethra
through a common duct (fig. 3).

In this series of sections the tubules are distinctly divided into
five different groups as follows:

The middle lobe is composed of. ten extensively branching
tubules which are separated from the two lateral lobes by a rather
thick layer of connective tissue. The branches of these tubules
which are situated in the posterior wall of the prostatic urethra
between the ejaculatory ducts and the bladder join and communi-
cate with the urethra by means of ten ducts whose mouths are
situated upon the floor of the urethra a considerable distance
bladderward from the openings of the ejaculatory ducts and are
grouped together in a very characteristic manner, being sur-
rounded and bound together by connective tissue and developing
smooth muscle fibers (fig. 4). The location of these tubules, their
course as graphically shown in fig. 3 by means of a composite
drawing which represents the change in location of the various

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 3

a2 OSWALD S. LOWSLEY

groups of tubules and not the individual branches, and the group-
ing of the duct openings of these tubules on the floor of the urethra
as they communicate with it, demonstrate clearly that this struc-
ture is an independent part of the prostate gland.

Ejaculatory

Duct.

Utriculus
Frostaticus

Fig.3 Composite drawing showing course of tubules of the middle lobe of pros-
tate; 12.5 cm. human fetus. Four months. X 30.

The two lateral lobes are composed of tubules which are larger
in size and branch more extensively than those of the middle
lobe. The units composing these lobes grow posteriorly as well as
laterally and occupy a region to the outside of the middle lobe and
the ejaculatory ducts. In this specimen there are forty-six
tubules composing the lateral lobes, the ducts of which communi-

THE HUMAN PROSTATE GLAND ; 313

cate with the urethra not only at its sides but also in the depres-
sions on each side of the verum montanum commonly referred to
as the prostatic furrows.

In the posterior wall of the urethra outerward from the entrance
of the ejaculatory ducts and widely separated from all of the other
lobes of the prostate, are found four large branching tubules which
form the posterior lobe of the prostate. The branches of these
tubules are not at any place in close touch with the lateral lobe
tubules which extend this low in the urethral wall and there seems
to be a definite layer of connective and muscular tissue forming
around the component parts of this lobe.

Extending along the ventral or anterior wall of the prostatic
urethra are observed the tubules forming the anterior lobe. Four-
teen ducts open into the urethra in this specimen. Most of the
tubules are quite small. Only two are as large and extensively
branching as those of the other lobes already described.

Below the outermost tubule of the posterior lobe the urethra
changes its shape from the inverted semilunar type shown in the
drawings to a stellate shape. Extending from the walls on all
sides are small finger-like evaginations. Some of these are simple
folds in the urethral mucosa but others are developing tubules,
some of which have extended quite deeply into the submucous
tissue and a few have small branches. These are considered to
be the developing urethral glands. While they are quite numer-
ous just below the apex of the prostate, below that they are not
found at all. ;

The seminal vesicles in this series are composed of very tortuous
tubes which have a thick muscular layer surrounding them and
communicate by means of one small duct or opening into each
vasa deferentia under the internal sphincter.

The two ejaculatory ducts bound together as described above
become attached to the wall of the urethra below the point just
mentioned and gradually become more deeply implanted in the
thick posterior urethral wall. The utriculus prostaticus is en-
tirely obliterated above but appears between the ejaculatory
ducts in the middle of the prostatic urethra. These three struc-

314 OSWALD S. LOWSLEY

tures bound together by connective and muscular tissue in a
very characteristic way approach the urethra on a gradual slant
maintaining the globe-like appearance shown in fig. 4, pushing
the floor of the urethra up into its lumen, forming the verum
montanum and causing the urethra to assume an inverted semi-
lunar appearance. The ejaculatory ducts run parallel to the
floor of the urethra for a considerable distance and then ascend
vertically to empty on the sides of the verum montanum. The

Ventral or Anterior Lobe.

Middle Lobe-

7 Lateral
Lobe.

Utriceulus
Prostaticus.

Ejaculatory Ducts.

Fig. 4 12.5 ecm. human fetus. Four months. Showing rather definite separa-
tion of the middle lobe *from the lateral lobes.

utriculus prostaticus in this series is very irregular in shape, being
throughout most of its length in the shape of an inverted Y and
giving the appearance of two tubes which have fused at their
upper pole and not at the lower. Its opening into the urethra
is below those of the ejaculatory ducts and there is a blind end
extending a short distance down in the verum montanum under
the floor of the urethra.

THE HUMAN PROSTATE GLAND 315

Fetus 12 cm: long: (sixteen weeks)

A series of sections cut in the longitudinal direction through the
bladder and prostatic urethra of a fetus four months of age shows
the arrangement of the variousstructuresin avery striking manner.
The smooth muscle of the bladder wall stains very deeply and is
bound in place by the lighter staining connective tissue. The
mucous membrane of the bladder and the submucous tissue is
thrown into folds and finger-like projections into the vesical lumen
except over the trigonum vesicae where it is tightly adherent to
the muscle forming this structure. The trigonum vesicae is made
up of muscular fibers which are finer in texture, more tightly
bound together, and have lighter staining characteristics than the
rather heavy, loose, deeply staining muscle bundles of the bladder
wall proper upon which it is superimposed. The trigonal tissue
extends down through the vesical neck and becomes lost among
the muscle fibers of the prostatic urethra. The sphincter of the
bladder appears as a large oval mass of circular fibers which sur-
round the neck of the bladder and are quite distinct from the sur-
rounding structures. There is a sharp line of differentiation
around its lower part, while above some fibers from the trigone
and vesical wall mingle with the fibers of the sphincter attaching
it intimately to them.

The distribution of the tubules of the prostate gland is similar
to that just described in the cross sections of the bladder and pros-
tatic urethra of a fetus at this age. On the floor of the urethra
just below the vesical sphincter several small evaginations repre-
sent the developing glands of Albarran. The middle lobe of
the prostate is composed of several large tubules with a number of
branches which extend back posterior to the sphincter and ante-
rior to the ejaculatory ducts. The lateral lobes are somewhat
larger than the middle lobe, being composed of numerous tubules
with many large branches. The posterior lobe tubules are seen
extending from the floor of the urethra below or outerward from
the entrance of the ejaculatory ducts to a position behind them.
On the roof or ventral wall of the urethra are found the tubules

a6 OSWALD S. LOWSLEY

which make up the ventral lobe of the prostate. In this specimen
they are not nearly so numerous or so large as in the series of
cross sections just described.

Fetus 16.5 cm. long: (twenty weeks)

The bladder and prostate of a fetus five months of age cutin
cross sections show the musculature of the bladder wall and the
superimposed trigonum vesicae much more clearly than any of
the previous specimens described. A striking thing is the great
thickness of the vesical sphincter whose many layers of circular
fibers form a mass about as thick as the bladder wall at its base.
At its upper and also outer borders the muscle fibers of the bladder
wall intermingle with it, thereby strengthening the vesical neck
to a considerable degree. The vesical orifice in this specimen
seems to be tightly closed. The muscular layers surrounding the
tubules of the prostate gland whose earliest differentiation from
the rest of the wall of the prostatic urethra was noted in the six-
teen weeks fetus, have now become quite thick and the main mass
of fibers are arranged circularly, although there is noted a certain
amount of muscular tissue extending longitudinally as described
by Walker.”

The same distribution of the prostatic gland tubules is noted
in this series of sections that has already been described in detail
in the prostate of a fetus sixteen weeks old.

Fetus 16 cm. long: (twenty weeks)

Sections of the bladder and prostate of a fetus five months old
cut in a longitudinal direction show particularly well the arrange-
ment of the structures at the vesical neck. The muscle bundles
of the bladder wall are well developed and arranged in the charac-
teristic manner already described. The bladder mucosa which
elsewhere is thrown into folds is bound tightly to the surface of
the trigone which is superimposed upon the vesical wall and is
composed of muscle bundles made up of very fine fibers which

22 Walker, Johns Hopkins Hospital Bulletin, vol. 11, p. 246.

THE HUMAN PROSTATE GLAND 317

do not stain so deeply as the muscle fibers elsewhere. The inter-
nal sphincter is sharply marked off and does not seem to be pro-
portionally so large as the one observed in the twenty weeks
fetus cut in cross sections.

Just under the mucosa of the trigonum vesicae is seen a small
collection of simple tubules which are taken to be the subtrigonal
glands (fig. 5). They are few in number and are very delicate
in structure.

Trigonum
Vesicae

Middle lobe Tubvle
z <—Ejaculatory Duct

Fig. 5 16cm.human fetus. Prostate. X 15.

Below the internal sphincter another collection of glands is
seen developing from the floor of the urethra. These have not
as yet grown deeply into the wall of the prostatic urethra. They
are identified as the subcervical glands of Albarran and in this
series are very few in number and very slight as regards their
architecture. -

Extending from the floor of the urethra backward and down-
ward is seen a tubule of the middle lobe of the prostate with num-
erous branches which extends into the region ascribed to the
middle lobe and is posturethral and prespermatic. The ejacula-
tory duct is observed passing up into the verum montanum.
Below its entrance there are several tubules of the posterior lobe

318 OSWALD S. LOWSLEY

with many branches and whose ducts communicate with the floor
of the outer portion of the prostatic urethra. The section which
shows this arrangement of structures is reproduced diagrammatic-
ally in fig. 5. It was cut alittle to one side of the middle line as
shown by the presence of part of the ejaculatory duct and its mus-
cular wall, and shows quite conclusively that tubules of the middle
and posterior lobes arise independently from the urethra and in
this instance could not be construed as extending in towards the
middle line from the lateral lobes.

Other sections of this series show the tubules developing from
the prostatic furrows and sides of the urethra to form the lateral
lobes. Others extending from the ventral or anterior wall make
up the anterior lobe of the organ. In this specimen these latter
are quite numerous and are extensively branched.

Fetus 19 cm. long: (twenty-two weeks)

The bladder of this specimen is relaxed and the lumen almost
filled with the loose folds of mucosa pulled into it by the con-
traction of the muscles of the wall, which are very large in this
case compared to the specimens previously described. The three
muscular layers are clearly made out and rather larger fibers
than those heretofore described are seen passing from the wall of
the entering ureters out upon the vesical wall, forming the trigo-
num vesicae. The upper border of this structure is about twice
as thick as the bladder wall and more compactly arranged. Lower
down. the trigone is less thick and at its middle is not as thick as
the bladder wall proper. The mucosa covering the trigonum vesi-
cae is tightly bound to it so that in relaxation of the bladder the
mucosa of the rest of that structure may be thrown into many
large folds but the trigonal mucosa does not so arrange itself.

Extending parallel with the long axis of the lumen of the blad-
der just outside of the mucosa of the roof or ventral wall of that
structure is seen a body which arises by two small branches and
ends blindly in the vesical wall. It extends from the upper border
to the middle of the trigonum vesicae and at its upper end
the branches become indistinguishable from muscle bundles.

THE HUMAN PROSTATE GLAND . 319

Throughout its lower portion after the two branches join it is
surrounded by a thick layer of circularly arranged fibers enclosing
this other structure, which in some places seems to be a lumen
filled with degenerated epithelial cells and at others a space filled
with bundles of fibers undergoing degeneration. This structure
is about ten times as large as any of the blood vessels seen in the
section and does not bear any resemblance to tissue seen in this
or other series studied.

There are five small tubules seen developing from the floor
of the bladder and extending down into the lower part of the tri-
gone. These are the subtrigonal glands.

The circular fibers composing the sphincter at the vesical neck
are quite thick in this case and are intimately connected with
the muscle bundles of the bladder proper, while the trigonal
fibers although they are very few in number down here are super-
imposed on the fibers composing the sphincter. This arrange-
ment persists until the upper portion of the prostatic urethra is
reached where they become lost in the musculature of the urethral
wall.

Just below the vesical sphincter there are found eleven evagina-
tions from the floor of the urethra which are lined with very fine
cylindrical epithelium and which in no case branch. These
tubules are recognized as the subcervical glands of Albarran and
their direction of growth seems to be upward toward the bladder
and in no instance do any of them extend deeply into the muscu-
lature of the urethra.

In the region usually occupied by the middle lobe tubules of
the developing prostate gland, there is in this series an entire
absence of such tubules (fig..6). Very careful study fails to
reveal any glandular tissue, except Albarran’s tubules described
above, developing from the floor of the urethra between the en-
trance of the ejaculatory ducts and the cervix of the bladder.
Extending from each of the lateral lobes is seen a large branch
from the tubule nearest the middle line, which, if it should con-
tinue to grow and send out additional branches, would ultimately
form a bridge of tissue extending from one lateral lobe to the

320 OSWALD S. LOWSLEY

other and binding them intimately together. This is the condi-
tion which Pallin believes to exist in all prostate glands.

The two lateral lobes are composed of 42 distinct tubules which
have numerous and large branches. These structures all extend
laterally, posteriorly, and towards the bladder, their uppermost
branches being found under the lower end of the vesical sphincter.
The muscular layers around the tubules, the development of which
has already been referred to, are very prominent in this specimen,
particularly where the tubules are beginning to push out laterally
and posteriorly to form the bulging lateral lobes.

Ventral lobe tubules.

Urethra,

Latera|

Lobe.

Lateral

Ejaculatory Duct; lp SR Ejaculatory Duct.

a

Tig. 6 19 cm. human fetus. Five and one-half months.

Below the entrance of the ejaculatory ducts and utriculus pros-
taticus there are observed ten large branching tubules which ex-
tend back behind the structures mentioned above and between
and behind the lateral lobe tubules. These tubules compose the
posterior lobe of the prostate and make up the main mass of its
apex, being distinctly separated from the two lateral lobes by a
considerable area of tissue. The posterior lobe does not extend
very far up behind the ejaculatory ducts in this specimen on
account of the huge size of the utriculus prostaticus which occu-
pies the space into which the tubules of this portion of the gland
usually extend.

THE HUMAN PROSTATE GLAND oom

The anterior lobe consists of seven tubules which are very much
smaller than those making up the other lobes and very few
branches are sent off. There seems to be a shrinking into insig-
nificance of these tubules as far as a comparison with the size of
similar structures in the younger fetuses is concerned.

The seminal vesicles are very tortuous in this series. The
ejaculatory ducts seem to be of normal size but in their course
through the wall of the urethra they are separated by an utriculus
prostaticus, the upper solid end of which is seen in fig. 6. This
structure is a little larger in size than the ejaculatory ducts in
most of the specimens studied, but in this case it is about twenty
times as large an an ejaculatory duct and at least three times as
large as the lumen of the urethra. It is as usual contained within
the same muscular sheath as the ejaculatory ducts as shown in
fig. 4 but here the muscular coat is very much thinned out. The
utricle extends obliquely through the urethral wall accompanied
by the ejaculatory ducts until one part of its circumference is
quite close to the floor of the urethra, where a small part about
the size of a normal utricle separates from the main part of the
organ and runs along under the floor of the urethra for some dis-
tance below the openings of the ejaculatory ducts. Itthen opens
in the midline on the crest of the verum montanum. The larger
part of the lumen runs forward on the same plane that it occupied
above and stops rather sharply in a blind end. There have not
been seen either prostatic or ejaculatory ducts opening into the
utriculus prostaticus of this or any other specimen studied.

The ejaculatory ducts are observed to run obliquely through
the prostate until they approach quite close to the urethra in the
crest of the verum montanum; then they run parallel with the
axis of the urethra for a considerable distance (800u in this case),
ultimately opening on a slant into the prostatic urethra on each
side of the verum montanum, so that any pressure within the ure-
thra would tend to close them very effectually and in a manner
similar to that observed in the ureters in cases of distention of the
bladder.

The verum montanum is formed in this case as already described
by the ingrowth of the utriculus prostaticus and ejaculatory ducts,

Bp OSWALD S. LOWSLEY

Their development under the floor of the urethra pushes it up into
the lumen causing the characteristic semilunar shape of the pros-
tatic urethra. The fact that the lumen of the prostatic urethra
below the entrance of the ejaculatory ducts is also crescentic
in shape is due to the continuance of the muscular fibers that
form the coats surrounding the utricle and ejaculatory ducts
under the mucosa of the urethra where they finally become lost
in the muscle of that structure.

Fetus 27 cm. long: (thirty weeks)

Inspection of the bladder of this specimen cut in cross section
corroborates the findings already described, the only difference
being that the muscle bundles seem to be larger in this thanin
any of those previously observed. The trigone is superimposed
in the characteristic way and its submucous blood vessels, which
in all cases are much more numerous than those of any other por-
tion of the viscus, are greatly increased in number and size in
this case. The sphincter of the bladder is thicker and composed
of heavier muscle fibers than the ones that have been described.
Outside of the sphincter are seen the muscle bundles of the vesical
wall which become blended with and are lost in the circular fibers
forming that structure.

The subtrigonal glands are present in very small numbers, only
four being detected. They extend through the submucous struc-
tures and only a very short distance into the musculature of the
trigonum vesicae. Their architecture is very slight and there
seems to be no particular differentiation of connective tissue or
muscle fibers around them. They resemble the urethral glands of
Littré more than any other tissue and probably are a continuation
upward of those structures.

The subcervical glands are present in this series, being nine in
number. They are all quite small and no branches are noted in
any of them. In addition to these nine tubules on the floor of
the urethra there are six found on its roof or ventral wall. These
tubules are similar in extent and architecture to the subcervical
group and are apparently not a continuation upward of the ven-

THE HUMAN PROSTATE GLAND 320

tral lobe of the prostate as they are widely separated from that
structure and are entirely different in their character from pros-
tatic tubules. All of these gland tubules extend from the urethra
upwards or towards the bladder under the mucosa, and in no case
do they extend deeply into the muscular tissue. They push out
into the submucous tissue between the numerous blood vessels
in this region and are not surrounded by a specially arranged
musculature. The lumina of these tubules are about three times
the size of the blood vessels found here, but are much smaller
than the prostatic gland tubules.

The middle lobe is made up of eleven.tubules, whose blind ends
extend backward and upward under the sphincter of the bladder,
each of which has a larger number of branches and communicates
with the urethra through quite small ducts which open upon its
floor some distance above the openings of the ejaculatory ducts
and utriculus prostaticus. In the specimen studied previously
there has been a wide and distinct separation of the middle lobe
from the two lateral lobes. This distinctiveness is well shown in
the fetus 16 weeks old (fig. 4). In this specimen it is easy to dis-
tinguish the middle lobe from its location, the direction in which
the tubules extend, and the fact that its duets communicate with
the urethra at the same point that ducts of middle lobe tubules
have in younger series. On the other hand, the layer of tissue
separating the middle from the lateral lobes is not very extensive
except at the base. There is no semblance of a special capsule
and in some places the tubules of the middle lobe are side by side
with the tubules of the lateral lobe, but in no place is there found
an intermingling of the branches of one lobe with those of another.
At the base of the prostate, up under the vesical sphincter where
the prostatic tubules have grown farthest away from their point
of origin, the separation between the middle lobe and the two
lateral lobes is quite wide and there is a great deal of connective
and muscular tissue between them. In no ease has it been possi-
ble to find any difference between the architecture of the tubules
of the different lobes with the exception that those of the middle
lobe are usually not quite so large as the lateral lobe tubules.
Finer histological studies of the glandular epithelium have not

324 OSWALD S. LOWSLEY

always been possible, as some of the tissue is quite old and con-
traction of the epithelium has taken place.

The two lateral lobes are made up of thirty-six tubules which
extend posteriorly and up under the internal sphincter as far as
do those of the middle lobe. At the sides of the lower part of the
sphincter there is a distinct bulging due to the further develop-
ment of the lateral lobe tubules, which with the middle lobe form
the base of the gland. This posterior and lateral bulging is noted
throughout the entire prostatic region, although the larger num-
ber of tubules are massed in the base of the gland. There are a
number of tubules whose ducts open into the sides of the urethra
and which send branches up towards the ventral wall. These
structures have many branches but they are much smaller than
either the rest of the lateral or middle lobe tubules and their mus-
culature is very slight compared to that of the other prostatic
tubules. In the apex of the gland the tubules of the lateral lobe
have a few branches which run downward or outward. Every-
where else the general direction of growth is towards the bladder.

The glandular tissue making up the posterior lobe is found at
the point where the ejaculatory ducts extend vertically towards
the urethra. It is between and behind the lateral lobes and this
part of the gland is distinctly separated from the lateral lobes by
a thick layer of connective tissue (fig. 7). This lobe in the series
under consideration is made up of nine branching tubules of large
size whose ducts communicate with the urethra on its floor or
posterior wall outerward from the openings of the ejaculatory
ducts and at no point do any of the branches of these tubules
intermingle with the tubules of the lateral lobes, being in all
places separated from them by a definite layer of connective tissue.
In the apex of the gland some of these tubules send branches
forward or outward but elsewhere they extend in a bladderward
direction.

The comparative decrease in size of the tubules of the anterior
lobe noted in the prostate of the twenty-two weeks old fetus is
also striking in this series. The number of tubules making up the
anterior lobe noted in the fetuses younger than twenty-two weeks
was about twice as great as that noted in the specimens twenty-

THE HUMAN PROSTATE GLAND 325

two weeks and older. In this prostate there are eight small
branching tubules communicating with the anterior urethral wall.
They are surrounded by muscular layers similar to the other pros-
tatic tubules and are quite limited in their extent and widely
separated from the lateral lobes.

Below the apex of the prostate there are noted a large number of
glands, most of which are simple tubules, extending into the sub-

Viricuius Prostaticus
Ventral Lobe

ee
c | \ ’ Urethra
jaculatory SNK
avai Bye S WO Lateral Lobe
NAS. ; y

Lateral Lobe

Avie
De

/ “od Sent
1 | Ute

\
2) hee \\
SAY, ZAIN

Ahi
'

—<

eA I " Gi
Spy MeN SHINN
Ranke
\ )j

\

ALS
ws

Posterior Lobe

Fig. 7 27cm. humanfetus. Seven and one-half months. x 14.

mucous tissue their ducts opening into the urethra. They are
found on all sides of its lumen but most of them open on the ven-
tral wall. They extend barely to the muscular bundles and none
of them are found within the musculature.

The seminal vesicles have grown back behind the base of the
bladder to a point opposite the middle of the trigonum vesicae.
They are made up of thick walled tortuous connected tubes which

326 OSWALD 8. LOWSLEY

communicate with the vasa deferentia behind the lower part of
the sphincter of the bladder rather deeply in the base of the pros-
tate gland.

The ejaculatory ducts surrounded by their thick muscular wall
and imbedded in the prostate traverse its posterior portion for a
considerable distance in a plane horizontal with the floor of the ure-
thra. <A little below the point where the lowest of the middle lobe
tubules opens out upon the floor of the urethra the ejaculatory
ducts turn sharply and mount almost perpendicularly towards the
urethra and with their thick coats and the utriculus prostaticus
form the verum montanum. By referring to fig. 4 it is seen that
the relative sizes of the urethra, the verum montanum and its
contained structures are much smaller in the older specimen
when compared with the size of the prostate gland. The ejacula-
tory ducts, which are greatly narrowed here, run for a considerable
distance in the verum montanum parallel with the axis of the
urethra and finally open on the sides of the verum montanum.
The course of the ducts in this compressible structure and their
lateral openings which are very small in size, are the features
which make total occlusion of these important structures possible
in case of distention of the prostatic urethra from any cause.

The utriculus prostaticus is about the size of an ejaculatory
duct in this fetus. Its uppermost end begins at the point where
the ejaculatory ducts assume their position in the verum monta-
num and extends outward in‘the apex of that structure opening
in the middle line near the apex of the prostatic urethra. Outer-
ward from the mouths of the ejaculatory ducts the utricle is sur-
rounded by a thick muscular wall which includes a continuation
of the fibers which surrounded the ejaculatory ducts. Below the
mouth of the utricle the muscular tissue continues for a short dis-
tance only.

New-born infant: (35 cm.)

The bladder and prostate of an infant who had died a few days
after birth were cut in cross section and mounted in the usual way,
being stained with haematoxylin and eosin. A very thorough
study has been made of the prostate of this baby, which has in-

THE HUMAN PROSTATE GLAND BH

cluded drawings, diagrams, and the construction of a wax model,
pictures of which are reproduced in this article.

The bladder was distended upon fixation and the natural curve
at its neck mechanically straightened, so it is possible to get a
good idea of the thickness of its walls at various points. At the
place where the ureters join the vesical wall and begin their oblique
course through to its lumen, there is an increase in the size of
the base to twice that observed elsewhere in its circumference.
This increased thickness of the base is maintained throughout the
trigonal region, being most pronounced between the ureters where
the interureteric ridge is made up of fibers extending from one
ureteral wall mingling with those of the other. The trigone in
this as in the previously described cases is made up of muscular
fibers which originate in the walls of the ureters and pass over the
musculature proper of the bladder, becoming lost among the fibers
of the urethra. Just below the interureteric ridge the wall of the
base of the bladder becomes comparatively smaller, being very
little larger than the rest of the wall. As the sphincter is ap-
proached the entire vesical wall increases greatly in size, the base
always being considerably larger than any other part and the
trigonal fibers are always distinguishable from the tissue of the
vesical wall proper. The mucosa over the trigone is much more
richly supplied with blood vessels than any other portion.

The internal sphincter is made up of circular fibers which are
very numerous at its upper part. It is larger at the floor of the
vesical cervix, and while many of the fibers pass entirely around
the orifice of the urethra the majority of them mingle with the
muscle bundles of the bladder on its ventral aspect which approach
much closer to the lumen of the orifice of the urethra here than in
the posterior wall. At the outermost portion of the sphincter
there is a great decrease in the number of fibers composing it and
it is about equal in size at all points of its circumference. In
its ventral portion a number of small longitudinally disposed
fibers are seen. The lower part of the sphincter ends considerably
outerward from the point where posterior lobe tubules have ex-
tended backward behind the two ejaculatory ducts.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 3

328 OSWALD 8S. LOWSLEY

In the mucosa at about the middle of the trigonum vesicae are
found nine very delicately constructed tubules which are recog-
nized as the subtrigonal glands. Most of them are simple tubules
that extend down to the muscle but a few of them have one or
two very small branches and these extend for a short distance
into the musculature of the bladder wall. The blind ends of
these tubules are a little closer to the base of the trigone than
the mouths of their tubules.
~ Commencing at about the middle of the vesical sphincter and
extending down to its lower border there are found on the floor of
the orifice the tubules forming the glands of Albarran. There are
nineteen of these structures in this series, most of which are simple
tubular glands lined with very small columnar epithelium and do
not extend very far beneath the mucosa. In a few cases there are
several branches springing from a tubule and these in many in-
stances extend a short distance into the musculature of the sphine-
ter. None of these tubules are found in the ventral mucosa of the
orifice but there are some very small evaginations in that region
which may later develop into tubules of the same sort. The
- blood vessels in the mucosa of this region are quite numerous and
large. The tubules of Albarran’s group are not surrounded by
any differentiated tissue as are the tubules of the prostate but
seem to be merely imbedded in the submucous structures. They
open for the most part near the middle line on the floor of the ure-
thra but a few open in the angular depressions at the sides of the
urethra which marks the beginning of the prostatic furrows.

The middle lobe tubules have extended up behind the sphincter
to its uppermost border. The end of the tubule that has extended
the highest is situated in the middle line just above the ampulla
of the vas and its branches are surrounded by rather dense mus-
cular tissue. The whole mass lies imbedded in the loose connec-
tive tissue beneath the bladder musculature. Lower down in the
series these branches are reinforced by others of a like nature and
a very short distance below become connected with the muscula-
ture surrounding the urethra. The tissues which envelop the
middle lobe tubules are thicker in this specimen than in any of
the younger ones observed being in places as dense as the walls

THE HUMAN PROSTATE GLAND 329

of the ejaculatory ducts. A few of the most lateralward of the
middle lobe tubules have branches which extend to the sides for
short distances but in no place is there noted an intermingling
with lateral lobe tubules. There are no branches of the lateral
lobes extending into the middle lobe region and the ducts of the
nine large branching tubules which form the latter structure
pour their secretions through mouths which empty upon the ure-
thral floor bladderward from the openings of the ejaculatory ducts.
An interesting thing noted in all of the prostates studied but more
particularly in this one is the fact that an enormous number of
branches join together to form one tubule which empties into the
urethra through a duct which is no larger in diameter than one of
its smallest branches. The ducts of the middle lobe tubules
just before their entrance into the urethra are quite widely sepa-
rated from other parts of the prostate, thereby retaining their
embryological characteristic of an independent origin.

The uppermost ends of the left lateral lobe tubules have ex-
tended back under the bladder and are found contained in their
thick muscular envelope adherent to the sides of the seminal
vesicles. The right lateral lobe tubules are found to have ex-
tended back only to a point above the opening of the seminal
vesicle into the ejaculatory duct. Lower down where the pros-
tate is broadest the branches of the lateral lobe tubules are exceed-
ingly numerous. By their great development they cause the
base of the prostate to bulge laterally and posteriorly. The two
lobes are made up of thirty-four tubules and are separated pos-
teriorly by the ejaculatory ducts, the middle lobe and the posterior
lobe. Anteriorly in this series of sections the branches of the
lateral lobes approach each other very closely, especially in lower
part of the gland at its apex. The outermost or caudalward por-
tion of the lateral lobes, represented by two large gland tubules,
-send branches in a caudalward direction and this fact is of interest
surgically because in nearly every successful enucleation these
forward branches must be cut with curved scissors in order to
free the lobes from the capsule as they seem to be particularly
adherent. The posterior borders of these lobes are separated
from the posterior lobe by a rather dense layer of connective tissue

330 OSWALD S. LOWSLEY

which also separates it from the middle lobe and ejaculatory ducts
(fig. 8). An increase in the amount of connective tissue at this
point would be of great interest surgically in cases of perineal
prostatectomy, because the incision must be made entirely through
it, otherwise the enucleation procedure would lead the instrument
to the capsule instead of to the inside of that structure and hence
could not be completed.

The posterior lobe is made up of eleven glands, some of which
have extended toward the bladder until they are almost as far

Ventral

Urethra

Lateral Lobe

Lateral Lobe £:""

Ejaculotory .

Duct —\

Posterior Lobe

Fig. 8 36cm. (new-born) baby prostate < 6, camera lucida.

back as the ends of the middle lobe tubules. Their course follows
rather closely the dorsal aspect of the ejaculatory ducts until the
latter structures ascend almost vertically towards the urethra,
immediately caudalward to which there is a small area free from
glandular elements. The ducts of posterior lobe tubules enter the
floor of the urethra caudalwards from the entrance of the ejacu-
latory ducts. The posterior lobe is the part of the prostate that
is palpated per rectum and presents in the middle of its posterior

THE HUMAN PROSTATE GLAND 331

surface a slight depression which is ordinarily termed the median
furrow. This depression in the specimen under discussion is
more pronounced in the region of the apex and gradually becomes
shallow at the middle of the gland and at the base assumes a
rounded contour. At the apex two of the posterior lobe tubules
send branches forward which appear outside of the muscular
walls of the urethra, which at this stage are quite well developed.

The ventral lobe is composed of two very small tubules with

_ Just a few branches whose ducts open upon the ventral wall of the

prostatic urethra at about its middle part. This lobe in the new-
born has atrophied to, almost complete insignificance (fig. 10).

Just at the apex of the prostate and a little below that point
there are noticed numerous gland tubules contained in the sub-
mucous tissue of the urethra which are easily differentiated from
prostatic tubules, being much smaller in size and lacking the mus-
cular coats of the latter. In most instances these tubules have
several branches, but none of them are at all extensive as are the
tubules of the prostate. The entire structures are contained
within the muscular walls of the urethra and their ducts open into
it on all sides. These tubules are considered to be the glands of
Littré, and while very numerous just below the apex of the pros-
tate are very few in number lower down in the urethra.

In the new-born the seminal vesicles have extended back
under the bladder almost to the base of the trigone. Its upper-
most portion consists of five lumina on each side with walls almost
as thick as those of the vasa deferentia. These lumina all com-
municate lower down and connect with the ejaculatory ducts
just below the point where they become imbedded in the muscula-
ture of the prostate gland. The ampullae of the vasa deferentia
are easily distinguishable in this specimen and are marked by a
considerable widening and great increase in the size of the lumen.

As the ejaculatory ducts become more deeply imbedded in the
prostate they become situated in more immediate contact with
one another until their musculature becomes intermingled as
shown in fig. 8, and while each contains its own walls intact they
are both contained within the same bundle. In their progress
through the prostate they take the course already described in

332 OSWALD 8. LOWSLEY

previous specimens and which is shown in fig. 10. In this as
in the other specimens studied the verum montanum is observed
to be composed of the ejaculatory ducts, utriculus prostaticus
and their walls. The ejaculatory ducts run for a shorter dis-
tance in the verum montanum in this than in the prostates previ-
ously discussed, finally opening on the lateral walls of that struc-
ture in the characteristic way already described.

The utriculus prostaticus is found in the tip of the verum mon-
tanum extending only a short distance before it opens into the .
urethra in the midline, and in this case has a very large wide open-
ing above the mouths of the ejaculatory ducts which is not the
usual arrangement, as in all of the other specimens studied its
mouth has been below those of the ejaculatory ducts. There
have not been found in this or in any of the other prostates studied
either an ejaculatory duct or a prostatic tubule opening into the
utriculus prostaticus.

DISCUSSION

1. The bladder of a fetus ten weeks old consists of a cylindrical
tube composed of embryonic connective tissue. Near its base
it is jommed by the two ureters which pass through its wall and
fibers from which are superimposed upon the bladder wall to form
the trigonum vesicae. There is no tissue resembling muscle
present and the future site of the vesical sphincter is not dis-
tinguishable except by the change in the size and shape of the
lumen.

At the thirteenth week the muscular development has begun
and is observed as circular interlacing and longitudinal strands of
tissue that take on a deeper pink color than does the connective
tissue which still forms the major portion of the bladder wall.
The entire base of the bladder is thicker than any other portion
of its circumference while the superimposed trigonum vesicae
and the wall under it are five times as thick as the anterior wall.
The muscle fibers observed in the trigonum are traced between
the two ureters and down on the rest of the trigone. They are
finer in texture than the muscle fibers in the bladder wall. The
site of the internal sphincter is marked by a sudden narrowing in

THE HUMAN PROSTATE GLAND 333

‘ the size of the lumen and a few circularly arranged muscular fibers
are made out.

By the sixteenth week there has been a very great develop-
ment of the musculature of the bladder, so that the fibers are
sharply defined from connective tissue. The muscle fibers form-
ing the trigonum vesicae are easily traced out into that structure
from the ureteral walls and the mucosa which is arranged in folds
everywhere else in this specimen is attached tightly to the tri-
gone and is smooth. The sphincter vesicae has become quite
sharply outlined by many muscular fibers arranged circularly at
the constricted lumen of the cervix.

At the twentieth week the muscular fibers have become very
much more prominent but the most striking change is the enor-
mous increase in the size of and the number of fibers forming the
internal sphincter which seems to be tightly closing the vesical
orifice. The longitudinal fibers of the bladder intermingle with
the outermost of the fibers of the sphincter. The mucosa in this
stage 18 smooth over the trigonum vesicae and folded elsewhere.

In the twenty-two weeks oid fetus there is noted still greater
increase in the size of the bladder musculature and a marked in-
crease in the size of the interureteric bar. The trigone and the
bladder wall under it do not show the great increase in size over
the rest of the vesical wall observed in younger fetuses. The
fibers forming the trigone are observed to be more tightly bound
together than the fibers of the bladder wall and the mucosa is
smooth over it. The sphincter is large and closely bound to the
rest of bladder musculature.

Further increase in the size of muscles forming the vesical wall,
trigonum and sphincter is noted at the thirtieth week and in the
bladder of the new-born. The sphincter in the latter is larger at
its upper margin than elsewhere and is very thick on its posterior
quadrant. Many of the fibers appearing there do not entirely
encircle the orifice but intermingle with longitudinally arranged
fibers from the anterior surface of the bladder. At the lower or
outermost part of the sphincter there are fewer fibers but they
all entirely encircle the orifice.

334 OSWALD S. LOWSLEY

2. There is found in some of the sections studied a small group ©
of glands which open upon the mucosa at the middle and lower
part of the trigonum vesicae and which barely extend into the
underlying musculature. These glands are in nearly every case
simple tubules of very delicate structure, no muscular or special-
ized tissue layers being found surrounding them. They are not
found until the twentieth week but are observed in every fetus
older than that and also in the new-born. These tubules form the
subtrigonal group of glands and are in every case few in number,
nine being the most found and four the fewest. They are appar-
ently insignificant but occupy a strategic position as a very slight
hypertrophy at this point might cause a considerable obstruction
to urinary outflow. Cases have been observed by Dr. H. H.
Young and Dr. J. T. Geraghty in which this group of tubules had
become enlarged and a further growth had caused them to become
almost free in the bladder lumen being connected to the original
site by a small pedicle. Upon attempted urination this globular
mass would: fall into the orifice of the urethra and blocking it
would cause more obstruction than an enormous hypertrophied
prostate. The structures are probably a continuation upward of
urethral glands.

3. On the floor of the prostatic urethra of the fetus sixteen Sielk:
of age at its commencement there are found eight small evagina-
tions which are observed to extend only a short distance into the
submucosa. These tubules are easily distinguishable from pros-
tatie gland tubules being of very slight architecture and lacking
the muscular layers which surround the prostatic tubules. This
group of tubules first described by Albarran is found in all of
the specimens studied here older than sixteen weeks. In fetal
life they have no branches at all but in the new-born a few very
small branches were made out. In all cases these tubules grow
back toward the bladder in the submucosa and never extend
deeply into the musculature. In the new-born they have grown
back within the sphincter. These tubules are found in one in-
stance growing from the roof of the prostatic urethra. They are
few in number, varying from eight to nineteen and are very simi-
lar in structure to the subtrigonal glands just described, except

THE HUMAN PROSTATE GLAND BD)

they are somewhat larger. The position of the sub-cervical glands
of Albarran is even more strategic than that of the subtrigonal
group. Growing back directly within the sphincter it is easy to
see that a slight increase in size would form a very considerable
obstruction to the passage of urine from the bladder.

4. The prostate gland begins to develop at the third month of
fetal life. The tubules which compose it make their first appear-
ance as solid epithelial outgrowths from five distinct parts of the
prostatic urethra. These solid masses of deeply staining cells
very soon become circularly arranged around lumen and branches
are found very early. The five foci from which groups of pros-
tatic tubules take their origin are located as follows: on the floor
of the urethra between the neck of the bladder and the openings
of the ejaculatory ducts and utriculus prostaticus, one in each
prostatic furrow and on the sides of the urethra, on the floor of
the urethra below the openings of the ejaculatory ducts and the
utricle, and on the ventral or anterior wall of the prostatic ure-
thra. The tubules originating from these five foci by their fur-
ther growth and the development of stroma around them become
‘the middle, right and left lateral, posterior, and anterior lobes
respectively. In early fetal life they are widely separated from
one another but in later stages the separation between the middle
and two lateral lobes is not very great. There is not an inter-
mingling of tubules in any of the specimens studied but in many
places in the new-born the tubules of the middle lobe are observed
side by side with those of the lateral lobes, there being no definite
capsule separating them. The separation of the posterior lobe
from the others is complete, as there is a rather dense layer of
fibrous tissue between it and the lateral lobes. The anterior
lobe is widely separated from the two lateral lobes.

The first appearance of muscular fibers developing around the
tubules of the prostate is found at the sixteenth week at which
time a slight differentiation in staining properties is noticed. At
the twenty-second week the muscular layers are quite well devel-
oped and are particularly noted where some of the lateral lobe
tubules have extended away from the base of the prostate. The
musculature continues to become thicker and thicker until at

336 OSWALD S. LOWSLEY

birth some of the tubules are surrounded by very dense muscular
layers.

In every case but one in this study the middle lobe of the pros-
tate develops independently from tubules which originate from
the floor of the urethra near the middle line between the bladder
and the entrance of the ejaculatory ducts. The tubules compos-
ing this lobe are separated from the lateral lobes by a considerable
area of tissue free from glandular elements, the younger the em-
bryo the greater the separation. In later development while the
tubules and their branches maintain their characteristic position,
there does not seem to be any definite formation of a fibrous cap-
sule separating the middle from the lateral lobes. The ducts of
the largest of the middle lobe tubules which were originally near
the middle line are pushed laterally by the further development
of the structures in the verum montanum so that they open
rather high up on its sides and not in the middle line, as that part
is occupied by the utriculus prostaticus and the ejaculatory ducts.
Rarely (once in our experience) there may be an absence of inde-
pendent middle lobe tubules, in which case branches from the
lateral lobes are seen approaching the middle line in the region:
ordinarily occupied by the middle lobe, and if the growth con-
tinues this region will be occupied by a glandular commissure such
as Jores declares all middle lobes to be. In most cases the middle
lobe is made up of nine or ten large tubules, the number in five
cases varying from seven to twelve, the average being ten.

As one of our embryonic series showed an absence of tubules
in the middle lobe region it was deemed advisable to determine
the percentage of prostates that have a middle lobe.

Dissecting room and autopsy subjects to the number of twenty
were examined and all demonstrated definite prostatic tissue in
the middle lobe region. ‘Ten sets of serial sections of embryonic
tissue demonstrated only one case without independent middle
lobe tubules. Thirty-three autopsy cases of enlarged prostate
in the Johns Hopkins Hospital Genito-Urinary Museum were
examined, with the result that in thirty-one, definite middle lobes
were identified. The existence of middle lobes in the other two
cases could not be determined, as the specimens were not sec-

THE HUMAN PROSTATE GLAND 337

tioned. Two of the specimens examined had enlarged anterior
lobes. Forty autopsy specimens of prostate gland enlargement
were examined in the pathological museum at Guy’s Hospital,
London, thirty-seven of them showed the existence of glandular
substance in the middle lobe region, while in three cases there was
no macroscopic evidence of such a structure.

TABLE 1

Showing the frequency of the occurrence of middle lobes

| NUMBER OF DEFI- QUESTIONABLE | MIDDLE LOBE
| NITE MIDDLE LOBES MIDDLE LOBES DEFINITELY ABSENT

SPECIMENS

Twenty cadavers in Dr. Mall’s| |
SAD OnATORVA cs. ac.) ole eee 20 0 | 0

Menwdetuses?.. Ue. <i. +. esa eee 9 0 | 1
Thirty-three enlarged prostates, | |
Dre Youness) Clini Cyanmeeenr oe | Dill 2 0

Forty autopsy specimens of en-|
larged prostates, Guy’s Hospi- |
tal, London. . Lae ee ee 3

7
Total.. A os ca 97 5 | 1

Reference to table 1 shows that there is a definite absence of
glandular tissue in the middle lobe region in one specimen out of
one hundred and three studied, and five specimens in which it
was not determined whether there was glandular tissue there or
not.

The lateral lobes are composed of tubules which are greater in
size and number than those of any of the other lobes. They
originate from the right and left prostatic furrows and the lower
parts of the lateral walls and extend backward and outward form-
ing the main part of the base of the prostate. In the younger
specimens these lobes are widely separated from each other and the
remaining lobes but the older the fetus studied the less is the sepa-
ration between the branches of the lateral and middle lobes,
although there is in all cases a definite separation observed where
the ducts communicate with the urethra. In the region of the
apex some of the lateral lobe tubules send branches forward, at
all other parts of the gland the direction of the growth of tubules
is back towards the bladder.

338 OSWALD S. LOWSLEY

In one specimen there was observed a growth of branches from
lateral lobe tubules into the region usually occupied by the middle
lobe, the latter structure in this case being absent.

In the prostate of the new-born near the apex the lateral lobes
have two extensively branching tubules which extend up into
the anterior commissure and practically form a glandular commis-
sure in this region.

The lateral lobes are in most cases very well separated from
each other anteriorly a considerable area of stroma being present,
in which are found the anterior lobe tubules. Mesially the ejacu-
latory ducts, urethra and middle lobe are interposed and there is
a definite plane of connective tissue which sharply separates their
posterior branches from the posterior lobe.

The number of tubules forming the two lateral lobes vary from
twenty-seven to forty-six, the average number being thirty-seven.

The posterior lobe is made up of tubules which begin to develop
with the other prostatic tubules at the third month. They are
found on the floor of the urethra below the openings of the ejacu-
latory ducts and their direction of growth is behind those struc-

tures back towards the bladder. They compose the main mass

of the apex of the gland and the posterior lobe is that part of the
gland which is palpated per rectum. Just anterior to this lobe is
found in the older stages a stroma of connective tissue free from
tubules which separates it from the posterior parts of the lateral
lobes and from the ejaculatory ducts. The tubules of this part
of the prostate send a few branches forward in the region of the
apex as do the most anteriorly arranged branches of the lateral
lobes. They are quite large with numerous branches’but are not
very numerous, the smallest -number being four, the greatest
eleven, and the average in six specimens is eight.

This lobe is present in all fetuses here studied and in the new-
born. It is an independent structure developing from tubules
which are separated from those composing the other lobes. and
being divided from them by a definite capsule which is laterally
connected with the capsule of the gland.

The posterior lobe is of considerable importance for several
reasons. In doing Young’s operation for perineal prostatectomy

THE HUMAN PROSTATE GLAND 339

operators have found it absolutely necessary to make their two
parallel incisions quite deep so that they go completely through
the connective tissue layer separating the posterior from the two
lateral lobes. In case the incision is only made through the cap-
sule of the gland into the posterior lobe, an attempted enucleation
leads the operator’s instrument or finger laterally into the outer
capsule again where the anterior capsule of the posterior lobe
becomes lost in it and enucleation of the real offenders in hyper-
trophy, lateral and middle lobes, is not possible until the incision
is made into the capsule containing them.

Recent studies by Dr. John T. Geraghty and Dr. Montague L.
Boyd on the pathology of the prostate gland have confirmed the
facts (1) that hypertrophy rarely or never occurs in this lobe, (2)
that primary carcinoma of the prostate usually arises init. This
knowledge coupled with the embryological fact that the tubules
forming this structure arise independently from a localized area
in the urethra and remain independent throughout, increasing
enormously in size in the normal adult prostate, suggests the possi-
bility that it may have a different function from the other parts
of the gland.

The tubules forming the anterior or ventral lobe begin to de-
velop at the same time as do those of the other lobes. They are
large and have numerous branches_at first, but in the sixteenth
week they are slightly smaller than the tubules of the other lobes.
At the twenty-second week these tubules have decreased in size
and number and very few branches are noted. There seems to
have been a shrinking into insignificance of the anterior lobe after
the sixteenth week, but the tubules persist until birth, at which
time there are found two very small tubules. We have evidence
of the fact that the anterior lobe may persist throughout life in
that among the ninety-three specimens of the adult prostate
examined two were found with hypertrophied anterior lobes. The
average number of anterior lobe tubules in the first half of fetal
life is thirteen while that of the last half is six. Two tubules in
the new-born were the fewest found, fourteen in the sixteen weeks
old fetus being the greatest number.

340 OSWALD S. LOWSLEY

In most descriptions of.the prostatic urethra it is stated that
there are from twenty to thirty duct openings upon its floor. My
studies have convinced me that this number is far too low. The
results of microscopic studies of the number of tubules composing
the prostate have been arranged in the form of a table.

TABLE 2

Showing number of tubules of each lobe opening into prostatic urethra, the number of
Albarran’s tubules, and the number of subtrigonal tubules.

SIZE OF FETUS | TOTAL NO. |SUBCERVI- |

CROWN-RUMP | MIDDLE LATERAL | POSTERIOR | ANTERIOR OF PROSTA- \CAL GLANDS) SUBTRIGONAL
aan | LOBE LOBES » LOBE LOBE ipI0 TUBULES OF eae? | GLANDS
cm. | a , ¥ ; i z i 7 ¥ |
(itso ae a a 39 HM 12 74 | Ome 0
Pe eeny fi 27 6 13 53 0 0
IQS Riri VO 46 | 4 14 74 | 8 0
19s 5c 0 42 10 ff 59) sali li 5
PAROS || 36 | 9 8 64 | 9 4
ope be 1) ais a Ra 2 11 2 56 19 9
Averages* 10 3M 8 9 G3) y7 WR | 6

* The averages are taken from the specimens in which the structure is present
‘in case of middle lobe and the groups of Albarran and the subtrigonal group.

By referring to table 2 it is seen that in no case were there fewer
than fifty-three prostatic ducts opening into the urethra, and in
two specimens there were as many as seventy-four, the average
for six specimens studied microscopically in series being sixty-
three, including one specimen in which the middle lobe was en-
tirely lacking. The number of middle lobe tubules vary from
seven to twelve, the average in five prostates being ten. The
lateral lobe tubules vary in number from twenty-seven to forty-
six, the average in six specimens being thirty-seven. The pos-
terior lobes show a variation of from four to eleven, eight being
the average number of tubules in the six specimens recorded.
The anterior lobes present a very interesting variation. It is
seen that up until and including the sixteenth week the tubules
composing the anterior lobe are quite large and numerous, but
after that time there is a decided decrease in the number, and, as
has already been stated, the size and branches of the individual

THE HUMAN PROSTATE GLAND 341

tubules. The greatest change in this structure is noted in the
new-born in which the anterior lobe is made up of two insignifi-
cant tubules situated on the anterior part of the gland at its middle
as shown in fig. 10.

The subcervical glands of Albarran are seen to occur in the
specimen sixteen weeks of age and in all of the older ones. Their
number varies from eight to nineteen, the average in the four
series being twelve.

The subtrigonal glands are not found until the twenty-second
week and they are in all cases few in number, varying from four
to nine, the average in three specimens being six.

5. The vasa deferentia descend behind the posterior surface of
the bladder, each one being situated near the ureter. . Lower
down they gradually approach one another and in the earlier
stages become contiguous behind the middle of the trigonum vesi-
eae. In their descent they increase enormously in size, so that
under the internal sphincter of the bladder they with their envel-
oping tissue are larger than the commencement of the urethra
with its surrounding tissue. At this point the lumen of the vasa
deferentia have widened considerably denoting the first appear-
ance of the ampullae. These structures remain comparatively
large until the sixteenth week after which they become relatively
very much smaller.

The seminal vesicles originate in the thirteenth week. They
appear first as an evagination lateralward from each vasa deferens
being covered by the same tissue that envelops the latter struc-
tures. They grow backwards and laterally consisting of a main
part which is convoluted and from which rather numerous, short,
convoluted branches grow out as described by Pallin.?? In the
further development of these structures the opening into the ejacu-
latory ducts becomes comparatively smaller and its component
parts become larger and more tortuous. In the thirty weeks old
fetus they are found back under the trigonum vesicae at about its
middle point and communicate with the vasa deferentia deeply
in the base of the prostate to form the ejaculatory ducts. At

23 Gustaf Pallin, Archiv fiir Anatomie und Physiologie, 1901.

342 OSWALD S. LOWSLEY

birth the ends of the seminal vesicles have extended back almost
as far as the base of the trigonum vesicae and four branches from
the main lumen of the organ are made out. Its walls are made up
of fibrousand muscular strands and are almost as thick as those of
the ejaculatory ducts.

The ejaculatory ducts are the continuations of the vasa deferen-
tia below the entrance of the seminal vesicles. They are like the
vasa deferentia very large in size in the younger fetuses but grad-
ually become comparatively smaller in size as the other structures
grow larger. They pass obliquely through the posterior wall of
the prostatic urethra accompanied by the utricle and the surround-
ing stroma layers the outer fibers of which attach this cylindrical
mass intimately to the urethral wall by intermingling with the
tissue forming it. As these structures approach the lumen of the
urethra its floor is pushed up into a mound and its lumen con-
verted from a triangular into a semilunar shape. This mound,
called the verum montanum, is made up entirely of the ejacula-
tory ducts, utricle and surrounding envelopes, and gradually
disappears below the openings of these several structures into the
urethra, the stroma cells becoming intermingled with those form-
ing the urethral wall.

The ejaculatory ducts pass through the posterior wall of the
urethra on a gradual slant in the younger fetuses but in older ones
their rise through the prostate is a very sudden one until they are
within a short distance from the lumen of the urethra, then they
run along parallel to its axis for a considerable distance so that
their course in the prostate is very much like that shown in fig.
10 in all of the older specimens. They open on the sides of the
verum montanum in such a way that there is a small area of
tissue over them which, if the slightest amount of pressure were
exerted upon it, would close them most effectively. In fact their
whole course in the verum montanum parallel with the axis of the
urethra near its lumen makes for a closure of the duct in case of
distention of the urethra.

6. The urticulus prostaticus in the thirteen weeks old fetus is
observed as a very small lumen between the vasa deferentia.
It descends between them, and just below the beginning of the

THE HUMAN PROSTATE GLAND 343

ejaculatory ducts becomes much larger than either of these struc-
tures. It passes through the wall of the prostatic urethra in
company with them and opens in the midline just below the open-
ings of the ejaculatory ducts. In the sixteen weeks old fetus the
utricle seems to be composed of two partially fused tubes. Its
upper end is obliterated and nothing is seen of it until the point
is reached where the ejaculatory ducts and their envelopes are
entirely within the posterior wall of the prostatic urethra. It
opens below the mouths of the ejaculatory ducts. The utricle
begins between the ejaculatory ducts before they enter the pros-
tate in the twenty-two weeks old fetus as shown in fig. 6. It is
very much larger than those observed in other specimens and
shows other peculiarities already described. In all of the speci-
mens older than twenty-two weeks the utricle appears only in the
tip of the verum montanum. It opens in the middle line and in
nearly every case below the openings of the ejaculatory ducts.
In none of the specimens studied has there been found a single
case in which an ejaculatory duct for a prostatic tubule opened
into the utriculus prostaticus. There is no evidence that its
mouth is protected in any way from invasion of substances or
organisms in the posterior urethra.

7. Below the apex of the prostate there is noted in some cases
a large number of glands of Littré. They are quite large and some
of them have branches. They do not extend into the muscular
layers of the urethra. They become very few in number lower
down in the urethra.

CONCLUSIONS

1. -The mucosa is always free from folds over the fetal trigo-
num vesicae. The musculature of the bladder wall, trigonum,
and sphincter begins to develop at the thirteenth week and by
the sixteenth week is very pronounced.

2. The subtrigonal glands begin to develop at the twentieth
week. They are found at all ages after that, are few in number
and insignificant in appearance, their only importance being that
they occupy a strategic position where a small pedunculated
enlargement might cause great obstruction to urinary outflow.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 3

344 OSWALD S. LOWSLEY

3. The subcervical glands of Albarran are constant after the
age of sixteen weeks. They are similar in structure but are more
numerous and larger than the subtrigonal tubules. Their impor-
tance also lies in their position. They originate from the floor of
the urethra within and below the internal sphincter and grow
back under the mucosa so that they lie directly within it. Hence
a slight increase in their size would cause very grave obstruction
to urinary outflow.

4. The prostate gland originates from five independent groups
of tubules which begin to develop at about the twelfth week as
follows:

(a) The middle lobe is made up of nine or ten large branching
tubules originating on the floor of the urethra between the bladder
and the openings of the ejaculatory ducts. There may be an
absence of the middle lobe in which case there may be an ingrowth
of tubules from the lateral lobes to form a commissure beneath the
urethra. Embryologically the middle lobe is an independent
structure. Practically it makes no difference because it is not
separated by a capsule from the lateral lobes. The middle lobe
is rarely absent, being found definitely present in ninety-seven
specimens examined, possibly lacking in five others and definitely
absent in one.

(b) The right and left lateral lobe tubules originate in the pros-
static furrows and from the lateral walls of the urethra. They
are composed of from twenty-seven to forty-six tubules which
erow back to form the main part of the base of the prostate.
They are well separated from each other by the anterior lobe and
commissure, the urethra, the middle lobe and the ejaculatory
ducts. Posteriorly they are separated from the posterior lobe
by a fibrous capsule.

(c) The posterior lobe is an independent structure being made
up of tubules which originate from the floor of the prostatic
urethra below the openings of the ejaculatory ducts. They grow
back behind the latter structures and are in no sense a glandular
commissure as they are definitely separated from the other parts
of the gland. The posterior lobe is the part of the prostate pal-
pated per rectum and is an important consideration in the per-

THE HUMAN PROSTATE GLAND 345

formance of perineal prostatectomy. Hypertrophy rarely or
never occurs in it and primary carcinoma of the prostate rarely
or never begins anywhere else (Boyd and Geraghty).

(d) The anterior lobe is fairly large until the sixteenth week
after which time it becomes greatly decreased in size and in the
number of its tubules. It was found in all of the microscopic
specimens studied but shrinks into insignificance at the twenty-
second week. There is evidence in the occasional finding of en-
larged anterior lobes at autopsy that this structure may persist
throughout life.%

The number of openings of prostatic tubules into the urethra
is usually said to be between twenty and thirty. My studies
have convinced me that this number is too low as in the specimens
here recorded the number of tubules opening into the urethra
varies from fifty-three to seventy-four, the average being sixty-
three.

5. The vasa deferentia in early fetal life are comparatively
speaking very large, being greater in size than the urethra at the
thirteenth week. Their lumina broaden out behind the vesical
sphincter at this time showing the earliest appearance of the
ampullae. They become relatively decreased in size after the
sixteenth week. :

6. The seminal vesicles begin to develop as lateral evaginations
from the vasa deferentia at the thirteenth week. They grow
backward and laterally becoming more or less tortuous and send
off as many as four short tortuous branches. In the later stages
they communicate through a very narrow duct with the vasa
deferentia just within the base of the prostate gland.

7. In the younger fetuses the ejaculatory ducts pass obliquely
through the posterior wall of the prostatic urethra forming with
their envelopes the verum montanum. In the older specimens
their course through the prostate is not so regular. At the base
they progress on a gradual slant until the middle of the gland is
reached, where they rise quite sharply until they lie in the top of
the verum montanum, after which their course lies parallel to

*4 Kuznitzky found a persistent ventral lobein one out of every fifteen prostates.

346 OSWALD S. LOWSLEY

the axis of the urethra for some distance. They open into the
urethra on the sides of the verum montanum, their mouths being
composed of a collapsible fold of tissue so that pressure within
the prostatic urethra very effectively closes them.

8. The fused Miillerian ducts may persist intact until the thir-
teenth week, after which time the lower end (utriculus prostaticus)
which has become quite large in size and surrounded by a rather
dense layer of stroma cells, contracts until after the twenty-second
week when it is found only in the tip of the verum montanum and
is relatively very small in size. It usually opens in the midline
just below the ejaculatory duct openings and rarely if ever is
there an ejaculatory duct or a prostatic tubule opening into it.

Description of a wax model of prostate of a new-born infant

The sections from which this model was constructed were cut
30 micromillimeters in thickness. Every fifth section was mag-
nified twenty times and drawn by means of the projection appa-
ratus in use at the Anatomical Laboratory of the Johns Hopkins
‘University. These drawings were traced upon wax plates 3 mm.
in thickness, the tubules and their branches first being identified
microscopically with great accuracy. The wax plates were then
cut in such a way that the bladder lumen and those of the pros-
tatic tubules were left with bridges of wax between to preserve the
exact contour. The wax plates were then piled, the axis of the
bladder and urethral lumen and the lumen of the ejaculatory
ducts being used as points upon which to build. The prostatic
tubules are represented with their branches grouped, it being
obviously impossible to represent every branch of each tubule.
The various parts of the model are held together by means of
pins and copper wire so that the exact position of the various
structures represented is maintained. The various parts of the
model are painted with several coats of different colored enamel
to make clearer the different structures reproduced.

In conclusion I wish to express my thanks to Drs. Clark,
Mall and Young for many suggestions which were of great value
to me in this investigation.

THE HUMAN PROSTATE GLAND PLATE 1

OSWALD 8. LOWSLEY

The diagram in the corner of each plate shows the exact region of the prostate
reproduced in the drawing.

Fig. 9 Dorsal view of a wax model of the prostate of a new-born infant.
«x 14. Lat., lateral lobes; P.L., posterior lobe; E.J., ampullae of vasa defer-

entia; S. V., seminal vesicles; A.B., anterior branches of lateral and posterior

lobes; U., urethra.
347

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 3

PLATE 2 THE HUMAN PROSTATE GLAND
OSWALD 8S. LOWSLEY

Fig. 10 Sagittal view of a wax model of the prostate of a new-born infant. X 14. Lat., anterior
branches of lateral lobes; P.L., posterior lobe; H.J., ejaculatory duct; S.V., seminal vesicle; A.L.,
anterior lobe tubule; U., urethra; U.P., utriculus prostaticus; A.G., subcervical glands of Albarran;
M.L., middle lobe tubules; L.Ur., left ureter; Bl., bladder; P.Gl., prostate gland.

348

THE HUMAN PROSTATE GLAND PLATE 3
OSWALD S. LOWSLEY

Fig. 11 View of wax model of the prostate of a new-born infant with posterior
and most of lateral lobe removed. X 14. Lat., lateral lobe; #.D., ejaculatory
duct; S.V., seminal vesicle; U., urethra; M.L., middle lobe tubule; P.L., cut
ducts of posterior lobe; L.Ur., left ureter; R.Ur., right ureter.

049

A ’ — ‘
af 3 . ah
i: . ‘
, 7)
f
Rs ' :
:
p \
- : u — J
B é, - ;
‘ &
= * * a
ra “ #
= - .
‘ ; - ie: [ 7 =
‘ > +] 7 r
P | : Poe
si
La n
~ “hy 7 '
Ae i =. » 4 _ rt '
. s -
J ; i
2 “ 4 ‘
% ‘ 7 = : A ,
a rap 7 i ~N " - 7
2 ‘ ; \ ti
~ '*
-
a 1 :
{ ~ 7
: 4
4 ~ w=) :
e - % ‘
. '
: .

FURTHER OBSERVATIONS ON LIVING GROWING
LYMPHATICS: THEIR RELATION TO THE
MESENCHYME CELLS

ELIOT R. CLARK

r From the Anatomical Department of The Johns Hopkins University

EIGHTEEN FIGURES

In an earlier paper,! read before the American Association of
Anatomists in December 1908, the results were given of a series
of observations on the method of growth of lymphatic capillaries
as revealed by a study of the transparent fin expansion of the tail
of living frog larvae. The present paper of which the essential
parts were presented to the American Association of Anatomists
in December 1910, represents the results of a newer set of studies,
made during the spring of 1910.

The methods employed in the second study were identical
with those used in the first; they deserve a fuller description than
has been given. As there stated, two factors are essential to
the success of the observations, an upright chamber and chlore-
tone anesthesia. The former allows the larva to remain in its
normal upright position while being watched; the latter keeps
it motionless, without seriously interfering with the circulation
of the blood or with the growth of the tissues.

The upright chamber which was used is by no means new to
the histological laboratory. The earliest description of such an
apparatus which I have found is by Cori.? In the article in
Zeitschrift fiir wissenschaftliche Mikroskopie he describes the first
apparatus, which he devised, as follows:

1}. R. Clark, Observations on living, growing lymphatics in the tail of the
frog larva. Anatomical Record, vol. 3, no. 4, 1909.

2 Cori, Lotos, Bd. 13, referred to in Zeitschrift fiir wissenschaftliche Mikros-
kopie, Bd. 10, 1893, p. 149.

351

Se ELIOT R. CLARK

Nach der in der Zeitschrift Lotos gegeben Beschreibung bestand aus
einem Objecttrager von Format 5:10 cm., auf welchem ein |_| formig
gebogener Glasstreifen aufgekittet war. Derselbe functionirte als Seit-
enwinde und Boden des Aquariumraumes. Die Ruckwand dagegen
bestand aus einem Deckglaschen vom Formate 30:40 mm. Die ganze
Vorrichtung wurde mittels Klammern wie ein Objecttrager auf dem
Tisch eines umgelegten Mikroskopes befestigt.

An upright chamber with the tube of the microscope horizon-
tal was used by Professor Mall in unpublished studies on the

Fig. A Apparatus used in making the observations; for explanation of the
letters see text.

blood vessels in the tail of the frog larva. Harrison? used a
‘glass cell having plane walls’ to hold living frog larvae in their
normal position while being photographed. Since the upright
chamber is quite important for the success of the observations,
a description of the one used will be given. It corresponds almost
exactly with the one described by Cori. On an ordinary 51:76
mm. glass slide fig. A (S), are fastened, with the aid of damar or
Canada balsam, three narrow strips consisting of two thicknesses
of window glass, in the position | | The uprights, (B) and (C),

3 R. G. Harrison, Archiv fiir Entwickelungsmechanik der Organismen, Bd.
7, 1898, p. 434.

GROWING LYMPHATICS AND THE MESENCHYME BB

are fastened to the horizontals (A) by damar. The fifth side of
of the chamber is formed by a cover-slip (£), while the sixth
side is left open. Paraffin is used in order that repair may be
quickly made if the cover-slip is cracked or broken. A paraffin
coating of the floor of the chamber was found to be useful. Such
a cell has a thickness of about 5 mm., and a depth of from 10 to
12 mm. For larvae of more than 20 mm. in length, a thicker
cell is necessary.

In use the slide is clamped to the stage of the microscope
which is arranged with the tube horizontal. The tadpole to be
examined is anesthetized in a small dish of chloretone solution,
and is transferred carefully from dish to cell by means of a pipette
or medicine dropper. The tadpole is brought carefully to the
side of the cell next the cover-glass by means of a blunt needle,
or by washing with a pipette. A narrow piece of cover-glass
(D) is set on the floor of the cell, leaning against the cover-glass
(£) which forms the front wall of the cell, and is brought against
the portion of the tail to be observed, by pressing gently with a
needle. In this way the part of the tail to be studied is held
firmly in contact with the cover-glass, and is protected from the
jarring of the water. Since the narrow piece of cover-glass comes
in contact with the thick central muscular portion of the tail,
pressure on the fin expansion is avoided. For older larvae, which
require more frequent changes of the chloretone solution, a special
apparatus was designed, by which a continuous circulation of
chloretone through the cell may be maintained. It is thismodifi-
cation which is shown in fig. A. The solution is poured into a
funnel, which may be raised or lowered as desired. A rubber
tube conducts the fluid to a fine glass canule (/’) which enters
the chamber at the lower corner. A small block of paraffin (7)
protects the tadpole from the direct stream. A second glass
canule (G) carries off the excess of fluid from the opposite upper
corner. It is possible, with a cell so constructed to make obser-
vations with the highest powers of the microscope, even with the
oil immersion, for a cover-slip, only, separates the microscope
objective from the object studied.

354 : ELIOT R. CLARK

For making records the camera lucida (E. Leitz drawing eye-
piece no. 112), with which the plane of the drawing-board makes
an angle of forty-five degrees with the tube of the microscope,
was used. A board, set at this angle to the horizontal, is placed
directly below the microscope, and is held firmly between clamps.
The outline sketches, made with the aid of the camera lucida
were usually corrected with a higher power ocular.

Chloretone (acetone-chloroform), the anaesthetic properties
of which were discovered by Abel, and which is rapidly replacing
other substances as an anesthetic for small organisms, because
of its efficiency and harmlessness, is a well-nigh perfect anesthetic
for tadpoles. At the stages on which these studies were made
it inhibits bodily movements and respiratory movements; while
the movements of the alimentary canal are unaffected, and, if the
proper strength is found, the force of the heart beat remains
undiminished. The strength of chloretone necessary to produce
the proper depth of anesthesia varies somewhat with the species
of tadpole. Even for the same species, individual differences
are met with, so that a set rule cannot be laid down. In general,
a solution made by dissolving 1 gram of chloretone in 5,000 parts
of tap water suffices. With larvae of hyla pickeringil, 1:3,000
may be used safely. For larvae of rana sylvatica, r. palustris,
and r. catesbiana, the necessary strength varies between 1 :4,500
and 1:6,000. It was found most convenient to make upa series of
dilutions containing 1:3,000 to 1:6,000 parts of chloretone. With
these on hand the strength may be varied according toindica-
tions. Since chloretone is somewhat volatile, fresh chloretone
must be added from time to time during the observations. The
‘indications’ which controlled the strength of chloretone to be
used, were, on the one hand, the return of muscular movement, on
the other, the weakening of the heart action. Evidence of the
latter is to be seen in the condition of circulation in the blood
vessels of the tail, which are in view throughout the observations.

The anesthetization of hyla pickeringii larvae is much easier
than that of the other three species mentioned. With a solution
of between 1:3,000 and 1:4,000 one of these larvae has been kept
under anesthesia five to twenty hours a day for more than four

GROWING LYMPHATICS AND THE MESENCHYME es

weeks without seriously interfering with heart action or the
growth of the tissues. The anesthetization of the other species
offers many difficulties. Often from a half-hour to an hour or
even more may be taken up in the attempt to bring the animal
to the proper point of anesthesia, where muscular movements
are lost and the heart beat is vigorous. Occasionally the heart
beat stops entirely. In such a case it will usually be resumed as
the result of a delicate massage of the heart produced by rhyth-
mically forcing a stream of water against the heart region with a
medicine dropper. The difficulties of anesthesia increase as the
larva grows older. It was thought that this might be due in
part, at least, to the greater need for oxygen. Whether this
explanation is true or not, certain it is that frequent changing
of the chloretone solution greatly assisted the anesthesia. This
is most conveniently done by employing the modified apparatus
described, with which a continous circulation of the chloretone
solution may be kept up. During the intervals between obser-
vations the larva is returned to fresh water.

A short review of the results of the observations previously
reported will now be given. It was possible to see clearly the indi-
vidual structures present in the fin expansion of the tail. Lym-
phatiecs, blood vessels, nerves, mesenchyme cells with their
branched processes, red and white blood cells, stand out with
remarkable clearness. The immobility of the object, which may
be maintained for hours, makes possible both prolonged obser-
vations of a small selected area, as well as careful records of small
or large areas. It is possible to watch, from minute to minute,
the changes which go on in a selected cell or process, and it is
also possible to make accurate drawings of the entire blood-
vascular or lymphatic plexus of both fin expansions.

Among the results of this first series of observations, in which
the structures were studied both extensively and intensively,
it was found that lymphatics grow by a process of sprouting, in
which, so far as concerns the material of which their endothelium
is formed, they maintain a complete independence of blood ves-
sels, mesenchyme cells, and wandering cells. In the species
studied the blood vessels precede the lymphatics, in their invasion

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 3

356 ELIOT R. CLARK

of the fin expansion. The lymphatics’ grow in as outgrowths
from the dorsal and ventral longitudinal lymph trunks, and soon
catch up to the blood capillaries; from then the two systems are
practically coextensive. Careful studies were made of the two
sets of vessels, as they developed side by side, and each was
found to be quite independent of the other. Blood capillaries were
watched through their various transformations. Some were
seen to be converted into arterioles or venules as the capillary
area increased, while others were watched through the stages of
atrophy, solidification and retraction. Neither new nor atrophy-
ing blood capillaries contributed tissue to the lymphatics.

The lymphatics are apparently quite uninfluenced by the blood
capillaries. They have a specific independent life of theirown. If
a capillary is selected and carefully observed, it is found to be ina
state of perpetual activity. Fine pointed protoplasmic processes
are being continuously sent out at the sides and tip. These proc-
esses reach varying lengths and most of them areagain withdrawn.
The lumen of the capillary may, however, extend into the base
of one of these fine processes, and, gradually extending, lead to
an increase in the length of the capillary, or to the formation
of a branch. A branch which is at first without a nuclear area,
will, if it continues to grow, receive one from the parent stem.
That these growth processes are intimately associated with func-
tional activity was suggested by the observation that extravasated
red blood cells may be taken into the interior of the capillary,
through one of these finely pointed processes.!

4 A. Daiurzynski (Untersuchungen tiber die Regeneration der Blut-und Lymph
gefiisse im Schwanze von Froschlarven, in Extrait du Bulletin des Sciences
de Cracovie, serie B: sciences naturelles, March 1911) states, in a footnote,
pp. 207-208, that he has failed to observe the taking up of red blood cells by the
lymphatic capillaries, although he sought eagerly for it. The accuracy of my
observation of this unexpected process is hardly open todoubt. It was seen many
times and with perfect clearness, and has been verified in the studies which are
to be reported in the present paper, on hyla pickeringii larvae. The illustrations
of the various stages, given in my earlier paper, (E. R. Clark, Anatomical Record,
vol. II, no. 4, 1909, figs. 6 and 7) are not exaggerated. It must be concluded, then,
that for some reason Dziurzynski has missed seeing the process. I am inclined
to believe that his failure is due to the crudity of his method of observation.
On page 189 he states that his larvae were placed on a slide, with a cover-glass
over the tail. Also that he was able to observe a larva for only one, or, at most,

GROWING LYMPHATICS AND THE MESENCHYME oon

A few observtions were made on the relation between the grow-
ing lymphatic and the mesenchyme cell. Here again, as in the
case of the blood capillaries, while this point was not subjected
to such rigid tests, nothing was seen which would suggest that
there is any transfer of tissue from mesenchyme to lymphatic.

The finding of larvae of hyla pickeringii, during the spring of
1910, made possible a new set of studies. The tadpoles of this
species offer peculiar advantages over those of rana sylvatica,
r. palustris, and r. catesbiana which were previously used. They
are smaller and more transparent, are free, for a longer period,
from toublesome pigment cells, and are very much easier to
anesthetize. But of much more importance was the discovery
that, in the posterior portion of the tail, the lymphatics grow
into the fin expansion of the tail in advance of the blood vessels.
In fact, specimens may be met with in which, over a region cor-
responding to the posterior twenty muscle segments, the lym-
phatics have grown out in abundance, reaching almost to the
fin margin, without a single blood capillary appearing beyond
the edge of the central muscle mass (fig. 1).

The importance of this discovery will be appreciated readily.
If, on further investigation, it be found that, in this area, no blood
vessel has at any time been present, during the growing out of
the lymphatics, the question as to the possibility of the growth
of lymphatics independent of blood vessels will be settled.
Obviously the only way to be certain is to select a portion of the
lymph trunk before the branch has appeared, to watch its growth
and see whether there is even a transitory appearance of a blood
capillary near it which might contribute to its growth. If it be
found that the lymphatics do indeed precede the blood vessels,

two hours, because of the return of movement. Had I used such a method I feel
quite certain that I should also have failed to observe this process, for it was only
by the use of the upright chamber, and by the most careful use of the chloretone
so that the same larva could be observed for many hours at a time, and over
periods of several days, that I made the discovery. Another possible explanation
of the divergence of our results may lie in the difference in age of larvae studied.
While Dziurzynski does not state the age of the ones used for this particular
study, it may be inferred from his paper that he studied much later stages than I.
The larvae which I studied were rana catesbiana larvae 8 to 9 mm. long.

358 2\)) BRIOT TRY CLARK

such a region offers an unparalled opportunity to study the rela-
tion between the growing lymphatic and the erowing mesenchyme
cell.

It seemed that both these points could well be taken up in
one set of observations, for it is easily possible, while making a
study of the growth of a lymphatic sprout from its very beginning,
to note the presence or absence of a blood capillary in the neigh-
borhood.

Fig. 1 Posterior half of tail of hyla pickeringi larva, showing region in which
the lymphatics precede the blood vessels. This larva is the one on which the
studies shown in figs. 2 to 14 were made. This drawing was made from the living
larva on June 3, at the close of the observations. Lymphatics are in solid black,
blood vessels in lines. The ventral caudal lymph trunk and a short stretch of the
dorsal caudal lymph trunk are shown. A, the lymphatic sprout shown in figs.
2 to 14. The larva is shown in this figure in its normal position. In the other
drawings the structures are shown in their reverse position, as seen under the
microscope. Enlargement 25 times.

It is most convenient to anticipate results at this point enough
to say that during the series of studies to be recorded, no blood
vessel entered the region under observation.

In order to discover the exact relationship which the lymphatic
endothelium and the mesenchyme cell bear to one another during
their growth, it is first necessary to find out just how each one
develops by itself, and then to study the relations between them
when they develop side by side. For the mesenchyme cell, a
separate study of the observable growth changes is easily possible,

GROWING LYMPHATICS AND THE MESENCHYME 359

since areas may be selected in which no lymphatic or blood vessel
has appeared.

The exact method chosen for making the desired studies was
as follows. In the tail of the tadpole the lymphaties first appear
as two longitudinal vessels, which have been many times described,
the dorsal and ventral caudal lymph trunks. In hyla pickeringii
the dorsal trunk, save for a short distance at the tip of the tail,
is concealed between the muscle layers. The ventral caudal
lymph trunk, on the other hand, may course for considerable
stretches, even for its entire extent out from under the muscle,
so that it may be clearly seen. The caudal vein is situated just
dorsal to the lymph trunk, while, further dorsally, ventral to the
notochord, runs the caudal artery. The vein may be seen quite
readily, especially toward the tip of the tail, where the muscle
layer is thinnest. A stage may be selected in which neither
blood nor lymph vessel has grown from the main trunks into
the ventral fin.

It was planned to select for study a portion of the ventral caudal
lymph trunk, at this early stage, before any branches had been
sent out, to make a careful record of each individual mesenchyme
cell around it, as well as of each mesenchyme cell in the entire
region in which a sprout of the lymphatic would be expected to
grow, and, in case a sprout were sent out in the region hoped,
to make numerous successive studies and records of each indi-
vidual mesenchyme cell, as well as of the lymphatic, as they
developed side by side.

My hopes were amply fulfilled. In the selection of a portion
of the lymph trunk, a place was chosen where a slight ventral
bend occurred, as shown in fig. 2. The larva, at this time, was
about 5.5 mm. long. At this spot the lymphatic was well away
from the muscle edge, so that both walls of the lymphatic were
clearly visible. It was also well separated from the caudal vein.
The vein was easily recognizable through the very thin muscle
layer. A drawing was made of the lymphatic, vein, muscle edge,
and of the neighboring mesenchyme cells, with their main proc-
esses. The tadpole was then placed in fresh water over night.
The following morning a lymphatic sprout had started to grow

360 ELIOT R. CLARK

out from the trunk at exactly the spot chosen. A new drawing
was made of the various structures (fig. 3). The larva was kept
under anesthesia the entire day, and frequent drawings made,
figs. 3, 4, 5, 6, 7 and 8, for the lymphatic was ina state of very
rapid growth. <A record was made of the mesenchyme cells well
in advance of the growing lymphatic, toward the edge of
the fin. The following day a new record was made (fig. 9,) in
which the area of recorded mesenchyme cells was extended to
the fin margin. From this time on, for thirty-one days, records
were made almost daily, and in each record there was incl ded
lymphatic, caudal vein, and all the mesenchyme cells, with their
main processes, in the region selected. It is unnecessary to re-
produce all the drawings. They were demonstrated at the meet-
ing mentioned. ‘Two later stages are selected (figs. 10 and 11)
which were drawn at eight and at twenty days later than fig. 9.
In addition to this series, numerous other studies were made on
other regions in this same tadpole, as well as on other larvae.
Fig. 15, for instance, in which are given the results of a study
primarily of the changes in the capillary wall, is taken from the
‘dorsal fin of another larva. The camera lucida drawings were
made at an enlargement. of 450x, and corrections were made at
a still higher enlargement.

In order to simplify the description, the two tissues—mesen-
chyme and lymphatie—will be taken upseparately. The relations
between the two as they grow together will then be indicated.

The mesenchyme cells at the time at which they first become
visible in the fin expansion of the larva of hyla pickeringil, consist
of an irregularly shaped thick central portion containing a
nucleus which is not clearly visible during life. From this central
portion extend a varying number of branched processes of differ-
ent lengths, and of a thickness which diminishes from center
to periphery. The branches in turn give off large numbers of
minute fibrillae, which form a richly anastamosing network ex-
tending from epidermis to epidermis. These minute fibrillae may
be seen and followed in the living larva, by careful focussing.
They usually appear as minute dots which seem to travel up or
down as the microscope is focussed. When any of these tiny

Figs. 2 to 11 Successive drawings of the same region (A in fig. 1) in the fin
expansion of the tail of hyla pickeringii larva, to show the relation between the
growing lymphatics and the growing mesenchyme cells. Drawings were made
while the larva was anesthetized with chloretone. The individual mesenchyme
cells, A to Z and 2 to 20, are represented by three forms; solid black: cells near
the surface toward the observer; cross-lined: cells in interior; dotted: cells near
the surface away from the observer. The same letters in the different drawings
indicate the same cells. Where a cell has divided the two daughter cells are
indicated as B! and B?. Between figs. 9 and 10, also between figs. 10 and 11,
records were made nearly every day, which have not been reproduced, which show
the intermediate positions of the cells and of the lymphatic. Lymphatic wall,
Lym LL, is represented as somewhat thicker than it actually is. The nuclear
areas in the lymphatic are represented by dots. In figs. 4,5 and 6 the three cells
shown Bb, J and K, are the ones in approximately the same layer as the lymphatic.
Fig. 8 is included because it shows well the wandering into the sprout of a nuclear
area (nucl), from the mainstem. In figs. 4to7 this nuclear areais seen just start-
ing into the sprout. Enlargement 300 times.

361

362

a

bm
wal zi feaank
<

ih A) 2. i

GROWING LYMPHATICS AND THE MESENCHYME 365

fibrillae is carefully followed, it is seen to lead to one of the larger
easily visible processes. The behavior of these minutest fibrillae
has not been considered in the present study—the drawings
including only the processes which are readily seen. Between
the larger processes communications are rare, though here and
there undoubted connections may be seen. The branched mesen-
chyme cells have a uniform distribution, an equal amount of
space separating the neighboring cells.

Near the edge of the muscle, where the distance between the
two layers is greatest, the cells are arranged in three or four layers.
Near the fin margin, where the fin is so thin that the two layers
of skin are almost in contact with each other, the cells form but
a single layer. Between these two extremes there are all grada-
tions. In making records of the cells a color scheme was used to
designate roughly the cells in the different layers. Thus, those
nearest the observer were drawn in black, those farthest away in
green, and those in the middle in red. In the reproductions,
the form has been changed, as indicated in the legend of figs. 2
to 11. With th’s crude scheme, it was found to be very easy to
identify each individual cell, with its main processes, in the suc-
cessive stages.

What now, are the visible changes which occur in these cells
during the growth of the tail?

As successive stages are compared it is soon to be noted that
changes are continually going on. On the one hand, branches
become thicker and longer, and new branches are sent out- On
the other hand, branches become shorter and thinner, until they
may entirely disappear. The thickened central portion, too, is
continually changing shape, now extending further and further
along one or more branches, while retreating from others. The
picture becomes much clearer when it is found that, by a summa-
tion of these processes, the cell may actually shift its position,
orwander. The evidence for thisconclusion does not rest upon the
observation of a single cell or a group of two or three cells; nor
does it depend upon the selection of some supposedly stable
structure to serve as a fixed point—such as a chromatophore
which may itself shift position. It is only certainly to be proved

366 ELIOT R. CLARK

by observing a considerable number of cells, by keeping records
of the distance of these cells from relatively fixed spots and from
one another, and by watching them over periods of time long
enough so that the shifting in position is beyond question. More-
over the shifting must be such as cannot possibly be accounted
for by the general growth expansion, of the entire organ, or by
a mere mechanical dragging of the cell.

That there is a genuine wandering of the cell is shown by a
study of the series of drawings from which figs. 2 to 11 are taken.
The region selected was large enough, the number of cells watched
was large enough, the length of time and number of successive
observations were great enough and the changes in shape and
position of cells were marked enough to exclude all other inter-
pretations. A general idea of the relative amount of shifting
is given in fig. 12, which shows diagrammatically the amount and
direction of change in position of cells between fig. 9 and fig. 11
‘(May 3 and May 23). Here we see that, while certain cells
such as A, P, U have remained in practically the same spot,
others such as B, C, J or their daughter cells B 12242, C 1and2,
J 1and2 have made long excursions while still others, such as
F, S have made short excursions. While the difference in loca-
tion of such a cell as B (B'), with reference to the muscle edge
may be in part accounted for by the general expansion of the
entire fin, measurements show that the change is very much
greater than could be accounted for in such a way. In fig. 9 (May
3) the width of the fin from muscle edge is 0.38 mm. The

Fig. 12 Diagram to show the amount and direction of change in position of the
mesenchyme cells shown in the figs. 2 to 11. The same form for the cells is used—
solid black, dots, and cross-lines to indicate the different layers. The squares
represent the position of the cells on May 8 (fig. 9), the circles their position
twenty days later, on May 23 (fig.11). Since the fin increased in size during this
time, the positions of the cells in fig. 8 were measured with reference to the muscle
edge as a horizontal and a vertical passing through the base of the lymphatic
sprout, and were referred to a similar horizontal and vertical in the diagram of
the later stage, with increases to correspond to the growth of the tail. Where
mitotic division has taken place, the line, which shows the change in position,
forks. The arrows represent the direction of wandering. In cases where cells
came into the field, which was being watched, after May 3, the square is omitted
(cf. cells 13, 15, 16, etc.)

GROWING LYMPHATICS AND THE MESENCHYME 367

368 ELIOT R. CLARK

distance of cell B from muscle edge is 0.078 mm., or 29.5
per cent of the entire width. On May 26, seventeen days later,
three days after the stage shown in fig. 11, cell B', one of the
daughter cells of cell B is 0.29 mm. distant from the muscle edge,
or 58 per cent of the entire width, 0.50 mm. Thus this cell has
moved through 37.5 per cent of the width of the fin, an actual
distance of 0.188 mm. Calculated in a similar way, the cell P
is found on May 3, 13.5 and on May 26, 14 per cent of the distance
from muscle edge to fin margin; it has, then, remained practically
stationary. Many striking instances of relative change in posi-
tion may be seen by following any two or three selected cells
through figs. 7 to 11, such as cells #, N and F.

C May 22

D May23 :

A. May. 20

Fig. 13 Four successive drawings of the same mesenchyme cell, taken from
the series of studies represented in figs. 2 to 12, to show the method of wandering
of the mesenchyme cell. The cell here shown is the one labelled G in figs. 2 to
12; stage D in this figure corresponds to cell G in fig. 10. The dotted cell in D
is the same as A. It is superimposed to show the change in position. Enlarge-
“ment 270 times.

The exact way in which this change in position of the cells
takes place is shown by a comparison of successive drawings
of the same cell. Cell B, in figs. 2 to 9, furnishes a good example,
since the several stages here are separated by short intervals of
time. In fig. 13 a single cell is shown, in four successive stages
at daily intervals. It will be seen that change in position is
effected by a movement of the protoplasm. Fine processes are
sent out from the main body of the cell, such as 3, in fig. 13. These
processes may be quite temporary, and may be withdrawn sub-
sequently, without attaining more than a hair-like thickness.

GROWING LYMPHATICS AND THE MESENCHYME 369

They may, however, grow thicker as new processes are sent out
beyond. This increase in thickness takes place at the expense
of other portions of the cell. Thus, while (fig. 13) processes 2,
3 and 4 successively increase in thickness, processes 1, 2 and 3
decrease. In fig. 13 A, the main body of the cell lies at /, in fig.
13 D, process 7 is merely a short hair-like filament. The mode of
progression, then, seems to be a true amoeboid one, since change
in position is brought about by the sending out on the one side
and the withdrawal on the other of cell processes, with a shifting
of the main body of the cell from the retreating to the advanc-
ing processes. That the wandering is an active process and not
the result of a purely mechanical pulling and pushing, is shown
by the character of the changes which the cell undergoes. It
is impossible that such changes as are shown in fig. 13 could be
produced by a merely mechanical dragging and pushing.

The rate of progression of the individual cells varies. Wander-
ing is most marked in the cells in the interior of the fin, while those
near the surface are almost stationary. ‘The former are repre-
sented with cross lines in the drawings, the latter are dotted or
solid black. ‘The cell B between fig. 2 and fig. 9, two days, wand-
ered about 22 micra. Cell G shown in fig. 18, wandered in three
days about 44 micra. Roughly speaking since the thicker por-
tion of the cell measures from 25 to 30 micra in length, the main
body of the cell may wander a distance equal to its own length
in two to three days. It is probable that this is rather low for
the normal maximal rate, since the prolonged use of chloretone
causes a slowing of the growth processes. Moreover, during most
of the time over which the observations extended, the weather
was unusually cold, which also causes a retardation in growth.
Since growth processes are much more rapid in the regenerating
than in the normal tail, it is probable that a very much meer rate
of wandering may be aed there.

The direction of wandering is principally toward the free mar-
gin of the fin, as may be seen by a glance at fig. 12. During the
observations many cells, such as cells 75, 16, 17, moved out into
the fin from the space between the muscle layers. A smallamount
of antero-posterior wandering occurred, as seen in cells H! and

370 ELIOT R. CLARK

H?, X! and N. A few cells such as cell D, (D! and D?) moved
from the interior toward the surface. One cell, R, which for a
time was spread out in part under one epidermal layer, and in
part under the other, gradually withdrew one set of processes
until it was eventually entirely on one side.

In addition to the amoeboid wandering, another striking change
takes place in these cells, namely, mitotic division. This process
has been figured and briefly described by Haidenhain.? My
findings agree in the main, with his, the chief difference being
that, while he states that the cells after division become fixed,
I find that, while movements are most rapid at the time of divi-
sion, yet after, as well as before, the cells are not ‘fixed’ but are
distinctly motile. The changes which take place during the divi-
sion of the cell are as follows (fig. 14). The thick central
portion of the cell swells and the processes become shorter and
thicker at the base and some are entirely withdrawn (fig. 14,
A and B). The spireme appears as dark spots which soon form
in a line across the equator of the cell (fig. 14, C), this line divides
rapidly and the chromosomes as two sets of lines move in opposite
directions to the poles, where they shorten and become moulded
into spheres, the two daughter nuclei. These nuclei are ciearly
visible for only a short time. The moving apart of the chromo-
somes takes place rapidly—so rapidly in fact that themovement
may be seen. Four minutes after their arrangement in theequa-
torial plane, they are well separated, and form the typical group-
ings at the poles. Five minutes later the rounded nuclei are
formed and fifteen minutes later the nuclear outlines are lost
and the nuclei can no longer be distinguished in the living cell.
The remainder of the cell with all of its processes shortened and
some entirely withdrawn has a clear, glassy appearance during
division. The concentration of the protoplasm in the neighbor-
hood of the nucleus is most marked at the time when the chromo-
somes are moving apart. As soon as the chromosomes are well
separated, the cutting in of the protoplasm commences. The
pear-shaped body of the cell lengthens to a cylindrical shape,

®> Haidenhain, Plasma und zelle 2e. Lief. pp. 721 and 722 in Bardelebens Hand-
buch der Anatomie des Menschen, Jena, 1911.

GROWING LYMPHATICS AND THE MESENCHYME 3/71

and, at the center, a ring-like depression occurs. ‘This increases
quickly in depth until, in three or four minutes, the two daughter
cells are connected by only a narrow process (fig. 14, #). At
this time the retraction of the processes is most pronounced,
while the ones not entirely retracted are extremely thin. Soon
however, what seems like a relaxation of the protoplasm occurs,
and the two daughter cells expand and commence to send out

C. 1:36 RM. DE P37 PINE

A.12:21PM. B :21PM.

E tSORM F 1:59PM. G. 5:00 P.M @pprox)

Fig. 14. Drawing to illustrate the mitotic division of a mesenchyme cell,
from same larva as shown in figs. 2 to 13; Cand Dare incomplete. This division
was observed on May 22. The position of the two daughter cells on the follow-
ing day, May 23, is shown in fig. 11, Z1 and Z?.

processes. For some time a slight process connects the two cells,
but after about two hours, they are completely separated. The
sending out of new processes takes place rapidly. For a time after
the disappearance of the nuclei the two new cells have a -finely
granular appearance, which distinguishes them from the other
cells. This is gradually lost, and eventually the new cells take

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 3

Si2 ELIOT R. CLARK

on the character of the non-dividing cells (fig. 14 G) and slowly
wander away from each other. Cells Z! and Z?, fig. 11, show
these same two daughter cells on the day following.

The position of the cell in the fin at the time of division is of
interest. A study of fig. 12, in which the position at the time of
division is indicated by the point at which the line showing the
course of the cells forks, shows that more divisions occur in
the thicker portion, of the fin near the muscle than in the thin
part near the edge. Here cells 2, 3, 4, 5, 6, 7, 8 and LZ remain
undivided. Of the cells in the thicker portion, those in the inte-
rior appear more likely to divide than those near the surface—
the cells A,-P, U, R, S and V remain undivided. In all, nearly
one-half of the cells observed underwent division during twenty-
three days. None of the daughter cells arising from division,
underwent further division during this period.

As a summary of the results of this series of observations on
the growing mesenchyme cells the following may be said. All
the mesenchyme cells in a selected strip of the fin expansion extend-
ing from muscle edge to fin margin were carefully watched and
~ recorded from day to day over a period of twenty-three days.
During this time the larva increased in length from about 5.5
mm. to 8.5 mm., while the width of the ventral fin expansion
increased from .38 to .50 mm., at the portion selected. It was
possible to identify each cell throughout this period, and tostudy
its growth changes. ‘The cells were found to possess two striking
properties, those of amoeboid movement and of mitotic division.
Every mesenchyme cell that was present in the area selected at
the beginning of the observations, as well as every cell which
wandered into this area from between the muscle layers or from
the more anterior portion of the tail during the observations,
maintained always its indentity asa mesenchyme cell. Although
the cells were actually watched for hours at a time, during periods
of active growth changes, and thoughanumber of mitotic divisions
were watched throughout, there was never the slightest indica-
tion of the transformation of one of these mesenchyme cells into
a cell of another type. The mesenchyme cell remained mesen-
chyme cell. The two daughter cells quickly resumed the typical

tROWING LYMPHATICS AND THE MESENCHYME ale

form and the characteristic power of amoeboid movement of the
non-dividing mesenchyme cell. There was nothing whatever to
indicate that either the non-dividing or the dividing mesenchyme
cell may give off, by bodily transformation or by budding, a leuco-
cyte or alymphatic endothelial cell. The mesenchyme cell, then,
during this period of growth, has a specific, independent life.

Let us now turn our attention to the growing lymphatic. To
former observations, which have been confirmed in the present
study, some new observations have been added which make the
story more complete. The newer work has been concerned par-
ticularly with a more careful study of the nuclear areas.

The nuclear areas occur at somewhat irregular intervals along
the wall. Whenseen in profile, they consist of a thick clear cen-
tral portion, surrounded by a granular zone, which extends for
a considerable distance longitudinally along the lymphatic in
both directions—gradually growing thinner. The term nuclear
area is used because the nucleus which may be clearly seen by
fixation and staining, merges imperceptibly into the perinuclear
area in the living larva (fig. 15). When seen en face the clear
central portion is invisible, but the position of the nuclear area
is clearly indicated by the granules surrounding the nucleus, so
that it is possible to keep track of all the nuclear areas in a sprout.

If a study is made of the behavior of the nuclear areas itis
found that there are two distinct processes going on, wandering
and mitotic division. These will be taken up separately.

The newly formed sprout receives its first nuclear area from the
parent stem by an in-wandering along the wall. As the branch
grows, it may receive a second and a third nuclear area from the
the main stem. How many may wander into a branch cannot
be stated; I have watched as many as five pass into a branch,
(cf. fig. 16, nuclei 4,5, 6 and 7). Once in the branch they do not
remain at rest, for they are continually moving up or down the
lymphatic, or from side to side, along a spiral course (fig. 16).
They have a tendency to arrange themselves in pairs on opposite
sides of the lymphatic (fig. 16, nuclei 3 and 4). They may how-
ever remain single, or from groups of three or four. In one in-
stance, in a sprout which was under observation, four nuclear

374 ELIOT R. CLARK

areas grouped themselves so closely together near the tip of the
sprout that they could hardly be recognized individually, until
they separated again. The nuclear areas do not maintain the
same relative position in a branch, for one may actually move

Fig. 15 Drawings of lymphatic in the tail of hyla pickeringii larva, to show
the relation between the nucleus and the rest of the nuclear area. A was drawn
while the embryo was in alcohol, unstained. B shows the same lymphatic, while
the embryo was still in alcohol, after staining with hematoxalin. The large black
dots represent pigment granules. Oil immersion.

past another (compare in fig.16 the relation between nuclei 2°
and 1» in stages J, K and L).

The nuclear areas which have once passed into a branch, do
not necessarily remain in that branch. They may pass into the

GROWING LYMPHATICS AND THE MESENCHYME oo

base of the branch for a short distance and then retreat, to con-
tinue their movement along the main stem. (fig. 16, nuclei
24 and 2» in stages F to J). Moreover, often in the developing
lymph system, a branch which is quite active at one time, is
apparently found to be superfluous later. It then becomes
narrower and shorter and the nuclear areas, which may be present,
retreat to the parent stem.

These nuclear movements furnish strong evidence for the view
that the endothelium of the new lymph sprouts is a syncytium.
That one nucleus could move past another and that such changes
in relative position as are shown in fig. 16 could take place, if
there were definite cell boundaries seems hardly conceivable.
Many observers have tried in vain to find, by the use of silver
nitrate, outlines of endothelial cells in the lymphatic sprouts in
the tail of the frog larva. I have injected them directly with
silver nitrate, but have failed to find endothelial markings. More-
over, the study of the wall of these capillaries in toto, in alcohol,®
as well as in sections stained with intense protoplasmic stains
such as acid fuchsin, reveals a network of fibrillae which connect
neighboring nuclear areas, thus forming a syncytium.

The other striking property possessed by the nuclear areas,
which has been observed in the living larva, is that of mitotic
division. Here, as in the case of the mesenchyme cells described
above, the picture was unmistakable (fig. 17). The granules
disappear, as the central portion becomes clear and spindle-
shaped. As in the dividing mesenchyme cells spireme formation,
arrangement of chrosomes, formation of daughter nuclei, and the
cutting apart of the remainder of the cell are to be seen with sur-
prising clearness. The chromosomes separate so rapidly that
the actual movement may be seen. The cutting-in of the cell
also takes place rapidly—from one to two minutes only being
consumed. After division the two daughter nuclear areas move
apart and maintain for a time a characteristic shape. Later

5K. R. Clark, An examination of methods used in the study of the development
of the lymphatie system, Anatomical Record, vol. 5, no. 8, 1911, p. 403, fig. 1,
and p. 406.

376 ELIOT R. CLARK

@ Pp, = On A
Tm: y aS Yoo a > :
APR. 27-4:00PM “ss Ny)
j « i APR. 22-3:00PM|

Wy APR 26-2 30PM.

\Ye
3D 23H 30M APR.23-10 SOAM. \

19H.50M.

APR.23-520PM \*s
ID 1H 20M

4 wy +b
APR 24-1130AM
1D 20H.30M

APR. 25-3:45 A.M.
20D 12H45M

<oTFs

,

A\{ APR.25-2:15 " / a
‘ 20 of St APR. 24-l0:1I5 PM. \\ \
. ZUP ApeloiMe

Fig. 16 Successive drawings of the same growing lymphatic sprout in the
dorsal fin expansion of the tail of a hyla pickeringii larva, to show the move-
ments of the nuclear areas. Each of the numbers, 1 to 7, indicates the same
nuclear area, which is dotted, in successive observations. In case of the division
of a nucleus, the daughter areas are indicated as 1* and 17>. Enlarged 180 times.

GROWING LYMPHATICS AND THE MESENCHYME She

Fig. 17 Lymphatic sprout from the hyla pickeringii larva on which studies
in figs. 1. to 14 were made, showing mitotic division of a nucleus. This division
took place on May 5, while the larva was anesthetized with chloretone. The num-
bers show the time. From 9:18 to 9:44 inclusive, the dividing nuclear area alone
is shown. Enlarged approximately 450 times.

378 ELIOT R. CLARK

they gradually resume the character of the non-dividing nuclear
areas. During the division the wall of the lymphatic remains
continuous.

Mitotic divisions have been observed in the main caudal lymph
trunk as well as in the branches. ‘There seems to be no definite
proportion between the number of nuclear areas a branch may
receive by in-wandering as compared with the increase by divi-
sion. I have seen the first nuclear area in a sprout divide, one
of the daughter nuclear areas remaining, and the other passing
back to the main stem. On the other hand, I have seen as many
as five nuclear areas wander into a branch.

Thus it is found to be possible to keep an accurate account of
all the nuclear areas in a growing lymphatic sprout, from its very
beginning. When this is done it is found that the sprout receives
its nuclear areas in two ways—by the in-wandering along the wall
of nuclear areas, which were present in the parent stem, and by
mitotic division of those which have wandered in. Moreover all
the nuclear areas which have been watched remain as nuclear
areas of the lymphatic. I have not seen them form leucocytes,
‘red blood cells, or mesenchyme cells. The sprout, then, receives
its nuclear areas from the preexisting endothelium. That the
protoplasm of the sprout is also derived from the preexisting
endothelium was shown in an earlier paper.’

What, now, are the relations between the growing lymphatic
and mesenchyme cells, which abound in the region which the
sprout is invading? In all my studies it has been observed that
the lymph sprout is so far from adding to itself these cells that it
actually remains at as great a distance as possible from the main
bodies and larger processes of the mesenchyme cells. At the tip
of the sprout, long fine processes are sent out by the lymphatic
in various directions. Occasionally such a thread pushes toward
or even to the body of the mesenchyme cell. But in all instances,
according to my observations, this process is eventually with-
drawn and the path selected is, as said, midway between these
cells (ef. the relations between lymph sprout and cell J, in figs.

7E. R. Clark, 1. ce. 1909.

GROWING LYMPHATICS AND THE MESENCHYME 379

4,5,6and7). It might be said that the mesenchyme cells keep
the lymphatic at their finger tips, or that the lymphatic avoids
the mesenchyme cells. I have seen a mesenchyme cell approach,
in its wandering, a lymphatic lying across its course. Processes
were sent out by themesenchyme cell on both sides of the lymphatic
so that the lymphatic lay within the U. In this case the processes
of the mesenchyme cell remained away from the wall of the lym-
phatic, and the cell moved past on one side, gradually withdraw-
ing the other arm of the U.

In summing up the results of these studies, there are certain
facts on which I desire to lay especial emphasis. It will be
recalled that the purpose of the investigation was to obtain a
complete history of the growing lymphatic sprout and of the
growing mesenchyme cell individually and side by side, in order to
determine their relationship to one another—to find out whether
the growth of the lymphatic is brought about by the addition
of mesenchyme cells or of spaces lined by transformed mesenchyme
cells, or whether the growth of the lymphatic is mdependent of
the mesenchyme cells. The observations have furnished answers
to these questions which are perfectly clear. Each of the two
tissues has a characteristic independent life. The mesenchyme
cell wanders and increases by mitotic division. It maintains
throughout its indentity as a mesenchyme cell, and is not trans-
formed into lymphatic endothelium. The lymphatic grows by
the sending out of fine protoplasmic processes which become
definite lumen-containing sprouts. Nuclear areas in the sprout
are provided by the in-wandering of nuclear areas from the main
stem, and by mitotic division. New lymphatic protoplasm and
nuclei, therefore, are formed by the extension of preexisting lym-
phatic endothelium. In its peripheral growth the lymphatic
endothelium is not formed by the transformation of mesenchyme
cells or blood vessels, nor does it give rise to mesenchyme cells or
blood vessels, it is a specific independent tissue. Throughout
its growth, the endothelial wall of the lymphatic capillary is
closed, there are no open communications between the lumen of
the lymphatic capillary and mesenchyme spaces.

ae ' gy a

x

r ‘

a ws inn: tae

, pads,
yo eh one 4 ys Gas naar

i

ee ie Nik a OU ae
ORE «aie
aire aN nad ld ah:
ouarny S oy eT iy

’

<b ried
Ba)

Sate ye" je Er
eb td ost

ast ette,
ree tiie
ty edt nae:
a) ve
A i. a ne
hi na othe wa oy, a Aes

‘i Pi 4 iy rie
es

THE ATTACHMENT OF MUSCLES TO THE EXOSKELE-
TON IN THE CRAYFISH, AND THE STRUCTURE
OF THE CRAYFISH EPIDERM

HAL DOWNEY

From the Histological Laboratory, Department of Animal Biology, University of
Minnesota

FIVE FIGURES

According to Leydig, Janet, Hecht and Holmgren the striated
muscles of Arthropods may be inserted without an intervening
tendon. In such cases the muscle fibrils are continuous with the
fibrous spongy network of the epithelial cells (Leydig), or the
muscle is attached directly to the chitinous exoskeleton, the hypo-
dermal epithelium being absent at the point of attachment (Janet,
Hecht, Holmgren). The usual mode of attachment, however,
is by means of an intervening tendon composed of straight fibrils
whose length is equal to the thickness of the hypodermal layer
within which they are located. The tendon fibrils are so numer-
ous that little remains of the original hypodermal cells but a
more or less distorted nucleus surrounded by a very small quan-
tity of cytoplasm. Without ontogenetic studies it is, of course,
almost impossible to determine whether these fibrils are within
the hypodermal cells or between them. The former view is the
one accepted by most authors, but Frenzel, Nicholas, Ide, Pantel
and Hecht claim that the tendon fibrils are between the hypoder-
mal cells. This is undoubtedly true in such cases as Holmgren
describes for some of the Diptera, where the tendon is divided
into several fine bundles of fibrils which pass between or spin
around large epithelial cells having definite cell boundaries.
However, this condition must be very exceptional, as most authors
describe the hypodermis in the region of muscle insertion as being

381

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 4
SEPTEMBER, 1912

382 HAL DOWNEY

composed largely of straight, stiff, coarse fibers which are so
numerous that little remains of the original hypodermal cell
except the nucleus, and cell limits are entirely obliterated.

A basement membrane or connective tissue ‘Grenzlamella’
between the hypodermal cells and the muscle fibrils, at the point
of muscle insertion, is described by Claus, MeMurrich, van Rees,
Schneider, Holmgren and many other authors. Lécaillon, Sneth-
lage and Riley also describe a basement membrane for the hypo-
dermal cells, but in the region of muscle attachment it bends
around and becomes continuous with the sarcolemma, and there-
fore it does not extend between the epithelium and the muscles.
Claus states that it is a cuticular formation in Branchipus and
Artemia, and that it may become chitinous. In some regions
it is absent. In the Decapods, however, it is a true connective
tissue ‘Grenzlamella.’ MceMurrich and Holmgren also claim that
it is merely a cuticular membrane formed by the epithelium, the
former describing thickenings of the membrane (in terrestrial
Isopods) at the cell boundaries from which supporting fibers
originate and pass up into the cells. van Rees published figures
which show that in Musca vomitoria the basement membrane
is formed from long processes of the hypodermal cells which sur-
round the muscle insertion. According to Emmel the first base-
ment membrane in the regenerating claw of the lobster is a
homogeneous cuticular formation of the ectodermal cells. A nu-
cleated membrane is formed later, the origin of which Emmel
was unable to determine. The intermediate membrane is of con-
nective tissue origin according to Schneider, and Claus states
that it is a connective tissue formation in the Decapods.

Claus found the basement membrane to be absent from some
regions of Branchipus and Artemia, and Henneguy never saw
it in the Insecta and claims that it is not present in other Arthro-
pods, but that the Z lines of the muscle fibers give the appearance
of a basement membrane. There is no basement or intermediate
membrane in the Millipedes according to Duboseq. Snethlage
finds no intermediate membrane at the point of insertion, but
pigment granules and anastomosing muscle fibrils give the appear-
ance of a basement membrane.

MUSCLE ATTACHMENTS AND EPIDERM IN CRAYFISH 383

The muscles are attached to the epithelial cells and not directly
to the tendon fibers according to List, Viallanes, Vitzou and
Weisman. Stamm declares that the muscles are inserted to the
basement membrane and that there are no fibrils passing through
the membrane. Bertkau (in spiders) finds muscle fibers to be
continuous with the tendon fibers, but he can always see a distinct
dividing line between the epithelium and the muscle.

Leydig, Duboseq and Snethlage declare that the sarcoplasm is
continuous with the protoplasm of the ectodermal epithelial cells,
but this is denied by Holmgren, excepting for the vagina of Sar-
cophaga, where he found direct continuity between chitin matrix
cells and the sarcoglia of muscles. This is explained as being the
persisting embryonic connection between ectoderm and mesoderm
which was described by Heatcote.

Tendon fibers and muscle fibers are directly continuous with
one another according to Claus, Duboseq, Leydig, Lécaillon and
Henneguy, but Riley declares that there is a splicing or fusion of
the muscle fibers into the tendon fibers and no direct continuity
between the two.

Many authors believe that the tendon fibrils are merely pro-
longations of the muscle cells. Frenzel, Nicolas, Ide, Pantel,
Hecht and Holmgren locate the tendon fibers between the epithe-
lial cells instead of within them. As the muscles approach the
epithelium they are resolved into fine bundles of fibrils which pass
between the epithelial cells to be inserted directly into the chitin.
According to another scheme of Holmgren’s the muscle fibrils
penetrate the epithelial cells and are then inserted to the chitin.
Nowikoff and Snethlage believe that both tendon and muscle
are developed from ectoderm, so the continuity of the two types
of fibrils is easily explained. Emmel found that in the regenerat-
ing lobster claw the new tendon fibrils are merely the non-striated
ends of the muscle fibrils which are developed in an ectodermal
syncytium. The striations may continue as far as the level of
of the chitin-forming outer later of this syneytium.

The tendon fibers are described by Claus as being very coarse,
while Henneguy states that they are identical in appearance
with fibrils in other parts of the epidermis where there are no

384 HAL DOWNEY

muscles. They correspond to M. Heidenhain’s ‘Tonofibrillen’
which are formed in many epithelial cells, according to Maziarki
and Labbé, and Janet observed that they may become chitinous.
Léeaillon’s observations show that a tendon may divide into two
parts and that two tendons may fuse.

The tendon fibers are developed from the ectodermal epithelium
according to Claus, Henneguy, Leydig, Duboseq, Maziarki, Labbé,
Janet, Snethlage, Reichenbach, van Rees, Emmel, Riley and
Bertkau, but Tullberg and Braun derive them from the connec-
tive tissue. Most of those authors who believe that the tendon
fibers are prolongations of the muscle fibrils also believe that the
muscles are products of the mesoderm, which compels the con-
clusion that the tendon fibers are also of mesodermal origin. This
view, however, is not shared by Nowikoff and Snethlage, nor by
Reed, Ost and Emmel. All of these authors derive the muscles
and the epidermis from the ectoderm, the three last named basing
their conclusions on a study of regenerating tissue.

A study of the development of the muscles has been made by
Henneguy (insects), Reichenbach (crayfish) and vanRees(insects).
The work of these authors shows the muscles to be of mesodermal
origin, while the investigations of Snethlage (Artemia salina) fur-
nish equally convincing proof for the ectodermal origin of the
muscles in normal ontogeny. Reed, Ost and Emmel have shown
that the regenerating muscles are of hypodermal origin. Henne-
guy studied insect embryos and found that the muscles were
developed from mesodermal cells which were in contact with the
hypodermis. Epithelial fibrils were differentiated at the same
time that the muscle fibrils were formed, and the two became con-
tinuous. Snethlage’s studies on developing Artemia muscles con-
vineed him that they are formed from the ectoderm at the point
of insertion. Development from the mesoderm would be impos-
sible, because a mesoderm in the sense of a distinct germ layer
does not exist in these forms. van Rees worked on the metamor-
phosis of Musea vomitoria and found that the basal portions of
the hypodermal cells become spun out into fine processes which
push the basal membrane away from the epithelial cells. ‘These
fine processes become the muscle tendons. Reichenbach con-

MUSCLE ATTACHMENTS AND EPIDERM IN CRAYFISH 385

cluded from his studies on the development of the crayfish that
the tendon fibrils are formed from the ectoderm, but that the
muscles are developed from mesoderm. Emmel has shown that
in the regenerating lobster claw the muscles and hypodermis are
developed from an ectodermal syncytium. The tendon fibrils
are the non-striated ends of the muscle fibrils. The basement
membrane which is formed at first is also a product of the ecto-
dermal syncytium.

The above gleanings from the literature show that the problem
of muscle attachment in Arthropods can hardly be regarded as
settled. We find a variety of opinions even on such a question
as the exact mode of attachment of the muscle fibrils to the tendon
fibrils, the solution of which depends chiefly on good technique
and careful observation of sections of adult tissue. It seems,
therefore, that any new observations in this field may be of value.

The following account of muscle attachment in the crayfish is
based on a study of material from the histological collection of
the department of Animal Biology of the University of Minnesota.
The sections were made through the region of muscle attachment
in the claw of a crayfish after the chitin had been stripped off.
The fixation of the material is good. The sections are stained in
a haematoxylin combination.which gives a splendid differentia-
tion of the finer cell structures.

The epiderm or hypodermis, as it is usually termed by entomolo-
gists, of the large claw of the crayfish may consist of one (figs.
4 and 5) or several (fig. 1) layers of extremely irregular cells
which are so closely associated with one another that it is usually
quite impossible to distinguish cell boundaries. A study of sections
through regions located between the points of muscle attachment
(figs. 1, 2, 4) shows very clearly that there are no cell boundaries,
and that the hypodermal layer is composed of a protoplasmic
syncytium in which the nuclei have a very irregular distribu-
tion, a condition which was described by Henneguy for various
Arthropods and by Emmel in the regenerating claw of lobster.
In some regions (fig. 1) the presence of small and large irreg-
ular spaces gives the syncytium the appearance of an extremely
irregular protoplasmic network the strands of which are extremely

386 HAL DOWNEY

variable in size and shape. Vacuoles and small spaces are nearly
always present, even in the thinner portions of the hypodermis
(figs. 1 and 4). The nuclei all have about the same relative
amount of chromatin, but their size and shape is subject to great
variations. They may be round, irregular or oval, and their long
axis may be in any plane. Very large nuclei are shown in figure
3 behind the tendon fibrils and at the right of figure 4, and figures
1 and 2 show the irregularities in the shape of the different nuclei.

Everywhere the hypodermis contains may sharply defined;
homogeneous fibrils which take a wavy course theough the syney-
tium, usually extending through several cell-territories. Sometimes
they unite into bundles (fig. 1), but usually each fibril is indepen-
dent. They show a general tendency towards an oblique course
through the syncytium from its outer to its inner surface, or they
may converge on the muscles, as is seen in figure 1, where the
fibrils of the syncytium are all directed towards a small area
on the bundles of striated muscle fibers shown in the lower right-
hand portion of the figure. Occasionally, the fibers will run for
some distance in the horizontal direction parallel to the outer sur-
face, before they take the oblique course downwards to join the
fiber bundles in the basal region of the syneytium (fig. 4). Figure
4 also shows considerable variation in the size and direction of the
fibrils. In the upper right hand portion of the figure one sees many
exceedingly fine fibrils whose course is more vertical than that of
the other fibers. Some of the thicker fibers permit a considerable
change of focus before they pass out of view, which indicates
that some of them are bands or membranes and not fibers.
Figure 2 shows a very interesting arrangement of the hypodermal
fibers. If we examine the bundle of fibers in the basal portion of
the hypodermis we notice that one fiber leaves this bundle, and
if we follow it to the left we see that it gradually gets further
away from the basal bundle until it very suddenly changes its
direction. It passes upwards, bends around a hypodermal nucleus
and can then be followed back to the right in the horizontal plane,
where it is joined by several other fibers running in the same direc-
tion, some of which again pass downwards to join the basal
bundle. That portion of the hypodermis which is some distance

MUSCLE ATTACHMENTS AND EPIDERM IN CRAYFISH 387

away from muscle attachments may contain a great many hori-
zontal fibers, some of which pass downwards to join the basal
bundle of fibers (fig. 4). Others unite with vertical and oblique
fibers which results in the formation of a network within the
syneytium.

Supporting fibrils in the hypodermis are mentioned by Henne-
guy, Maziarki, Labbé and Emmel, Maziarki and Labbé believing
that they correspond to M. Heidenhain’s ‘'Tonofibrillen’ which
are formed in many epithelial cells. In insects Lécaillon could
see no other fibers in the epidermal epithelium but the tendon
fibers which are continuous with the muscle fibrils. MeMurrich
found supporting fibrils passing up into the epithelial cells of the
mid-gut of terrestrial Isopods from thickenings of the basement
membrane.

In the crayfish the supporting fibrils of the hypodermis show
a tendency to collect in the basal portion of the syncytium to
form. a horizontal layer of fibers which, when composed of many
fibers not clearly differentiated by good technique, may have the
appearance of a basement membrane. However, a good haema-
toxylin stain will show exactly what this ‘basement membrane’
is. Figures 1, 2 and 4 show clearly that most of the supporting
fibrils of the hypodermis eventually reach the basal portion of
the syncytium where they take a horizontal course. In some
regions (fig. 1) the horizontal fibers form a dense bundle at the
inner margin of the syncytium and therby produce a definite limit-
ing structure between the hypodermis and the underlying tissues,
but in other cases (fig. 2) the basal fibers are not so closely asso-
ciated with one another as to form a fibrous membrane. Figure
4 shows that the basal fibers may be widely separated from one
another and that they may occupy the entire cell-territory between
the lower level of the nucleus and inner margin of the syncytium.
The same section (left-hand portion of figure 4) shows also that
there may be about as many horizontal fibers above the nuclei
as below them. In still other regions of the hypodermis horizontal
fibers are entirely absent and there is absolutely no limiting struc-
ture of any kind between the hypodermal syncytium and the
underlying connective tissue. In such places it is impossible to

388 HAL DOWNEY

determine the boundary between the hypodermis and the connec-
tive tissue, and the supporting fibrils are continuous from one
tissue to the other. In the deeper portions of the connective
tissue syncytium the fibrils pass to the margins of the syncytial
strands, thus entering into the structure of true Leydig’s cells of
the first order. The fibers or bands are continuous from one
Leydig cell to another but are always located at the margin of
the cells. Some of the fibers join the outer adventitial membrane
of the blood vessels the wall of which is composed of Leydig’s
cells which are so arranged that they form simple epithelial tubes
which have a very definite inner and outer membrane. The inner
membrane appers homogeneous, but it has the same structure
and staining reactions as the fibers of the hypodermis and support-
ing tissue. The outer membrane is frequently fibrous, its fibers
being continuous with those of the surrounding Leydig’s cells.
That the hypodermis is continuous with the underlying sup-
porting tissue is not a new observation. Leydig, in 1851, called
attention to the similarity between Arthropod integument and
connective substance and M. Braun, in 1875, states that it is
impossible to separate the products of the ectoderm from those
of the mesoderm in the crayfish, and many other authors have
made the same observation. In view of these facts it is not sur-
prising to find some of the fibrils from the supporting tissue Joining
the horizontal layer of hypodermal fibers (figures 1 and 2, of which
S. T. is supporting tissue). We must admit, therefore, that it
is quite possible that connective fibrils from the supporting tissue
take part in the formation of the fibrous limiting layer where such
a structure is formed but, on the other hand, it is also possible
that the fibers which we see running down into the supporting
substance are derived from the hypodermis. The presence of
occasional flattened oval nuclei among the horizontal fibers has
been interpreted by Schneider and others as proof for the existence
of connective tissue in this region. Further speculation on this
point is useless, especially so in view of the fact that we are by no
means certain of the mesodermal origin of the supporting tissue
or of the muscles in Arthropods. The fact remains, that in the
crayfish, the layer of horizontal fibers is composed largely of hypo-

MUSCLE ATTACHMENTS AND EPIDERM IN CRAYFISH 389

dermal fibers. There is no cuticular basement membrane, such
as is described by McMurrich, Holmgren and Emmel, and no
connective tissue ‘Grenzlamella’ as stated by Claus and Schneider,
although fibers from the supporting tissue may enter into the
fibrous layer in the basal portion of the hypodermis.

Where the layer of basal fibers is well developed it may pass
between the muscle and tendon fibrils in the region of muscle
attachment (figs. 1 and 2), or it may bend downwards and run
along the surface of the muscle or continue for some distance
between the muscles (fig. 3). A horizontal layer of fibers between
the muscles and their tendon fibers may also be produced by con-
tinuous branches of the tendon fibers themselves (fig. 5). This
is probably what Snethlage saw when he stated that between the
muscles and their tendon fibrils an apparent basement membrane
is produced by anastomosing muscle fibrils and pigment granules.
Small, highly refractive pigment granules are also found in the
crayfish in this region, but they are seen only when the light is
cut down. For the sake of clearness they have been omitted
from the drawings.

Figures 1 and 4 show that bundles of fibers are not limited to
the basal layer of horizontal fibers. At the outer border of the
hypodermis these bundles are resolved into their individual fibers
which spread out fan-shaped and probably penetrate the chitin.
In the present investigation no attempt was made to determine
the relation of the fibrils to the chitin, but many authors state
that they are continued into the chitin.

Close study of the tendon fibers (7. F., figs. 1, 2, 3, 5) shows
that they are coarse, straight bundles of fibrils which seem to be
identical with those which are found in other regions of the hypo-
dermis. ‘This is in accord with the findings of Henneguy, but
is contrary to Emmel. The outer ends of the tendon fibers, where
they are attached to the chitin, are frequently spread out in such
a way that they form a fan-shaped figure in sections. This is
seen especially well in the isolated smaller bundles of fibers such
as are shown in figure 1, 7’. F.’ and on the left of figure 2. Here
one sees that the fibrils correspond in size, structure and staining
reactions to the supporting fibrils in other regions of the hypoder-

390 HAL DOWNEY

mis. The tendon fibers stain darker than the smallest fibers
found in other parts of the epiderm, but not any darker than
thick bundles of fibers (fig. 1). Where the tendon fibers are
frayed out at the ends the staining reactions of their fine consti-
tuent fibrils are identical with those of fibrils in other parts of
the hypoderm. It is probable, therefore, that the dark staining
of the tendon fibers is due to their density, rather than to changes
in their chemical nature.

A condition similar to that described above is seen also at the
inner ends of the tendon fibers, where they are attached to the
muscles (figs. 1, 3,5). The fine inner branches resulting from the
division of the coarser fibers into their constituent fibrils may fol-
low a comparatively straight course, as in figure 3, or they may
anastomose freely with neighboring fibrils at the lower border of
the tendon, as in figure 5. The latter condition may produce a
more or less horizontal layer or network of fibers composed of anas-
tomosing branches of the tendon fibers (fig. 5). Such a structure
might easily be mistaken for a basement membrane. Fibrils
from the surrounding epidermal regions, and probably also from
the supporting tissue, usually join with the branches of the tendon
fibers in the formation of this layer (fig. 1). However, this is
probably not the case in the region shown in figure 5, and certainly
not in the parts shown in figure 3. The latter figure shows that
a horizontal layer of fibers may be absent at the point of attach-
ment of the muscle to its tendon, but even in this case there is
more or less branching of the finer fibrils at their inner ends.

The exact mode of attachment of the muscles to their tendons
is shown in figures 1,3, and 5. In figures 1 and 3 we see that the
fine branches of the tendon fibers are continued down into the
muscle for a variable distance, and that the muscle fibrils run up
between these processes. Figure 1 also shows that fibers from the
horizontal layer which are derived from other regions of the hypo-
dermis may penetrate the muscle for some distance. The long
fiber on the right of the figure is probably derived from this source.
There are no horizontal fibers in the greater part of figure 3, and
therefore the muscle attachment here is entirely by means
of fine branches of the tendon fibers. In figure 5 most of the fibrils

MUSCLE. ATTACHMENTS AND EPIDERM IN CRAYFISH 391

which project down into the muscle are derived from the network
which is produced by anastomosis of tendon fibrils at the base of
the tendon. Some of the fibrils pass down into the muscle for a
considerable distance. The: ‘splicing’ or ‘dove-tailing’ of the
muscle fibrils and tendon fibrils is seen very clearly in this figure
and in figure 1.

The findings of previous workers in regard to the union of tendon
and muscle in Arthropods have been presented in the first part
of this paper and it is unnecessary to restate them here. Riley
seems to be the only author who has observed ee like
the process described here. According to Riley, “‘
there occurs a slicing or fusion of the two types of fibrils, the Huset
ment membrane being lacking at the point of contact.”’ If the
network or layer of fibers at the base of the tendon and in the lower
portion of the hypodermis is to be interpreted as a basement
membrane, then Riley’s statement in regard to this structure is
not correct for the crayfish, as is seen by an examination of figures
1 and 5. However, the writer can see no reason for calling this
structure a basement membrane. ‘The variations to which it is
subject are shown in the five figures presented here, and it hardly
seems reasonable to call such an indefinite structure a basement
membrane.

For the crayfish it seems almost certain that the tendon fibers
are within the hypodermal cells and not between them, as is
claimed by Frenzel, Nicolas, Ide, Pantel and Hecht. The fact
that the tendon fibers seem to be bundles of ordinary supporting
fibrils is in favor of the former view, and figures 1, 3 and 5 can be
interpreted in no other light. The fibers occupy the greater part
of the cell, but a small quantity of cytoplasm can always be seen
between them, and the original hypodermal nuclei are still to be
seen. The final solution of this question will depend on ontogene-
tic studies, but for the present the view presented here seems to
be the most reasonable one.

The origin of the muscles in the Arthropoda is still being debated,
some authors claiming that they are of mesodermal origin, others
that they are derived from ectoderm. It is very likely that they
may be derived from both sources, which is probably also true

392 HAL DOWNEY

of some vertebrates, Goronowitsch having shown that in birds
the musculature of the visceral arches is derived from cells of the
neural crest and from the outer ectoderm. Julia Platt does not
admit the development of muscles from wandering extoderm cells
in Necturus, but she does claim that the branchial cartilages and
the anterior portions of the trabecular bars are developed from
ectoderm, while Goronowitsch claims that in birds a portion of
the mesenchyme, cutis and skull in the region of the midbrain
are formed from ectoderm. According to Kastschenko the mesen-
chyme is derived from all three germ layers, from cells which
are not. used up in the formation of epithelial structures, and
v. Kupffer derives the branchial cartilages of Petromyzon from
the deeper layers of the ectoderm.

Facts like these show that we must be careful in our state-
ments regarding the origin of various structures from certain germ
layers until we have exact ontogenetic studies of those structures,
and this is especially true of the invertebrates.

The peculiar arrangement which provides for union between
muscle and tendon fibrils in the crayfish suggests independent
origin and secondary fusion of those structures, but the ontogene-
tic proofs for this are lacking. In any case, the method of union
between the two types of fibrils remains quite different from what
it isin vertebrates. In the latter there is direct continuity between
muscle and tendon fibrils, as has been recently proven by Oskar
Schultze! for man and several groups of vertebrates. His results
have been confirmed by other investigators (see Schultze’s Leipzig
paper for further literature on the subject), and the writer can
testify that his preparations show all that is claimed for them.
That many authors believe that the same conditions obtain .in
Arthropods has already been pointed out. However, the present
investigation does not warrant this conclusion for the crayfish.
Here the ends of the muscle fibrils are surrounded by fine branches
of the tendon fibrils, some of which may pass in between the muscle
fibrils for a considerable distance.

1 Verhdl. der Phys.-Med. Gesellsch. zu Wiirzburg N. F. Bd. 41 and Verhdl.
der Anat. Gesellsch, 25. Vers., Leipzig, 1911.

MUSCLE ATTACHMENTS AND EPIDERM IN CRAYFISH 393

~ Coagulated blood plasma is found in some of the spaces of the
hypodermal syncytium and in occasional spaces which occur in
the layer of horizontal fibers at the base of the epiderm. This
‘ indicates that the vascular circulation is in very close relation
with the hypodermis.

SUMMARY

The hypodermisof the crayfish consists of a protoplasmic syncy-
tium containing one or more layers of nuclei which have a very
irregular distribution within the syncytium.

Supporting fibers and fibrils are found everywhere within the
syncytium. In some regions they all take about the same course
and eventually reach the basal portion of the syncytium where
they run in a horizontal direction.

The basal layer of fibers shows great variation in the number
of fibers entering into its composition and in their proximity to
each other. In some cases the basal fibers occupy the entire
territory between the nuclei and the lower border of the syncytium,
and in others this layer is absent altogether. On account of these
variations this layer of fibers can not be classified as a true base-
ment or intermediate membrane.

Fibers from the supporting tissue are continuous with those of
the hypoderm and some of them join the basal layer of hypodermal
fibers which, in the regions of muscle attachment, usually pass
between the ends of the muscle fibrils and the tendon fibers by
means of which the muscles are attached to the chitinous exoskele-
ton.

The tendon fibers are bundles of fine fibrils, similar to those
found in other regions of the epiderm. At their inner ends the
tendon fibers are resolved into their constituent fibrils which usu-
ally anastomose with neighboring fibrils in such a way as to pro-
duce a network at the base of the tendon. Fibers from other
regions of the hypoderm and from the supporting tissue frequently
join this network (fig. 1).

Tendon fibrils, and fibrils from the intermediate network or
basal layer of hypodermal and supporting tissue fibrils penetrate

394 HAL DOWNEY

the muscle for variable distances in such a way that the outer
ends of the muscle fibrils are surrounded by them.

Muscle fibrils and tendon fibrils are not directly continuous
with one another. The muscle is ‘dove-tailed’ or ‘spliced’ into‘
its tendon.

The evidence obtained from a study of the adult structures is
in favor of the view that the tendon fibers are located within the
hypodermal cells and not between them.

LITERATURE CITED

BertKatu, Ph. 1885 Ueber den Verdauungsapparat der Spinnen. Arch. f.mik.
Anat., Bd. 24.

Braun, M. 1875 Ueber die histologischen Vorginge bei der Hiutung von Asta-
cus. Arb. d. Zool. zootom. Instituts, T. 2, Wiirzburg.

Cuiaus, C. 1886 Untersuchungen tiber die Organisation und Entwicklung von
Branchipus und Artemia. Arb. aus d. Zool. Inst., Bd. 5, Wien.

Dusosce. O. 1898 Recherches sur les Chilopodes. Arch. Zool. exper. (3), T. 6.

Emmet, V. BE. 1910 A study of the differentiation of tissues in the regenerating
crustacean limb. Amer. Jour. Anat., vol. 10.

FRENZEL, J. 1885 Ueber den Darmkanal der Crustaceen nebst Bemerkungen
zur Epithelregeneration. Arch. f. mik. Anat., Bd. 25.

GoronowitscH, N. 1892. Die axiale und die laterale( A. Goette) Kopfmeta-

merie der Vogelembryonen.—Die Rolle der sog. ‘Ganglienleisten’ im
Aufbaue der Nervenstimme. Anat. Anz., Bd. 7.
1893 Untersuchungen iiber die Entwicklung der sog. ‘Ganglienleis-
ten’ im Kopfe der Vogelembryonen. Morpholog. Jahrb., Bd. 20.

Hecut, E. 1899 Notes biologiques et histologiques sur la larve d’un Diptére.
Arch. Zool. exper. (8), T. 7.

Hennecuy, F. 1906 Les modes d’insertion des muscles dur la cuticule sur les
Arthropodes. Compt. rend. de L’Assoc. des Anatomistes, 8. Réunion
Bordeaux.

HotmGRren, Nits 1901-02 Ueber das Verhalten des Chitins und Epithels zu den
unterliegenden Gewebearten bei Insecten. Anat. Anz. Bd. 20.

Ipe, M. 1892 Le tube digestif des Edriophthalmes. La Cellule, T. 8.

Janet, C. 1896 Etudes sur les Formis, les Guépes et les Abeilles, T. 12, note.

KastscHENKO, N. 1888 Zur Entwicklungsgeschichte des Selachierenbryos.
Anat. Anz... Bd. 3.

vy. Kuprrer,C. 1895 Ueberdie Entwicklung des Kiemenskelets von Ammocoetes
und die organogene Bestimmung des Exoderms. Verhdlg. der Anat.
Gesell.

Lassk, A. 1902 Sur la continuité fibrillaire des cellules épithéliales et des
muscles chez les Nebalia. C. R. Acad. des Sc., T. 135. ;

Lécartton, A. 1907 Recherches sur la structure de la cuticule tégumentaire
des insectes et sur la maniére dont s’attachent les muscles chez ces
animaux. Bibliographie Anatomique, T. 16.

MUSCLE ATTACHMENTS AND EPIDERM IN CRAYFISH 395

Leypic, F. 1851 Ueber Artemia salina und Branchypus stagnalis. Zeitschr f.
wiss. Zool., Bd. 3.

1885 Zelle und Gewebe.

List, Th. 1897 Morphologisch-biologische Studien iiber den Bewegungsappa-
rat der Arthropoden. 2. Teil. Die Decapoden. Mitth. der zool. Station
zu. Neapel, Bd. 12. ;

MazrarklI, St. 1903 Sur les rapports des muscles et de la cuticule chez les Crus-
tacés. Bull. Acad. des Sc. de Cracovie.

McMuraricu, J. P. 1897-98 The epithelium of the so-called midgut of terres-
tial Isopods. Jour. Morph., vol. 14.

Nicouas, A. 1898 Sur les rapports des muscles et des éléments epithéliaux dans
la pharynx du Péripate (Peripatus capensis). Rev. biol. du Nord de
la France, T. 2.

Nowrkorr, M. 1905 Untersuchungen iiber den Bau der Limnadia lenticularis
L. Zeitschr. f. wiss. Zool., Bd. 78. ;

Ost, J. 1906 Zur Kenntnis der Regeneration der Extremitiiten bei den Arthro-
poden. Arch. f. Entwgsmch. der Organism, Bd. 22.

PanTEL, J. 1898 Thrixion halidayanum Rond. Essai monographique sur les
caractéres extérieures, la biologie et l’anatomie d’une larve parasite du
groupe des Tachinaires. La Cellule, T. 15.

Pratt, Jura B. 1898 The development of the cartilaginous skull and of the
branchial and hypoglossal musculature in Necturus. Morpholog. Jahrb.
(Gegenbauer), Bd. 25.

Reep, M. A. 1904 The regeneration of the first leg of the crayfish. Arch. f.
Entwgsmch. der Organism, Bd. 18

vAN Rees, J. 1889 Beitraege zur Kenntnis der inneren Metamorphose von
Musca vomitoria. Zool. Jahrb., Abth. f. Anat. u. Ont. Bd. 3.

ReicHENBACH, H. 1888 Studien zur Entwicklungsgeschichte des Flusskrebses.
Abhand. der Senkenberg. Naturforsch. Gesellsch., Bd. 14.

Ritey, W. A. 1908 Muscle attachments in insects. Annals of the Entomolo-
gical Soc. of America, vol. 1.

ScHNEIDER, K.C. 1902 Lehrbuch der vergleichenden Histologie der Tiere. Jena.

SNETHLAGE, E. 1904-05 Ueber die Frage vom Muskelansatz und der Herkunft
der Muskulatur bei den Arthropoden. Zoot. Jahrb. (Anat. u. Ont.),
Bd. 21.

TULLBERG, 1882 Studien tiber den Bau und das Wachstum des Hummerpanzers
u. d. Molluskenschalen. Stockholm.

VIALLANES, H. 1882 Ann. Se. nat. Zool. (6), T. 14.

Virzou, A. 1882 Recherches sur la structure et la formation téguments chez les
crustacés Décapodes. Arch. Zool. exper. T. 10.

WetsMann, A. 1864 Die Nachembryonale Entwicklung der Musciden nach Beo-
bachtungen in den Musca vomitoria und Sarcophaga carnaria. Zeitschr.
f. wiss. Zool., Bd. 14.

All drawings were made with the camera lucida under Zeiss apochrom. obj.
2mm. and compens. ocular 6, drawing board at the height of the stand.

PLATE 1
EXPLANATION OF FIGURES

1 Striated muscle and its tendon on the right, thick layer of hypodermal
syncytium on the left. Notice the numerous supporting fibrils within the syncy-
tium, the irregular distribution of its nuclei and the irregular spaces. A hori-
zontal layer of fibrils in the basal portion of the syncytium passes between the
muscle and its tendon fibers. The figure shows that the fibrils of this layer are
derived from the hypodermis and also from the supporting tissue, S. 7. Tendon
fibrils and fibers from the horizontal layer penetrate the muscle for varying dis-
tances.

2 Shows an arrangement of supporting fibers in the hypodermal syncy-
tium which is quite different from that shown in figure 1. Horizontal fibers are
found both above and below the nuclei. The inner layer of horizontal fibers
occupies almost the entire cell-territory between the nuclei and the inner border
of the syncytium. This arrangement of the fibres shows that the basal layer of
horizontal fibers does not form a true basement or intermediate membrane.

396

MUSCLE ATTACHMENTS AND EPIDERM IN CRAYFISH PLATE 1
HAL DOWNEY

Helen A. Sanborn, del.

397

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, No. 4

PLATE 2
EXPLANATION OF FIGURES

3 Very long tendon fibers which show a rather loose arrangement. The
tendon fibers branch at their inner ends, the fine branches surrounding the ends
of the muscle fibrils. Large nucleus of the hypodermal syncytium back of the
tendon fibers. This figure also shows that the horizontal layer of fibers may be
absent at the point of attachment of the muscle to its tendon.

4 A clear demonstration of the fact that the basal layer of horizontal
fibers is composed largely of fibers derived from the hypodermal syncytium. In
this case the basal layer can hardly be interpreted as a true basement membrane
or as a connective tissue intermediate membrane. Horizontal fibers are seen
both above and below the nuclei. Those above eventually join the lower layer
which occupies most of the cell-territory between the nuclei and the inner border
of the epiderm.

5 Striated muscle and its tendon. This section is especially favorable
because the tendon fibers show a very loose arrangement which provides for a
clear demonstration of their relation to the muscle fibrils. The network of fibers
at the base of the tendon is formed by anastomosing branches of the tendon fibers.
Tendon fibrils and fibrils derived from the network penetrate the muscle between
the ends of the muscle fibrils. This arrangement shows that the muscles are
‘spliced’ or ‘dove-tailed’ into their tendons, and that there is no direct continuity
between the tendon and muscle fibrils.

398

PLATE 2

MUSCLE ATTACHMENTS AND EPIDERM IN CRAYFISH

HAL DOWNEY

Helen A. Sanborn, del.

399

=
i ’ (
b= ~~
.
q
ni , Soe
i
n : \
*
° ’ ‘
os .
- ¢ a ee
ms
:
\
i ae
be =
¥) nd 7 ct
' SA’ sse¢ a. owe
¥ >

THE DEVELOPMENT OF THE THORACIC DUCT IN
THE PIG

OTTO F. KAMPMEIER

From the Laboratory of Comparative Anatomy, Princeton University
THIRTY-FIVE FIGURES (FIVE COLORED PLATES)!

I. INTRODUCTION

At the present day the question of the origin of lymphatics has
become one of the most interesting problems and is holding the
attention of perhaps as great a number of investigators as any
other problem in the field of embryology and anatomy. This is
partly due to the fact that the lymphatic system as a whole is
the last of the organ-systems to be taken up for more thorough
investigation, and partly it is the result of the impetus given by
modern physiology which has emphasized the question of the
relationship between lymphatics and blood vessels and their func-
tional significance in the economy of the organism. Although this
problem has been attacked at various times in the history of
anatomy, relatively few important advances have been made,
and it is only recently during the last decade that it has been
attacked with renewed vigor and has been pushed to the critical
point of its solution. The question of the individuality of a
lymphatic, especially, has developed a most animated contro-
versy between two schools diametrically opposed in their conten-
tions. One of these maintains that all lymphatic channels are
a direct product of the venous system, and the other that they
arise independently of it, and that if they do enter into certain
structural venous relations during their genesis such relations are
purely of a secondary character.

1 Expense of illustrations partly borne by author.
401

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 4

402 OTTO F. KAMPMEIER

In 1909 Sabin? published a paper on the origin of the lymphatic
system in the pig which reinforced and extended the theory of
Langer? and Ranvier,‘ that lymphatic vessels arise from the veins
by a process of sprouting and centrifugal growth. Because her
results and conclusions did not present fully all of the evidence or
agree with the results of later investigations, Professor McClure
suggested to the writer the expediency of repeating her work on
the same kind of material, namely, pig embryos, but restricting
this research to one lymphatic channel and concentrating upon
it alone allattention. The thoracic duct was selected on account
of its size, importance, and definite position in the body, thus
serving better as a test or control than perhaps any other lymph
channel.

For the completion of this work, carried on during the last two
years, I am indebted to Professor McClure for his advice and
valuable criticisms, and to Dr. A. G. Brown of Columbia Univer-
sity and to Mr. Charles F. Silvester for their hints in the prepara-
tion of the microphotographs. I acknowledge my gratitude to
Professor Houser for permitting me to use the laboratory at
the Iowa State University during the summer of 1910. My
thanks are also due to Professor Sabin for the use of series 23a
of the Johns Hopkins University Embryological Collection, and
for the privilege of publishing any information or evidence which
might be derived from this series.

II. MATERIAL AND TECHNIC

Most of the embryos used in this work were fixed in Zenker’s
fluid and the sections stained on the slide with Delafield’s haema-
toxylin and orange-G. Sabin’s series 23a was preserved in the
same fixative but stained with Congo red. A few embryos were

2 Florence R. Sabin: On the origin of the lymphatic system from the veins, and
the development ‘of the lymph hearts and thoracic duct in the pig. Am. Jour.
Anat., vol. 1, 1902, pp. 367-389.

3C. Langer: Ueber das Lymphgefass-systems des Frosches. Sitzb. d. Akad. d.
Wissensch., Bd., 57, 1, Abth., 1868.

47, Ranvier: Morphologie et developpement des vaissaux lymphatiques chez
les mammiferes. Comptes Rendus, 1895, 1896. Archives d’Anatomie Muicro-
scopique, tome 1, 1897.

THORACIC DUCT DEVELOPMENT IN THE PIG 403

fixed in picro-sublimate and chrom-aceto-formaldehyde, and
several others were stained with borax carmine in toto and coun-
terstained with picric acid or blue de Lyon. Picro-sublimate is
very unsatisfactory in its preservation of tissue and should not be
used. For excellency of fixation and differentiation of vascular
structures the first method mentioned proved by far the best.
Not only did the sections show a beautiful transparency of color
and a strong contrast when this method was carefully followed in
detail, but they also produced the most favorable microphoto-
graphs.

A few embryos were injected with India ink. by the writer.
Series 23a was prepared and injected by Professor Sabin and
serves as an excellent critical stage in the development of the
thoracic duct.

Table 1 which represents only a partial catalogue of the Prince-
ton Collection of pig embryos enumerates the series studied and
reconstructed. They are arranged and grouped according to the
developmental status of the thoracic duct region in each, and
therefore certain embryos precede others of slightly lesser length.
On the whole, however, this system of gradation corresponds very
closely to gradations by length, the little inconsistencies here
and there being of trivial account when we consider that measure-
ment can be at most of only approximate accuracy, and that
fluctuations in growth are not at all infrequent.

The length of each embryo, excepting Sabin’s series 23a, was
obtained after fixation, when the danger of possible mutilation to
the embryo is least, and represents the crown-rump measurement,
viz., the distance between the crown of the head and the base of
the tail. Sabin’s series being measured before fixation is con-
sequently longer; but when it is considered that the processes of
fixation, hardening and dehydration reduce the absolute length
of such an embryo by 1 or 2 mm., this discrepancy in size dis-
appears and it is seen to fit in smoothly with related series.

All of the embryos were sectioned transversely to 15 and 20
micra in thickness. The significant sections of the series selected
for reconstruction were carefully drawn with the aid of the Edinger
drawing apparatus, the details then confirmed by the high power

404

OTTO F. KAMPMEIER

of the microscope, and the wax models made after a slight modi-

fication of Born’s method.

In this way the regions of immediate

interest in nine series were reproduced at a magnification of 100
diameters, excepting series 69 which was magnified to 50 diameters.
Besides these wax-models, a number of graphic reconstructions

List of material examined

TABLE 1

SERIES NO.

LENGTH OF

REMARKS

Injected

| Reconstructed in wax, X 100

Reconstructed graphically, x 100

Reconstructed in wax, X 100
Reconstructed in wax, X 100

Reconstructed graphically, X 100

| From the Johns Hopkins U. Emb. Coll.; injected;

reconst. in wax, X 100

| Reconstructed in wax, X 100

| Posterior region reconstructed in wax, X 100
| Reconstructed in wax, X 100
| Injected

Posterior region reconstructed in wax, X 100

| Injected
| Anterior region reconstructed in wax, X 100

EMBRYOS
- mm.
214 | 14
215 14 |
PAI 15 |
210 15
212 16 |
151 16 |
216 16
217 16
Unnumbered 16 |
150 18
167 17
141 lz
168 19
101 19
104 18
194 | 20
106 20
102 20
193 20
186 21
227 20
Unnumbered 21
23a, 23
103 21
191 21
105 22
192 | 21.
Unnumbered 23
67 23 |
Unnumbered 24
69 26
Unnumbered | 28

THORACIC DUCT DEVELOPMENT IN THE PIG 405

were carefully made, but on account of the more restricted accu-
racy of this method and the probable errors in the expression of
relations and proportions none were used as plate figures.

The series labelled ‘unnumbered’ represent incomplete series,
embryos of which only the anterior two-thirds or middle regions
were sectioned. |

Ill? REVIEW OF LITERATURE

It is needless to survey all of the former investigations which
have been concerned with the genesis of the lymphatic system ;
suffice it to say that the fundamental question in all of them has
been: Do the lymphatics arise from the veins, or are they a direct
product of the mesenchyme? Only the more recent literature
bearing on this query and intimately related to the writer’s own
observations will be cited and discussed here.?

In 1902 Sabin published her first work on the development of
the lymphatic system. By the method of injection she found
that all lymph vessels, both peripheral and systemic, arise at
four centers, two anterior and two posterior, and that they invade
the skin as well as the deeper-lying regions of the embryo by a
process of centrifugal growth. In other words, consecutive in-
jected stages will show the channels springing apparently as a
few simple sprouts from these points of radiation and then gradu-
ally growing longer and branching in an intricate manner until
they have spread throughout the whole body. The two anterior
centers, situated one on each side of the neck in the fork of the
jugulars, and the posterior ones, inguinal in position, represent the
locations of the anterior and posterior lymph hearts, respectively.
With the aid of a series of clear diagrams she has mapped out the
general course of lymphatic development. Unfortunately, how-

5 For a comprehensive list of the literature bearing on the development of the
lymphatic system, the reader is referred to the following two papers: George S.
Huntington: Die Entwickelung des lymphatischen Systems der Vertebraten vom
Standpunkte der Phylogenese des Gefiss-systems. Anat. Anz., Bd. 39, 1911.
Florence R. Sabin: A critical study of the evidence presented in sevéral recent
articles on the development of the lymphatic system. Anat. Rec., vol. 5, no. 9,
1911.

6 Florence R. Sabin: 1902; loe cit.

406 OTTO F. KAMPMEIER

ever, she did not describe or figure the details or factors of this
growth, except to say that the lymph vessels at first bud from
the veins and thereby derive their endothelium from them and
then continue to grow distally apparently by the proliferation of
their cells. The sprouting and elongation of the thoracic duct is
indicated in three diagrams constructed from embryos of 20, 27, °
and 30 mm. In the first of these schemes the mght lymphatic
and thoracic ducts are seen as two short caudal extensions from
the jugular lymph sacs. In the second, they have become longer
and the left one or thoracic duct has divided into two branches,
one of which passes to the right side of the aorta as the rudiment
of the right thoracic duct, and the other remains on the left side
as the left duct. In the diagram of the 30 mm. embryo these two
branches have extended far back and have established continuity
with the cisterna chyli and the posterior lymph hearts, thus com-
pleting the thoracic duct system.

A few years later, in 1906, F. T. Lewis made public a short
account of the development of the lymphatic system in rabbit
embryos.’ His results are of great interest because they repre-
sent a new conception of the genesis of the lymphatics, being
neither identical with the centrifugal growth theory of Sabin
nor with the theory of their direct mesenchymal origin, but in a
sense standing between these two. Furthermore, he was the
first investigator to determine the principle involved in the forma-
tion of the lymph sacs and to point out certain definite events
preparatory to the completion of the thoracic duct in the mamma-
lian embryo. In this studies on the transformation of the venous
system in the posterior regions of rabbit embryos, he noticed that
portions of the subeardinal veins became isolated and seemingly
converted into lymphatic vessels. He found more and more of
these so-called detached ‘lymphatics,’ and led by this suggestion,
he took up a more systematic investigation of the pathways of
the larger systemic lymphatics of the body. He made a number of
serial graphic reconstructions and thereby brought to light some
very interesting results. He observed that the jugular lymph

7 Frederic T. Lewis: The development of the lymphatic system in rabbits. Am.
Jour. Anat., vol. 5, 1905, pp. 95-111.

THORACIC DUCT DEVELOPMENT IN THE PIG 407

hearts arise by the coalescence along the internal jugular vein
of several venous outgrowths which become detached to form a
large isolated sac. A similar process was observed in the posterior
part of the body in the region of the subeardinal and mesenteric
veins. He also discovered a chain of discontinuous ‘lymphatic
spaces’ or endothelial-lined anlagen, apparently detached venous
outgrowths, situated along the azygos veins and in the path of the
future thoracic duct. From their position and consecutive arrange-
ment he concluded that they fuse with one another and the
jugular and mesenteric sacs and thus produce the continuity of
the duct. 'To quote his own words: ‘‘The study of the specimens
seems to show that the lymphatics along the aorta (thoracic
ducts) are derived in part from the azygos veins; below from the
subeardinal; and above from the jugular sacs.’’ It is also of
importance to state here that the ‘lymphatic spaces’ which he
described are ‘scarcely distinguishable from blood vessels.’

In 1907, shortly after the work of Lewis, Huntington and
McClure made a more detailed and extensive study of the develop-
ment of the jugular lymph sacs and confirmed in the main his
results.’ They had formerly believed that the sacs arise by the
formation, enlargement and fusion of perivascular spaces,’ but
now, from the data of a large number of beautiful and accurate
wax reconstructions of cat embryos, they established the opinion
that they come from the precardinal and in part from the post-
cardinal veins by the confluence of a series of outgrowths or
derivatives which they called veno-lymphatics, as suggesting
their venous origin and their subsequent transformation into a
lymphatic structure. For a short time McClure also carried
this conception to the developing thoracic duct,!° agreeing with
Lewis in the formation of a chain of discontinuous anlagen along

8 George S. Huntington and Charles F. W. McClure: The anatomy and develop-
ment of the jugular lymph sac in the domestic cat. Anat. Rec. vol. 2, 1908, pp.
1-18. Am. Jour. Anat., vol. 10, no. 2, April, 1910, pp. 177-311.

9 George 8. Huntington and Charles F. W. McClure: The development of the
main lymph channels of the cat in their relation to the venous system. Am. Jour.
Anat., vol. 6, 1907. Abstr. Anat. Rec., vol. 1, pp. 36-41.

10 Charles F. W. McClure: The development of the thoracic and right lymphatic
ducts in the domestic cat. Anat. Anz., Bd. 32, nos. 21 and 22, 1908, pp. 534.

408 OTTO F. KAMPMEIER

the azygos veins and derived from them, but he later withdrew
this view since it was based on the study of an insufficient num-
ber of critical stages.

Concerning all lymphatic vessels, not including the lymph
hearts, Huntington"! maintained the theory advanced jointly
by McClure and himself in 1906, that lymphatics have their origin
in the fusion of extra-intimal spaces which arise irregularly and
disjointly along primitive temporary venous channels.” Thus
he says:

The peripheral general lymphatic channels appear to be developed by
confluence of spaces independent of the venous system, although closely
associated with the same. The histological picture presented by them
differs radically from that of the jugular veno-lymphatic derivatives.
They begin as minute extravenous vacuoles closely applied to the sur-
face of the veins which they accompany. They enlarge as the lumen of
the veins diminishes... They become confluent with each other but
they never from their first inception contain red blood cells, nor do they,
as far as I have been able to ascertain in numerous carefully studied
series of excellent preservation.and fixation, communicate with the blood
channels.

_ In 1908, Sabin published a short paper! in which she reviews the
several positions held relative to the genesis of lymphatic channels
and attempts to turn the evidence in favor of the centrifugal
growth theory. Concerning Lewis’ multiple anlagen she says:

Since these spaces are lined with a definite endothelium, they form a
much more serious obstacle to the theory of growth of the lymphatics
from the endothelium of the veins than the more indefinite spaces to be
found in earlier embryos, and I cannot but think that if these multiple
endothelial-lined isolated spaces do exist along the veinsin the later stages,
they would form serious evidence against the theory of the origin of the
lymphatics from the veins. Or at least if the lymphatics, in their growth,
do pick up isolated endothelial-lined spaces, we shall again be left with-
out a clue as to the origin of the lymphatic system.

11 George 8S. Huntington: The genetic interpretation of the development of the
mammalian lymphatic system. Anat. Rec., vol. 2, 1908, pp. 19-45.

12 This theory was presented by Huntington and McClure before the Associa-
tion of American Anatomists in 1906 and published as a preliminary account in the
Anatomical Record no. 3 and in the American Journal of Anatomy, vol. 6, 1907.

13 Florence R. Sabin: Further evidence on the origin of the lymphatic endo-
thelium from the endothelium of the blood vascular system. Anat. Rec., vol.
2, 1908, pp. 46-54.

THORACIC DUCT DEVELOPMENT IN THE PIG 409

However, she firmly believes that the ‘lymphatic anlagen’ of
Lewis appear isolated only in the study of serial sections, and that
their continuity can be demonstrated by the method of injection.
In other werds, ‘‘in complete injections there are no vessels
which have not received the injecting mass,” but ‘‘in partial
injections and uninjected specimens there are endothelial-lined
vessels’? which appear to be broken up into segments, so that con-
tinuity ‘‘can be traced only with difficulty or not at all.” In
this same article she admits the presence of true mesenchymal
spaces ‘“‘which undoubtedly contain lymph,” but tacitly assumes
that they ‘‘are to be excluded from the lymphatic system morpho-
logically.” They are isolated and cannot be injected and they
do not possess a clearly defined intima.

A year later, in 1909, Sabin published her observations on the
development of the lymphatic system in human embryos."
In this work she reaches and emphasizes essentially the same
points as in her previous investigations. In the case of the.
thoracic duct, however, she hesitates to take a definite position.
She believes that it originates as outgrowths of the jugular lymph
sac and cisterna chyli, but she states in this connection that the
“thoracic duct has proved to be the most difficult part of the
lymphatic system to work out for this reason, we have not yet
found a way to inject it in early stages and uninjected sections
are not adequate.” Further:

The question is, does the duct develop from multiple anlagen from
the azygos veins for which there is no proof except that lymphatic ves-
sels can be seen in sections adjacent to these veins, or does the duct
grow from the two sacs, the cisterna chyli and the jugular one. For
the second view the evidence is also weak, it consists in this, that other
lymph ducts wherever we can study them grow from the sacs; and sec-
ondly in pig embryos and in human embryos one can trace a duct for-
ward from the cisterna chyli and caudalward from the jugular sac, and
in later stages these two ducts have joined. The weakness of this evi-
dence lies in the fact that in earlier stages the picture is always liable to
be confused by Lewis’ multiple anlagen.

M4 Florence R. Sabin: On the development of the lymphatic system in human
embryos, with a consideration of the morphology of the system as a whole. Am.
Jour. Anat., vol. 9, 1909, pp. 48-90. ,

410 OTTO F. KAMPMEIER

In 1910 Huntington and McClure! read two papers at the
International Congress of Anatomists at Brussels and, on the
basis of a study of cat embryos, presented striking evidence for the
theory, that the lymph ducts of the body are developed by the
confluence of mesenchymal spaces which are largely extra-
intimal, that is, formed around the lumen of an embryonic venous
channel which subsequently disappears completely. Further-
more, they drew the distinction very clearly that the lymphatic
system of mammals may be divided into two morphological
components: the lymph ducts which arise as indicated, and the
lymph hearts which form the connecting segment between the
systemic lymphatics and the veins and are transformed from a
venous plexus derived from the veins in their respective regions.
Concerning the former, which at present are of the greatest inter-
est, Huntington says:

Die Lymphgefiisse des ganzen KG6rpers enstehen durch den Zusam-
- menfluss einer grossen Anzahl von Hohlriumen, welche sich intercel-
lulir in Mesoderm entwickeln, in sehr genauer Anpassung an die Wand
der Embryonalen venésen Bahnen und in ganz derselben Weise wie die
ersten Anlagen des Blutgefiiss-systems, aber unabhaingig vom demselben.
~ Das Endothel, welches diese Hohlriitume die ersten lymphatischen An-
lagen auskleidet, ist von Anfang an unabhiingig vom Endothel der
Blutbahen und entwickelt sich mit dem ersten Auftauchen der lympha-
tischen Hohlriume aus den indifferenten Mesodermalzellen, welche
diese Hohlriiume begrenzen. Mit anderen Worten, das lymphatische
Endothel hat dieselbe genetische Herkunft wie das Endothel der Blut-
gefiisse, nimlich es besteht aus modifizirten Mesodermalzellen, welche
in die Wandung der intercelluliiren Hohlriume eintreten. Die erste
Stufe des histogenetischen Verlaufes ist ganz die gleiche, ob nun das
resultierende Hohlraumsystem in der Folge der lymphatischen oder der
hiimalen Abteilung des Gefiss-systems zugeteilt wird. Es gibt dem-
nach zwei Generationen der embryonalen vaskuliren Endothelzelle,
eine lymphatische und eine himale. Beide enstehen auf gleiche Weise
und infolge gleicher genetischer Einfliisse aus indifferenten Mesoder- >
malzellen. Beide sind vom Anfang des Vorganges an unabhiangig
voneinander.

165 George S. Huntington: Ueber die Histogenese des lymphatischen Systems
beim Sduger-embryo. Verhandl. d. Anat. Gesellsch., Bd. 24, 1910.

16 Charles F. W. McClure: The extra-intimal theory and the development of the
mesenteric lymphatics in the domestic cat. Verhandl. d. Anat. Gesellsch., Bd.
24, 1910.

THORACIC DUCT DEVELOPMENT IN THE PIG All

In order to emphasize the principle involved in the extra-intimal
development of the mammalian lymphatics, Huntington selected
the thoracic duct as an example for the reason that the histoge-
netic processes which enter into its inception and completion are
clearly expressed and easily followed, and also because the duct
retains a more definite and constant position relative to sur-
rounding structures than perhaps any other lymph channel. He
observed that the continuity of the thoracic ducts is realized by
the confluence of a large number of spaces which have sprung from
the mesenchyme immediately in contact with decadent venous
channels. Their first appearance is as numerous intercellular
isolated fissures which then coalesce to form larger spaces, and
these in turn become confluent to produce the continuous vessels.
Lined with undifferentiated tissue cells at their beginning, they
gradually assume the flattened and delicate endothelium of the
adult lymphatic.

McClure has described the same process of development in the
formation of the mesenteric lymphatics of the cat.!7 He showed
that at a certain period, a plexus of veins situated in the dorsal
mesentery becomes detached from the postcava and soon after
manifests signs of atrophy. With the aid of several clear micro-
photographs he further pointed out that the mesenteric lympha-
ties follow topographically, or better, appropriate these aban-
doned venous channels by a process of extra-intimal replace-
ment. The haemal endothelium collapses and large mesenchymal
spaces appear around it. In this connection he states:

These lymph spaces which lie external to the intima of the veins gradu-
ally encroach upon the territory formerly occupied by the veins and
finally fill it completely; the result being that the original intima of the
vein, no longer serving in the capacity of lining a functional venous
channel, gradually degenerates and disappears. Traces of this intima
can often be observed, however, in older embryos, clinging to the walls
of the lymph channels within which a new lymphatic intima has been
established.

That these pictures are real, and not artifacts induced by poor
preservation of tissue, is conclusively shown by the fact that they

17 Charles F. W. McClure: 1910; loc. cit.

412 OTTO F. KAMPMEIER

occur only at a definite period and place, and only in connection
with those venous channels which have become detached from the
main venous trunks and no longer serve in the economy of the
blood vascular system.

Last year there appeared in monograph form the first two parts
of Huntington’s investigations on the anatomy and development
of the systemic lymphatic vessels in the cat.18 Besides its great
detail, the work is profusely illustrated with convincing micro-
photographs and reconstructions and is a very positive and elabor-
ate confirmation of the theory of the direct mesenchymal origin
of alllymph ducts. Once more he emphasizes sharply the analogy
between the blood vascular and lymphatic systems in their earliest
anlagen, both beginning their history in a similar manner and in
the same soil. The first blood vessels arise in and amongst the
strands and ‘blood islands’ of the mesoderm as intercellular clefts
and fissures which enlarge, elongate and flow together to create a
network of intercommunicating channels. Their boundaries at
first are the unspecialized and cuboidal mesodermal cells among
which they lie. The fluid which fills their cavities and which is
perhaps secreted by these cells is evidently under a certain pres-
sure and exerts its influence in the modification of the immediate
or limiting walls into a vascular endothelium. The cells by a
mechanical adaptation to this pressure lose their cuboidal form
and become flattened and scale-like. Likewise the lymphatic
anlagen begin as intercellular spaces and enlarge, elongate and
coalesce into continuous vessels, and like the intima of the blood
vascular anlagen their intima is a differentiation of the cells
among which they are formed.

After such general considerations, Huntington enters into a
very complete description of the development of the thoracic
duct. Because a résumé of this history, as determined by him,
has already been given in the review of an earlier paper, it need not
be repeated here. Suffice it to say that nowhere has he found the

18 George 8. Huntington: The anatomy and development of the systemic lym-
phatic vessels in the domestic cat. Part I. The development of the systemic
lymphatics in their relation to the blood vascular system. Part II. The develop-
ment of the pre-azygos and azygos segments of the thoracic duct. Memoirs of the
Wistar Institute of Anatomy and Biology, May, 1911.

THORACIC DUCT DEVELOPMENT IN THE PIG 413

slightest evidence for ‘centrifugal growth’ as the fundamental
principle in the genesis of the thoracic duct, nor for its origin
from multiple venous anlagen.

In her most recent paper Sabin!® takes a position plainly at
variance with her earlier view of the development of the thoracic
duct, in that she restricts to a considerable degree the importance
of centrifugal growth by budding as the active principle or factor
in its formation. After briefly mentioning the conditions found
by her in two 23 mm. embryos and one measuring 25 mm., she
Says:

It is not possible to set limits to the transformation of veins into lym-
phatics making the cisterna chyli and thoracic duct, for by comparing
the two specimens measuring 23 mm. it can be seen that vessels which are
clearly branches of the azygos veins in the one specimen do not seem to
connect with the vein in the other. The thoracic duct develops in part
as a down growth of the jugular sac and in part, especially its dilated
portion or cisterna chyli, as a direct transformation of the branches of
the azygos veins.

This quotation would seem to indicate that she now believes
that the longer part of the thoracic duct is produced as a caudal
extension of the jugular lymph sae alone, and not, as she formerly
held, from two growing sprouts which subsequently meet, one of
them derived from the jugular sac, and the other from the cis-
terna chyli as an extension cephalad. Her failure to mention
either the absence or presence of this last or second sprout, which
she claimed to have found in her earlier investigations on the ori-
gin of the thoracic duct in human and pig embryos, confirms
strongly enough her change of view in this respect.

IV. OBSERVATIONS AND DISCUSSION

It is clear now, that there are three distinct views in the field
concerning the development of lymphatic vessels:

1. They spring from the veins at four centers of radiation and
by continuous elongation, centrifugal growth and branching
invade practically the entire body.

19 Florence R. Sabin: A critical study of the evidence presented in several

recent articles on the development of the lymphatic system. Anat. Rec., vol. 5,
no. 9, 1911.

414 OTTO F. KAMPMEIER

2. They are derived from the embryonic venous system either
by a direct transformation of certain of its channels or by the
fusion of multiple derivatives which have become detached from it.

3. They arise by the confluence of mesenchymal spaces, which
in the mammalan embryo are frequently perivenous in formation.

After a thorough and prolonged study of an extensive series of
pig embryos, the writer is forced by direct evidence to ally him-
self with the third position which holds that the thoracic duct has
an independent origin, that it is not a product of the veins either
by centrifugal growth or by the direct fusion and transformation
of venous derivatives, and that its intima is a gradual differentia-
tion from the mesenchymal reticulum. Essentially then he is in
harmony with and can confirm Huntington’s conclusions. Indeed
so positive and convincing are the results and so perfect in their
agreement among themselves that the possibility of doubting
their accuracy would seem to be entirely excluded.

In further corroboration of this view is the fact that some of the
individual stages of pig embryos show very clearly how the views
of the venous origin of the thoracic duct sprang into existence and
securedastrong foothold. During the formation of this duct the use
of the injection method can produce conditions, which, disregard-
ing all other details, would seem to corroborate the theory of its
centrifugal growth. Again, in certain stages there are structures
and data which would seem to furnish the necessary basis for the
other view, that it arises directly from a transformed vein or its
detached derivatives. But in both cases a comparative study of
a sufficient number of closely graded series will prove that these
appearances are due to the examination of an inadequate number
of embryos, to the tyranny of one method of investigation, or to
a faulty codrdination of all the facts available. The appearance
of centrifugal growth is plainly given by injection, but this method
with all its advantages can only indicate the regions of completed
channels, or the direction of their growth, and at best serve as a
control by supplying negative or indirect evidence; it can never
portray the actual genesis of a channel or reveal the histogenetic
processes which are at work from the beginning. Furthermore,
the fact that in certain stages there are venous channels which

THORACIC DUCT DEVELOPMENT IN THE PIG 415

have a course or position subsequently occupied by the thoracic
duct does not imply or-prove that the latter is a transformation
of the former. Embryologists are learning that in order to
select and understand all of the steps of embryonic changes,
one embryo of each consecutive age will not suffice as was
formerly the practice, but a number of embryos of the same
stage are requisite. Variations do not begin with the fin-
ished organism, but they are potent throughout ontogeny from
the beginning onward. Fluctuations of growth either of the
whole or of parts are not infrequent, and should we base conclu-
sions on a scanty number of embryos the chancesare that they will
be fragmentary or distorted. There are genetic changes of intrin-
sic importance, but so evanescent that we may not even catch a
glimpse of them unless we have a series of embryos which ap-
proaches the ideal of complete continuity. To produce such a
closely graded series may prolong the investigation and make
the technic more tedious, but the end result will be more certain
and will justify the labor expended. In the following descrip-
tions of typical and consecutive stages of the thoracic duct his-
tory, and in the discussion of the data presented by them, these
general considerations, as well as the various points suggested in
the review of earlier investigations, will become more significant.

The fundamental genetic period of the duct, that is the time
between the first appearance of its anlage and the acquisition of
continuity throughout its entire extent, is of very short duration.
Embryos of 18 or 19 to 23 or 24 mm., depending on individual
variations, are the important ones and almost the only ones
necessary for this study. In order to demonstrate more clearly,
however, the relation which the development of the duct bears to
the remainder of embryonic history, we must be aware of the fac-
tors and events that lead up to it or, in other words, believe it to
be already potential in the period just preceding its actual incep-
tion and realization. With this supposition in mind we may
artificially divide its history into three phases.

1. A veno-lymphatie phase, in which a system of provisional
venous channels, or ‘veno-lymphatics,’ is laid down throughout
the entire distance subsequently occupied by the thoracic duct.

416 OTTO F. KAMPMEIER

2. A transition phase, characterized by the atrophy of these
veno-lymphatics and the genesis of discontinuous lymphatic
anlagen.

3. A lymphatic phase, in which continuity is established
along the whole thoracic duct anlage and secondary growth proc-
esses bring about its completion.

Because these developmental changes proceed in a general
antero-posterior direction, more than one phase may be present
in the same embryo at the same time, although at different levels,
but so definite is the succession of events that the above division
into three phases will invariably suggest itself.

1. The veno-lymphatic phase (15-19 mm. pig embryos)

In the early embryonic history the venous plan of the thoracic
region is composed of two strictly symmetrical and bilateral
halves which are practically disconnected from each other except
through the heart. Later by the formation of plexuses, anasto-
moses and fusions, certain channels acquire more of the blood
current and thereby gain supremacy over others, which gradually
dwindle in size and vanish and consequently give rise to the start-
ling asymmetries of older embryonic stages and of the adult. At
the time when the first of these profound transformations are
initiated, a series of vessels are developed which function only
temporarily in this scheme and then disappear completely. Ref-
erence is here made to the ‘veno-lymphatics’ which have their
origin and consummation in those stages approximately between
15 and 19 mm. in length and are a part or product of the supra-
cardinal or azygos system during its early transformations.

For want of a better descriptive term, the word ‘veno-lymphatics’ has
been extensively used throughout this paper but nevertheless with some
reluctance. A veno-lymphatic, as defined by Huntington and McClure
in their work on the development of the jugular lymph sac in the eat,
is a venous derivative which by confluence with other such channels is
directly transformed into the lymphatic structure. Instead of restrict-
ing himself to this original meaning, the writer has employed this term
in a somewhat different and a wider sense, as designating those tem-
porary embryonic venous channels which occupy topographically the

position of the future thoracic ducts, or other lymph ducts, but atrophy
and disappear during the genetic period of these lymphatics. If this

THORACIC DUCT DEVELOPMENT IN THE PIG 417

distinction is clearly grasped all possible confusion will be avoided.
There is a firm suspicion, however, that all veno-lymphatics vessels,
whether apparently direct or indirect antecedents of some lymphatic,
are fundamentally identical or homologous structures.

A brief account of the early history of the supracardinal or
azygos system of veins will simplify the explanation of the source
and character of the veno-lymphatics in the thoracic duct area
and their grouping into three divisions. In 14 and 15 mm. em-
bryos, the preecardinal and jugular veins give rise, in the region of
the anterior lymph sac, to a number of dorsal tributaries which
are continued back to the posterior part of the body as two slender
channels immediately above and parallel to the pre- and post-
cardinal veins. On account of their position and subsequent
history, these longitudinal channels may be ealled the supra-
eardinal lines. Throughout their course they are joined to the
pre- and posteardinals by numerous cross-anastomoses. They
also possess branches which may be deseribed as dorsal segmental
veins, because they drain the regions of the back on each side of
the vertebral column and spinal cord and appear to be arranged
metamerically. The disposition and fate of the supracardinal
lines in a later stage (19 mm. embryo, fig. 28) can be indicated as
follows: Their precardinal division (A), that segment extending
between the levels of the jugular lymph sacs and the Cuvierian
ducts, is complicated into a plexus, some channels of which become
veno-lymphaties (6a), to be considered presently, and others are
absorbed by the dorsal branches (8) of the precardinal veins.
Their middle or posteardinal division (B, 12ld, 12ls), between the
Cuvierian ducts and the anterior extent of the mesonephroi, fuses
longitudinally with the postcardinal veins. The posterior divi-
sion (C) furnishes the true supracardinal veins?® (12d, 12s),

20 George S. Huntington and Charles F. W. McClure: The development of the
posteava and tributaries in the domestic cat. Am. Jour. Anat., vol. 6, 1907,
Abstr. Anat. Rec., vol. 1: “‘A bilateral and originally symmetrical venous channel
develops dorso-medial to the primitive postcardinal vein by longitudinal anas-
tomoses between somatic postcardinal tributaries. This secondary vein channel
forms what we have termed the supracardinal system of veins. It extends from
the level at which the posterior limb veins open into the postcardinals to a point

cephalad where it joins that portion of the postcardinal which alone persists to
form the anterior end of the adult azygos.”’

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 4

418 OTTO F. KAMPMEIER

which proceed caudally as two large and important vessels and
only very much later take a further part in the production of the
asymmetrical venous plan of the trunk, especially in their trans-
formation with the posteardinals to form the azygos system of
veins; this segment may therefore be termed the supracardinal
division. Since the pre- and posteardinal divisions (A, B) of the
longitudinal supracardinal lines disappear or lose their independ-
ence, either by being transformed into transient venous plexuses
as in the first division or by fusing entirely (12/d, 721s) with the
posteardinals as in the second division, it is evident that the dorsal
segmental tributaries (8), which in the beginning spring from
these lines, must shift their roots so that they arise directly from
the pre- and posteardinal veins in the two divisions mentioned.
In the third or posterior division (C), however, these segmental
tributaries (8) continue to return their blood to the supracardinals,
for the latter exist as independent venous trunks.

Because the three divisions in the transformations of the supra-
cardinal lines coincide perfectly with, or better map out, three
well-defined regions in which the history of the thoracic duct is
enacted, the terms, pre-, post-, and supracardinal divisions are fully
as applicable here, the events in this history occurring immedi-
ately along the pre-, post-, and supracardinal veins, respectively.

(A) Precardinal division. In 17 to 19 mm. embryos the
anterior segment of the left supracardinal line, originally a simple
longitudinal channel as in the 15 mm. embryos, has been trans-
formed into an intricate plexus the roots of which now appear as
numerous dorsal tributaries of the internal and common jugular
and precardinal veins. The branches of these tributaries extend
dorsad and vascularize the areas on both sides of the sympathetic
nerve trunk, but in number and complexity the internal branches
exceed the external ones. As will become evident later, an im-
portant distinction is potentially present between these two sets
of branches, and, although the lack of differentiation at this
stage would not warrant the use of specific terms, in the light
of future events they may be described as precardinal veno-
lymphaties, lying internal and mesial (6a, figs. 1 and 28), and
precardinal segmental veins (8), functionally related to the ter-

THORACIC DUCT DEVELOPMENT IN THE PIG 419

Fig. 1 Transverse section through the left jugular lymph sac region in a 19 mm.
pig embryo (series 168, slide 16, section 16), 100. 1, lymph sac; 2, thoracic
duct approach; 6a, precardinal veno-lymphatic; 8, root of a precardinal segmental
tributary; 9, internal jugular vein; 13c, carotid artery; 14, sympathetic nerve
trunk; 15, vagus; 16, recurrent laryngeal nerve; 17, oesophagus; 18, trachea.
(Reconstruction, fig. 28.)

ritory lateral and dorsal to the sympathetic trunk. In figure
28, which represents a reconstruction of a 19 mm. embryo, these
two kinds of tributaries (6a, 8) and their relations to the neigh-
boring structures can be clearly distinguished.

Continued back from the terminals of the precardinal veno-
lymphatics (6a, fig. 28) is a vessel which passes obliquely over the
aorta and oesophagus to enter the right postcardinal vein at the
level of the Cuvierian ducts (7, fig. 28). Although it is morpho-
logically a part of the precardinal veno-lymphatics, it will often
be treated separately and called the ‘oblique vessel’ on account of
its diagonal course. The simplicity and size of this vessel varies
with the individual, but it is a constant factor in all of the early

420 OTTO F. KAMPMEIER

Fig. 2 Transverse section taken near the level of the left Cuvierian duct in
a 17mm. embryo (series 167, slide 15, section 22), * 100. 6a and? (oblique vessel),
precardinal veno-lymphatics located in the positions of the later right and left
thoracic ducts respectively; 8, dorsal segmental vein; 10, left precardinal; 11d,
right posteardinal; 12/d, 12ls, right and left supracardinal lines; 13, aorta; 14,
sympathetic nerve trunk; 15, vagus; 17, oesophagus; 18, trachea.

stages of the thoracic duct history. In figure 2,2! it (7) is illus-
trated in section, taken just in front of its Junction with the
middle segment of the right supracardinal line, which in this
embryo (17 mm.) had not yet fused throughout its entire extent
with the right posteardinal.

21 The image of the sections being reversed by the microscope lens, the structures
of the right side of the embryos are seen on the left side in the microphotographs,

and vice versa. However, in the dorsal views of the reconstructions and in figure
13 the right and left sides correspond with the respective sides of the page.

THORACIC DUCT DEVELOPMENT IN THE PIG 421

(B) Postcardinal division. Small venous derivatives are formed
by a process analogous to fenestration along the mesial border of
the posteardinal (66, figs. 12, 28 and 32) of each side and appar-
ently always in the line of fusion of this vein with the middle
segment of the originally independent supracardinal channel,
thus suggesting their derivation from the supracardinal system
also in this division. During their development consecutive
ages may be distinguished among them by the amount of individ-
uality they manifest. Some are merely little bulging irregulari-
ties in the circumference of the parent vein, others describe
the first step of separation by the presence of thin strands or
partitions, and still others are complete throughout a number of
sections but open to the veins at one or both ends. Being in the
direct axis of the precardinal veno-lymphatics and homologous
with them, these venous spurs or derivatives may be called the
posteardinal veno-lymphaties. Ordinarily they are exhibited
more distinctly on the right side. Later they constitute longer
and shorter venules (60, figs. 29, 32) parallel to the posteardinal,
but they are never quite independent of the latter, remaining
joined to it here and there until their reduction when they break
up into degenerating segments and disappear in the mesenchyme.
Such a final procedure will become clearer in the consideration of
the second or transition phase.

(C) Supracardinal division. In the region of the mesonephrol,
plexuses of vessels spring from the ventral aspect of the supra-
cardinal veins, and in some embryos they become so extensive
as to encompass the aorta almost completely. For this reason
and the fact that they are the equivalents of the anterior veno-
lymphatics in function, they were named the supracardinal peri-
aortic veno-lymphaties. A further description of them would be
superfluous considering their clearness in the accompanying micro-
photograph and reconstruction (6c, figs. 8 and 32).

Caudally, at the level of the superior mesenteric artery, sub-
sidary channels arise from the supracardinals ventro-medially,
become more and more plexiform, and appreach one another from
both sides to anastomose in the area dorsad of the aorta (6c,
fig. 4). Since eventually they will be concerned in the pro-

422 OTTO F. KAMPMEIER

Fig. 3. Transverse section through the region just in front of the mesonephroi
in a 19 mm. embryo (series 168, slide 23, section 17), X 150. 6c, supracardinal
periaortic veno-lymphatics in the positions of the later right and left thoracic
ducts; 11d, 11s, right and left postcardinals; 12d, 12s, right and left supracardinals;
13, aorta; 17, oesophagus. (Reconstructions, figs. 28 and 32.)

duction of the most posterior segment of the thoracic duct, or
cisterna chyli, they may be labelled posterior supracardinal
veno-lymphatics.

Situated topograpically in the pathway of the future thoracic
duct, all of the veno-lymphatics mentioned in the three divisions
would be said to give rise to it directly, or to be transformed into
it, were the second or transition phase of its history disregarded.
This leads up to the suggestion that Lewis’ and temporarily also
MecClure’s multiple endothelial-lined anlagen derived from the
veins along which they lie are nothing more and nothing less than
the veno-lymphaties or precursors of the thoracic duct as deline-

THORACIC DUCT DEVELOPMENT IN THE PIG 423

Fig.4 Transverse section taken near the level of the superior mesenteric artery
in a 20 mm. embryo (series 194, slide 40, section 16), * 200. 6c, posterior supra-
cardinal veno-lymphaties, the precursors of the subsequent cisterna chyli; 12d, 12s,
right and left supravardinals; 13, aorta; 13ds, dorsal segmentalartery; 14, sympa-
thetic nerve trunks.

ated by the writer. According to the descriptions, their character
and derivation in the rabbit and cat embryos accord accurately
with their tendencies in the pig embryos and allow of no other
conclusion than that they are homologous structures; for not only
do they occupy identical positions but they also have their sources
in the same venous trunks. This similarity is even more strik-
ingly emphasized in somewhat older pig embryos where they are
on the verge of evanescence and break up into isolated segments.

Until they have reached their culmination these veno-lympha-
tics can be followed and reconstructed with remarkable ease and
accuracy. Their caliber or lumen is as uniform as a plexiform
condition permits, and constrictions, irregular shrinkage, collapse,

424 OTTO F. KAMPMEIER

or any other characteristic which might suggest artifacts due to
the action of the preservative are nowhere in evidence. Their
endothelium is tense and clear and the enclosed blood corpuscles
take a clean and transparent stain.

During the veno-lymphatic phase the mesenchyme is evenly
woven and fairly compact, and lymphatic anlagen, or conspicu-
ous and discontinuous spaces, are not present throughout the
entire thoracic duct area. Nor can vacuities, fissures, or rents
be observed which might be ascribed to unequal fixation, or
which differ in any way from the regular intercellular lacunae of
the tissue reticulum.

2. The transition phase (19-22 mm. embryos)

Having arrived at the second phase of the thoracic duct his-
tory, we are confronted with the paramount point at. issue,
namely, the source and formation of lymphatic anlagen. A
critical examination of the stages belonging to this period will
disclose three facts of major importance which are impressed
upon the observer firmly and constantly. In the first place the
longer portion of the thoracic duct anlage arises discontinuously
from mesenchymal lymphatic spaces, but it may present various
aspects according to its genetic levels or to the degree of relation
it bears to immediately surrounding structures, that is, it may be
instituted either by extra-intimal spaces, or by spaces in the near
environment of the veno-lymphatics but not in contact with them.
Secondly, the wall of the entire thoracic duct is a differentiation
in situ from the mesenchyme. Thirdly, the development of the
thoracic duct proceeds in a general antero-posterior direction;
for example, in series 194, a 20 mm. embryo, lymphatic develop-
ment has made considerable progress in the anterior or precardinal
division, has just been initiated inthe middle or posteardinal divi-
sion, and is totally lacking in the supracardinal division.

(A) Precardinal division. In the collection of pig embryos
studied by the writer the first instances of incipient lymphatic
anlagen are found in series 168 (19 mm.) along a limited extent of
the precardinal veno-lymphaties as far back as the anterior half
of the oblique vessel (fig. 28). Located in the path of the poten-

THORACIC DUCT DEVELOPMENT IN THE PIG 425

Fig. 5 Transverse section through the left lower cervical region in a 19 mm.
pig embryo (series 168, slide 17, section 9), & 150. 1, posterior tip of the jugular
lymph sac; 2, thoracic duct approach; 4, lymphatic space formed against the intima
of a precardinal veno-lymphatic, (6a); 9, internal jugular; 13, wall of the aorta;
14, sympathetic trunk; 15, vagus; 16, recurrent laryngeal nerve; 17, oesophagus.
(Reconstruction, fig. 28.)

tial thoracic duct there is a distinct vacuolation of the tissue as
shown in figure 6 (4), medial to a veno-lymphatic tributary (6a).
Although exceedingly difficult to describe, these vacuoles or
spaces are seen to stand out conspicuously, perhaps by the greater
clearness of their cavities, even if their boundaries are ill-defined,
and by their preponderance in size over the more regularly dis-
posed openings of the ground substance. In a section taken fur-
ther forward another lymphatic anlage occurs as a large lenticular
space (4, fig.5) against the intima of aretrogressive veno-lymphatic
(6a), but 1t can be distinguished definitely only in two sections.
In the same figure the thoracic duct approach (2) is indicated.
This ends blindly but is followed shortly by large clear-cut spaces
(4) which are in no way continuous with it (fig. 28). Similar
and widely separated anlagen occur along the anterior or upper
half of the oblique vessel (7) and immediately ventral to it.

426 OTTO F. KAMPMEIER

Those that were not hidden in the dorsal view of the reconstruc-
tion are illustrated in the drawing (4, fig. 28). These spaces are
short capsule-shaped vesicles running through four or five sec-
tions and are quite distinct in outline, so that their extent and
discontinuity can be easily determined. A transverse section
of one of them is represented in figure 7 (4). The nicety with
which it can be discriminated from the interstices of the sur-
rounding mesenchyme and from the lumen of the neighboring
venule which is filled with blood removes all doubt as to its reality
and individuality.

On the evidence of the few microphotographs inserted here,
the presence of the spaces pointed out as lymphatic anlagen
can not be denied. In the embryo just described they exist only
in the foremost region of the thoracic duct line; outside of this
path there are no spaces which might invalidate the significance
of these anlagen.

Extra-intimal replacement of evanescent venous channels, a
method of lymphatic development just hinted at in figure 5 (4)
finds a most convincing expression in series 194 (20 mm.), where
the entire anterior precardinal veno-lymphatic plexus (6a) is
being replaced by large perivenous spaces (3, fig. 29). The
veno-lymphatics designated have been detached from their
parent veins and thus abandoned by the systemic blood vascular
circulation, and they now display successive steps towards com-
plete collapse. The section represented in figure 8 is typical, for
most of the other sections taken at random from this region offer
equally decisive illustrations. Besides revealing the shriveled
and discarded venous intima and its gradual disintegration in the

Fig.6 Transverse section through the left lower cervical region in a 19 mm. pig
embryo (series 168, slide 18, section 4), X 150. 4, vacuolation of the mesenchyme
in the formation of lymphatic spaces; 6a, precardinal veno-lymphatics, beginning
to degenerate; 9, internal jugular; 13, aorta; 14, sympathetic nerve and branches;
17, oesophagus. (Reconstruction, fig. 28.)

Fig. 7 Transverse section through the left upper thoracic region in a 19 mm.
pig embryo (series 168, slide 19 section 1), * 150. 4, lymphatic space in the
line of the future right thoracic duct; 7, oblique vessel; 8, dorsal segmental vein;
10, left precardinal; 13, aorta; 14, sympathetic nerve trunk and branch; 17, oeso-
phagus. (Reconstruction, fig. 28.)

428 OTTO F. KAMPMEIER

lumina of the replacing lymph vessels (3a, b, c, d, e), the photo-
eraph suggests the method by which the latter enlarge, as well
as their mode of origin. The anlage at 3e (fig. 8) presents an
excellent initial stage in which the endothelium of the blood chan-
nel has receded from the original circumference and two small
mesenchymal vacuoles have appeared one on each side of the
points of weakness. On the other hand, the anlage at 3a, the
channel of the lymphatic plexus nearest to the lymph sac and
internal jugular vein (9), is very large and irregular and has
increased in size obviously by the coalescence of several closely-
crowded spaces, as indicated by the extremely ragged periphery
of its lumen and the remnants of tissue traversing it.

The lymphatic plexus just described is widely confluent with
the jugular lymph sae (fig. 29) through the thoracic duct approach
(2). The reader will recall that this structure, the approach, is a
part of the sac and has its origin with it. In the preceding or
veno-lymphatie phase it exists normally as one (fig. 28) or two,
and sometimes three, short prolongations between the roots of
the dorsal tributaries of the jugular vein. At the time when
lymphatic spaces are appearing along the precardinal veno-
lymphatics, the sharply defined venous endothelium of the
approach (2, fig. 8) retracts from its former circumference, evi-
dently as the result of a stagnation in its growth, and becomes
surrounded by a clear and larger cavity (4) which is lined with
ordinary unmodified mesenchymal cells, the progenitors of the

Fig. 8 Transverse section through the left lower cervical region in a 20 mm.
pig embryo (series 194, slide 23, section 21), & 120. 2, thoracic duct approach,
and its original venous intima replaced by alarge space (4); 3a, b, c, d, e, lymphatic
plexus replacing as extra-intimal spaces the precardinal veno-lymphatic plexus;
9, internal jugular; 13, wall of the aorta; 15, vagus; 17, oesophagus; 18, trachea.
(Reconstruction, fig. 32.)

Fig. 9 An accurate camera lucida drawing of a highly magnified area of the
section represented in figure 8, X 266 (reduced from X 400). 3a, b,c, d, cross-sec-
tions of the extra-intimal lymphatic plexus replacing the anterior precardinal
veno-lymphatics; the absence of any specialized endothelium in the wall of the
lymphatic spaces is plainly evident; the strands of mesenchyme jutting into their
cavities suggest the breaking down of contiguous spaces in the formation of the
lymphatic plexus; 6a, the collapsed venous intima of the abandoned veno-lympha-
tics; 9, wall of the internal jugular; 14, sympathetic nerve trunk and branch.

T DEVELOPMENT IN THE PIG 429

~”

DUC

THORACIC

430 OTTO F. KAMPMEIER

later more specialized lymphatic endothelium. Synehronously,
the reorganized approach becomes confluent with the contiguous
lymphatic spaces of the duct anlage by the breaking down of
tissue partitions and septa between them. The vestiges of the
eld vascular intima may persist throughout a number of stages
clinging to the wall of the new cavity, but it gradually fades and
vanishes as the thoracic duct acquires more and more of its func-
tional activity.

Histologically, all incipient lymphatic anlagen, whether they
are spaces independent in position or spaces following, transform-
ing and expanding the discarded pathways of redundant venous
channels, are decidedly different from either a —ctive vein or a
mature lymphatic. They lack definition and possess vague and
undifferentiated outlines; for the cells of their walls are not
arranged in that end-to-end fashion so characteristic of vascular
endothelia. Instead, many instances were observed under strong
magnification where the tissue cells in their longest diameter
stand perpendicular to the periphery of the anlagen and project
far into the lumen with their cytoplasmic filaments, a condition
unquestionably brought about by the addition or fusion of con-
tiguous spaces. Figure 9, which is an accurate camera lucida
drawing of a highly enlarged portion of the area pictured in figure
8, should be carefully examined as depicting clearly the features
here mentioned. With the most critical observation one is not
able to detect differences at this stage between those cells con-
stituting the boundaries of the lymphatic anlagen (3a, b, c, d) and
those of the mesenchymal reticulum, either in regard to their
arrangement and shape or to their staining attributes.

While there is sufficient evidence for the atrophy of veno-
lymphaties (6a, fig. 9) in their elimination from the blood stream
and the recession of their intima, further evidence is revealed at
this stage by their reaction to the stain. Treated with heama-
toxylin and orange-G, the defunct intima takes an opaque brownish
color as compared with the transparency of a functional vessel.
Their lumina also contain the débris of blood cells. That these
conditions are not induced by poor fixation is evinced by the nor-
mal appearance of the veins in the immediate vicinity. For

THORACIC DUCT DEVELOPMENT IN THE PIG 431

example, the endothelial lining of the jugular vein (9, fig. 9)
stains clearly and is sharply defined, and there are no spaces
external to it.

Extra-intimal replacement occurs only among those venous
channels which are immediate antecedents of lymphatics in time
and place. This is attested by the fact that other veins in their
atrophy are not surrounded by spaces but disappear by the grad-
ual reduction of their caliber, or by a process of constriction
cutting the channel into segments which become smaller and
smaller to form dense masses or islands of cells ultimately to be
lost in the mesenchyme. As an instance of such a process may be
described the reduction and dismemberment of a large portion of
the plexuses uniting the original supracardinal lines (25, figs. 28,
29 and 30). The writer has often noticed these temporary venous
plexuses in the various stages of degeneration. In 20 mm. pig
embryos, for example, such retrogressive venous channels fre-
quently reveal constrictions at irregular intervals along their
course. In later stages they begin to break up into segments, which
at first, however, are still connected with one another by densely
staining cell-strands, the remains undoubtedly of the endothelial
walls, and in this way indicate the pathways of the originally
functional vessels. The segments at the beginning possess a
distinct cavity or lumen, but subsequently the cavity becomes
filled up with a solid cell mass which is apparently due to a pro-
liferation of the former endothelial or lining cells. Such cell
aggregations gradually vanish in the mesenchyme perhaps by the
regression of their elements to undifferentiated tissue. In other
words, the cells which at one time functioned as the limiting walls
of a haemal vessel, after the elimination of such a vessel from the
blood channel system, lose their specialized characteristics and
possibly return to the mesenchyme by assuming the qualities and
functions of the ordinary tissue cell. By comparing the recon-
structions illustrated in figures 28 and 35, it is seen that the venous
plexuses which are so profusely developed in the 19 mm. embryo
have almost entirely disappeared in an embryo of somewhat
greater length. In the stained sections of the latter, however,
some of the dense cellular masses are still visible here and there in

432 OTTO F. KAMPMEIER

yi ae 5 oP
Ra £2. K

ven ~~

eS
,
%
>! S. 4
8
pAtes8
‘ as
wah Oe

eos

sel

mnt

Fig. 10 Transverse section through the upper thoracic region in a 20 mm. pig.
embryo (series 194, slide 26, section 21), * 150. 4d, potential lymphatic anlage in
the right thoracic duct line; 4s, lymph space in the left duct line; 7, oblique vessel,
here plexiform; 8, dorsal segmental vein; 10, left precardinal; 13, aorta; 17,
oesophagus. (Reconstruction, fig. 29.)

the territory formerly traversed by the temporary venous plex-
uses.

In series 194 (20 mm. embryo) the atrophy of the oblique ves-
sel (7) and its replacement by lymphatics has begun anteriorly,
where it is supplanted to a greater or lesser degree by perivenous
spaces (4d, fig. 29) which reflect the features of those further for-
ward but are smaller and less conspicuous by reason of the greater
simplicity of the vessel replaced. Along the second half of its
course lymphatie anlagen are met with which are not formed in
such intimate relations but le ventral to and not closely in contact
with it. Figure 10 illustrates a section from this region. The
clear area (4d) subjacent to the plexiform oblique vessel (7) in
the triangular or wedge-shaped territory between the aorta and
the oesophagus unhesitatingly suggests a thoracic duct anlage in
the making, as an examination of its character and a comparison
with later stages seem to affirm. That this clear area is less com-

THORACIC DUCT DEVELOPMENT IN THE PIG 433

pact in texture than the tissue surrounding it is evident at a
glance. Within it the cells are fewer in number, and the tissue
fibrils, which appear to be more delicate than those of the mesen-
chymal reticulum elsewhere, enclose larger interstitial openings or
tissue spaces. In longitudinal extent this potential duct-anlage,
as we may eall it, occurs along a considerable portion of the pos-
terior half of the oblique vessel but varies from section to section
in its definition. Often, distinct vacuoles appear suddenly in
it, continue through several sections, and as suddently disappear.
Only these ‘centers of space-formation,’ however, can be repro-
duced in a model (fig. 29), the remainder of the anlage being as
vet too indefinite to warrant reconstruction.

The atrophy of the oblique vessel in this specimen, series 194,
serves also as a typical example of the atrophy of all temporary
or redundant venous pathways, both of the veins which are the
immediate antecedents of lymphatics and of those veins which are
not so intimately associated with the development of a lymphatic
channel. In the embryo from which figure 28 was drawn the
oblique vessel (7) is still complete and continuous with the main
venous trunks; in figure 29, on the other hand, it is seen to be
broken. up into irregular segments, some of which are replaced by
extra-intimal spaces, and others gradually diminish in size and
disappear in the mesenchyme adjacent to an incipient lymphatic
anlage but with an appreciable amount of tissue between them.
Thus it is evident that the vanishing segments of redundant
venules and the growing segments of potential lymphatics may
exist side by side in the same section. But now it may be asked,
what distinguishes the one from the other, what basis is there for
naming this one a lymphatic rudiment and that one a venous
remnant, and how can both be followed to their ultimate fates
without confusion? The distinction between these two vascular
structures can easily be recognized beyond the possibility of a
doubt. A lymphatic segment, here specifically a thoracic duct
anlage, is invariably characterized by a very clear lumen and, if
it is in the formative stage, by the absence of a clear-cut and
specialized lining, as shown in the microphotographs already
mentioned (figs. 6, 7, 8 and 10); whereas, the segments of a ven-

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO

434 OTTO F. KAMPMEIER

Fig. 11 Transverse section through the left lower cervical region in a 20 mm.
pig embryo (series 193, slide 14, section 20), * 200. This figure shows the anterior
lymphatic plexus (3) of the thoracic duct not developed from extra-intimal spaces,
but formed in the mesenchyme independently of the precardinal veno-lymphaties
(6a); 9, internal jugular; 14, sympathetic nerve trunk; 14, vagus; 17, oesophagus.
(Compare with fig. 8.)

ous channel undergoing atrophy almost constantly contain
numerous blood corpuscles, or the deeply staining débris of cel-
lular material (figs. 6, 7,8 and 10), and without exception possess
a sharply defined endothelium.

To strengthen the evidence of figures 6, 7 and 10, illustrating
the formation of lymphatic spaces without the features of extra-
intimal replacement, figure 11 from series 193, another 20 mm.
embryo, may be introduced here as a very definite specific case
to show conclusively that such replacement of venous channels
is not inevitably a requisite in the genesis of lymphatics, and that
its absence is quite as common as its occurrence. In this embryo
the most anterior portion of the thoracic duct anlage (3) has not
been modelled, so to speak, upon the anterior precardinal veno-

THORACIC DUCT DEVELOPMENT IN THE PIG 435

lymphatic plexus (6a), but has formed a plexus independently
of it in a position closely parallel and medial to it. Histogeneti-
cally, however, this lymphatic plexus is the same as that of the
preceding embryo (series 194, figs. 8, 29), but the redundant
verous lines of which it seems a shadow picture are still quite
regular, although occasional constrictions do suggest their decline
and subsequent atrophy.

The foregoing descriptions of the relation between the develop-
ment of lymphatic anlagen and the degeneration of the veno-
lymphatics determine clearly that this relationship possesses
only secondary significance. These two processes are necessary
events in the embryonic history, and if they occur simultaneously
and the abandoned venous derivatives occupy a position identical
with that of the potential thoracic duct, then the anlagen of the
latter will follow the path of least resistance or, better, follow
a hydrostatic tendency and collect around their weakened intima
and cause its collapse. If, however, these venous lines do not lie
in the pathway of the duct, or if their degeneration is slightly
retarded so that they are still joined to the systemic blood circula-
tion, and are under the influence of its pressure and their intima
is still tense, then the lymphatic anlagen will arise independent of
any contact with them.

Fluctuations in the amount of progress attained by the right
and left branches of the duct at any given moment during the
critical stages of their development are not infrequent; indeed
there appears to be a reciprocal action, for when one is large and
long, or well represented in the number of its anlagen, the other
is only scantily represented. All transition stages show these
variations to a greater or lesser degree, but especially favorable
examples are series 103 and 191 (21 mm. embryos), which can be
regarded as complements of each other, the former being promi-
nently dextral, and the latter sinistral in lymphatic growth. In
series 103 the right limb of the duct extends as an unbroken chan-
nel far back into the posteardinal division of the thoracic region,
but the left limb is just visible in its earliest rudiments as a few
minute and isolated spaces. In series 191, on the contrary, we
meet with a complete reversal of conditions so that the descrip-

436 OTTO F. KAMPMEIER

Fig. 12 Transverse section taken immediately posterior to the left Cuvierian
duct in a 20 mm. pig embryo (series 194, slide 29, section 4), X 150. 4d, isolated
lymphatic spaces in the right thoracic duct line; 6b, posteardinal veno-lymphatics;
11d, 11s, right and left posteardinals; 13, aorta; 17, oesophagus. (Reconstruction,
fig. 29.)

tion of the right thoracic duct limb of series 103 applies with
almost equal force to the left limb of this embryo, and vice versa.

(B) Postcardinal division. In the region of the posteardinal
veno-lymphatie channels the discontinuity of incipient lymphatic
anlagen and their origin in situ from the mesenchyme can be
plainly demonstrated. In a destined course and at intervals,
though not metamerical in sequence, spaces and fissures are pres-
ent in the evenly meshed tissue along the channels mentioned.
For instance, they can not be mistaken in figure 12 (series 194—
20 mm.), where two of them are shown as clear crevice-like spaces
(4d) quite sharply defined, while several others are invading the
environs of the veno-lymphaties (6b) and are beginning to enclose
one of their smaller branches. In other sections this process of
circumclusion has proceeded further, but in every case venous

THORACIC DUCT DEVELOPMENT IN THE PIG 437

ean readily be distinguished from non-venous, the first by the
presence of blood cells and heavier walls, and the second by clear
cavities and more delicate walls.

In series 23a (23 mm. before fixation), an embryo somewhat
older than series 194, the discontinuous lymphatic spaces of the
posteardinal division are much larger and more conspicuous in
the figures. Series 23a is from the Johns Hopkins University
Embryological Collection and was injected and prepared by Pro-
fessor Sabin. The fixation and preservation of its tissue is excel-
lent, and the injection was successfully carried out and is as per-
fect as a developing lymphatic channel permits. It was sent to
the Princeton Laboratory as a crucial stage in favor of the ‘cen-
trifugal growth’ theory of the origin of the thoracic duct, and
therefore the evidence derived from it will seem more significant
perhaps than that derived from any other series described. For
this reason it will here be dealt with in greater detail.”

The right and left limbs of the thoracic duct anlage which are
joined to the jugular lymph sae are continuous channels as far
back as the points X, XY on the drawing, figure 30, and being
injected are shown. in black on the diagram, figure 18. The fore-
most portion of the embryonic duct is in the form of a broad and
extensive lymphatic plexus, a typical section of which is repro-
duced in figure 14 to illustrate the large size of the channels (3)
and the extravasations or leakage (Hx) of the injection massinto
the surrounding mesenchyme. ‘To consider the right limb of the
duct first, the injected vessel, or vessel confluent with the lymph
sac, extends unbrokenly backward and dextrad towards the right
posteardinal vein as a slender channel, subjacent to the oblique
vessel (7, fig. 830) and at intervals applied to its wall, and termi-

22 The description of series 23a was presented by the writer before the last or
28th session of the American Association of Anatomists (1911) and published in
the June number of the Anatomical Record, vol. 6, no. 5, 1912, as a part of a pre-
liminary paper, entitled The value of the injection method in the study of lym-
phatic development. In this connection it is also well to call attention to Profes-
sor McClure’s article, A few remarks relative to Mr. Kampmeier’s paper on the
value of the injection method, etc., which appeared in the same number of the
Anatomical Record, and which is a critical analysis of some of the papers published
by Professor Sabin and Dr. Eliot Clark on the development of the lymphatic
system.

438 OTTO F. KAMPMEIER

nates just below the level of the right Cuvierian duct at X. There
is also a long and spindle-shaped space which les lateral to it
and in the pathway of a tributary of the anterior lymphatic
plexus were it prolonged downward, as suggested in figures 13
and 30

Passing to the posteardinal division (B, fig. 30) of this embryo,
we meet with the most decisive evidence in favor of the non-venous
origin of the thoracic duct, namely, a clear case of discontinuity
in its anlage than which nothing could be more conclusive. Im-
mediately following the injected portion of the right duct (Sd) is a
long fusiform mesenchymal space (4d), but in no way connected
with it, as exemplified by the drawings, figures 13, 30 and 31
(ventral view), and the microphotographs, figures 15, 16 and 17
which represent transverse sections taken at this level. Espe-
cially the ventral view of the reconstruction (fig. 31) illustrates the
abrupt break (X) in the duct anlage, the position of the terminal
portion of the injected channel (5d), and the independent fusi-
form space (4d) and its longitudinal extent. The injected channel
ends obscurely in a ‘mossy’ area produced by slight extravasa-
tions, the position of which is indicated at X in figure 15 just
ventral to the broad lumen of the anterior tip of the independent
space (4d). That there is absolutely no open communication
between. these two segments of the duct-anlage is strikingly con-
firmed by both observation and experiment. In the first place
the most critical examination with the high powers of the micro-
scope was not able to detect continuity, and secondly, not a
particle of the injection mass was found to have entered the cavity
of the blind fusiform space (4d), although the pressure of the
injection was sufficiently great to produce the extravasations
referred to above.

Fig. 13 A simplified or schematic drawing of an accurate reconstruction of
the thoracic duct region in series 23a (Johns Hopkins University Embryological
Collection) represented in figure 33. The lymph sac and the injected portion of
the thoracic duct anlage were drawn in black; the uninjected lymph spaces are
discontinuous but are located in the axes of the injected channels and consequently
in the paths of the future complete thoracic ducts. The cross lines indicate the
levels at which figures 14, 15, 16 and 17 were taken.

JUGULAR LYMPH SAC
(INJECTED)

EXTERNAL JUGULAR

Fig. 14

LEFT SUBCLAVIAN LYMPHATIC PLEXUS

L. THORACIC DUCT

UNINJECTED LYMPHATIC
SPACE

LEFT PRECARDINAL

R.~THORACIC DUCT

UNINJECTED LYMPHATIC

SPACES RIGHT DUCT OF CUVIER

Fig. \6 16
Fig.17 17

LEFT POSTCARDINAL

UNINJECTED LYMPHATIC
SPACES

LEFT SUPRACARDINAL
( Azycos )

RIGHT SUPRACARDINAL
( azycos )

439

440 OTTO F. KAMPMEIER

3 t

: at gees ‘
we oh Rat
nak sere ; al

Fig. 14 Transverse section through the left lower cervical region in a 23 mm-
pig embryo (series 23a, J.H.E.C., slide 21, section 16), * 200. 3, anterior lympha-
tic plexus of the eee ic duct injected; Hx, extravasations of the injection sub-
stance into the surrounding mesenchyme; 9, internal jugular; 14, sympathetic
nerve trunk; 17, oesophagus. (Reconstruction, fig. 30.)

The discontinuous fusiform lymphatic space (4d) is of consider-
ablelength, capable of being followed through thirty-seven sections
(thickness of sections: 20 micra), and it is variable in diameter (figs.
31, 15, 16 and 17), at times being very broad and at other times
narrow and not so sharply demareated from the intercellular
lacunae of the mesenchyme surrounding it. In form it is very
irregular, and its lumen is often bridged by tissue strands of
greater or lesser thickness which give to it a multilocular appear-
ance as shown in cross-section. in figures 16 and 17 (4d). This
condition and the fact that it is bounded by ordinary mesen-
chymal cells supply strong proof against its venous origin. The
difference between its lining and that of the neighboring venules
(25) and veins is strikingly expressed even in figure 15, in which its
boundary is quite regular and clear-cut but the greater delicacy

THORACIC DUCT DEVELOPMENT IN THE PIG 441

Fig. 15 Transverse section taken shortly beyond the right Cuvierian duct
in a 23 mm. pig embryo (series 23a, J.H.E.C., slide 26, section 10), X 200. 4d,
anterior tip of the long fusiform lymphatic space in the right thoracic duct line;
X, position of extravasated particles from the injected portion of the right thora-
cic duct anlage; 6b, posteardinal veno-lymphatic; //d, right postcardinal; 13,
aorta; 17, oesophagus; 25, venules or branches of the posteardinal. The more deli-
cate lining of the lymphatic space as compared with that of the veins and venules
can be clearly distinguished in the figure. (Reconstruction, fig. 30.)

of its wall can be distinguished without the least difficulty.
Figure 17, again, illustrates the occasional circumclusion of the
precardinal veno-lymphaties (6b) by this space and draws more
plainly, perhaps, the distinction between lymphatics and venous
channels, where the latter are replete with blood and possess
sharply defined boundaries as compared with the often ill-defined
outlines of the lymphatic space. Caudally this long space after
a course which can be easily pursued through thirty-seven sec-
tions, as already stated, becomes more indistinct until it vanishes
by the loss of its cavity in the confusion of the interstices of the
tissue reticulum, but after a number of sections it is followed by a
second space, which, though shorter and simpler (figs. 13 and 31),

44? OTTO F. KAMPMEIER

x, | = 1%

Z : 4 bad : [a Sees, : be at sede Jee

tz 21 4.5

Fig. 16 Transverse section taken a few sections beyond the one represented by

the preceding figure from a 23 mm. pig embryo (series 23a, J.H.E.C., slide 26,

section 12), * 200. 4d, long fusiform space in the right thoracic duct line, and

mesenchymal bridges traversing its lumen; 4s, lymphatic space at the level of the

left Cuvierian duct (21) and in the pathway of the left thoracic duct; 11d, 11s,

right and left postcardinals; 13, aorta; 17, eosophagus; 25, branch of the post-
eardinals. (Reconstruction, fig. 30.)

exhibits the same peculiarities of character. This again is fol-
lowed by tissue which is still undifferentiated but coarsely retic-
ulate and persistently suggests the potentiality of further lympha-
tic anlagen. Both of the spaces described and figured are situated
in the axis of the injected channel and consequently in the axis
of the ultimately complete thoracic duct.

On the left side in series 23a the principle of lymphatic develop-
ment is the same and is expressed fully as well as on the right
side. The injected segment of the left thoracic duct limb (és,
figs. 30 and 13) is much shorter than that of the right, but it is
slender and often it can only be traced by a ‘mossy’ path due to

THORACIC DUCT DEVELOPMENT IN THE PIG 443

shght extravasations. At intervals beyond the farthest extent
to which the injection mass has penetrated (X, fig. 30), and
located in a line destined to become the pathway of the future
thoracic duct, are a number of small blind mesenchymal vacuoles
(4s), the largest one of which extends through eight sections at the
level of the Cuvierian duct. Being hidden by the veins in a dorsal
view of the reconstruction, their positions are indicated on the
drawing by dotted circles (4s, fig. 30; see also fig. 13). The con-
spicuous size of the lumen of the last space (4s) and the mesen-
chymal strand bisecting it, as illustrated in the microphotograph,
figure 16, require no further comment.

Pee. at ae af ee
ae ease
aes ;

a
Fas

Fig. 17 Transverse section through the thoracic region in a 23 mm. pig embryo
(series 23a, J.H.E.C., slide 27, section 18), * 200. 4d, long fusiform lymphatic
space in the right thoracic duct line, and tissue bridges traversing its lumen;
6b, postcardinal veno-lymphatic surrounded by this space; 11d, right postcardinal;
13, aorta; 17, oesophagus; 25, venule, branch of the postcardinal. (Reconstruc-
tion, fig. 30.)

444 OTTO F. KAMPMEIER

In series 192 a 21.5 mm. embryo slightly older than the preced-
ing embryo 23a, the right thoracic duct anlage (5d) extends as a
continuous channel back to the point X, as indicated on figure 35.
On comparing this with figure 30, it can be observed that the
longer portion of the posteardinal division of the right anlage,
which in Sabin’s series 23a exists in the form of the long fusiform
space described above, has established connection with the pre-
cardinal thoracic duct segment, and in this way it has increased
considerably the length of the channel joined tothe jugular lymph
sac. Like the fusiform space of that embryo, the posteardinal
thoracic duct division of series 192 is irregularly beaded or vari-
cose, the constrictions or nodes suggesting more recent fusion
between successive internodes. That this suggestion is a fair
one is substantiated by the fact that toward the region of the peri-
aortic veno-lymphaties, or beyond the point X (fig. 35) where the
continuous anlage ends blindly, it is followed at intervals by a few
large mesenchymal vacuoles (4d) between which no communica-
tion is as yet noticeable save through the indifferent tissue net-
work.

‘The continuous portion of the left thoracic duct anlage in series
192 is perhaps no longer than in Sabin’s series 23a, but the blind
lymphatie spaces (4s, fig. 35) following it are far more extensive
in length, especially those in the region of the left Cuvierian duct.
Here there are two long spindle-shaped spaces parallel to each
other, the shorter one of them being that portion of the anlage
of the future mediastinal lymphatic vessel situated near the point
of its subsequent junction with the other, or longer space, which
represents an anlage of the left thoracic duct. In a 22 mm.
embryo, a slightly older stage, all of the blind spaces of the pre-
and post-cardinal divisions have become confluent to form the
uninterruped duct and its mediastinal tributary.

(C) Supracardinal division. In the third division of the thor-
acic duct area there may berecognized an anterior and a posterior
half, those regions, respectively, in which the duct-anlage during
its initial development is associated with the periaortic (6c, fig.
3) and the posterior supracardinal (6c, fig. 4) veno-lymphatics.

THORACIC DUCT DEVELOPMENT IN THE PIG 445

4a

Fig. 18 Transverse section through the region of the mesonephroi in a 21 mm.
pig embryo (series 103, slide 32, section 9), X 200. 4, lymphatic spaces or dis-
continuous thoracic duct anlagen forming near or against the periaortic veno-
lymphatics (6c); 4a, potential lymphatic spaces; 8, dorsal segmental vein; 12d, 12s,
right and left supracardinal veins; 13, aorta.

In series 103 (21 mm.) is foreshadowed the decadence of the
periaortie veno-lymphaties and the transference of supremacy to
their successors. Frequently throughout their course the veno-
lymphatic channels have lost their former fullness and their endo-
thelium has been thrown into a slightly wavy and uneven contour,
this being especially true of that side of the plexus facing the
aorta where these areas of weakness are more abundant and accen-
tuated. Coexistent with this condition is an incipient vacuolation

446 OTTO F. KAMPMEIER

of the mesenchyme by which large and small fissures and spaces
(4, fig. 18) arise irregularly and indiscriminately along thechan-
nels designated. Sometimes these spaces (4) cling closely tothe
receding walls of the venules (6c), and at other times they are
separate with a perceptible amount of tissue intervening, but they
are always discontinuous and non-venous in character. Like the
more anterior lymphatic anlagen in their inception, no visible
difference either of form or arrangement can be discerned
between the cells which comprise their circumference and the
cells of the intricate meshwork of the mesenchyme. Strands and

Fig. 19 Transverse section immediately in front of the mesonephroi in a 22
mm. pig embryo (series 105, slide 36, section 1), X 150. 4, lymphatic spaces
replacing the periaortic veno-lymphatics (6c); 8, dorsal segmental vein; 12d, 12s,
right and left supracardinal veins; 13, aorta; 14, sympathetic nerve trunk; 19,
embryonic vertebral column. (Reconstruction, fig. 33).

THORACIC DUCT DEVELOPMENT IN THE PIG 447

fibrils or their spur-like remnants often bridge the spaces or pro-
ject into them and seem to suggest the existence of stresses and
strains, as well as their direction, in the production of the anlagen
by the breaking down of barriers and the fusion of interstitial
spaces.

A later stage in the transformation of the periaortic region is
offered by series 105, a 22 mm. embryo; but a minute account is
hardly necessary considering the clearness of the appended figures
which are self-explanatory and almost sufficient in themselves.
Figure 19 is from a section. taken just in front of the mesonephroi
and is representative of the conditiors active along the whole
range of the periaortic plexus. The vero-lymphatics (6c) have
lost most of their connections with the supracardinal veins and
throughout the greater part of their course present a shrunken
cavity filled with the deeply staining débris of blood cells. The
lymphaties (4) which enmesh them are either broad spaces and
have obliterated the verous core almost completely, or small
crevices hugging one side of a vessel which has just begun to
manifest degeneration. They are irregular in arrangement but
they may always be distinguished by their clear lumina andun-
specialized walls. A segment of the proximal portion of the
supracardinal division in a 22 mm. embryo was reconstructed,
a drawing of which is reproduced in figure 33, illustrating in three
dimensions the conditions described.

The cisterna chyli is the outcome of a number of changes which
proceed in very rapid succession and at the beginning often occur
simultaneously: the detachment of the posterior supracardinal
veno-lymphaties from their venous trunks; the condensation of
these abandoned channels progressively toward the production
of a plexiform or multilocular channel; the recession of their
intima; the breaking down of broad partitions of tissue between
them; the expansion of the resultant cavity by the addition of
spaces from the mesenchyme; the simplification of its lumen and
the acquisition of a lymphatic endothelium. In other words, a
large part of the cavity of the cisterna chyli is derived from the
combined cavities of preéxisting venous channels, but its wall is
newly differentiated from the mesenchyme. <A consideration of

448 OTTO F. KAMPMEIER

“iy eas
nari Be nay)
oat ae ie

4p Pata

why "ted ht
bie .

2

13
Fig. 20 Transverse section through the region of the future cisterna chyli
in a 21 mm. pig embryo (series 103, slide 39, section 12), * 200. 4, clear areas
representing the formation of lymphatic spaces around the posterior supracardinal
veno-lymphatics (6c); 13, aorta; 14, sympathetic nerve trunk.

the features revealed in 21, 22 and 23 mm. embryos will give sup-
port to these observations.

While the veno-lymphaties of the cisterna chyli region are anas-
tomosing extensively in the median line dorsal to the aorta, they
are beginning to lose their connections with the supracardinals of
both sides so that they come to exist in the form of a condensed
and abandoned multilocular venous plexus. Coincident with this
process, the mesenchyme located between the plexiform channels

THORACIC DUCT DEVELOPMENT IN THE PIG 449

often become less compact, vesicles and fissures arise in it, and the
venous intima, contiguous to these rarefied tissue-areas, retracts
and breaks down; thus initiating continuity between the originally
distinct channels. <A section of sucb an incipient stage is illus-
trated in figure 20 which shows very clearly the vesiculated mes-
enchyme (4) in and amongst the veno-lymphatic plexus (6c).
Typical examples of extra-intimal replacement also become very
abundant and are significant as confirming the evidence already
given for the reorganization of the intima whenever a lymphatic
channel appropriates the pathways of redundant veins or venules.
A photograph (fig. 21) of such a condition in a 22 mm. embryo
describes more plainly than an extended narrative the important
features in the transition period of the ciserna chyl. The com-
partments or loculi of the potential cisterna are traversed by dis-
tinct, delicate and devious lines which upon closer examination
are found to be composed of compressed or scale-like cells, placed
end to end, and to represent the discarded endothelium of the
former venous derivatives (6c). This is shown when the endothe-
lium is followed either forward or backward and it can be found
occasionally lining a sharply defined cavity containing blood and
then again to be pushed far into the lumen apparently by the
pressure of the fluid within large mesenchymal spaces (4) on its
external surface. At still other places where fusion of several
parallel channels has occurred simultaneously, this evanescent
venous intima is visible in cross-section as torn fibrils pendant
from the irregular and frayed walls, or lying isolated in the lumen
of the new or compound cavity. Besides these vestiges of the
venous intima there are broader and thicker shreds of tissue,
which are composed of a mass of ordinary mesenchymal cells
jutting into the cavity and which indicate therefore the position
of former boundaries between separatechannels. Examination
of the sections will also show distinctly that the outlines of the
perivenous spaces are ill-defined and radically unlike those of the
venous channels which they surround. In the confines of the
transitional cisterna-anlage the irregular elliptical or cuboidal
mesenchymal cell is the prevailing type and exists in strong con-
trast with the flattened and dense endothelial cell of a normal

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 4

450 OTTO F. KAMPMEIER

Fig. 21 Transverse section through the cisterna chyli region in a 22mm. pig
embryo (series 105, slide 42, section 8), * 200. 4, anlage of the cisterna chyli
showing extra-intimal replacement of the supracardinal veno-lymphaties (6c) ;
14, sympathetic nerve trunk; 13, wall of the aorta.

vein or of the rejected and defunct intima lying in the lumen
of this embryonic lymphatic.

The enlargement and perfection of the developing cisterna chyh,
as well as of the other segments of the thoracic duct, will be con-
sidered in the treatment of the succeeding and final phase.

3. The lymphatic phase (22-28 et seq. embryos)

At the period when lymphatic spaces are appearing around the
periaortic veno-lymphatics and are growing in volume, the thor-
acic duct anlage has already become a continuous structure in the

THORACIC DUCT DEVELOPMENT IN THE PIG 451

pre- and posteardinal divisions. Almost concurrently the changes
producing the cisterna chyli and its connection with the posterior
and mesenteric lymph sacs have been active, so that the segment
of the duct anlage in the territory of the periaortic vessels, at the
level of the mesonephroi, is the last one to acquire continuity
among the lymphatic spaces and to make the thoracic duct an
unbroken tube from one extremity to the other. The reason
that those portions of the duct nearest to the lymph sacs are
developed first, that the vacuolation of the mesenchyme, the
formation of isolated spaces, and their confluence proceeds in
a general centrifugal direction is not far to seek, being probably
inherent in the explanation which would make the accelerating
pressure of the lymph stream towards its points of entry to the
veins sufficient to account for, or at least sufficient to furnish
the stimulus for, the progressive occurrence of such phenomena.

The elongation of lymphatic spaces and their fusion finally
into a continuous channel, as well as the growth of their cavities
in diameter, is accomplished by the same process which gave
origin to them, namely, by the disintegration of tissue fibrils and
the concentric addition of spaces. Figure 22 represents a typical
transverse section from the precardinal division of an early lym-
phatic stage and illustrates very plainly how the increment in
size of the thoracic duct anlage is effected. An accurate camera
lucida drawing, figure 23, of a portion of the same section is also
inserted here to bring out more distinctly some of the details
which may be obscure in the microphotograph due to the differ-
ences in focus. Both of these figures demonstrate that the duct
(5d) at this embryonic period enlarges by a process of growth not
from within outward but from without inward, by the admission
of adjacent mesenchymal spaces toits channel. The strands inter-
secting the lumen are therefore indicative of successive lines of
fusion or a measure of its gradual growth. From the facts just
stated and the exceedingly indefinite boundaries of the anlage,
we should expect the absence at this time of any kind of demarca-
tion membrane between the cavity and the interstices or lacunae
of the surrounding tissue. That such is really the case is borne
out by experiment. Both the veins and the thoracic duct of an

452 OTTO F. KAMPMEIER

early lymphatic stage (283 mm. embryo) were injected and there-
fore are favorable for a comparison of their reactions to the injec-
tion mass. The veins (//d, 11s, 25, fig. 24), possessing perfect
endothelial walls, did not admit of any extravasations or even
of blurred outlines. In the case of the thoracic duct (5d), on the
contrary, the injection mass (Ha) passed freely from the lumen
into the surrounding tissue reticulum, as pictured clearly in figure
24, showing the absence of any definite wall at this early embryonic
period.

Figure 25 is from the lower postcardinal division in a 26 mm.
embryo and reveals essentially the same features as portrayed in
figure 22 but shows even better perhaps the multilocular character
of the thoracic duct anlage. The formation of spaces from the
mesenchyme and their addition to the anlage is very clearly
expressed on the right side in the wedge-shaped territory (5d)
between the oesophagus (17) and aorta (73), and also on the left
side (5s) immediately ventral to the left posteardinal vein (1/s).

A later stage (23 mm.) in the development of the cisterna
chyli (5) is shown in figure 26. The vestiges of the antecedent
veno-lymphatics have completely disappeared, and only occa-
sional trabeculae indicate the originally extensive tissue parti-
tions between the early rudiments of the cisterna. What is of.
greater significance, however, is the ragged outline of the cavity,
the absence of any specialized endothelium, and the addition of
small mesenchymal spaces (4) to its lumen, upholding therefore
in every respect the writer’s contention that the cisterna chy,
concordant with the anterior divisions of the thoracic duct, is
primarily and fundamentally a product of mesenchymal differ-
entiation.

Fig. 22. Transverse section through the upper thoracic region in a 21.5 mm.
pig embryo (series 192, slide 21, section 11), & 200. 4d, right thoracic duct anlage
showing its enlargement by the addition of adjacent mesenchymal spaces; 11d,
right posteardinal; 13, wall of the aorta; 17, oesophagus; (compare with fig. 23).

Fig. 23 An accurate camera lucida drawing of a highly magnified area repre-
sented in fig. 22, * 266 (reduced from 400). 4d, right thoracic duct anlage
showing its concentric growth from enlarged tissue spaces and the absence of a
specialized intima: //d, right posteardinal vein filled with blood and possessing
a well-defined endothelial lining; 13, wall of the aorta; 17, oesophagus.

THORACIC DUCT DEVELOPMENT IN THE PIG

454 OTTO F. KAMPMEIER

Fig. 24 Transverse section through the upper thoracic region in a 23 mm. pig
embryo (injected series, slide 15, section 11), < 100. 4d, right thoracic duct
anlage injected; Ex, extravasations of the injection mass from the right duct-
anlage into the loose surrounding mesenchyme; ds, left thoracic duct anlage; 11d, 11s,
right and left postcardinals injected; 13, aorta; 17, oesophagus; 25, branches of
the posteardinals injected.

As shown in figures 22, 23, 25 and 26, at the beginning of the
lymphatic phase the thoracic duct in transverse section resembles
a condensed plexus and may be said to be at the height of its
complexity. For from now on there is a gradual reduction of
this condition until the two limbs of the thoracic duct exist nor-
mally as single simple channels. By the breaking down of the
abundant tissue bridges which had divided its channel into a
labyrinth of loculi, the ducts assume more and more the appear-
ance of unobstructed tubes. In figure 27, taken from a 26 mm.
embryo, the outlines of the ducts (4d, 5s) are assured, and only
the vestiges of former septa and trabeculae are still visible here
and there in the form of few small tissue spurs and filaments pro-
jecting into the cavities.

The most startling change, however, occurs in the precardinal
division in the region of the jugular lymph sac. The reader will

THORACIC DUCT DEVELOPMENT IN THE PIG 455

—_

SSS . SENSES oe —
SSS =

oe SS SSS
SSS

* SS
ie
La Nts

Fig. 25 Transverse section through the postcardinal region in a 26 mm. pig
embryo (series 69, slide 44, section 11), * 150. 4d, 4s, right and left thoracic
ducts showing the manner of their growth by the addition of spaces from the mesen-
chyme; 11d, 11s, right and left posteardinals; 13, aorta; 17, oesophagus; 19, embry-
onic vertebral column.

recall that in the later transition stages the most anterior portion
of the thoracic duct anlage is characterized by an extensive and
complicated lymphatic plexus, especially well developed in Sabin’s
series 23a (3, fig.30). During the lymphatic phase such a plexus is
completely reduced and converted into a simple channel (22—26
mm. embryos) whose fork or division into the right and left limbs
of the duct becomes shifted far back of the lymph sae. Series
192 (3, fig. 35) presents an intermediate stage in which the orig-
inal plexiform condition is still suggested but has been almost
entirely obliterated by the transverse fusion and consequent
reduction of the number of interanastomosing channels.

No matter whether the thoracic duct anlagen arise as extra-
intimal spaces or entirely apart from the veno-lymphatics, dur-

456 OTTO F. KAMPMEIER

ing their genesis their walls are composed of the undifferentiated
mesenchymal cells (figs. 6, 9, 18, etc). We have also seen the
method of concentric addition of tissue spaces by which the
anlagen enlarge during the later transition and early lymphatic
stages, implying thus a continual shifting of their boundaries so
that an intima may be said to be temporarily established, replaced,
and reorganized a number of times (figs. 22, 25, ete.) Ultimately,
however, we can speak of a definite and permanent endothelium

5
papa FREES ae s i 1
fe: 3 ;

be gee

Fig. 26 Transverse section through the region of the cisterna chyli in a 23 mm.
pig embryo (series 67, slide 43, section 14), X 150. 45, anlage of the cisterna chyl
showing the indefinite and ragged outline of its cavity, the absence of a specialized
intima, and the addition of small spaces (4) from the mesenchyme; the vestiges
of the former veno-lymphatics have completely disappeared; 13, aorta; 14, sym-
pathetic nerve trunk.

THORACIC DUCT DEVELOPMENT IN THE PIG ADT.

12d 12s
cag etree ol at
4s oe " e a <n a

Fig. 27 Transverse section through the anterior supracardinal region in a
26 mm. pig embryo (series 69, slide 53, section 10), & 150. 4d, 5s, right and left
thoracic ducts; 12d, 12s, right and left supracardinal veins; 13, aorta; 14, sympa-
thetic nerve trunk.

only after the thoracic duct has lost its multilocular and _ plexi-
form character, and its channel approaches more nearly to a
clear-cut and simple tube (4s, fig. 27). The manner in which the
mesenchymal cell is transformed into an endothelial cell occurs
undoubtedly, as Huntington has suggested, by a mechanical
adaptation tc the pressure of the fluid within the lymphatic
cavity. This view is entirely in harmony with the conditions
observed in the extra-embryonic area of the chick?? where the

23 At the last session of the American Association of Anatomists, December 27,
1911, at Princeton, N. J., John E. McWhorter and Allen O. Whipple of the College
of Physicians and Surgeons, Columbia University, presented a report on the devel-
opment of the blastoderm of the chick in vitro. This report appeared as a short
preliminary paper with twelve figures in the Anatomical Record, vol. 6, no. 3,
March, 1912. These investigators, on the basis of a study of the living growing

458 OTTO F. KAMPMEIER

earliest anlagen of the blood vessels arise between. the mesodermal
cell-strands as isolated spaces and fissures, which at first are
bounded by ordinary cuboidal cells but later acquire the charac-
teristic vascular endothelium by the modification of these cells.
Being plastic, the cells lining either a haemal or a lymphatic
anlage must be regarded as obeying the internal pressure of the
cavity and becoming more and more flattened and endothelial-
like as the pressure of the fluid or plasma increases.

Although this work deals primarily with the source of the thor-
acie duct, attention was not confined to it exclusively but also
considered briefly two other lymph ducts, the mediastinal chan-
nel draining the mediastinum and its organs, and the right lym-
phatie duct, which in earlier stages of phylogenesis undoubtedly
composed a part of the thoracic duet system but now ordinarily
remains independent of it and receives tributaries only from the
cephalic, cervical and upper thoracic regions of the right side.
In development these channels repeat in all details the history of
the thoracic duct, arising as isolated mesenchymal or perivenous
spaces which subsequently become confluent.

After the above description of the development of the thoracic
duct and a consideration of the evidence presented, attention
may be directed towards several criticisms advanced recently by
those investigators opposed to the view of the direct mesenchymal
origin of lymphatic vessels. These opponents would dismiss as
artifacts all of the ‘lymphatic anlagen’ described by the writer.
Sabin, referring to the figures of extra-intimal replacement dis-
covered by Huntington and McClure in their investigations on the
genesis of the lymphatic system in the cat, maintains that “‘they
are all in the center of the embryo where the fixing fluid pene-

chick embryo, brought forth conclusive evidence that the haemal channel system
of the extra-embryonic area is developed from isolated spaces, which arise blindly
in the undifferentiated mesenchyme and which subsequently change their shape
by expansion and elongation and become confluent with other such spaces to pro-
duce the complicated blood plexuses in the area designated. The ‘“‘spaces are
frequently bounded by a mere line, more or less refractile in character. In others
the lumen is lined with rounded or oval cells which later become fusiform and
flattened.

THORACIC DUCT DEVELOPMENT IN THE PIG 459

trates last;’’4 and concerning all ‘mesenchymal spaces,’ Clark
says: ‘‘ They occur most often around blood vessels, and are almost
certainly to be interpreted as shrinkage. spaces, or spaces caused
by the retraction of the mesenchyme processes made possible by
slight rents produced in the preparation of the sections.’2> That
these objections are wholly without foundation, at least in the
case of the developing thoracic duct in the pig, is conclusively
shown by the following observations: In the first place, the mesen-
chymal spaces termed lymphatic anlagen spring into existence at a
definite period of embryonic history and invariably in a definite
position. The embryos grouped under the veno-lymphatic phase
are without an exception the younger embryos, and, although they
were preserved in the same fixatives and the preparation of the
sections followed the same methods as those employed for the
specimens of the second phase, they do not show or even suggest
instances of such anlagen. Secondly, in all of the earlier stages
of the transition phase these lymphatic spaces can only be ob-
served in the precardinal division, namely, in the territory of the
most anterior segment of the thoracic duct which is formed first;
in the posteardinal and supracardinal divisions the mesenchyme
is still uniform, and the veno-lymphaties are functional and
joined to the parent veins. In the later stages of this phase the
last two segments repeat the history of the first. Thirdly, all
of the extra-intimal spaces in the thoracic region of the writer’s
specimens occur only in connection with those venules which
have been severed from their venous trunks and which lie topo-
graphically in the pathway of the potential duct. As far as the
writer was able to determine, other abandoned veins existing
in the same general areas and in the ‘center of the body,’ but not
antecedent in position to the duct or to any of its tributaries,
never manifest extra-intimal figures in their atrophy. All of
the functional veins possess a normal and distended endothelium.

*4 Florence R. Sabin: A critical study of the evidence presented in several recent
articles on the development of the lymphatic system. Anat. Ree., vol. 5, no. 9,
1911.

5 Kliot R. Clark: An examination of the methods used in the study of the devel-
opment of the lymphatic system. Anat. Rec., vol. 5, no. 8, 1911.

460 OTTO F. KAMPMEIER

Fourthly, these mesonchymal perivascular spaces may have fused
into a profuse plexus and become widely open to the jugular
lymph sac as in the ease of series 194 (fig. 29), but they, as yet, do
not show a well-defined or specialized wall (figs. 8 and 9). In the
fifth place, all of the discontinuous mesenchymal spaces follow
one another in. a succession practically undeviating which repre-
sents an outline or fragmentary picture of the future duct. Out-
side of this line there are no lymphatic spaces. Sixthly, in the
third or lymphatic phase, when continuity of the duct and its
branches has been established, no perivenous or other isolated
vacuities can be discovered. If the discontinuous lymphatic
anlagen were artifacts we should expect to find the largest num-
ber of them in this last phase because the diameter and the bulk
of the embryos are greater, and therefore the longer time required
for the fixing fluid to penetrate to their centers would make
possible greater uneveness of fixation and consequently greater
shrinkage.

Because the elongation of the thoracic duct is effected by a pro-
gressive summation or centripetal addition of large mesenchymal
spaces to that part of the anlage already confluent with the lymph
sacs, the injection of successive transition stages up to the time
when continuity has been acquired throughout its entire course
will show a gradual increase in the length to which the injection
mass has penetrated; but the study of serial sections will also
reveal anlagen which lie beyond the farthest point of the injec-
tion and are inaccessible to it on account of their discontinuity,
or because they have not as yet become confluent with the anlage
into which the injecta were introduced.

This leads to a second contention of Sabin, namely, thatthe
study of serial sections alone is inadequate, and that continuity
of the apparently discontinuous lymphatic anlagen can be demon-
strated by complete injection. A more radical refutation of this
argument than that furnished by her own series 28a is scarcely
possible. The abrupt break between the precardinal injected
segment of the right thoracic duct anlage in this embryo and the
posteardinal uninjected segment (fig. 31) bears out in a striking
manner the evidence derived from the writer’s series. Notwith-

THORACIC DUCT DEVELOPMENT IN THE PIG 461

standing the inability of the eye to discover a connection between
these two segments with the aid of high magnifications, 1t might
be urged by those prejudiced that the injection may have been
only a partial one. But this objection becomes groundless when
the reader recalls that the pressure of the injecting fluid was of
sufficient force to produce extravasations, which, as Clark main-
tains, signify an excess of pressure in filling the cavity completely ;
for he says, ‘‘With too great pressure there is produced a mossy
appearance around the capillary (lymph), as has been pointed out
by Hoyer, due evidently to forcing the injection mass through the
lymphatie wall.’ If an opening had been present between these
two anlagen the injecting substance would certainly have obeyed
the direction of least resistance and passed into the second one.
Nor is the objection valid which would exclude this large blind
fusiform space from taking any significant part in the production
of the thoracic duct; for not only is the distinct character and
position of this space contrary to such a view but also the fact that
the left side discloses similar spaces located in the identical line of
the future left duct. Somewhat later embryonic stages make
these observations conclusive; for example, in series 192 the post-
cardinal segment of the right duct duplicates or agrees in all of its
features with that of series 23a, except for its continuity with the
anterior or precardinal segment (figs. 30, 31 and 35) and conse-
quently with the jugular lymph sac. Moreover, during the prog-
ress of his investigation the writer has tentatively assumed the
possibility of a centrifugal growing of thoracic duct buds through
the large mesenchymal spindle spaces situated only in the thoracic
duct pathway, and he has searched for such hypothetical sprouts
but has not sueceeded in finding a trace of evidence in their favor.

Sabin’s and Clark’s contention that discontinuities in a lym-
phatic channel are due to artifacts, resulting during fixation
from the unequal shrinkage here and there of its caliber, is easily
controverted by the observed facts. In the case of the develop-
ing thoracic duct such discontinuities only occur in the stages of
the transition phase, in those embryos measuring approximately
between 20 and 23 mm. The discontinuous segments or anlagen
begin as minute mesenchymal vacuoles which gradually enlarge

462 OTTO F. KAMPMEIER

and elongate with the increasing age of these embryos; in other
words, in a 22 mm. embryo the blind segments of the duct anlage
will be much longer and more conspicuous than in a 20 mm. em-
bryo for instarce. Further, there is a positive regularity in the
progressive reduction of the number of these blind lymphatic
anlagen in a general antero-posterior direction by their addition
to the continuous anlage, which, as a consequence, gradually
becomes elongated. Were these lymphatic spaces artifacts, or
segments cut off from a continuous channel by shrinkage,then
the determinate sequence of genetic changes pointed out in the
descriptions of the individual stages could not exist, and we should
find them in slightly older embryos or in those portions of the
duct-anlage definitely known to be complete, for the same methods
of technic should produce similar effects. The embryonic tho-
racic ducts when fully formed and indeed all lymphatic vessels
possess a varicose channel constricted and dilated alternately into
irregular nodes and internodes. Such a condition, however, is
not brought about by fixation but is a characteristic peculiar
to a lymph vessel and obviously harks back to the period when it
was composed of a varying number of irregular fusiform or oblong
meser.chymal spaces succeeding one another with distinet inter-
ruptions. Accordingly, the nodes or constrictions of a thoracic
duct just completed would indicate the areas of final fusion
between consecutive anlagen.

It should be emphasized here that Sabin and Clark base their
criticisms chiefly upon the latter’s investigations on the develop-
ment of the lymphatic capillaries in the tail fin of the larval frog.
The fallibility of their argument becomes therefore further evi-
dent when we find them comparing the reaction to the fixatives
of these terminal lymphatics with that of other lymph channels,
especially the larger ducts and trunks; for, although the principle
of development probably is the same in both cases, the details
of their behavior during the preparation of the sections may be
quite different. It would be just as logical to describe a large
systemic artery or vein entirely in terms of their terminal arterioles
or venules. The writer will not deny that careless or imperfect
fixation may cause the delicate capillaries of the fin of a tadpole

THORACIC DUCT DEVELOPMENT IN THE PIG 463

to shrink into seemingly isolated segments so that they can be
pursued only with great difficulty, as described by Clark, but he
does deny, supported by the decisive evidence of the injected
series 23a and reinforced by all of the transition stages, that the
discontinuous anlagen observed by him and invariably found to
be concomitants in the formation of a large lymphatic trunk like
the thoracic duct are artifacts, produced by the preserving or
fixing reagents.

V. RESUME OF OBSERVATIONS AND CONCLUSIONS

1. Derived from the supracardinal or azygos system of veins,
a series of venous channels, called veno-lymphatics, are formed
in the pathway finally occupied by the thoracic duct, and at the
culmination of their development they exist as plexuses of vessels
abundantly connected with the parent veins.

2. The actual genesis of the thoracic duct is initiated by the
appearance of blind mesenchymal lymphatic spaces either around
or not immediately in contact with the venous derivatives, or
veno-lymphatics, which become detached from their venous
trunks and break up into degenerating segments. The lympha-
tic spaces or anlagen arise by the local disintegration of the fibrils
of the tissue reticulum and the fusion of the interstitial lacunae,
and they enlarge and elongate in a similar manner. If they are
of the nature of extra-intimal spaces the endothelium of the evan-
escent abandoned veno-lymphatics, which they replace, collapses
as the result perhaps of the increasing influence of the lymph
pressure on its external surface after its release from the blood
pressure. During their inception and growth the walls of the
discontinuous thoracic duct anlagen are composed of the ordinary
unmodified mesenchymal cells. That such lymphatic anlagen
are not artifacts is shown by their definite position and period of
formation and the determinate sequence from their first appear-
ance as mesenchymal vacuoles, through the phase of their growth
and elongation, to their final fusion into a continuous channel.
In the production of the most posterior portion of the thoracic
duct, or cisterna chyli, veno-lymphatic channels by fusion with

464 OTTO F. KAMPMEIER

one another give rise to the larger part, perhaps, of its cavity;
but at the same time their endothelium recedes and degenerates,
and the cisterna-anlage increases in size by the addition of spaces
from the mesenchyme, so that, like the more anterior segments of
the thoracic duct anlage, it is bounded by ordinary embryonic
tissue cells during this early developmental period.

3. The elongation and final continuity of the thoracic duct
anlage is effected by the progressive confluence of discontinuous
fusiform lymphatic spaces in a general centrifugal direction, prob-
ably determined by the impulse of the lymph flow towards the
radiation centers or lymph sacs. Injected specimens of the
early lymphatie stages certify the reality of blind uninjectible
anlagen beyond the farthest points to which the injecta have
penetrated, demonstrating that discontinuities in a developing
lymphatic channel are not ‘appearances’ found only by the study
of uninjected embryos. Not a shadow of evidence was discovered
in favor of the theory which maintains the centrifugal growth of
the duct by budding from the lymph sacs or the derivation of the
lymphatic endothelium from the veins. During the period of its
initial growth the thoracic duct increases in diameter by the con-
centric addition of enlarged and immediately surrounding tissue
spaces to its lumen. The intima of the thoracic duct is a differen-
tiation in situ of mesenchymal cells as an adaptation probably
to the pressure of the lymph flow within the cavity.

PLATES

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 4

465

PEATE: 1

EXPLANATION OF FIGURE

28

Reconstruction of the vascular channels of the lower cervical ead thoracic

regions in a 19mm. pig embryo (series 168, slides 16-24 inclusive), & 50. Dorsal

view.

Arrows indicate the levels at which the microphotographs were taken.

Cross lines, not labelled, indicate the extent of the divisions, A, B, and C.

A, precardinal division

B, postcardinal division

C, supracardinal division

1, left jugular lymph sac

2, thoracic duct approach

4, lymphatic spaces, incipient thoracic
duct anlagen

6a, precardinal veno-lymphatics

6b, posteardinal veno-lymphatics

6c, supracardinal veno-lymphatics

7, oblique vessel

8, dorsal segmental veins of the pre-,
post-, and supracardinals, respec-
tively

9, internal jugular vein

10, left precardinal vein

11d, 11s, right and left postcardinal
veins

12d, 12s, right and left supracardinal
veins

12ld, right supracardinal line beginning
to fuse with the right posteardinal

466

12ls, left supracardinal line, plexiform

15, aorta

13a, aortie arch

13c, left carotid artery

13ds, dorsal segmental arteries

13s, left subclavian artery and branches

14, left sympathetic nerve trunk

15, vagus

16, recurrent laryngeal nerve, and ac-
companying vein

17, oesophagus

20, mesonephroi

21, left Cuvierian duct

22, left subclavian vein

23, cephalic vein

24, external jugular vein

25, venous plexus between supracar-

dinal lines
26, subclavian

lymph sae
N65, fifth spinal nerve

approach of jugular

PLATE 1

THORACIC DUCT DEVELOPMENT IN THE PIG

OTTO F. KAMPMEIER

wo eo) n o
'é) (©) o) O
=n HX

fy a fe

PLATE 2

EXPLANATION

OF FIGURE

29 Reconstruction of the vascular channels of the lower cervical and thoracic
regions in a 20 mm. pig embryo (series 194, slides 22-31 inclusive), * 50. Sinistro-

dorsal view.

A, precardinal division

B, posteardinal division

C, supracardinal division

1, left jugular lymph sac

2, thoracic duct approach

3, anterior lymphatic plexus or sub-
sequent common trunk of the right
and left thoracic ducts replacing the
precardinal veno-lymphatics

4d, lymphatic spaces in the right thor-
acic duct line replacing the oblique
vessel

4s, lymphatic spaces in the left thoracic
duct line

6a, precardinal veno-lymphatics

7, oblique vessel, degenerating and
breaking up into segments

6b, posteardinal veno-lymphatics

6c, supracardinal veno-lymphatics

8, dorsal segmental veins of the pre-,
post-, and supracardinals

9, internal jugular vein

10, left precardinal vein

12d, 11s, right and left postecardinal
veins

468

12d, 12s, right and left supracardinal
veins

/2ls, left supracardinal line; right line
has completely fused with the right
posteardinal

13, aorta

13b, ductus arteriosus Botalli

13s, left subclavian artery and branches

14, left sympathetic nerve trunk

15, vagus

16, recurrent laryngeal nerve

17, oesophagus

21, left Cuvierian duct

22, left subclavian vein

23, cephalic vein

24, external jugular vein

25, degenerating remnants of the former
venous plexus between the supra-
cardinal lines.

26, subclavian approach of the jugular
lymph sac

N5, fifth spinal nerve

THORACIC DUCT DEVELOPMENT IN THE PIG
OTTO F. KAMPMEIER

23 24

N5 1 9 15 14 16

IG. 3:

SES £__-, ,

Lf

lls

12s 13 6c

469

PLATE 3

EXPLANATION OF FIGURE

30 Reconstruction of the lower cervical and thoracic regions in a 23 mm. pig
embryo (injected series 23a, Johns Hopkins University Embryological Collec-

tion, slides 21-30 inclusive), X 50.

A, precardinal division

B, posteardinal division

C, supracardinal division

1, left jugular lymph sac

3, anterior lymphatic plexus or subse-
quent common trunk of the right
and left thoracic ducts.

5d, 6s, continuous and injected portions
of the right and left thoracic duct
anlagen

X,X, extent of the continuity of 5d and
§s, and the farthest points to which
the injection mass has penetrated

4d, long and short lymphatic spaces in
the axis of the injected anlage and
in the path of the future right thor-
acie duct

4s, lymphatic spaces in the left thoracic
duct line hidden by the veins, but
indicated by the dotted circles

7, oblique vessel, degenerating ante-
riorly

8, dorsal segmental veins of the pre-,
post, and supracardinals

9, internal jugular vein

Dorsal view, slightly from the left.

10, left precardinal vein

iid, 11s, right and left postcardinal
veins

12ls, left supracardinal line, plexiform;
the right line has fused with the right
posteardinal vein

12d, 12s, right and left supracardinal
veins

13, aorta

13s left subclavian artery and branches

14, left sympathetic nerve trunk

15, vagus

16, recurrent laryngeal nerve, and ac-
companying vein

17, oesophagus

20, anterior tip of left mesonophros

21, left Cuvierian duct

22, left subclavian vein

23, cephalic vein

24, external jugular vein

25, degenerating segments of the for-
mer extensive plexus between the
supracardinal lines

26, subclavian approach of the jugular
lymph sac

470

THORACIC DUCT DEVELOPMENT IN THE PIG PLATE 3
OTTO F. KAMPMEIER

1 9 15 14 8 16
23 ; uy
: >i
- at A
FIG.14 26 7 uy
10 7 ny

iD

=
PLES

128 13 12d

PLATE 4

EXPLANATION OF FIGURE

31 Ventral view of the lower half and right side of the reconstruction repre-
sented in figure 30 (23 mm. pig. embryo, series 28a, J.H.E.C., from section 12,

slide 25 to slide 30 inclusive), 50.

5d, extremity of the injected portion of
the right thoracic duct

X, blind end of 5d and the farthest
extent to which the injection mass
has penetrated

4d, discontinuous and uninjected lym-
phatic spaces in the right thoracic
duct line

6b, posteardinal veno-lymphatics

7, posterior portion of the oblique
vessel and its junction with the post-
cardinal

8, dorsal segmental veins

11d, right posteardinal vein

12d, right supracardinal vein

13, aorta

32 Sinistro-ventral view of the lower third of the reconstruction shown in figure
31 (19 mm. pig embryo, series 168, slides 22-24 inclusive), 50.

6c, periaortic supracardinal veno-lym-
phaties
11s, left posteardinal vein

12s, left supracardinal vein
20, right mesonephros

33. Reconstruction of a segment of the anterior supracardinal division or region
of the periaortic veno-lymphaties in a 22 mm. pig embryo (series 105, slides 34-36
inclusive), X 50. Dextro-ventral view.
4d, 4s, lymphatic spaces replacing the degenerating veno-lymphatics (6c). Other

explanations the same as above.

34 Reconstruction of the same region, represented in the preceding figure,
in a 23 mm. pig embryo (series 67, slides 36-37 inclusive), 50. Sinistro-dorsal
view.

5, right and left thoracic duct; veno-lymphatics have been completely replaced.

Other explanations the same as above.

THORACIC DUCT DEVELOPMENT IN THE PIG PLATE
OTTO F. KAMPMEIER

—

473

PLATE 5

EXPLANATION OF FIGURE

35 Reconstruction of the vascular channels of the lower cervical and thoracic
regions in a 21.5 mm. pig embryo (series 192, slides 16-24 inclusive), X 50. Sinistro-

dorsal view.

A, precardinal division

B, posteardinal division

1, left jugular lymph sac

2, thoracic duct approach

3, anterior lymphatic plexus or subse-
quent common trunk of the right
and left thoracic ducts

4d, 4s, discontinuous thoracic ductan-
lagen

5d, 5s. right and left thoracic duct an-
lagen continuous with lymph sac

X, X, extent of continuity in the duct
anlagen connected with the lymph
sac

7, spur or vestige of the former oblique
vessel

8, dorsal segmental veins

9, internal jugular vein

10, left precardinal vein

11d, 11s, right and left posteardinal
veins

12ls, plexiform remnants of the left
supracardinal line

13, aorta

13b, ductus arteriosus Botalli

13d, right subclavian artery

13s, left subclavian artery and branches

14, left sympathetic nerve trunk

15, vagus

16, recurrent laryngeal nerve

17, oesophagus

18, trachea

21, left Cuvierian duct

22, left subclavian vein

23, cephalic vein

24, external jugular vein

26, subclavian approach of the jugular
lymph sae

474

THORACIC DUCT DEVELOPMENT IN THE PIG PLATE 5
OTTO F. KAMPMEIER

10 | Cera

475

.
e
2 ¢ ;
” ae
p45
|
a?
“ats

F : ew ea a tt Pa
* pt Ret fea os ee 3 5c

_ 2 val Pe ema
: Pee ee wee

THE FORM OF THE STOMACH IN HUMAN EMBRYOS
WITH NOTES UPON THE NOMENCLATURE
OF THE STOMACH

FREDERIC T. LEWIS

Harvard Medical School, Boston, Massachusetts

TWELVE FIGURES

X-ray examinations of the stomach, in adults and especially in
children, have led clinicians to inquire whether the stomach has
a characteristic embryonic form which may sometimes persist.
Figures of the typical embryonic stomach have, indeed, been pub-
lished; but it must be remembered that the stomach changes in
shape as the embryo grows older and, as Broman has found, its
individual variations in embryos of the same stage of develop-
ment is very great. Nevertheless certain fundamental sub-
divisions are strikingly distinct. These primary subdivisions, in
which the embryologist is most interested, were keenly discussed
by the early anatomists. In their writings many suggestive
questions are raised, at the same time that the fundamental
features of the organ are successively recognized and defined.
In the following historical notes, taken from such works as are
at hand, provisional definitions are offered for certain terms
adopted at Basle but at present loosely employed, and attention is
called to the features of the adult stomach which will be examined
in the embryos.

The human stomach was first considered to be a simple sac
with an orifice of entrance above and to the left, and an orifice
of exit below and to the right. Vesalius (1543) in his figures
designates the orifices as the ‘superius ventriculi orificium’ and
‘inferius ventriculi orificium,’ respectively. In his text, however,
both are said to be placed superiorly, so that food shall not escape
by its own weight, but when completely changed to chyme, shall

477

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 4

478 FREDERIC T. LEWIS

be propelled by force of the discharging stomach into the intes-
tine. Fabricius ab Aquapendente (1618) likewise states that the
orificium inferius is not inferior at all, and Spigelius (1627) places
it in the highest part of the stomach; so that the term ‘orificilum
dextrum’ was preferred, and finally the less objectionable Greek
name ‘pylorus’ (Latin, janitor), which had been introduced by
Galen, became the accepted designation. Winslow, however, in
1732, insisted that the position of the orifices is such that ‘‘we
ought with the ancient anatomists to call one of them superior, the
.other inferior.”

The significance of ‘cardia’ (Latin, cor), as applied to the
oesophageal orifice, was discussed by Fabricius, who cites Galen
as stating that the upper orifice of the stomach is called the heart
because the symptoms to which it gives rise are similar tothose
which sometimes affect the heart, sometimes even the brain; but
for Fabricius, cardia, as applied to this orifice, merely indicates a
chief part of the body. Spigelius describes the cardia as consist-
ing of circular fleshy fibers, so that the stomach, after having
received food, may be closed perfectly, thus preventing fumes from
rising, with consequent loss of heat. The cardia and pylorus are
intimately associated with their respective sphincter muscles,
but they do not include the adjacent portions of the stomach.

For the stomach as a whole these anatomists use the Latin
‘ventriculus,’ rather than the Greek ‘gaster,’ and the Latin term
has been adopted at Basle. Since however, the adjective gastri-
cus has been chosen instead of ventricularis, it seems desirable
that gaster should be used in. place of ventriculus, especially since
cardia and pylorus are of Greek origin.!

1T am indebted to Prof. Albert A. Howard for the following note regarding these
terms: Gaster is a Greek word meaning belly (the whole abdominal cavity) but
was often used by the Greeks in the more restricted sense of stomach. It is not
found in Latin with this meaning until very late (only after the literary period).
Ventriculus is used quite consistently for stomach by Celsus and at times by Pliny
the Elder. Cicero in one passage speaks of ventriculus cordis, but does not use
ventriculus for stomach. If gaster is adopted I think the genitive gastri is prefer-
able to gasteris, though as a matter of fact the genitive does not happen to occur in
any Latin that is preserved to our time. Petronius has used the ablative plural
gastris which would be the reason for deciding as I have.

THE FORM OF THE HUMAN STOMACH 479

The stomach, as described by Vesalius, is rounder and more
spacious on the left side, and more slender on the right; to which
Fabricius adds that it is not unlike a gourd with larger belly and
narrower neck. On its dorsal side Vesalius found two swellings,
separated by a vertical impression which was fitted against the
trunks of the aorta and vena cava and the projecting bodies of the
vertebrae. When the stomach was inflated, the impression and
swellings were lost in an even rotundity. It was not until Willis
(1674) described the pyloric antrum in the following passage,
that a permanent subdivision of the stomach was established.

The other orifice, commonly called the pylorus, on the right side of
the stomach, having a capacious and long, gradually narrowed antrum,
ends in a small foramen and thence bent back is continued into the duo-
denum. Here the coats are much thicker than in any other part of
the stomach.

Indeed the long and capacious antrum seems to be a sort of recess and
diverticulum in the stomach, into which the more elaborated and per-
fected portion of the chylous mass may withdraw and there remain,
while the other cruder and more recently ingested portion may be further
digested in the fundus of the stomach (ed. of 1680, p. 13-14)

Aecompanying this description Willis published four lateral
views of the stomach, with its coats successively removed. All
of them show the antrum, but in a fifth figure, representing the
everted stomach, its limits are most satisfactorily indicated (fig.
1). In this figure the antrum is shorter and broader than in one
of the others, in which it has been stretched out so as to forma tube.
In all of the figures it is clear that the antrum extends to the pylo-
rus, which is referred to as its orifice.

Bidloo (1685) published a more accurate figure of the stomach,
here reproduced as figure 2, but he failed to describe it adequately.
He states that the base is provided with two swellings, C and D.
In another figure, showing the same stomach partly laid open, the
portion of the duodenum near the stomach (A) is labelled pylo-
rus, but Bidloo does not refer in any way to the subdivision which
in figure 2 has been labelled B. Cowper (1698), who republished

2 For verifying and revising the Latin translations, the author is under obliga-
tion to Mr. 8. R. Meaker.

480 FREDERIC T. LEWIS

Bidloo’s plates, states that A is the part of the duodenum arising
from the pylorus and adds that B is the antrum pylori.

In 1732 Winslow described the large arch running along the
greatest convexity of the stomach, and the small one directly
opposite, and named them the great andsmallcurvatures. Bichat
(1802) states that ‘‘the great curvature ends simply at the pyloric
orifice, without presenting anything of note unless it be the elbow
(le coude) formed by this pyloric orifice, and named the small
cul-de-sac; but there is no particular swelling at this place and the
bend is precisely in the direction of the pylorus.’’ Cloquet
(1831) repeated this description and Cruveilhier (1834) made it
more explicit. He states that at about 2 or 3 em. from the pylorus
“the stomach, bending sharply upon itself, forms a very pro-
nounced elbow (coude de l’estomac) on the side of the greater
curvature, and presents an ampulla, corresponding to an interior
excavation, named by Willis the pyloric antrum, by others the
small cul-de-sac.”

As pointed out by Miiller (1897), Cruveilhier was unjustified
in identifying a pouch about an inch from the pylorus with the
pyloric antrum of Willis; but he was correct in stating that “‘it
is not rare to see a second ampulla beside the first, and a third
but smaller one, on the side of the lesser curvature’? (compare
with figs. 2 and 3). These had not been recognized by Willis,
but Cowper, in describing Bidloo’s plate, was confronted with
the question whether one or more of these parts was to be regarded
as the antrum. In applying the term to the part adjacent to the

Fig. 1 Willis’s figure of the inverted stomach re-drawn and reduced one-half.
‘‘A, Orificium sinistrum, sive os ventriculi. B, Pylori Antrum, in quo, Tunicae
crassiores existunt. C, Orificium ejus, cuo Duodenum annectitur.”’

Fig. 2 Bidloo’s figure of the unopened stomach, re-drawn and reduced two-
thirds, with lettering added from Bidloo’s drawing of the same stomach opened,
and from Cowper’s edition of Bidloo’s plates. A, pylorus (Bidloo) ; portion of the
intestinum duodenum (Cowper). 8, antrum pylori (Cowper). C, D, two bunch-
ings out in the lower part or fundus of the stomach (Cowper) ; in fundo Gibbis orna-
tur duobus (Bidloo).

Fig. 3 Home’s figure of ‘‘the human stomach inverted, to show the contraction
which divides the cavity into two portions.’’ Re-drawn and reduced two-thirds.
aa, the cardiac portion. 6, the contraction dividing the cardiac from the pyloric
portion. c, the pyloric portion. d, the pylorus.

481

THE FORM OF THE HUMAN STOMACH

NG

MY

My
i
Yih
i

|

IM

Wy

Ty
j
Hy

j

\
‘wie IY

482 FREDERIC T. LEWIS

pylorus but not extending to the bend of the stomach, Cowper was
justified by Willis’s figure here reproduced as figure 1. According
to Cunningham (1906) ‘‘no part of the organ is more definite and
distinct’? than the region which Cowper designated ‘antrum
pylori’ and which, rediscovered by Jonnesco (1895), was named the
pyloric canal. It may be defined as the part of the stomach adja-
cent to the pylorus, averaging 3 em. in length, cylindrical when
empty, bulbous when distended, separated from the remainder of
the stomach by a groove on the greater curvature—the ‘sulcus
intermedius’ of His (1903). For the small cul-de-sac of Cruveil-
hier the term ‘pyloric vestibule’ (Jonnesco, 1895) may be adopted.

Unfortunately Cowper’s use of pyloric antrum has been over-
looked by later anatomists, and the term has been so variously
employed, as tabulated by Miiller, that Miller, His and Cunning-
ham have proposed to abandon it altogether. His has suggested
an entirely new nomenclature for the pyloric region, as follows:
for pyloric vestibule, camera princeps; for the swelling on the
lesser curvature opposite the sulcus intermedius (fig. 3), camera
minor; and for pyloric antrum, camera tertia. But these terms,
as stated by Cunningham, are not in every respect satisfactory,
and it may be well to retain the appropriate name ‘pyloric antrum’
in the sense of Cowper, following Meyer (1861) and Hasse and
Strecker (1905).

The normal division of the entire stomach into two parts, car-
diac and pyloric (of which the latter presents the subdivisions Just
described), was first recognized by Home (1814). He wrote as
follows:

I found also, in the necessary examinations, that the dog’s stomach,
while digestion is going on, is divided by a muscular contraction into two
portions; that next the cardia is the largest, and usually containing a
quantity of liquid, in which there was some solid food; but the other,
which extended to the pylorus, being filled entirely with half-digested
food of an ordinary consistence. I shall, therefore, in my future descrip-
tion call that part which constitutes the first cavity the cardiac portion,
and that which constitutes the second the pyloric portion (p. 140).
The cardiac portion is in length two-thirds of the whole, but in capacity
much greater (p. 139).

Home distinguished these two portions not only in the dog
but, with varying distinctness and permanence, in many animals,

THE FORM OF THE HUMAN STOMACH 483

including the horse, pig, rat, rabbit and man. He did not
describe or label the subdivisions of the pars pylorica, which
however are clearly shown in his figure of the everted human
stomach (fig. 3).

In connection with Home’s work, the following more recent
physiological observations are of interest. Schiitz (1885) found
that in the dog’s stomach, contraction waves travel from the car-
diac end to a place about 1 cm. from the beginning of the pyloric
antrum (pars pylorica?), which in the isolated resting stomach
may be recognized by a shallow annular constriction about 2 em.
behind the pylorus, and there end with a deep ‘prae-antral con-
striction.’ The second phase, which follows the first immediately,
concerns the antrum only. The motions of the two parts may
take place independently of one another. Moritz (1895) deter-
mined the pressure within the two parts of the stomach, and stated
that the difference between them was greater than would be
inferred from merely observing their motions. Cannon (1898)
found that the stomach of the cat, as shown by X-ray examina-
tions, 1s composed of two physiologically distinct parts—a
‘busy antrum’ and a cardiac reservoir. In 1911 he states that
during normal digestion ‘‘slight constrictions appear near the
middle of the body of the stomach, and pressing deeper into the
greater curvature, course towards the pyloric end.. When a wave
sweeps round the bend into the vestibule, the indentation made by
it increases.”’ He adds that when vomiting occurs, a strong con-
traction at the angular incisure completely divides the gastric
cavity into two parts. Thus the observations of Home have been
amply confirmed and extended. Other X-ray observers, however,
have considered that the antrum, or pars pylorica, of anatomists
is merely the part of the stomach marked off by a passing peri-
staltic wave (Hertz; Kaestle, Rieder and Rosenthal; Barclay). In
this they follow Sappey (1874), who was of the opinion that
Home’s subdivision of the stomach was based on fortuitous mus-
cular contractions. This will be disproved by showing that the
two divisions of the stomach are well marked in embryos in
which the muscle-layers are still scarcely differentiated.

484 FREDERIC T. LEWIS

When a peristaltic wave remains fixed after death, the stomach
may appear as ‘‘two joined together’ (Riolan 1618), in which
case the subdivisions need not correspond with the anatomical
parts already described. Usually the constriction is near the mid-
dle of the stomach, and falls within the cardiac portion. Mor-
gagni (1761) observed five cases, allin women. One of the stom-
achs was from a patient who had been troubled with excessive
vomiting since birth, but none of the stomachs showed any sign
of disease. Since several cases had been reported in men, Mor-
gagni concluded that the double stomach was not a deformity
due to stays, but had existed from the first formation of the organ.
Sandifort (1777-1781), as quoted by Bettman, described a typical
case in.a fetus, the age of which is not stated in the citation. Del-
amare and Dieulafé (1906) reported a case in a new-born syphilitic
infant, in which they describe an hypertrophy of the circular
muscle at the place of constriction. The thickened muscle-layer
may, however, be due to contraction, as indicated by the folded
and thickened overlying layers shown in their figures. Cunning-
ham (1906) holds that there is not an atom of evidence that the
hour-glass stomach ever arises as a congenital deformity, but he
is not prepared to state that the strictures which separate the two
sacs of the bilocular stomach are always temporary and fleeting.

The change in the direction of the lesser curvature is more
dependable as a boundary between the pars pylorica and pars
cardiaca, than the constriction which is present in certain cases
but ‘‘not as a rule’ (Huschke). The lesser curvature, which is
coneave along the cardiac portion, becomes convex along the
pars pylorica (Meckel 1820; Huschke 1844). Retzius (1857)
figured a deep stricture in the lesser curvature at the beginning of
the bulbous pars pylorica, where Luschka (1869) frequently
found an acute angle directed toward the gastric cavity. This
notch has been named by His (1903) the ‘incisura angularis,’
and it occurs between the two parts of the stomach. Along the
greater curvature the boundary is less clearly marked. It is
indicated by the change in direction already described as the elbow
of the stomach, and referred to by Home as ‘‘an angle formed at
the part where the temporary contraction takes place.”” Some-

THE FORM OF THE HUMAN STOMACH 485

times the constriction is slightly to the cardiac side of the elbow,
as shown in figure 3.

The angle which separates the two parts of the stomach is
obscure in the older drawings in which the organ is almost hori-
zontally placed. According to Bichat (1802) ‘‘when the stomach
is filled its obliquity increases considerably ; often it appears almost
perpendicular, so that the right extremity . . . . isstrongly
recurved upward, and forms a very acute angle with the body of
the organ.” Luschka (1869) similarly found that the greater
part of the stomach, as a rule, has a precisely vertical position,
but that the pars pylorica is directed almost transversely. Both
of these forms, with vertical body and transverse or ascending
pars pylorica, will be seen in the embryos to be examined.

The pars pylorica and its subdivisions having been described,
the pars cardiaca may next be examined.’ It is divided into the
‘saeccus caecus,’ now called the ‘fundus; the ‘corpus’ or body;
and the gastric canal. The term fundus was appropriately applied
by Vesalius to the lower part of the stomach, which in the trans-
verse position of the organ, extends well toward the pyloric region.
It was so used by Willis (1674); and by Cowper (1737), as seen in
figure 2. Caldani (1804) makes fundus synonymous with greater
curvature. The bulging left or upper extremity of the stomach
received the special name ‘saccus caecus’ (Haller, 1764; Caldani,
1804). But Meckel (1820) considered fundus and saccus caecus
as synonyms, and preferred fundus; Huschke (1844) likewise
“made them synonymous, but adopted saccus caecus, which Henle
used in 1866. Nevertheless, fundus has become adopted for the
highest part of the stomach and saccus caecus has been rejected.
The fundus lies at the left of the cardia, being separated from the
oesophagus by a notch, the ‘incisura cardiaca’ of His (1903).
Below, as described by Cloquet, the fundus terminates almost
imperceptibly in. the greater curvature. It is therefore bounded
arbitrarily by a horizontal plane at the level of the inferior border
of the cardia (Jonnesco), or by a line prolonging the axis of the
abdominal part of the oesophagus (Keith and Jones, 1902).

According to Keith and Jones the fundus arises in human em-
bryos as a localized outgrowth or diverticulum of the stomach,

486 FREDERIC T. LEWIS

and in its manner of origin has much in common with the caecum
and vermiform process. From numerous observations they con-
clude that ‘‘it is not uncommon to find in the stomach of the
anthropoids, and to a lesser degree in that of the apes (especially
in Mycetes) clear indications of three chambers, namely, a fun-
dus, a body, and a pyloric part; and that therefore the stomach of
the Primates (excluding the Lemuroidea) is probably tripartite in
nature.” It should be noted that the fundus as defined by Keith
and Jones is a larger part of the stomach than that set off by Jon-
nesco, and that their boundary is justified by comparison with the
stomach of Semnopithecus which they have figured. If the fun-
dus corresponds in any way to a first stomach or rumen, it may be
regarded as the globular upper end of the organ which is often
marked off by the contraction of the corpus.

The body of the stomach (corpus gastri), as defined by Riidin-
ger (1873), is its middle subdivision, situated between the fundus
and the pars pylorica. Froriep (1907) proposed to rename it
the pats intermedia; but since it is a portion of the pars cardiaca,
and is not intermediate between the pars cardiaca and pars pylor-
ica, the proposed term would lead to confusion. Jonnesco (1895)
defined the body as including the pyloric vestibule, but in the
same paragraph he described the boundary between the vestibule
and ‘‘le corps proprement dit.’ Miiller (1897) included the fun-
dus with the body, making corpus and pars cardiaca synonymous.
It is only by accepting Riidinger’s earlier definition that corpus
becomes a useful term. The corpus may be contracted at any
point, as in the hour-glass stomach, in which case part of it appears
to belong with the fundus and the remainder with the pars
pylorica. Sometimes it is contracted as a whole, but more often
it is relaxed, and its boundaries are then ill-defined.

The gastric canal is a channel which follows the lesser curva-
ture, appearing as a groove when seen from the inside of the stom-
ach. It suggests a continuation of the oesophagus, split open
toward the gastric cavity, and has been named the sulcus oesopha-
geus, sulcus gastricus, sulcus salivalis and canalis salivalis. It is
confusing, however, to refer to this channel as a sulcus, since the

THE FORM OF THE HUMAN STOMACH 487

external grooves of the stomach are so designated (sulcus inter-
medius, suleus pyloricus), and it is undesirable to name a part of
the stomach oesophageal or salival. Therefore the term gastric
canal, ‘canalis gastricus,’ is here proposed, and canalis is used as
in Latin for an open canal, which in this case may become a tube
during its physiological activity, by the approximation of its
lips.

The gastric canal has long been known in ruminants, but in
its less highly developed condition in the human stomach, it
has attracted little attention. In man it is generally supposed to
be due to the arrangement of the oblique muscle fibers, which were
first described by Willis (1674), in connection with a figure of the
stomach in the position shown in figure 1. The ‘top’ of the
stomach is accordingly toward the lesser curvature, and the ‘fun-
dus’ is toward the greater curvature. Willis wrote as follows:

These muscle fibers, which are seen to arise behind the cardia and to
pass around its left margin, are carried forward to the right portion of
the stomach. A notable bundle of them, proceeding in straight lines
along the top of the stomach on either side, encounters the antrum, and
spreading over the length of its cavity in a scattered manner, terminates
in the pylorus. Moreover the remaining fibers of this layer extend
obliquely over the walls of the stomach on both sides, and then directly
toward the fundus where they come together. The function of the for-
mer (the straight bundles) seems to be to bring one orifice toward the
other in emptying, by making them lower and higher respectively (ed.
of 1680, pp. 11-12).

Retzius called attention to this description by Willis and, as
reported by Gyllenskoeld (1862), he supplemented it as follows:

The upper portion of the oblique fibers of the human stomach serves
to form a sort of trough along the lesser curvature which, under the con-
trol of the motor nerves, becomes more or less closed; along this path
possibly fluids and soft things, saliva, ete., may proceed directly from the
oesophagus to the pars pylorica, passing by the cardiac portion, which
corresponds to the first two stomachs of ruminants and the non-glan-
dular part of the stomach in rats.

The correctness of this conjecture concerning the passage of
fluids was established by Cohnheim (1908), who was surprised

488 FREDERIC T. LEWIS

to find that water or salt-solution passed rapidly through the
full stomach of a dog, without mixing with the gastric contents.

Gyllenskoeld (1862) states that the oblique fibers extend only
to the pars pylorica, and not to the pylorus as described by Willis.
This has been confirmed by Kaufmann (1907). He found that
there is no sphincter of circular fibers separating the pars cardiaca
from the pars pylorica, but that the furrow between them has a
special structure, since it is the place where the oblique fibers
terminate and interlock with the circular fibers.

Hasse and Strecker (1905) have named the folds which bound
the gastric canal the ‘plica hepatica’ and ‘plica aortica’ respec-
tively, and state that they are connected with one another by the
‘plica cardiaca’ which passes around the cardia, projecting into
the stomach beneath the incisura cardiaca. According to Hasse
and Strecker the plica cardiaca does not form a valve for the
cardia, as Braune (1875) thought possible from the result of experi-
ments on a cadaver. The hepatic, cardiac and aortic plicae
together form a U-shaped structure, across the open end of which
is the ‘plica angularis.’ This is beneath the incisura angularis,
at the beginning of the pars pylorica.

Waldeyer, who describes the channel from cardia to pars
pylorica as the ‘Magenstrasse’ (1908), considers that its formation
depends upon the oblique muscles, rather than upon folds which
arise in relation with adjacent organs. In the following pages
evidence will be offered to show that the gastric canal is a distinct
epithelial structure, arising independently both of the muscle and
the surrounding organs.

There remain to be considered two structures which are bevond
the limits of the stomach—the ‘antrum duodenale’ and the ‘antrum
cardiacum.’

Retzius (1857) states that the beginning of the duodenum is
often specially rounded, not only in man, but in a large propor-
tion of mammals; in dolphins it has been considered a part of the
stomach. Owen (1868) remarks that in all Artiodactyles the
duodenum is dilated at its commencement; it there forms a dis-
tinct pouch in the camel. For this pouch Retzius proposed the

THE FORM OF THE HUMAN STOMACH 489

names ‘‘antrum or atrium duodeni”’ but used the former in his
figures. Luschka (1863) refers to a flask-shaped expansion at
the beginning of the duodenum, which in his figure is called the
‘antrum duodenale.’ This structure will be seen to be far more
distinct in human embryos than it appears to be in adults.

The cardiac antrum was first described by Luschka (1863) as
follows:

At the junction of fundus and lesser curvature the oesophagus enters
the stomach, forming a funnel-shaped expansion—the cardia. Although
ordinarily the cardia is continued into the rest of the stomach without
definite boundary, in rare cases the funnel-like expansion is sharply
marked off by an external depression and corresponding internal eleva-
tion, thus forming a sort of cardiac antrum (p. 179).

In 1869 Luschka adds that if this funnel is to be regarded as
part of the stomach, the beginning of which is not rather to be
considered at the base of the funnel where the stratified epithe-
lium ends in a zig-zag line (fig. 3), ‘‘then the funnel-shaped expan-
sion must be specially designated as the pars cardiaca.”’

Thus Luschka proposed two names for a single structure; first,
cardiac antrum; and-later, in case the antrum is to be regarded
as part of the stomach, pars cardiaca. The latter may be rejected, —
since it is generally agreed that the cardia is at the base of the
cone, and that therefore ‘cardiac antrum’ is ‘‘merely another
name for the intra-abdominal part of the oesophagus” (Cunning-
ham). Moreover the earlier use of pars cardiaca, or cardiac por-
tion, for the fundus and corpus taken together, was overlooked by
Luschka, and by certain later anatomists who have proposed -to
substitute Hauptmagen (His), saccus ventriculi (Hasse and
Strecker) and pars digestoria (Froriep).

The fundamental subdivisions of the stomach and adjacent
parts of the digestive tube, as they have been defined in the pre-
ceding pages, are presented in figure 4 and in the following table,
with authority for certain of the definitions adopted:

490 FREDERIC T. LEWIS

Antrum eardiacum (Luschka, 1863)
Gaster
Cardia
Pars cardiaca gastri (Home 1814)
Fundus (Meckel. 1820)
Corpus (Riidinger 1873)
Canalis gastricus
Pars pylorica gast1i (Home 1814)
Vestibulum pyloricum (Jonnesco 1895)
Antrum pyloricum (Willis 1674 (?); Cowper 1698)
Pylorus
Antrum duodenale (Retzius, 1857)

As boundaries between these parts, the following may be
recognized: Between the cardiac antrum and fundus, the ‘inci-
sura cardiaca;’ between cardiac and pyloric parts, the ‘incisura
angularis;’ between pyloric antrum and pyloric vestibule, the
‘sulcus intermedius’ (all of His 1903); at the pylorus, the ‘sulcus
pyloricus’ (Luschka 1863).

THE STOMACH IN HUMAN EMBRYOS

The embryonic stomachs to be examined are five in number,
from embryos between 10 mm. and 45 mm. in length. Thus
they are all smaller than the specimens studied by Miller, but
similar stages have been described by Broman in his extensive
work on the omental bursa. Broman modelled not only the
gastric epithelium, but also entire stomachs, including the meso-
dermal portion. In the models to be described, only the epithe-
lium has been included, since this is the portion having character-
istic shape, to which the other layers subsequently conform.

In the youngest embryo (10 mm., fig. 5) the stomach is no
longer a simple sac with superior and inferior orifices, but is
already divided into an expanded pars cardiaca and a tubular
pars pylorica. Between the two, and almost exactly in the middle
of the stomach, is the incisura angularis. Since the incisure in the
adult is perhaps twice as far from the cardia as from the pylorus,
it is evident that the pars pylorica is relatively long in early stages.
This is strikingly shown in other models of the series (figs. 6-9).

THE FORM OF THE HUMAN STOMACH 49]

Incisura cardiaca ‘

Antrum cardiacum

Fundus

Cardia

Canalis gastricus
Incisura angularis

Sulcus pyloricus
(Pyiorus)

Antrum
pyloricum

Antrum
duodenale

Vestibulum
pyloricum

Sulcus intermedius

Fig. 4 Diagram showing the subdivisions of the human stomach

It is true also in the cat, if one may judge by comparing Thyng’s
model of the stomach of a 10.7-mm. embryo (this Journal,
vol. 7, p. 496) with Cannon’s tracings from the adult. In rumi-
nants a constriction early separates the rumen and reticulum from
the psalterium and abomasum; according to Ellenberger and
Baum the abomasum is larger than the rumen in embryos and
very young animals, but later this relation is reversed. The
relatively large size of the pars pylorica in early stages is therefore
not limited to human embryos.

The cardia cannot be definitely located in the 10-mm. embryo
(fig. 5) since the oesophagus, in joining the stomach, expands into
a flattened cone, one margin of which extends to the angular
incisure. A similar extension of the oesophageal cone to the
incisure is clearly seen in Broman’s model of the stomach of the
seventh embryo in his series (11.7 mm.). At 16 mm. (fig. 6)
the body of the stomach may be recognized along the lesser cur-
vature, separating the oesophageal cone from the angular incisure ;
but a canal, distinctly marked out above and indicated below,

492 FREDERIC T. LEWIS

passes along this curvature from the oesophagus to the pars
pylorica. A more distinct canal in this position is seen in two of
Broman’s models, from embryos of 10 mm. and 16.2 mm. respec-
gively. Apparently this canal has not been previously described
in embryos, although Toldt (1879), referring to the general
direction of the oesophagus in a 23-mm. specimen, states that it
descends into the stomach “‘in such a way that the lesser curva-
ture forms, as it were, a continuation of the ventral border of the
oesophagus.”’

In the embryos of 19.3 mm. and 19.0 mm. shown in figures
7 and 8 respectively, the canal is not seen. The first of these
stomachs is abnormal, but the second specimen is unobjectionable.*
Moreover in Broman’s figure of the stomach from an embryo of
21 mm., there is no trace of the channel. Its obliteration, if
normal, appears to be temporary however, for in the 44.3-mm.
specimen shown in figure 9, it is more distinct than in preceding
stages. It passes from the conical cardiac antrum to the angular
incisure. This embryo, owing to its large size, was not perfectly
preserved, and the epithelium has separated from the mesen-
chyma; but whether one, or the other, or both of these tissues has
shrunken is uncertain. The model may, however, be accepted
as giving an essentially correct idea of the shape of the stomach,
since the separated mesenchyma presents corresponding ridges
and furrows. The distinctness of the gastric canal is strikingly
shown when the model is viewed from the inside (fig. 10). It takes
a slightly S-shaped course from the stellate cardia to the orifice
of the pars pylorica, and is bounded by a rounded plica aortica,
and a more prominent and angular plica hepatica. These folds
are not formed, as described in the adult by Hasse and Strecker,
through compression of the border of the stomach between the
aorta behind and the caudate lobe of the liver in front; for the
outer layers of the stomach are not indented. Moreover at this
stage there are no bands of oblique fibers to account for the canal.
If the channel proves to be a constant feature of embryos of this
stage, and it is present in an embryo of 37 mm. which was not
modelled, it may be that the arrangement of the oblique fibers is a
consequence rather than the cause of the gastric canal.

THE FORM OF THE HUMAN STOMACH 493

In the same way that the gastric canal accords with the ‘oesoph-
ageal suleus’ of ruminants, which is described by comparative
anatomists as a continuation of the oesophagus open on one side,
the cardiac antrum may correspond to the ‘atrium ventriculi.’
This, according to Ellenberger and Baum is ‘‘a dome-shaped
swelling on the dorsal side of the reticulum and thoracic end of the
rumen, which is only indistinctly marked off from them by a
shallow groove; ventrally its cavity passes directly into that of
the reticulum, and caudo-ventrally into the vestibule of the rumen;
toward the thorax it rests against the diaphragm near the hiatus
oesophageus.”’ From the general ‘atrium’ seen in figure 5, the
lower part is set off as the gastric canal, and the upper part
remains as the cardiac antrum (figs. 8 and 9). From studies of
the adult stomach it may be assumed that the cardia is at the
base of this antrum, which therefore belongs with the oesophagus.

The development of the fundus of the stomach has been
described by Broman (1911) as follows:

By the beginning of the second month the cranial part of the greater
curvature begins to bulge out. But not until the third month, or later,
is this outpocketing generally directed so strongly craniad that its blind
end comes to lie above the orifice of the oesophagus. Only from this
time, therefore, can we speak of a distinct gastric fundus (pp. 3826-828).

Similarly Keith and Jones state that the outgrowth is best
marked in embryos of the third and fourth month. But as shown
in figures 8 and 9, and by the fact that Toldt, in an embryo of 48
mm., found a well marked fundus projecting toward the concavity
of the diaphragm, it is clear that the fundus may be well devel-
oped in the second month. In the model shown in figure 9, the
fundus when seen from above, presents a curious appearance,
since seven prominent ridges converge toward its apex. Two of
them come from the cardiac antrum, sweeping in a semicircular
curve beneath the cardiac incisure, thus resembling the ridges seen
in figure 8. There is normally no boundary between the fundus
and corpus, but in an abnormal embryo of 18.5 mm., described
by Broman, the fundus is cut off by a rather deep constriction.
Broman states that this specimen suggests an hour-glass stomach,
from which, however, it is essentially different, since the oesopha-

THE AMERICAN JOURNAL OF ANATOMY, VOL. 13, NO. 4

494 FREDERIC T. LEWIS

Figs. 5 to 9 Models of the gastric epithelium in human embryos, as follows:
figure 5, 10 mm., Harvard Embryological Collection, Series 1000, 50 diam.;
figure 6, 16.0 mm., H. E. C. 1322, x 35 diam; figure 7, 19.3 mm., H. E. C. 1597,
xX 30 diam.; figure 8, 19.0 mm., H. E. C. 819, X 26 diam.; figure 9, 44.3 mm., H.

THE FORM OF THE HUMAN STOMACH 495

I. ang.

E. C. 1611, X 18 diam. A.du., antrum duodenale. C., corpus gastri. D.ch.,
ductus choledochus. D.p.d., ductus pancreatis dorsalis. F., fundus gastri.
I.ang., incisura angularis. Oe., oesophagus. P.py., pars pylorica gastri.

496 FREDERIC T. LEWIS

gus enters the part toward the pylorus. The fundus is best
marked when the pars cardiaca is in an approximately vertical
position, and this is the case in figures 6 to 9. Broman, however,
has found a greater variety of positions. In an embryo of 21 mm.
he figures the stomach as horizontal, so that both orifices are
superior, as described in the adult by Vesalius; but this position
must be regarded as exceptional.

The body of the stomach requires no comment other than that
its ridges appear to be rather definitely placed. The shelf-like

Car.

P.ao

O. p. py-

Fig. 10 Model of the interior of the stomach, from an embryo of 44.3 mm.,
H. E. C. 1611, X 25 diam. Car., cardia. C.g., canalis gastricus. O.p.py., ori-
ficium partis pyloricae. P.ao., plica aortica. P.hep., plica hepatica.

prominence at the base of the oesophageal cone in figure 5, is
evidently represented by the chief fold which extends horizon-
tally across the base of the fundus, and bends down parallel
with the lesser curvature in figure 6. Such an angular fold (with
a subdividing furrow) is seen in figure 9, and it is clearly shown in
embryos of 10 and 16.2 mm. figured by Broman. Why the ridges
are absent from other specimens, as in figures 7 and 8, and in
several of Broman’s embryos, is not apparent.

THE FORM OF THE HUMAN STOMACH 497

The position of the pylorus could not be determined with cer-
tainty in the 10-mm. embryo (fig. 5); and Tandler (1900) states
that in an embryo of 11 mm. the pylorus is not marked. At
14.5 mm., ‘‘where the stomach passes into the duodenum, there-
fore at the place of the future pylorus” he saw ‘‘a considerable
thickening of the epithelium.” The epithelial proliferation,
which Tandler describes, is seen throughout the upper part.of
the duodenum. It is not evident that he recognized the local
swelling, chiefly on the upper side of the digestive tube, which is

A. du.

Fig. 11. Frontal section through the pylorus of a 19-mm. embryo, H. E. C.
828, section 330, X 40 diam. A.du., antrum duodenale. B.om., bursa omentalis.
P.py., pars pylorica. T.musc., tunica muscularis.

shown in figures 6 to 8. This swelling, which distinetly marks the
position of the pylorus when the muscle-layers are still undiffer-
entiated, and scarcely to be recognized, is apparently the duo-
denal antrum of Retzius. In a frontal section through the pars
pylorica of a 19-mm. embryo, it appears as shown in figure 11.
At this stage the musculature of the pars pylorica is considerably
thicker than that of the duodenum, but in this it conforms to the
shape of the epithelial tube. In figure 9 the duodenal antrum is
seen to be smoother than the more distal part of the duodenum,
recalling the statement of Retzius that here, in the adult, the
valvulae are absent and the villi are short. In this embryo the

498 FREDERIC T. LEWIS

gastric epithelium is seen to be slightly invaginated into the duo-
denal tube, as observed by Cunningham at birth. Toldt found
that the sulcus pyloricus could be seen externally in an embryo of
48 mm., and presumably it could have been found in this specimen
by dissection.

The pars pylorica, even in the 44.3-mm. embryo, fails to show
distinet subdivision into antrum and vestibule. Miller, who
studied dissections of the embryonic stomach, states that in the
‘first fetal period’ the pyloric antrum is a direct continuation of
the pyloric vestibule, but that later, when the pars pylorica is
bent convexly upward, the general direction of the vestibule is
upward, and of the antrum, downward. In the earlier period the
antrum is characterized by “‘its cylindrical form and the great
development of its muscle-layer.”’ In the still earlier stages under
discussion, neither distinction is applicable, for the entire pars
pylorica is cylindrical, and as shown in figure 11, its musculature
is thick. It is possible that the short and relatively smooth
terminal portion of the pars pylorica, which in figure 9 is seen to
be directed upward, represents the antrum; but this cannot be
affirmed without further investigation.

In conclusion, the abnormal stomach shown in figure 7 may be
considered. It is of special interest since Gardiner (1907) has
described the stomach of a child of three months, which presents
a very similar condition. In the embryo there is a round nodule
of epithelial cells near the angular incisure. In sections (figure
12) this nodule appears as a compact ring of radiating cells
arranged about a lumen. ‘Toward the gastric epithelium there
is one section in which this structure fails to appear, so that it is
apparently detached, but a short stem projects towards it from
the adjacent epithelium. Both the nodule and its stalk are inside
of the muscular coat. A comparable but larger structure was found
by Lewis and Thyng in the duodenal region of a 20-mm. pig
(figured in this Journal, vol. 7, p. 509). In that case, however,
the detached portion, which had become cystic, lay outside of the
tunica muscularis. That the nodule in the human embryo is an
accessory pancreas, is made certain by Gardiner’s specimen, in
which a well developed gland with typical islands occurs in a

THE FORM OF THE HUMAN STOMACH 499

corresponding position. Similar epithelial nodules were fre-
quently found by Lewis and Thyng in young pig embryos, but
they hesitated to interpret them as pancreases because of their
abundance, and because they were never seen to branch like true
pancreases. They may, however, as Elze has shown, be dis-
tinguished from the epithelial pockets of the gall-bladdet and

Du.

P. py.

Fig. 12. Section through an abnormal stomach of an embryo of 19.3 mm., H.
E. C. 1597, section 730, X 35 diam. A.m.s., arteria mesenterica superior. B.om.,
bursa omentalis. C.,corpusgastri. D.ch.,ductuscholedochus. Du., duodenum.
L., lien. P.ac., pancreas accessorium. P.py., pars pylorica gastri.

small intestine which these authors described, and which seem
to be transient irregularities of the expanding tubes. Accepting
the small, round, compact nodules as accessory pancreases, we
may conclude that they arise at about the time when the normal
pancreases become established, and usually at no great distance
from them, either up or down the intestine. Subsequent elonga-
tion of the tube may carry them farther away. They may be

500 FREDERIC T. LEWIS

assumed to develop slowly, since in the early stages they fail to
produce branches like the adjacent normal pancreases; and as
they are frequently seen to be detached, probably many of them
degenerate without becoming functional glands.

Taken as a whole the stomach which Gardiner described is
shaped like a retort. It has a globular cardiac end, 7 to 8 em. in
diameter; ‘a constriction about its middle; and a tubular pyloric
portion, 3 to 4em.in diameter. If the cardiac half of the stomach
shown in figure 7 should be pressed down, so that the lesser curva-
ture became horizontal and the pars pylorica seemed to leave the
upper portion of the corpus, then the form shown in Gardiner’s
case would be duplicated. Although Gardiner describes his case
as an hour-glass stomach, it should not be classed with those
which are due to muscular contraction. It is an arrest of develop-
ment, in which the pars pylorica remains clearly set off from the
pars cardiaca, and as in the 19.3-mm. embryo, the line of separa-
tion is in the middle of the stomach.

CONCLUSIONS

‘In addition to suggestions in regard to the nomenclature of
the stomach, presented in tabular form on p. 490, the following
conclusions may be drawn.

In the stomachs of embryos from 10 to 45 mm. in length, the
division into pars cardiaca and pars pylorica is well marked; the
latter is relatively long, constituting one-half the length of the
stomach.

The oesophagus in joining the stomach in 10-mm. embryos
forms a cone extending to the angular incisure. Later this cone
gives rise to the cardiac antrum above, and to a downward pro-
longation of the antrum below. This prolongation, which extends
along the lesser curvature, constitutes the gastric canal (canalis
gastricus). It was found to be well developed in an embryo of
44.3 mm.

The fundus develops during the second month as a conical
pouch; its boundary toward the corpus is arbitrary.

THE FORM OF THE HUMAN STOMACH 501

The position of the pylorus is first indicated by the antrum
duodenale. The pylorus, like the gastric canal, is primarily an
epithelial differentiation, to which the musculature conforms.

The occurrence of an accessory pancreas near the angular
incisure is shown in an embryo of 19.3 mm., in connection with
a stomach which would probably have presented a permanent
stricture between the pars cardiaca and the pars pylorica, thus
giving rise to one form of the so-called hour-glass stomach.

BIBLIOGRAPHY

(When translations or two editions are included in the following list, the earlier
edition was not consulted).

Barcuay, A. E. 1910 The normal and pathological stomach as seen by the X-
rays. Brit. Med. Journ., vol. 11, p. 537-541.

BetrrmMann, H. W. 1899 The shape and position of the stomach. Philadelphia
Monthly Med. Journ., vol. 1, p. 121-153.

Bicuat, X. 1802 Traité d’anatomie descriptive. Tome 3. Paris.
Brptoo, G. 1685 Anatomia humani corporis. Amstelodami.
Braung, W. 1875 Topographisch-anatomischer Atlas. Leipzig.

Broman, I. 1904 Die Entwickelungsgeschichte der Bursa omentalis. Wies-
baden.

1911 Normale und abnorme Entwicklung des Menschen. Wiesbaden.
Caupant, L. M. A. anp CaupAnt, F. 1804 Icones anatomicae, vol. 2, Venetiis.

Cannon, W. B. 1898 The movements of the stomach studied by means of the
Réntgen rays. Amer. Journ. Physiol., vol. 1, p. 359-382.

1911 The mechanical factors of digestion. London.
Cioquet, J. 1831 Anatomie de homme. Tome 5. Paris.

CouHNHEIM, O. 1907 Beobachtungen tiber Magenverdauung. Miinchen. med.
Wochenschr., Jahrg. 54, pp. 2581-2583.

Cowper, W. 1698 The anatomy of human bodies. Oxford. (2d ed., Leyden,
lfeyie)

CRUVEILHIER, J. 1834 Anatomie descriptive. Tome 2. Paris.

CunnineuaM, D. J. 1906 The varying form of the stomach in man and the
anthropoid ape. Trans. Roy. Soc. Edinburgh, vol. 45, Part I, pp. 1-47.

DELAMARE, G. AND DrnuLars&, 1906 Estomac de nouveau-né a tendance bilo-
culaire. Journ. del’Anat. et de la Physiol., vol. 42, pp. 624-629.

502 FREDERIC T. LEWIS

ELLENBERGER, W.AND Baum, H. 1908 Handbuch der vergleichenden Anatomie
der Haustiere. Zwolfte Auflage. Berlin.

Euze,C. 1909 Beitrag zur Histologie des embryonalen Saiugetierdarmes. Inaug.
Diss. Freiburg.

FABRICIUS AB AQUAPENDENTE, H. 1618 De gula, ventriculo, intestinis tractatus.
Patavii. (Opera omnia. Lugduni Batavorum, 1737.)

Froriep, A. 1907 Ueber Form und Lage des menschlichen Magens. Verh. d.

Gesellsch. deutsch. Naturf. u. Aerzte., Vers. 78, Teil 2, 2 Halfte, pp.
312-314.

GARDINER, J. P. 1907 A case of congenital hour-glass stomach with accessory
pancreas. Journ. Amer. Med. Assoc., vol. 49, pp. 1598-1600.

GYLLENSKOELD, O. 1862 Ueber die Fibrae obliquae in dem Magen. Arch. f.
Anat. Physiol. u. wiss. Med., Jahrg. 1862, pp. 132-139.

Hasse, C. anp Strecker, F. 1905 Der menschliche Magen. Arch. f. Anat. u.
Entw., Jahrg. 1895, pp. 33-54.

Hertz, A. F. 1910 The motor functions of the stomach. Quart. Journ. Med.,
vol. 3, pp. 373-394.

His, W. 1903 Studien an gehirteten Leichen iiber Form und Lagerung des
menschlichen Magens. Arch. f. Anat. u. Entw., Jahrg. 1903, pp. 345-
367.

Home, E. 1814 Lectures on comparative anatomy. Vol. 1. London.

Huscnke, E. 1844 Lehre von den Eingeweiden. Bd. 5 of Von Sémmering’s
; “Vom Baue des menschlichen Korpers.’”’ Leipzig.
Jonnesco, T. 1895 Appareil digestif. Traité d’anatomie humaine, ed. by P.
Poirier. Tome 4, Paris.
KarstLe, C., Rreper, H. anp Rosenruat, J. 1910 The bioroentgenography of
the internal organs. Arch. of the Roentgen Ray, vol. 15, pp. 3-18.

KaurMANN, R. 1907 Ueber Kontraktionsphinomene am Magen. Wein. klin.
Wochenschr., Jahrg. 20, pp. 1063-1069.

Keira, A. AND Jones, F. W. 1902 A note on the development of the fundus
of the human stomach. Proc. Anat. Soc. Gr. Brit. and Ire., pp. xxx1v-
XXXVIII, in Journ. Anat. and Physiol., vol. 36.

Lewis, F. T. anp Toyna, F.W. 1908 The regular occurrence of intestinal diver-
ticula in embryos of the pig, rabbit and man. Am. Jour. Anat., vol.
7, pp. 505-519.

Luscuxa, H. 1863 Die Anatomie des Menschen. Bd. 2, Abt. 1. Tubingen.

1869 Die Lage des menschlichen Magens. Vierteljahrsch. f. d. prakt.
Heilkunde, Prag., Bd. 101, pp. 114-126.

MecxeL, J. F. 1820 Handbuch der menschlichen Anatomie. Bd. 4. Halle
und Berlin.

THE FORM OF THE HUMAN STOMACH 503

Meyer, G. H. 1861 Lehrbuch der Anatomie des Menschen. Zweite Aufl.
Leipzig.

Moreaent, J. B. 1761 De sedibus et causismorborum. Venetiis. (Tr. by B.
Alexander, London, 1769.)

Moritz, 1895 Studien iiber die motorische Thiatigkeit des Magens. Zeitschr.
f. Biol., Bd. 32, pp. 313-869.

Mi.uuer, E. 1897 Beitrige zur Anatomie des menschlichen Foetus. Kongl.
Evenska Vetenskaps-A kad. Handl., Bd. 29, no. 2. Stockholm.

OweENn, R. 1868 On the anatomy of vertebrates: Vol. 3. London.

Rerzius, A. 1857 Bemerkungen itber das Antrum Pylori beim Menschen und
einigen Thieren. Arch. f. Anat. Physiol. u. wiss. Med., Jahrg. 1857,
pp. 74-87.

Riouan, J. 1618 Anthropographia. Parisiis. (De Vanthropographie. Tr. by
P. Constant. Paris, 1629.)

Rtpinepr, N. 1873 Topographisch-chirurgische Anatomie des Menschen.
Stuttgart.

SAnpiFoRT, HK. 1777-1781 Observationes anatomico-pathologicae. T.3. Lugd.
‘Bat. (Cited by Bettman.)

Sapprey, P.C. 1874 Traité d’anatomie descriptive. 2me édit., tome 4. Paris.

Scutrz, E. 1885 Zur Kenntniss der motorischen Function des Magens. Zeit-
schr. f. Heilkunde, Bd. 6, pp. 467-478.

Spiagevius, A. 1627 De humani corporis fabrica. Venetiis.

TanpterR, J. 1900 Zur Entwicklungsgeschichte des menschlichen Duodenum.
Morph. Jahrb., Bd. 29, pp. 187-216.

Tuyne, F. W. 1908 Models of the pancreas in embryos of the pig, rabbit, cat
andman. Am. Jour. Anat., vol. 7, pp. 489-503.

Totpt, C. 1879 Bau und Wachsthumsverinderungen der Gekroése des mesch-
lichen Darmkanales. Denkschr. kais. Akad. Wiss., Wien, math-
naturw. Classe, B. 41, Abt. 2, pp. 1-56.

Vesauius, A. 1543 De humani corporis fabrica. Liber 5, Caput 3. Basileae.

WaLpEYER, W. 1908 Die Magenstrasse. Sitz.-ber. d. kén. preuss. Akad. d.
Wiss., Bd. 29, pp. 595-606.

Wiuuis, T. 1674 Pharmaceutice rationalis. Hagae-Comitis. (Genevae, 1680).

Winstow, J. B. 1732 Exposition anatomique de la structure du corps humain.
Paris. (Tr. by G. Douglas, London, 1734.)

tetany
abn

Se

Sere

=a

bs

is port

yea eee

ug
ee ae

xy eee L

—NT
Bet

im.

E 0474

Sch beneneee ee neeees

Rep eaeeerea oa

spade teeayeuren

Promeneresrtes

ecperrereercers:

eetrees
SSI SS ee

egesetaseqaass SSeS

OI a teh as hina hegen ane te leone hee ge

arora ete wemnensineoee

x Saree
SSS oan

ers
woseeroses

SSS eee

atarpeermenre tees

Saeteaseegnicioes

Sinaia
SEEET Sinan

ates