' 7
ib oe
Patel i :
—
PREPERS
ae oe me.
=
=
= stowed
= SSSS>
a
natn mayan |
weaatnatectstetnctectas
eR ey Senne eS
; f hie
% R : tes " aU J 7
Da pera ie
y
oi
fs
a on t\ As.
iN hy Lia 2
BPR te yah
A ie avAs
ANT ik
Mi UPN TAY
Cais i!
TAA
Re SCR
PVRS
Reps vA i
Ne v4
y
'
i & :
i
y
'
i
ft
‘|
j
.
.,
-
~ 0
a
r i
i
“a
—e.
eras: ; fuel ¥)
———
vk
alsa
ee
ms
i ae :
; 5
f ; : Ne
n aap
- i)
ai
t b
» ’ 4
f=
Gen
—
mn ;
ie
{ J
a U y
ioe
¢ nit) pe
Pe a
=F 4
THE AMERICAN JOURNAL
OF
ANATOMY
EDITOR LAL BOA R
CHARLES R. BARDEEN G. Cart HUBER
University of Wisconsin University of Michigan
Henry H. DoNAaLDson Grorce S. HUNTINGTON
The Wistar Institute Columbia University
Simon H. GaGeE Henry Mck. KNower,
Cornell University Secretary
University of Cincinnati
VOLUME 19
1916
FRANKLIN P. Mau
Johns Hopkins University
J. PhuAyrain McMourric#
University of Toronto
GroraE A. PIERSOL
University of Pennsylvania
THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY
PHILADELPHIA, PA,
a woalty
“S ae
: oe Ne. ° era
r
‘
i) Wa i
a ewer
CONTENTS
No. 1. JANUARY
H. Hays Buuiarp. On the occurrence and physiological significance of fat in the mus-
cle fibers of the normal myocardium and atrio-ventricular system (bundle of His),
interstitial granules (mitochondria) and phospholipines in cardiac muscle. Sixteen
HU OMLTEC Sewer Rr ee ee ore os ce aes cher ere tha le SL ree Baker tes Saeco eee UN ots 1
R. R. Benstey. The normal mode of secretion in the thyroid gland. One plate in
R. R. Benstey. The influence of diet and iodides on the hyperplasia of the thyroid
PlANG, Gh GHOSSUMS IN, CAPLEVILY «Geena Boece a Satean sa sole se aemeay ee cae seprnuwnar eee 57
GrorGce L. Streeter. The vascular drainage of the endolymphatic sae and its topo-
graphical relation to the transverse sinus In the human embryo. Six figures....... 67
L. Botx. Problems of human dentition. Twenty-eight figures.....................0.. 91
No. 2. MARCH
M.R. Kine. The sino-ventricular system as demonstrated by the injection method.
DIKLeCH fIPUres (Vie DIA UCS). 5s 27 tation aes aaewe £ ab eal Se eoa es Seacoast a eieere 149
Joun Lewis Bremer. The interrelations of the mesonephros, kidney, and placenta in
different classes of animals. Twelve figures (two plates)................---eeeeee- 179
E. A. BAUMGARTNER. The development of the liver and pancreas in Amblystoma punc-
tatum. wHorty-sixngures (four plates)! 2. esc. «ssl. es cue clccins oeteineie seinieeeiee ieee 211
H.E.Jorpan. Themicroscopic structure of the yolk-sac of the pig embryo, with special
reference to the origin of the erythrocytes. Thirty-five figures (two plates)........ 277
C.M. Jackson. Effects of inanition upon the structure of the thyroid and parathyroid
glands of the’albino rat. Hourteen figures!)......+.2ss ses. ee sense cesses es ce sew ec 305
No. 3. MAY
J. A. Myrrs. Studies on the mammary glands. I. The growth and distribution of the
milk-ducts and the development of the nipple in the albino rat from birth to ten
WEEKS MO lp A Guys, se Mien chat Sood c toy cin arcgeh ate etsuaid ors extauay eres ta eraser cheusueveravere ec) etefae < Gucatae ecole sven ed 353
C. H. Danrortu. The relation of coronary and hepatic arteries in the common gonoids.
HOUT OUNTESiseapecatcyererere mercy costar sae ciate, Su erele aca revererate Vera csete ciel ny cnctolayoelciovsicista siecle 2 aeeer este ier 391
Vicror KE. Emmeu. The cell clusters in the dorsal aorta of mammalian embryos. Eleven
LUO UMESH (UW ON IAEER)| arcctarers cre ohh s sisisieusueveter siete Sea vars 40008 ave (o cieke erste caseie seve eon Gi cnememtercicys GieiGre aie 401
KE. V. Cowpry. The general functional significance of mitochondria.................-- 423
Wiuuiam A. Locy AND OLor LarsELut. The embryology of the bird’s lung. Based on
111
The
OMe
1
ON THE OCCURRENCE AND PHYSIOLOGICAL SIG-
NIFICANCE OF FAT IN THE MUSCLE FIBERS
OF THE NORMAL MYOCARDIUM AND ATRIO-VEN-
TRICULAR SYSTEM: INTERSTITIAL. GRANULES
(MITOCHONDRIA) AND PHOSPHOLIPINES IN CAR-
DIAC MUSCLE
H. HAYS BULLARD
From the Anatomical Laboratories, School of Medicine, University of Pittsburgh
SIXTEEN FIGURES!
CONTENTS
[OGG TIOM A ee Rete eee Lerten et ee PY oe ee ee age 1
INCRE AVOXS ISS s OGRE & Sey Cee eee eee a en) CG CREME Sp a OPE ane Sines karan nme Oe Ln oO RM 2 4
Dthhenentiabiontormenmtnalshaten. com cesses, eee te ee Beane 5
Occurrence of neutral fat in the heart muscle of different mammals under
WEAOUE MIAN KS Kono hn OMIBS va we oo oeem sive obocevanGanecadegbaeco sor 10
Physiological significance of neutral fat in cardiac muscle.................. 19
Occurrence and significance of neutral fat in the muscle tissue of the atrio-
VEMUTTCULareSV SUC) =6 no Me ie ea o Merc Sate( dco ex east ee tae eee ot Cee 21
Interstitial granules (mitochondria) and phospholipines in cardiac muscle 24
SAUCONY 3 oe ENE EOP RRR Poe en PEE MAT) RS NTC Geir SAE Perea ay ood 28
INTRODUCTION
In this paper I shall consider the occurrence and physiological
significance of microscopically demonstrable neutral fat in the
fibers of normal cardiac muscle and shall also discuss briefly the
so-called true interstitial granules and their relation to the
phospholipines of the heart. I have examined more than two
hundred apparently normal mammalian hearts. Of these, the
one hundred and forty-four last studied are listed below in
tabular form. . ;
Normal cardiac muscle fibers, like other tissues of the body,
contain a very considerable percentage of fats. This important
fact has been determined by Krehl (’93), Rosenfeld (’01), Leick
‘A large part of the cost of illustrations was borne by the Anatomical
Laboratories, University of Pittsburgh.
1
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, No. 1
JANUARY, 1916
2 H. HAYS BULLARD
and Winckler (’02), Rubow (’04—’05), Erlandsen (’07), Rosen-
bloom (713 a) and others who have analyzed portions of the myo-
eardium which had been carefully freed from adipose tissue. The
work of the biochemists just mentioned indicates that from two
to four per cent of the weight of the undried substance of the
myocardium is fat, half of which is present as phospholipines
while the other half is mainly neutral fat. The entire fat con-
tent of normal cardiac muscle is commonly supposed to be
‘invisible,’ that is, in such a form that it cannot be micro-
scopically demonstrated. A few investigators would appear to
make the total fat content not only ‘invisible’ but ‘combined’
or ‘masked;’ others would limit the latter terms to the compara-
tively small fraction of the total fats which can be extracted
only after protein digestion. According to Leathes (’10) there
is not sufficient evidence to justify us in assuming the existence
of a chemical combination of fat and protein. From Rosen-
bloom’s (13 b) recent review of the literature of the subject
one is impressed by the vague character of the information as
yet vouchsafed concerning this supposed fat-protein compound.
Virchow (47) taught that visible fat appears in cardiac muscle
only as the result of a pronounced retrogressive process, the de-
generation of cell protein into fat. By many clinicians, ‘fatty
degeneration’ in the heart was thought to lead to serious disturb-
ances of function of the organ with death as the inevitable out-
come. Welch (’88) demonstrated that this extreme view is not
tenable. He observed that rabbits kept for some time at a high
temperature and therefore known to possess fatty cardiac muscle,
show no symptoms of cardiac derangement either at the time of
the experiment or thereafter. He also found that cardiac fibers
which are crowded with fat will nevertheless contract rhythmic-
ally. Flexner (’94-95) showed that fat in the heart muscle
appears more frequently in certain infections than in others. He
pointed out that many clinical conditions formerly supposed to be
due to fatty degeneration of the heart are now known to be due
to other causes. Hasenfeld and Fenyvessy (’99) demonstrated
that even when cardiac muscle is extremely fatty, following phos-
phorus poisoning, the heart action is apparently unimpaired.
FAT AND MITOCHONDRIA IN CARDIAC MUSCLE 3
They interpreted this apparently normal action of the heart, in
the presence of a marked fatty change, as due to the very large
reserve force of the cardiac fibers. Pratt (’04) concluded that
there is no evidence to support the once commonly accepted
theory that fatty metamorphosis of the heart muscle is often
the cause of myocardial insufficiency. Many other observers
have arrived at similar conclusions and nearly all now agree with
the view advanced by Rosenfeld (01) and Herxheimer (’02)
that the fat in so-called fatty degeneration or metamorphosis is
in reality of infiltrative, not degenerative origin. Of importance
in this connection are the observations of Lambert (14) upon
tissue cultures of the chick heart. He finds that the source
of the fat in the embryonic muscle cells is the medium in which
the cells grow and that the fat droplets are not the result of cell
degeneration.
Referring especially to fat in the heart, Mallory (14) has re-
cently well expressed the view now held by most pathologists:
The fat makes its appearance in visible form because of diminished
utilization (oxidation) of the fat normally brought to the muscle cyto-
plasm. The fat accumulates as the result of two different causes, (a)
disturbances of nutrition and (b) toxemia. The accumulation of fat
in the cytoplasm of the muscle-fibers has of itself little significance.
It may in time unquestionably be utilized and removed. Its im-
portance pathologically lies in the evidence its presence gives of dis-
turbed cell metabolism and in its frequent association with necrosis.
Ostertag (’89), Ricker and Ellenbeck (’99), Fibiger (’01),
Arnold (’03), Keinath (04) and Babes (’08) have referred to the
isolated occurrence of visible fat in the apparently normal cardiac
fibers of certain species or individuals either without advocating
or without offering sufficient proof that fatty droplets are of usual
occurrence in this situation.
Hofbauer (05) describes visible fat in normal human fetal
heart muscle. Bell (11) was the first to show clearly that visible
fat (‘liposomes’) is normally present in the cardiac muscle of the
common laboratory mammals. He also demonstrated that the
quantity of visible fat is increased when fatty foods are given and
decreased when animals are starved. Bell’s work, as well as my
own observations in confirmation of his results (712 a), had ref-
4 H. HAYS BULLARD
erence mainly to skeletal muscle but we both stated that our
results also applied to cardiac muscle. Wegelin (13), working
independently, obtained experimental results similar to those
which Bell and I had previously published. Wegelin gives one
drawing showing fat in the heart muscle of a normal rat. With
this single exception there are, in the literature, no figures pur-
porting to show the normal fat content of cardiac muscle.
In the recent literature are two important communications,
by Eyselein (14) and Borchers (14). Both of these authors
are familiar with the work of Wegelin but neither is able to
confirm his results. In general, the observations recorded in
the present paper are in agreement with those of Hofbauer
(05), Bell (1112), Wegelin (13), and with my own previous
work (712-14).
I am indebted to Prof. R. E. Sheldon for kind encouragement
and valuable criticism in connection with the work of this paper.
In making the figures I have received a number of helpful sug-
gestions from Miss 8. E. Watson, artist of the Department of
Anatomy.
METHODS
Technique of fat demonstration. In the demonstration of fats
in tissue sections, methods and technique are of the utmost im-
portance. ‘The wide divergence of opinion among different in-
vestigators concerning the occurrence of fat within the cells of
normal and of pathological tissues is due primarily to differences
of technique in preparing the sections for examination. I have
elsewhere (712 b) treated this subject in some detail and shall
here give only a brief outline.
Fixation. Fat containing material is usually fixed in formalin.
Bell (11) and Bullard (’12) have pointed out that this fixative,
although frequently giving excellent results, is not to be relied
upon under all circumstances. Frozen sections of the heart or
other tissues which have been fixed in formalin may, when care-
fully stained (Scharlach R.), appear to contain little fat, while
sections of the same material when stained fresh without previous
fixation may be loaded with droplets. This apparent disappear-
~
FAT AND MITOCHONDRIA IN CARDIAC MUSCLE o
ance of fat may be noticed when the blocks of tissue have re-
mained in the fixative only a few hours, at other times it occurs
only after several weeks or months, if at all.
As a fixative I now employ formalin which has been neutralized
and distilled according to the method given by Mann (’02) in his
Physiological Histology, p. 88. A twenty per cent solution
of formalin is prepared and is rendered isotonic by the addition
of 0.75 gm. of sodium chloride to each 100 ee. of the fluid. With
short fixation in this solution the quantity of fat usually does not
differ from that seen in fresh tissue. Blocks are fixed for thirty
minutes to five hours and are then cut on thefreezing microtome.
Frozen sections of fresh unfixed tissue are employed as controls.
Staining. In this study I have employed all the fat stains in
common use but principally Herxheimer’s alkaline alcoholic solu-
tion of Scharlach R., which stain usually shows much more fat
than the simple alcoholic solutions of the same dye. The latter
solutions sometimes fail to stain a large part of the fatty drop-
lets seen in the fresh unstained tissue. This is not the case with
Herxheimer’s solution. Herxheimer’s stain is a saturated solu-
tion of Scharlach R. in seventy per cent alcohol to which sodium
hydroxide, 2 gm. per 100 ec. has been added. It is essential
that care be taken to avoid precipitates. Also the excess stain
must be thoroughly washed out of the sections before the nuclei
are stained with dilute hematoxylin. For the details of this
method the reader is referred to Herxheimer’s papers (’01—’02), or
to my former papers (712).
DIFFERENTIATION OF NEUTRAL FAT
' Among recent contributions to our knowledge of the technique
and chemistry of fat demonstration may be mentioned those of
Herxheimer (’01), J. Lorraine Smith (0607-710), Smith, Mair
and Thorpe (’08), Smith and Mair (10), Fischler (04), Fauré-
Fremiet, Mayer and Schaeffer (’10), Eisenberg (’10), Klotz (06),
Aschoff (’09), Kawamura (711) and Hanes and Rosenbloom (’11).
These observers have introduced a number of valuable staining
methods but more important still they have established a large
6 H. HAYS BULLARD
number of facts relating to the optical, chemical and physical
properties exhibited by pure fats and fatty mixtures when ob-
served.after being artificially introduced into the tissues or when
studied as smears on tissue paper. Application of the knowledge
thus gained makes possible, in certain eases, the identification of
various fats as they occur in the tissues.
Figures 1, 2, 3, 4, and 5 represent sections of the myocardium
of rats, figure 10 is from a dog and figure 9 from a fattened hog.
The preparations were stained with Herxheimer’s Scharlach R.
The number of normal hearts which I have examined by this
method is more than two hundred and always with results simi-
lar to those represented in the figures. It is of course well recog-
nized that Scharlach R. is not specific for neutral fat. Never-
theless, I believe that the colored droplets shown in these figures
are, at least in very large part, neutral fat. The reasons for this
belief as outlined below are essentially those advanced in a former
paper (12a). Concerning similar fatty droplets in cardiac
muscle Wegelin (713), likewise, has come to the conclusion that
they are neutral fat and for much the same reasons.
In unstained preparations these droplets are to be seen as
approximately spherical, highly refractive, isotropic bodies which
do not disappear in acetic acid or in dilute alkalies, ‘orm no
myelin figures, but are completely dissolved by absolute alcohol
and other fat solvents. They stain characteristically with
Scharlach R. and stain red, not blue, with nile blue sulphate and
nile blue chlorhydrate. They do not stain with basic anilin
dyes and are not rendered insoluble by the action of potassium
bichromate. The above combination of properties makes it
appear certain that the droplets under consideration are neutral
fat and not any of the other fats occurring either normally or
pathologically in cardiac muscle as, phospholipines, cholester-
inester, ete.
FAT AND MITOCHONDRIA IN CARDIAC MUSCLE re
RELATION OF NEUTRAL FAT TO STRUCTURE OF CARDIAC MUSCLE
Position of fat in muscle fibers. Virchow (’47) pointed out that
the pathological fat droplets of cardiac muscle fibers are situ-
ated in the sarcoplasm, not in the fibrillae or muscle columns.
This view has been confirmed by nearly all subsequent observers.
Welch (’88) observed that the droplets are arranged in rows be-
tween the fibrillae and according to Wegelin (713) transverse rows
are found in segment J on either side of the membrane of Krause.
In well stained preparations of normal cardiac muscle, made
by the Herxheimer method, it is very clearly seen that the fat
droplets are found inthe sarcoplasm, never in the muscle ¢éolumns
or myo-fibrillae. The droplets are arranged in both longitudinal
and transverse rows, figure 5. When the muscle fiber is con-
tracted the larger droplets, about 2 » in diameter, are spherical
and each occupies an entire segment of the fiber extending be-
tween adjacent Krause’s membranes. When the fiber is extended
the large droplets are usually somewhat elongated and are found
in the anisotropic segment Q. Small droplets, 1 » or less in diam-
eter, are arranged in transverse rows in the isotropic segment
J on either side of Krause’s membrane.
Light and dark fibers. As regards distribution of affected
fibers, Ribbert (’97) recognized three types of fatty degeneration:
1) diffuse general degeneration in which all the fibers contain fat;
2) mottled peri-arterial degeneration affecting areas immediately
surrounding the smaller arteries; 3) mottled degeneration oc-
curring in areas most distant from the smaller arteries. This
latter type gives the well known tiger-lily or thrush-breast
appearance usually most marked in the papillary muscles.
In normal cardiac muscle the fatty fibers do not give rise to the
mottled appearance but conform closely to the diffuse general
type of Ribbert. All of the fibers of a specimen and of the whole
heart may show a uniform amount of fat, figure 1, and figure 6 at
C. In an equal number of individuals certain of the fibers con-
tain much more fat than others, figures 3, 4, 5, 9, 10, and 13.
This latter distribution of the fat droplets, at times observed in
all the species (rat, cat, dog, hog, sheep, ox, man) here studied, is
S H. HAYS BULLARD
similar to that occurring in the so-called light and dark fibers of
skeletal muscle.
The existence of light, dark and intermediate fibers (by trans-
mitted light) has been known in skeletal muscle for a long time.
Knoll (8991) and Schaffer (’93) described the dark fibers as
containing many interstitial granules and fatty droplets while
the light fibers contained comparatively few granules and little
fat. Figure 14 shows a section of striated muscle fibers from the
diaphragm of a dog. At D is seen a fatty or dark fiber, and at
L a slightly fatty or light fiber. The work of Knoll and Schaffer
was done before Scharlach R. was extensively used as a fat stain
so that they did not obtain the exact picture shown in figure 14
but there can be little doubt that the types of voluntary striated
fibers here shown (fig. 14) correspond to the light and dark fibers
of Knoll. In skeletal muscle the fibers are of uniform type,
either light or dark throughout their entire length, and as was
pointed out by Knoll it is easy, in certain cases, to observe mor-
phological differences between the two types. A goodexample
of this is found in the breast muscle of the pigeon. Here the
light fibers are large, with nuclei placed within their substance and
contain little fat, while the dark fibers are small, with nuclei
peripherally situated, and contain a great deal of fat. It is cer-
tain that in the pigeon light and dark fibers are definitely fixed
types and not morphologically identical. In the skeletal muscles
of mammals a morphological difference is often observed in that
the dark fibers are of less diameter than the light. In a former
paper (712 a) I reported having observed the two types of fibers
in the skeletal muscles of the human fetus and in the fetal calf.
I have not been able to differentiate the two types during the
first half of fetal life but in the human fetus they are well marked
as early as the sixth and seventh month. The relative number
and arrangement of the dark (fatty) fibers in the different muscles
of the human fetus is so similar to that found in the adult that it
seems certain that the dark fibers of the fetus remain true to
type in post-uterine life. In the mammalian fetal heart usually
the different types of fibers are not clearly marked although
they contain fat droplets.
FAT AND MITOCHONDRIA IN CARDIAC MUSCLE 9g
According to Knoll (91) and Schaefer (12) cardiac muscle
fibers correspond to the dark fibers of skeletal muscle. It has
been my experience, however, that in many cases the different
types of fibers, dark, hght and intermediate are quite as well
marked in cardiac as in skeletal muscle. Figures 3, 4, and 5
show cardiac fibers of the rat, figure 10 those of a dog and figure
9 those of afattened hog. The different types of fibers are clearly
shown in each of these figures. Dark fibers are designated D,
light fibers .L Figure 5 shows a longitudinal section from the
same specimen as the transverse section shown in figure 4. In
the longitudinal section it is seen that after a brief course fibers
of one type, dark or fatty, pass abruptly into those of different
type, light or shghtly fatty. This change of type occurs along the
transverse lines marked out by the intercalated disks. The
greater number of intercalated disks, however, mark no change
of type. A cardiac fiber of any given type includes, therefore, a
variable number of so-called cardiac cells. In inanition, as we
shall see, fat gradually disappears from heart muscle. All the
fibers then appear light and it is frequently impossible to dis-
tinguish one type from the other, figure 1. Similarly when the
muscle fibers, as in certain fat fed animals, are loaded with fat
the light fibers may be so crowded with fat droplets as to present
the same appearance as dark fibers, figure 6.
In the skeletal muscles of nearly all apparently normal mam-
mals including man, dark or fatty fibers are found, almost invari-
ably, side by side with others which are light or non-fatty and
as | have set forth the two types are also of frequent occurrence
in apparently normal cardiac muscle. This indicates that dark
or fatty fibers are to be considered normal, not pathological.
The occurrence of light and dark fibers is usually accounted for
on the theory that dark fibers have undergone pathological
‘fatty degeneration’ while light fibers have escaped the patho-
logical process. This explanation we cannot accept for reasons
given, as well as for others to be stated later.
Distribution of fatly fibers in the heart. The hearts of the rats
here used were usually prepared for study by making frozen
sections extending transversely across both ventricles. Such
10 H. HAYS BULLARD
preparations bring out the fact that the fatty fibers are distrib-
uted with approximate uniformity throughout the myocardium
of both ventricles. This is also true in the cat and doubtless in
other animals although I have not studied the question ex-
haustively. In the auricles the fat content of the fibers appears
normally to parallel that found in the ventricles. Figure 13
represents a transverse section of fibers from the right auricle of
adog. Figure 10 at Land D shows cardiac fibers from the inter-
ventricular septum of the same heart. The number of fat drop-
lets and the distribution of light and dark fibers in the auricle is
similar to that in the ventricle.
OCCURRENCE OF NEUTRAL FAT IN THE HEART MUSCLE OF DIF-
FERENT MAMMALS UNDER VARIOUS NUTRITIVE CONDITIONS
The data upon which this investigation rests are, in part,
given below in tabular form. The animals are grouped ac-
cording to species and also with respect to character of food.
Although in nearly all animals some cardiac fibers hold much more
fat than others, the distribution of fatty fibers is so uniform that
in any given heart the quantity of fat in sections taken at ran-
dom from different parts of the ventricles is approximately the
same in all sections. This makes it possible, in any given indi-
vidual, to represent fairly accurately the amount of fat in the
ventricular fibers by one of the following five designations viz:
very large, figure 6; large, figure 4; moderate, figure 3; small, figure
2; very small, figure 1. An acquaintance with the literature makes
it appear that what is ‘very large’ to one author is but ‘large’
to another and ‘moderate’ toathird. In order to show with some
clearness what is here intended to be conveyed by the various
designations just given a type drawing is referred to in each case.
From the tables it will at once be noted that animals, especially
rats, kept for a short time on a fatty diet, have much more fat
in the heart muscle than do those which are on a diet of carbo-
hydrate and protein with only a small amount of fat. In inani-
tion, however, animals show a comparatively small amount of
fat in the cardiac fibers. The fat findings here given are based
upon preparations stained by Herxheimer’s alkalin-alcoholic
FAT AND MITOCHONDRIA IN CARDIAC MUSCLE 11
Scharlach R. The various items of the tables will be more fully
explained in the discussion which follows.
Group 1 (table 1) consists of fifteen adult rats fed for three
weeks with an abundance of raw grain, wheat bread and boiled
beef. The quantity of fat allowed was somewhat less than the
animals appeared to desire. This group of rats may be consid-
ered as having received food, fat, protein, and carbohydrate,
suitable for normal maintenance and growth. ‘These rats were
kept in large well ventilated cages and when killed were all
in good nutritive condition. Of the fifteen rats in the group,
(table 1) ten have a moderate amount of fat in the cardiac fibers
as in figure 3; two have a large amount of fat in the fibers as in
figure 4, three have a small amount as in figure 2. Figure 3,
showing as it does a moderate amount of fat, may be taken as
representing the average condition of the group.
TABLE 1
Albino rats, normally fed group, kept
for three weeks on wheat bread, raw
grain, and boiled beef with a small
quantity of fat
ANIMAL NUMBER
WEIGHT IN GRAMS
AMOUNT OF FAT IN HEART MUSCLE
(LEFT VENTRICLE)
(SMALL AS IN FIG. 2)
(MODERATE AS IN FIG. 3)
(LARGE AS IN FIG. 4)
KILLED
WHEN
—
bo
[@7)
moderate
178} large
137| moderate
185} moderate
176| small
196} moderate, fig. 3
135| moderate
8 | 187) large
NOOR WW FH
9 | 139} moderate
10 | 187} small, fig. 2
11 | 196} moderate
12 | 185} moderate
13 | 143] small
14 | 156} moderate
t
163} moderate
_
Co
12 H. HAYS BULLARD
Group 2 (table 2) consists of ten rats in various stages of inani-
tion. As is seen from the table these rats had been without food
for forty-eight to ninety-six hours and had lost from twelve to
twenty-four per cent in body weight. The three members of the
group which had lost as much as twenty per cent in weight
showed, upon section, little or no subcutaneous fat and but slight
traces in the omentum. Several of these animals are to be re-
garded as in the last stages of inanition. Of the ten members
of the group five have a very small quantity of fat in the cardiac
fibers, as in figure 1, four have a small amount, as in figure 2,
and one a moderate amount. Figure 1 from rat no. 23 (loss of
weight twenty per cent) shows a very small amount of fat and
this figure may be taken as characteristic of the inanition group.
The animals of group 3 (table 3) were fed fats in the form of
butter, olive oil, pork fat and egg yolk. In order to increase the
quantity of fat consumed no food was given for twenty-four hours
preceding the initial feeding. When fats were to be given for
several days, feeding was not preceded by a fast. A little grain
TABLE 2
Albino rats, inanition group, rats 16—
20 inclusive, no foed for 48 hours;
rats 21 and 22, no food for 60 hours;
rats 23-25 inclusive, no food for 96
hours, water supplied
D a
ao 2 alae
alice 4 & & | AMOUNT OF FAT IN HEART
5 Pa] - & | MUSCLE (LEFT VENTRICLE)
Zz es ere (MODERATE AS IN FIG. 3)
S) E is oR (SMALL AS IN FIG. 2)
3 Sk], %| (VERY SMALL 4s IN FIG. 1)
5 | ae| @e
< z a
16 | 147) 12 | small
17 | 173} 14 | very small
18 | 116) 14 |} small
19 | 148} 13 | small
20 | 137] 13 | moderate
21 | 176] 15 | small
22 99| 16 | very small
23 | 185] 20 | very small, fig. 1
24 96| 24 | very small
25 90) 23 | very small
FAT AND MITOCHONDRIA IN CARDIAC MUSCLE £3
was given in nearly all cases for the reason that it appears to
stimulate the appetite of the animals for fats. It is known, more-
over, that fat metabolism does not proceed normally in the
absence of carbohydrates. As shown in table 3 the thirty-
TABLE 3
Albino rats, fat fed group. The rats of this growp were given all the fatty food that
they would eat plus a small amount of raw grain
. AMOUNT OF FAT IN HEART
anual, | axans wares reo is MUSCLE (EFT VENTRICLE)
(LARGE AS IN FIG, 4)
26 156 Butter for 10 hours large
27 109 Butter for 10 hours large
28 188 Butter for 20 hours large
29 167 Butter for 20 hours large
30 179 Butter for 20 hours large
31 112 Butter for 36 hours moderate
32 146 Butter for 48 hours large
33 179 Butter for 14 days large
34 183 Butter for 14 days moderate
35 186 Butter for 14 days large
36 194 Butter for 14 days moderate
30 173 Butter for 7 days large
38 115 Butter for 7 days moderate
39 147 Olive oil for 20 hours moderate
40 179 Olive oil for 20 hours large
41 168 Olive oil for 20 hours moderate
42 -114 Olive oil for 20 hours moderate
AS 164 Pork fat for 10 hours large
44 183 Pork fat for 15 hours large
45 198 Pork fat for 20 hours moderate
46 162 Pork fat for 20 hours large, figs. 4 and 5
47 174 Pork fat for 36 hours large
48 195 Pork fat for 14 days moderate
49 181 Pork fat for 14 days moderate
50 60 Egg yolk for 20 hours moderate
51 169 Egg yolk for 20 hours large
52 81 Egg yolk for 20 hours large
53 114 Egg yolk for 36 hours large
54 209 Egg yolk for 36 hours moderate
55 78 Egg yolk for 6 days moderate
56 234 Egg yolk for 14 days large
57 193 Egg yolk for 14 days moderate
58 101 Egg yolk for 14 days moderate
59 164 Kgg yolk for 14 days large
14 H. HAYS BULLARD
four individual rats of this group were fed as follows: four,
olive oil; seven, pork fat; ten, egg yolk; thirteen, butter. Fif-
teen individuals of the group have a moderate amount of fat
in the myocardial fibers, as in figure 3, while nineteen have a
large amount as in figure 4. The cardiac fibers of the rats in
this group contain an unusual amount of neutral fat due no doubt
to the fatty character of the food. The animals would eat but
sparingly of olive oil and the effect produced was less marked
than in the case of butter and pork fat which were eagerly
consumed. A number of rats (nos. 48, 49, 57, 58) would eat
but ‘little fat after the first few days and the fat content of
the myocardial fibers was no more than in animals living on
carbohydrates and protein. Figure 4 represents a transverse
section of ventricular fibers from a rat (no. 46) which was
killed twenty hours after consuming eight or ten grams of
pork fat. The cardiac fibers are loaded with droplets. As will
be seen from table 3 the large amount of fat makes its ap-
pearance in the heart with astonishing rapidity, reaching a
maximum in from twelve to twenty-four hours after but one or
two large feedings of a fatty food. Even when a fatty diet is
continued for as long as twelve or fourteen days the amount of
fat in the cardiac fibers is no more than after a single large fatty
meal, although the animal may show a marked increase of adi-
pose tissue.
Figure 4 may be taken as representative of the fat fed group,
just as figure 1 was considered representative of the inanition
group and figure 3 of the normally fed group. It is quite clear
that the cardiac muscle fibers of the rat normally contain a very
considerable quantity of fat in a microscopically visible form.
Also in inanition the normal quantity is diminished almost to
the point of complete disappearance while in animals on a fatty
diet the cardiac fibers are usually loaded with droplets.
Cats. Table 4 shows a group of normal cats which were fed
for three to ten days on a well balanced ration of bread, milk and
moderately fat boiled beef. Of the twenty animals in this group
eleven have a moderate amount of fat in the cardiac fibers
(similar to fig. 3), six have a large amount (similar to fig. 4),
FAT AND MITOCHONDRIA IN CARDIAC MUSCLE 15
TABLE 4
Cats, fed three to ten days on wheal bread, whole milk and moderately fat boiled beef
AMOUNT OF FAT IN HEART
ee ee |) eamerrme 2) | Suomen rag me moses
KILLED So inee ene ciara ey (LEYT LIMB )
(VERY LARGE AS IN FIG. 5)
60 332 large large
61 337 large moderate
62 368 moderate large
63 375 very large very large
64 416 moderate very large
65 445 moderate moderate
66 504 large very large
67 525 moderate moderate
68 592 moderate large
69 1054 moderate large
70 1376 large very large
71 f6H64 large large
72 1830 large moderate
73 2206 moderate moderate
74. 2235 small moderate
75 2245 moderate moderate
76 2364 moderate large
edi 2432 moderate moderate
78 2750 moderate moderate
7§ 2816 small small
two have a small amount (similar to fig. 2) and one has a very
large amount (similar to fig. 6). It is evident that the myocar-
dial fibers of normal cats contain fat in visible form and to an
extent exceeding that found inrats. After ninety-six hours with-
out food, rats are usually in a state of extreme inanition and show
very iew droplets in the cardiac fibers (table 3) but the same does
not hold for cats. Cats (nos. 80 and 83, table 5) kept for ninety-
six hours without food still show a moderate amount of fat in the
heart. As is well known eats reach the last stages of inanition
only after having been kept for about three weeks without food.
For the purposes of this investigation, I have thought it un-
necessary to subject cats to a long period of starvation. My
results show only that during a fast of three or four days fat does
not disappear from the cat heart. It is very probable that in the
lt
)
H. HAYS BULLARD
last stages of inanition the cardiac fibers of the cat, like those of
the rat, would contain little fat.
TABLE 5
Cats, fed as indicated
foo] g . AMOUNT OF FAT
Alea IN HEART MUSCLE (LEFT
2 : : Geena a AMOUNT OF FAT IN
ash FOOD GIVEN FIG. 3) MUSCLE FIBERS OF HIS
a 5 cs (LARGE AS IN FIG. 4) BUNDLE (LEFT LIMB
a oe (VERY LARGE AS IN
Z > a FIG. 6)
80| 465| No food for 96 hours moderate moderate
81| 620| 4 day fast, given butter, killed af-
ter 20 hours very large very large
82| 576| 3 day fast, given pork fat 15,
grams, killed after 20 hours. very large large
83| 666} No food for 96 hours large large
84| 810} 3 day fast, given pork fat large
meal, killed after 17 hours very large moderate, fig. 6
85| 885] 3 day fast, fed 7 gms. butter,
killed after 17 hours moderate small
86|1137| Bread and water 7 days moderate moderate
87|1194| Pork fat 1 day large large
88/1259} Bread and water 10 days moderate moderate
89}1285| Bread and water 10 days moderate small
i
TABLE 6
Dogs, killed as soon as brought to the
laboratory, animals in good nutri-
tive condition, no special feeding
& | AMOUNT OF FAT IN
|HEART MUSCLE (LEFT 7
Bl) a ieannyatene 0 akreecree ete on
7 | (SMALL AS IN FIG. 2) STR aS 5
4 (MODERATE AS iaaiet anes
3 IN FIG. 3)
= (LARGE AS IN FIG. 4)
500.
3 The same as figures 1 and 2 from an albino rat fed as in figure 2 (animal
6, table 1). Moderate amount of fat in the muscle fibers; L, ‘light’ cardiac fibers;
D, ‘dark’ cardiac fibers. 500.
4 The same as figures 1, 2 and 3 from an albino rat, killed twenty hours after
consuming 8 or 10 grams of pork fat (animal 46, table 3). The amount of fat
in the cardiac fibers is large; L, ‘light’ fiber; D, ‘dark’ fiber. > 500.
5 Longitudinal section from same preparation as in figure 4; L, ‘light’ fibers;
D, ‘dark’ fibers (animal 46, table 3). X 500.
6 From the interventricular septum of a cat killed 17 hours after receiving
a large meal of pork fat (animal 84, table 5). Very large amount of fat in cardiac
fibers, C; moderate or large amount in Purkinje fibers of the left limb of bundle of
His; 22) <500:
7 From the interventricular septum of an eight months’ human fetus (animal
141, table 8). There is a moderate amount of fat in the cardiac fibers, C, and also
in the Purkinje fiber of the left limb of the bundle of His, P. > 500.
8 From the interventricular septum of the heart of a woman aged 53, fatal
lobar pneumonia, cardiac fibers, C, contain a moderate amount of fat; fibers of
left limb of bundle of His, P, show a very large amount of fat. > 300.
32
_FAT AND MITOCHONDRIA IN CARDIAC MUSCLE PLATE 1
H. HAYS BULLARD
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, NO. ]
PLATH 2
EXPLANATION OF FIGURES
9 From the moderator band of a fattened hog (animal 104, table 7). Fibers
of right limb of bundle of His, P, and light cardiac fibers, L, contain a very small
amount of fat, dark cardiac fibers, D, a large amount. > 300.
10 From the interventricular septum and left limb of the bundle of His of
a well nourished dog (figs. 10 to 16 inclusive from same animal, no. 90, table
6). Dark cardiac fibers, D, and light cardiac fibers, L, contain a moderate
amount of fat; Purkinje fibers, P, contain a small amount of fat. > 500.
11 Muscle fibers from the sino-auricular node, heart of a normal dog (figs.
10 to 16 inclusive from same animal, no. 90, table 6). Fat content of muscle
fibers small. X 500.
12. Muscle fibers from the atrio-ventricular node, heart of a normal dog (figs.
10 to 16 inclusive from same animal, no. 90, table 6). Fat content of muscle
fibers small. X 500.
13. Cardiac fibers from the right atrium of a normal dog (figs. 10 to 16 inclu-
sive from same animal, no. 90, table 6), fat content moderate. >< 500.
14 Skeletal muscle fibers from the diaphragm of a normal dog (figs. 10 to 16
inclusive from same animal, no. 90, table 6), large amount of fat in dark fibers, D,
and small amount in light fibers, L. > 500.
15 Cardiac fiber from interventricular septum of a dog (fig. 10 to 16 inclu-
sive from same animal, no. 90, table 6). True interstitial granules, G (mito-
chondria, sarcosomes, @ granules), muscle columns M, Krause’s membranes Z.
« 1600.
16 Same as figure 15, transverse section. X 1600.
34
PLATE 2
FAT AND MITOCHONDRIA IN CARDIAC MUSCLE
H. HAYS BULLARD
2,0
oP es, = oer ee
22a Ser .: aos ? ste
:@ * gy! re
‘ FF,
oy .
pt
‘ he, ei
7
‘can
OF) Tp
y
f
al}
earits (ty) eee
i} 4 is
pls dt) ae
uf ia
Pt i £
‘ v
j 5 =
at |
Pell
’ be —
j ~=— - 1) ;
fe. | ‘
An y a Ny j
vial mM
ve. Le
> » F th.
. ae
: ees a
an : i
= a
Eg i tes is
=a Vv .
J i
‘ ' ae
i = :
'
i
J *
ie
:
‘
'
:
7
an
’
cd
é
ry
a
. i)
if)
‘
at
re
i
Ta
ORM,
> , i
vit
i
t -) ‘
[
rm
; ,
ave
,
oe
r
be
THE NORMAL MODE OF SECRETION IN THE
THYROID GLAND
R. R. BENSLEY
From the Hull Laboratory of Anatomy, University of Chicago
ONE PLATE IN COLOR
In the glands of the alimentary canal the process of secretion is
associated with definite changes in the structure of the secreting
cells, and with the accumulation in them of products, granular
or otherwise, which may be interpreted as the organic antece-
dents of the secretion itself. Even in some of the internal secret-
ing glands, as, for example, the islets of Langerhans of the pan-
creas, functioning is associated with the storage or exhaustion of
intracellular products which may be similarly interpreted. By
means of these secretion antecedents an observer, who has, by
experiment and observation, acquainted himself with the secre-
tory mode, may form an estimate of the secretory potential at
the time of observation.
In the thyroid gland, on the other hand, the search for such
evidences of secretory activity, has been, as regards the nature
of the intracellular, secretion antecedents, of so contradictory a
nature, and of such doubtful functional import, that, at present,
we are unable to state from the examination of a thyroid gland
whether the gland was active or inactive. Accordingly, differ-
ent observers, as, for example, in Grave’s disease, in discuss-
ing the same results, have arrived at diametrically opposed
conclusions. .
One of the features of the thyroid gland, in particular, which
baffled interpretation was the presence in it of a storage product,
the so-called colloid, the route and rate of resorption of which
have remained problematical, though chemical and physiologi-
‘al studies indicated that it contained the physiologically active
QF
ol
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, No. 1
‘
38 R. R. BENSLEY
thyroid substances. Some observers have even doubted the
resorption of this material, and have suggested that the function
of the thyroid gland was primarily to withdraw toxic substances
from the blood. Others have conceived the colloid as a sort of
menstruum in which the real thyroid secretion was received and
from which it might be withdrawn without visible change in the
colloid itself. Still others have held the view that the colloid
was the real secretion of the thyroid gland and that the normal
mechanism of thyroid secretion was by this indirect route, first
secreting into the centre of the follicle, and then withdrawing
this ware-housed material, as functional needs required, by some
unknown method and route.
The determination of the true significance of the colloid in the
secretory cycle of the gland, and of the ways in which it is formed,
and of its intracellular antecedents, is of fundamental importance
in the physiology and pathology of the thyroid gland. The con-
viction that it is by this indirect method that the thyroid gland
produces its internal secretion lies at the bottom of ali of our
more or less speculative interpretations of pathological conditions,
and in view of the strong physiological evidence supporting this
conviction few have had the courage to question its accuracy.
Many authors have tried nevertheless to influence experimentally
the rate of secretion in the gland, and to read in the changes so
produced the true history of its secretory process. In this way
many interesting facts have been discovered, which at present
seem to some extent contradictory of one another, but which
nevertheless must be found to be in accord when the true history
of the process is revealed.
Our earliest knowledge as to the origin of the intrafollicular
colloid of the thyroid gland is due to Biondi and Langendorff.
Biondi $B) showed that this substance was a true product of
the secretory activity of the thyroid epithelial cells, inasmuch
as he found globules of similarly staining substances in the cells
themselves. He conceived the process of secretion as follows:
the cells of the thyroid gland produce the colloid, since one can
see in them little globules having the same microchemical reac-
tions; the vesicle has a tendency to increase in size partly by
NORMAL MODE OF SECRETION IN THYROID GLAND 39
multiplication of the epithelial cells, partly by increase of the col-
loid; after filling itself the vesicle discharges into the nearest
lymphatic vessel; finally the collapsed vesicle disposes itself in the
form of a number of little acini which repeat the process.
Langendorff (’89) using the method of comparative study for
the elucidation of the secretory process in the cells of the thy-
roid gland, reached conclusions which, in some respects, confirm
and extend those of Biondi. He described two sorts of cells in
the gland which he designated, respectively, principal cells, and
colloid cells. The principal cells constituted the main mass of
the epithelium. They were cylindrical or columnar cells, of
variable height in different species and in different ages of the
same animal species. They possessed a reticular protoplasm,
with granules at the nodal points, and an oval or round nucleus
situated at the basal end of the cell. Like Biondi he saw occa-
sionally in these cells small hyaline spherules, but considered
them to occur very rarely. The colloid cells differed from the
principal cells by the hyaline, transparent appearance of their
cytoplasm. This cytoplasm browned with osmic acid, and, in
dyes, stained the same as the colloid content of the follicles.
He found all grades of transition between the colloid cells and the
principal cells. He regarded the colloid cells as elements en-
gaged in the secretion of colloid but did not commit himself
definitely to the opinion that, after a period of secretion, they
might return to the state of the principal cells. He was like-
wise in doubt whether they degenerated or not after secretion.
V. Wyss (89) studied the effects on the thyroid gland pro-
duced by poisoning with pilocarpine. He found in cats and dogs
that the gland after pilocarpine was large, turgid, and filled
with blood, and that the cells were larger, the nuclei less appar-
ent. The free ends of the cells were prolonged into processes
which were continuous with the colloid mass, and between these
processes were brilliant spherules of apparently fluid nature.
Anderson (’94) confirmed V. Wyss’ conclusions relative to the
effect. of pilocarpine on the gland, and studied the structure of
the epithelial cells in young cats and rabbits at different periods
of time after injections of pilocarpine. He described, in the
40 R. R. BENSLEY
earlier phases of pilocarpinisation, the appearance of clear drop-
lets in the cytoplasm, which collected at the free pole of the
cell, to be extruded in the form of small droplets into the cavity
of the vesicle. These, therefore, he regarded as the antece-
dents of the clear vacuoles of the margin of the colloid and on
account of their lack of affinity for dyes, designated chromo-
phobe secretion.
.
'
- t a
i - ;
:
+, t i
, a a
7 . |
re
‘ (
6] s t t
zs ¢
?
»
x i
; 7
.
: a vee
i Z >
,
i A
‘
. — ;
al '
,
*
ty
THE VASCULAR DRAINAGE OF THE ENDOLYM-
PHATIC SAC AND ITS TOPOGRAPHICAL RELA-
TION TO THE TRANSVERSE SINUS IN THE HUMAN
EMBRYO
GEORGE L. STREETER
Department of Embryology, Carnegie Institution of Washington
SIX FIGURES
CONTENTS
linGROCMCHOM ee aesemecee cose oe misige eee cess sh sae eee e ces cece esas sane e 67
Miaterialvand, methods... ...2 seuss esses ee oe se stole te tcl sree teenies eects 68
RDS ORLCAN cya N ed ce cies aratiemee Heater poe «1 ie Pear POET SR Uaear meine Hee ae te 69
Endolymphatic appendage at different stages
1. Human embryos one and two months old........................0.... id
2. Human embryos during third month.........................6002-0-0- ii
3. Human embryos four months old.................. 0. ccc cece sees cease S1
4. Human embryos during seventh month.......................2.....0. 85
TPT TEU ewe erecta ere aie no) aioe ceseaee se woe taint oe Ae odere ime Gare ene te 87
IGHETeNGES CIECAE. Ma. Senn se saan. ote oases ene gate von Oe oe 89
INTRODUCTION
In a previous paper (Streeter 714) dealing with some experi-
mental studies on amphibian larvae, it was shown that in the
tadpole the endolymphatic sac always lies in close apposition to
the membranous roof of the hind-brain. This relation exists not
only in normal specimens, but it was also found that in specimens
where the ear vesicle had been rotated or transplanted by opera-
tive procedure, the endolymphatic sac in the subsequent self-
correction of posture, succeeds in most cases in attaching itself
to the membranous chorioidal roof in the normal manner.
This interesting topographical relation of the endolymphatic
sac in the tadpole, induced the writer to examine more closely the
endolymphatic sac in later human embryos, and it is the purpose
of the present paper to outline the results of such a study in
embryos from 20 mm. to 240 mm. crown-rump length.
67
68 GEORGE L. STREETER
It has long been known that in elasmobranchs the endolym-
phatic appendage opens directly on the surface of the body and
that the surrounding sea-water can thereby pass directly through -
the endolymphatic duct to the cavities of thelabyrinth. The
arrangement that we have referred to as existing in the tadpole,
suggests that we have there quite a different source of access for
the endolymph. At any rate, it is evident that the contact ex-
isting between the endolymphatic sae and the membranous roof
of the hind-brain affords favorable structural conditions for an
interchange of substances between the cerebro-spinal fluid and
the endolymph, either by diffusion or by a secretory activity
of the separating epithelial membranes. The endolymphatic ap-
pendage also in the human embryo serves as an absorption-appa-
ratus or one for regulating the endolymph, that is, if we may
judge from its structural and topographical characteristics. The
condition, however, in human embryos becomes somewhat more
complicated than that in the tadpole in that here the sac is sepa-
rated very early from the chorioidal membrane by the develop-
ment of the dura mater and the intervening arachnoid-pial
membrane. Instead of attaching itself to the membranous roof
of the hind-brain, the sac projects against one of the large veins
of the dura mater. Furthermore, it does not apply itself directly
against the vein wall, but is separated from it by an intervening
capillary plexus, which in turn drains into the vein. As far as
the writer knows the character and connections of this endo-
lymphatic capillary plexus is described here for the first time.
As to its functional significance we must for the present limit
ourselves to the above suggestion and in the following paper
attention will be directed only to its morphology as seen in the
typical stages of its development.
MATERIAL AND METHODS
The specimens which were examined microscopically in con-
nection with this study consist of a group of human embryos,
measuring from 21 mm. to 240 mm. (crown-rump) long, that is,
from about the eighth to the twenty-eighth week of fetal life.
They all belong to the Collection of the Department of Embry-
VASCULAR DRAINAGE OF ENDOLYMPHATIC SAC 69
ology of the Carnegie Institution of Washington. The speci-
mens in most cases had been injected with India ink through the
umbilical vein and had been prepared in serial sections. In
some cases after injection and fixation they were dissected so
as to make total preparations which were rendered transparent
in wintergreen oil and were examined under the binocular micro-
scope. For purpose of topographical determinations, profile re-
constructions were made of several of the embryos that had been
cut serially and in some instances the structures were modelled
after the Born wax-plate method. ‘These will be specified under
their separate descriptions. Although other embryos were ex-
amined the following list includes those that were chosen as best
representing the stages of growth of the endolymphatic sac and
its blood-vessels.
TABLE 1.
: ae CROWN-RUMP THICKNESS AND : 4 Z
EMBRYO NO. LENGTH DIRECTION OF SECTION VASCULAR INJECTION
460 21 mm. 40 » trans. India ink. Wax-plate reconstruction
632 24 mm. 100 uw sagit. India ink. Profile reconstruction
449 | 34mm. 100 pw sagit. | India ink. Serial examination
96 50 mm. | 100 » sagit. 0 Profile reconstruction
448 52 mm. 100 uw sagit. India ink. Serial examination
458 54 mm. 0 India ink. Cleared specimen
lee side
50 » trans. India ink. Profile reconstruction
1018 130 mm. Re : ae SOI I nate : ae aoe
Right side
{ 0 India ink. Cleared specimen
1131 240 mm. 100 » trans. 0 Serial examination
HISTORICAL
In the opinion of the earlier embryologists the endolymphatic
appendage represents the last portion of the ear vesicle that is
attached to the skin, and which becomes drawn out into a stalk-
like elongation as the vesicle recedes from the surface. They
further pointed out that it corresponds to the narrow tube found —
in Selachians that passes dorsally through the cartilagenous skull
to reach the surface of the head where it opens and thereby con-
stitutes a canal that leads from the outside directly to the laby-
70 GEORGE L. STREETER
rinth. In this instance the ear vesicle remains attached to the
skin throughout the whole period of its development. In
other vertebrates it persists only as an embryological remnant of
varying size that terminates as a blind sac under the dura mater
and is apparently of no further use (Balfour ’81, Hoffman ’90,
Hertwig ’98).
This was the prevailing view regarding the endolymphatic
appendage until results that conflicted with it were reported by
Poli ’97 and Netto ’98. These investigators found that in rep-
tiles and amphibians it is the lateral surface of the ear vesicle that
is last to be detached from the skin, at a place clearly remote
from the dorsal tip that gives origin to the endolymphatic duct.
It was also found that in some cases the endolymphatic append-
age does not make its appearance until after the detachment
from the ectoderm is completed. Keibel ’99 was strongly influ-
enced by the condition existing in the embryo of the chick,
where the separation of the otic vesicle from the ectoderm occurs
relatively late and in fact the last point of attachment does occur
at the dorsal tip of the endolymphatic appendage, and he there-
fore supported the original view of Balfour ’81. He quite cor-
rectly defends the opinion that the tube in Selachians connect-
ing the inner ear with the ectoderm is the same as the endo-
lymphatic duct of the vertebrates. The conditions found in
amphibians by Netto ’98, where the endolymphatic duct does
not develop until a considerable time after the complete detach-
ment of the ear vesicle, he explains as a shifting in the time of
occurrence of the ontogenetic as compared with the phylogenetic
processes.
Subsequently the origin of the endolymphatic sac was care-
fully reviewed by Krause ’01 who had an abundance of material
for a comparative anatomical study. He showed that in reptiles
the point of separation of the ear vesicle from the ectoderm has
nothing to do with the dorsal pointed end of the vesicle from
which the endolymphatic duct arises. While in birds, as de-
scribed by Keibel ’99 and others, it corresponds exactly to the
tip of the endolymphatic duct. In mammals it also corresponds
approximately to the tip of the endolymphatic duct, but here the
VASCULAR DRAINAGE OF ENDOLYMPHATIC SAC 71
duct does not form until after or just at the completion of the
detachment of the ear vesicle. In other words the separation
point of the ear vesicle is a variable one and is not to be con-
fused with the question of the homology of the endolymphatic
duct. As regards the latter, Krause concludes that the endo-
lymphatie duct of higher vertebrates is completely homologous
with the canal that connects the labyrinth in Selachians with
the surface of the head.
This, in brief, is the present status of our information regard-
ing the endolymphatic duct in its general embryological aspects.
As to its histology and blood supply we are primarily indebted
to Boettcher 69. This investigator made razor-serial sections of
the endolymphatic appendage of the adult cat and new-born
babe. He, first of all, established the fact that it does not de-
generate in mammals as was thought by contemporary investiga-
tors, but develops further and persists through life as an epithelial
canal that connects with the two vestibular sacs, and forms an
important part of the labyrinth. The terminal part spreads out
(new-born babe) into a flattened sac 0.6 mm. wide, and is em-
bedded in the connective tissue of the dura. This sac he de-
seribes as made up of cuboidal pavement epithelium, closely
under which, and sometimes resting directly against it, are found
capillary loops filled with red blood cells. The walls of the sac
are somewhat irregular, due to the presence of small epithelial
pockets which project outward into the periosteum or bone, and
also papilla-like processes or folds which extend into the lumen
of the sac. Both varieties are provided with capillary vessels.
The capillaries are described as losing themselves in the perios-
teum. In another place he describes the small vessels of the ves-
tibular aqueduct at its bony exit as uniting to form a common
stem that empties into the inferior petrosal sinus. These signifi-
cant observations of Boettcher have received scant attention
from subsequent writers and do not seem to have resulted in
further investigation of these interesting conditions.
Hasse ’73 to whom we owe the generally accepted terms
‘endolymphatic duct’ and ‘endolymphatic sac,’ and who con-
tributed many observations on the anatomy of the labyrinth,
de GEORGE L. STREETER
speaks of the endolymphatic appendage as a tube extending from
the labyrinth to the cranial cavity where it either ends blindly
as an ‘epicerebral lymph space’ or opens into the general epicere-
bral lymph space (p. 768). Elsewhere (p. 792) he describes a
small funnel-shape flaring process of the endolymphatic sae that
penetrates through a small opening in the dura and there fuses
with the arachnoid, thus establishing a communication between
the ‘cavum endolymphaticum’ and the ‘cavum epicerebrale.’
The function of the endolymphatic appendage, according to
Hasse, is threefold: 1, the sac, during embryonal life, is an epi-
thelial secretory organ that furnishes the endolymph; 2, in the
adult, it is either a closed sac that secures new materials for the
endolymph by endosmosis from the epicerebral spaces, or it is
an open sac through which the epicerebral fluid flows directly
into the chambers. of the labyrinth; 3, the endolymphatic sac is a
reservoir for endolymph which serves as an expansion tank that
relieves the pressure when it becomes too great in the labyrinth.
The investigators who have studied the blood supply of the
labyrinth do not seem to have directed much attention to the
vascularization of the endolymphatic appendage. They have
done little more than to confirm the observation of Cotugno,
made a century and one half ago, that a vein draining the vesti-
bule and the canals accompanies the endolymphatic duct and
empties into one of the dural sinuses. The most careful descrip-
tion is that of Siebenmann ’94 who showed, as others had done
for the aquaeductus cochleae, that the veins of the vestibular
aqueduct (endolymphatic appendage) though originally accom-
panying the duct, become separated later in their own bony
canal, which he designated as the ‘canalis accessorius aquae-
ductus vestibuli.’, Eichler’92 who studied the blood-vessels of the
human labyrinth confined his attention to the cochlea.
Shambaugh ’03 describes the endolymphatic duct as incased by —
capillaries which are supplied by an arteriole coming usually from
the posterior vestibular artery, and are drained by a vein that
empties into the transverse vestibular vein. Where the endo-
lymphatic sac was preserved it was found to be drained by a
small dural vein.
VASCULAR DRAINAGE OF ENDOLYMPHATIC SAC (iE:
ENDOLYMPHATIC APPENDAGE DURING FIRST TWO MONTHS
The features with which we are chiefly concerned in the pres-
ent paper do not become established until toward the end of the
second month (embryos over 30 mm. long). A review, however,
will be briefly made of the form and relations of the endolymphatic
appendage prior to that time. For a more detailed description
PLEXUS ANT.
PLEXUS
SAGITTALIS
SIN. TRANSVERSUS
b>
SAC-
ENDOL.
V. CEREBRAL. INF. PLEXUS
POST.
SIN, PETROS. SUP.
x
V. OPTHAL. oh \ Vv. JUG.
NT.
SIN. CAVERNOSUS
Fig. 1 Profile reconstruction showing the topography of the membranous
labyrinth and the endolymphatic appendage in a human embryo 24 mm. long
(No. 632, Carnegie Collection). The principal head veins are shown in solid
black. Enlarged about 4 diameters.
with illustrations the reader is referred to a paper previously
published on the development of the membranous labyrinth
(Streeter 06) and to a recent paper on the dural sinuses in which
special attention is given to the topography of the labyrinth at
its different stages (Streeter 715).
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, No. 1
74 GEORGE L. STREETER
In embryos 4 mm. long the ear vesicle consists of a simple
slightly elongated spherical sac that lies in the space between the
primary head vein and the lateral wall of the hind-brain. At its
dorsal end can be recognized a rounded pouch-like projection
which is quite distinctly marked off from the rest of the vesicle.
This is the early endolymphatic appendage. It is in relation
both with the brain wall and the skin, but is separated from
them by a scant amount of mesenchyme, in which can be seen
minute blood-vessels that communicate with the middle and pos-
terior dural plexuses. The appendage points toward the rhombic
lip, but does not quite reach its dorsal margin.
In its subsequent growth the endolymphatic appendage rap-
idly becomes more clearly differentiated from the remainder of
the labyrinth. It takes on a slender tubular form, whereas the
vestibular part of the labyrinth expends into a voluminous tri-
angular pouch. The tubular character of the endolymphatic
appendage is pronounced in embryos from 9 mm. to 14 mm.
long. By its elongation it passes over the rhombic lip and in
14 mm. embryos we find the tip of it overlapping the ventro-
lateral part of the thin chorioidal roof of the fourth ventricle.
It, however, does not lie in direct contact with this membrane
as is the case in tadpole larvae, but is always separated by a
thin layer of the surrounding mesenchyme.
At about the time of the closing-off of the semicircular canals
(embryos 15 mm. long) the simple tubular form of the endolym-
phatic appendage is gradually modified by the expansion of its
distal half into a flattened fusiform sac, which from then on is
recognized as the endolymphatic sac as distinguished from the
remaining proximal part, the endolymphatic duct, that connects
it with the rest of the labyrinth. The endolymphatic sac lies
lateral and caudal to that part of the chorioidal membrane that
is to form the lateral recess of the fourth ventricle. It lies close
against it, but is always separated from it by the tissue that is to
form the arachnoid and dural membranes.
Simultaneously with the formation of the semicircular canals
and the differentiation of the endolymphatic sac there occurs an
alteration in the large dural veins in this neighborhood that plays
.
VASCULAR DRAINAGE OF ENDOLYMPHATIC SAC ao
an important part inits topography. This consists in the replace-
ment of the primary head vein by a more dorsally situated longi-
tudinal channel. The middle dural plexus instead of draining, as
SIN. RECTUS
SIN. SAGITTALIS SUP.
PEEXUS> ANit.
/ SIN. TRANS.
'V. CEREBRAL. INF. - SAC.
ENDOL.
SIN. PETROS. SUP.
EMISSAR.
MAST,
V. OPTHAL.
SIN. CAVERNOSUS FORAMEN
JUGULARE
SIN. PETROS. INF.
Fig. 2. Profile reconstruction showing the topography of the membranous
labyrinth and endolymphatic appendage in a human fetus 50 mm. long (No.
96, Carnegie Collection). The endolymphatic sac is partly covered by the trans-
verse sinus, which with the other head veins is shown in solid black. Enlarged
about 4 diameters.
formerly, into the primary head vein drains caudalward into the
posterior dural plexus. Soon afterward the anterior dural plexus,
in a similar manner, changes its direction of drainage and instead
of continuing to drain into the cephalic end of the primary head
76 GEORGE L. STREETER
vein, it unites with the middle dural plexus and they both drain
into the posterior dural plexus and through it into the internal
jugular vein. Due to these alterations in the drainage of the
anterior and middle dural plexuses the greater part of the pri-
mary head vein disappears and we find it replaced by the more
dorsally situated channel that is to become the transverse sinus.
This channel forms in a groove in the dorsal margin of the otic
capsule. ‘Topographically it passes longitudinally in the space
between the two vertical canals and the endolymphatic sac. The
general relation of these structures is shown in figures 1 and 2
which are reproduced from the paper previously referred to
(Streeter 715). The canals are separated from the sinus by their
cartilagenous envelope. The endolymphatic sac, however, like
the transverse sinus itself does not become encased by cartilage
and lies against the median wall of the latter, separated from it
only by a small amount of loose embryonic connective tissue in
which both are embedded. This close relation which becomes
established between the endolymphatic sac and the transverse
sinus in 18 mm. embryos, continues as a permanent condition.
At first (fig. 1) when the endolymphatic sac has a vertical posi-
tion, it completely overlaps the median surface of the sinus.
Subsequently as the cranium enlarges, this part of its wall is
crowded outward and downward into a more horizontal position
and partakes in the formation of the floor of the posterior cerebral
fossa. We then find the endolymphatic sac resting on the dorsal
surface of the sinus and furthermore the sinus becomes relatively
larger than the sac and is then only partly overlapped by the
latter.
Though closely related to the chorioidal membrane of the lat-
eral recess, the endolymphatic sac becomes more and more clearly
separated from it as the dural and arachnoidal tissues become
differentiated. On the other hand, though resting against the
transverse sinus, there is a scant amount of loose embryonic
connective tissue separating the two. Running through the
meshes of this connective tissue can be seen blood capillaries
that form a plexus which empties into the transverse sinus. This
plexus anastomoses with the vessels of the labyrinth by com-
VASCULAR DRAINAGE OF ENDOLYMPHATIC SAC 77
munications along the endolymphatic appendage. It also an-
astomoses with the posterior dural plexus. These blood-vessels
and their communications can be recognized in embryos 20 mm.
Pix. Chor.
Saccus endolymph.
Cerebellum
Sinus
petros. Sup.
Cartilage
Utriculus
Tympanum
|
Ductus cochlearis
Fig. 3 Sagittal section through the ear region of a human fetus 52 mm.
long (No. 448, Carnegie Collection). The blood vessels are injected with India
ink and are represented in solid black. The endolymphatic sac is outlined as a
clear space and surrounding it can be seen its dense capillary plexus and the
manner in which this drains into the transverse sinus. Enlarged about 10
diameters.
long, but subsequent to that they rapidly increase in size and
importance, and in embryos 50 mm. long obtain a characteristic
appearance which we shall now proceed to describe.
78 GEORGE L. STREETER
ENDOLYMPHATIC APPENDAGE DURING THE THIRD MONTH
The topography and vascular drainage of the endolymphatic
sac in embryos about 50 mm. (crown-rump) long are shown in
figures 2,3 and 4. In figure 2 can be seen the general posture of
the labyrinth and the relation of its component parts to the dural
sinuses. This figure is drawn from a profile reconstruction of the
labyrinth, dural veins and central nervous system in an embryo
50 mm. long (No. 96, Carnegie Collection). The reconstruction
was prepared by projecting the serial sections on transparent
papers which were then superimposed and all traced on one sheet.
It will be noted that the endolymphatic sac passes upward so
that its dorsal one-third rests against the median surface of the
transverse sinus, opposite the chorioidal roof of the ventricle of
the hind-brain. It does not project above the sinus as in the
younger stage shown in figure 1.
A section through this region is shown in the accompanying
figure 3. This is a portion of a sagittal section through a human
enbryo 52 mm. long (No. 448, Carnegie Collection). Before the
embryo was prepared in serial sections its vascular system was
injected with India ink through the umbilical vein. This in-
jection mass is shown in the drawing in solid black. The section
passes antero-posteriorly through the lateral part of the cere-
-bellum, and includes a portion of the ventricle with the chorioidal
villi projecting into it. At the base of the villi there is a col-
lection of the injection mass which apparently is an extravasa-
tion. This is separated from the endolymphatic sac and its
vessels by the dura which is already fairly well outlined, though
it is not represented in the drawing.
The feature to which particular attention should be given is
the capillary plexus surrounding the endolymphatic sac. Its
general character is indicated, and it can also be seen that it
drains by several outlets into the transverse sinus. On following
it through the sections of the series it is found that it completely
envelops the endolymphatic sac and duct. It can be traced cen-
trally within the cartilage as a finely meshed tubular covering of
the duct extending to the region where the duct arises from the
Plexus endolymphaticus
Sinus dur. matr.
transversus
Sac. endolymph.
(cut off)
Fig. 4 Camera lucida drawing showing the endolymphatic plexus and its com-
munications in a human fetus 54 mm. long (No. 458, Carnegie Collection).
The blood vessels were injected with India ink and the whole rendered transparent
with wintergreen oil. A portion of the plexus was removed to show the contained
endolymphatic sac, and part of the sac was also removed in order to show the
drainage of the plexus. The vessels of the rest of the labyrinth are only filled
in far enough to show their communication with the endolymphatic plexus. En-
larged about 17 diameters.
79
SO GEORGE L. STREETER
utricle and saccule. At this point it anastomoses with the
vessels of the vestibular part of the labyrinth. The vessels be-
longing to the vestibule and canals are more sparse; a portion of
the cochlea, however, seems equally as well provided as the
endolymphatic appendage. It is to be remembered that we are
dealing with an injected embryo and the meshes of this plexus
are doubtless distended, so that the picture we obtain shows
them more prominently than would be the case in uninjected
material. The topography and communications of the endo-
lymphatic blood plexus are shown more completely in figure 4.
This is an outline drawing of the labyrinth and its blood-vessels
in a human embryo 54 mm. long (No. 458, Carnegie Collection).
The blood-vessels were injected with India ink and after fixa-
tion the head of the embryo was dissected and the desired portions
of it were dehydrated and cleared in wintergreen oil. Figure 4
shows the right labyrinth as seen in such a specimen. The in-
jected vessels in the region of the vestibulo-cochlear junction are
shown in solid black and also their continuation into the endo-
lymphatic plexus inclosing the endolymphatic duct. The con-
tinuation of the plexus toward the lateral sinus is shown in stipple.
In the region of the endolymphatic sae a part of the plexus is
represented as cut away. The greater part of the sac is also cut
away in order to expose more completely the outer leaf of the
plexus, that intervenes between the endolymphatic sac and the
sinus, and its characteristic communications with the sinus.
The sac is quite flat and when it is intact it corresponds in contour
to that portion of the plexus that has been left. The reader will
be able to form a picture of the whole apparatus by imagining
the rest of the sac back in place and covered in by the inner leaf
of the plexus.
From an examination of figures 2, 3 and 4, we see, therefore,
that in embryos about 50 mm. long the endolymphatic append-
age consists of a narrow duct that widens out into a broad flat-
tened sac that lies between the chorioidal membrane of the lateral
recess and the transverse sinus. It is separated from the former
by the dura and is separated from the latter by the endolym-
phatic plexus. This plexus consists of thin walled capillaries
VASCULAR DRAINAGE OF ENDOLYMPHATIC SAC 81
which everywhere inclose the duct and sac. In the distended
state, as in Injected specimens, they virtually constitute a sur-
rounding sheet of blood inclosed in endothelium, since the open-
ings in the mesh are, as a rule, narrower than the blood channels
themselves. There is some tendency at this time, and it becomes
more marked later on, to the formation of principal channels
in this plexus. The plexus anastomoses centrally with the other
blood vessels of the labyrinth. Distally it drains by several
openings into the transverse sinus. In addition it anastomoses
with a coarser plexus of veins that les between the dura and the
cartilaginous skull in the neighborhood of the sinus. In this
same region there are some small arteries of the dura mater that
seem to communicate by minute branches with the endolymphatic
plexus. There were very few of these and their arterial nature
could not be determined with certainty.
ENDOLYMPHATIC APPENDAGE AT END OF FOURTH MONTH
The endolymphatic plexus gradually changes its character as
we advance to older fetuses. Instead of a fairly uniform mesh-
work that envelops evenly all parts of the appendage, part of it
takes the form of larger and simpler channels that become more
or less separated from the remainder of the plexus while the latter
continues as a fine meshwork closely applied to the surface of the
appendage. The finer plexus drains into the larger channels
which in turn drain into the transverse sinus.
In order to determine the topography and vascularization of
the endolymphatic appendage at this period, a well hardened
fetus, 130 mm. crown-rump length, was selected in which the
blood-vessels had been injected through the umbilical vein with
India ink (No. 1018, Carnegie Collection). The part of the
skull on each side containing the labyrinth was removed, care
being taken to preserve the dura. The specimen from the right
side was dehydrated and cleared in wintergreen oil and studied
as a transparent specimen. The left one was decalcified and
cut in serial sections and a profile reconstruction was made of the
labyrinth and larger vessels. By combining the reconstruction
with the study of the transparent specimen it was possible to
82 GEORGE L. STREETER
ascertain very definitely the relations of the structure with which
we are concerned.
A camera lucida drawing of the endolymphatic plexus and its
connecting vessels is shown in figure 5, as they are seen in the
cleared specimen mentioned above. In the same drawing is
introduced a profile reconstruction of the endolymphatic append-
age prepared from serial sections of the other labyrinth. From
an examination of this figure it will be seen that the endolym-
phatic appendage is divisible into a duct and a sac. The duct is
further divisible into a proximal flaring portion and a narrow por-
tion that connects this with the sac. It can be seen in sections
that the proximal flaring portion possesses thin walls that show
a tendency to be thrown in folds. The endolymphatic sae con-
sists of a flattened blind pouch with a rounded contour. Micro-
scopic examination shows that its walls consist of a single layer
of cuboidal epithelium which is uniform throughout the sac
except at its distal extremity where it narrows into a tubular
process whose epithelium retains the embryonic character. In_
its general topography the endolymphatic sac maintains its former
relations and its distal part is found overlapping the dorso-
median wall of the transverse sinus.
On examining the endolymphatic plexus in figure 5 it will be
seen that it has undergone certain changes as compared with the
younger stage shown in figure 4. A vascular plexus still envelops
the appendage everywhere. This consists of a thin walled endo-
thelial network whose meshes vary in size and pattern and lie
closely against the epithelial wall of the appendage. In the
drawing only the more prominent loops are shown; besides these
there are everywhere small anastomosing capillaries that inter-
vene between them. The network as a whole is richer over the
sac and over the proximal flaring portion of the duct and is more
secant over the narrow portion of the duct. Running through
the plexus there are a few larger channels that have been sepa-
rated out. These form main drainage channels that become
partially detached from the general plexus, though the latter con-
tinues to anastomose with them at frequent intervals. One of
these is the so-called ‘vena aquaeductus vestibuli.’ This forms
Sinus dur. matr.
transversus
Saccus
endolymph:
Plexus
endolyn.ph.
V.aquaeduct. vestib.,
4
Fig. 5 Profile reconstruction of the endolymphatic appendage in a human
fetus of 130 mm. crown-rump length (No. 1018, Carnegie Collection). Com-
bined with it is a camera lucida drawing of the endolymphatic plexus, with its
connections, made from the other labyrinth of the same specimen which had been
cleared in oil. The numerals indicate communications of the endolymphatic
plexus with other veins: 1 and 2, dural veins; 3, vein draining plexus on dorsal
surface of utricle; 4, from plexus on median surface of utricle; 5, from posterior
ampulla and adjacent part of utricle and saccule; 6, veins from median surface
of saccule and cochlea. Enlarged 1734 diameters.
83
84 GEORGE L. STREETER
along the borders of the endolymphatic duct. It may be regarded
as having a group of tributaries from the remainder of the laby-
rinth. These are numerically indicated in figure 5 as follows:
‘3’ is a vein draining the dorsal surface of the utricle from where
it curves around at the base of the crus commune to join the endo-
lymphatie system; ‘4’ drains the plexus belonging to the medial
wall of the utricle; ‘5’ drains the plexus of the posterior ampulla
and the adjacent posterior surfaces of the utricle and saccule;
‘6’ indicates a group of anastomosing vessels from the median
wall of the saccule through which it also communicates with the
cochlear system. Opposite the narrow part of the endolymphatic
duct these various channels are assembled into two vessels of
which the one along the posterior margin of the duct is the prin-
cipal one, and the one that persists as the v. aquaeductus ves-
tibuli. Tracing it upward we find it receiving large tributaries
from the plexus of the endolymphatic sac and at the same time
enlarging into a wide channel along the caudal margin of the sac.
In addition to the tributaries from the endolymphatic plexus it
receives several tributaries from the plexus underlying the sur-
rounding dura, such as ‘1’ in figure 5. It empties into the trans-
verse sinus by one or two openings in conjunction with adjacent
dural veins.
In describing this plexus and the vena aquaeductus vestibuli
it is simpler to think of the blood stream as flowing all in one
direction, that is, toward the transverse sinus. In reality it is
‘quite possible that, due to mechanical’ conditions, the plexus of
the proximal part of the duct drains backward into the vessels of
the rest of the labyrinth and in common with them through the
veins of the cochlear aquaeduct. The natural drainage of the
sac, however, is toward the transverse sinus. Under these con-
ditions the narrow part of the duct is a ‘divide’ from which the
blood flows in both directions, and through the same v. aquae-
ductus vestibuli.
VASCULAR DRAINAGE OF ENDOLYMPHATIC SAC 85
ENDOLYMPHATIC APPENDAGE IN EMBRYOS DURING SEVENTH
MONTH ;
To represent late fetal conditions of the endolymphatic sac, a
fetus was selected weighing 948 gms., in formalin, and measuring
240 mm. crown-rump length (No. 1131, Carnegie Collection).
The head of the fetus was removed and divided in bilateral halves.
On one side a dissection was made exposing the endolymphatic
sac which was done by carefully reflecting the dura. The form
of the sac and its relation to the transverse sinus was found
to be essentially the same as that shown in figure 5, so a draw-
ing of it will not be repeated. On the other side of the speci-
men the dura was raised in one mass together with all the soft
tissues between it and the bone; this included the endolymphatic
sac, the periosteal vessels and part of the terminal portion of the
transverse sinus. This was then embedded and prepared in
serial sections, in a plane longitudinal to the duct and transverse
to the sac. A simplified drawing of one of these sections is
shown in figure 6.
In the drawing the endolymphatic sac is shown in heavy black
stipple. It consists of a flattened sac embedded in the connec-
tive tissue that forms the substratum of the dura. Its distal
portion overlaps the dorso-median surface of the sinus as in the
previous stage. One new feature is found that was not present
in the younger stages and that is that the epithelial wall of the
sac projects irregularly in small longitudinal folds apparently .
thereby offering greater surface area. A characteristic fold of
this kind is cut through in the section shown in figure 6. Such
a fold gives the appearance of a double sac but tracing it through
the sections shows that it is only an out-pocket whose lumen
communicates with that of the main sac.
The dura mater merges gradually into a somewhat loose sub-
stratum of connective tissue that attaches it to the bony skull.
This is schematically represented in the drawing and the ragged-
ness of the bony surface of the dura is due to the difficulty in
detaching the dura from the bone and also in part to the irregu-
larity of the bone. In the meshes of the connective tissue of the
Art.
Facies arachnoidalis
durae matris
x*——— Sinus dur. matr.
transversus
Plexus endolymph,
Ei selehnps® aaa
i
Fig. 6 Section through the endolymphatic sac showing its relation to the dura
and blood vessels in a human fetus measuring 240 mm. crown-rump length
(No. 1131, Carnegie Collection). Endolymphatic sac is stippled dark. Blood
vessels are shown in plain white. The endolymphatic plexus is more dense on
the median or upper surface of the sac; on the lateral or lower surface the plexus
is partly replaced by the main channel through which it drains into the trans-
verse sinus. ‘V.d.,’ a large dural vein; ‘Art.,’ artery. The arachnoidal surface
of the dura is intact, but the bony surface was torn in the removal of the specimen
from the bone. Enlarged 15 diameters.
; 86
VASCULAR DRAINAGE OF ENDOLYMPHATIC SAC 87
dura are found numerous blood vessels which are shown in the
drawing as white spaces. The largest of these is transverse sinus.
A portion of its wall is missing having been injured in the removal
of the dura from the bone. Around the endolymphatic sac is a
thick plexus of thin walled veins which apparently is the same as
the endolymphatic plexus which we have studied in the younger
specimens. At the caudo-lateral surface of the sac they open
into a large channel which in turn drains into the transverse
sinus. This is the channel that follows along the endolymphatic
duct and is known as the vena aquaeductus vestibuli. Other
dural veins anastomose with it, but its primary communication
is with the venous plexus of the endolymphatic sac. As this
specimen did not include the intraosseus portion of the endo-
lymphatic appendage the proximal connections of these veins
could not be studied.
SUMMARY
From the above study of the endolymphatic appendage in
human embryos the principal features in its development, to-
pography and vascularization may be summarized as follows:
The endolymphatic appendage makes its appearance at the
dorsal tip of the otic vesicle in embryos about 4 mm. long, where-
upon it rapidly enlarges, forming an elongated tube that extends
upward toward the chorioidal roof of the hind-brain. As it
does this it becomes differentiated into two subdivisions: the distal
half spreads out forming a broad flattened blind pouch, the saccus
endolymphaticus; the proximal half, the ductus endolymphaticus,
forms an elongated narrow tube connecting the distal part with
the remainder of the labyrinth. The main features in this dif-
ferentiation are completed in embryos 30 mm. long and at the
same time the topographical relations of the appendage have
assumed practically the adult conditions.
A prominent factor in the topography of the endolymphatic sac
is its relation to the transverse sinus. The characteristic flat-
tened form of the sac and the establishment of the sinus are to be
seen at about the same time. From then on the sac always lies
with its flat surface applied against the median wall of the sinus,
SS GEORGE L. STREETER
or the dorso-median wall as the base of the skull becomes more
flattened out. The sac does not become incorporated with the
rest of the labyrinth in the cartilaginous capsule, but like the
sinus lies exposed in the floor of the posterior cerebral fossa and
is covered in only by the dura mater.
Throughout the greater part of foetal life the endolymphatic
appendage is ensheathed by a vascular plexus, the plexus endo-
lymphaticus, which anastomoses on the one hand with the vessels
of the rest of the labyrinth and on the other hand with the
transverse sinus into which it drains through several openings.
This plexus makes its appearance at about the time of the dif-
ferentiation of the appendage into its adult subdivisions of duct
and sac. It can be plainly recognized in embryos 30 mm. long.
In embryos 50 mm. long, it is well developed and at that time
it forms a closely meshed web completely investing the append-
age, whereby the latter is virtually inclosed in a sheet of blood
from which it is separated only by the endothelium of the blood
spaces.
In the course of its further enlargement and development in
embryos 100 mm. long and over, the endolymphatic plexus be-
comes resolved into a few principal channels connected with
which there remain parts of the original plexus. The plexus
persists notably in the neighborhood of the endolymphatic sac.
One of the most constant channels that are developed through
the endolymphatic plexus is the one forming the so-called vena
aquaeductus vestibuli. This forms along the side of the endo-
lymphatic duct and the posterior margin of the endolymphatic
sac, and it constitutes a direct communication between the vas-
cular plexus surrounding the labyrinth on the one hand, and the
transverse sinus on the other. It may be a single or multiple
channel. Through it is drained the plexus of the endolymphatic
sac and also some of the dural veins of the immediate neigh-
borhood.
VASCULAR DRAINAGE OF ENDOLYMPHATIC SAC SY
REFERENCES CITED
Batrour, F. M. 1881 A treatise on comparative embryology. London, Mce-
Millan, vol. 2, p. 426.
Boerrcuer, A. 1869 Ueber den Aquaeductus vestibuli bei Katzen und Men-
schen. Archiv fiir Anat. u. Physiol., p. 372.
Coruano, D. 1761 De aquaeductibus auris humanae internae anatomica dis-
sertatio. Neapoli. (Quoted from Eichler 792.)
Ercourer, O. 1892 Anatomische Untersuchungen itiber die Wege des Blut-
stromes im menschlichen Ohrlabyrinth. Ixgl. Sachs. Gesell. d. Wiss.,
Bd. 18, Abhandl. Math. Phys. Classe.
Hassp, C. 1873 Ductus endolymphaticus. Anatom. Studien, Hft. 4, p. 792.
Herrwie, O. 1898 Lehrbuch der Entwicklungsgeschichte. 6th Edit. Jena,
Fischer.
Horrman, C. Kk. 1890 Entwicklungsgeschichte der Reptilien. Bronn’s Klas-
sen u. Ordnungen d. Their-reichs. Bd. 6, Abth. 3, p. 2012.
Keiser, F. 1899 Ueber die Entwickelungdes Labyrinthanhanges. Anat. Anz.,
Bd. 16.
Krause, R. 1901 Die Entwickelung des Aquaeductus vestibuli s. Ductus en-
dolymphaticus. Anat. Anz., Bd. 19.
Nerro F. 1898 Die Entwickelung des Gehérorgans beim Axolotl. Dissert.
Berlin.
Pour C. 1897 Zur Entwickelung der Gehérblase bei den Wirbeltieren. Archiv
f. mikr. Anat., Bd. 48.
SHamBauGH G. E. 1903 The distribution of blood vessels in the labyrinth of
the ear of sus scrofa domesticus. Decennial Publications, Univ.
Chicago, vol. 10.
SIEBENMANN, F. 1894 Die Blutgefiisse im Labyrinthe des menschlichen Ohres.
Wiesbaden.
Streeter, G. L. 1906 On the development of the membranous labyrinth and
the acoustie and facial nerves in the human embryo. Amer. Jour.
Anat., vol. 6.
1914 Experimental evidence concerning the determination of posture
of the membranous labyrinth in amphibian embryos. Jour. Exper.
Zool., vol. 16.
1915 The development of the venous sinuses of the dura mater in
the human embryo. Amer. Jour. Anat., vol. 18.
PROBLEMS OF HUMAN DENTITION
PROF. DR. L. BOLK
Director of the Anatomical Institule, University of Amsterdam
TWENTY-EIGHT FIGURES
INTRODUCTION
During the last few years [ have been occupied with ana-
tomical and embryological researches upon the dentition of
mammals and reptiles. The memoirs dealing with the results
of these investigations are published in the Dutch or German
language. My researches on the mammalian dentition were
principally executed on human and other primate material. In
consequence of these investigations, | have arrived at conclu-
sions regarding some fundamental odontological problems, differ-
ing somewhat strongly from those generally accepted. However,
as my conclusions are founded partially on the examination of a
large amount of material and partially on the observation of here-
tofore unknown facts and relations, I believe that my points
of view in some respects throw new light upon odontological
problems.
The present paper discusses only some purely morphological
problems of the primate dentition, especially with reference to
the dentition of man. In the next essay I hope to treat of some
embryological phenomena.
T have demonstrated in my ‘‘Odontologische Studien’! that
the ontogenesis of the mammalian teeth shows peculiarities pre-
viously undescribed, the knowledge of which makes their devel-
opmental history in some degree different from the generally
' Odontologische Studien I. Die Ontogenie der Primatenziihne. Jena, Gustav
Fischer, 1913. Odontologische Studien II. Die Morphogenie der Primatenziihne.
Jena, Gustav Fischer, 1914.
91
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, NO. 1
Q? PROF. DR. L. BOLK
accepted scheme. The embryological development of mam-
malian teeth is a process more complicated than the description
of it as given in text-books of anatomy or odontology would
lead one to suppose.
First I shall treat of the manner in which the dentition of man
(and of all other platyrrhine Primates) is a development of the
more primitive form of platyrrhine dentition. Then the prob-
lem will be discussed as to which set of teeth our molars belong
(to the first or milk dentition, or to the second or permanent set
of teeth); and finally I shall give my conception as to the future
changes which will occur in human dentition.
FIRST PROBLEM: THE RELATION BETWEEN THE DENTITION OF
PLATYRRHINE AND CATARRHINE PRIMATES
As is generally known, one of the most striking anatomical
differences between the two groups of Primates (the Platyrrhinae
and the Catarrhinae) is that the monkeys of the New World
possess three premolars and three milk molars in each jaw,
whereas in catarrhine monkeys and likewise in man, there are
only two of each of these teeth. There is some difference be-
tween the two families of American monkeys, the Hapalidae
possessing two molars only, whereas in the Cebidae there are three
of these teeth, as in all other Primates. Therefore the majority
of New World monkeys have a set of teeth whose post-canine
portion possesses one tooth more than the corresponding portion
of the Old World monkeys, or of man.
It is a common view of anatomists and zoologists that the
dentition of the latter evolved from that of the former group by
the loss of one of the premolars. But there is no agreement as to
the premolar which was reduced. Most investigators maintain
that the first premolar of the platyrrhine monkeys is wanting in
the catarrhine group. Therefore the first premolar of the latter
should be considered the homologue of the second premolar of
the former. Other investigators, on the contrary, assert that it
is the third premolar of the Platyrrhinae which is wanting in the
other families of the Primates. I do not agree with either of
these views. My opinion as to the relation between the den-
PROBLEMS OF HUMAN DENTITION 93
tition of the Platyrrhinae and the Catarrhinae is wholly differ-
ent. But there is one point common to both hypotheses worthy
of special reference because it constitutes a weak side of each
of the hypotheses.
The dental formula of the platyrrhine monkeys (save the
Hapalidae) runs as follows:
lm ton CI Wo» 1s,
lee ie Gr ie RS: eee M,. M.. M3.
and that of the ecatarrhine Primates:
in lee. aw Ts:
T,. Iz. C. Pi Pe. Mi. Me. Mz.
In these formulas the elements of the milk dentition are written
in small print and those of the permanent set of teeth in capital
letters. The formulas as above written are not intended to give
expression to any homology between the teeth of the two groups
of primates. In accepting either of the hypotheses mentioned
above one must keep in view the fact that the catarrhine denti-
tion not only arose from the reduction and final loss of a premolar
in the permanent set of teeth, but also of its predecessor in the
milk dentition. It seems to me to be very important that in
none of the recent genera of the New World monkeys is a reduc-
tion of a tooth to be seen either of the first or third milk molar,
or of the first or third .premolar. It is altogether probable that
in the extinct ancestors of the catarrhine Primates such a reduc-
tion really took place. But one looks in vain both in the upper
and lower jaw of this group of Primates for any proof that a
tooth in the premolar region has been lost. In its whole extent
this portion of the dental arch is always regularly constructed, the
teeth standing very closely approximated. Furthermore in catar-
rhine monkeys a diastema, especially between the canine and first
premolar is wanting. These facts surely are not very favorable
to the opinion that the reduction in the number of the premolars
in Primates happened in the common way—in consequence of the
loss of a premolar and its predecessor in the milk dentition.
For, in case such a process really happened, the situation of the
lost tooth would be indieated by a diastema. The objection to
Q4 PROF. DR. L. BOLK
the usually accepted hypothesis is further strengthened by the
consideration that the diminution from the primitive number of
four premolars to three, which happened in the eocene Primates,
‘an be followed step by step in the different well-known genera
of their group of the common ancestors of all recent Primates.
The following objection may also be advanced. If really the
first post-canine tooth in both dentitions should be reduced and
lost, there is very strong ground to expect that in the embryo-
logical evolution of the dentition of man, apes or catarrhine
monkeys, the anlage of this tooth or occasionally the tooth
itself should be found in a rudimentary form and size. With-
out doubt the teeth are reduced and lost during the last phase
of development in all mammals. Although the development of
the human dentition in human embryos has been examined by
a great number of investigators, there has never been found a
single vestige of the anlage of a rudimentary milk molar imme-
diately behind the canine.
T shall not consider in a detailed manner the current hypothe-
ses, being of a wholly different opinion with regard to the rela-
tion between the dentition of American monkeys and that of
Old World monkeys. This opinion may be briefly expressed as
follows: The dentition of the catarrhine Primates (including
man) with its two premolars is derived from an ancestral form
with three premolars in two phases. The first phase was char-
acterized by the reduction and final loss of the third molar of
that ancestor. By this process a form resulted with three pre-
molars and only two molars, just as we actually find in the group
of the recent Hapalidae. The second phase was of an entirely
different nature: during the same, the third milk molar devel-
oped into a permanent tooth, whilst the development of its
successor (the third premolar) was suppressed, in consequence
of which the number of permanent molars increased to three,
as in the primitive form.
By this hypothesis the Hapalidae, with regard to their dentition,
are placed on a higher level in the phylogenetical system than
they occupy in the common systems of Primates. One is accus-
tomed to consider the Marmosets the most primitive recent
PROBLEMS OF HUMAN DENTITION 95
representatives of the primate stem. It was never very clear
to me upon which points in their anatomical structure this
opinion is grounded. It is true that their nails, except those on
the hinder thumbs, are formed like claws. But this peculiarity
is a phenomenon of less value with regard to the problems of
phylogenetical evolution than the indications supplied by the
structure of the dentition. And in comparing the anatomy
of the molars of the other platyrrhine monkeys—the Cebidae—
with those of the Hapalidae, it becomes clear that nearly all
Cebidae show a tendency to attain a developmental stage already
accomplished by the Hapalidae, or the total reduction of the
hindmost molar. In most genera of the New World monkeys
Cebus, Ateles, Chrysothrix, Pithecia, Nyctipithecus—-the third
molar is already reduced in a very large degree, having only a
single root and a very small crown without cusp-differentiation.
Usually this hindmost tooth of the Cebidae is a far more reduced
element of the dentition than the third molar in man. The fact
that the Cebidae approach a structure of their dentition already
acquired by the Hapalidae, is to me a sufficient ground to place
the latter on a higher level of phylogenetical evolution than the
former. It is worthy of mention, that the investigation of
Weber showed the brain of the Marmoset, although a lissen-
cephalous one, to be relatively heavier than even that of man.
Therefore, | consider the dentition of the Hapalidae an inter-
mediate form between that of Cebidae and catarrhine Primates,
notwithstanding the reduced number of their molars. The loss
of the hindmost molar, was the first step which led the plat-
yrrhine ancestor of man to the more progressive dental struc-
ture peculiar to all Old World Primates. We will return later
on to the cause of this reduction.
The primitive number of three molars was regained in conse-
quence of the third milk molar becoming permanent and of the
suppression of the development of the third premolar. This is,
Tl admit, somewhat unusual in the evolution of dentition. But
there are other examples, well-known to us, in which the same
phenomenon took place, and in consequence of which the func-
tional set of teeth became a mixed one, composed partially of
96 PROF. DR. LL. BOK
milk teeth and partially of teeth of the second dentition. The
best known case is that of the Erinaceus. According to the
very exact and ample researches of Leche, the functional set of
teeth of this Insectivora is composed of elements of both den-
titions. The above point of view therefore does not introduce
a wholly new principle in odontology.
Before developing the different arguments upon which my
hypothesis is based, I wish to summarize its esssentials by means
of some dental formulas. In these the symbols of the milk
dentition are printed in small letters, the permanent teeth in
capitals.
Dental formula of the Cebidae:
Nfs doe'@. tile, gs Mle.
Tie te. Gs Pix Pe Ps, Ma.) Ma. Mis.
Dental formula of the Hapalidae:
Pear ee Guei ic spuRaa Men aa
lee Te: Cr lee Ps: Pe: M;,. Mae. [M3]
Dental formula of the Catarrhinae:
Tha Ips Ge Ty. Mise Wie
I. dee GuiPy. Ps 3.) Meo Ma. Ma)
In the last formula, relative also to man, the elements, whose
development is suppressed, are placed in brackets, as is also
done in the second formula. In my hypothesis two principles
are involved, which will be discussed separately, viz., the belong-
ing of our first permanent molar to the socalled milk dentition,
and, secondly, the disappearance of two elements in our perma-
nent set of teeth, the third premolar and the molar originally
hindmost. A further consequence of this hypothesis is that the
three molars of the catarrhine Primates are not homologous with
the three molars of the Platyrrhinae, our second molar corres-
ponding with the first of the latter. I question the nature of our
first permanent molar. My opinion is that this tooth is homolo-
gous with the third milk molar of the more primitive Primates.
Evidently the process by which this tooth became the first
molar of our permanent dentition is composed of two factors, to
PROBLEMS OF HUMAN DENTITION Q7
wit: the loss of its sueceeding tooth (the third premolar of the
lower forms), and secondly, the fact that an originally decidu-
ous tooth became a persisting element. It is clear that these
phenomena stand in a close relation to each other, for a milk
tooth cannot acquire the character of a persisting tooth, so
long as the evolution of its successor is not suppressed. The
two events must have happened simultaneously. As to the
question whether the evolution of the milk tooth or the re-
gression of the permanent element was the leading factor in
this process, I incline to the first of these two possibilities, on
the following ground: I consider that the evolution of the dental
structure of the catarrhine Primates commenced with the reduc-
tion and final loss of the hindmost molar of an ancestor with a
platyrrhine dentition, perhaps in consequence of the shortening
of the jaws. By this reduction the grinding surface of the set of
teeth underwent ashortening. This circumstance (without im-
portance in small animals such as the Hapalidae, who live princi-
pally on soft food or insects) became disadvantageous as the
species grew taller and the nature of the food required a larger
erinding surface. It is very important that the third milk
molar of the platyrrhine monkeys is a larger tooth, with a greater
surface and more cusps than its successor, the third premolar.
Especially in Hapalidae is the difference notable. And so it
was advantageous to the grinding function of the dental arch of
the historically succeeding larger forms of monkeys, that the
third milk molar with its four or five cusps was not replaced by a
tooth with two cusps only. Thus the grinding surface of the
dental arch regained at its anterior end what it had lost in an
earlier period of evolution at its posterior end.
These considerations lead me to the supposition that in the proc-
ess of evolution, the alteration of the character of the third milk
molar occurred first, the loss of the third premolar being a
necessary consequence of it. Because the third premolar was
reduced, the third milk molar became a persisting tooth; but
because the permanence of this milk molar brought a functional
advantage, the evolution of the third premolar was suppressed.
QS PROF. DR. L. BOLK
My hypothesis regarding the origin of the dental formula of
the catarrhine Primates explains in a very simple manner the
otherwise incomprehensible fact that the diminution of the pre-
molars could occur without a gap in the dental arch. And in the
post-canine portion of the set of teeth of the higher Primates a
diastema is never found. That the continuity of the set of teeth
by the above hypothesis was never interrupted, surely does not
tell against the justice of it. The foregoing, however, are mere
theoretical considerations, let us now proceed to some more
practical arguments.
Regarding the embryological evolution of our dentition and
the succession of the eruption of our teeth, I believe our first per-
manent molar was, in an earlier stage of phylogenetical evolu-
tion, a deciduous tooth, belonging to the first or milk dentition.
In reality this tooth appears (in man, as in all other catarrhine
Primates) before the first permanent incisor. And during the
nearly two years between the eruption of our first permanent
molar and that of our permanent first incisor, the structure
of our dentition is identically the same as in young platyrrhine
monkeys. During this period of the platyrrhine phase of our
dentition, there are three molars immediately behind the canine
tooth.
The affinity of our first permanent molar to the set of milk
teeth is more clearly shown the moment the first anlage of this
tooth appears in human embryos. According to the investiga-
tions of Rose, the anlage of milk teeth commences in the ninth
week of embryological development. Immediately afterward the
germ of the second milk molar is produced by the dental lamina,
the latter is prolonged backwards, and the enamel-organ of the
first permanent molar is formed. This happens in the sixteenth
week of embryological development. Therefore, there is no dis-
continuity in the suecession of the first anlage of the enamel-
organs of our deciduous teeth and that of our first permanent
molar. cirea 640 diameters.
4 Portion of renal glomerulus of cat fetus of 8.5 em. Orientation as in
figure 1. Epithelial plates already present. X 640 diameters.
6 Part of labyrinth of placenta of a rabbit of 27 days, showing the endo-
thelium of the fetal capillaries and the succession of thin plates and thicker
nucleated portions of the ectoderm, between capillary and maternal blood stream.
Copied from Duval’s Atlas, Placenta des Rongeurs, figure62. > 470 diameters.
7 Portion of placenta of a guinea pig of the second month, similar to figure
6. Copied from Duval’s Atlas, figure 262. X 300 diameters.
206
INTERRELATIONS OF THE MESONEPHROS, ETC. PLATE 1
JOHN LEWIS BREMER
PLATE 2
EXPLANATION OF FIGURES
5 Portion of placental chorion of human embryo of 29.0 mm. (H. E. C.
No. 389). Above, the chorionic mesoderm; the basal layer of the ectoderm and
the syncytial layer are both interrupted by a fetal capillary, separated from
the maternal blood stream only by an ectodermal plate, pl., which is closely
adherent to the endothelium of the capillary. X 640 diameters.
8 Portion of chorion and labyrinth of placenta of a rat of 13 days (H.
E. C. no. 1930, sect. 143). The same production of epithelial plates separating the
endothelium of the fetal capillaries from the maternal blood stream. The two
streams are recognizable by their blood corpuscles. It will be noticed that the
plates occur against both fetal arteries and veins. The basal layer of fetal
ectoderm has partially disappeared. 250 diameters.
9 Villus of human placenta of 3 months. Note the complete syncytial
layer of the fetal ectoderm, and the basal layer interrupted by a fetal capillary,
over which the syneytium has developed a plate. > 480 diameters.
10and 11 = Villiof human placenta at term. The basal layer of ectoderm is
no longer present. The syncytial layer shows a succession of thick granular
nucleated portions and thin epithelial plates in direct contact with the fetal eapil-
laries. The maternal blood stream surrounds the villi.. & 480 diameters.
12. Model of the blood-vessels and the ectodermal syncytium of a villus
of the human placenta at term. It will be noticed that two small villi have
fused, making a ring formation, around which capillaries pass. One artery and
two veins pass into the villus. In addition to the areas seen in profile where the
ectodermal covering is of plate-like thinness, the blood-vessels are also covered
by plates between x and x, and at y, and z. This, with figure 11, shows the rela-
tive extent of the plates and the thicker svneytium. X 250 diameters.
208
INTERRELATIONS OF THE MESONEPHROS, ETC. PLATE 2
JOHN LEWIS BREMER
THE DEVELOPMENT OF THE LIVER AND PANCREAS
IN AMBLYSTOMA PUNCTATUM
EK. A. BAUMGARTNER
Institute of Anatomy, University of Minnesota
FORTY-SIX FIGURES (FOUR PLATES)
CONTENTS
eealinGreo chu C1 O Tiere sence ere eter nae even pee eee” ee eee a eae, eee ae 211
II. The development of the liver, hepatic ducts and gall-bladder.......... 218
UEC Crates we eae ees es eee caste eras Wee ee a ranhea Sr 213
2. Early development of the liver................ 0c ccc cece e eee eeees 218
3. Position of the organ during development.....................+.- 223
4. Development of the biliary apparatus........................... 226
a. Description of hepatic ducts in the adult..................... 226
b. Development of the ductus choledochus..................... 230
ce. Development of the major hepatic ducts.................... 282
d. Development of the minor hepatic ducts.................0.. 236
e. Development of the gall-bladder and cystie duct............. 288
f. Summary of the development of the biliary apparatus....... 242
III. The development of the pancreas and pancreatic ducts................ 247
MAMIE OPAL teats nag Pour Pica ate akan oes rrake fork bre ea citer ar ore wid a eI Canesten 247
2. Early development of the pancreas and pancreatic ducts......... 250
3. Description of the adult pancreas............... cece cece cence eens 260
AM DISCUSSION paeceee tne ace weet tee eco re eee ee eee ere oe 261
Vis CREME ALS UATE Yost ewan ees ae Scie aie tee crf dates oe 2 eee Bide ore ween 264
Wess oli o payin eve nave srevarersrerieee ctstee tere chs ik sc eeu susie ae teed suatiaie toitialigeey Sheu susie cre 266
I. INTRODUCTION
Comparatively little work has been done on the morphology
of the biliary and pancreatic duct-systems in vertebrates. The
arrangement of these structures has been worked out in the
adult forms of a few species but no attempt has been made to
correlate these scattered observations or to determine what
may be considered the typical arrangement in vertebrates and
the major variations which may occur in the various groups
of the phylum. The development of these systems is also
211
2 E. A. BAUMGARTNER
almost unknown. Although the formation of the anlagen of
the liver and pancreas has been investigated in almost every
group of vertebrates, the later history of duct systems of these
structures has been quite neglected. The two exceptions to
this statement are furnished by the work of Corner (’13) who
investigated the development of the pancreatic ducts of the
pig by means of injection methods and Scammon’s study of the
biliary system of selachians.!
The following study is an attempt to follow in detail the de-
velopment of these duct-systems in the tailed amphibia, and
to point out the embryologic significance of the principal varia-
tions which are encountered in the adult and the mechanical
influences which are, in part at least, responsible for them.
Although we are not as yet in possession of sufficient data to
formulate a statement of the typical vertebrate plan of biliary
and pancreatic duct-systems, it is hoped that this description
of these structures in a representative amphibian may add to
the material upon which such a schema must eventually be based.
The material used for this work consisted of embryos of Am-
blystoma punctatum from 4 mm. to 20 em. in length. These
were sectioned serially in transverse and sagittal planes. Graphic
and wax reconstructions were made of the hepatic ducts, gall-
bladder, liver and pancreas of different embryos and adults.
It is a pleasure to express my thanks to Dr. Richard E. Seam-
mon for his constant interest and helpful criticisms throughout
this study.
A correlation of the embryos employed in this study with
those described in the Normal Plates of Necturus maculosus
by Eyecleshymer and Wilson may be desirable. This correla-
tion is based on a comparison of the digestive system including
liver and pancreas, as well as partially on the external form.
Probably the greatest difference in the development of the
digestive tract between these two forms is in the time of union
of the dorsal and ventral pancreatic anlagen which had taken
' The terms used by Scammon ('13) in describing the ducts of Elasmobranchs
have been used in this paper.
DEVELOPMENT OF LIVER AND PANCREAS 2035
place in most of the 13 mm. Amblystoma embryos which I
have observed, and is described in stage 42 (29 mm.) in the
Normal Plate series of Necturus. Also .the limbs, particularly
the caudal ones, appear comparatively later in Amblystoma.
Such a table, of course, can be only an approximate comparison.
TABLE 1
Correlation of Amblystoma embryos with the Normal-plate series of Necturus
FIGURES EMBRYOS NORMAL-PLATE SERIES
Figure Length in mm. Stage No. Length in mm.
11 eel APSE oer ae 4.5 21 8
Vee ee a OE ae ae 5 22-23 9
Bie OU Cee on re Eee 7 25 12
de Erin CERO RO ECOG 9 28 15
Ss EAC eee Ee 9 29 16
WMeonccsk fe eemeee ens 11 30 ily/
GSS tote acces abs eis 15 31 18
7d) DRIES yee ae Oe en eee 12.5 34 21
Ble Ao ssc lnnis «aie vie ois 13
See re eee ae See eee: 13 38 25
Ar aE NPN ite SMe cres 13.5 39 26
BHSOR AQ ice chest Gactrn 14 42 29
(1 Ce | ee 13.5 43 30
10038, 42: AG. -.2.. <5 15 45 32
DO MAS IR Sse sec tele 20 49 39
II. THE DEVELOPMENT OF THE LIVER, HEPATIC DUCTS AND
GALL-BLADDER
1. Literature
The literature of the development of the great glands of the
digestive tract of Amphibia can be conveniently divided into
two parts covering two fairly distinct periods: first, the work
of the early investigators who determined the position of these
glands in the embryo and their relation to the lower germ layer;
second, the series of contributions beginning with Goette’s
large monograph upon the development of Bombinator (’75)
and dealing mainly with the detailed developmental anatomy
of these organs. :
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, NO. 2
.
214 E. A. BAUMGARTNER
The following table gives a list of the authors, the dates of
their publications and the material upon which their work on
the development of the liver and pancreas was based.
Steinheim (’20) studied older embryos and observed the
attachment to the gut. Rusconi (’26) investigated younger
‘embryos and, as did Reichert (’40) and Vogt (’42), described
the ventral growth of the intestine to form the liver. Remak
(55) and v. Bambecke (’68) differed from the above only in
the number of lobes formed and noted the close relation of
the gall-bladder to the right lobe.
According to Goette (’75), the liver in Bombinator originates
as a ventral outpouching of the foregut posterior to the heart.
This diverticulum becomes separated from the gut by a gradual
eranio-caudal constriction, and the narrow connection which
remains forms the common hepatic duct. The outpouching
then grows by the production of folds or buds from its sides which
form the primary hepatic columns. The lumina remain in these
columns although they may be very small. Goette regarded
the early anastomoses and formation of the net-like hepatic
cylinders as aided by the ingrowth of a capillary network. The
gall-bladder develops as an outpouching of the posterior part
of the primitive hepatic duct caudal to which the ductus chole-
dochus is formed.
Balfour (81) made the statement that there is a single ven-
tral diverticulum from the gut which later develops into two
secondary branches and so forms the liver.
Shore (’91) in his study on the frog found that the liver takes
origin as a ventral lengthening of the gut lumen into the mass
of yolk-cells which lies posterior to the heart. The yolk-cells
lining this lumen are transformed into hepatic cells and this
mass becomes partially separated from the gut. This constric-
tion is aided by the caudal growth of the sinus venosus. Later
there is formed at the expense of the yolk-cells and by cell-divi-
sion a large cell-mass into which the blood-vessels tunnel form-
ing a tubular gland whose columns divide and anastomose pro-
ducing a network interlacing with that of ‘blood-lacunae.’
DEVELOPMENT OF LIVER AND PANCREAS 215
Marshall (’93) gave a brief account of the development of
the liver in the frog in his vertebrate embryology. He described
a caudo-ventral projection from the anterior part of the mesen-
TABLE 2
Table of authors and the forms studied
Baumgartner........... 1914 Amblystoma punctatum
AUTHOR DATE MATERIAL
Steimmhemmienyes.. sss. ss 1820 Rana
PRUSCONM 4320-75 crutenghon cs 1826 Rana
eT Cher treat take ce id a 1840 Rana temporaria
Rana esculenta
IRAISC OMI eeeeeir anaes o. 1854 Salamandra
Wo gies st tet aan iad ae 1842 Alytes obstetricans
emma ke fan sistve seta otic 1855 Rana temporaria
Rana esculenta
Rab OKem een cans 2c5 abe. 1861 “Vertebrates”’
Bamibecke. 2.22... osse os 1868 Pelobates fuscus
(GOGTCCRA Ha iiesteacs 1875 Bombinator igneus
Wiedersheim........... 1875 Salamandra perspicillata:
PSAMOUT: soc. o.e canex eens 1881 “Amphibia”’
SNORE ye hidas cote ter. 1891 | Rana
LE C0) 0) 0. oe 1891 | Salamandra maculata, ete.
Bufo vulgaris, ete.
VES esa s,ce5e ob .40.6 7s ane 1893 Rana
WENIOb. 20 chic ce chee Seal * 1893 ““Amphibia”’
WIG YSSOlRS 3k Calicanels ood 1895 Rana temporaria
Rana esculenta
DOME Ma Sines 4.t.s, dita Sachs 1895 Rana temporaria
UCT OWE ia... scaths sioke stan als 1896 ““Amphibia”’
BTACHEbs, asi eeak ease: 1896 Review
EVEL TINIAN I ee sen ss chore, 2 teens 1897 Rana
WV Otome Sch M i niah cy can oh 1897 Rana temporaria
Triton taeniatus, ete.
Kollman. :.2°.27.2 aps: 1898 ‘“Amph bia”’
(Guamve lea 88%, See5 515 4 1899 Triton cristatus
Choronshitzky......... 1900 Rana temporaria
Salamandra maculosa, etc.
CCUGCE, nse se cates Ss 1900 Alytes obstetricans
Grratre liv ais .chet is MY sees 1902 Triton
JEN YS) an Ge oer rE ns 1902 Review
NMC) oc) eae ee 1903 Review.
SrA ates seers cities = 1906 Alytes obstetricans
Eycleshymer and 1910 Necturus maculosus
Wil somisay 8 Sele ai re
216 E. A. BAUMGARTNER
teron. The anterior wall of this depression is thrown into
folds, blood-vessels penetrate between these structures and
outgrowths from the hypoblast form the hepatic cylinders.
Weysse (95) found in the frog that the liver-anlage is a dorso-
ventral cleft extending into the yolk-mass from the gut lumen.
A caudal extension of this cleft forms the posterior hepatic
duct, while the cranial hepatic duct is formed by a folding of
the anterior wall of the hepatic anlage. The yolk-cells are
transformed into the true hepatic cells and can be early recog-
nized by the deposit of pigment within them.
Hertwig (’96) and Kollman (’98) gave only short descriptions,
stating that in Amphibia the hepatic anlage is a single out-
pouching from the ventral wall of the duodenum.
Hammar (’97) who worked on the development of the frog’s
liver, has named the entodermal cell-mass posterior to the
heart the ‘Leberprominenz.’ Into this extends an early length-
ening cavity which is continuous with the lumen of the gut.
This he termed the ‘Leberbucht.’ By a cranio-caudal con-
striction this hepatic anlage is separated from the gut. The
cell-mass about the fundus of this anteriorly directed sac develops
into trabeculae of the adult organ and the posterior part forms
the ductus choledochus. The gall-bladder is developed very
early as a diverticulum of the ventral wall of the common bile
duct, and by further growth comes to be a pedunculated organ,
consisting of a cystic duct and gall-bladder proper. He re-
garded the origin of the trabeculae as perhaps due partially to
the developing capillary network tunnelling into the hepatic
cell mass as suggested by Shore.
Choronshitzky (’00) showed the anlage of the liver in the sala-
mander in a figure of a sagittal section of a 9 mm. embryo, in
which there is a ventral fold in the wall of the foregut. This
fold is lined with yolk-laden cylindrical cells which posteriorly
pass gradually over into the polygonal yolk-cells which form
a mass projecting into the lumen of the gut. In the anterior
ventral wall of the gut is -a second slight pouch which later
forms the gall-bladder. The two omphalo-mesenteric veins
crowd in on either side of the liver outpouching, thereby aid-
DEVELOPMENT OF LIVER AND PANCREAS 217
ing the constriction of the lateral walls of the gut. These veins
unite anteriorly and form the ductus venosus. The liver-anlage
therefore first grows ventrally and then anteriorly below the
horseshoe-shaped union of the omphalo-mesenteric veins and
the ductus venosus. A similar sagittal section of a later stage
shows the liver at the cranial end of a short ductus hepaticus
which is continuous caudally with the ductus choledochus. From
the ventral wall of the ductus choledochus there is now a very
marked outpouching, the gall-bladder, which is united with the
common duct by a short cystic duct. The primitive liver-an-
lage has thus grown cranialward and become separated from
the gut. Choronskitkzy believes this process to be due to the
growth and differentiation of the gut. The walls of the primi-
tive liver-anlage have folded and these folds later develop into
solid liver-columns. The liver grows around the developing
ductus venosus even to its dorsal surface and in so doing produces
many folds and columns which grow through the ductus venosus
and divide it into sinus-like branches.
Reuter (00) in his studies on the development of the intestine
of the Alytes obstetricans made mention of the early origin of
the liver. This develops from the ‘Anfangsdarm’ division of
the midgut. In later embryos the liver develops very rapidly
and is divided into three lobes.
Gianelli (01 and ’02) described the hepatic anlage in Triton
as developing in two parts, the anterior giving rise to the he-
patic tissue proper and the caudal forming the hepatic duct.
The gall-bladder arises from a mass of cells belonging to the
primitive hepatic outpouching. By the development of the
intestinal folds the hepatic duct becomes attached to the dorsal
side of the gut.
Weber (’03) stated that the observations made on the develop-
ment of the liver in the frog and in Triton differ but little. In
the latter the intimate relation of the anterior end of the hepatic
outpouching and the blood-vessels account for the develop-
ment of this part into the hepatic tissue proper.
Bates (04) in a paper on the histology of the digestive tract of
Amblystoma has described the hepatic and pancreatic ducts.
218 E. A. BAUMGARTNER
He has described a bile-duct which les free in the body-cavity
for a short distance and then enters the pancreas which lies
between the liver and the intestine. Here it is joined by two
hepatic ducts and just as this enters the intestine it is joined
by two other hepatic ducts.
To summarize briefly, the early investigators described the
liver and pancreas as developing at the same time from the
ventral wall of the gut, and also considered that they were parts
or lobes of the same organ. Remak (’55) first noted that the
liver is separate and distinct from the pancreas. Goette first
gave a detailed account of the development of the liver in am-
phibia. Most of the investigators from that time have agreed
that the liver begins as a single ventral outpouching of the gut-
wall caudal to the heart. The question as to the origin of the
gall-bladder, whether from the caudal end of the ductus chole-
dochus or from the wall of the intestine in this region may be,
as Piper (’02) stated, one of interpretation rather than one of
observation. Whether the hepatic cylinders divide and the
blood-capillaries then grow between them, or whether the capil-
laries grow into the solid hepatic anlage so forming hepatic cyl-
inders seems not to have been definitely determined. Shore’s
(91) observations support the latter theory. According to
the observations of Weysse and others the yolk-cells are trans-
formed directly into hepatic cells. Very little has been written
about the development of the hepatic ducts. The common
bile-duct is described as the constricted attachment of the
hepatic anlage, or the posterior end of the hepatic outpouching.
2. Early development of the liver
The liver in Amblystoma first appears in embryos about
4.5 mm. in length, which corresponds roughly to no. 21 of Keibel’s
Normal-plate series. The digestive tract at this stage is quite
simple. The pharyngeal cavity is large and extends anteriorly
to the oral cavity. Caudally it opens widely into the mesenteron
which is composed of a large mass of yolk-cells and extends
backward to the proctodaeum. The yolk-mass extends dorsally
to the notochord and bulges ventrally.
DEVELOPMENT OF LIVER AND PANCREAS 219
Posterior to the anlage of the heart a sagittal section shows
a ventrally and somewhat caudally directed projection of the
gut-lumen, (fig. 1) which extends backward near the dorsal side of
the yolk-mass. The anterior wall of the ventrally directed exten-
sion of the gut-lumen is lined by yolk-laden columnar cells and
its posterior wall is formed by the cells of the large yolk-mass.
This cavity is quite wide transversely and is connected to the
gut-lumen above by a wide cleft.
Fig. 1 Sagittal section of an Amblystoma embryo 4.5 mm. long taken at
about the median plane. X 30. F.g., foregut; He, heart; Li, liver; Y, yolk mass.
Fig. 2 Sagittal section of an Amblystoma embryo 5 mm. long, taken to the
right of the median line. X 30. F.g., foregut; G, caudal extension of gut; He,
heart; Li, liver; Y, yolk mass.
Weysse (’95) has described this cavity in frog as a cleft in the
ventral mass of yolk-cells, and Hammar (’97) has termed it
the ‘Leberbucht.’ From the study of a slightly more advanced
stage Weysse concluded that the caudal and ventral end of this
cleft finally formed a caudal hepatic duct. He correlated this
with the caudal hepatic duct described in the chick. That the
caudal projection does not form a caudal hepatic duct in amphibia
seems clear from a study of the later development. The reason
for this error was probably, as Hammar has pointed out, that
220 E. A. BAUMGARTNER
Weysse did not follow the development beyond a very early
stage.
In an embryo approximately 5 mm. (fig. 2) long the anterior
wall of this early ventro-caudal projecting cavity has become
more prominent. The extension of the gut-lumen into this out-
pouching is a large cone-shaped cavity somewhat flattened in
transection. The columnar epithelial cells lining it are now
found farther caudalward than in the preceding stage.
Fig. 3 Sagittal section of an embryo almost 7 mm. long. X 30. D.chol.,
ductus choledochus; F.g., foregut; G. caudal extension of gut; G.B., gall-bladder;
He, heart; Li, liver; Y, yolk mass.
In a sagittal section of an embryo 7 mm. long there is shown
a more advanced stage of the condition just described. From a
comparison of this stage (fig. 3) with the previous one and the one
following, it will be seen that the hepatic anlage has become more
prominent by a cranio-caudal constriction from the gut. Folds
have begun to form on the outer surface of the liver. The cav-
ity of the hepatic diverticulum is widely connected with that
of the gut. In the ventral wall there is a slight median depres-
sion (GB) which is the earliest indication of the gall-bladder.
This depression is at the caudal end of the liver-anlage in the
region where the primitive ductus choledochus is forming.
DEVELOPMENT OF LIVER AND PANCREAS 22
The liver of another embryo 7 mm. long appears as an anterior
and ventral outpouching of the gut. Figure 36 is of a plastic
reconstruction of this region of the archenteron. That the con-
striction from the gut has proceeded caudally will be apparent
by comparison with earlier and later stages. The cavity pro-
jecting into the liver-anlage from the lumen of the gut is now
much longer, and there are indications of further projections
from it on the right side as the lumina of ducts.
Choronshitzky noted this transverse extension of the lumen
in the hepatic anlage of the salamander but did not follow its
further history. At the posterior end in the median ventral
wall is a marked outpouching which is the gall-bladder (GB,
fig. 36). The opening of this outpouching into the gut is still
very wide laterally and shows no differentiation into cystic duct
and gall-bladder. The evagination is wide transversely though
not extending as far laterally as the liver. In ventral view the
gall-bladder appears as a wide transverse outpouching. ‘There
is a slight furrow separating it anteriorly and laterally from
the liver proper, and a more pronounced one separating it from
the caudally placed yolk-mass.
In an embryo approximately 9 mm. in length (fig. 37) the liver
is distinctly further advanced than in the preceding one. The
caudal constriction from the gut has progressed rapidly (fig. 4).
The original anterior convex surface of the liver has become
markedly irregular showing numerous depressions or furrows be-
tween projecting masses of cells. Greil (05) figures a model
of the liver in a Bombinator embryo 7.5 mm. long with many
secondary buds. A network of veins already occupies the
spaces between the hepatic buds but Greil only states that
it is present. The anteriorly directed cavity has become con-
stricted dorso-ventrally and the division into ducts is more dis-
tinct. On the left side (fig. 37) there is a ventral (vl) and a
dorsal (dm) projection of the lumen. On the right side the ven-
tro-lateral extension is prominent. The median ventral evagina-
tion (GB) has become more pronounced. There is now the
beginning of a lateral constriction of this evagination represent-
ing the formation of a cystic duct. The anterior lip of the
222 E. A. BAUMGARTNER
evagination has developed into quite a ridge separating the gall-
bladder from the developing hepatic ducts. On the ventral
surface the anterior furrow separating liver and gall-bladder
from yolk-mass is, as before, the more marked.
According to Shore (91) in the frog the furrows found in the
liver-mass are caused by the ‘tunnelling in’ of blood vessels.
That it is not due only to this is apparent in Amblystoma where
sections of this and other embryos show furrows in which there
are no blood-vessels (fig. 4). It is important to note that Shore
saw no vascular endothelium in these spaces which he regarded
as blood-vessels.
6
Fig. 4 Sagittal section of an embryo almost 9 mm. long. X 30. Dhd.,
ductus choledochus; F.g., foregut; G.B., gall-bladder; He, heart; Li., liver; Lu.,
lung; Y, yolk mass.
In another embryo of.9 mm. in length the liver in cross sec-
tion (fig. 5) appears as a large oval mass with an irregular surface
showing deep furrows separating the developing ducts. There is
also a very marked dorso-ventral furrow separating the liver-
mass into two unequal lateral portions of which the left is the
smaller. The right portion is marked by two lesser furrows,
one ventral, the other lateral.
In 10 mm. embryos a beginning of the network of anastomos-
ing trabeculae can be seen. The development of the sinusoidal
DEVELOPMENT OF LIVER AND PANCREAS 223
capillary circulation in this network has progressed. In the
11 and 12 mm. embryos there is a confusing network of trabeculae
and it is difficult to differentiate the main ducts from the hepatic
columns. Shore believed that in the frog the tubules were first
solid and that later a lumen developed. Goette expressed the
opinion that a lumen was present from the earliest formation,
though he admitted this was hard to demonstrate. The reason
of the difficulty of proving this either way is apparent. How-
ever, from a study of sections of Amblystoma it would seem
that a lumen is present from the earliest stages.
Fig. 5 Transverse section of embryo 9 mm. long. X 30. F.g., foregut; L,
left portion liver; FR, right portion liver.
Fig. 6 Transverse section of anembryo 11.5mm. long. X 30. F.g., foregut;
GB., gall-bladder; Z., liver.
_ 38. Position of the organ during development
At a stage represented by 11.5 mm. embryos there is a shift-
ing to the right particularly of the caudal end of the liver (fig. 6).
Such a shifting of the posterior part of the liver was noted at a
later stage in Necturus by Eycleshymer and Wilson (’10) and
others. The reason for this lateralward shifting is probably the
pressure of the rapidly growing stomach and duodenum which are
beginning to take a ventral and sinistral position. It is possible
also that the spleen which is now a prominent organ in the left
dorsal region of the body cavity has some influence on this
Fig. 7
of 13.5 mm.
HNC 3 So) araiera, = satatsrers
ASIN, ore craves eee:
At level of an-
terior end of
liver
A®
A series of transverse sections in the region of the liver.
< 20; B, embryo of 20mm. X15; C, embryo of 35 mm.
G.b., gall-bladder; LZ, liver; P., pancreas; Sp., spleen; St., stomach; x, ostia
of ductus choledochus into gut.
About midway
between first
andad_ third
drawing
Anterior end of
gall bladder
224
Level of
attach-
ment of
cystic
duct to
gall
bladder
“
A, embryo
x 10;
Level of osti-
um of ductus
choledochus
DEVELOPMENT OF LIVER AND PANCREAS 225
movement. Then, too, the ventral pancreas forms quite a
mass in the median ventral region. Figures 7 A, B and C show
the lateral and upward shifting of the posterior portion of the
liver. The first drawing in each of the series shows a section
taken near the anterior end of the liver which here is median
and ventral in position and occupies somewhat more than one-
half of the area of a circle. The second drawings in figure 7
A and B show a beginning of a depression on the left side caused
largely by the change in shape and position of the stomach and
duodenum as mentioned above. Figures 8 to 12 are cross sec-
tions of embryos 13.5 to 35 mm. in length showing the position
of the liver at the level of the junction of gall-bladder and cystic
ducts. Here the lateral and dorsal growth of the liver is marked.
A somewhat further shifting is shown in the third drawing of
figure 7 A and B and the second of 7 C. These sections were
taken near the anterior extremity of the gall-bladder. In all
of these the liver is crescentic in transsection and extends up-
ward almost to the level of the dorsal wall of the stomach. The
last drawing in figure 7 shows the relation of parts at the level
of the opening of the ductus choledochus in the gut. In all
cases a small portion of the liver is found dorsal to the duo-
denum in this region of the embryo. In an embryo 45 mm.
long the anterior end of the liver is median and ventral as de-
scribed above. There is a marked lateral and dorsal growth
of the caudal end but in this embryo there is also quite a marked
ventral growth which would indicate that from now on the
shifting to the right will not be so noticeable, and that there is
a growth to the left also.
4. Development of the biliary apparatus
a. Description of the hepatic ducts in the adult. A description
of the fully formed biliary apparatus may be of interest before
describing the development of the hepatic ducts.
The liver in the adult Amblystoma is a large organ extending
fully one-half the length of the abdominal cavity (fig. 13). It has
a ventral convex surface conforming to the wall of the abdomen
226 E. A. BAUMGARTNER
and is divided by an indefinite median line into a right and a
left part of which the left is the longer and covers the left ven-
tral surface and a part of the lateral wall of the stomach. The
right portion or lobe though somewhat shorter, covers the ven-
Fig. 8 Transverse sections of an Amblystoma embryo 14 mm. long, taken
at level of attachment of cystic duct to the gall-bladder. X35. D, duodenum;
G.B., gall-bladder; Li., liver; P., pancreas; Sp., spleen; Sé., stomach.
Fig. 9 Transverse section of an Amblystoma embryo 13.5 mm. long, taken
at the same level as figure 8. X35. For abbreviations, see figure 8.
Fig. 10 Transverse section of an embryo 15 mm. long, taken at the same
level as figure 8. 35. For abbreviations see figure 8.
DEVELOPMENT OF LIVER AND PANCREAS Dé
tral surface of the stomach to the right of the midline and lat-
erally extends well toward the dorsal wall of the stomach. There
Fig. 11 Transverse section of an embryo 20 mm. long, taken at the same
level as figure 8. X 30. For abbreviations see figure 8.
Fig. 12 Transverse section of an embryo 35 mm. long, taken as in figure 8.
x 15. For abbreviations, see figure 8.
228 E. A. BAUMGARTNER
are usually one or two lesser indefinite furrows dividing the right
lobe into two or three parts. The gall-bladder is embedded in
the caudal end of the right lobe some distance from its ven-
tral surface. Only a small part of its rounded fundus appears
beyond the hepatic tissue. From the notch in the liver caused
by the gall-bladder the one or two lesser furrows of the right
lobe extend forward. The gall-bladder is a pear-shaped ;isac
with its larger end extending laterally and somewhat pos-
Fig. 13 A dissection of an Amblystoma 12 em. long. X 1. The ventral
abdominal wall has been cut away and the gall bladder and main hepatic ducts
dissected out. D, duodenum; D.chol., ductus choledochus; D.cy., cystic duct;
D.h.d., right hepatic duct; D.h.s., left hepatic duct; L.L., left lobe liver; R.L.,
right lobe liver; St., stomach.
teriorly. The smaller, medial and ventral end projects for-
ward and connects with the short cystic duct. Only the large
blind end of the gall-bladder receives a peritoneal covering,
the remainder is embedded in hepatic tissue.
There are two main hepatic ducts. These unite to form a
common bile-duct of variable length which may be joined by the
pancreatic duct just before opening into the gut (fig. 14). Quite
often, however, the pancreatic duct opened into the gut immedia-
ately beside the ostium of the common bile-duct. The ductus
DEVELOPMENT OF LIVER AND PANCREAS 229
choledochus is embedded for some distance in the long narrow
pancreas lying on the anterior surface of the duodenum and
finally empties into the anterior side of the gut near the ventral
surface.
The right hepatic duct is divided into lateral and medial
rami. The lateral ramus divides into medial and _ lateral
branches. Generally the cystic duct opens into the latter
(fig. 14 and 16). However, sometimes the cystic duct is one or
LRGs
Lateral branch of sexs
medial rarrnus
MR IT?S
Media! branch of left raedia/ rornus
MR/d
Medial branch of right latera/ rarms
MRmd
Medial branch of right
medial raraus
LARmd
Lareral Lranch of right
medal rarnus
M.P/ s: s
Medial branch of
left Jatera! rarrus.
DCy-Cystic_gucr
GB
Gal! bladder
L.A/ s ~Loteral branch
Of ler¢ Jateral rarmus eed
Lateral branch of right
lotera/ rarrus
-erm.s -Left redial rarrus
ILS
Lert soferdl rarnus
Rim d-Right medial rans
RI d-Kight lateral rarnus
Di pret.
Lett hepatic duct Dhd-kight hepatic duct
D cho! ~ Ductus Choledochts
D.P.
Fancreatic Cucr.
14
Fig. 14 Diagrammatic drawing of the gall-bladder and hepatic ducts of an
Amblystoma.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, NO. 2
230 E. A. BAUMGARTNER
even two divisions further removed from the common duct,
as shown in figure 17 and 44. In a graphic reconstruction of
the biliary apparatus of a 7 em. embryo (fig. 15) the cystic
duct joins the right lateral ramus as is shown also in figure 48.
The hepatic radicle to which the cystic duct is attached shortly
divides into trabeculae beyond this point. The right medial
hepatic ramus divides and subdivides into branches as shown
in figure 14. Its branches sometimes anastomose with the
branches of the right lateral or left medial ramus (fig. 17).
The left hepatic duct is generally shorter and of slightly smaller
diameter than the right one, as well as more ventral in position.
It is divided as the latter into lateral and medial rami. The
left medial ramus sometimes joins the right medial ramus as
shown in figure 16, and this duct then subdivides as a single
one. Frequently, however, the left medial ramus runs anteriorly
subdividing into smaller branches of which some may anasto-
mose with those of the right medial (fig. 17). The left lateral
ramus is shortly divided into two of which the lateral either
turns caudally (fig. 44) or sends out branches that go to the
posterior portion of the longer left lobe.
b. Development of the ductus choledochus. The ductus chole-
dochus in 9 mm. embryos is still very wide and short. The origi-
nal caudalward projection from the gut cavity has disappeared
and there is only the anteriorly directed common duct. In a
model of an embryo 9 mm. long the ductus choledochus is wide
transversely but constricted dorso-ventrally (fig. 37 and 38).
It is attached at the anterior side of the now ventrally directed
gut. At 11mm. the duodenum has turned ventrally and folded
to the right. A very much constricted and short common
duct is attached to its superior anterior surface. In a 13 mm.
embryo the common duct is attached to the anterior surface
of the cranial fold of the duodenum. As before, the duct is
small and short, soon dividing into right and left hepatic ducts.
The epithelial lining of the duct still contains yolk-granules
and except for a quite irregular but prominent lumen is very
much like the hepatic ducts. Indeed the difference in the
'
DEVELOPMENT OF LIVER AND PANCREAS 23
lining cells of this duct and those of the hepatic trabeculae is
not great.
x \7
Fig. 15 Graphic reconstruction (lateral view) of an Amblystoma 7 cm.
long. X 15. D.chol., ductus choledochus; D.cy., cystic duct; D.h.d., right
hepatic duct; D.h.s., left hepatic duct; D.P., pancreatic duct; G.B., gall-bladder;
L.Br., left branch of common ramus; L.R.l.d., lateral branch right lateral ramus;
L.R.l.s., lateral branch left lateral ramus; L.R.m.d., lateral branch right medial
ramus; L.R.m.s., lateral branch left medial ramus; M.R.l.d., medial branch
right lateral ramus; M.R.l.s., medial branch left lateral ramus; M.R.m.d., medial
branch right medial ramus; M.R.m.s., medial branch left medial ramus; R.Br.,
right branch of common ramus; &.l.d., right lateral ramus; R.l.s., left lateral
ramus; R.m.d., right medial ramus; R.m.s., left medial ramus.
Fig. 16 Graphic reconstruction (ventral view) of an Amblystoma 10 cm.
long. X 15. For abbreviations, see figure 15.
Fig. 17 Graphic reconstruction (lateral view) of an Amblystoma 15 cm.
long. X15. For abbreviations, see figure 15.
Dips 4 E. A. BAUMGARTNER
In another embryo of approximately 13 mm. length, which
is somewhat more advanced, the ductus choledochus is longer
and of larger caliber (figs. 18 and 19). It is, however, still
attached to the cranial surface of the anterior fold of the duo-
denum. The epithelium here is now definitely columnar in
type, though yolk-granules are still present. In this case the
pancreatic duct is attached near the gut to the common duct.?
In an embryo 13.5 mm. long the ductus choledochus (fig. 7—A)
is attached in a fold to the left side of the gut. The duct here
is large but shortly divides into the right and left hepatic ducts.
The attachment of the duct to the left wall of the gut is to be
seen in a less completely developed embryo 14 mm. long. From
now on the common duct is attached to the left side of the gut
which is faced somewhat cranialward, due to its growth anteriorly
and to the right. The length of the common bile-duct before
its division varies. In a 35 mm. embryo modelled the com-
mon duct is quite long and has a distinct turn shortly before
it entered the gut. Here again the pancreatic duct opens into
the common duct. There has been a continual change of posi-
tion of the two ducts from the earliest stage to the fully developed
one. In an embryo 13 mm. long a distinct pancreatic duct is
seen ventral to the common duct. In the further development
with the gradual rotation of the liver to the right there has
been a change in position of the common duct until in the 35
mm. embryo it lies to the left of the pancreatic which is the
condition found in the adult (fig. 44).
c. Development of the major hepatic ducts. The earliest indi- |
cation of the hepatic ducts was pointed out in the description
of the formation of the liver. In a model of an embryo approxi-
mately 5 mm. long, as previously stated, the cavity of the early
hepatic anlage extends far laterally. On either side the cavity
is constricted dorso-ventrally. From the drawings shown by
2 In the further study of the pancreas it was found that this duct was attached
by means of a small tubule to the left side of the ventral duct of the pancreas.
The epithelial lining resembled that of the gall-bladder, for which this duct
was mistaken at first. It might very well be a pancreatic bladder. The pan-
creatic duct in this embryo was to the right of the enlarged duct.
DEVELOPMENT OF LIVER AND PANCREAS De
Choronshitzky it is probable his lateral cylindrical extensions
are the early hepatic ducts. In Amblystoma these lateral ex-
tensions form only the lateral rami of the hepatic ducts. The
medial rami are shown in the model of an embryo about 7 mm.
els.
Fig. 18 Graphic reconstruction (lateral view) of the biliary apparatus of
an Amblystoma embryo 13 mm. long. X 100. D., duodenum; D.chol., ductus
choledochus; D.h.d., right hepatic duct; D.cy., cystic duct; G.b., gall bladder;
R.l.d., right lateral ramus; R.l.s., left lateral ramus; R.m.d., right medial ramus;
R.m.s., left medial ramus; P.D., pancreatic duct.
Fig. 19 Graphic reconstruction (lateral view) of the biliary apparatus of
an embryo approximately 13.5 mm. long. X 100. For abbreviations see figure
18.
long (fig. 36). On the right side in this model there is a lateral
extension of the hepatic lumen. A longitudinal ridge in the
floor of this side shows a beginning constriction into lateral and
medial rami. The medial ramus is more dorsal in position and
appears as a swelling on the outer surface. On the left side there
234 E. A. BAUMGARTNER
is a wide cavity. On the external surface there is a slight dorso-
ventral furrow, an indication of the beginning division into
lateral and medial rami.
In an embryo approximately 9 mm. long the right side shows
a more marked lateral ramus. The medial still somewhat dorsal
ramus is to be seen (fig. 37). Here the left side shows a marked
dorso-medial and a ventro-lateral prolongation. The outer
surface of both sides of the organ shows many projections, the
beginning of tubules from these main rami. The cystic duct
though slightly to the right shows more of a constriction from
that side. The anterior lip of the cystic evagination also is
very prominent. |
The rami are formed from the early hepatic ducts by a cau-
dalward constriction and by elongation. Mitotic figures are
to be seen at this stage but are more numerous in later ones. As
is true of fishes (Seammon 713) there is a relative and actual
reduction in the size of these ducts.
In another 9 mm. embryo the development of the ducts is
seen to have progressed rapidly (fig. 38). Numerous mitotic
figures are to be seen in different sections indicating a rapid
growth of the ducts. There are distinct right and left hepatic
ducts which show a marked growth. There is a medial longitud-
inal ridge in the ventral wall of the ductus choledochus indi-
cating a caudalward progressing constriction and division (fig.
38). The cystic duct (D. cy.) is distinctly differentiated and
attached to the right of the beginning constriction in the common
duct. It extends ventrally and somewhat towards the right.
The right hepatic duct as seen in figure 38, and in a figure of a
model of the cavity of ducts (fig. 20) is divided into a lateral and
a dorso-medial ramus. ‘The lateral ramus is further divided into
lateral dorsal and medial ventral branches. The left ramus also
has medial and lateral divisions.
In embryos from 10 to 12 mm. in length, the trabeculae pre-
sent a confusing network. The epithelium of both the hepatic
ducts and trabeculae are heavily laden with yolk-granules, and
that of the ducts is not yet differentiated into a distinct columnar
type. However, the right and left hepatic ducts are clear. In
DEVELOPMENT OF LIVER AND PANCREAS 235
an 11 mm. embryo the right duct is distinctly divided into lateral
and medial rami. A short cystic duct is attached to the caudal
end of the lateral ramus and on its ventral side. In an embryo
somewhat less than 13 mm. long the same arrangement of a
short common duct and right and left hepatic ducts is present.
The right duct is divided into the medial and lateral rami. The
cystic duct here projects somewhat to the left and dorsalward
connecting as before with the right lateral ramus.
In a graphic reconstruction of a 13 mm. embryo (fig. 18)
the right hepatic duct is divided into lateral and dorso-medial
rami. The short cystic duct extends upward and opens into the
Fig. 20 Anterior view of a reconstruction of the lumina of hepatic ducts
and gall-bladder of a9 mm. embryo. X 100. D.h.d., right hepatic duct; D.h.s.,
left hepatic duct; G.b., gall bladder; R.l.d., right lateral ramus; R.l.s., left lateral
ramus; R.m.d., right medial ramus; R.m.s., left medial ramus.
right lateral ramus. A short lateral branch is the only other
division of the right lateral ramus. The dorso-medial branch
shortly breaks up into trabeculae. The left duct is also divided
into rami. The differentiation of hepatic ducts from trabeculae
is now clearer as the epithelium of the former is columnar in
type.
In figure 19 from an embryo less than 1 mm. longer than the
above, the formation of ducts is seen to have continued. The
right hepatic duct is divided into lateral and medial rami, each
of which is further divided into dorsal and yentral branches.
The same holds true in a general way for the left hepatic duct
and its divisions.
236 E. A. BAUMGARTNER
In the ventral view of the model of a 14 mm. embryo (fig. 39)
the relation of pancreatic duct to the common duct is shown.
The short thick common duct divides into right and left hepatic
ducts (figs. 39 and 40). They lie in almost the same horizontal
plane and are of about the same diameter, but the right is the
shorter, dividing almost immediately into its lateral and medial
rami. In a 13.5 mm. embryo (fig. 41) the right hepatic duct is
of larger diameter than the left. In a15 mm. embryo the com-
mon duct is very short (fig. 42). The right and left hepatic
ducts here are very long as compared with those in other embryos.
The left duct has come to lie in a more ventral plane due to the
shifting of the whole posterior part of the liver and gall-bladder
to the right. The same is true to a greater extent for the left
ducts in the 20 and 35 mm. embryos (figs. 43 and 44). In a 20
mm. embryo the right hepatic duct is the shorter as it is in a
35 mm. embryo. In a 35 mm. stage the left hepatic duct is
almost ventral to the right. The same holds true for a 45 mm.
embryo. In the adult, however, the left duct is again more
lateral to the right, but still somewhat more ventral.
d. Development of the minor hepatic ducts. Right lateral ramus.
The right hepatic duct in a 14 mm. stage is divided into lateral
and medial rami and the right lateral ramus is subdivided into
lateral and medial branches (fig. 39). The short cystic duct
is attached to the lateral branch. The medial branch (fig. 40)
gives off several tubules in an oblique dorso-ventral plane. In
a 13.5 mm. embryo the right lateral ramus is quite ventral to
the medial one (fig. 41). As in the earlier stage, it is divided
into lateral and medial branches. The cystic duct which is now
directed almost horizontally, is attached to the right side of
the lateral branch. The anterior portion of the lateral branch
anastomoses with a duct from the right medial ramus. In a
15 mm. embryo (fig. 42) the right lateral ramus is shorter than
in the preceding specimen. The right hepatic duct is, however,
longer so that the cystic duct is attached to the lateral branch
farther from the-gut. The lateral branch here divides into dor-
sal and ventral branches. In a 20 mm. embryo the right lateral
ramus is very short (fig. 43). In position it is now somewhat
DEVELOPMENT OF LIVER AND PANCREAS 237
dorsal to the right medial ramus. It soon breaks up into dorso-
lateral and ventro-medial branches. Both of these branches
are very long. At the attachment of the cystic duct to the lateral
branch there is a further division of the lateral again into medial
and lateral radicles. The medial branch has anastomoses with
the right medial hepatic ramus. Its further division is in a
dorso-ventral plane. In a 35 mm. embryo the right lateral
ramus divides into dorsal and ventral branches (fig. 44). There
is another division of the dorsal branch and the eystic duct is
attached to the dorsal one of this last division. Frequent an-
astomoses are formed between the tubules of the dorsal and ven-
tral branches, and between those of the dorsal branch and those
from the right medial hepatic ramus, as also of the left medial
ramus.
Right medial ramus. The right medial hepatic ramus of a
14 mm. embryo as shown by model is very simple (fig. 39). It
joins with the left medial ramus, the further division of this
common ramus is into right and left branches. The division of
the medial ramus is very short and its lateral and medial branches
long. Caudally directed tubules are given off from the lateral
branch. The medial branch here is connected with the right
lateral ramus. The medial branch divides dorso-ventrally
into tubules. In a 15 mm. embryo (fig. 42) the medial hepatic
ramus is again very simple. It is short and divides into lateral
and medial branches of which the latter is given off almost at
right angles and from its anterior surface are given off several
tubules. The medial hepatic ramus in a 20 mm. embryo as in a
14 mm. one is joined with the left medial ramus (fig. 43). The
resulting common ramus divides into a right dorsal (R. Br.)
and a left ventral branch (L. Br.). From the right dorsal branch,
dorso-lateral tubules are given off some of which are directed
caudally. In a35 mm. embryo (fig. 44) the right medial ramus
is on the same horizontal plane as the right lateral. Its divisions
are also into dorsal and ventral branches. Many anastomoses
are found between the tubules of this ramus. Tubules from
this ramus join those from the right lateral and from the left
medial ramus.
238 E. A. BAUMGARTNER
Left medial ramus. ‘The left medial ramus is joined to the
right medial in a 14 mm. embryo (fig. 39). In a 13.5 mm.
embryo the left medial is long and divides into dorsal and ven-
tral branches (fig. 41). Also in a 15 mm. embryo is the left
medial ramus quite long (fig. 42). It divides into medial and
lateral branches both of which have dorsal and ventral tubules.
The left medial ramus in a 20 mm. embryo (fig. 43) is joined to
the right. The left ventral branch of this combined duct
divides shortly into dorsal and ventral radicles. In a 35 mm.
embryo (fig. 44) the left hepatic ramus is quite long. Its anas-
tomoses with the other rami have been noted. There are also
several anastomoses with the left lateral ramus.
Left lateral ramus. In a14mm. embryo the left lateral ramus
is very simple, dividing into medial and lateral branches (fig. 40).
The left lateral ramus in the next stage shows further develop-
ment and growth (fig. 41). In a15 mm. embryo this ramus has
lateral branches given off at quite an angle (fig. 42). It is
shorter than the left medial ramus and divides into medial and
lateral branches, the latter sending tubules far out to the side.
The left lateral ramus in a 20 mm. stage is given off nearly at
right angles to the left hepatic duct (fig. 43). It divides into
dorso-medial and ventro-lateral branches. In this case the
lateral branch is the longer. Several tubules go out laterally
almost at right angles and from these tubules hepatic columns
go posteriorly as well as anteriorly. In a 35 mm. embryo (fig.
44) the left lateral ramus forms quite a network of ducts. The
ventral branch makes an arch forward and is then divided into
anterior and posterior branches. In an embryo 45 mm. long the
main hepatic ducts are more nearly on the same horizontal plane.
Of these ducts the left hepatic has extended farther to the left.
e. Development of the gall-bladder and cystic duct. The gall-
bladder appears somewhat later than the liver as noted by
Hammar (’97). It arises as a median ventral outpouching
caudal to or in the posterior end of the hepatic anlage. Choron-
shitzky has figured the anlage of the bladder in a median, sagit-
tal section. The structure is shown as a slight depression de-
veloping from the gut, at the entrance of the common duct, and
DEVELOPMENT OF LIVER AND PANCREAS 239
a definite fold is shown between this and the ventrally extending
lumen of the hepatic anlage. Greil (’05) showed the gall-bladder
in a Bombinator embryo of 7 mm. length caudal to the hepatic
tissue but more closely connected with the liver than with the
yolk-mass behind it. In an embryo approximately 7 mm.
long, which is undoubtedly an earlier stage in Amblystoma
(fig. 3) there is no distinct fold between the gall-bladder and
liver-anlage. Only a slight median depression of the floor at
the posterior end of the hepatic diverticulum is present. No
difference is shown by ordinary stains in the epithelium lining
this early cystic evagination and that of the liver. Not until
later does the epithelium change into the low cuboidal type
characteristic of the adult gall-bladder.
A little later the depression in the floor of the hepatic divertic-
ulum is considerably increased (fig. 36). The position of the
gall-bladder with reference to the opening of the hepatic anlage
has not changed. In a model of liver and gall-bladder of a
9 mm. embryo (fig. 37) the evagination is quite deep. There is
a distinct lateral constriction of the dorsal opening of the gall-
bladder and distinct anterior and posterior lips to the evagi-
nation, indicating the formation of a cystic duct (fig. 4). There
is also a deep furrow anterior to the evagination separating the
gall-bladder from the hepatic anlage. The posterior furrow is
even more marked. The gall-bladder is, however, still very wide
laterally.
In another embryo approximately 9 mm. long the gall-bladder
has a long cranio-caudal diameter. The furrow marking off
the gall-bladder from the hepatic tissue laterally is distinct.
The cystic duct is short and of large diameter and it, as well as
the gall-bladder, lies to the right of the midline. The cystic
duct projects upward and to the left (fig. 21).
A section of the gall-bladder of an embryo 11.5 mm. long
shows there has been a continual shifting to the right (fig. 6).
The cystic duct has become longer but is still of wide diameter.
It projects more to the left and upward. The gall-bladder,
though embedded between hepatic tissue and caudal yolk-mass,
is completely separated from both (fig. 22). In figure 23 is
240 E. A. BAUMGARTNER
shown an increased cranio-caudal diameter, although the trans-
verse is still the greater. The cystic duct here projects more
to the left, still somewhat dorsally and slightly backward.
The cranio-caudal diameter increases rapidly from now on, and
the position of the cystic duct would indicate that there is a more
rapid caudal growth. Figure 23 shows the model of a gall-
Oo.
GB
Fig. 21 Transverse section of an Amblystoma embryo 9 mm. long, taken
in the region of the gall bladder. X 30. D.chol., ductus choledochus; F.g.,
foregut; D.cy., cystic duct; G.b., gall-bladder.
Fig. 22 Sagittal section of an embryo 12.5 mm. long. X 30. F.g., foregut;
G.b., gall-bladder; L7., liver.
A
Fig. 23 Drawing of a model of the gall-bladder of an Amblystoma 14 mm.
long. A, anterior view; B, left lateral view. > 40.
bladder of an embryo almost 14 mm. long. The cystic duct
attached near the anterior end, projects to the left and dorsally.
In two graphic reconstructions of embryos 13 and 13.5 mm.
in length respectively (figs. 18 and 19), the gall-bladder is at-
tached by a short and constricted cystic duct to a radicle of the
right hepatic duct. In figure 18 the cystic duct leads from the
anterior dorsal end of the gall-bladder to the left, caudally and
DEVELOPMENT OF LIVER AND PANCREAS 241
somewhat dorsally, the gali-bladder being distinctly to the right
of the midline. In figure 19 the larger of these two embryos the
cystic duct is not quite at the anterior end, but the cranio-
caudal length of the gall-bladder is distinctly greater. The
general direction of the cystic duct is the same. The gall-
bladder is relatively as far caudally here as the one shown in
figure 18. From the connection of the cystic duct to the gall-
bladder, it appears that there has been a marked growth
cranialward.
In an embryo 14 mm. (fig. 39) long the gall-bladder has
decidedly increased in its cranio-caudal diameter. In trans-
verse section it is almost circular. The cystic duct is of very
small diameter as compared with its earlier size. It projects
now somewhat upward but almost directly to the left, due to
the increased lateral shifting of the liver and the gall-bladder.
In this embryo the cystic duct is attached to the extreme ante-
rior dorsal end of the gall-bladder.
Figure 41 is of a model of a 13.5 mm. embryo. In this the
general shape of the gall-bladder is the same as of the one just
described, except that there is a slight increase in the vertical
diameter (fig. 9). The cystic duct, however, is not attached
at the extreme anterior end but to the left upper side. It ex-
tends towards the left as before but is now almost horizontal.
In a 15 mm. embryo the attachment of cystic duct to the
gall-bladder is further caudalward than the previous one (fig.
42). This seems to mark the limit in its caudal attachment for
all sizes examined. It would be difficult to say whether this
shifting in attachment of the duct to the gall-bladder were due
to a difference in the antero-posterior growth of the gall-bladder
or to the rapidity of differentiation and growth of hepatic ducts.
The cystic duct in this embryo extends toward the left, but
now slightly ventrally, which can be taken as evidence of con-
tinued rotation to the right and dorsalward of the entire biliary
apparatus (fig. 10).
Marshall (’93) has described the gall-bladder of amphibians
developing as a lateral outgrowth from the bile ducts. From
242 E. A. BAUMGARTNER
its position at this stage one could easily be led to such a
conclusion.
The gall-bladder of a 20 mm. embryo shows a very distinct
dorso-ventral increase in diameter (fig. 11). With this there
has been a marked cranio-caudal lengthening (fig. 43). The
relative size of the gall-bladder is now greater. As before indi-
cated, the cystic duct is here again nearer the anterior end, it
extends towards the left and now distinctly ventralward (fig. 11).
A right lateral and slightly ventral view of the gall-bladder is
shown in figure 43.
In a 35 mm. embryo (fig. 44) the vertical diameter of the gall-
bladder has greatly increased. The cystic duct is now in the left
anterior ventral end extending ventrally and to the left. Ina
45 mm. embryo the gall-bladder has the same general shape as
in the preceding, and the cystic duct has not changed in position
(fig. 12):
In a graphic reconstruction of the biliary apparatus of a 10
cm. Amblystoma the cystic duct extends to the left, somewhat
ventrally and anteriorly (fig. 16). The gall-bladder is pear
shaped (fig. 13) with its large, blind end projecting slightly dor-
sally and to the right but mainly caudalward.
f. Summary of the development of the biliary apparatus. In
summarising the development of the hepatic ducts a table of
the ducts as found in the various models will bring out more
clearly their relations to the main duct. Table 3a to 3d shows
the principal variations found in the hepatic and cystic ducts.
TABLE 3a
Ductus choledochus
|
Left hepatic duct Right hepatic duct
| | |
Lt. lat. ramus, Lt. med. ramus Rt. med. ramus, Rt. lat. ramus
alt he | 4
Branches—Lat., Med. lLat., Med. Med. Lat. Med. Lat.
Cystic duct
DEVELOPMENT OF LIVER AND PANCREAS 243
Or, in case of anastomoses of the medial rami, as was found in
two embryos of 14 and 20 mm. length and two older Amblystoma
of 7 and 10 cm. length respectively, the following table is given:
TABLE 3b
Ductus choledochus
|
Left hepatic duct Right hepatic duct
| | a |
Lt. lateral ramus Common medial ramus Rt. lateral ramus
|
| a
Branches,—Lat. Med. Left Right Med. Lat.
|
Cystie duct
The cystic duct is attached as here shown:
TABLE 3c
Right lateral ramus
Medial branch Lateral branch
Cystie duct
TABLE 3d
Right lateral ramus
Medial branch Lateral branch
Medial radicle Lateral radicle
Cystic duct
or as found in a 35 mm. embryo and one of the larger Ambly-
stoma.
From these tables it will be seen that sometimes the right
and left medial rami are joined. The division of the common
medial ramus is into right and left branches. In their position
244 E. A. BAUMGARTNER
and final division these branches are the same as the right
medial and left medial rami. As will be seen in figures of
the different models, the smaller embryos did not have all of
the divisions and subdivisions marked in the tables. In figure
39, for instance, the right branch of the common medial ramus
shows no further division, the left branch only one. Further
division of both is seen in the 20 mm. stage (fig. 43). The
division here, however, is more into dorsal and ventral radicles,
due to the more marked lateralward shifting of the liver and the
ducts. The extreme of this lateral shifting is seen in figure 44,
where the left hepatic duct is almost ventral to the right. The
left lateral ramus in a 45 mm. embryo does not hold such a
ventral position with reference to the left medial.
There seems to be no definite rule in regard to the anastomos-
ing of ducts. In a 35 mm. embryo they are the most frequent
and here apparently because the ducts were crowded so close
together. That the right and left medial rami sometimes join
and form one duct is seen in the models of a 14 and a 20 mm.
embryo, also in the graphic reconstruction of a 7 em. and 10
cm. Amblystoma. It would seem this fusion of the ducts is
quite probably due to crowding.
The definite position of the hepatic ducts with reference to
the portal vein is seen for all embryos (figs. 8 to 12). The same
relation is also found in the adult. As a rule there is a branch-
ing of the hepatic ducts corresponding to the division of this
vessel. In the developing embryo the ducts are found usually
to the right of and ventral to the portal vein.
From the usual description of the biliary apparatus in the
frog it would seem that there is a fairly close correlation in the
main features between these two amphibians. The figures of
Ecker, Wiedersheim and others show a gall-bladder connected
to a right hepatic duct. There is also a left hepatic duct, the
two uniting in the pancreas and forming a ductus choledochus
which, as usually described, is joined by the pancreatic duct.
In no ease in Amblystoma were two cystic ducts found as is shown
for the frog. The division into rami in the frog as far as the ducts
have been figured, seems to be somewhat different from that
DEVELOPMENT OF LIVER AND PANCREAS 245
found in Amblystoma. The more marked divisions of the liver
into several lobes may partially explain this. The duct-system as
found in Necturus is quite different. Kingsbury here described
three hepatic ducts opening into the gut. These anastomosed
with each other and two were joined by the ventral pancreatic
ducts. The third is a duct direct from the gall-bladder which,
however, anastomoses with the other hepatic ducts. Grdnberg
(94) described three hepatic ducts which unite with the cystic
duct and form a ductus choledochus in Pipa americana.
Bates (04) has described the hepatic ducts in Amblystoma.
According to his description there are four heaptic ducts, two
of which join the bile-duct in its course through the pancreas
and the other two just as it opens into the intestine. It is possi-
ble that the two he found joining the bile-ducts are the right
medial and lateral rami, and the other two, the left medial and
lateral rami. In that case the ductus choledochus and the right
and left hepatic ducts were very short as was found in some of
the material used in this work. Or it may be that the two ducts
which joined the bile-duct as it opened into the intestine are the
two pancreatic ducts which have not fused until just at the
ostium of the hepatic duct. The first two ducts then would be
the right and the left hepatic ducts. I have never seen the
eystic duct (bile-duct as Bates terms it) open directly into the
common hepatic duct.
From the models and drawings it will be seen that the gall-
bladder at first has a wide dorsal communication just caudal
to the hepatic lumen. As this communication constricts there
is formed a short large cystic duct extending dorsally into the
right hepatic duct. With further growth and division the cystic
duct extends more and more to the left until at the 15 mm.
stage it is almost horizontal and at the 20 mm. stage projecting
ventrally and somewhat anteriorly. Its earliest attachment is
to the ventral surface of the common bile-duct, but in the lateral-
ward shifting of the whole liver its attachment goes to the left
side of a right hepatic radicle. The connection of the cystic
duct to the gall-bladder in early stages is to its dorsal surface
about midway between cranial and caudal pole. Somewhat
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, NO. 2
246 E. A. BAUMGARTNER
later the connection is nearer the cranial end and usually reaches
the extreme anterior end. The cranio-caudal growth of the
gall-bladder has kept pace with the lengthening and differentia-
tion of ducts in the 13 to 14 mm. stage. From the relations in
a 15 mm. embryo it appears that the gall-bladder has shifted
anteriorly. In this case the hepatic ducts have lengthened more
than the gall-bladder. At 20 mm., however, there has been a
marked increase in cranio-caudal growth of the gall-bladder so
that it is almost as long as the ducts.
Beginning about at this stage the cystic duct is again attached
nearer the anterior end of the gall-bladder. This may be taken
as evidence that the cystic duct really shifts in its attachment
to the gall-bladder. This seems to be borne out in some cases
by the fact that its attachment to the hepatic ducts is to a
division of the lateral branch of the right lateral ramus instead
of to the lateral branch proper. In some cases where the lateral
branch is quite long the attachment may have remained to it.
Whether the gall-bladder originates from the early hepatic
anlage or from the gut has caused much discussion. As said
before Piper (’02) thought this a matter of interpretation. The
more marked furrow caudal to the gall-bladder might be taken
as evidence of its belonging to the hepatic anlage, also the fact
that the same type of yolk-laden cells form hepatic tissue and
gall-bladder. That it, at least is directly caudal to the hepatic
anlage is proven by the early connection of its duct to the common
bile-duct.
The connection of the cystic duct probably depends to some
extent on the extent of growth and division of the hepatic
ducts. It will be remembered that in the earlier stages the
cystic duct opens into the common duct, then into the early
right hepatic. In the further growth and division of the right
hepatic duct the cystic duct becomes attached to one of its radi-
cles. As noted above, the cystic duct opens into the lateral
branch of the right lateral ramus in all of the embryos studied
except one, which was 35 mm. long.
That there is considerable variation in the relative dorso-
ventral position of these main hepatic ducts is to be expected.
DEVELOPMENT OF LIVER AND PANCREAS 247
However, in general, a study of the models shows a close simi-
larity in their positions. Theze is a constant rotation of the
liver towards the right and with this is a similar one of the
hepatic ducts. In this rotation the right ducts come to be more
dorsal in position, the left more ventral. The right lateral
divisions would thus be dorsal to the right medial and the reverse
should be true for the left. In general such an arrangement is
found. A variation in the length of the different ducts is present.
However, there is qute a definite relation in the total lengths
of ducts in the different embryos. In a 15 mm. embryo the
common duct is quite short but the greater length of the hepatic
ducts compensates for this reduction. In a 35 mm. embryo the
common duct is long, the hepatic ducts and their radicles divide
shortly.
Ill. THE DEVELOPMENT OF THE PANCREAS AND PANCREATIC
DUCTS
1. Literature
The literature concerning the development of the amphibian
pancreas like that regarding the liver is divisible into two periods,
and Goette’s work (75) may again be said to mark the begin-
ning of the newer one. The older observers mainly considered
the pancreas as a part of the liver, or a modified lobe of that
organ.
A list of the investigators describing the development of the
pancreas will be found included in the tabular classification of
the literature on the development of the liver (table 2).
Goette (75) in his studies on the development of the Bombin-
ator recognized three distinct pancreatic anlagen, two ventral
and one dorsal. The dorsal one he described as placed just
caudal to the gastroduodenal loop. The two symmetrical ven-
tral anlagen develop from the primitive hepatic duct. Of
these the right grows dorsalward to join the ventral growing
dorsal anlage. The right duct changes in position until it opens
into the left side of the hepatic duct. The united right and left
duct then separates from the common bile-duct. Apparently
248 E. A. BAUMGARTNER
Goette considered the left outpouching as a rudimentary one.
Later the dorsal duct disappears, thus leaving but one perma-
nent pancreatic duct.
Balfour (’81) and Hertwig (’88) described a dorsal outpouching
of the gut wall caudal to the level of the common bile-duct.
The development of the pancreas in both Urodela and Anura
was described by Goeppert (’91). In both he found as Goette
had described, one dorsal and two symmetrical ventral out-
pouchings. 100.
41 Ventral view of a reconstruction of the hepatie ducts and gall-bladder of
an embryo 13.5 mm. long. » 100.
42 Dorsal view of a reconstruction of the hepatic ducts and gall-bladder
of an embryo 15 mm. long. »_ 100.
43 Right ventral view of a reconstruction of the hepatic ducts and gall-bladder
of an embryo 20 mm. long. 100.
D., duodenum
D.chol., ductus choledochus
D.cy., cystic duct
D.h.d., right hepatic duct
D.h.s., left hepatie duct
D.P., pancreatic duct
g.b., gall-bladder
L.Br., left branch of common ramus
L.R.l.d., lateral branch right lateral
ramus
GRAS,
ramus
L.R.m.d., lateral branch right medial
ramus
LUIS,
ramus
lateral branch left lateral
lateral branch left medial
M.R.l.d., medial branch right lateral
ramus
Metialsse,
ramus
M.R.m.d., medial branch right medial
ramus
M.R.M.s.,
ramus
medial branch left lateral
medial branch left medial
R.Br., right branch of common ramus
R.l.s., left lateral ramus
R.l.d., right lateral ramus
R.m.s., left medial ramus
R.m.d., right medial ramus
Z., extra duct in 13.5 mm. Amblystoma
embryo
DEVELOPMENT OF LIVER AND PANCREAS PLATE 2
E. A. BAUMGARTNER
PLATE 3
EXPLANATION
OF FIGURES
44 Right ventral view of a reconstruction of the hepatie duets and gall-
bladder of an Amblystoma embryo 35 mm. long.
D., duodenum
D.chol., ductus choledochus
D.cy., cystic duct
D.h.d., right hepatic duct
D.h.s., left hepatic duct
D.P., pancreatic duct
g.b., gall-bladder
L.Br., left branch of common ramus
L.R.l.d., lateral branch right lateral
ramus
L.R.l.s., lateral branch left lateral
ramus
L.R.m.d., lateral branch right medial
ramus
70.
L.R.m.s., lateral branch left medial
ramus
M.R.1.d., medial branch right lateral
ramus
M.R.l.s., medial branch left lateral
ramus
M.R.m.d., medial branch right medial
ramus
M.R.m.s., medial branch left medial
ramus
R.Br., right branch of common ramus
R.l.d., right lateral ramus
R.l.s., left lateral ramus
R.m.d., right medial ramus
R.m.s., left medial ramus
DEVELOPMENT OF LIVER AND PANCREAS PLATE 3
E. A. BAUMGARTNER
972
273
PLATE
EXPLANATION
45 Anterior view of the pancreas
embryo. X 60.
46 Anterior view of the pancreas
D., duodenum
D.pan., dorsal pancreas
G.B., gall-bladder
Lt.pan.d., left ventral pancreatic duct
Rt.pan.d., right ventral pancreatic duct
St., stomach
and ducts of a 15 mm. embryo.
274
4
OF FIGURES
and ventral pancreatic ducts of a 13 mm.
x 60.
V.pan., ventral pancreas
Y, yolk-gut
D.chol., ductus choledochus
V.pan.d., ventral pancreatic ducts.
For other abbreviations, see figure 45.
DEVELOPMENT OF LIVER AND PANCREAS PLATE 4
E. A. BAUMGARTNER
THE MICROSCOPIC STRUCTURE OF THE YOLK-SAC
OF THE PIG EMBRYO, WITH SPECIAL REFERENCE
TO THE ORIGIN OF THE ERYTHROCYTES
H. E. JORDAN
Department of Anatomy, University of Virginia
THIRTY-FIVE FIGURES (TWO PLATES)
CONTENTS
NSM CROGUCHIOM Soh ecw aa eee ere. Sane eRe aI hs cet we e 207
MiPViaterraleand methods) pes scehejeert 2 ete ae ee coe Sans Sh sesus rae 278
UR DESCrUPORV Gtk ata e joe Man Ack ean ee eta See eles fae chk Urabe eee 279
Qe NC LeM OG CTIA re wise neat se aca Sty en eee. do os 279
be hesmies othe live ecpen niece sere orate: art Pe pe cel ee ea se)
Gut MO GMNESONC UVTI sn. acne tome 6 yeas So he eater seo ste Doe ae 281
ceaduierendotheliumis 5 Aan bitin spoke eee Get tena AC ee 281
emmulehiesbl OOdKCElIS: grees oasis cassie Se erie eee es ac cane ok ee, 283
Le ROTIATN GLORY xia. caren ea.atate OG a uae aes VE hoe eases ok eae 283
DHE @ OlAStS eee wee Gai se eae oe ee ee 283
Sey COTO ASUS: m0 to actos tie a Se tons -nictin od Se en 286
AN OTM Ob laStS a -er.-wwice oe eerste asst oes + oe eee 287
bee Grant Cells sans aes Ae Pane etn: Aen 10> |e Py eee ee IS 7]
VE IDISCUSSION -2%.5 5 acon oe RR aN oy nae aR RANT, MON NERS, VD Le, SET ew 289
a2 bhestunction, of they olkesac #144 senses. oa... - a ctasee eee eee 289
DD IS ESCLV Cray. 5 silat ate ARI ate econ oes Ae SES eee bia Ee 289
Dee ENA CHO DOLE LICL 308 os. Harn ates Sasha Salas cams iis espa oe MAE ATR oe ee 292
bathe grant cells: 4.f4 «nue Me mnenes Mes tae S00 i ee et 295
WAS NSLUN AGN TOE ah gerne Beye a arene, Saeed ne Rie rs. Ce E nS, Se 297
VI. Literature cited...... RO EN er aOR, | 4, Oe en eens «OS
I. INTRODUCTION
The chief purpose in view in this study of the yolk-sae of the
pig embryo was the acquisition of further data regarding the
earliest stages in blood cell origin and development in mammals.
The yolk-sac was believed to be the most favorable material
for the search for evidence concerning the disputed relationship
between mesenchyma, primitive endothelium and haemoblasts.
The pig embryo was selected for study on account of its ready
availability. It was hoped that information could be contrib-
277
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, NO. 2
278 H. E. JORDAN
uted to the following debated questions in haemopoiesis:
1) Does the angioblast bear any direct genetic relationship to
the entoderm? 2) Does the yolk-sac mesothelium produce
haemoblasts? 3) Does the mesenchyma differentiate in part
into endothelium? 4) Do haemoblasts arise directly from mes-
enchyma? 5) Do haemoblasts differentiate from endothelium?
6) What is the origin and function of the giant cells of the yolk-
sac? The first question involves a careful consideration of the
structure of the entoderm; which in turn raises the question:
7) What is the function of the entoderm in yolk-sacs which
contain little or no yolk?
A preliminary report of this study appeared in the Proceedings
of the thirty-first session of the American Association of Anato-
mists (Anat. Rec., 9:1, ’15, pp. 92-97). In the present paper
more extensive observations, with illustrations, are recorded.
Moreover, a further study of the entoderm compels a reinter-
pretation of the cytoplasmic filaments of these cells; my earlier
conclusion that they are mitochondrial in nature no longer seems
warranted.
A portion of this investigation was done at the Marine Bio-
logical Laboratory, Woods Hole, Massachusetts during the sum-
mer of 714. I take this opportunity to gratefully acknowledge
my indebtedness to the institution for the privileges of a research
room.
II. MATERIAL AND METHODS
The material consists of pig embryos ranging in length from
5 to 25 mm. Zenker’s and Helly’s fluids were used for fixation.
The stains employed were the Giemsa blood stain, and the
haematoxylin and eosin combination. Sacs of stages within the
limits specified differ essentially only with respect of relative
abundance of the various types of early blood cells. The 10
to 15 mm. stages were soon discovered to be most favorable
for this study, since here was included in the same sections
both earliest and later stages in haemopoiesis. Haemopoietic
phenomena seem to be at their height in the yolk-sac of the pig
embryo at about the 10 mm. stage of development.
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO 279
III. DESCRIPTIVE
a. The entoderm
It seems preferable to begin the description of the histology
of the yolk-sae with the entodermal constituent of its wall.
In the 5 mm. stage the entodermal cells are cuboidal, and
arranged in a single layer; there is as yet no trace either of solid
or tubular evaginations into the enveloping mesenchyma.
In the 10 mm. stage of development the lining cells are colum-
nar, the taller being about twice the length of the tallest cells
in the earlier. stage; they are still arranged in a single layer.
However, there is great variation in the form of the cells; the
predominating type of entodermal cell is columnar, but all
transitional forms appear from very low cuboidal to tall colum-
nar cells. At certain points the entoderm invaginates the
mesenchyma in the form of short cords and tubules. The
‘tubules’ are scarcely more than shallow folds, but recall the
larger branched tubules of the yolk-sac of human embryos of
this length (Meyer (18); Jordan (10).) The condition is probably
to be interpreted in terms of a mechanical adjustment on the
part of the entoderm to the exiguous confines delimited by the
enveloping mesenchyma, or it may perhaps be merely a shrink-
age phenomenon.
At the 25 mm. stage of development the entodermal cells
appear shorter columnar but are still almost invariably arranged
in only a single layer. Occasional small stratified areas occur
similar to those characteristic of the human yolk-sac of even
much earlier stages, but they are perhaps most correctly inter-
preted as short stout entodermal buds or cords. At this stage
the very sparse enveloping mesenchyma is extensively invaded
by very numerous robust solid cords and irregular tubules of
entodermal cells. In tangential sections the yolk-sac wall of
this stage looks strikingly like reptilian liver tissue.
The cytology of the entoderm is essentially identical for the
5 to 25 mm. stages (figs. 2 and 31). The vesicular nucleus is
relatively large and spherical, and is generally placed nearer
the basal pole. It contains one or several large, spheroidal,
280 H. E. JORDAN
chromatic nucleoli (fig. 2) and a delicate wide-meshed granular
reticulum. Many of the cells are undergoing mitosis. The
cell wall appears distinct. But there is no indication of terminal
bars nor brush borders, such as have been described for the ento-
dermal cells of the human yolk-sae by Branca (2). In Giemsa-
stained preparations the cytoplasm is colored dark blue, the
nuclei light blue, and the nucleoli bluish orange or lilac.
The most striking feature of these cells is the presence of a
generous amount of delicate filaments (basal filaments; ergas-
toplasmic filaments) scattered throughout the finely granular
basophilic cytoplasm. They are oriented in general parallel
to the long axis of the cell. They may be coarser or finer, in
length equal to that of the entire cell or much shorter; and they
may be apparently homogeneous or segmented (fig. 31). The
latter condition would seem to indicate the possibility that they
may fragment into secretion granules, but the evidence for this
conclusion is not wholly satisfactory. Their probable signifi-
cance and nature will be discussed in a later section. It may
suffice here to state that the cells of the liver (fig. 32) and those
of the mesonephric tubules contain apparently identical cyto-
plasmie threads; and that in no case do they bear any direct
relationship to mitochondria, which must have been dissolved
by the fixing fluids used. —
b. The mesothelrum
The outer surface of the yolk-sac wall, like the homologous
layer of the splanchnopleure generally, is characterized by a
layer of greatly flattened cells each bulging more or less at the
point where the nucleus is located. The cytoplasm is delicately
reticular like that of the underlying mesenchyma, with which the
mesothelial cells are apparently in syncytial continuity (figs. 2
and 4). The nuclei are generally relatively large, oval, vesicular
structures, with one or several small irregular net-knots, and a
delicate wide-meshed nuclear reticulum (figs. 9 and 20). In
their general form, structure and light staming capacity they
are practically identical with the nuclei of the mesenchyma and
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO 281
the endothelium (figs. 13 to 18). In the Giemsa stain the nuclei
of these three tissues are similarly colored bluish orange, while
the cytoplasm stains a lighter blue. Occasional cells may be
seen in mitosis, but there is no clear evidence to indicate that
their proliferation products may differentiate into haemoblasts.
The proliferation is most probably related only to the extension
of the mesothelial covering. Certain cells, however, are more or
less rounded, simulating early stages in the formation of haemo-
blasts from endothelium (fig. 20).
c. The mesenchyma
The mesenchyma is of very variable amount in different
portions of the wall (figs. 2, 4, 29 and 30); in certain regions it is
so sparse as to be barely discernible between the entoderm and
the mesothelium; in other regions it may greatly exceed in width
that of the tallest portions of the entoderm. It is a loose-meshed
syncytium containing numerous spaces and occasional small
blood islands, and larger and smaller blood vessels or sinusoids
(figs. 29 and 30). Around certain spaces the mesenchymal
cells may become arranged so as to very closely simulate endo-
thelial cells. Indeed it seems impossible to differentiate between
such a cell and certain endothelial cells from blood-cell-contain-
ing channels. It seems difficult to avoid the conclusion that
endothelium is thus differentiated from the mesenchyma, the
differentiation depending here as in the case of the structurally
apparently identical mesothelium, upon the mechanical factor
of pressure (fig. 29). Many of the mesenchymal nuclei are in
some phase of mitosis, and occasional nuclei appear to be divid-
ing amitotically.
d. The endothelium
The cells lining blood-cell-containing channels are flattened
elements, of fusiform shape in sections. ‘The commonest type
of cell contains a vesicular oval nucleus, practically identical
with that of the mesenchyma and the mesothelium (figs. 4, 13
and 29); and also the delicate reticular cytoplasmic structure
282 H. E. JORDAN
of the endothelial cells is ike that of these cells. Moreover,
the endothelial cells appear to be in direct syncytial continuity
with the mesenchyma. Many are in mitosis, and occasional
nuclei appear to be dividing amitotically. It seems most prob-
able that they are actually mesenchymal cells modified in shape
‘by the pressure of the confined blood stream. Endothelial
cells which lie next the entoderm are sharply separated there-
from (figs. 29 and 30). The entodermal cells rest upon a deli-
cate but distinct basement membrane, with which the endothe-
lium is not in structural continuity (fig. 2). The vascular an-
lages (angioblast) are at certain points in direct continuity with
’ the mesenchyme, but are sharply demarked from the entoderm
(fig. 29). There is no evidence here that the angioblast has any
direct genetic relationship to the entoderm; all the available
morphologic data are opposed to the idea of such a relationship.
The endothelium includes, however, numerous cells which
may be arranged into a complete series connecting the above
described endothelial cell with a haemoblast (figs. 4, 18, 14, 15
and 16). The transition steps consist of a progressive rounding
up of the nucleus and a gathering of the cytoplasm around it.
At the same time the nucleus enlarges and the cytoplasm ap-
pears to increase in amount. Moreover the cytoplasm becomes
more highly basophilic and appears finely granular. The cell
as a whole, of fusiform shape, becomes progressively shorter
and finally separates from the endothelial wall either as a short
fusiform cell, or frequently as a spherical cell flattened at its
proximal pole and drawn out laterally into delicate processes
which gradually separate from the vessel wall (figs. 5 and 6).
Such cells may even become multinucleated before separation
(figs. 8 and 9), and undergo cytoplasmic differentiation, even
elaborating haemoglobin, as will be described below. The
multinuclear condition appears to be the result of amitotic
nuclear division (figs. 9 and 35). The observation of the differ-
entiation of endothelial cells into haemoblasts is of cardinal
importance, and will be more fully discussed in a later section.
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO 283
e. The blood cells
1) Terminology. Four distinct types of cells may be recog-
nized: 1) The haemoblasts, or blood mother-cells. These cor-
respond with the primitive ‘lymphocytes’ of Maximow (16),
and the ‘mesamoeboid cells’ of Minot (19). 2) The erythro-
blasts, corresponding with the ‘ichthyoid’ blood cell of Minot,
and in part with the ‘megaloblast’ of Maximow. 3) The
normoblasts, corresponding with the ‘sauroid’ cell of Minot.
The last two may be designated inclusively as erythrocytes.
4) The giant cells, both megakaryocytes and polykaryocytes.
The majority of the blood cells can be classified under one or
the other of the above heads. However, between typical primi-
tive haemoblasts and erythroblasts, and between the latter and
normoblasts, as also between haemoblasts and giant cells, com-
plete series of transition forms occur.
Up to the 15 mm. stage no cell is present that can be certainly
identified as a leucocyte. The haemoblasts are structurally
very similar to the lymphocytes of the adult, and if they are
indeed in part at least, functionally identical, as claimed by Maxi-
mow in support of the monophyletic theory of haemopoiesis,
they may be properly designated ‘lymphocytes.’
2) Haemoblasts. This terminology implies that the cell
designated ‘haemoblast’ is the common mother-cell of both leu-
cocytes and erythrocytes. No evidence, besides its very close
similarity to a lymphocyte, accrues from this study to indicate
that the cell in question is also a leucocyte progenitor. It may
be noted, however, that this cell would apparently have to under-
go less differentiation in becoming a mononuclear, or even a
polymorphonuclear, leucocyte, than in becoming an erythro-
plastid. Moreover, there is now a very considerable body of
embryologic data to show that this cell in certain mammals
(rabbit, Maximow (16); birds, Dantschakoff (5); reptiles, Dant-
schakoff (6), and Jordan and Flippin (14); and selachii and
amphibia, Maximow (17)) is indeed the parent cell of both red
and white blood corpuscles. Thus while the haemogenic proc-
284 H. E. JORDAN
ess here to be described is purely erythropoietic, the primitive
cell is nevertheless properly termed ‘haemoblast.’
The haemoblast is in its youngest form a relatively small cell,
ranging from about half to approximately the full size of the
definitive normoblast, with a larger nucleus and much less eyto-
plasm (figs. 1, 2 and 3). It has a relatively enormous nucleus,
which is enveloped by a narrow shell of cytoplasm generally
wider at one point over an area of from less than a quarter to
more than a half of the surface (fig. 1 a). The cytoplasm is
finely granular and deeply basophilic. The nucleus is vesicular
with one or several spheroidal chromatic masses (nucleoli),
scattered irregularly through a wide-meshed, delicate, fre-
quently granular reticulum containing larger chromatin granules
peripherally on the nuclear membrane. In Giemsa-stained prep-
arations the nucleoli are colored lilac, the nuclear sap bluish
pink, the cytoplasm deep blue. Thehaemoblast may show several
blunt pseudopods indicating amoeboid capacity (figs. 2 and
27). The young haemoblasts are more generally peripherally
placed in the blood vessels, the later differentiation stages more
centrally.
The haemoblasts show a very wide range of size variations and
nuclear forms, while at the same time adhering to a very close
structural similarity both nuclear and cytoplasmic (figs. 1, 2, 3
and 7). By growth the primitive haemoblast may become very
large; this growth may be chiefly nuclear (fig. 34) or chiefly
cytoplasmie (fig. 7). It does not seem possible to draw a sharp
line between large haemoblasts and certain so-called ‘giant
cells,’ to be described below. Their essential nuclear and cyto-
plasmic features are very similar.
By division a larger haemoblast gives rise to smaller, structu-
rally identical, haemoblasts. The mode of division may be
mitotic, and apparently also amitotie (figs. 3, ec, d, and e, and 22).
Cytoplasmic division frequently does not directly follow nuclear
division, thus giving rise to binucleated cells (fig. 3 d and e).
Similarly, tripolar spindles may produce trinucleated cells (fig.
21), or the same may be probably produced also by direct di-
vision (figs. 11 and 12). Multinuclear cells are probably simi-
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO 285
larly formed (figs. 25 and 35). The bi- and multinucleated
types will be further described under ‘giant cells.’
Haemoblasts have a double source of origin: 1) from the
mesenchyma (fig. 30); 2) from the endothelium of the earliest
blood vessels (fig. 4). Since this endothelium, however, also
originally arose from mesenchyme, the primary, in part indirect
source, is the same, namely the original mesenchyma.
The endothelial origin of haemoblasts has already been
partially described above under ‘endothelium.’ It need merely
be emphasized here that the evidence on this point seems un-
equivocal; transition stages are practically innumerable; their
abundance is so great as to make it difficult to adhere to a
reasonable limit in the selection of illustrations. Possible
objections to the interpretation here given to the observations
will be considered below. ‘The above description pertains only
to intravascular haemopoiesis; the endothelium contributes
also, but apparently much more rarely (except in the mesoneph-
ric glomeruli of the body of the embryo), extravascular haemo-
blasts. The continuity of such with the endothelial wall counter-
vails the possible objection that these are migrants (fig. 4).
The direct mesenchymal origin of haemoblasts concerns it-
self with the blood-islands and certain isolated cells separating
from the mesenchymal syncytium. Peripherally the blood-
islands are in continuity with the mesenchyma, where endothelial
cells are differentiated; centrally the cells are haemoblasts in
various earlier stages of metamorphosis into erythroblasts;
some of these may be binucleated (fig. 29).
The unique and crucial evidence for mesenchymal origin of
haemoblasts pertains to certain isolated cells caught in the
actual process of differentiation and separation from the syncy-
tium. These are admittedly rare, but the evidence they furnish
is of prime importance. It supplies the link in the monophyletic
theory of haemogenesis concerning which there has been the
greatest scepticism. Figure 30 is an illustration of the clearest
case of the condition referred to. Here is shown an area of
mesenchyma in which two of the nuclei, as well as their envelop-
ing cytoplasm, have mesenchymal features; the third nucleus
286 H. E. JORDAN
(h) and its enveloping cytoplasm are of typically haemoblast
character. A delicate chromatic nuclear bridge still connects
the haemoblast nucleus with the mesenchyma nucleus. The
significance of this nuclear bridge is uncertain, but it plainly
reveals genetic relationship whatever its meaning in terms of
type of cell division. Such instances should definitely dispose
of the objection that all mesenchymal haemoblasts are migrants
from adjacent blood vessels. Haemoblasts are very variable
in form, due to the variable number and form of their pseudo-
pods (fig. 27). They must be regarded as capable of extensive
amoeboid motility.
It is a matter of sufficient importance to warrant special
emphasis at this point, that between typical haemoblasts and
typical erythroblasts, next to be described, transition forms
exist abundantly (fig. 1b). The marks of transition pertain both
to the nucleus and the cytoplasm. The change is perhaps
most marked in the staining capacity of the cytoplasm. This
loses its intense basophily, and in Giemsa preparations becomes
a much lighter pink or grayish blue. This chemical alteration
inheres principally in the elaboration of a small amount of
haemoglobin. The cytoplasm shows also faintly a coarse wide-
meshed reticulum. And a distinct cell wall is now evident
(fig. 1 b), whereas the haemoblast is apparently a naked cell.
The nucleus becomes relatively smaller and more chromatic;
the nucleoli tend to disappear, and the nuclear reticulum be-
comes coarser, more granular and more chromatic.
3) Erythroblasts. ‘Thesecells are characterized by their slightly
smaller spherical nuclei and an acidophil cytoplasm generous
in amount (fig. 1c). The nuclei generally lack distinct nucleoli
but contain a coarsely granular, intensely chromatic, nuclear
reticulum. The cytoplasm has frequently a finely granular
appearance (fig. 3 f). In Giemsa preparations the nucleus
stains blue, the cytoplasm a faint brownish pink. These cells
are much more uniform in size than the haemoblasts and are
generally mononuclear. They undergo very extensive mitotic
proliferation. The transition stages (figs. 1 b and 3 f) between
the haemoblast and the erythroblast, characterized by a bluish
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO 287
pink color in Giemsa preparation, correspond to the ‘megalo-
blast’ described by Maximow in the rabbit.
Occasionally a disintegrating erythroblast may be seen in-
gested by an endothelial cell (fig. 28). This observation indi-
cates a phagocytic function on the part of the endothelium of
the yolk-sac vessels. An alternative interpretation will be dis-
cussed below.
4) Normoblasts. The normoblasts differ from the erythro-
blasts in that they have a smaller more compact and chromatic
nucleus, and a more acidophilic cytoplasm (figs. 1 d, 2 e and 3 g).
These cells are very uniform in size. In this character of size
uniformity they differ markedly from the similar cells in certain
lower forms, for example, in turtles. They multiply extensively
by the indirect method of cell division. In Giemsa preparations
the cytoplasm stains a brilliant red, the coarsely granular
nucleus a deep blue. The nucleus frequently has an irregular
lobed contour. The chromatin is frequently gathered into sev-
eral large and many smaller clumps, the reticulum being delicate
and only slightly chromatic. In preparations fixed in Zenker’s
fluid, the haemoglobin content has become dissolved, and the
cytoplasmic area reveals a coarse wide-meshed reticulum,
bounded peripherally by a coarse cell membrane (fig. 3 g).
By abstriction of the portion of the cytoplasm containing the
excentric nucleus, in the manner described by Emmel (7), the
erythrocyte becomes an erythroplastid. These stages in plas-
tid formation are still extremely rare in 10 mm. embryos.
5) Giant cells. These cells include a great variety of different
forms and sizes. The extremes include: 1) An enormous cell
consisting almost wholly of nucleus, the naked cytoplasm con-
stituting a mere shell (figs. 33 and 34). The cytoplasm is baso-
philic. The vesicular nucleus is generally extensively lobed and
contains many large spheroidal and irregular chromatic masses;
its nuclear reticulum is wide-meshed, granular, and intensely
chromatic. 2) A cell of similarly large size with generally two
or three relatively small, spherical, oval or irregular, pale stain-
ing, granular nuclei (figs. 23, 24 and 25). The nuclei may
contain one or several nucleoli; and the reticulum is more regular,
288 H. E. JORDAN
more delicate, sometimes double (fig. 23) and less deeply chro-
matic. The cytoplasm is slightly acidophilic. Both nuclear
and cytoplasmic features resemble those of the erythroblasts
(‘megaloblasts’). 3) A cell of similar or even larger size with
numerous nuclei (as many as eight are common) of various shapes
and sizes and differing in structure between the two extremes
above described (figs. 11, 12 and 35). The cytoplasm of such a
cell is also more or less basophilic.
The origin of giant cells can be definitely traced by means of
transition stages to the haemoblasts. Type 1, above described,
is simply a giant haemoblast (compare figs. 3.a, 7 and 33). Type
2 is a giant haemoblast with several nuclei (compare figs. 3 a
and 11) derived by nuclear amitotic division—occasionally
possibly also by nuclear mitosis—unaccompanied by cyto-
plasmic division. ‘The cytoplasm has entered upon the early
stages of differentiation into erythroblast cytoplasm. Type 3
is derived from type 1 by extreme and irregular fission of the
single nucleus, accompanied by slight differentiation in the
cytoplasm (figs. 12, 25 and 35).
Frequently a typical giant cell with two or even three nuclei
may be seen in continuity with the endothelial wall of the
blood vessel, and in late stages of separation (figs. 8 and 9).
This observation further supports the conclusion of haemoblast
derivation of giant cells. There is no evidence in favor of an
entodermal origin of giant cells as held by Graf. v. Spee (22)
in the case of the human yolk-sac.
A small number of giant cells contain one or several normo-
blasts. The normoblast periphery may be separated from
the enveloping giant cell cytoplasm by a narrow space (fig. 10);
or such space may be lacking, in which event the continuity
between the two cytoplasms seems complete (fig. 26). Two
possibilities of the origin of these intracellular normoblasts at
once suggest themselves: 1) ingestion; 2) differentiation from
the nuclei and portions of the surrounding cytoplasm of the giant
cell. The fact that endothelial cells (potential haemoblasts)
may ingest erythroblasts (fig. 28), as above described, lends
much weight to the first suggestion. The further facts, how-
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO 289
ever—1) that in certain cells with more than one normoblast no
haemoblast nucleus remains (fig. 26); 2) that giant cells of the
yolk-sae are simply modified haemoblasts whose cytoplasm
undergoes a chemical alteration, as indicated by staining re-
actions, similar to that of haemoblasts in becoming erythro-
blasts; 3) that no multinucleated giant cells could be found in
process of fragmentation into mononucleated cells; 4) that the
cytoplasmic relationship between the two cells is frequently very
intimate; 5) that such intracellular normoblasts are occasionally
in mitosis, an unexpected phenomenon in ingested degenerating
cells; and the possibility 6) that the cells interpreted as phago-
eytic endothelia may indeed be cells differentiating normo-
blasts intracellularly while still attached to the blood vessel
wall—all indicate that the structure in question is one repre-
senting actual intracellular differentiation of normoblasts within
a giant cell. This matter will be further discussed below.
In the yolk-sac of the 25 mm. pig embryo the blood vessels
are relatively much larger. No blood-islands occur. The blood
cells are predominantly of the normoblast type; there are also
some erythroblasts and a few haemoblasts. Giant cells are
apparently lacking; and the endothelium of the blood channels
is apparently no longer capable of haemoblast formation.
IV. DISCUSSION
a. Function of yolk-sac
1) Digestive. The yolk-sac entoderm is of course continuous
with the epithelial lining of the gut through the yolk-stalk.
Originally similarly undifferentiated, the yolk-sac entoderm
already at the 5 mm. stage has far outstripped the gut entoderm
in differentiation. Even at the 10 mm. stage the cells lining
the gut are relatively little differentiated. The chief mark of
functional activity on the part of the yolk-sae entodermal cells
is the presence of a generous amount of basal filaments. Such
are lacking in the gut entoderm of this stage. These filaments
resemble very closely mitochondria; they may be long or short,
straight or variously curved, delicate or coarse, apparently
290 H. E. JORDAN
homogeneous or segmented. While structurally very like
mitochondria—on the basis of which. characters I previously so
interpreted them—TI now feel compelled to give them a different
interpretation, and for the following reasons: 1) Identical fila-
ments appear in the cells of the hepatic cords (compare figs. 31
and 32) and those of the mesonephric tubules of these embryos.
These cells are functionally active in a secretory way, strength-
ening the presumption that the filaments in the yolk-sac ento-
derm also have secretory significance. 2) If these filaments
were really mitochondria, many other cells should show such
elements, for it is well established that mitochondria are practi-
cally universally present in embryonal cells. But no other cells,
besides those mentioned, contain similar filaments in these
embryos. It is quite unreasonable to suppose that the technic
should have preserved mitochondria only in selected types of
cells. The filaments in question most probably have nothing
directly to do with mitochondria. 3) The filaments are appar-
ently identical with the ergastoplasmic filaments described by
Bensley (1) for the parenchymal cells of the pancreas of the
adult guinea pig, readily distinguishable from mitochondria
demonstrable by appropriate technics. Similar filaments have
been described for other secretory cals, as for example, salivary
glands and kidney.
On the basis of the above considerations the conclusion seems
unavoidable that these filaments in the yolk-sac entoderm are
of secretory significance. The manner in which they function
in the secretion process is uncertain, but there is some evidence
that they segment distally into granules. These filaments, then,
may be presecretion filaments. In the 25 mm. stages, filaments
are relatively less, and granules relatively more, abundant than
at the 10 mm. stage.
Similar structures have been described in the human yolk-
sac of about this same stage [Jordan (10, 11, and 12); Branca (2)].
Branea indeed interpreted them as ‘functional protoplasm.’
I first designated them by the term ‘mucinous masses,’ since
they reacted to the specific stains for mucus. In my first study
(1907) I inclined to the belief that they were degeneration prod-
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO 291
ucts. In the light of the data from pig embryos my _ subse-
quent (710) interpretation as secretory structures appears to
have been correct. The filaments have a basophilic staining
reaction, hence stain well in specific mucous dyes. In later
developmental (functional) stages they are limited to the basal
ends of the cells, where they may become clumped into a deep
staining irregularly oval mass. The ‘mucinous masses’ described
for the yolk-sac of 9 and 13 mm. human embryos are essentially
the same structure as the presecretion filaments of the 10 mm.
pig embryo; and their functional réle is most probably secretory.
What then may be the meaning of the yolk-sac entoderm in
terms of function? The additional evidence from the yolk-sac
of the pig, further supports my earlier conclusion (’07) that
this cell structure is to be interpreted in terms of the ancestral
history of higher mammals. In the ancestors with yolk laden
eggs the entodermal cells undoubtedly had the function pri-
marily of elaborating a digestive fluid for the liquefaction and
assimilation of the yolk. In yolkless umbilical vesicles, the
entoderm apparently still develops and differentiates in accord
with an ‘ancestral memory,’ though it can perform no true
digestive function. The umbilical vesicle of the pig, as of man,
is in large part—that is, as concerns digestive significance—a
vestigial structure. But it has taken on a secondary function,
now apparently become of great importance, as an early, perhaps
original, center of haemopoiesis.
The above discussion would seem to dispose of Paladino’s
(20) suggestion that the yolk-sac entoderm of higher mammals
has a hepatic function. The form and structure of the two
classes of cells are indeed very closely similar (figs. 31 and 32),
but this need not necessarily imply identity of function. The
similarity is due more probably to the fact of common origin
from the primitive gut, and the further fact that both are func-
tionally active, and in a secretory manner. Nor need the pres-
ence of glycogen in both types of cells be interpreted in terms of
functional identity, since many types of cells of embryos contain
glycogen [Gage (8)].
292 H. E. JORDAN
Neither brush borders nor terminal bars occur on these cells.
Such structures have been described for the entodermal cells
of the human yolk-sac by Branca (2). However in my own
specimens of the human yolk-sae (10, 11, and 12), I could never
convince myself of the presence of these structures. Lewis (15)
likewise was unable to find them in human yolk-saes of similar
ages.
The entodermal cells in the yolk-sac of the 10 mm. pig embryo
are undergoing extensive mitotic proliferation. This fact,
viewed in conjunction with the good cytologic preservation of
the cells as indicated primarily by the abundance and character
of the presecretion filaments, should remove all doubt as to the
normal and healthy condition of these specimens.
Not a single entodermal cell can be found in process of amitotic
division. Nor are any of these cells binucleated. This is signi-
ficant in view of the fact that all types of cells in the mesenchyma
and its derivatives show abundant examples which admit of
interpretation in terms of direct division.
2) Haemoporetic. The first question under this caption con-
cerns the origin of the angioblast. The term ‘angioblast’ is
employed here to designate the original anlage of the vascular
tissue in the yolk-sae. It is obvious that no sharp line can be
drawn between original and secondary angioblast. Suffice it
to note that angioblast is still in process of formation in the
yolk-sac of the 10 mm. embryo. No information accrues from
this study touching the question of the origin of the first mass of
vascular anlages. Once formed, angioblast can of course spread
by process of growth. However, it is also still being added to
by previously discrete moieties. If these additions can be
shown to be made from the mesenchyma, it would seem to afford
a strong presumption against the derivation of the original
angioblast from entoderm [Minot (19)].. Such anlages do arise
by differentiation within the mesenchyma in the shape of dis-
crete blood-islands, as described above. I conclude for the
mesenchymal origin of the angioblast on the basis, then, mainly
of these two observations: 1) the common origin of endothe-
lium and haemoblasts, as described above, from mesenchyma;
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO 293
2) the sharp demarcation between mesenchyma and entoderm
in the embryos here considered. Where blood vessel and ento-
derm abut, the basement membrane of the entoderm and the
endothelial cells of the vessel are never in direct continuity
(fig. 2).
The close detailed structural similarity between the meso-
thelial cells and the endothelial cells, and between the nuclei of
both and those of the mesenchyma, was noted above (figs. 15
to 20). The criteria which Clark (4) applied in the chick embryo
for the differentiation between endothelial nuclei and mesenchy-
mal nuclei are inapplicable to the yolk-sac mesenchyma of pig
embryos of the 5 to 15 mm. stages of development. Number of
nucleoli, character of nucleolar contour, and depth of tingibility
of nucleoli are not features by which mesenchyma nuclei can be
differentiated from endothelial nuclei. These are marks which
characterize different cells (probably representing different
functional phases) of mesenchyma, mesothelium and endothelium
alike.
The morphologic evidence seems to force the conclusion that
endothelium and mesothelium are both very similar differenti-
ation products of mesenchyma, the factor chiefly operative in
the differentiation being the mechanical factor of pressure, as
maintained by Huntington (9), Schulte (21) and others. The
pressure exerted upon the mesothelium operates from the rela-
tively more rapidly growing entoderm; that upon the endo-
thelium from the confined blood cells and plasma. The further
fact that haemoblasts arise from both mesenchyma and endo-
thelium supports the conclusion of their essential identity.
If the above is correct then one would expect that the meso-
thelium also could produce haemoblasts. My material yields
no data in support of this view. Indeed very careful study of
the mesothelium both of the yolk-sae and the chorion, with this
point in view, gave only negative evidence. The mesothelial
cells proliferate both mitotically and apparently amitotically
but nothing appears closely similar to the phenomena described
by Bremer (3) for the chorion of the young human embryo,
where the mesothelium is said to invaginate the underlying
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, NO. 2
294 H. E. JORDAN
mesenchyma of the body stalk in the form of cords and tubules
(angiocysts) the cells of which differentiate into haemoblasts and
endothelium. Bremer’s observations, however, are a further
very strong support to the claim that angioblast is of mesenchy-
mal origin, and that mesenchyma, mesothelium and endothelium
are originally identical structures.
The monophyletic theory of blood cell origin considers the
haemoblast the common parent of both erythrocytes and leuco-
cytes. Its correspondence with fact, at least in essential out-
lines, is now widely accepted. The point which has stimulated
most discussion concerns the origin of isolated haemoblasts
within the mesenchyma. Are such differentiation products of
the mesenchyma, or are they migrants from the blood vessels?
The latter view was held by Minot (19); Maximow (16 and 17)
and others champion the opposing view. In the case of the yolk-
sac of the pig, the evidence seems definite in favor of the in situ
differentiation of haemoblasts from the mesenchyma. The
observations both from blood-islands and single cells have been
given above. Haemoblasts of course are capable of amoeboid
activity, and undoubtedly do leave the blood vessels under
certain conditions, and invade the surrounding mesenchyma.
But that the cell (h) illustrated in figure 30 cannot be interpreted
as such is clear from: 1) the connection of its nucleus, through a
delicate chromatic bridge, with the nucleus of the mesenchyma;
and 2) its perfectly healthy condition, both from the viewpoint
of its nucleus and its cytoplasm. Nor can there remain any
doubt that it is actually a haemoblast when its cytoplasm and
nucleus, in contrast to the cytoplasm and nucleus of the mesen-
chyma, is compared with an intravascular haemoblast.
The evidence given above for the extensive origin of haemo-
blasts from the endothelium seems conclusive for the 10 mm. pig
embryo. Neither at earlier nor later stages is this process so
evident.
The haemogenic activity of the endothelium in the yolk-sac
of the pig is of cardinal significance especially in view of Stock-
ard’s (23) findings in the case of the Fundulus embryo, where the
problem was approached by the experimental method. This
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO 295
consisted in the stoppage of the embryonic circulation by means
of anaesthetics. Stockard’s observations led him to conclude
that in the Fundulus embryos investigated (up to 20 days) the
endothelium plays no haemogenic role. In the pig embryo, on
the contrary, the data leaves no escape from the opposite con-
clusion, a conclusion arrived at also by many investigators of
various embryo forms. |e.g., certain chelonians, Jordan and
Flippin (13)].. This conclusion is supported by the further
important fact that the endothelium of the sinusoids of the liver
and of the glomerular capillaries of the mesonephroi also produce
haemoblasts.
The sole alternative interpretation that has any appearance
of plausibility respecting the haemoblasts of the yolk-sac vessels
here described as separating from the endothelium, is that they
have become pressed against the wall and thus modified in shape
and caused to adhere intimately to the endothelium, so as to
stimulate endothelial continuity and derivation. This sugges-
tion is rendered inapplicable by 1) the possibility of tracing a
complete series of transition stages between a true endothelial
cell, through intermediate haemoblast stages, to a free haemo-
blast; 2) the possibility of tracing a similar series through to
multinucleated giant cells; 3) the fact that such haemoblasts in
apparent continuity with the endothelium are quite as abundant
in essentially empty vessels as in vessels crowded with blood
cells, where alone an adequate factor of pressure would seem to
prevail, and 4) that haemoblasts, though apparently naked
cells, do not in general exhibit adhesive properties except among
themselves.
b) Giant cells. The derivation and the morphologic and cy-
tologic variations of the giant cells are clear, as described above.
These cells are simply modified haemoblasts, capable of under-
going a similar differentiation into giant erythroblasts, and appar-
ently ultimately differentiating normoblasts intracellularly. This
last point may be thought perhaps to remain somewhat doubt-
ful, and even if the interpretation is accepted, the significance
and economy of this process—for it is clearly not essential, since
it is not the exclusive method for yolk-sac haemopoiesis still
remains obscure.
296 H. E. JORDAN
That the giant cells have no genetic relationship to the ento-
derm, as urged by Graf. v. Spee (22), is certain. That the
method of nuclear multiplication is largely a matter of budding
and fission is also demonstrable (fig. 35). It may be stated also
that these cells are much more abundant at about the 10 mm.
than at earlier and later stages; and that while all the other
types of erythrocytes are found in the embryonie circulatory
system, giant cells are practically limited to the yolk-sace vessels.
A few smaller varieties appear in the liver and the mesonephroi,
and occasionally one appears in a capillary in the mesenchyma
next the brain. Haemoblasts also are only sparingly found
outside of the yolk-sac, liver, and the glomerular sinusoids of
the mesonephroi. The normoblasts are in the vast majority in
the intraembryonie circulatory system.
On the basis of their occasional normoblast content the giant
cells might be interpreted 1) as erythrophages or 2) as multiple
erythroblasts. The latter interpretation was urged by Graf
v. Spee (22). The former interpretation is supported by the
fact that endothelial cells—which are potential haemoblasts—
may apparently function as phagocytes for erythroblasts. The
latter more plausible conclusion rests upon my observations
that in a giant cell with two normoblasts (fig. 26) no additional
nucleus is present; and the further fact that frequently the cell
membrane of the normoblast is not separated from the eytoplasm
of the giant cell by any space, but the two structures appear
continuous (fig. 26). Moreover in certain multinuclear giant
cells the several nuclei and their enveloping cytoplasmic areas
are at different stages of development. In figure 25 two of the
nuclei are typical haemoblast nuclei, two are typical erythroblast
nuclei. The upper right hand nucleus (x) is differentiated more
than the other, and the enveloping cytoplasm is beginning to
take on normoblast characteristics. Nevertheless this inter-
pretation must perhaps still be regarded as more or less tentative.
But the fact that mega- and polykaryoecytes are present in all
haemopoietic foci, embryonic, foetal and adult, strongly supports
the conclusion that they are closely associated with the haemo-
poietic process. As such, however, their function does not seem
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO 297
to be an essential one; they may represent simply atypical or
possibly ancestral phenomena. Erythrocytes commonly de-
velop from mononuclear haemoblasts; binucleated haemoblasts
apparently sometimes divide to form two haemoblasts (fig. 3, e);
multinucleated haemoblasts (polykaryocytes) do not break up
into mononuclear haemoblasts, but may produce erythrocytes
(normoblasts) intracellularly.
V. SUMMARY
1. In pig embryos of about 10 mm. length, the yolk-sac
attains its highest stage of progressive histologic differentiation.
This statement pertains both to the entoderm and to the angio-
blast.
2. The entodemal cells are characterized chiefly by abundant
presecretion filaments, in which feature they agree with the
cells of the liver and mesonephroi.
3. Angioblast arises from the mesenchyma.
4. The mesothelium of the yolk-sae of pig embryos between
5 and 12 mm. does not produce haemoblasts. Nor is there any
satisfactory evidence that the mesothelium of the body stalk
and chorion function to this end.
5. The mesenchyma may differentiate directly into endothe-
lium or into haemoblasts.
6. Haemoblasts arise extensively at the 10 mm. stage from
the endothelium of the yolk-sac blood vessels. The endothelia
of the hepatic sinusoids and mesonephric glomeruli of this stage
also show extensive haemopoietic capacity.
7. Giant cells, both mono- and polynuclear, are abundantly
present in the yolk-sac only at about the 10 mm. stage of develop-
ment. They may arise from endothelium or directly from haemo-
blasts. They are giant haemoblasts, and apparently function
as multiple erythroblasts in which normoblasts differentiate
intracellularly.
8. The several stages in haemopoiesis, represented successively
by haemoblasts, erythroblasts and normoblasts, with transition
stages, are abundantly present in the yolk-sac of embryos from
5 to 15 mm.
298
(17)
(18)
H. E. JORDAN
LITERATURE CITED
Benstey, R. R. 1911 Studies on the pancreas of the guinea pig. Am.
Jour. Anat., voll. 12, mo: 3.
Branca, A. 1908 Recherches sur la vesicule ombilicale de Vhomme. Ann.
de Gynéc. et Obst., Paris, T. 2, p. 577.
Bremer, JoHn Lewis 1914 The earliest blood-vessels in man. Am.
Jour. Anat., vol. 16, no. 4.
Crark, Exuior R. 1914 On certain morphological andstaining characteris-
tics of the nuclei of lymphatic and blood-vascular endothelium and of
mesenchymal cells, in chick embryos. Anat. Rec., vol. 8, no. 2, pp.
81-82.
DantscHakorr, W. 1908 Untersuchungen iiber die Entwicklung des
Blutes und Bindegewebes bei den Végeln. I. Die erste Entstehung
der Blutzellen beim Hiihnerembryo und der Dottersack als blut-
bildendes Organ. Anat. Hefte, Bd. 37.
Dantscuakorr, W. 1910 Uber den Entwicklung der embryonalen Blut-
bildung bei Reptilien. Anat. Anz., Bd. 37 (Erginzungsheft).
Emmet, Vicror EK. 1914 Concerning certain cytological characteristics
of the erythroblasts in the pig embryo and the origin of non-nucleated
erythrocytes by a process of cytoplasmic constriction. Am. Jour.
Anat. vol.; 16, no; 2.
Gace, 5S. H. 1906 Glycogenin a 56-day human embryo and in pig embryos
of 7to 70mm. Am. Jour. Anat., vol. 5, no. 2; Proc. Am. Assoc. Anat.,
Due lise
HuntTInGcron, GrorGce 8. 1914 The development of the mammalian
jugular lymph-sac, ete. Am. Jour. Anat., vol. 16, no. 2.
Jorpan, H. E. 1907 The histology of the yolk-sac of a 9.2 mm. human
embryo. Anat. Anz., Bd. 31, nos. 11 u. 12, p. 291.
Jorpan, H. E. 1910 A further study of the human umbilical vesicle.
Anat. Rec., vol. 4, no. 9.
Jorpan, H. E. 1910 A microscopic study of the umbilical vesicle of a
18mm. human embryo, with special reference to the entodermal tubules
and the blood islands. Anat. Anz., Bd. 37, no. 1.
Jorpan, H. E. 1915 Haemopoiesis in the yolk-sae of the pig embryo.
Proc. Am. Assoc. Anat., Anat. Rec., vol. 9, no. 1.
Jorpan, H. E. ano Furpprn, J. C. 1913 Haematopoiesis in Chelonia.
Folia Haematologica, Bd. 15.
Lewis, Freperick T. 1912 Chap. XVII, Human embryology, Ixeibel and
Mall, vol. 2, p. 318. Lippincott Company.
Maximow, A. 1909 Untersuchungen iiber Blut und Bindegewebe. I.
Die frihesten Entwicklungsstadien der Blut- und Bindegewebszellen
beim Saiigetierembryo, etc. Arch. f. mikr. Anat., Bd. 73, p. 444.
Maxrmow, A. 1910 Uber embryonale Entwicklunge der Blutzellen bei
Selachiern und Amphibien. Anat. Anz., Bd. 37 (Ergiinzungsheft.)
Mryper, ArtHur W. 1904 On the structure of the human umbilical
vesicle. Am. Jour. Anat. vol.3.
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO 299
(19) Minor, CHartes S. 1912 Chap. XVIII, Human Embryology, Keibel and
Mall, vol. 2. Lippincott Company.
(20) Panapino, G. 1901 Contribuzione alla conoscenza sulla struttura e
funzione della vesicola ombelicale nell’uomo eneimammiferi. Arch.
Ital. Ginecol., Napoli, vol. 8, p. 127.
(21) Scuuutze, H. von W. 1914 Early stages of vasculogenesis in the cat (Felis
domestica) with especial reference to the mesenchymal origin of
endothelium. Memoir, Wistar Inst. Anat. and Biol., Philadelphia.
(22) Sper, Grar V. 1896 Zur Demonstration iiber die Entwicklung der Driisen
des Menschlichen Dottersackes. Anat. Anz., Bd. 12, p. 76.
(23) SrockarpD, CHARLES R. 1915 An experimental study of the origin of blood
and vascular endothelium in the Teleost embryo. Proc. Am. Assoc.
Anat., Anat. Rec., vol.9,no.1. (Complete paper in Am. Journ. Anat.
18:2 and 3; and as Memoir, No. 7, of The Wistar Institute of Anatomy
and Biology.)
PLATE 1
EXPLANATION OF FIGURES
(Unless otherwise specified the illustrations are from a single specimen of the
10 mm. stage, the magnification 1000, the fixation with Zenker’s fluid, and the
stain employed the haematoxylin-eosin combination).
1 A group of blood cells from one of the larger yolk-sac vessels of a 6 mm.
pig embryo (Helly’s fixation; Giemsa stain; magnification, 1500 diameters).
a) various types (differentiation stages) of haemoblasts; the sparse naked cyto-
plasm has a vague irregular granular character and stains intensely blue; the
large vesicular nucleus stains a very light blue and contains a delicate, finely
granular reticulum and one or several spheroidal or irregular nucleoli staining
like the chromatic granules, a bluish orange. b) Young erythroblasts (‘megalo-
blasts’); the nucleus is relatively smaller and the cytoplasm more voluminous
than in the smaller younger haemoblasts; the cytoplasm stains a light blue
(brownish gray or bluish pink) and contains fine, uniform, spherical granules
(probably haemoglobin) ; a cell wall is distinet; the still vesicular nucleus contains
a coarsely granular reticulum which stains blue; some of these nuclei still contain
a nucleolus. ¢) Older erythroblasts; the nucleus has become still smaller and
more chromatic; the homogeneous cytoplasm is relatively more voluminous and
now stains pink. d) Normoblast; the nucleus is small, granular and chromatic;
the cytoplasm stains brilliant red (in Zenker fixed tissue the cytoplasm consists
merely of a coarse irregular unstainable reticulum enclosed by a robust cell
membrane. )
2 Narrow portion of wall of yolk-sac including all of its layers. EH, entoderm;
the cells contain many presecretion filaments. Between the entoderm and periph-
eral mesenchyma is a large blood vessel containing a few blood cells at various
stages in the metamorphosis into a normoblast (e);a) endothelial cell; b) haemo-
blast; ¢) binucleated haemoblast with long pseudopod; d) binucleated erythro-
blast. MM, mesothelium; bm., basement membrane; end., endothelium.
3 A group of developing blood cells from a yolk-sac blood vessel. a and b)
young haemoblasts; ¢) haemoblast with nucleus in process of amitotie division;
d) binucleated haemoblast; e) binucleated haemoblast in process of cytoplasmic
amitotic constriction, a fairly common form of cell; f) erythroblast (Maximow’s
‘megaloblast’); ¢) normoblast.
4 Portion of wall of yolk-sae of 10 mm. pig embryo including mesothelium,
endothelium and the intervening mesenchyma. a) endothelial cells from wall
of a blood vessel; b) endothelial cell in early stage of separation from wall of
blood vessel to become a haemoblast;c¢) later stage; d) extravascular haemoblast,
separating from the endothelium.
5 Haemoblast at late stage in process of separation from the endothelium.
6 Haemoblast, of spindle shape, just about to separate from the endothelium.
7 Uninucleated giant cell; large haemoblast.
Trinucleated haemoblast (giant cell) in final stage of separation from the
endothelium (e).
010)
9 Trinucleated giant cell, immediately after separation from endothelium.
Note the lateral basal projections, the points of, final separation. One of the
nuclei is apparently undergoing amitosis.
10 Binucleated haemoblast in which one of the nuclei and the surrounding
cytoplasm have differentiated into a normoblast.
llandi2 Trinucleated giant cells.
300
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO PLATE 1
H. E. JORDAN
PLATE 2
EXPLANATION OF FIGURES
13 Haemoblast (A) in final stage of separation from endothelium; e, endo-
thelial cell.
14, 15 and 16 Three successive stages in the transformation of an endothelial
cellinto a haemoblast. #, towards entoderm; V, towards blood vessel.
17 Nucleus of endothelial cell in phase of amitotic division. Many nuclei
also can be seen in mitosis.
1S Nucleus from mesenchyma. Note the similarity between nuclei of endo-
thelium, mesothelium and mesenchyma.
19 and 20 Two mesothelial cells. s, towards surface. Occasional cells can
be seen in mitosis.
21 Haemoblast in mitosis. The spindle is apparently tripolar. Such irreg-
ular mitoses if sufficiently common would explain the multinuclear haemoblast
with nuclei of various sizes. Haemoblasts apparently divide both mitotically
and amitotically.
22 Haemoblast with nucleus apparently dividing amitotically.
23 Large binucleated giant cell; the cytoplasm is at an early phase of differ-
entiation into the erythroblast type; the nuclei also are in early, but different,
stages of differentiation.
24 Smaller binucleated giant cell (haemoblast); the nuclei are of the typical
haemoblast type.
25 Giant cell from yolk-sac of 10 mm. pig embryo with four nuclei, which,
with their enveloping cytoplasm, are at different stages of differentiation. Two
of the nuclei have haemoblast characters, one erythroblast and one (x) early
normoblast characters. The cytoplasm also around x has normoblast char-
acteristics.
26 Binucleated haemoblast (giant cell) in late stage of process of direct
intracellular differentiation into two normoblasts.
27 Haemoblast with one long and several shorter stubby pseudopods.
28 Endothehal phagocytic cell (perhaps a_ differentiating haemoblast)
having ingested an erythroblast whose nucleus is undergoing karyorrhexis, the
cytoplasm appearing normal.
29 Portion of wall of yolk-sac of 10 mm. pig embryo showing a small blood
island. Thecellsare allof the early haemoblast stage, and closely related periph-
erally to the surrounding mesenchyma, from which they have apparently
differentiated. One haemoblast is binucleated. H, entoderm, schematically
represented; V, blood vessel.
30 Portion of wall of yolk-sac of 10 mm. pig embryo showing the differenti-
ation of a haemoblast (h) from the mesenchyma. “The nucleus of the definitive
haemoblast is still connected through a chromatic nuclear strand with the nucleus
of its sister mesenchymal cell. #, entoderm, schematically represented; V,
blood vessel; mes., mesothelium.
31 Four adjacent entodermal cells to show especially the ‘basal’ or presecre-
tion filaments.
32 A group of four adjacent liver cells from the same embryo, to show the
close similarity in nuclear and cytoplasmic structure and form between the
hepatic and yolk-sae embryonic epithelium. Many of the hepatic cells (not
here represented) show mitotic figures; amitotic divisions apparently do not
yet occur.
33, 34 and 35 Various types of giant haemoblasts. Figure 35 is typical of a
large group of giant cells whose nuclei proliferate amitotically.
302
ERYTHROPOIESIS IN YOLK-SAC OF PIG EMBRYO PLATE 2
H. E. JORDAN
1 AY §
EFFECTS OF INANITION UPON THE STRUCTURE
OF THE’ THYROID “AND PARATHYROID
GLANDS OF THE ALBINO RAT
C. M. JACKSON
Institute of Anatomy, University of Minnesota, Minneapolis
FOURTEEN FIGURES
CONTENTS
LEGO CLULC U1 O Lease ett ee ce ae RN ena ll
Material and methods................ 0.0... cc cece cece eee eee Bt eee UO
iheathvnoidsc:lonceew ee ee ee ee en 5 SoM acaak ee eee . 307
a. Normal strue as of the chenord a PHOKG Lae eee ONO ROMO es a cy “HONS
b. Structure in young rats held at reaneenPMIce ae EEE 3e 2c. hee LD
ce. Structure in adult rats after acute and chronie inanition........... S22
d. Discussion and conclusions................:.....s..:: 324
The parathyroid gland......................: co RS hn 52 os, OO
a. Normal structure of the pagai engi BI es pst By
b. Structure in young rats held at meeevemnice eS Reet Mea a ae 342
ce. Structure in adult rats after acute and chronic inanition.............. 344
d. Discussion and conelusions..................... Suds le eC 345
ho}611 010.0121 ey (ene ee cas Pee CEs gE erg are See Eee 348
leTGeraGunerCited aq etek oe eres chat eae oe es a eee 30)
INTRODUCTION
The thyroid gland presents an interesting and difficult bio-
logical problem. Although the morphology of the thyroid has
been extensively studied, there are still many doubtful and un-
settled questions concerning its development and its normal adult
structure. Even more uncertainty exists concerning its physio-
logical significance and its pathological changes. Some light
may be thrown upon the various phases of this problem by a
study of the changes produced in the thyroid gland by inanition.
In previous papers (Jackson 715 a, 715 b) it was shown that
during inanition the various organs of the albino rat suffer very
unequally in loss of weight, the loss also varying according to
305
306 C. M. JACKSON
the length and character of the inanition. In acute inanition of
adult rats, with loss of about one-third in body weight, the thy-
roid gland had apparently lost but little if any in absolute weight;
in chronic inanition of adults, the average apparent loss of the
thyroid was about 22 per cent; while in young rats held at main-
tenance (constant body weight) by underfeeding for several
weeks the average loss was about 24 per cent. In order to de-
termine what histological changes are correlated with these
changes in gross weight, material was preserved for further
study. On account of its intimate association with the thyroid,
the parathyroid gland was also included in this investigation.
The results are presented in the present paper, which is the third
of a series of studies upon the effects of inanition. The work is
being carried on with the assistance of a special grant from the
research fund of the Graduate School of the University of Min-
nesota.
MATERIAL AND METHODS
The material used included the thyroid (and included para-
thyroid) glands of the albino rat (Mus norvegicus albinus) from
previous studies (Jackson “15a, 715 b), together with some col-
lected since. In all, more than 50 normal glands were sectioned
and studied, varying in age from newborn to adult (15 months).
These were chiefly controls from the same litters as those used
for experiments (including several controls used in experiments
by E. R. Hoskins and C. A. Stewart). In addition, I am in-
debted to Professor Bensley and Mr. Burgett, of the University
of Chicago, Professor Addison, of the University of Pennsylvania,
and Professor Evans, of the University of California, for mate-
rial kindly furnished in order to investigate possible local varia-
tions in normal thyroid structure of the rat.
Of the animals subjected to inanition, the thyroid and para-
thyroid were obtained from 14 of the younger rats held at main-
tenance for various periods (chiefly beginning at 8 weeks and
ending at 10 weeks of age). Of the adult rats, 6 glands were
INANITION OF THYROID IN RATS 307
studied from those subjected to acute inanition, and 3 from those
with chronic inanition.
The material was obtained at the autopsy held immediately
after the animals were killed, and was fixed chiefly in Zenker’s
fluid, 12 to 24 hours. In a few cases formalin or Flemming’s
fluid was used, but the results were less satisfactory.
The glands were embedded in paraffin, and cut at 5 micra
(occasionally 7 to 10 micra) in thickness. In the great majority
of cases, the glands were cut and mounted in complete serial
sections. This was found to be important, not only to make
certain of including the parathyroids, but also because the struc-
ture frequently varies in different parts of the thyroid and para-
thyroid glands.
The sections were stained in most cases with haematoxylin and
eosin; In a few cases with iron-haematoxylin, safranin, Mallory’s
anilin-blue connective tissue stain, ete.
All of the drawings were made with a Zeiss 2 mm. 1.30 N. A.
apochromatic objective and compensating ocular No. 6, with
the aid of a camera lucida. A wheel-micrometer eyepiece was
used for the measurements. It was the original intention to
measure systematically a large number of cells and nuclei in the
glands of the controls and of the animals subjected to inanition.
On account of the great irregularity in the size and shape of cells
and nuclei, however, it was found that the results, though of
limited value, could not be obtained with sufficient accuracy to
justify any very extensive series of observations. Therefore the
number of measurements was restricted to that judged sufficient
to give merely an approximation of the apparent average and
range observed.
THE THYROID GLAND
The normal structure of the thyroid gland in the albino rat
at various ages will first be considered. Then the changes found
in the various types of inanition will be deseribed and their sig-
nificance discussed.
308 CG. M. JACKSON
a. Normal structure of the thyroid gland
The general form and topography of the thyroid gland in the
albino rat is shown in cross section in figure 1. Each lateral
lobe presents the typical relations—convex external surface coy-
ered by the infrahyoid musculature; concave internal surface in
contact with the lower larynx and upper trachea; and narrower
posterior surface (or border) in relation with the oesophagus
Fig. 1 From a photograph (retouched) of a cross section of the thyroid gland
in situ at the level of the isthmus, showing relations to infrahyoid musculature,
uppermost trachea, oesophagus, ete. One parathyroid is visible, on the left side
of the figure. From albino rat No. S 9.47, age 22 days, gross body-weight
25.5 grams. (X 28.)
medially the carotid artery, ete., laterally. The isthmus is fre-
quently a very thin somewhat fibrous band, almost invisible
when the fresh gland is exposed in situ. It invariably contains
thyroid follicles (contrary to Sobotta 715), although these may
become scattered and more or less atrophied in adult rats.
The minute structure of the normal thyroid gland at 3 weeks
(the age when the experiments began with the younger rats) 1s
shown with slight magnification in figure 1, and under high power
in figure 2. No attempt will be made to describe in detail the
INANITION OF THYROID IN RATS 309
minute structure of the gland, but some of the essential features
especially affeeted by inanition will be considered briefly.
Fig. 2 A small portion of the thyroid gland shown in figure 1 (rat 8 9.47, age
22 days) magnified to show the details of the normal histological structure. Sev-
eral follicles containing colloid are shown. Follicular epithelium cuobidal; cy-
toplasm abundant and granular, with a few scattered vacuoles. Apparent ori-
gin of the colloid ‘vacuoles’ from the epithelial cells is shown in a few places.
Four interfollicular epithelial cells are seen between the two lower follicles.
Fibrous stroma scanty, with rich blood-vascular plexus. ( 750.)
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, NO. 2
310 C. M. JACKSON
The follicles at this time are chiefly oval or rounded in outline
and vary mostly from 20 to 70 micra in diameter. The larger
follicles are rather infrequent and somewhat uniformly scattered,
but are usually more frequent near the surface of the gland.
Those shown in figure 2 are of average size.
The finer structure of the thyroid gland is shown in figure 2.
The cells of the follicular epithelium are approximately cuboidal
in form (in some cases low columnar; in others, especially in the
larger peripheral follicles, somewhat flattened). In height, they
range chiefly between 8 and 15 micra, the average being 10 to
12 micra. The inner and outer cell walls are sharply distinct;
the intercellular boundaries are less distinct and sometimes
absent.
The cytoplasm of the follicular cells (fig. 2) is filled with mod-
erately fine granules, reddish violet in color (with Zenker fixation
and haematoxylin-eosin stain). The granules are usually some-
what uniformly distributed. They are not densely packed, but
are sometimes arranged so as to give an indefinite reticular form,
apparently intermingled with small, clear vacuoles. Some of
these vacuoles (though not all) may correspond to the minute
fat droplets or granules described in the thyroid cells by Erd-
heim (’03) and Traina (’04). The cells present a fairly uniform
appearance, and there is nothing to indicate any division into
the ‘chief’ and ‘colloid’ cell types of Langendorff.
The nuclei of the follicular cells are spherical or slightly ovoidal
(ellipsoidal) in form, the diameters varying from 4 to 7 micra,
usually 5 or 6 micra. The nuclear membrane is distinct and
stains deeply. There are usually one or two larger nucleoli
(karyosomes) and several smaller granules; and a fine, paler nu-
clear network, often indistinct, with a very pale bluish, homo-
geneous nuclear background, corresponding to the nuclear sap
(karyolymph). The nuclei shown in figure 2 are typical, though
in the larger peripheral follicles with shghtly flattened cells the
nuclei may also be somewhat more flattened and slightly hyper-
chromatic. Cells in mitosis are relatively frequent, 5 having
been noted in one entire cross-section of one lobe, and 6 or 8 in
another.
INANITION OF THYROID IN RATS gael
The colloid (fig. 2) appears typical in form though somewhat
variable in staining reactions. In some cases it fills the follicu-
lar cavity completely, in other cases it is retracted somewhat
with either smooth or serrated margin. This retraction is prob-
ably an artefact in most cases, due to shrinkage produced by
the reagents used. But I cannot agree with those investigators
who explain the vacuoles (some of which are shown in figure 2)
in a similar manner. These vacuoles are usually small (4 micra
or less) and spherical in form, and are most frequent near the
surface of the colloid. Occasionally they are found in intimate
relation with the adjacent cells, from which they are apparently
extruded, as shown in the follicle on the right in figure 2. They
are probably connected in some way with the process of colloid
formation, as described by Anderson (794) and Miiller (’96).
Desquamated epithelial cells, which sometimes dissolve leaving
clear vacuoles in.the colloid in older rats, are extremely rare at
this stage.
A variable amount of interfollicular epithelium appears, which
sannot be distinguished from tangential sections of follicles, ex-
cept in serial sections. In structure, these interstitial epithelial
cells are similar to those of the follicles. .A few appear in the
lower part of figure 2.
The interfollicular connective tissue forms a delicate fibrous
stroma (fig. 2), relatively small in amount, but containing a
rich capillary plexus of blood-vessels. The nuclei visible are
mostly of capillary endothelium. They are elongated or flat-
tened in form, and stain somewhat deeply.
At 10 weeks (the age when the inanition experiments ended in
most of the younger rats) the thyroid gland has normally under-
gone but slight changes, the structure being essentially the same
as that just described at three weeks. Therefore no detailed
figures are considered necessary. A photograph with low mag-
nification is shown in figure 3, representing a cross section of one
lateral lobe. The follicles have increased somewhat in size,
the maximum diameter now reaching about 100 micra. The
larger follicles are more frequent, and are sometimes rather uni-
formly distributed, as shown in figure 3, though very frequently
ou Cc. M. JACKSON
5
Fig. 3. From a photograph of one lobe of the thyroid gland from normal al-
bino rat No. 8 5.3, age 74 days, gross body-weight 172 grams. Parathyroid in-
cluded. Compare with figures 4, 5 and 6, representing thyroids in rats of the
same age, but held at maintenance from age of 3 weeks. (X 28.)
Fig. 4 From a photograph of one lobe of the thyroid gland from albino rat
No. 8 11.63, age 72 days, gross body-weight 23.8 grams (held at maintenance from
age of 3 weeks). Follicles much larger at periphery. Some extra-capsular tis-
sue is included. Parathyroid relatively large. (X 28.)
INANITION OF THYROID IN RATS ie
there is a distinct tendency to larger follicles in the superficial
layers.
In finer structure, the thyroid at 10 weeks is very similar to
that shown at 3 weeks (fig. 2). The height of the follicular
cells varies considerably, however. While the maximum is about
the same as at 3 weeks, the average (about 8 to 10 micra) is some-
what lower. In other words, the cells are usually more flattened,
especially toward the periphery of the gland. In the larger sur-
face follicles, the cells are usually distinctly flattened, 6 micra or
less in height. The nuclei in these cells are also correspondingly
flattened, their diameters averaging 4 x 6 micra. The nuclei in
general are similar to those at 3 weeks in size and structure, usu-
ally nearly spherical in form, and averaging about 6 micra in
diameter. Mitosis is very much less frequent in the cells of
the thyroid at 10 weeks than was found in the gland at 3 weeks.
While the typical normal structure of the thyroid cells at 10
weeks is like that at 3 weeks (fig. 2), cells of abnormal appear-
ance are also found. These atypical forms vary from shght
modifications up to marked cellular degenerations, and should
be carefully noted in order to avoid confusion with the changes
during inanition to be described later.
The flattening of the epithelium in the larger peripheral fol-
licles is almost constant, though variable in extent. The flat-
tened nuclei are hyperchromatic in type, and often present a
more or less deeply-staining, homogeneous background, which
may obscure or obliterate the nuclear network and granules.
The cytoplasm is reduced in amount, more deeply-staining, and
often somewhat homogeneous in appearance. This type of cell
occurs so constantly in the peripheral follicles (and occasionally
elsewhere) that it can hardly be considered abnormal. I inter-
Fig. 5 From a photograph of one lobe of the thyroid gland from albino rat
No. 8 5.10, age 67 days, gross body-weight 22.7 grams (held at maintenance from
age of 3 weeks). Gland small, with relative increase of stroma. Parathyroid
included. (xX 28.)
Fig. 6 From a photograph of both lobes and a portion of the isthmus of the
thyroid gland from albino rat No. 8 11.64, age 73 days, gross body-weight 24.2
grams (held at maintenance from age of 3 weeks). tollicles larger at periphery
but irregular, many degenerated. Parathyroids relatively large. (X 28.)
314 Cc. M. JACKSON
pret it as an atrophic type, due perhaps largely to the pressure
on the gland from adjacent organs. Although these follicles are
filled with dense, deeply-staining colloid, it is unlikely that the
flattening is due entirely to consequent endofollicular pressure,
as follicles are sometimes seen in which the epithelium on the ex-
ternal surface is much more flattened than that on the inner
aspect of the follicle. The socalled ‘colloid’ cells of Langendorff
(frequently described by various authors) probably belong to
this atrophic type, and have no specific functional significance.
In addition to these peripheral flattened atrophic cells, men-
tion must be made of more advanced types of degeneration, al-
though the latter appear much less frequent in the young rat at
10 weeks than in older animals. These degenerative cells may
occur in any part of the gland, either singly or involving an en-
tire follicle (occasionally a regional group of follicles). The de-
generating cells may remain in the follicular wall or may be
desquamated into the follicular cavity. In rare cases the desqua-
mated epithelium may replace the colloid with an irregular mass
of cells in various stages of degeneration.
In the degenerating cells, the cytoplasm loses its typical light
granular structure and becomes vacuolated and reticular in ap-
pearance, later disintegrating into irregular, usually deeply-stain-
ing (eosinophile) masses. The nucleus may be hypochromatic
(karyolytic) in type, but more frequently presents various grades
of pyenosis (rarely karyorrhexis), especially in the desquamated
cells.
As to the frequency with which these degenerative types of
cell occur in the thyroid of (apparently) normal rats at 10 weeks,
it may be stated that of 9 glands carefully examined in serial
sections, one showed rather extensive degeneration, one a well
marked area (much less extensive), and four showed traces or
small areas in early stages of degeneration. Thus in a majority
of the glands, at least slight traces of degeneration could be
found, even in normal, apparently perfectly healthy animals.
In older rats (from 3 to 15 months of age), the normal struc-
ture of the thyroid gland is essentially similar to that described
for the younger rats. The follicles average slightly larger, the
€
INANITION OF THYROID IN RATS Abs)
maximum sometimes reaching 150 micra, the largest follicles be-
ing frequently found at the periphery of the gland. While cubi-
eal cells still occur to a variable extent, the average cell is some-
what more flattened than at 10 weeks, especially in the larger
peripheral follicles. The colloid is variable in appearance. Oc-
easionally the follicular content may be clear and unstainable.
The stroma remains as previously described.
The most striking difference in the thyroid of the older rats is
found in increasing prevalence of the degenerative process al-
ready described as appearing at 10 weeks. In the older control
rats it was found not only more frequently, but more pronounced
in character. In extreme cases, the follicles in some regions (see
fig. 10) are entirely obliterated, being replaced by masses of epi-
thelial cells with irregularly disintegrated cytoplasm and nuclei
in various stages of pyenosis or karyorrhexis. Among 20 glands
from apparently normal older rats, only 3 thyroids appeared en-
tirely normal; 6 showed slight degeneration, 5 were moderately
involved, while in 6 the degeneration was extensive, involving
the larger portion of the gland. The significance of this appear-
ance of degeneration in normal animals will be discussed later.
b. Structure of thyroid gland in young rats held at maintenance
The appearance of the thyroid gland in young rats held at
constant body-weight from 3 to 10 weeks of age is shown under
low magnification in figures 4, 5 and 6, and more highly magni-
fied in figures 7, 8 and 9. The follicles in general appear in av-
erage size smaller than the normal at 10 weeks, although it is
somewhat difficult to make exact comparisons, on account of the
irregularity in their size. The maximum diameter found in 9
normal glands was about 100 micra, whereas in 14 maintenance
rats it was 85-90 micra. The average in both cases was of course
much lower, as follicles are found of all sizes down to those with
only a minute cavity. Although in the maintenance rats at LO
weeks a few follicles are larger than found in the normal rat at
the age of 3 weeks (where the maximum was 70 micra), it 1s
doubtful whether the average follicle is any larger. It is proba-
bly shghtly smaller.
316 Cc. M. JACKSON
In size, the epithelial cells of the thyroid follicles in the young
rats held at maintenance are variable but distinctly smaller than
normal. Even the largest cells found rarely reach the average
Fig. 7 A portion of the same thyroid gland shown in figure 4 (rat No. 11.63,
maintenance from 21 to 72 days of age), magnified to show details of histological
structure. The area represented shows the hypochromatic (incipient karyo-
lytic) type of nuclear structure, which is relatively infrequent. Cells reduced in
average height (ef. fig.2). Cytoplasm reduced in amount and vacuolated in
structure. Stroma here normal. (x 750.)
INANITION OF THYROID IN RATS 317
normal height (8-10 micra) at corresponding age. The average
height in the maintenance rats is about 6 to 7 micra. Since the
Fig. 8 | 46 50E—) 0.22 40. (For lettering, see fig. 3.
third tho-
)
STUDIES ON THE MAMMARY GLAND 363
lel to the surface of the skin. In most cases observed the pri-
mary duct extends only a short distance until it divides into two
branches (secondary ducts) nearly equal in size.
The extent of the primary duct varies considerably. For in-
stance, in the first thoracic and the last inguinal glands the pri-
Fig. 5 Internal view of a wax model reconstructed from the left first inguinal
gland of a newborn albino rat. X 40. (For lettering, see fig. 3.)
mary ducts present a rather extensive course before dividing,
while in the remaining glands they divide almost immediately
after making a sharp turn in the tela subcutanea. In figure 4
(last thoracic gland) the primary duct is seen to divide into three
branches. This is an exception to the general rule that the pri-
mary duet divides into two branches.
364 J. A. MYERS
As compared with the primary duct, the secondary ducts pre-
sent a rather extensive course, after which they break up each
into two or more branches (tertiary ducts). It will be noticed
that at birth (figs. 3 to 6) the terminal branches of each tertiary
duct vary from one to three in number. On the end of most
terminal branches is a small bud-like enlargement. These en-
largements were described as true alveoli by earlier investigators,
but this was found later to be incorrect. Billroth (according to
Berka 711) doubts whether completely formed end-vesicles occur
in young human virgins. While he called the terminal enlarge-
ments ‘real end-vesicles,’ yet he adds that they later develop
into ‘true end-vesicles’ and further multiply during pregnancy.
Berka (11) states that true alveoli donot occur in young (human)
virgins. Similarly the terminal enlargements found on the milk-
ducts of young rats are not true alveoli, but are only enlarged
erowing processes corresponding to the end-buds found in other
developing glands. The microscopic structure of these enlarge-
ments and the development of true alveoli will be discussed in
a later paper dealing with the histology of the mammary gland.
The question often arises as to whether the ducts of glands
branch dichotomously or otherwise. From the various figures it
will be seen that the more proximal parts of the terminal seg-
ments usually follow the dichotomous method, but the distal por-
tions, as stated above, may terminate as a single duct or divide
into two or three branches. In the last thoracic gland (fig. 4)
the secondary branches approach true dichotomous division.
Anastomoses occur between ducts, but they are not very fre-
quent in the newborn rat. In the reconstructions made from
glands at birth, only two distinct anastomoses occur (fig. 3).
However, others have been observed in cleared preparation at
the same stage.
It will be noted that along the secondary and tertiary ducts
numerous lateral buds occur (figs. 8 to 6). Many more of them
are present on the distal than on the proximal ducts. Such buds
later form collateral branches destined to develop into ducts
similar to those already present. This point will be more clearly
brought out in the older stages.
STUDIES ON THE MAMMARY GLAND 365
Fig. 6 Internal view of a wax model reconstructed from the left second in-
guinal gland of a newborn albino rat. > 40. (For lettering, see fig. 3.)
366 J. A. MYERS
A point which has been discussed at some length recently and
one which has proved to be of considerable importance in experi-
mental work is that of the variation in the relative size and de-
velopment of the various glands in the same individual, and of
glands from different individuals of the same age. Lane-Clay-
pon and Starling (06) in working on the growth and activity of
the mammary gland concluded that. breast hyperplasia of preg-
nancy is caused by chemical substances formed in the embryo.
Such substances passing through the placenta into the maternal
blood-strezm cause growth of the mammary gland. ‘To decide
definitely as to just what tissues cause this growth Lane-Claypon
and Starling injected extracts of placenta, placenta and uterus,
ovaries, fetus, fetus together with the placenta and membranes,
and mucous membrane of the uterus into virgin rabbits. Some
of the extracts when injected caused very little apparent change
in the size of the mammary gland of virgins, while others (fetus
extract, for example) seemed to cause a marked development of
the glands.
Frank and Unger (11) in repeating certain of Lane-Claypon
and Starling’s experiments obtained different results, and further-
more found that their own series of experiments did not show
uniform results. Thus they concluded that some disturbing fac-
tor remained to be accounted for, so they decided to study more
carefully the anatomy and the physiology of the normal mam-
mary glands of the rabbit. For such study they selected a num-
ber of apparently virgin adult female rabbits and under the
necessary precautions removed a mammary gland from each.
At various intervals of time other mammary glands from the
same animal were removed and studied. From these experi-
ments Frank and Unger were able to demonstrate in virgin rab-
bits changes which were indistinguishable from those seen at the
end of the first third of pregnancy. Thus some physiological
factor must be involved. Frank and Unger found a partial ex-
planation for this condition in an article by Bouin and Ancel
(09) who deseribe variation in the size and appearance of the
rabbit’s mammary gland corresponding to the development of
the corpus luteum. A little later O’ Donoghue (12) showed that
STUDIES ON THE MAMMARY GLAND 367
there is a decided change in the structure and size of the mam-
mary glands of Dasyurus viverrinus when ovulation is not suc-
ceeded by pregnancy.
A comparison of the individual glands of the rat at birth and
at two weeks (figs. 2 to 6) will show that there is considerable dif-
ference in the size and development of the various glands in the
same rat, sometimes even in the same pair of glands (fig. 2).
It has also been observed that corresponding glands from differ-
ent rats of the same age and approximately equal weights show
considerable variation in size and complexity of structure. The
differences in size and development observed by me in the rat
are not so marked as those described by Frank and Unger,
Bouin and Aneel, and O’Donoghue. Yet they are worthy of
mention and are certainly sufficient to prove that the normal
structure and variability under different conditions of any part
of the animal body should be thoroughly investigated before
conclusions are drawn from experimental work. It is quite pos-
sible that such knowledge of the mammary gland of the rabbit
would have changed decidedly the conclusions of Lane-Claypon
and Starling.
2. Growth of the ducts
In the newborn rat, models were reconstructed showing one
gland of each of the six pairs (figs. 3 to 6). At two weeks, all
the glands are represented in figure 2, to show the general topog-
raphy of the ducts. At the other stages (1 week, 2, 3, 4, 5, 7
and 9 weeks, figs. 7 to 13) it is found unnecessary to reproduce
all the glands, so only the abdominal and inguinal glands of the
left side are shown. 67.
15 Drawn from a section through the nipple area of the second inguinal gland
of an albino rat one week old. X 67.
16 From a section through the nipple area of the second inguinal gland of an
albino rat two weeks old. X 67.
17 From a section through the nipple of the second inguinal gland of an albino
rat nine weeks old. X 67.
388
PLATE 4
STUDIES ON THE MAMMARY GLAND
A. MYERS
J.
389
: y i
>_> 7 :
— = .
it <-
ye
. /
THE RELATION OF CORONARY AND HEPATIC
ARTERIES IN THE COMMON GANOIDS
C. H. DANFORTH
From the Department of Anatomy, Washington University Medical School
FOUR FIGURES
The regular occurrence of anterior and posterior coronary
arteries in several of the Rajidae was definitely recognized by
Hyrtl (’58) and T. J. Parker (’84). Subsequent investigations by
numerous students of fish morphology have shown that while
both sets of these arteries are characteristically present in skates,
posterior coronaries do not occur in either the sharks or the tele-
osts. This fact has led to the prevailing idea that the posterior
coronary artery is peculiar to the Batoidei. Such a vessel, how-
ever, has been found in the ganoid, Polyodon, by Allen (07) and
Danforth (712).
In the latter paper the writer also pointed out the existence
of anterior hepatic arteries arising in connection with the pos-
terior coronaries. ‘These anterior hepatic arteries were found to
have a relation to the hepatic vein similar, in general, to the rela-
tion of posterior hepatics to the portal system. They supply
the anterior half of the liver and usually part of the gall bladder,
anastomosing here and elsewhere with the posterior hepatic ar-
teries. That they may also occur in skates is indicated by a
single observation on Raja ocellata where a typical anterior
hepatic artery was found extending to the gall bladder.
The rather unexpected occurrence of posterior coronary ar-
teries in Polyodon and the discovery of the anterior hepatics,
has prompted the present investigation into the status of these
vessels in other ganoids. For this study the crossopterygian
Polypterus has not been available, but the laboratory has fur-
nished abundant material representing at least one genus of each
of the remaining four orders of recent ganoids. The following
391
392 C. H. DANFORTH
species, one from each order, have been studied: Seaphirhyn-
chus platyrhynechus, Polyodon spathula, Lepidosteus osseus,
Amia calva. Specimens of all these fish have been kept in the
aquarium to be killed and injected as needed. Some of them
were operated upon and certain vessels tied several days or weeks
before the fish was finally killed. About thirty specimens have
been studied. A few of them were dissected without injection,
but in most cases a colored starch mass was forced into the dorsal
aorta as soon as the fish was killed. By this method all the
vessels considered in the present paper are easily filled. The
dissections have been supplemented by a study of serial sections
of embryos and pieces of adult heart and liver.
The account of the conditions found in these four ganoids may
be prefaced by a brief statement of the relations of hepatic and
coronary arteries in lower forms. In this connection, it may be
stated at once that the hepatic arteries described for sharks and
skates (e.g. by Cavalié, ’04) are the posterior hepatics. These
are rather variable vessels, two of which usually arise from the
coeliac artery and accompany the portal vein. In addition, other
small arteries, from a similar source or from the oesophageal
trunk, generally enter the liver. The number and arrangement
of these arteries is far from constant even within a single species.
In the present study, except for the few points mentioned below,
I have found nothing about them calling for special comment.
A systematic interpretation of the relations of the coronary
arteries in fishes was first proposed by T. J. Parker (84, ’86).
According to his view, the whole region beneath the pharynx
and in front of the pectoral girdle is supplied, typically, by a
pair of vessels arising from the subclavian arteries. These ves-
sels he designates as hypobranchials. In their course forward
they give rise first to posterior (in skates) and then to anterior
coronary branches, finally ending in anastomoses with the re-
current branches of the efferent branchial arteries. The pos-
terior coronary arteries reach the heart by passing into the sep-
tum transversum, the anterior by following the aortic bulb
backward. Later writers, among whom may be mentioned in
particular G. H. Parker and K. Davis (’99) and Ferguson (’11),
RELATION OF ARTERIES IN COMMON GANOIDS 393
have pointed out that the connection between the subclavian
and hypobranchial arteries is generally shght or actually lacking.
This is interpreted to mean that the two systems are essentially
distinet with only secondary anastomoses. This being the case,
the posterior coronary arteries of skates are derived from the sub-
clavian artery, the anterior coronaries from the true hypobran-
chial system. Parker and Davis (’99) designated the branch of
the subclavian from which the posterior coronary arises as the
coracoid. They also differentiated clearly between median and
lateral hypo-branchial arteries and recognized the existence in
Carcharias and other forms of dorsal and ventral anterior coro-
naries.
It will be apparent from the foregoing that a discussion of the
morphology of the coronary and hepatic arteries of ganoids will
involve some consideration of the subclavian and hypobranchial
arteries. The accompanying schematic drawings show the es-
sential relations of these vessels in the fishes studied.
The principal branches of the subclavian artery are fairly con-
stant. After reaching the level of the lateral line the main trunk
gives rise above to a dorsal and below to a ventral superficial
branch to the lateral body musculature. Following Pitzorno (’05),
these may be designated as Rr. thoracico-dorsalis and thorac-
ico-ventralis. Between these two vessels arise branches to the
abductor and adductor muscles of the fin and a branch which
runs forward beneath the heart to supply the Mim. coracoarcualis
and pharyngoclaviculares. This is the coracoid artery of Parker
and Davis (l. ¢.).
The hypobranchial system of vessels is somewhat rudimentary
and highly variable. It does not lend itself readily to a general
description.
The condition found in skates is most closely approached by
Scaphirhynehus. his form will therefore be described first.
SCAPHIRHYNCHUS
In Seaphirhynehus anterior and posterior hepatic and posterior
coronary arteries are present. | have not found anterior coro-
naries. ‘The coeliaco-mesenteric artery passing down on the
394 C. H. DANFORTH
right side of the oesophagus becomes embedded in the liver for a
short distance near the pyloric end of the stomach. From the
part of the vessel which les within the liver and from its immedi-
ate subdivisions the posterior hepatics arise. A few small twigs
also enter the dorsal side of the liver from the oesophageal ar-
al
| |
md. coca. va,
Fig. 1
Scaphirhynchus.
abd, artery to abductor muscle of pec-
toral fin
add, artery to abductor muscle
aoe, oesophageal artery
ac, arteria apicis cordis
br 3, br 4, third and fourth efferent
branchial arteries
ch, common trunk of posterior coronary
and ant. hepatic arteries
coca, coraco-arcualis (epigastric) artery
come, coeliaco-mesenteric artery
sub. da. come oe. hp.
eT ATE UEECCCEECEECEC neat
/ | \ \
cor. Cp. ch. ha. ac. abd. add. thy. thd. liv
|
Sketch to show the relations of coronary and hepatic arteries in
cor, coracoid artery
da, dorsal aorta
ha, hp, anterior and posterior hepatic
arteries
liv, liver
md, mandibular artery
oe, oesophagus
sub, subclavian artery
thd, thv,
tralis arteries
thoracico-dorsalis and -ven-
va, ventral aorta
teries which are derived from the coeliaco-mesenteric and fourth
efferent branchial arteries.
The anterior hepatic and posterior coronary arteries arise from
a common trunk (ch) which in turn comes from the coracoid
branch (cor) of the subclavian.
skates, which suggests that the vessels are homologous.
This is the condition found in
Both
the anterior hepatic and posterior coronary arteries, however,
RELATION OF ARTERIES IN COMMON GANOIDS 395
are better developed in Scaphirhynchus than in the skates that
have been described. The common trunk (ch) usually divides
into coronary and hepatic divisions, each of which may again
divide before entering its respective organ. The posterior coro-
nary artery supplies the base of the heart and the aortic bulb;
the anterior hepatic branches supply the very abundant lymphoid
areas within the anterior part of the liver.
The apex of the heart may be supplied by a typical arteria
apicis cordis (ac) such as Spalteholz (08) found in certain turtles
and lizards. This observation is of interest since Spalteholz
thought the vessel did not occur in fish, birds or mammals. In
the Seaphirhynchus used for the accompanying sketch the ar-
tery was as well developed as in the case of some of the turtles
shown in Spalteholz’s own figures. I have not been able, how-
ever, to demonstrate it in all cases, so the vessel is probably
variable in its occurrence.
The points in which Scaphirhynchus departs from the typical
condition found in skates are, briefly, the greater development of
posterior coronary arteries with the associated anterior hepatic
vessels, the absence of anterior coronaries, and the potentiality
of an arteria apicis cordis.
POLYODON
The same hepatic and coronary arteries found in Scaphirhyn-
chus are present in Polyodon, except for the arteria apicis cordis
which has not been observed. The relation of these vessels to
the heart and liver are essentially the same in the two species,
but the origin of the hepatico-coronary trunk is quite different.
In Polyodon this vessel arises from the fourth efferent branchial
artery, a fact which it was found difficult to interpret when the
arteries of this fish were first described.
This peculiar arrangement is easily explained by reference to
the condition found in Seaphirhynchus. In the latter (fig. 1)
there is an artery arising from the fourth aortic arch which sup-
ples a part of the oesophagus and the region about the pericar-
dium. In Polyodon this vessel has anastomosed with the he-
596 Cc. H. DANFORTH
patie and coronary vessels and become their main trunk. The
original hepatico-coronary trunk (ch) is now reduced to a small
branch connecting the coronary artery with the subclavian. It
was originally interpreted as merely a secondary anastomosis.
As in Seaphirhynchus no anterior coronary arteries are present.
What might be interpreted as a rudimentary dorsal coronary is
lost on the ventral aorta before reaching the heart. Polyodon
differs from Seaphirhynchus chiefly in that the origin of the he-
br.4 da.
Treen ua en eee Tea
VX | |
md. coca. va. cp. ha. ch/abd.cor thy add. liv hp
Fig. 2 The coronary and hepatic arteries of Polyodon. Letters as in figure 1.
patico-coronary trunk is transferred from the coracoid branch
of the subclavian to the dorsal part of the fourth efferent bran-
chial artery.
LEPIDOSTEUS
In Lepidosteus, posterior hepatic and anterior coronary ar-
teries alone are present. ‘The latter arise from a vessel which
comes off on the right from the large oesophageal trunk, which
arises in turn from the subclavian artery. The anterior coro-
nary arteries arise from a short median hypobranchial which de-
rives its supply partly from the A. mandibularis (md) and partly
from a commissural artery at the level of the second gill. These
two vessels unite laterally and the resulting common trunk reaches
RELATION OF ARTERIES IN COMMON GANOIDS 397
the median line below the aorta. The coronary arteries, which
follow the aorta back, give rise to branches that may be definitely
identified as ventral coronaries and others that may, with less
certainty, be classed as dorsal coronaries.
Lepidosteus shows little in common with the preceding forms,
but does present some similarity to the condition found in Amia.
br3 br4 sub aoe da. oe hp.
ww
a
| \
coca. va. Ca. cor. abd. thd, add. thv. liv
Fig. 3) The coronary and hepatic arteries of Lepidosteus. ca, anterior coro-
nary artery. Other letters as in figure 1.
AMTIA
Like Lepidosteus, Amia has neither posterior coronary nor
anterior hepatic arteries. The anterior coronaries arise from a
median hypobranchial artery, as described by Parker and Davis
(99). The hypobranchial itself, however, may be derived, as
shown in the figure, from two commissural arteries instead of
the one usually described. Dorsal and ventral coronary arteries
may be recognized in the branches that run along the upper and
lower aspect of the aorta, but there seems to be no fundamental
difference between them. The remaining arteries in Amia call
for no special mention in this connection.
It will appear from the deseriptions that there are among the
ganoids two different schemes for supplying arterial blood to the
heart and liver. One, represented by Lepidosteus and Amia,
398 Cc. H. DANFORTH
approximates the teleostean type and presents no new points of
special interest. The other, represented by Scaphirhynechus and
Polyodon, appears to be a modification of the more primitive con-
dition found in skates. Functionally, the latter arrangement
would seem to be superior to the former, particularly in Polyo-
don where the whole supply to the heart and a large part of that
to the liver is carried by a special vessel directly from the gills
to these organs.
br.3 br.4 aoe, come. oe hp.
ggssssuut A
S
S
SS
i
OQ
NN eee
t Soars cee
175200 NNT Ip a gat
Sim
sim
Yj
Corr taLarrn,
ta ZAK *
md. coca. va. Cai “cor, abd. add. thv. thd.
Fig. 4 The coronary and hepatic arteries of Amia. Lettering as in the pre-
ceding figures.
Morphologically, how much significance can be attached to
these blood vessels is an open question. In different individuals
of the same species of shark, Carazzi (04) found the various ar-
rangements of cardiac and oesophageal vessels that had been
supposed to characterize several different species. The dis-
section of a few specimens of almost any fish will serve to show
how greatly the vessels vary. On the processes involved in the
production of such variations considerable work has been done.
In other groups of animals blood pressure has been thought to
play an important rdle in determining the interrelations of ar-
teries (Thoma, ’01). It seems probable to the writer that the
distribution of blood vessels in the fish is primarily a functional
RELATION OF ARTERIES IN COMMON GANOIDS 399
matter, dependent largely upon the interaction of blood pres-
sure within the vessel and various forces that are brought to bear
on its outer surface. Besides these passive factors, is the tend-
ency of the vessel itself to spread into new regions.
This hypothesis receives some support from the results of ex-
periments made on Polyodon and Secaphirhynchus in which the
coeliaco-mesenteric artery can easily be tied.!. There is, as above
deseribed, a normal anastomosis between the anterior and the
posterior hepatic arteries thus connecting the branchial and in-
testinal arteries. There is also an anastomosis, smaller than the
other, between the intestinal artery and a branch of the aorta
that passes over the posterior side of the swim-bladder. Now
when the coeliaco-mesenteric artery is tied off, the pressure must
be lowered in the intestinal arteries and raised in all others. It
would be expected that this would cause any secondary con-
nections existing between the two systems to enlarge, and as a
matter of fact a new artery adequate to supply the whole intes-
tine is promptly formed through the posterior anastomosis. In
neither form, however, have I been able to increase appreciably
the caliber of the artery that passes through the liver, even by
a second operation in which the new posterior mesenteric artery
is tied. The difference in the behavior of the two anastomoses
is very probably due to the fact that one occurs in the substance
of the liver while the other is in the loose tissue of the mesentery.
A somewhat analogous set of interacting factors might easily
determine the distribution of arteries in a growing embryo, and
indeed, the evidence, so far as it goes, suggests that the arrange-
ment of vessels is determined anew in each individual. That
the same scheme tends to prevail in a given species or in related
species is due, on this assumption, to the constant relations of
' Owing to its lack of scales Polyodon is especially favorable for this kind of
experimentation. The procedure has been to open the abdomen by cutting
along the linea alba and then turning the fish on its left side, when it becomes an
easy matter to pass a ligature around the long free strand in which runs the coe-
liaco-mesenteric artery. The operation may be done without loss of blood. The
fish usually recovers quickly and shows no bad after effects except that the cut
in the linea alba is slow in healing. Several operated specimens were kept in
the aquarium for three weeks before they were finally killed.
400 Cc. H. DANFORTH
other organs rather than to an inherent morphological individu-
ality of any given vessel. On this basis similar vascular arrange-
ments might be produced by very different causes, and it is not
impossible that this fact may explain the above described simi-
larity between skates and the cartilaginous ganoids.
LITERATURE CITED
Auten, W. F. 1907 Distribution of the subcutaneous vessels in the head region
of the ganoids, Polyodon and Lepisosteus. Proc. Wash. Acad. Sci.,
vol. 9, pp. 79-125.
Auuis, HE. P. 1900 The pseudobranchial circulation in Amia calva. Zool. Jahr.
Abt. Anat., Bd. 14; pp. 107-134.
1911 The pseudobranchial and carotid arteries in Polyodon spathula.
Anat. Anz., Bd. 39, pp. 257-262, 282-293.
Carazz1, D. 1904 Sulla circolazione arteriosa cardiaca ed esofagea della Scyl-
lium catulus. Intern. Monatschr. Anat. Physiol., Bd. 21, pp. 1-20.
Cavauib, M. 1904 La vesicule bilaire et sa circulation arterielle, chez quelques
poissons de mer (Torpedo galvani, Scyllium catulus, Galeus canis.),
C. R. Soe. Biol. Paris, T. 55, pp. 13886-1388.
Danrortu, C. H. 1912 The heart and arteries of Polyodon. Jour. Morph.
vol. 23, pp. 409-455.
Frerauson, J. 8. 1911 The anatomy of the thyroid gland of Elasmobranchs,
with remarks upon the hypobranchial circulation of fishes. Amer.
Jour. Anat., vol. 11, pp. 151-208.
Hyrti, J. 1858 Das Arterielle Gefiisssystem der Rochen. Denkschr. d. Kais.
Akad. d. Wissens., Wien, Bd. 15.
Parker, G. H. anp Davis, Kk. 1899 The blood vessels of the heart in Carchar-
ias, Raja, and Amia. Proc. Boston Soc. Nat. Hist., vol. 29, pp. 163-
178.
Parker, T. J. 1884 A course in zootomy. London.
1886 On the blood-vessels of Mustelus antarcticus. Philos. Trans.
Soc., London, vol. 177.
Prrzorno, M. 1905 Richerche di morphologia comparata sopra le arterie suc-
clavia ed ascellare Selachii. Monit. Zool. Ital., vol. 16, pp. 94-103.
SpauTEuoiz, W. 1908 Zur vergleichenden Anatomie der Aa. coronariae cordis.
Verh. d. Anat. Gesell. 22 Versam., pp. 169-180.
THoma, R. 1901 Uber den Verzweigungsmodus der Arterien. Arch. f. Ent-
wickelungsmech. d. Organ., Bd. 12, pp. 352-413.
THE CELL CLUSTERS IN THE DORSAL AORTA
OF MAMMALIAN EMBRYOS!
V. E. EMMEL
Assistant Professor of Anatomy, University of Illinois, College of Medicine
From the Department of Anatomy, Washington University Medical School
TWO PLATES (11 FIGURES)
CONTENTS
I. General structural characteristics of the aortic cell clusters... ........ 401
Lie The question. oftheir origin... 22... .a 5-2 ond a ayos Beats SAA cal asin bedale 402
Lo Statement ol proplemic... 2.264.410. a0 fe ews odin cadeeen ee ox ts ead ie 402
2. Grounds for regarding these clusters as of greater significance than
merely incidental structures..............600.0000c ees cseenes 403
3. Evidence as to their origin from the vascular endothelium........ 405
III. Concerning the correlation of the clusters with certain aortic develop-
THEN GaN DEOCESSES gs. xpos 2 ast, Me Ge aera es a ae ee ee re)
1. Degeneration and caudal wandering of aortic rami.............. . 407
2. Correlation of the aortic clusters with these vascular changes.... 409
IV. Discussion concerning endothelial tissue as contributing to the cellular
elements oly the vbhood. 2.4 vers wile 6 ee ee ee ee ee ee 412
Ne nes ae ce Menace Saute arteie ie aie oe wei ed aa ee x 9 geaeee ae eh ae 413
Vile MEN CETACUTERCIGC Cin: Dita cz Geek am aco nashayete ia SOE eked oho eet Sen 415
I. GENERAL STRUCTURAL CHARACTERISTICS OF THE AORTIC
CELL CLUSTERS
Figure 1 represents a portion of the ventral wall of the dor-
sal aorta as seen in a median sagittal section of a 9 mm. pig
embryo. Along its endothelial surface may be observed five
or six darkly stained cellular masses or clusters (ac,-ac;). Fig-
ures 2, 3 and 4 represent similar clusters as seen in transverse
‘Some of the present observations were made while engaged in research work
at the University of Strassburg and I wish here to express my indebtedness to
Professor Weidenreich for the generosity with which the facilities of his labo-
ratory were placed at my disposal. An abstract of the work was also published
in the Proceedings of the American Association of Anatomists, Anatomical
Record, Vol. 9, p. 77, 1915.
401
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, NO. 3
402 V. E. EMMEL
sections of the dorsal aorta. These figures illustrate the gen-
eral appearance, maximum size, and anatomical relations of
these structures. In consequence of their basophilic reaction
to Giemsa’s stain they stand out in sharp contrast to the red
stained erythrocytes of the aortic circulation.
Cytological details in the structure of these clusters are il-
lustrated in figures 6 and 9. The cells usually approximate
a spherical form. The nuclei may be round but they more
frequently present an indented or kidney shape and generally
occupy an eccentric position in the cell body (figs. 6 and 9, m,, mz).
Usually, the nucleoplasm takes a lighter stain than that of the
cytoplasm. Chromatin granules are rather evenly distributed
throughout the nucleus. The cytoplasm takes a basophilic
stain of varying intensity. Occasionally the cytoplasm is
also vacuolated and contains phagocytized cellular inciusions,
the latter being evidently chiefly of an erythrocytic nature.
In some instances a reddish tinted or what appears to be a cen-
trosphere is observed in the region of nuclear indentation (fig.
9, m,). Aside from certain structural variations to be subse-
quently considered, the nuclear and cytoplasmic characteristics
of the majority of these cells are comparable to the basophilic
and phagocytically active cells or macrophags (mesamoeboids?)
present in the circulating blood of the same embryos. ‘That
the component cells of these clusters become gradually dis-
sociated and as detached cells contribute to the circulatory
elements of the blood is indicated by those relationships in
which certain cells are attached by only a slender cytoplasmic
pedicle (figs. 6 and 9, m,) and others are apparently entirely
free (figs. 6 and 9, mz).
These clusters were studied in pig, mouse and rabbit em-
bryos and at certain developmental stages found in each of
these mammals.
Il. THE QUESTION OF THEIR ORIGIN
1. Statement of problem. In endeavoring to ascertain the
nature and origin of these cell clusters in the aorta certain ques-
CELL CLUSTERS IN MAMMALIAN EMBRYOS 403
tions arise for consideration. Are they merely more or less
agglutinated masses of circulatory blood corpuscles incidentally
resting against the vascular surface or are they structures aris-
ing directly from the vascular wall? Furthermore, can their
occurrence be correlated with any special developmental pro-
cesses of the embryo?
Of preceding investigators who have recorded the observa-
tion of similar structures may be noted Maximow (’09. p. 157)
for rabbit and cat embryos, Dantschakopff. (’07) for the chick
embryo and Minot (712, p. 525) for the human embryo. Maxi-
mow and Dantschakopff interpret these cell masses as endothe-
lial derivatives differentiating in situ from the vascular wall.
Minot on the other hand did not find the evidence sufficiently
convincing to justify the conclusion that they are of endothelial
origin. Beyond the immediate question of their origin no data
has been advanced as to why and under what conditions these
cell clusters occur in the aorta. The present status of the
problem and certain hematological questions associated with
the subject are consequently such as to justify a further exten-
sion of the investigation.
Of the mouse, rabbit, and pig embryos studied in the present
work the cell structures in question were found most pronounced
and striking in the pig embryo. Consequently the following
account is based chiefly on the results derived from the latter
mammal, in which it appears that these clusters have not been
previously described. The pig material for this purpose con-
sisted of seventeen embryos varying in size from 6 mm. to 25
mm. fixed in Zenker-formalin (Helly’s combination), embedded
in paraffin or celloidin and the serial sections stained with Giem-
sa’s fluid. Several of the 12 mm. embryos were also stained
in toto with borax carmine and the sections counter stained
with Lyon’s blue.
2. Grounds for regarding these clusters as of greater signifi-
cance than merely incidental structures. Upon first impression
one may be inclined to discredit any special relationship between
these cell masses and the aortic wall. As the result of further
investigation it appears clear, however, that the phenomenon
404 Vv. E. EMMEL
is evidently one of greater significance than that of merely
incidental cellular accumulations.
First, the clusters are In many cases at least, evidently rather
firmly attached to the aortic wall. This is indicated by the
fact that these structures may be found in the aorta even though
the vessel is practically empty as may occur for example through
the loss of blood during the preparation of the material. Again
in some cases in which during the process of fixation almost the
entire blood content settles toward one side of the aorta, the
clusters instead of being carried along with the rest of the blood,
continue attached in what appears to be their original position
on the aortic wall. Finally, it may be observed that many
of the larger cell masses present a rather elongated shape (fig. 1)
with one end of the long axis adherent to the aortic wall and
the other end free and directed caudalward (towards the left
in the figure); i.e., the relations are such as might be expected
in an attached cell mass one end of which was free to be carried
down stream by the force of the blood current.
Second, the clusters are of constant occurrence in 6 mm. to
15 mm. embryos. Thus in a count for four 12 mm. embryos
there was a total of 45 clusters distributed in the proportion
of 11,9, 13 and 12. On the other hand in embryos beyond about
the 15 mm. stage, these cell masses are absent. Such condi-
tions do not appear readily accounted for on the basis of acci-
dental agglutination.
Third, in all the embryos studied the clusters were confined
to the ventral half of the aortic tube and with greatest frequency
toward the median region of its ventral wall (figs. 2, 3 and 4).
In no case was a cluster found in the dorsal aortic wall. Such
facts certainly appear indicative of a deeper relationship be-
tween these cell structures and the aortic tube.
In connection with this conclusion it is of course to be recog-
nized that an occasional cell from the circulating blood may
possibly now and then have been fixed in sueh a manner by the
killing reagents as to be adherent to the vascular surface. On
the other hand sections other than these passing through the
base of the cluster may also present the deceptive appearance
CELL CLUSTERS IN MAMMALIAN EMBRYOS 405
of an absence of attachment, Finally with the gradual disap-
pearance of the clusters in older embryos an intimate histologi-
cal relationship with the aortic wall may become less and less
evident.
3. Evidence as to their origin from the vascular endothelium.
The preceding considerations rendered necessary a more de-
tailed investigation of the cytological relations and origin of
the aortic clusters. On the basis of the following results the
conclusion is drawn that they are endothelial derivations and,
as will be subsequently more fully elaborated, arise in relation
to certain vascular conditions in the ventral portion of the aorta.
The first notable feature to which attention may be directed
is the absence in the majority of cases of a definite continuity
of the vascular endothelium in the region of contact between
the clusters and the aortic wall. Cell boundries are not clearly
defined and the cells of the endothelium are in evidently syn-
eytial relation (illustrated in figure 9 but not clearly evident
in the low power drawings for figures 2 and 3). The cluster
in figure 9 cannot be said to be resting upon a continuous sheet
of typical flattened endothelium. On the contrary it can in the
second place be shown that endothelium at the base of the
cluster presents marked cytological modifications. In the
vicinity of the cell masses the endothelial cells can be observed
usually closer together than normally and the rounded nuclei
frequently at one side present a more or less marked indentation
or concavity giving it a kidney shaped appearance. (fig. 9).
It is also to be noted that such endothelial conditions are most
evident in the case of the clusters presenting the more intimate
relationship with the vascular wall. Third, what appears to
be transitional cytological changes can be traced from the
somewhat lighter stained endothelial cells at the base of the
cluster to the more sharply outlined and more deeply basophilic
cells at the periphery of the mass (figs. 9, 6). It may also be
observed in the same figures that many of the more peripheral
cells still present a clearly defined cytoplasmic elongation or
pedicle attaching them to the more central regions of the clus-
ter. Fourth, that these conditions represent active cellular
406 Vv. E. EMMEL
differentiation rather than cellular disintegration appears dem-
onstrated by the not infrequent evidence of mitotic activity
as well as phagocytic function in the component cells of the
clusters (fig. 9).
Fifth, areas of the aortic wall are found in which the endo-
thelial cells present structural characteristics comparable to
those already described in the endothelium adjacent to the
larger cell masses (fig. 10a). Such areas are not, however, in
any direct juxtaposition to the aortic clusters nor are they to
be explained as deceptive appearances due to oblique or tangential
sections of the endothelial surface. It may be observed that the
endothelial cells are closer together, project above the general
level of the vascular surface, and not infrequently take a more:
basophilic stain. The nuclei may be either rounded in form or
approximate a kidney-shaped contour. No evidence was ob-
served of mitotic cavity indicative of merely an incidental in-
crease of endothelial cells in such regions through ordinary
endothelial growth and cell multiplication. It is also note-
worthy that such conditions were not found in the dorsal aortic
wall as is illustrated in a comparison of figures 10a and 10b.
Indeed the cytological structure and vascular relations of these
cells appear identical with that of cells occurring in the aortic
clusters. The fact that the aortic cell clusters present a great
difference in size varying from the large masses in figures, 2 6
and 9 to these smaller accumulations of only a few cells and that
they are also no longer present beyond certain stages of em-
bryonic development, suggests that such areas as shown in
figure 10a may represent end stages in the gradual dissociation
and final disappearance of the aortic clusters, in which, how-
ever, there still remain indications of the endothelial reaction
which has given rise to these structures. Sixth, of the two fixed
tissue elements of the aortic wall which could possibly take
part in the cell activities in question it is evidently primarily
the endothelium rather than the mesenchyma which partici-
pates in the formation of these cell masses. The demarcation
between the mesenchyma and the aortic clusters is fairly well
defined, nor was there obtained any conclusive evidence of a
EL
CELL CLUSTERS IN MAMMALIAN EMBRYOS 407
possible migration of free cells from the mesenchyma into the
clusters.
Finally, it is important to note that not infrequently the aortic
clusters are in direct relationship or continuity with cell masses
situated within certain atrophying arterial branches of the
aorta. These intra-arterial masses have evidently arisen in
situ from the endothelium of the artery in question and as
described elsewhere (pp. 409-411) probably constitute the pri-
mary source of origin of the aortic clusters.”
III. CONCERNING THE CORRELATION OF THE CLUSTERS
WITH CERTAIN AORTIC DEVELOPMENTAL PROCESSES
1. Degeneration and caudal wandering of aortic rami. In
the development of the mammalian aorta there occur two im-
portant vascular changes involving a shifting or caudal wander-
2 An additional observation which may be conveniently recorded here relates
to a type of structure illustrated in figure 5. This group of cells is attached to
the ventral aortic wall and projects into the lumen of the vessel but differs from
the typical aortic cluster through its enclosure by a more or less definitely marked
peripheral membrane (en). In close relation to the membrane are a number
of cells some of which present the flattened endothelial form while others are
more rounded in shape. In other respects the component cells appear similar
to those of the clusters. Such structures are apparently of rare occurrence for
in the present material they were found in only one embryo, a 12 mm. specimen
(W. U. coll. No. 3), in which there were two of these bodies, both ventrally lo-
eated. (This embryo had been stained with borax-carmine and the present
statements are made without having data derived from Giemsa stained material. )
It is of interest to observe that this same embryo was also deficient in the usual
number of aortic clusters, for only three of the latter were found instead of the
9-13 clusters in each of four other embryos of the same size.
The aorta of the same embryo also contained three elongated cellular strands
evidently of endothelial nature. Two of these strands (about 60 micra in length)
were attached to the left umbilical artery near its origin from the aorta. The
third strand about 300 micra in length and varying from one to several cells
in thickness, was connected by only a slender cytoplasmic strand to the aortic
wall. Two such structures were also found in a second 12 mm. embryo (W. U.
coll. No. 5), one in the aorta and the other in the region of origin of the umbilical
arteries. Nothing conclusive was ascertained as to the significance of these
strands, but the suggestion merits further investigation as to whether they may
possibly be associated with the fusion of the two original dorsal aortae. With
reference to our present purpose, however, it is of interest to note that many
of their component cells present rounded form, kidney shaped nuclei, and phago-
cytic activities comparable to that of the component cells of the aortic clusters.
408 Vv. E. EMMEL
ing of certain arterial branches of the aorta and an extensive
degeneration of others.
Directing attention first to the degenerative changes it will
be recalled that the early embryonic aorta has three sets of
eighteen to twenty or more paired branches—dorsal, lateral,
and ventral. Of these rami, nearly all the dorsal vessels per-
sist throughout embryonic development, whereas practically
all the remaining vessels, with certain exceptions, subsequently
atrophy and disappear. Thus in the case of the human em-
bryo it has been shown that the primitive lateral branches of
the aorta which form an extensive system in the twenty-three
somite embryo, have in the 16 mm. to 19 mm. stage embryos
become largely atrophied. The single median arterial stems
which have replaced the extensive series of paired ventral ar-
teries of younger stages (Tandler, 03) and in a 5 mm. embryo
extend as a ‘“‘complete series of unpatred or median ventral
segmentals from the seventh cervical to the second lumbar
segment inclusive,” in a 7 mm. embryo “‘have been reduced to
three main trunks” the coeliac, superior, and inferior mesenteric
arteries. (Keibel-Mall, ’12, pp. 603, 611, 648, 653).
Second, it is to be observed that the three remaining arterial
trunks to which the ventral segmental aortic arteries have been
reduced, undergo a remarkable shifting or caudal wandering
as first described by Mall in 1891 and subsequently confirmed
by Tandler (’03) and Broman (’08). In the human embryo
for example ‘‘the coeliac artery thus wanders from the seventh
cervical to the twelfth thoracic segment, a displacement of some
eleven segments, and the superior mesenteric artery almost
equally as far (ten segments, second thoracic to third lumbar) ;
whereas the inferior mesenteric artery wanders through but
three segments (twelfth thoracic to third lumbar)’’ (Evans 712,
p. 647).
Referring again to the human embryo it is of especial interest
with reference to our present purpose to note that all of these
vessels “usually attain the adult levels by the time the embryo
is 17 mm. long.’ Furthermore this caudal wandering of the
intestinal arteries is not by a displacement of the aorta on the
—»
— 842°) Wr wt
CELL CLUSTERS IN MAMMALIAN EMBRYOS 409
vertebral column, but is an actual shifting of these ventral
branches when compared with the dorsal branches of the same
trunk (Evans, 712, p. 648).
Although the details have not, so far as I am aware, been
as carefully ascertained as in the case of the human embryo,
essentially the same conditions evidently maintain for the pig
embryo as in other mammals. The ventral and lateral aortic
rami undergo a similar degeneration. The shifting of the
ventral or intestinal vessels is indicated by the fact that the
coeliac artery which in a 6.5 mm. pig embryo is at the level
of the eighth cervical segment, in a 12 mm. embryo is at the
level of the fifth thoracic segment. Again, the superior mesen-
teric artery in the 6.5 mm. embryo is at the level of the third
thoracic segment, but has descended to the level of the eighth
thoracic segment in the 12 mm. embryo. Both the atrophy
of vessels and caudal wandering appear practically complete
at about the 15 mm. stage.
2. Correlation of the aortic clusters with these vascular changes.
In a comparative analysis of the preceding data for the vascular
changes and cell clusters in the aorta certain striking relation-
ships become evident. First, both phenomena occur within
the same period of embryonic development—between the stages
of about 5 mm. to 15 mm. in the pig embryo. Again, both
the formation of the clusters and the degenerative changes and
caudal wandering of the arteries, as already indicated, are con-
fined to the same region of the aorta, namely, its ventral wall.
Third, in a linear direction within this ventral region the cell
clusters are furthermore fairly evenly distributed between the
coeliac and umbilical arteries as shown in the following data
for four 12 mm. embryos:
NUMBER OF CLUSTERS BETWEEN:
TOTAL NUMBER
SPECIMEN 7. U. COLL. NO. : ; :
SE eIMEN (W. U. COLL. NO.) The coeliac and The superior OF CLUSTERS
superior mesenteric] mesenteric and
arteries umbili¢al arteries
1 (4) 5 6 11
2 (1) 3 6 9
g (2) 7 6 13
4 (5) 5 7 12
410 Vv. E. EMMEL
Fourth, there are certain important cytological conditions
to be considered in the degenerating arteries themselves. Many
of these vessels are found compactly filled with basophilic stain-
ing cells (cf. figures 1 da, 7 and 11). Such conditions are found
near the aortic origin of the artery and may extend for short
distances into the ramus, occasionally continuing to a point where
the degenerating vessel is lost in the mesenchyma. Eryth-
rocytes are strikingly deficient in such regions and may in-
deed be entirely lacking throughout the vessel, a condition evi-
dently indicative of the reduction if not complete cessation of
the circulation through these retrograding arteries. The com-
ponent cells of these intra-arterial cell masses may take a some-
what lighter stain but they otherwise appear cytologically iden-
tical with those of the aortic clusters. They are phagocyti-
cally active (fig. 11, in) and undergo cell multiplication (fig.
7,d). In regions of the artery not thus occluded the endothelial
cells are frequently rounded up or swollen and project into the
lumen of the vessel. The transitional stages to be found be-
tween the still intact endothelial cells and the intra-arterial
masses seem to leave no doubt but that the latter have arisen
in situ from the lining endothelium of the retrograding vessel.
Finally, there remains the crucial fact of an intimate relation-
ship between these intra-arterial masses and the aortic clusters.
This is illustrated in figure 6 in which the intra-arterial cell
mass (tam) when followed toward the aorta is found to terminate
in an aortic cluster situated within the lumen of the aorta. It
may be observed that the component elements merge into each
other with no evident line of demarcation between them. In-
deed the cytological conditions and morphological relations are
such as to justify regarding the phenomenon as essentially
comparable to a partial evisceration of the contents of the de-
generating artery into the lumen of the aorta. Such a relation
of aortic clusters and ‘intra-arterial masses is of frequent occur-
rence. It is also nuteworthy that in many cases where such
a relationship is apparently lacking the aortic cluster is, how-
ever, situated in a well marked depression or concavity in the
aortic wall, and that some of these depressions are in relation
CELL CLUSTERS IN MAMMALIAN EMBRYOS 411
to the atrophied remnant of a small artery which soon ter-
minates blindly in the adjacent mesenchyma. (Such depres-
sions are inadequately shown in ac2 and ac4 of figure 1, but can
be readily demonstrated in serial sections.) Not infrequently
in instances where such depressions are lacking there may still
be observed a clearly evident irregularity, sometimes of a more
or less whorled character, in the arrangement ofthe mesenchymal
cells at the base of the cluster as compared with the adjacent
regions of the aortic wall (figs. 2, 3, s). Occasionally the clus-
ters occur in pairs (fig. 4) as if they had arisen in connection
with the simultaneous atrophy of two paired aortic rami. In
apparent corroboration of these results certain conditions are
occasionally found at the aortic entrance to an as yet relatively
intact arterial ramus in which the cytological structure, form
and relations of the component elements of the vascular sur-
face suggest an early stage in the endothelial activities involved
in the production of the intra-arterial and aortic cell masses
(fig. 8).
On the basis of the preceding data the conclusion is drawn
that the formation of the cell clusters in the aorta are not only
intimately associated with, but are also: evidently correctly
interpreted as a direct result of the developmental processes
involved in the atrophy of the ventral and lateral aortig rami
and the establishment of the permanent intestinal arteries of
the adult organism.?
3’ Concerning the remarkable caudal wandering of the visceral rami of the
aorta a number of hypotheses have been advanced to account for the phenomenon
(Broman, ’08, Tandler, (93), Evans, ’12) but the exact manner in which the
process takes place has apparently as yet not been established. In connection
with the present study it may be observed that the depressions in the aortic
wall and the relations of the atrophied arterial stems and the aortic clusters
are such as to suggest that arterial remnants of the former aortic rami are
ultimately incorporated into the aorta itself. Such additions and consequent
inequalities of growth in the ventral region of the aortic wall as contrasted with
its dorsal portion may in a final solution of the problem be found to contribute
materially to the caudal shifting of the coeliac and mesenteric arteries. Evans
(12, p. 649) does indeed express the opinion that a primary factor in these vas-
cular changes is an unequal growth of the dorsal and ventral walls of the aorta
but has not, so far as I am aware, elaborated the specific nature of the process.
412 Vv. E. EMMEL
IV. DISCUSSION CONCERNING ENDOTHELIAL TISSUE AS
CONTRIBUTING TO THE CELLULAR ELEMENTS
OF THE BLOOD
On the basis of the present results it appears evident that the
cell clusters in the embryonic aorta furnish an instance of the
vascular endothelium contributing cellular elements to the
circulating blood. Since this is not entire agreement with
one of the postulates of the angioblast theory, namely that all
the blood cells of the organism are direct descendants from
the early embryonic blood islands, it becomes of interest to
note the conditions under which this endothelial activity is
taking place. The close association of the aortic clusters with
degenerating vessels directs attention to the occurrence in these
retrogressive vessels of stimulative factors to which the en-
dothelium reacts in the manner under consideration. As stated
by Thoma (’93) and elaborated by Mall (06) in the case of the
embryonic liver, a vessel in which there is a reduction of the
circulation below normal tends to shrink and disappear. The
occurrence of such a retardation of circulation in the atrophy
of the aortic rami is indicated by the marked absence of red
blood corpuscles (p. 410). With the circulation practically at a
standstill it is not improbable that diminished oxidation and
inhibition of gaseous interchange are in part at least produc-
tive of mildly abnormal chemical and toxic conditions conducive
to the phagocytic activities, endothelial proliferation and con-
sequent formation of intra-arterial cell masses and aortic clusters.
In support of this conclusion attention may be called to the
emphasis being more recently attached to such abnormal intra-
vascular conditions as stimulative to endotheilal activity.
Thus Mallory (00 and 714, pp. 165-166, 183) advances grounds
for the conclusion that certain dilute-and weak toxines stimulate
endothelial proliferation and phagocytes and maintains that
in this manner arise the macrophags encountered in many dis-
eases. Batchelor (14) and Scott (14) record marked prolifera-
tive endothelial changes and phagocytic activity in hepatic
vessels occluded by artificially produced emboli and wounds.
CELL CLUSTERS IN MAMMALIAN EMBRYOS 413
Finally the experimental results of a number of investigators
of whom may be mentioned Tschaschin (13, p. 370) and Mae
Curdy and Evans (712, p. 1695) ,may be adequately summarized
in the recent statement by Evans (715, p. 254) “that occurrences
which place the endothelium of the most various vessels under
conditions, such, for instance, as a direct injury of the endothel-
lium, cessation of the adjacent current, in short in all cases of
thrombosis or embolism, lead to the proliferation of endothe-
lum ”’ “Probably no area of the body can be ex-
cluded in this respect.’”!
In conclusion, therefore, it may be stated that while the original
assumption of the angioblast theory—that vascular endothelium
does not give rise to cellular elements of the blood—may under
normal conditions be true for the general systemic vascular
system, it appears that in both embryo and adult mammals,
endothelial tissue ordinarily passive may under certain abnormal
conditions, however, assume proliferative activities contributing
to the free cellular elements of the circulating blood.
V. RESUME
1. During the development of mouse, rabbit, and pig em-
bryos certain well defined cell masses or clusters are found in
the aorta of these mammals.
2. The majority of the component cells of these clusters are
in their cytological characteristics comparable to the basophilic
and phagocytically active cells or macrophags (mesamoeboids?)
in the embryonic circulation.
3. Their constancy of occurrence, firm attachment and restric-
tion to the ventral wall of the aorta indicate that these cell
clusters are not merely chance cellular accumulations but struc-
tures having a significant relationship to the vascular conditions
in the aortic artery.
4’The participation of the mesothelium in the origin of macrophags in the
embryonic coelom (Emmel, 715) is not improbably also a reaction to stimulative
conditions arising in part at least through degeneration and disintegration of
erythrocytic and other foreign elements escaping into these cavities.
414 Vv. E. EMMEL
4. Certain structural modifications and frequent discontinuity
in the endothelium at the base of these masses, the cytological
characteristics transitional between these endothelial cells and
the component cells of clusters, the evidence of mitotic activity,
variation in size, and relationship to certain degenerating aortic
rami, support the conclusion that the aortic clusters have arisen
from the vascular endothelium.
5. An intimate association and fundamental causal relation-
ship can be demonstrated between the formation of the aortic
clusters and the developmental processes involving the atrophy
of certain aortic rami and the establishment of the permanent
intestinal arteries of the adult mammal. The endothelium
in degenerating stems of the aortic rami is stimulated, (evidently
through certain toxic conditions arising in the retrogressive
vessels), to phagocytic and proliferative activities giving rise’
to infra-arterial cell masses constituting a primary source of
origin of the aortic clusters.
6. On the basis of the cumulative evidence of various recent
investigators it appears evident that the original assumption
of the angioblast theory that the endothelium of the general
systemic vascular system does not contribute to the cellular ele-
ments of the blood, while possibly true under normal conditions,
requires the qualification that under certain abnormal conditions
endothelial tissue ordinarily passive may in both embryo and
adult assume such proliferative activities.’
5 While the present paper was in press an article appeared in the Anatomical
Record, Vol. 10, p. 417, by Jordan on the ‘‘Evidence of Hemogenic Capacity
of Endothelium.”’ It is of especial interest to note that Jordan records the
observation of cellular structures in the aorta of mongoose and turtle embryos
apparently similar to the clusters occurring in the pig, mouse and rabbit em-
bryos. Here again the clusters are confined to the ventral region of the aorta.
In the mongoose and turtle, just as in the pig, ‘‘the clusters show a progressive
increase in size corresponding with the age of the embryos, between 5 and 10
mm., indicating an intrinsic growth” p. 419. It is emphasized that ‘‘similar
clusters are found nowhere else either in the yolk sac or the embryonic vessels
or sinusoids” (p. 418) and in agreement with the results of the present paper
cogent reasons are advanced for regarding these clusters as being not meiely
chance cellular accumulations, but as structures arising from the vascular en-
dothelium.
CELL CLUSTERS IN MAMMALIAN EMBRYOS 415
VI. LITERATURE CITED
BarscuELor, R. P. 1914 Demonstration of preparations showing the behavior
of endothelium after the introduction of emboli in the portal vein.
Anat. Rec., vol. 8, no. 2, p. 139.
Broman, I. 1908 Uber die Entwicklung und ‘Wanderung’ der Zweige der
Aorta Abdominalis beim Menschen. Anat. Hefte, Bd. 36, p. 407.
DanrscHakorr, W. 1907 Uber das erste Auftreten der Blutelemente im
Hithnerembryo. Folia haematolgica, IV, p. 159.
Emmet, V. E. 1915 The cell clusters in the dorsal aorta of the pig embryo.
Proceed. Am. Assoc. of Anatomists., Anat. Rec., vol. 9, p. 77, no. 1.
1916 Concerning certain cellular elements in the coelomic cavities
and mesenchyma of mammalian embryos. Am. Journ. of Anatomy..,
(In press for vol. 19.)
Evans, H. M. 1912 The development of the vascular system. Keibel-Mall
Human Embryology, vol. 2, p. 570.
1914 The physiology of endothelium. Anat. Ree., vol. 8, no. 1,
p. 99.
1915 The macrophags of Mammals. Am. Jour. of Phys., vol. 37,
p. 248.
His, W. 1901 Lecithoblast und Angioblast der Wirbelthiere. Histogenetische
Studien. Abhandl. der Math.-Phys. Classe Gesellsch. de Wissens-
chaften., Bd: 26, p. 171.
Jorpan, H. E. 1916 Evidence of homogenic capacity of endothelium. Anat.
Rece., vol. 10, p. 417.
Kerpet-Mati 1912 Human Embryology.
MacCurpy AND Evans 1912 Experimentelle Lisionen des Zentralnerven-
systems, untersucht mit Hilfe der vitalen Farbung. Berl. Klin.
Wochenschr., no. 36.
Matt, F. P. 1891 A human embryo twenty-six days old. Jour. Morph., vol.
5, pp. 459-480.
1906 A study of the structural unity of the liver. Am. Jour. Anat.,
vol. 5, p. 227.
Mauuory, F. B. 1900 Proliferation and Phagocytosis. Journ. Exp. Medicine,
vol. 5, pp. 1-18.
1914 The principles of pathologic histology. Saunders Company,
Philadelphia.
Maximow, A. 1909 Untersuchungen itiber Blut und Bindegewebe. I. Die friihes-
ten Entwicklungsstadien der Blut- und Bindewegebezellen beim
Saiugetierembryo, bis zum Anfang der Blutbildung in der Leber. Ar-
chiv f. Mikr. Anat. Bd. 73, p. 444.
Minot, C. 8. 1912 The origin of the angioblast and the development of the
blood. Keibel-Mall Human Embryology, vol. 2.
Scorr, K. J. 1914 Preparations showing the vital stain applied the study of
wound healing. Anat. Rec., vol. 8, no. 1, p. 141.
TanvbLeR, J. 1903 Zur Entwicklungsgeschichte der Menschlichen Darmar-
terien. Anat. Hefte, Bd. 27,. Heft. 71, p. 189.
416 Vv. E. EMMEL
Tuoma, R. 1893 Untersuchungen iiber die Histogenese und Histomechanik
des Gefiissystems. Stuttgart.
Tscuascuin, 8. 1913 Uber die ‘‘ruhenden Wanderzellen’”’ und ihre Bezie-
hungen zu den anderen Zellformen des Bingedewebes und zu den
Lymphozyten. Folia Haemat. Bd. 17, p. 317. Archiv.
PLATE I
EXPLANATION OF FIGURES
Figures 1, 6 to 11, inclusive, are from 9 mm. pig embryos fixed in Zenker-
formalin and stained with Giemsa. Figures 2 to 5, inclusive, are from 12 mm.
embryos fixed in Zenker-acetic and stained with borax carmine and Lyon’s blue.
All figures are from sagittal sections, except figures 2 to 5 which are from trans-
verse sections of the aorta. The drawings, reduced one-fifth in reproduction,
were originally made with the magnifications obtained in the following combina-
tions of Zeiss apochromatie lenses and compensating oculars:
Figs. 1 to: 4, oc. 8, No.3, obj.
Fig. 5, oc. 4, 2 mm. obj.
Figs. 6 to 11, oc. 6, 2 mm. obj.
Recognition is due the artist, C. D. Jarrett, for faithful reproduction of cyto-
logical details.
417
ABBREVIATIONS
ac, aortic cluster la, lateral aortic ramus
d, mitosis m, the more highly differentiated baso-
philic cells (or macrophags) in the
da, degenerating aortic ramus aortic clusters
dw, dorsal aortic wall r, erythrocyte
e, endothelium s, mesenchyma
iam, intra-arterial cell mass va, ventral aortic ramus
in, phagocytic inclusion vw, ventral wall of the aorta
The direction of circulation and the long axis of the aorta is indicated by
an arrow.
A portion of the ventral region of the aorta as seen in longitudinal section,
showing the general appearance and morphological relations of the aortic clus-
ters (acj-acs;). A caudalward projection of the free end of the clusters is illus-
trated in acs. Cluster acs is situated at the entrance to a root of the superior
mesenteric artery. da is a section of a degenerating aortic ramus packed with
basophilic cells similar to those of the clusters.
2and3 Show the form and structural relations of the aor tic clusters as seen
in transverse sections of the aorta. Adjacent to the base of these clusters may
also be observed a variation in the general arrangement of the mesenchymal
cells (s).
4 I[llustrates the double aortic clusters occasionally found. .
5 One of two spherical masses of cells found ina 12mm. embryo. The mass
appears in some respects comparable to the aortic clusters, but is surrounded
by a more or less definite endothelial membrane (en). Situated in the ventral
wall of the aorta.
6 Section of an aortic cell cluster situated at the extrance.to a ventral aortic
ramus and in intimate relationship with the intra-arterial cell mass (cam) fill-
ing the latter vessel. Also illustrates the more definite differentiation of the
peripheral cells (m,) of the clusters, some of which appear entirely detached from
the main mass (mz).
7 Section of an intra-arterial cell mass in a degenerating ventral aortic
ramus. Note absence of a definite lining endothelium and the evidence of active
-cell multiplication (e).
418
CELL CLUSTERS IN MAMMALIAN EMBRYOS PLATE 1
Vv. E. EMMEL
~
ao,
a .
® . > Re
NE DES Area
en A nae x 9
2 2 wae gene
Or, wee Rents gee : Peake .
>
a ~ Ce, => 7%
m — -
Se so e «
iv ~. - .
57 ATI antiets a =“ ESSA ap
‘ on Pn PSY
10a
I1
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, NO. 3
——
THE GENERAL FUNCTIONAL SIGNIFICANCE OF
MITOCHONDRIA!
Ek. V. COWDRY
From the Anatomical Laboratory, Johns Hopkins University
Years ago Schultze defined protoplasm as being a glass-like,
semifluid material in which granules are embedded. Only
now are we beginning to see more clearly, for investigators have
succeeded, in the last few years, in separating from this baffling
heterogeneous complex a class of granulations which are more
or less distinct chemically, morphologically and physiologically
and which occur in almost all protoplasm. These granules
are usually called ‘mitochondria.’ They are not a recent dis-
covery for both Flemming and Altmann observed and described
them, but recent studies have enabled us to define them more
accurately and have in consequence forced us to revise our
views of cell structure.
MORPHOLOGY AND NOMENCLATURE
Mitochondria vary in form from granules (0.2u-2u) to rods
and filaments, which may be straight, curved or even forked.
Sometimes they are ring-shaped, pear-shaped and possess bleb-
like swellings. Networks rarely oceur. That mitochondria
collectively undergo changes in form was early recognized through
the study of successive stages of histogenesis, particularly of
spermatogenesis. But it remained for the Lewises (’15, p.
352) to actually follow the form changes of individual mito-
chondria by the observation of living cells in tissue cultures,
to see filamentous mitochondria changing to granular ones, ete.
The Lewises also observed that the mitochondria moved
freely, and quite rapidly, from place to place in the cytoplasm.
1 Aided by the Carnegie Institution.
423
424 E. V. COWDRY
Of course it does not follow that they behave in the same way
in the cells in the organism, but the observation nevertheless
strengthens our suspicions of the ancient hypothesis of a proto-
plasmie reticulum, which still persists in the form of misleading
diagrams in all our textbooks of histology.
Following along the lines of Maximow’s (713, p. 244) studies
N. H. Cowdry has observed the changes in the shape of mito-
chondria in the streaming protoplasm of living plant cells. He
has seen filaments assume the form of loops and spirals in res-
ponse to currents and eddies in the stream indicating clearly
that they are flexible and that their form is in a measure deter-
mined by their environment.
In spite of observations such as these the literature is still
befogged by the use of a specific term for each of the different
forms of mitochondria. Thus, some would restrict the term
‘mitochondria,’? which was originally applied to them, to gran-
ules only; when the granules are arranged in rows they would
be called ‘chondriomites’; the filaments, ‘chondriocontes’; the
word, ‘chondriosomes,’ would be used as a generic term to in-
clude all the forms: and, finally, the cytoplasmic content of
mitochondria would be styled the ‘chondriome.’ This sys-
tem of terminology was very popular for a few years. It was
partially supplanted by the short-lived ‘Plastochondrial’ nomen-
clature advocated by Meves (710, p. 150), the chief objection to
which is that it was devised to proclaim the view that mito-
chondria play an important part in histogenesis. American
investigators have, with few exceptions, from the beginning
employed the term ‘mitochondria,’ exclusively, recognizing
well that the same material, under different conditions, may
assume special forms. Even the word ‘mitochondria’ leaves
many things to be desired, but it is in general use, it is descrip-
tive of morphology only and it does not commit the user to any
hypothesis of the functional significance of the materials in
question. True, we cannot use the name in the exact sense
that Benda, who introduced it, used it no more than we can
employ the term ‘cell’ with anything like its original meaning.
2 From uiros, a thread and yévépos, a grain.
FUNCTIONAL SIGNIFICANCE—MITOCHONDRIA 425
Yet no one would invite us to give up the word ‘cell’ and to sub-
stitute a new and more appropriate term in its place.
Mitochondria may be provisionally defined as substances
which occur in the form of granules, rods and filaments in almost
all living cells, which react positively to janus green and which,
by their solubilities and staining reactions resemble phospholipins
and, to a lesser extent, albumins.
RELATION OF MITOCHONDRIA TO OTHER CYTOPLASMIC
, CONSTITUENTS
It must again be emphasized that mitochondria are by no
means a recent discovery. Altmann and Flemming undoubtedly
observed and described them, but Altmann grouped other struc-
tures, like secretion granulations, under the same heading of
‘bioblasts’ just as Flemming did with his ‘fila.’ The interstitial
granules of Koelliker and the plasmosomes of Arnold as well
as the cytomicrosomes of Heidenhain are also in part mitochon-
dria. What is new about mitochondria is that they are with-
out doubt a more concrete class of cell granulations than any
of the above mentioned. Investigators were not slow to ap-
preciate the value of the work of the older authors, in fact the
confusion which is so woefully apparent, at the present time, .
in our ideas of the relation between mitochondria and other
formed bodies is due to the fact that they went altogether too
far in identifying mitochondria with previously described eyto-
plasmic constituents to which they are in no way related. This
tendency to carry things to an extreme, to overstep the mark
completely, is manifested in the study of mitochondria over
and over again.
A case in point is that of the chromidial substance. Gold-
schmidt (09, p. 107) and his pupils have persistently asserted
that mitochondria belong to the category of the chromidial
> One must be on the lookout for descriptions of mitochondria under the fol-
lowing headings also: chondriospiren and chondriorhabden (Benda), neuro-
somes (Held) karyochondria (Wildman), fuchsinophile granules, plasmafaden
(Retzius), plastidulen (Maggi), perinéme and pericaryonime (Renaut), sub-
stantia granulo-filamentosa, etc.
426 E. V. COWDRY
apparatus. Duesberg (710, p. 652) and I (’12, p. 497), among
others, have shown that they are two separate and distinct
substances. The wonder is that they could ever have been
confused for we have ample evidence that the chromidial sub-
stance (Nissl substance) is a nucleoprotein containing iron
(Scott ’05, p. 507), formed at least in part through the activity
of the nucleus, and the mitochondria a phospholipin albumin
complex. In this connection, also, must be mentioned the
attempts of Bouin (’05, p. 917) and others to identify mitochon-
dria with the previously discovered ‘ergastoplasme’ (proto-
plasme supérieur, kinoplasme, archoplasme, etec.). The term
‘ergastidions’ which Laguesse (711, p. 276) used for some years
instead of mitochondria, and later abandoned, is a relic of this
tendency. Regaud and Mawas (’09, p. 229) have vigorously
combatted the view that the mitochondria and ergastoplasme
are identical and the justice of their arguments is apparent
when we remember that the terms of ‘ergastoplasme’ and ‘chro-
midial substance’ are usually applied to one and the same material
Another instance is that of the reticular apparatus of Golgi‘
which Hoven (710, p. 479), Rina Monti (’15, p. 40) and others
believe to be, in some cases, identical with mitochondria. But
here the question is a far more complicated one, because it is
still impossible to define the reticular apparatus in any other
terms than in the appearance of cells fixed and stained by notori-
ously capricious methods of technique. The relation between:
the two substances cannot be profitably discussed before refine-
ments in technique are made and we learn more about the retic-
ular apparatus.
CHEMISTRY
It is an interesting and rather unusual occurrence, in the study
of mitochondria, for three independent lines of investigation to
yield similar results, yet Regaud (’08, p. 720), in the first place,
in the study of mammalian tissues; Fauré-Fremiet (’10, p. 622),
who worked on protozoa; and the botanist, Léwschin (’13, p.
4 Synomyms: apparato reticolare interno, binnennetz (Kopsch), netzapparat.
saftkanalchen and trophospongium (Holmgren)? spiremes (Nelis)? conduite
de Golgi-Holmgren (Cajal)? canalicular apparatus? ete.
FUNCTIONAL SIGNIFICANCE—MITOCHONDRIA 427
203; 714, p. 269), have all arrived at the conclusion that*mito-
chondria are chemically a combination of phospholipin and
albumin, which, in itself, speaks very strongly in favor of the
unity of the class of granules under consideration. The evidence
is briefly this:
1. Mitochondria are almost completely soluble in alcohol,
ether, chloroform and dilute acetic acid. They are rendered
insoluble by chromization. They are not doubly refractile
and they do not stain with Sudan III or Scharlach R. They
are only sometimes blackened with osmic acid.
2. It is said that part of the mitochondrial substance is not
soluble in these fat solvents and it is supposed that this portion
is albumin (see also Bullard 716, p. 26) for formalin and bichro-
mate, which are used as fixatives for mitochondria, are energetic
coagulants of albumin. Miullon’s reagent is the only color test
for protein which can be applied to material in sections. So far:
as can be ascertained? it is negative, but this cannot be stated
positively because, even if there were a change in color, it might
not be of sufficient intensity to be appreciated in filaments of
such extreme fineness as mitochondria (0.24 in diameter) em-
bedded in a colored cytoplasm. Mitochondria do not give any
of the color reactions of polysaccharides.
3. Artificial mitochondria have been made by Léwschin of
lecithin, and albumin solutions (resulting in the formation of
lecithalbumin?) which apparently present the same form and
solubilities as true mitochondria. They form granules, rods
and filaments which multiply by division. He embedded them
in glycerin-gelatin, fixed them and found that they stained in
the usual way by the various mitochondrial methods.°®
This evidence is good (being apparently accepted by the
Kochs 713, p. 427 and Mathews 715, p. 102, as far as the phos-
pholipin fraction is concerned) but it cannot be considered as
> Bensley, personal communication.
° Mayer, Rathery and Schaeffer (’14, p. 612) have been able to alter the mito-
chondria experimentally in liver cells. In stages with more mitochondrial sub-
stance chemical analysis shows an increase in phosphorized lipoid; in stages with
less, a diminution.
428 E. V. COWDRY
absolutely conclusive because it is subject to all the multi-
tudinous objections, which are very justly raised, against the
results of analyses of intracellular material. As yet no direct
chemical analyses of mitochondria have been made. The eggs
of fishes may prove favorable material because they are fairly
large, and, since their cytoplasm is quite hquid, the mitochondria
can be easily collected in a compact mass to one side by centri-
fuging, and, perhaps, be dissected out or removed by means of
a capillary pipette.
Hoppe Seyler pointed out that lecithin (a typical phospholipin )
and cholesterol are to be found almost anywhere that life phe-
nomena exist. In fact a great wave of revived interest is mani-
fested in recent chemical and pathological literature in these
complex compounds of fatty acid, phosphorus and _ nitrogen.
Mathews very aptly says that the phospholipins are the most
important substances in living matter:
For they are found in all cells, and it is undoubtedly their function
to produce, with cholesterol, the peculiar semifluid, semisolid state
of protoplasm. The latter holds much water in it but it does not dis-
solve. Indeed it may be said that the phospholipins with cholesterol
make the essential substratum of living matter.—This physical sub-
stratum of phospholipin differs in different cells and probably in the
same type of cells in different animals, but everywhere, from the low-
est plants to the highly differentiated brain cells of mammals and of
man himself, it possesses certain fundamental chemical and physical
properties. In all cases the phospholipin substratum is soluble in
aleohol containing some water: (715, p. 88).
In view of these considerations it is interesting to enquire
whether the distribution of mitochondria in cells corresponds
with that of the phospholipins. It is certainly true that mito-
chondria are more widely distributed than any other kind of
cytoplasmic granulation now known to us. They occur in
almost all cells. Certain cells, like the fully differentiated non-
nucleated red blood cell, undoubtedly contain a large amount
of phospholipin though no formed mitochondria can be seen.
The mitochondrial substance is probably present in solution
(Cowdry ’16), because it would be obviously absurd to state
that it must always occur in a certain state of condensation
which makes it visible with the microscope.
FUNCTIONAL SIGNIFICANCE—-MITOCHONDRIA 429
Chemically, then, we may for the present regard mitochondria
as being a combination, in varying amounts, of phospholipins
and protein. The phospholipins probably differ in quality
as well as in quantity and this is in all likelihood the case with
the protein also. It is probably the chemical basis of the per-
plexing differences in solubility and staining reaction, and to
a lesser extent, of the differences in form, which mitochondria
exhibit in different cells.
THE JANUS GREEN REACTION
That mitochondria in living cells’stain specifically with janus
green was originally discovered by Michaelis (’99, p. 565).
Furthermore, when the Janus green is reduced by the tissue,
a red diethylsafranin is formed which also colors the mitochondria
specifically. The delicacy of the reaction is shown by the dilu-
tion of the stain which will give it. I have found that mito-
chondria will stain in human lymphocytes in a dilution of janus
green in physiological saline of 1: 500,000. This is very remark-
able when one reflects that the mitochondria are only about 0.2u
in thickness. But the most significant fact is that the reaction
depends upon the presence of two ethyl groups in the safranin
portion of the janus green molecule. There are three janus
greens of the following formulae :7
1. Janus green (Griibler) safraninazodimethylanilin chloride
OOO TAT.
N
oS
v
’ Diazingriin S(K) and Halbwollgriin B(M) are also janus greens but it is
not clear which of the following formulae they possess.
430 * E. V. COWDRY
2. Janus green G(Farbwerke Hoechst Co.) dimethylsafraninazodimethyl-
anilin chloride ;
N
eon)
(CH3)2N Gy N=N N (CHs)2
N
/Nol
ea
|
VY
3. Janus green B(Farbwerke Hoechst Co.) diethylsafraninazodimethylanilin
chloride
N
ey xa
(C2Hs)2N L yy, N=N ‘ Part II of this paper will appear in the July issue.
447
448 WILLIAM A. LOCY AND OLOF LARSELL
It is commonly recognized by morphologists that our knowl-
edge of the development of the avian lung including its air-sacs
is both inadequate and defective in several important respects.
The notable observations of Schulze (11) and of Juillet (12) have
brought forward a newly recognized structural element, the re-
current bronchi, known only in the lung of birds, and which
imparts a renewed interest in the structural peculiarities of the
avian lung and the physiology of its air-sacs. It is now more
imperative than heretofore that we should have a review of the
embryological history of the lung with a more precise study of
the development of the bronchial tree, of the air-sacs and their
recurrent bronchi.
The assumption that, except for air-sacs, the lungs of birds
and of mammals are essentially similar as to architecture has
retarded the recognition of the structural peculiarities of the
bird’s ung. The beginnings are similar in these two classes of
vertebrates but the end-products are very different. There is
needed an embryological study to determine the way in which
the avian lung departs from the mammalian type and to de-
termine the precise nature of the intercommunications between
its bronchioles. The development of recurrent bronchi from the
air-sacs and the establishment of labyrinthine communications
between all parts of the bronchial tree, imparts to the avian lung
a unique architecture not found in any other class of vertebrates.
There is no ending of the ultimate twigs of the bronchial branches
in culs-de-sac, as in the lungs of other vertebrates, so that the
facilities for ventilation of the avian lung are very complete.
The absence of alveoli in which a portion of the air is retained as
residual air, permits the air current to sweep unimpeded through
the minutest air passages and affords great opportunities for
respiratory exchanges between the blood capillaries and the air
capillaries. The air-sacs receive their supply through direct ori-
fices, during inspiration, and the air passes from these reservoirs
into the lung by way of the recurrent bronchi during expiration.
It is essential to understand the intereommunications of the air
passages in order to comprehend either the morphology or the
physiology of the bird lung.
THE EMBRYOLOGY OF THE BIRD’S LUNG 449
In early stages of development the outgrowths of the bron-
chial tree end blindly and this condition is maintained in the
adult mammalian lung, but in the bird lung, the terminals come
into contact and anastomose during embryonic development so
that in the adult lung there are no culs-de-sac. This condition
of anastomosis affects also the air capillaries that are radially
arranged around the parabronchi. Thus in following the devel-
opment of the air passages of the bird’s lung we pass from the
primary condition of a bronchial tree to the modified condition
of uninterrupted bronchial circuits.
Some of the points that require elucidation for understanding
the morphology and the physio'ogical action of the avian lung
may be enumerated:
As a background, a knowledge of the phenomena of extra pul-
monary development, or the general course of its embryology.
The intra-pulmonary development of the bronchial tree, its
ramifications and the establishment by anastomoses of unbroken
communications between the parabronchi and the air capillaries.
To determine the method of formation of the air-sacs and of
their outgrowths, the recurrent bronchi.
To observe the formation of the air capillaries and the estab-
lishment of anastomoses among them.
To observe the origin and mode of development of the pul-
monary artery and of the general course of circulation within the
lune.
In addition to the above there should be observations on the
diaphragmatic membranes and the muscular means of producing
respiratory movements accompanied by physiological experi-
ments, but observations of this nature have not been included
in our studies.
Our observations are confined to the embryology and mor-
phology of the lung and air-sacs, and in this study of limited
range we do not presume to have found answers to the ultimate
questions of morphology of the bird lung. We have assembled
our results merely as an objective account of what we have been
able to observe in the time and with the material at our disposal.
450 WILLIAM A. LOCY AND OLOF LARSELL
The observations are brought under consideration in the fol-
lowing order:
1. The external aspects of lung development.
2. The development of the bronchial tree.
3. The air-sacs and the recurrent bronchi.
4. The development of the pulmonary artery.
Followed by comments on the steps of progress in the ana-
tomical analysis of the bird’s lung and comparison of some of
our results with those of previous observers.
Comments on the literature. In dealing with the extensive lit-
erature of the avian lung one is confronted with the dilemma. of
choosing between a comprehensive chronological mention of the
observations of the different investigators or a very condensed
selective review of the results of a few workers. The latter plan
on the whole seems better, since the literature has been repeatedly
reviewed (as in Flint’s contribution, ’06, in Juillet’s, 712, and in
the papers of others); moreover, genuine advances are contained
in a limited number of papers.
As to embryological observations, the chief contributions are
by Rathke, ’28; Von Baer, ’28; Remak, ’55; containing the first
figures of the buds of the ecto- and entobronchi; Selenka, ’66, on
development of the air-sacs; His, ’68, laryngo-tracheal groove and
trachea; Zumstein, ’00, bronchial tree and air-sacs; Moser, ’02,
method of growth; Bertilli, ’05, air-saecs; Juillet, ’12, compre-
hensive treatise; besides text-books, as Foster and Balfour ’74;
Marshall, ’93; Lillie, ’08, ete.
As to intra-pu monary anatomy of adult stages: Sappey, ’47;
the bronchial passages especially analyzed by Campana ’75; Hux-
ley, 83 bringing the terms mesobronchium, ecto, ento, and para-
bronchia into common use; F. E. Schulze, ’09, 710, *11, bronchial
tree and air-sacs; Miller 93, comparative structure of lungs in-
e:uding birds; Guido Fischer, ’05; Juillet, ’12.
Histology: F. E. Schulze, ’71; Oppel, ’05.
The air-sacs have been extensively described in the adult with-
out involving the anatomy of the lungs as by Guillot, ’75, com-
parative; Bruno Miller, ’07, pigeon; Schulze, ’10, ete.
THE EMBRYOLOGY OF THE BIRD’S LUNG 45]
As to methods of growth: Aeby, ’80, monopodial; Miller, 793,
comparative, budding predominates in birds, septum formation
secondary; Moser, 702, budding the uniform principle of growth
in birds and other vertebrates; Flint, ’06, paper on mammals but
reviews the literature on other vertebrates and comments on the
method of growth in birds.
Campana’s thorough and extensive paper of 1875 requires sepa-
rate mention. It is part of a general plan designed to illustrate
the laws of genesis and evolution, and the primary title of his
memoir is Physiology and Respiration of Birds. Nevertheless,
the anatomical part is of chief importance, and it 's the most
critical and comprehensive treatise on the structure of the adult
bird’s lung to which we have had access. This memoir is not
easily accessible, and although it is commonly mentioned in the
literature lists, it has, unfortunately, been little read. Campana
makes an illuminating analysis of the bronchial passages, tracing
their ramifications in detail and making an especial point of the
bronchial circuits which unite the various divisions of the bronchi
into a plexus of intercommunicating passages. He also noticed
the recurrent bronchi but without understanding their signifi-
cance. Further mention of this point will be made later, and,
also, his classification of bronchi will be explained in our section
on the bronchial tree.
IF. E. Schulze in 1911 published an important paper on the com-
parative anatomy of the air-sacs in the adult and for the first
time (’09) deseribed the recurrent bronchi and pointed out their
physiological office. His excellent methods of injection with
metal and celloidin are described in detail.
The most recent important contribution to the morphology of
the bird’s lung is the paper of Juillet published in 1912. This is
a comprehensive treatise embracing an anatomical, embryologi-
cal, histological and comparative study of the avian lung. It
contains a review of previous work and a list of the literature.
Its most significant feature is the description of recurrent bron-
chi (discovered by Schulze, 09 and 711) growing from the air-
sacs into the lungs and anastomosing with the parabronchi of
ecto, ento and laterobronchi. He used metallic injections of
452 WILLIAM A. LOCY AND OLOF LARSELL
Wood’s metal and Darcet’s metal besides plastic reconstructions
and the usual embryological and histological methods. An ade-
quate review of this excellent paper would require much space
and it should be read in the original. Our observations differ in
some particulars from those of Juillet (especially as regards the
origin of the interclavicular air-sac) and these differences will be
commented on later.
Technique. Chick embryos from the close of the second day
up to the time of hatching were used in observations on the de-
veloping lung. The stages were compared with the figures in
Duval’s Atlas d’Embryologie and his chronology adopted.
For dissecting, fresh embryos were first immersed in a solution
of 8 per cent formalin and preserved in a 4 per cent solution.
While the heart was still pulsating a large number of the embryos
for dissection were injected with India ink through the vascular
area or the liver according to the age of the embryo. Dissec-
tions were made of stages from three days to hatching, of young
chicks one, two and three days after hatching and of adults.
For imbedding, the embryos were fixed in Kleinenberg’s picro-
sulphuric solution and in formalin. Stages from 48 to 96 hours
were sectioned from eight to ten microns in thickness and sections
were also made of older stages and of the lung parenchyma of
the adult.
It would have been impossible to work out with any degree of
satisfaction the development of the bronchial tree and of the re-
current bronchi without the use of a method originated by Hoch-
stetter (Zeitschr. fiir Wissenchft, Mikr. und Mikr. Tech., Bd.
XV, 798) of using clove oil and chloroform. This method was
modified by using rather thick cedar oil instead of clove oil which
ras found to give clearer preparations and those of longer dura-
tion.
In stages subsequent to 96 hours, the lungs and air-sacs were
dissected out of the previously fixed and hardened specimens,
then cleared in cedar oil, after which the organs were placed in
a mixture of one part cedar oil and two parts chloroform. On
becoming permeated with this fluid, the preparation was re-
moved from the mixture and placed on a filter paper until the
THE EMBRYOLOGY OF THE BIRD’S LUNG 453
chloroform might evaporate. The evaporation of the chloro-
form served to draw the cedar oil from the lumina of the various
branches of the bronchial tree into the lung tissue and to fill
the spaces thus made with air. When this preparation was re-
placed in pure cedar oil, the difference between the refractive
index of the imprisoned air and the surrounding medium gave
the lung tubes the appearance of being filled with a metallic cast.
Thus the minute air passages that could not be injected by other
means were made clear. The finer details would disappear af-
ter a few minutes as the cedar oil percolated into them, but the
same specimen, if carefully manipulated can be treated repeat-
edly without apparent injury, and a complete picture could fin-
ally be obtained. This method was successfully used in tracing
the development of the bronchial tree up to the eighteenth day
of incubation.
For later stages, celloidin and Wood’s metal injections af-
forded the most helpful preparations. Such injections were also
attempted of earlier stages, but the uncertainty of successful
preparations and the destruction of the specimen employed
made the air injections more satisfactory. This was especially
true since the air injections showed fine points of detail that were
not revealed by the more limited penetration of the fluid cel-
loidin and the heated Wood’s metal. Several preparations of the
adult lung were made with the Wood’s metal injections. Pre-
ceding the Wood’s metal casts, the lungs of the freshly killed
fowl were distended under pressure with 80 per cent alcohol
until the air-sacs were fully expanded, after which the entire
bird was immersed in alcohol for twenty-four hours or more be-
fore attempting metallic injection.
For histological study of the air capillaries of the adult bird,
the pulmonary apparatus was injected under pressure with cor-
rosive acetic fluid and by this means the lungs were fixed in a
distended condition and the air capillaries were not collapsed.
454 WILLIAM A. LOCY AND OLOF LARSELL
aE CLERE pry
Fig. 1 Cross section through pharynx and lung pouches of a chick embryo
incubated 51 to 52 hours.
Fig. 2. Similar section of a slightly older embryo (52 to 53 hours) showing the
well defined lung pouches. Figures 1 and 2 drawn by Gilbert H. 8. Rech.
Fig. 3 Two consecutive sections through the pharyngo-tracheal groove of
the same embryo. These sections are respectively 120 and 128 microns in front
of the one sketched in figure 2.
Fig. 4 Sections through the same region of an embryo incubated 55 to 56
hours.
THE EMBRYOLOGY OF THE BIRD’S LUNG 455
1. THE EXTERNAL ASPECTS OF LUNG DEVELOPMENT
Under this heading the external features of lung formation
will be described while the intra-pulmonary changes will receive
separate consideration in the following section.
The time, the place and the method of formation of the primi-
tive lung of the chick has been well described by various ob-
servers. In reference to the time, it should be remembered that
in all embryonic development there is individual variation as
well as variable methods of estimating stages. It is not, how-
ever, So important to establish an exact correspondence in chro-
nology of different observers as to determine the method of lung
formation and the normal sequence of changes.
The first external appearance of the lung of the embryo chick
comes in the early part of the third day. Many specimens of
30-31 somites show a slight ridge-like enlargement on each side
of the latero-ventral surface of the pharynx just behind the
fourth gill-pouch. This is in the narrowed respiratory divi-
sion of the pharynx, as distinguished from the broadly expanded
branchial division.
Cross sections show that the ridge-like formation is owing to
an evagination of endoderm into the surrounding mesenchyma.
Figure 1, from a specimen of 30 somites, estimated as 50 hours’
development, is cut through the more prominent part of this out-
growth. Figure 2 is from a slightly older specimen, estimated
as the 52-hour stage. Both sketches are from camera lucida
tracings, so that the outlines are correctly represented, but in
finishing, the details, especially the nuclei of cells, have been
made diagrammatic. The shallow pockets on the ventral bor-
der of the pharynx are the beginnings of lung pouches; they push
out into the mesenchyma which is bordered by a very pronounced
mesothelium. At their beginning, therefore, the primitive lungs
are paired, and consist of two shallow pouches that open widely
into the floor of the pharynx. The surrounding mesoderm is
also a part of the lung anlage and increases pari passu with the
growth of the endodermal part. The endoderm by budding gives
rise to the lining membrane of the bronchial tree, the mesoderm
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, NO. 3
456 WILLIAM A. LOCY AND OLOF LARSELL
providing material for blood vessels, lymph spaces, muscles, con-
nective tissue and like elements, while ingrowths from the ecto-
derm provide the nerve supply.
In front of the bulges the walls of the pharynx are compressed
laterally and the tube is narrowed on its ventral border to form
the laryngo-tracheal groove which is the forerunner of trachea
and larynx.
Figure 3, showing the pharynx and the ventral groove, is a
camera tracing of two consecutive sections taken 120 and 128
microns in front of the one sketched in figure 2. In still earlier
stages the narrowing of the respiratory portion of the pharynx
is easily seen as well as the incipient stages of the bulges from
which the lung pockets are produced.
The cavity of the pharynx is also narrowed just above the
lung pouches (fig. 4) so that in cross section, the outline is simi-
lar to that of figure 3 in an inverted position.
Immediately the lung pouches begin to elongate by growth of
the endodermal lining in a caudo-lateral and somewhat dorsal
direction, and prior to the 60th hour, their divergent distal ends
become separated from the oesophagus (fig. 4). By the end of
the third day (72 hours) the embryonic lung can readily be ex-
posed by dissection.
At the close of the fourth day (96-hour stage) the lungs and
adjacent territory present the appearance shown in figure 5. At
this stage the lungs are small, smooth pouches extending cau-
dally and dorsally along each side of the oesophagus. In this
specimen, the sixth arch, from which the proximal end of the
pulmonary artery is shortly to develop, has not been completed
although a ventral spur of the arch is shown and a shorter dorsal
spur from the aorta.
Figure 6 represents a ventral view and figure 7 a lateral view
of the lung of a chick embryo near the close of the fourth day
of development. The lung pouches are divergent and their dis-
tal portions extend caudad, laterad and dorsad. Their cavities
are lined by the endodermal diverticula from the pharynx, and
these are surrounded by mesoderm, so that the surface exposed
by dissection is mesodermic. The walls of the endodermal tube
THE EMBRYOLOGY OF THE BIRD’S LUNG 457
do not as yet show any buds (fig. 24). Although the differen-
tiation of the trachea has begun it is not visible from surface
views. Seen from the lateral aspect as in figure 7, the lung
pouches, closely united with the walls of the oesophagus, pass
below it and unite with the laryngo-tracheal ridge on the ventral
rf
Fig. 5 Dissection of an ink injected chick embryo of 96 hours incubation ex-
posing aortic arches and the left lung. Drawn by G. H. A. Rech.
““ Figs. 6 and 7 Surface views of the lungs of a chick embryo at the close of the
fourth day of development. Figure 6 from the ventral and figure 7 from the
lateral aspect.
458 WILLIAM A. LOCY AND OLOF LARSELL
part of the pharynx. The lung pouches are smooth and do not
as yet exhibit surface irregularities.
Injected specimens of this stage frequently show blood vessels
running along the ventral surface of each lung and uniting in the
median plane at a point where the lung pouches join the pharynx
(fig. 6). From this place of union a vessel leads into the left
atrium of the heart. Another vein, coming from the front,
passes along the ventral surface of the laryngo-tracheal groove
and joins the stem vessel that leads into the left atrium. The
blood vessels on the lungs are the beginnings of pulmonary veins
and they are commonly injected before the pulmonary artery is
established. Sections show however that vascular spaces for the
formation of the distal extremity of the pulmonary artery are
already present in the lung walls.
In the closing hours of the fourth day the trachea becomes
differentiated from the posterior portion of the laryngo-tracheal
groove. It may be definitely distinguished in an embryo with
39 somites (estimated as in the 94-hour stage), and, by the 100th
hour, it is well defined. This is not readily evident in surface
views but in optical section (fig. 25, 4 days, 4 hours) the connec-
tion of the trachea with the pharynx and with the bronchi is
well exhibited. At the distal end of the lung tube is an enlarge-
ment that foreshadows the abdominal air-sac. At its proximal
end, on each side, a short portion of the bronchus lies between
the anterior limits of the lung and is the first appearance of the
extra-pulmonary bronchus. These extra-pulmonary bronchi join
the trachea which is of larger calibre than the bronchi. The
oesophagus makes a rather abrupt dorsal bend away from the
trachea, and then, with a more gentle curvature continues cau-
dally and, bending downward, passes between the lungs.
During the fifth day the lung grows larger and begins to show
surface irregularities. Figure 8 shows the appearance of the left
lung territory as exposed by dissection in a specimen of the 45-
day stage. The lung pouch of this specimen has grown dorsally
so as to extend across the path of the aorta. Its distal extremity
exhibits a protuberance which is the beginning of a lobe in which
lies the expanded end of the mesobronchus. At about this
THE EMBRYOLOGY OF THE BIRD’S LUNG 459
stage the pulmonary artery is usually established and, in injected
specimens forms one of the external anatomical landmarks. In
the subsequent descriptions the pulmonary artery will be included
as a feature of external anatomy but the details of its formation
are separately considered under another heading.
In the specimen sketched in figure 8, the proximal end of the
pulmonary artery shows as a spur from the sixth aortic arch, but
owing to imperfect injection, the distal part that is formed in the
lung wall is not seen. The rudimentary fifth arch is present in
this specimen as a short vessel arising from the truncus arteriosus
and joining the lower half of the sixth arch. Much variation
exists as to the presence or absence of the rudimentary fifth
arch and as to its dimensions when present. The degree of de-
velopment of the pulmonary artery also varies in different speci-
mens of this age.
Figure 9 represents a side view, and figure 10 a ventral view of
the lung in the last half of the fifth day of development. The
lobe at the posterior end of the lung pouch is shown in figure 9
and the trunk of the pulmonary is fully established. When
viewed from the ventral surface (fig. 10) the right lung forms a
somewhat greater angle (fig. 6) with the oesophagus than the left,
producing an appearance of asymmetry. This asymmetry, how-
ever, is not owing to a difference in size of the tung (as in mam-
mals and some reptiles) but to the pressure of the stomach (ven-
triculus) enlargement which begins at about this period. The
greatest asymmetry comes about the middle of the fifth day; it
is gradually rectified with the change in relative position of the
viscera and the symmetry is restored by the eighth day. In
figure 10 the pulmonary veins and the laryngo-tracheal branch
are also shown on the ventral surface of specimen.
Near the close of the fifth day of development well injected
specimens (fig. 11) show a network of blood vessels near the sur-
face occupying a small area on the anterior dorsal part of the
lungs. Sections and transparencies of this stage show that the
network of vessels within the lungs occupies chiefly the dorsal
region and is more extensive than appears from the surface.
This figure shows also a branch from the pulmonary artery ex-
460
WILLIAM A. LOCY
AND OLOF LARSELL
B,
J
Wg
Fig. 8
4> days development (114 hours). Drawn by G. H. A. Rech.
Fig. 9
Fig. 10
metry of the lungs.
Dissection exposing the lung and adjacent territory in a specimen of
Side view of a dissection of the lung territory during the last half of
the fifth day of incubation.
Ventral aspect of the same specimen illustrating the apparent asym-
] £ i
THE EMBRYOLOGY OF THE BIRD’S LUNG 461
tending towards the trachea and passing through a network of
capillaries which communicate with the vein, mentioned above,
as running on the ventral surface of the laryngo-tracheal groove.
The immediately following external features of development
may be rapidly passed over since, for some time, there is no sig-
nificant change in the external appearance of the lung.
Figure 12 shows a dissection of the lung territory in an embryo
of 53 days incubation and figure 13 a similar dissection of an
embryo during the last half of the sixth day. In both these fig-
ures the anterior dorsal (cephalic) part of the lung is protuberant
and the hook-like process at the caudal extremity is more evi-
dent than in earlier stages. They both exhibit the course of the
pulmonary artery and pulmonary vein as seen in surface views,
and figure 12 also shows in addition, a short ventral spur from
the pulmonary artery.
A more comprehensive view of the superficial blood vessels of
the lung is shown in figure 14, sketched from a specimen in the
early part of the seventh day. The shape of the lung and the
external appearance of both pulmonary artery and pulmonary
vein are well shown. Especially to be noted is the trunk of the
vein on its way to enter the left atrium of the heart, and the
juncture with this trunk of the pulmonary vein and of the vein
(larvngo-tracheal) running along the ventral surface of the
trachea. A short arterial branch leavesthe pulmonary artery
on its ventral border in front of the lung and, passing through a
capillary network, connects with the laryngo-tracheal vein. The
anterior part of the lung above the pulmonary artery shows a
superficial network of blood capillaries.
Although there is relatively little change in the surface appear-
ance on the sixth day, it is to be understood that internal changes
of great significance are taking place. The first branches of the
bronchial tree arise on the sixth day of development and the
network of internal capillaries is moulded over them. These
internal changes are described in later divisions of this paper.
The seventh day stages, as seen from the side, show the lung
approximating a rectangular outline with a protuberance from
ys WILLIAM A. LOCY AND OLOF LARSELL
~
a
Fig. 11 Dissection of the lung territory of an injected specimen near the
close of the fifth day of development, showing network of blood vessels on the
anterior dorsal area of the lung, also the opening of the pulmonary vein into the
left atrium.
Fig. 12
Increase in size of lung and of pu'monary artery is evident.
trally from the pulmonary artery just above the heart is a small blood vessel
which anastcmoses through a capillary network with the laryngo-tracheal vein
as shown in figure 14. Modified from a sketch by G. H. A. Rech.
Vig. 18 Dissection of the left side of an injected embryo of the last half of
the sixth day of development. Projections on the lung, pulmonary artery and
pulmonary vein shown. Drawn by G. H. A. Rech.
Dissection of the left side of an injected embryo of 53 days incubation.
Projecting ven-
THE EMBRYOLOGY OF THE BIRD’S LUNG 463
the cephalic end and another at the caudal extremity. These
mark the points of emergence of the cervical and of the abdomi-
nal air-sacs. Figure 15 A and B, from a specimen of the last
half of the seventh day, show the surface of the left lung and of
the right lung of the same embryo. Early on the seventh day
the abdominal air-sac projects beyond the border of the lung
proper, but the cervical lags behind the abdominal in its devel-
opment. ‘These are the first two sacs to develop and the others
follow shortly except that the posterior intermediate sac is the
last to emerge outside the lung wall. The course of the pul-
monary artery and of the pulmonary vein is shown in both lungs.
Pul.Vn.
Fig. 14 Lung territory of a well injected specimen of the early seventh day
of development. Note especially the laryngo-tracheal artery and its capillary
network succeeded by a vessel that connects with the trunk of the united pulmo-
nary veins.
The net work of blood vessels of the anterior dorsal surface of
the lung was not easily seen in this specimen, partly on account
of imperfect injection and partly because the outer covering is
thickened, but in heavily injected specimens, the network is
seen (as in fig. 14) to occupy the dorsal anterior half of the lung.
The eighth and ninth days are important periods in the em-
bryonic development of the lung, not only on account of inter-
nal changes, but also because the air-sacs emerge, and on the
ninth day of development, project beyond the surface of the
lung, and thus one of the characteristic structural features of the
bird lung is established.
Figure 16, A and B, shows the surface appearance of the lungs
on the ninth day of development. The lung has increased in
464 WILLIAM A. LOCY AND OLOF LARSELL
dimensions dorso-ventrally and when viewed from the side is
rectangular in outline. It also occupies a more lateral position
in the thoracic cavity. The lung has begun to press against the
ribs and exhibits shallow furrows where the lung substance has
grown around the bodies of the ribs. The five air-sacs, two on
the anterior and three on the ventral margin, are formed and
project beyond the surface of the lung. As indicated above, the
>
NG
3
Fig. 15 (A) Dissection of the left side of an embryo during the last half of
the seventh day of incubation, the fourth aortic is much atrophied on this side.
(B) Heart and right lung of the same specimen. The fourth aortic arch is larger
on this side and separated from the third. Projections of the cervical and ab-
dominal air-sacs are exhibited. Drawn by G. H. A. Rech.
posterior intermediate is the last of the air-sacs to expand and
project beyong the lung wall. The cervical, and interclavicular
sacs are smaller, but both project from the lung wall earlier than
the posterior intermediate. The part of the interclavicular sac
showing in figure 16 is only the lateral moiety of the sac, the
mesial moiety, which at this stage ‘s separate and independent,
can not be seen from this aspect.
THE EMBRYOLOGY OF THE BIRD’S LUNG 465
Well injected specimens of this age show that the blood supply
predominates in the dorsal region of the lung. After the air-
sacs are well projected their walls are relatively thin and they
do not exhibit any blood vessels that will take the India ink in-
Fig. 16 (A) Dissection exposing the left lung of an embryo of the ninth day
of incubation showing five air-sacs projecting from the lung. The mesial moiety
of the interclavicular is hidden from view. (B) Heart and right lung of the same
specimen. Modified from a sketch by G. H. A. Rech.
466 WILLIAM A. LOCY AND OLOF LARSELL
jection. As is well known, their blood supply in later stages is
derived from arterial branches coming from the aorta. The
superficial distribution of blood vessels is shown in figure 17 which
is from a specimen somewhat younger than that sketched in figure
16. It is to be understood that the internal plexi of capillaries
are extensive and the blood vessels represented in figure 17 are
those visible through the translucent walls of the lung and are
mentioned here merely as a feature of external anatomy.
Figure 18 is a surface view of both lungs of an embryo at the
close of the ninth day of development. The bronchus of the
right side has been severed and the right lung rotated so as to
expose more fully the mesial facet. The five air-sacs are now
well projected beyond the lung wall andin this figure a new struc-
tural feature is brought into evidence. This is the mesial moiety
(Mes.moz.) of the interclavicular air-sac. At this stage it is con-
nected through the interclavicular canal with the anterior in-
termediate air-sac, and the mesial moiety is widely separated
from the lateral moiety. At a later period (fifteenth day, fig.
51) the mesial moiety comes into contact with the lateral moiety
and subsequently the two moieties fuse into one sac. In the
published sketches of surface views of embryonic stages (with the
exception of a figure by Selenka, ’66) the mesial moiety has not
been represented. It is commonly hidden from view between
the two lungs. There are however some published sketches of
section of the lungs, as Lillie, ’08, Juillet, ’12, etc., in which it
-has been represented as a forward projecting diverticulum of
the anterior intermediate air-sac, but in these sections it has
heretofore been interpreted as a portion of the anterior inter-
mediate sac. Also in a diagram of Juillet, ’12, (cf. his fig. X, p.
313) the mesial moiety of the lung of the embryonic chick is
represented to the exclusion of the lateral moiety. In the genus
Larus, even in the adult, a separate lateral sac of the interclavicu-
lar is present in addition to the mesial portion of that sac (Juillet’s
fig. XVII, p. 351). For the further history of these moieties of
the interclavicular air-sac see figures 47, 49, 50 and 52 and the
accompanying comments.
THE EMBRYOLOGY OF THE BIRD’S LUNG 467
Fig. 17 Camera tracing of the capillary network on the dorsolateral surface
of the left lung of an embryo at the close of the eighth day of incubation. Drawn
by Mary Head.
Fig. 18 Surface view of the lungs of an embryo at the close of the ninth day
of development. Right bronchus cut and lung rotated so as to expose the mesial
surface. Abd.sc., abdominal air-sac: A.intr.sc., anterior intermediate air-sac, with
primordia of recurrent bronchi (Rec. Br.) ; Cerv.sc., Cervical air-sac; Lat.mot., lat-
eral moiety and Mes.moi., mesial moiety of the interclavicular air-sac; P.intr.sc.,
posterior intermediate air-sac.
468 WILLIAM A. LOCY AND OLOF LARSELL
The obvious external feature of the tenth day is the indenta-
tion of the ribs on the dorso-lateral border of the lungs. Figure
19 shows a dissection exposing the right lung and adjacent or-
gans of an embryo at the beginning of the eleventh day. There
are four well marked indentations of the ribs. As mentioned
above, during the ninth and tenth days the lungs undergo a
change in position passing from a more ventral to a more dorsal
position, and in so doing come close against the ribs, and the
dorsal margin of the lung comes to lie along the vertebral column.
The air-sacs are enlarged and the proximal ends of the two pos-
terior ones are constricted to form a sort of neck. The abdomi-
nal air-sac has increased relatively faster than the others.
The trunks of recurrent bronchi are also shown in connection
with the two posterior air-sacs. The recurrent bronchi are the
most important structural feature that we have yet had occasion
to mention. They begin on the ninth day as buds from the
proximal ends of the abdominal and the posterior intermediate
air-sacs, and, later, the other air-sacs, with the exception of the
cervical, g-ve rise to similar ourgrowths ‘They are destined to
develop ramifications that anastomose with parabronchi in vari-
ous parts of the lung and play a very important part in its physi-
ology. They are so important that they receive separate treat-
ment in section 3 to which reference should be made for figures
and further details.
Figure 20 is a diagram made from a study of the left lung of
an embryo, incubated 95 days, to show especially the relations
of the mesial and the lateral moieties of the subbronchial sac at
this stage of development. In the preceding sketches (except
figure 18) the mesial moiety of the interclavicular sac has not
been shown chiefly because the sketches were executed before
we had learned to look for the two moieties of this sac, and,
further, the aspect from which the specimens were drawn did not
bring the mesial moiety into view.
The diagram (fig. 20) was made from observations by re-
flected as well as by transmitted light and the connections of the
alr-sacs with the bronchi are indicated. On the lateral border
of the lung is seen the lateral moiety (Lat.moz.) of the interclavic-
THE EMBRYOLOGY OF THE BIRD’S LUNG 469
ular sac connected with the transverse branch of the first ento-
bronchus. The mesial moiety (Wes.moz.) arises in connection
with the anterior intermediate air-sac, and these two have a
common opening into the third entobronchus. The distal end
of the mesial moiety is forked and partly encircles the main
bronchus.
/ \ en
. \ a
\
HY MS Hf
/ \ ! =
/ ; i oe ~\
NY
/ \ ! i i)
/ \ 1% {
/ \ \ G
/ ‘ ‘ \ ae
rf \ \ \
, aN ‘ ak
/ \ me
4 g N \ ~
\ ~
' ‘ \ ‘. See
a sees > \ iV era a
/ Th : \ \ \)
' “———['ra, \ \ \'
i] . \ \ ' oo
] Pe \ \\ 7
! \ \ 4i Pit
' \ t a
Vic yLat.mol. \\ art -
Gerv. (ones N 1 | Pe.
NS \ 41 /
yi ; 2A.Int-Sc. ws x) ft /
: é \ Sh \e / ‘
: ge ‘ : \ Sa \ _
! ao See
t Rs VotN
Lg.-=: x NS
“ Boalt
crn | My
\
- \
iRec.Bra= \ ;
pas AY ts \
‘ RSENS of \ \ ‘
1P.Int.Sc,- S33 ts \
SATE i \ \
| ' rs ¥ \ ~~ \
\
\ Abd.Sc = Secon aaa ie \ SS
\ \ . a
1 ( va
\ =
\ AS ; rh
\ 4
\ : 7
\ aw ’
. -
\
A.lnt-Se.s<
Fig. 23 Enlarged view of the lungs of the specimen sketched in figure 22
removed from the body. Five air-sacs are shown. The mesial moiety of the
interclavicular not visible. Modified from a sketch by G. H. A. Rech.
blance is closer to the capillary connection between arteries and
veins than to a tree.
From their earliest formation, the lung pouches are lined by
endoderm and this internal cavity is the basis from which hollow
buds arise to form branches of the bronchial tree. The endoder-
mal tube lies in a layer of mesenchyme that is bordered on the
surface towards the pleuro-peritoneal cavity by a well defined
THE EMBRYOLOGY OF THE BIRD’S LUNG 475
layer of mesothelium. Accordingly, the external boundaries of
the lung are formed by a wall of mesoderm which gives no indi-
cation of the internal configuration of the endodermal lining.
The formation of the branches of the tree must be traced by re-
constructions or by air injections, since the study of the lungs as
transparencies, without such injections, is very limited in its ap-
plication and does not show the outgrowths with any degree of
completeness as to detail.
At the 96-hour stage the simple cylindrical tube of endoderm
extends into the mesenchyme of the lung primordium and is
slightly expanded at its distal extremity (fig. 24, A and B).
The trachea has already begun to differentiate from the pharyngo-
tracheal groove, and in a few hours’ time is definitely formed.
The stomach enlargement (ventriculus) pushes on the left lung
so as to throw it out of alignment and give an appearance of
asymmetry to the lungs.
At the proximal end, a short portion of the lung tube lies out-
side the boundary of the mesenchymal swelling of the lung pouch
and forms the primordium of the extra-pulmonary bronchus
(fig. 24).
By the 100-hour stage the trachea is distinctly formed (fig.
25); it is of larger calibre than the bronchi that connect with it.
The oesophagus curves abruptly from it in the dorsal plane, and
then, with a more gentle curvature, it passes caudad for some
distance and finally curves ventrally and passes into the stom-
ach enlargement between the lungs.
On the second half of the fifth day of development a spindle-
shaped expansion arises within the lumen of the lung tube (fig.
26). This is a convenient anatomical landmark of the central
lung tube of the embryo and may be designated as the embry-
onic vestibulum. It does not correspond however to the vestib-
ulum of the adult, into which the entobronchi open, neither is it
the genetic forerunner of the adult vestibulum. It is a relatively
thin-walled dilation of the central lung tube located further cau-
dad than the adult structure of that name. Figure 26 shows its
dimensions at four days and twenty hours’ incubation. Its po-
sition varies from about two-fifths to two-thirds the space from
476 WILLIAM A. LOCY AND OLOF LARSELL
Fig. 24 Side and ventral view of the central lung tube of an air injected
specimen of the 96-hour stage. Shows an early stage in the development of the
trachea.
Fig. 25 Ventro-lateral view of the air injected lungs of an embryo incubated
100 hours. The trachea is definitely formed.
Fig. 26 Cedar oil transparency of the lungs of a specimen incubated 4 days
20 hours, showing the expanded embryonic ‘‘vestibulum’’ and the occluded por-
tion of the bronchus.
THE EMBRYOLOGY OF THE BIRD’S LUNG 477
the anterior to the posterior end of the lung tube. The ectobron-
chi arise from this expanded region. Later the more cephalad
region of the lung tube increases in diameter and, owing to un-
equal growth, the embryonic dilation merges into the rest of the
lung tube and is no longer conspicuous. The region of the adult
vestibulum is, in reality, occluded at this stage of development.
As shown tn figure 26, the greater part of the future lumen be-
tween the trachea and the anterior end of the lung becomes oc-
cluded early on the fifth day and this condition persists through-
out that day of incubation. This occluded portion embraces
the region of the adult vestibulum.
The embryonic dilation referred to was first figured by Selenka
in 1866. Its rise served to divide the central lung tube into
three regions—an interior, a middle and a posterior. The lung
tube as a whole is in the shape of an elongated S and the dilata-
tion of the middle region arches so as to come close to the dorsal
surface of the lung.
Shortly after the appearance of the embryonic vestibulum, the
primary lung tube, early on the sixth day, begins to give off buds
which form the beginnings of the secondary branches of the bron-
chial tree. The first bud to be formed (fig. 27, ent. 1, 5 days, 9
hours) is from the internal (mesial in the adult) wall of the intra-
pulmonary bronchus This is the primordium of the first ento-
bronchus. Behind this (in the same figure) is the smaller bud
of the second entobronchus. It will be noted that both are in
front of the embryonic vestibulum.
Two similar hernia-hke enlargements follow in quick succes-
sion and form the primordia of the third and fourth entobronchi.
It results that, at the stage of five days twenty hours’ incubation
there are present four bladder-like outgrowths (fig. 28, ent. 1,
2, 3, 4) which are connected with the bronchus by slender stalks.
The third and fourth have their attachment tothe bronchus
somewhat more mesially than do the first and second. These
four entobronchi, although arising on the internal wall, curve
around the bronchus as they grow so as to oceupy the ventral
face of the lung.
478 WILLIAM A. LOCY AND OLOF LARSELL
Since the bronchial branches are known under a variety of
names, it will be advantageous before proceeding further to in-
dicate the nomenclature employed in this paper. The different
secondary outgrowths from the primary lung tube have the same
histological structure, but they vary in position, in size, and in
the profusion and distribution of their branches.
THE EMBRYOLOGY OF THE BIRD’S LUNG 479
Sappey, in 1847, described two principal kinds, the diaphrag-
matics (bronches diaphragmatiques), branching towards the ven-
tral face, and the costals (bronches costales), extending dorsally
below the ribs.
In 1875, in his extensive memoir, Campana designated three
categories: primary, secondary and tertiary bronchi. The pri-
mary (La bronche primaire, ou Souche) is the central lung tube
commonly called mesobronchus. He distinguished four groups
of secondary bronchi which we have adopted with some modi-
fications. Campana’s divisions are: (1) The system of five large
divergent bronchi (the diaphragmatiques of Sappey); (2) The
system of eight internal bronchi (the costales of Sappey); (3)
The system of six external bronchi, for which we have adopted
Schulze’s term of laterobronchi (lateribronchi); (4) The system
of posterior or dorsal bronchi for which we employ the term dor-
sobronchi (dorsilateribronchi of Schulze).
The tertiary bronchi, or terminal branches of subdivisions of
the secondary, are commonly known as parabronchi or air-pipes.
Huxley, in 1882, introduced the terms mesobronchium for the
central lung tube, and ento- and ectobronchia for the diaphrag-
matics and the costals; we have adopted these simplified terms.
While on anatomical grounds some critical objections may be
made to certain groups, we have, nevertheless, purely for descrip-
tive purposes adopted the following terminology:
For the primary lung tube, mesobronchus with its three divi-
sions, anterior, middle and posterior.
Fig. 27 Air injected lung of an embryo of 5 days 9 hours incubation showing
the bud of the first entobronchus and the beginnings of the second.
Fig. 28 Lateral aspect of the air injected lung of an embryo of 5 days 20 hours
incubation.
Fig. 29 The same specimen rotated so as to be viewed from the dorso-lateral
aspect. This figure and the preceding show the establishment of the four entc-
bronchi in front of the dilatation of the mesobronchus.
Fig. 30 Dorsal aspect of both lungs of a chick embryo of 6 days 6 hours incu-
bation. Shows the appearance of the bronchial tree when injected with air as
explained in the text.
Fig. 31 Lateral view of the air injected lung of an embryo of 6 days 6 hours
incubation. This specimen was slightly younger than the one sketched in figure
30. Shows the four entobronchi and the buds for the first four ectobronchi.
480 WILLIAM A. LOCY AND OLOF LARSELL
The secondary bronchi, or the original branches, from the cen-
tral lung tube are designated under four divisions: entobronchi
(ventribronchi of Schulze); ectobronchi (dorsibronchi of Shulze) ;
laterobronchi and dorsobronchi.
For tertiary bronchi, the terminal branches of the subdivisions
of the four kinds of secondary bronchi, we use the term para-
bronchi. These are tubes of uniform calibreand unite the dif-
ferent systems of bronchi into bronchial circuits.
In addition to the above, coming from the air-sacs, are the
recurrent bronchi of Schulze and Juillet and the air capillaries
that are radially arranged around the parabronchi.
After this digression on the terminology we continue the de-
scription of embryonic stages. When the four entobronchi are
well started, another series of hernia-like buds arise early on the
seventh day. These are the primordia of the ectobronechi and
they arise from the wall of the embryonic vestibulum. While in
the chick there are usually six, in birds in general they vary
from six to ten. It should also be noted in passing, that the four
entobronchi enumerated in the chick, may in other birds be as
many as six (pelican, Schulze).
The first eetobronchus springs from the widest part of the em-
bryonic vestibulum. By six days six hours it is already of con-
siderable length (figs. 830 and 31) and projects forward and
slightly laterally. The second, third and fourth ectobronchi are
at this stage only papillate buds that protrude dorso-mesially
(dorsally in the adult) from the vestibular wall and caudad to
the first.
Figure 31, from a different specimen of the same age as figure
30, shows the first ectobronchus in a more favorable position.
A fifth and sixth ectobronchus are subsequently developed
from the wall of the mesobronchus, but further consideration of
both ento and ectobronchi will be deferred to a later page.
Reference to figure 82, which is a sketch of the condition of
the bronchial tree in the last half of the seventh day (six days,
twenty hours), discloses another series of incipient secondary
branches from the central lung tube. These are formed on the
lateral wall of the embryonic vestibulum and were called by
THE EMBRYOLOGY OF THE BIRD’S LUNG 48]
Campana (’75) ‘bronches secondaires externes.’ Although they
do not attain in the adult lung the dimensions of ento- and ecto-
bronchi they are, nevertheless, coordinate with them (1) in that
they arise directly from the central lung tube; (2) that they are
the principal branches supplying the lateral part of the ung;
(3) that they make their appearance about the same time as ec-
tobronchi. Accordingly the term laterobronchi (lateribronchi,
Schulze) seems appropriate.
The first three laterobronchi are illustrated in figure 32, lat.
1, 2, 3. Subsequently three smaller ones arise caudad to the
anterior three making a total of six laterobronchi.
Other relatively small bronchi, somewhat more dorsal in posi-
tion, arise similarly, but at a later stage. These correspond to
Campana’s fourth kind of secondary bronchi and we have called
them dorsobronchi (dorsilateribronechi of Schulze). Deserip-
tions of the numerous dorsobronchi and of the six laterobronchi
will be taken up later.
Attention may be briefly directed to the configuration of the
central lung tube. It forms a figure resembling an elongated
S, the posterior bend of which is more marked. The terminal
end is inflated and projects into a protuberance of mesenchymal
tissue, and form the primordium of the abdominal air-sac.
The central expanded part is the embryonic vestibulum. As in-
dicated above, the embryonic vestibulum is not to be confused
with that of the adult.
The bronchial tree from this stage on, continues to grow and
its branches to ramify so profusely, that for the sake of clearness
it is necessary to describe its subsequent development under
separate headings.
The entobroncht. As noted above, the first entobronchus makes
its appearance early on the sixth day of development as a papil-
late bud from the dorsal wall of the anterior division of the meso-
bronchus (mesial in the adult lung). In order to maintain a
correct orientation the following should be noted. Beginning
late on the ninth day the lungs begin to rotate about their longi-
tudinal axis and pass through an angle of about 30 degrees be-
fore reaching their adult position. This rotation carries the
482 WILLIAM A. LOCY AND OLOF LARSELL
Lae 34) —— A.Int.Se.
Abd.Sc. — — \e
THE EMBRYOLOGY OF THE BIRD’S LUNG 483
previously ventro-lateral border laterad, and the previously dor-
sal border towards the median plane.
During the last half of the sixth day (fig. 28, 5 days, 20 hours),
the slender stalk of the first entobronchus extends dorsally for
a short distance, then curves mesially and expands, the trend is
again dorsad, and finally, laterad. The enlarged bladder-like
extremity lies directly above the bronchus (fig. 29) and some-
what anterior to the point where the stalk joins the bronchus.
Soon the distal enlargement of the entobronchus begins to
divide as shown in figures 30 and 32. One bud-like outgrowth
(Cerv.sc.) extends cephalad and constitutes the primordium of
the cervical air-sac. The other bud, visible from this position,
extends laterad and ventrad.
It is not, however, until a later period that the divisions of the
first entobronchus can be clearly seen. On the ninth day of
development (figs. 36 and 37), the first entobronchus exhibits
three principal divisions, or branches, each of which is subdivided
by lobular branches. The cranial branch (Cr.br.) extends for-
ward and bears at its tip the cervical sae (Cerv.sc.) which at this
period extends beyond the lung wall. The cervical sac is not
terminal in the adult as in the embryo, on the contrary, its orifice
opens upon the surface of the main part of the cranial branch.
The transverse, or lateral branch (figs. 36 and 37, Lat.br.) extends
Fig. 32. Dorsal view of the lung of an embryo incubated 6 days 20 hours. L-
lustrates ento- and ectobronchi and the early condition of laterobronchi.
Fig. 33 Mesial view of the same specimen.
Fig. 34 Ventral view. Abd.sc., abdominal air-sac; A.intr.sc., anterior in-
termediate air-sac; Bd., bud of recurrent bronchi; Cerv.sc., cervical air-sac; Dor.,
dorsobronchi; Dr., dorsal ramus of second entobronchus; Ect. 1, 2, 3, ete., the
corresponding ectobronchi; Hnt., 1, 2, 3, 4, the corresponding entobronchi; Lat.,
1, 2, etc., the corresponding laterobronchi; Lat.moz., lateral moiety of the inter-
clavicular air-sac; Mes.moi., mesial moiety of the interclavicular air-sac; P.intr.sc.
posterior air-sac.
Note: Figures 34, 35, 36 and 37 represent the air injected right lung of an em-
bryo of the early ninth day of incubation viewed from different aspects. At this
period the principal divisions of the bronchial tree are established, although still
in a relatively simple stage of development, and the air-sacs with the exception
of the posterior intermediate are well outlined. The reference letters are the
same for all figures.
Fig. 35 Dorsal view. (Reference letters as above.)
484 WILLIAM A. LOCY AND OLOF LARSELL
towards the lateral border and a slender outgrowth at its tip
(Lat.moi.) foreshadows the lateral moiety of the interclavicular
air-sac. The medial branch extends transversely in the oppo-
site direction (fig. 37 Mes.br.).
The further development of the first entobronchus consists in
-the subdivision of its lobes into smaller lobules. These, in turn,
elongate and give off branches that develop into the parabronchi
of the cephalic region of the lung. The first entobronchus is a
eh ae SOE DY.
ee a LN Pas ot Ba acy cas
“7 Gs y
Ss
SOS ft —Ent. 3----\- Pp
ROSS Ent, 4 ----\.----— ge
ee --Abd,Se.---~- Pe
36 37
Fig. 36 Lateral view. (Reference letters as in figure 34)
Fig. 37. Mesial view. (Reference letters as in figure 34)
very large and important part of the bronchial passages. Its
subdivisions supply the summit of the lungs (fig. 44) and also
give rise to two air-sacs, the cervical and the lateral moiety of
the interclavicular air-sac. |
Entobronchus number two (figs. 28 and 30, Hnt. 2) has a simi-
lar origin to that of number one. It starts about the middle of
the sixth day of incubation as a bud from the dorsal wall (mesial
in the adult) of the intra-pulmonary bronchus. It grows rapidly,
so that, at five days, twenty hours (figs. 28 and 29) it has an
THE EMBRYOLOGY OF THE BIRD’S LUNG 485
expanded distal end attached by a slender stalk to its point of
origin. The stalk first grows mesiad, then enlarges and bends
dorsad, and finally, expands into the distal sae mentioned above.
This expansion, like that of the first entobronchus, lies directly
above the bronchus (fig. 29), but slightly towards the mesial
border of the lung, It gives off from its lateral border a slender
elongated branch which at six days, twenty hours, crosses the
bronchus dorsally (fig. 32) and at its tip bends slightly ventro-
lateralward. This is the dorsal ramus of the adult second ento-
bronchus. It is first noticeable in the latter part of the sixth
day.
Aside from increase in dimensions, but little change oceurs in
the second entobronchus or its dorsal ramus until the ninth day
of development. At this time, as illustrated in figures 35 and 37,
the extremity of the entobronchus has divided into two unequal
lobe-like branches, the more posterior of which is bifurcated (fig.
37). The dorsal ramus has elongated and taken on a more ven-
tral curve (fig. 37) and is a distinct anatomical feature.
The second entobronchus is smaller and of less wide-spread
distribution than the first. The three lobes, formed on the eighth
day, subdivided into branches, which on the ninth day, begin to
differentiate into parabronchi that ultimately supply the an-
terior mesial region of the adult lung. The parabronchi of the
dorsal ramus supply the interior of the cranial part of the lung
between the stem of the first entobronchus and its transverse
branch.
Exceptionally the second entobronchus gives off near its base
a mesial branch (fig. 38) from which develops the mesial moiety
of the interclavicular sac. The mesial moiety, however, as de-
scribed below, usually has its origin from the third entobronchus.
This exceptional point of origin may account for the fact that
some observers have ascribed the origin of the interclavicular
sac to the second entobronchus while others have claimed that
it arises from the third entobronchus.
The third entobronchus arises (fig. 28) somewhat more mesially
on the intra-pulmonary bronchus than the first two. On the
second half of the sixth day it extends dorsally a short distance,
486 WILLIAM A. LOCY AND OLOF LARSELL
then enlarges and bends caudally, descending on the mesial side
of the central lung tube so that the expanded distal end lies at
the side of the main bronchus, not above it, as is the case with
entobronchi numbers 1, 2.
Figures 30, 32 and 33 represent the beginning of a bud project-
ing ventrally from the third entobronchus which is to play an
important part in the later history of the lung. This hollow bud
—--Abd.Sc.
Fig. 88 Dorso-mesial view of right lung of an embryo slightly younger than
the one sketched in figures 34 to 37. Illustrates the exceptional position of ori-
gin of the mesial moiety of the interclavicular air-sac, from the second instead of
the third entobronchus.
gives rise to two branches, one cephalad and the other ventrad.
The more cephalad branch forms the mesial moiety of the inter-
clavicular air-sac (figs. 33 34); the ventrad, bladder-like portion,
differentiates into the anterior intermediate air-sac. By the
close of the eighth day of incubation both these air-sacs project
beyond the boundary of the lung.
It should be noted that under the exception indicated above,
namely, when the second entobronchus gives rise to the mesial
THE EMBRYOLOGY OF THE BIRD’S LUNG 487
moiety of the interclavicular sac, the branch of the third ento-
bronchus just described does not divide and produces only the
anterior intermediate air-sac.
The other division of the third entobronchus elongates and
expands caudo-dorsally until, early on the ninth day, its dorsal
face becomes tri-lobed (fig. 37). At the ten-day stage these
lobes branch and later give rise to the parabronchi of the middle
portion of the mesial lung facet.
In the adult, the third entobronchus extends obliquely toward
the caudal extremity of the lung on its ventral face (fig. 44).
From its medial border six or seven parabronchi are given off
and on its lateral border is a transverse branch connecting it
with the fourth entobronchus.
The fourth entobronchus has an embryonic history very simi-
lar to that of the third, but it is not connected with any of the
air-sacs. Its origin is on that part of the mesobronchus where
the swelling of the embryonic vestibulum begins, or where the
anterior part of the mesobronchus merges with the middle part.
Its orifice is somewhat more mesial in position than that of the
third entobronchus.
Shortly after the beginning of the seventh day the fourth en-
tobronchus is short and extends mesially with an expanded end
which is directed caudally (fig. 30).
On the eighth day a branch is budded off very close to its base,
which early on the ninth day forms a sickle (figs. 34 and 37)
with its point directed laterally. In the adult this sickle-shaped
branch becomes the great transverse branch, extending obliquely
across the ventral face of the lung and forming a prominent ana-
tomical landmark (fig. 44). Campana (’75) enumerated this large
transverse branch as a fifth entobranchus, but since it has no
direct connection with the central lung tube, even in the embryo,
we have, as most previous observers, considered it a branch of
the fourth entobronchus.
The principal trunk of the fourth entobronchus extends cau-
dally and mesially in a course parallel to the caudal branch of
the third entobronchus. The territory included between the
main trunk and the transverse branch is overrun with small
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, NO. 3
488 WILLIAM A. LOCY AND OLOF LARSELL
branches springing from both the main division and the trans-
verse branch.
It is evident that the entobronchi begin very early to play a
leading part in the formation of the bronchial tree and although
limited in number they make a large showing on the metal casts
of the adult lung. Generally speaking, the entobronchi of the
adult are distributed on the ventral surface of the lung as shown
in figure 44. The parabronchi that spring from their branches
bend around the mesial border of the lung and curve on to the
dorsal surface where they meet parabronchi coming from the
ectobronchi and thus a connection is established between these
opposed bronchi and the characteristic bronchial circuits.
The ectobronchi. The ectobronchi are outgrowths from the
central lung tube arising somewhat later than the entobronchi
and more mesial in position upon the wall of the lung tube.
These are the bronches costales of Sappey and the secondaires
internes of Campana. While the ento- come from the anterior
division of the mesobronchus, the ectobronchi spring from the
wall of the embryonic vestibulum in the middle division of the
central lung tube.
The first (fig. 30, Hct. 1) arises as a forward and upward pro-
jecting bud from the anterior part of the expanded region of the
tube at the beginning of the seventh day. By rapid growth it
becomes elongated, and, at six days, twenty hours, may be seen
the first indication of branching in the form of a ventral bud
springing from the distal end of the ectobronchus (fig. 32). The
two lobes produced in this manner rapidly increase in size and
give rise to other lobe-like branches. As illustrated in figure 36
there are five such lobules at the beginning of the ninth day.
The two most dorsal are the result of the division of the original
dorsal lobe, and the three ventrally placed lobules come from
the other division of the first bifurcation.
These lobular outgrowths foreshadow the parabronchi of the
anterior lateral and of the dorsal region of the lung. By the
close of the ninth day the lobules have greatly increased in num-
ber and, on the tenth day, show the beginnings of some parabron-
THE EMBRYOLOGY OF THE BIRD’S LUNG 489
chi (fig. 39, Par.). The original bifurcation can be readily seen
in the later stages (fig. 40).
The second ectobronchus (figs. 31 and 33, Hct. 2) arises early
on the seventh day of development as a bud from the dorso-mesial
surface of the vestibulum (embryonic). Its point of attach-
ment is slightly more mesial than that of the first ectobronchus
and at first its course is nearly parallel with the first ectobronchus,
but in late stages it becomes transverse. A bud arises on its dis-
tal extremity towards the close of the seventh day but there is
no definite branching until early on the ninth day at which time
the distal end is divided into three lobes (figs. 35 and 36). These
lobules are the first indications of the subdivisions whose branches
become parabronchi as shown in figure 39. In the adult the
parabronchi of the second ectobronchus supply the dorso-medial
face between the third and fourth ribs.
The embryonic history of the remaining four ectobronchi is
so similar that only brief descriptions are required. It is suf-
ficient to note that each has its origin slightly more mesially than
the preceding on the dilatation of the lung tube and, after
branching has begun, they incline more towards the caudad
border of the lung. Beginning with ectobronchus number three
this caudal bending becomes more marked. In the adult their
parabronchi supply the mesial region of the posterior half of the
lung. The parabronchi that originate from the sixth ecto-
bronchus spread out in such a manner as to be distributed to
the dorso-caudal part of the lung.
None of the ectobronchi have any direct connection with air
sacs.
A seventh ectobronchus is sometimes present in the chick and
becomes connected with recurrent bronchi from the abdominal
alr-sac.
As to their number in birds in general, Schulze has pointed out
that they vary from six to ten and that seven is the usual num-
ber.
The laterobronchi. The laterobronchi correspond to those
designated ‘bronches secondaires externes’ by Campana. Six
490 WILLIAM A. LOCY AND OLOF LARSELL
laterobronchi are developed, but not simultaneous y, and they
arise as buds from the lateral wall of the embryonic vestibulum.
The first two (fig. 30, Lat. 2 and 3) arise at the stage of six days
six hours. These are, in reality, numbers two and three—num-
Saat EN as foe
Mes.moi.----fas--®e
Lat.mol. --- f)
Rec. Br. <\
Fig. 39 Latero-posterior view of the right lung of late ninth day of incuba-
tion. Shows the beginning of recurrent bronchi on the abdominal and the pos-
terior air-sacs.
Fig. 40 Lateral view of the left lung of the same specimen. Illustrates the
five air-sacs and primordia of recurrent bronchi of the four air-sacs possessing
them in the adult. Rec.Br., recurrent bronchi, other reference letters as under
figure 35.
ber one arising later in front of number two. They elongate in
the latero-ventral plane, so that near the close of the seventh
day they are distinctly prominent (fig. 32, 6 days, 20 hours).
At the beginning of the ninth day (fig. 36) both two and three
THE EMBRYOLOGY OF THE BIRD’S LUNG 49]
make a sharp bend caudally and their distal extremities extend
to the lateral border of the lung. Just before reaching this
wall, the second latero-bronchous bifurcates, and later forms a
number of divisions which result in supplying the middle part of
ventro-lateral tung region with parabronchi (figs. 39 and 40).
The third laterobronchus does not divide like the second, but
projects beyond the lung wall and forms the primordium of the
posterior intermediate air-sac (figs. 39 and 40). This sae makes
its appearance at the beginning of the ninth day.
Campana says that this a‘r-sac arises on the second latero-
bronchus, but he did not study development and a correct un-
derstanding is possible only by following embryonic stages.
A number of branches are later given off, from the third latero-
bronchus between the air-sac and the point of connection of the
laterobronchus with the mesobronchus. ‘The first branches of
this kind appear on the tenth day and the others later. These
branches which extend ventro-laterally, must not be confused
with the dorsally and forward projecting buds fromthe air-sac
itself (fig. 40). These latter are the recurrent bronchi and will
receive separate consideration on a following page.
We may now consider the first laterobronchus—so-called be-
cause it is the most anteriorly situated, although arising later.
It is first distinguishab!e on the second half of the seventh day
as a boss-like bud from the antero-lateral wall of the embryonic
vestibulum (fig. 32, Lat. 1). It grows forward a short distance
and then turns ventral (fig. 34). Later a number of branches
are given off its anterior and lateral surfaces which produce short
parabronchi that eventually anastomose with the deep-lying
branches of the transverse branch of the first entobronchus and
with parabronchi of the dorsal ramus of the second entobronchus.
The side branches of this laterobronchus anastomose also to
some extent with the anterior tips of the recurrent bronchi of
the abdominal air-sac.
Laterobronchi numbers four and five take their origin from the
posterior division of the mesobronchus, just caudad to the ex-
panded part or embryonic vestibulum. They appear during the
eighth day of incubation, and at the beginning of the ninth day
492 WILLIAM A. LOCY AND OLOF LARSELL
(fig. 36, Lat. 4 and 5), they follow a course parallel to that of
laterobronchus number three. In their subsequent history they
form branches by the customary bifurcation and give rise to para-
bronchi in the posterior lateral region of the lung.
A small inconspicuous bronchus arises more caudally on the
mesobronchus, and although it resembles quite closely some of
the larger dorsobronchi, we have considered it as the sixth latero-
bronchus (fig. 36).
The dorsobroncht. These embrace the bronchi of Campana’s
fourth division of secondary bronchi and by him designated
‘‘bronches posterieurs ou dorsales”’ (dorsilateribronchi of Schulze,
10). They are smaller and more numerous than the other bron-
chi arising from the central lung tube. ‘Their orifices are some-
what variable as to position, and some members of the group
arise on the stems of the latero- and ectobronchi, and this im-
parts to them the character of being transitional between the
bronchi of the second and of the third order. On strictly ana-
tomical grounds those that spring from the latero and ectobron-
chi are of the third order, but there are always many, especially
of the smaller ones (and frequently large ones), that connect
directly with the mesobronchus. Campana thinks that it in-
troduces a useless morphological complication to exclude them
from the group of the secondaries. On account of their transi-
tional character and their (sometimes) connections with latero
and ectobronchi, Juillet sets them at one side. Our observations,
however, lead us to agree with Campana, that, after recognizing
that some of them merge into bronchi of the third order, and
that they are so small as to resemble parabronchi, nevertheless,
there are always a considerable number arising from the walls
of the mesobronchus, and therefore, they should be retained in
the group of the secondaries.
So far as we are aware no previous embryological observations
have been recorded on these bronchi.
There are two groups of the dorsobronchi, four or five larger
ones and approximately twenty smaller ones. Figure 36 exhibits
three small spur-like projections from the lateral side of the meso-
bronchus just dorsal to the bases of the first, second, and third
THE EMBRYOLOGY OF THE BIRD’S LUNG 493
laterobronchi. These are the beginnings of the larger or prin-
cipal dorsobronchi and they arise on the eighth day of incubation.
The stage figured shows their degree of development early on the
ninth day.
After this stage additional buds appear, in rather rapid suc-
cession, within the space on the wall of the mesobronchus between
the bases of the latero- and ectobronchi and, also, more caudad.
By the eleventh day so many have developed that, when they
are injected with air, they tend to obscure the other air passages.
of the lungs. Accordingly, all but the larger dorsobronchi have
been omitted in the sketches subsequent to the ninth day, in
order not to confuse the more important features of the sketches,
In the adult lung they form a very important area on the
dorsal surface of the lung which will be described later. As il-
lustrated in figure 55, the bases of the dorsobronchi are arranged
roughly in three rows. The middle row is composed of the
larger ones, which are the first to develop, and arranged, in a
general way, alternating with those of the middle row, mesially
and lateral to it, are the more numerous smaller ones. In the
diagram (fig. 55) the bases of twenty-one dorsobronchi are shown
—five of the larger, opposite the bases of the laterobronchi, and
sixteen of the smaller ones. This diagram is made from the study
of a Wood’s metal cast in which the dorsobronchi were clipped off
at their bases. There is individual variation as to the total num-
ber present as well as to the pattern of their arrangement on the
mesobronchial wall. Campana cites a number of variations in
this arrangement. We have examined specimens with twenty-
five dorsobronchi—five larger and twenty smaller ones—but to
what mit the number of smaller ones may go in the adult we
have not determined.
These dorsobronchi occupy a position along the dorsal wall
of the central lung tube from the stem of the first eetobronchus to
the posterior end of the mesobronchus,
The branches of the dorsobronchi project towards the dorsal
surface of the lung, and arriving there, their parabronchi form a
well-marked, nearly circular area, or network, in the center of
the dorsal face. After the tenth day this area can readily be
494 WILLIAM A. LOCY AND OLOF LARSELL
seen from surface views of untreated embryo lungs. Figure 41
represents the condition at the age of ten and one-half days.
The small circles in the middle of the dorsal area represent end
views of the dorsobronchi and their subdivisions. At the pos-
terior and lateral parts of the circular area are several dorso-
bronchi curved in such a way as to show other side walls. Ob-
Fig. 41 View of the dorsal surface of the right lung of an embryo of the ten
and one-half day stage, showing the area at which the dorsobronchi reach the sur-
face. Also recurrent bronchi from the abdominal and posterior air-sacs.
viously these do not belong to the ectobronchial system, branches
of which surround the central patch of dorsobronchi.
In adult stages, as Campana has pointed out, the parabronchi
of the dorsobronchi form a network wnich he calls ‘‘réseau bron-
chique superficial de la face externe,’’ covering precisely the
mass of parabronchi of the smaller dorsobronechi. The dorso-
bronchi have longer parabronchi that reach the surface and
shorter ones distributed to the interior of the lung.
In Wood’s metal casts of the adult (fig. 45) the réseau anasto-
motique shows as an easily identified anatomical landmark.
THE EMBRYOLOGY OF THE BIRD’S LUNG 495
This dorsal plexus, or network, is an important one and in one
respect it is unique. Instead of uniting two different groups of
parabronchi (the usual situation) this network is composed only
of dorsobronchi. It is central in position and at its periphery it
connects with the parabronchi of the different systems with which
it is surrounded. In this manner there is established by the in-
termediation of the réseau anastomotique a general communica-
tion between the bronchial circuits. In addition to this it forms
connections with recurrent bronchi from the abdominal air-sac.
The parabronchi. The parabronchi, frequently designated lung
pipes (Bronchuli respiratoril of Schulze) belong to the third order
of bronchial tubes. In order to make this classification clear we
must bear in mind that the central lung tube is primary; and that
all air-tubes opening directly on the primary are secondary.
Branches of small and nearly uniform calibre, arising from subdi-
visions of the secondaries, are tertiary. For a specific illustra-
tion take the ento- and ectobronchi: The subdivisions of the
ento- and ectobronchi form fan-like and feather-like groups,
the branches of which diminish in diameter until they reach a
certain size, which thereafter they maintain, and continue as
cylindrical pipes; these tubes of uniform diameter are the ter-
tiaries, or parabronchi. By their anastomosis they form the net-
work of passages so characteristic of the avian lung. Since
parabronchi are incomparably more numerous than the second-
aries, the great mass of the lungs is composed of tertiary bronchi
and the air-capillaries that spring from them.
With the exception of the plexus of dorsobronchi, mentioned
above, the parabronchi unite two opposite systems of secondary
bronchi. Owing to this union, as Campana pointed out, there
is no bronchial tree, but instead bronchial circuits, the middle
part of which is multiplied into small tubes (tertiaries) while the
extremes (secondaries) open on the primary bronchus. The pri-
mary bronchus is tracheal, its branches only are pulmonary.
The network of parabronchi is formed relatively late and,
chronolog cally, as well as with respect to their anatomical rela-
tions, the account of the parabronchi might well follow that of
the air-sacs and their recurrent bronchi but unity of treatment
makes it advantageous to consider them at this point.
496 WILLIAM A. LOCY AND OLOF LARSELL
Their embryonic development may be briefly summarized.
The formation of lobular branches of ento- and ectobronchi, at
the close of the ninth, and during the tenth day, has already been
described. In spreading out on the ventral (entobronchi) and
dorsal (ectobronchi) facets of the lung, these branches elongate
towards the mesial border and undergo further subdivision.
Numerous outgrowths of ento- and ectobronchi also occur within
the interior of the lung.
The digitations of the ento- and ectobronchi gradually approach
each other, and, in order to get from the ventral to the dorsal
face, the branches of the entobronchi bend a’most at right angles
around the mesial face of the lung. On that facet they run
nearly parallel to one another. These are parabronchi and are
diagrammatically represented in figure 56.
The branches of the ectobronchi, on the dorsal face, grow to-
wards the approaching entobronchi, and by the eleventh day,
there is only a narrow lane-like area of mesenchymal tissue be-
tween the two groups that is unoccupied by bronchial tubes.
This area extends along the dorso-mesial face from the caudal
to near the cranial border of the lung, where it is turned to one
side owing to the fact that the branches of the first entobronchus
are in the path. These branches extend posteriorly and gradu-
ally approach the similar branches of the first ectobronchus and
of the first laterobronchus, and, shortly these opposite systems
of tubes become connected by parabronchi. Parabronchi, un-
like other bronchial tubes, are substantially of uniform calibre.
Those of the periphery are of somewhat larger diameter than
those within the interior of the lung.
The manner of the anastomosis of parabronchiis interesting.
On the twelfth day of development the tips of the aproaching
parabronchi are nearly in contact. They now bifurcate, forming
slender twigs which come into contact and anastomose with simi-
lar twigs coming from the opposite direction. Complete union
has occurred by the fifteenth day of incubation, but the anasto-
mosing parts are as yet very slender (fig. 42). These twigs at-
tain substantially the diameter of their parent branches by the
eighteenth day (fig. 43) and remain as the parabronchial network
THE EMBRYOLOGY OF THE BIRD’S LUNG 497
of the adult lung, along the dorsal margin of which they form a
distinct line. This line also marks the region of anastomosis,
on the dorso-lateral border between the first ento- and the first
ectobronchus.
During the eighteenth day of development, and subsequently,
similar connections are established between parabronchi, both ex-
ternally and internally, in other parts of the lung. There is ¢
well defined area along the ventro-mesial margin of the adult
lung, but this is not the result of terminal anastomoses, as in the
case just described, but the union effected between adjacent
parabronchi by the sending out of lateral buds.
Figures 42 and 43 are surface studies of air injected specimens
to show the nature of the peripheral anastomoses. Figure 42
exhibits the condition on the fifteenth day of development, with
the slender parabronchial tips coming into contact. A camera
tracing of the eighteenth day stage (fig. 43) shows the parabron-
chi of the same region having attained their full size. As before
indicated they are of substantially the same diameter throughout
their length and they are united by frequent transverse branches.
The diameters of parabronchi vary in different birds; they are
relatively large in the domestic fowl where they attain a diameter
of about 2mm. on the surface and about 1 mm. in intrapulmonary
situations. These figures show the formation and union of para-
bronchi at the surface, and it is to be understood that similar anas-
tomoses occur at different levels within the interior of the lung.
The branches of latero- and dorsobronchi give rise to parabron-
chi which are relatively short, and which anastomose, at differ-
ent levels within the lung, with those of adjacent latero- and
dorsobronchi, and to some extent with recurrent bronchi. By
these anastomoses an elaborate network of air-tubes is formed in
the dorso-lateral and lateral portions of the lung and, also, in the
caudal region of the ventral face. In the latter region, the sec-
ond laterobronchus and the first ramus of the fourth entobron-
chus play the most important part.
The recurrent bronchi from the two posterior air-sacs, also, as
more fully described below, help materially in forming the net-
work of this region.
498 WILLIAM A. LOCY AND OLOF LARSELL
Thus are established the bronchial circuits which serve to make
a network of communicating air-passages throughout the lungs.
Bronchial circuits of the adult lung. After the fifteenth day the
anastomosis of parabronchi goes on rapidly and results in a rich
profusion of intercommunicating air passages, imparting tothe
lung tissue a structure that may be designated sponge-like.
Cran.
Fig. 42 Portion of the mesial facet of the lung of an embryo of the fifteenth
day showing early stages of the anastomosis of parabronchi. X 73, reduced 3.
Fig. 43 Similar view of the lung surface of an embryo of the eighteenth day.
xX 48, reduced 3.
The passage ways are intricate and numerous and only the
chief sources of the parabronchial network of different lung
regions will be indicated. In observing the main features ofthe
distribution of branches of the bronchial circuits we have de-
pended chiefly on Wood’s metal casts. Campana’s detailed ac-
count of bronchial connections is impressive, but owing to the
THE EMBRYOLOGY OF THE BIRD’S LUNG 499
great space required for the description of each branch, we have
undertaken to give merely a condensed account based on our
studies of metal casts.
As shown in photographs of Wood’s metal casts (figs. 44 and
45), the large divisions of the entobronchi are sread on the ven-
tral and mesial facets of the lung, and have already been gener-
ally described.
The subdivisions of the eectobronchi occupy most of the dorsal
surface except the central area of dorsobronchi.
The entobronchi with their various ramifications form bron-
chial circuits of approximately the anterior one-third of the lung,
and of all that part which lies dorso-mesially to the central lung
tube (the cranial pente of Juillet) (figs. 44 and 45). In addition
to this, the transverse branch of the fourth entobronchus sup-
plies parabronchi for the middle region of the latero-ventral lung
facet.
It is evident that the four entobronchi are of wide distribution,
as the air passages arising from them occupy substantially one-
half of the lung, and they also give rise to three of the air-sacs.
The superficial branches of the six ectobronchi, as already in-
dicated, anastomose with similar branches from the entobronchi
along the dorso-mesial border of the adult lung. There are also
internal branches that form parabronchi which unite at different
levels with those coming from entobronchi. A number of minor
ectobronchial branches extend laterally and anastomose, around
the reseau, with branches from the dorsobronchi.
So far as we could determine the branches of the ectobronchi
do not exhibit a direct connection with any of the recurrent bron-
chi. There is, however, an indirect connection through the dorso-
bronchi, which anastomose on the one hand with ectobronchial
branches, and on the other hand, with twigs of the recurrent
bronchi from the posterior and abdominal air-sacs.
Besides the connections indicated, the cephalad parabronchi of
the first ectobronchus also unite, through intermediation of
twigs from the dorsal branch of the second entobronchus. with
recurrent bronchi of the interclavicular air-sac.
s*-=-=Rec.Br.
45
Fig. 44 Photographs of a Wood’s metal cast of the right lung of the adult
fowl. (A) Dorsal view showing the “‘réseau anastomotique’’ of the dorso bron-
chi. (B) View of the ventral face of the same, illustrating the superficial distri-
bution of entobronchi. The relations of the fourth entobronchus and its trans-
verse branch are well exhibited.
Fig. 45 Photograph of a Wood’s metal cast of the lung of a small adult
domestic fowl (A) seen from the latero-dorsal aspect, (B) mesial aspect. The
anterior and posterior air-sacs are expanded but the east of the abdominal is
shriveled. Recurrent bronchi are seen connected with the three air-sacs. (B)
shows well the large cranial branch of the summit of the lung.
500
THE EMBRYOLOGY OF THE BIRD’S LUNG 501
It is now necessary to point out the distribution in the adult
lung of the branches of the six laterobronchi. The parabronchi
of the first laterobronchus occupy the ventral part of the lateral
face of the lung and anastomose with those of the transverse
branch of the fourth entobronchus. They also connect, both di-
rectly and indirectly, with recurrent bronchi from the anterior
intermediate air-sac.
The second laterobronchus is the largest in the embryoand
also in the adult. It gives off branches between its origin and its
terminal bifurcations which send parabronchi to the caudo-ven-
tral part of the lung. The parabronchi of the second laterobron-
chus anastomose especially with those of the adjacent latero-
bronchi on each side of it, and with recurrent bronchi from the
two posterior air-sacs.
The third laterobronchus, as described in its embryonic his-
tory, gives origin to the posterior intermediate air-sac. Like the
other laterobronchi, its parabronchi are distributed in the ventro-
lateral part of the lung, and they establish connections with the
network of air-pipes in this general region.
The remaining laterobronchi are intermediate in size and im-
portance and may be included under one description. Their
branches serve the caudo-lateral part of the lung and form anas-
tomoses with the internal branches of the recurrent bronchi from
the two posterior air-sacs, as well as with other air-pipes of this
part of the lung.
In addition to the anastomoses specifically indicated above,
branches from all the laterobronchi form connections with
branches of the dorsobronchi, which may now claim our atten-
tion.
Speaking, general, the branches of the dorsobronchi extend to-
wards the dorsal surface of the lung, and since this is relatively
near by, the dorsobronchi are short as compared with the latero-
bronchi. Before reaching the lung surface, they bifurcate and
the branches thus produced again divide one or more times.
The parabronchi from these sources are internal and unite with
those of neighboring laterobronchi, and, in the case of the more
posterior ones with recurrent bronchi.
502 WILLIAM A. LOCY AND OLOF LARSELL
The most significant feature of dorsobronchi is the circular
plexus of the dorsal region (réseau anastomotique of Campana)
already described.
As Campana pointed out, in 1875, some bronchi of identical
appearance are inserted on the stems of ectobronchi, instead of
on the mesobronchus, and this circumstances indicates a transi-
tional form between secondary and tertiary bronchi. We have
designated as dorsobronchi only those (about twenty-five) that
have an orifice communicating directly with the mesobronchus.
It is to be noted that there are no bronchial stems on the oppo-
site (ventral) wall of the mesobronchus.
A eareful study of the intereommunications of the air passages
is very convincing that there is a universal anastomosis of para-
bronchi, and if there be any blind endings or culs-de-sac in the
lungs, they are very exceptional. [In our observations we have
never been able to find an undoubted blind ending of air pas-
sages. This labyrinthine communication extends also to the air-
capillaries that radiate around the parabronchi.
Besides the terminal anastomoses of parabronchi there are fre-
quent lateral communications by short canals. This kind of
communication is especially well seen on the parabronchi of
ento- and ectobronchi (figs. 42 and 43).
The air capillaries. The ultimate divisions of the bronchial
circuits are the air-capillaries that are radially arranged about
the parabronchi. It is to be kept in mind tht even these micro-
scopic branches do not end in culs-de-sac.
The lung parenchyma, beginning on the ninth day, becomes
arranged around parabronchi into prismatic columns which on
cross-section are hexagonal. In the middle of these hexagonal
areas lie the circular parabronchi, and in the adult lung a large
number of minute branches project radially into the lung paren-
chyma. The intereommunicating distal ends of these branches
constitute the air-capillaries.
All these parts arise in succession in the embryo between the
fourteenth and the sixteenth day of development. The vesti-
bules appear about the fourteenth day as hollow buds projecting
from the walls of the parabronchi. They develop into short cyl-
THE EMBRYOLOGY OF THE BIRD’S LUNG 503
indrical conduits measuring from 0.10 to 0.14 mm. in diameter
and there are about twenty of them within the circumference of
a cross-section of a parabronchus. Soon the vestibules divide
at their tips by unequal dichotomy and the branches thus formed
may subdivide. Asa rule, subdivision does not occur more than
twice, but, in the adult lung, there are numerous examples of a
third and even a fourth division. The first bifurcations occur
Fig. 46 Simplified diagram to illustrate the anastomosis of the air capillaries.
early in the fifteenth day, and by the close of that day branches
are well defined. By continuous growth these branches pene-
trate further into the tissue surrounding the parabronchus and
the air capillaries are soon formed on their distal extremities. At
first these terminate blindly, but between the nineteenth and the
twenty-first days of development they anastomose profusely
and thus establish a network of intercommunicating air pas-
sages. In the chick the anastomosis between air-capillaries is
confined to the limits of the hexagonal prisms. In good flyers,
THE AMERICAN JOURNAL OF ANATOMY, VOL. 19, No, 3
504 WILLIAM A. LOCY AND OLOF LARSELL
however, as Guido Fischer (’05) has shown, the air-capillaries of
adjacent parabronchi cross the boundaries and anastomose.
By the final anastomosis of these ultimate divisions of the bron-
chial circuits there is established the unique feature of the avian
lung—the lack of culs-de-sac in any part of the air circuits.
Figure 46 is a much simplified diagram to show the relation of
the air capillaries to the parabronchus and to the hexagonal
prism of lung parenchyma in the adult chick.
Intermingled with the network of air capillaries is a comple-
mental network of blood capillaries so that the facilities for rapid
aération are very complete. The lung of birds is not large in
extent, but it is highly vascular and the continuous air current
from the air-saes through the recurrent bronchi makes it an ef-
fective apparatus for aération of the blood.
SUBJECT AND AUTHOR INDEX
LBINOrat. Effects of inanition upon the
A structure of the thyroid and parathyroid
Slandsiot the:.cs:s0 de sceeieee rca see:
Albino rat from birth to ten weeks of age.
The growth and distribution of the milk-
ducts and the development of the nipple
IT, CHEW ss ee eee eee ee en tee ated snes
Amblystoma punctatum. The development
of the liver and pancreas in.............
Arteriesin the common gonoids. Therelation
of coronary and hepatic.............
Atrio-ventricular system (bundle of His),
interstitial granules (mitochondria) and
phospholipines in cardiac muscle. On the
occurrence and physiological significance
of fat in the muscle fibers of the normal
MYyOcardium ANG: 4.5.60 .4escese se eset ae
AUMGARTNER, E. A. The develop-
ment of the liver and pancreas in Ambly-
Stoma PUuNctatumM 2.5 «0. ccs acce ke esses 2
Benstey, R. R. The influence of diet and
iodides on the hyperplasia of the thyroid
gland of opossums in captivity..........
Benstey, R. R. The normal mode of secre-
tion in the thyroid gland..............
Bird’s lung. Based on observations of the
domestic fowl. PartI. The embryology
OUP GNG Bere ee a ie vere eee oe
BREMER, JOHN Lewis. The interrelations of
the mesonephros, kidney and placenta in
different classes of animals..............
Bouk, L. Problems of human dentition... .
Butviarp, H. Hays. On the occurrence and
physiological significance of fat in ihe
muscle fibers of the normal myocardium
and atrio-ventricular system (bundle of
His), interstitial granules (mitochondria)
and phospholipines in cardiac muscle...
ELL clusters in the dorsal aorta of mam-
malian embryos. The..................
Clusters in the dorsal aorta of mammalian
embryos. The cell........
Coronary and hepatic arteries in ‘the common
gonoids. The relation of........
Cowpry, E. V. The general functional - sig-
nificance of mitochondria
ANFORTH, C. H. The relation of coro-
nary and hepatic arteries in the common
SON OLG Settee ye eee eee eee
Dentition, Problems of human...............
Diet and iodides on the hyperplasia of the
thyroid gland of opossums in captivity.
The influence of.................
Distribution of the milk-ducts and the dev el-
ment of the nipple in the albino rat from
birth to ten weeks of age. The growth
ANG nae Aa ape a eee et aes arene
Drainage of the endolymphatic sac and its
topographical relation to the transverse
aos in the human embryo. The vascu-
Bee naa en NO een rane as A ee
MBRYO. The vascular drainage of the
endolymphatic sac and its topographi-
cal relation to the transverse sinus in the
human
305
391
179
91
57
353
67
_Embryos.
Embryo, with special reference to the origin
of the erythrocytes. The microscopic
structure of the yolk-sae of the pig...... 2
Embryology of the bird’s lung. Based on
peer ancre of the domestic fowl: Part
The cell clusters in the dorsal
aorta.ot mammaliane. 2.04. sseeeeee see
Emmet, Vicror E. The cell clusters in the
dorsal aorta of mammalian embryos....
Endolymphatie sac and its topographical
relation to the transverse sinus in the
human embryo. The vascular drainage
Ol, UN Gir erent rieen cei ieier vere rt cae teeter ne
Erythrocytes. The microscopic structure of
the yolk-sac of the pig embryo, with spe-
cial reference to the origin of the........
AT in the muscle fibers of the normal
myocardium and atrio-ventricular sys-
tem (bundle of His), interstitial gran-
ules (mitochondria) and phospholipines
in cardiac muscle. On the occurrence
and physiological significance of......
ONOIDS. The relation of coronary and
hepatic arteries in the common........
Growth and distribution of the milk-ducts
and the development of the nipple in the
albino rat from birth to ten weeks of age.
RS 0 ee era ee Se ye Oro
EPATIC arteries in the common gonoids.
The relation of coronary and...... .
Human dentition. Problems of..............
Hyperplasia of the thyroid gland of opossums
in captivity. The influence of diet and
LOGIGEStonMthe Ysa... seco sos ce eemnneee es
NANITION upon the structure of the thy-
roid and parathyroid glands of the albino
Tatey MectsiOt esse. sae usenet
Injection method. The sino-ventricular sys-
tem as demonstrated by the............
Interstitial granules (mitochondria) and phos-
pholipines in cardiac muscle. On the oc-
currence and physiological significance of
fat in the muscle fibers of the normal myo-
ecardium and atrio-ventricular system
(bundletof Is) ccenns- setae areas:
Iodides on the hyperplasia of the thyroid
gland of opossums in captivity. The in-
fluence’ of diet and on..9-.4062.2tseatanes
ACKSON, C. M. Effects of inanitition
upon the structure of the thyroid and
parathyroid glands of the albino rat,.
Jorpan, H. E. The microscopic structure of
the yolk-sac of the pig embryo, with spe-
cial reference to the origin of the erythro-
CY LOS. Sesvaze aclagene teeta antyard avs fencnen c= aloha apahearers fe
y IDNEY and placenta in different classes
of animals. The interrelations of the
MESONEDHTOSw sa. eeees oes see cee seis ;
Kine, M. R. The sino-v entricular system as
demonstrated by the injection method.
67
277
305
977
“dd
179
149
506
ARSELL, Ovor, Locy, WittraAm A, and.
The embryology of the brid’s lung.
Based on observations of the domestic
fOWla, (eacuel eae.
Liver and pancreas in
A mbly stoma punc-
tatum. The development of the.......
Locy, Wiiv1am A. and Larse.., OLor. The
embryology of the bird’s lung. Based on
observations of the domestic fowl. Part
Lung. “'Pased on ‘observations of the Domestic
fowl. Part 1. The Embryology of the
bind’ss.).40- seek eee eee oe eee eases
AMMALIAN embryos. The cell clus-
M ters.in the dorsal aorta of............
Mammary gland, studies on the. :
Mesonephros, kidney and placenta in 1 differ-
ent classes of animals. The interrelations
Milk-ducts and the development of the nipple
in the albino rat from birth to ten weeks
of age. The growth and distribution of
POSTAL See wrts aiteoens cc attite tor
Mitochondria. The general functional
nificance of
Muscle fibers of the normal myocardium and
atrio-ventricular system (bundle of His),
interstitial granules (mitochondria) and
phospholipines in cardiac muscle. On the
occurrence and physiolonie ul significance
Olsateim theses so. ee ts oe eticemakercer te
Myers, J!. A. Studies ~ on the mammary
gland. I. The growth and distribution
of the milk-ducts and the development
of the nipple in the albino rat from birth
to tembweeks'otvagee..... 2.2 ea eivareevetays eee
Myocardium and_ atrio-ventricular system
(bundle of His), interstitial granules
(mitochondria) and phospholipines in
cardiac muscle. On the occurrence and
physiological significance of fat in the
muscle fibers of the normal..............
slg-
TJIPPLE in the albino rat from birth to
ten weeks of age. The growth and dis-
tribution of the milk-ducts and the
development of the...........0..0-0000
POSSUMS in captivity. The influence of
diet and iodides on the hyperplasia of
the thyroid sland of. ..¢s:2i..¢... e+...
ANCREAS in Amblystoma punctatum.
The development of the liver and.....
Parathyroid ¢lands of the albino rat. Effects
of inanitition upon the structure of the
EN YVOIMUAN Glee orice cee seems semas ee
INDEX.
447
211
447
447
353
353
57
211
Phospholipines in cardiac muscle. On the
occurrence and physiological significance
of fat in the muscle’ fibers of the normal
myocardium and atrio-ventricular sys-
tem (bundle of His), interstitial granules
(nitochondria)sand..:eress ansae = wees ee
Pig embryo, with special reference to the origin
of the erythrocytes. The microscopic
structure of the yolk-sae of the.......... 2
Placenta in different classes of animals. The
interrelations of the mesonephros, kidney
Cis (6 MBPS a CIE tae rls Gece eee etal ore
AT. Effects of inanition upon the struc-
ture of the thyroid and parathyroid
elands-of the albinos... 0..0.5s:4sscene
Rat from birth to ten weeks of age.
growth and distribution of the milk-ducts
and the development of the nipple in the
UOINO. 4 dred ou, Seok re ale one ee
ECRETION in the thyroid gland. The
normal MOdEeVOls sass hese cee wee eee
Sino-ventricular system as demonstrated by
the injection method. The..............
Sinus in the human embryo. The vascular
drainage of the endolymphatic sae and
its topozraphical relation to the transverse
SrreETEeR, Georce L. The vascular drain-
age of the endolymphatic sac and its topo-
eraphical relation to the transverse sinus
in the ume emibryOuss..- see. deem eee
Structure of the yolk-sac of the piz embryo,
with special reference to the origin of the
erythrocytes. The microscopic..........
GEES JID gland of opossums in captivity.
The influence of diet and iodides on the
hyperplasia of the
Thyroid gland. The normal mode of seecre-
TIOMUITN ENC cossasmcc endo ee Se ci gmeo gies aso
Thyroid and parathyroid glands of the albino
rat. Effects of inanition upon the struc-
ture of the..
Transverse sinus in the human embryo.
vascular drainage of the endolymphatic
sac and its topozraphical relation to the. .
ASCULAR drainage of the endolym-
phatic sac and its topozraphical relation
to the transverse sinus in the human
embryo. THE. ccas0ce8 neues Mewes feels
OLK-SAC of the pig embryo, with special
reference to the origin of the erythro-
cytes. The microscopic structure of the
179
305
353
37
149
67
67
277
67
67
277
ao 141t
1 |
i
PES A ae OE OD OE ;
Ah Ll ela AS A NE EMITS St BOG
See es
Se
a! mate
eer
SSS .
cieenveter
ai
bt
—s
—
=
oe
ace epetnnaige pans
Socsscoss=
eC ene
Serer neaek a
sco Sa
Ss eS
SSS :
panera ty a
aon SS een a = ners
en ee
SoS
Se
= ==>
=
ae
a
i
oa: - —— -
Sas SE SES Se SSS Se See
225-525-555
===>
== <—