View source for Paper - A contribution to the early development of the heart in mammalia, with special reference to the guinea pig

{{Header}}
{{Ref-Yoshinaga1921}}
{| class="wikitable mw-collapsible mw-collapsed"
! Online Editor&nbsp;
|-
|[[File:Mark_Hill.jpg|50px|left]] This historic 1921 paper by Yoshinaga describes the development of the {{guinea pig}} {{heart}}.
<br><br>
See also:
<br>
<br>

'''Modern Notes:''' {{guinea pig}} | {{heart}}
<br>
{{Guinea Pig links}}
<br>
{{Heart Links}}
|}
{{Historic Disclaimer}}

{{Ref-Yoshinaga1921}}

Yoshinaga T. [[Paper - A contribution to the early development of the heart in mammalia, with special reference to the guinea pig|A contribution to the early development of the heart in mammalia, with special reference to the guinea pig]]. (1921) Anat. Rec. 21(3): 239


A Contribution to the Early Development of the Heart in Mammalia, with Special Reference to the Guinea Pig 

239 

THE ANATOMICAL RECORD, VOL. 21, NO. 3 



Tanzo Yoshinaga 

Department of Anatomy, University of Michigan 

Twenty-Three Figures 

Introduction 

Since the fundamental investigations on the development of 
the heart in mammals by His, Born, and others were published, 
many prominent investigators have contributed to our knowledge 
concerning the earlier stages of development of the heart and of the 
pericardial cavity in representatives of almost all classes of vertebrates. The earlier workers began their investigations after the 
developmental stage in which the embryonal heart tube had already assumed the complete S form. From the results of their 
work, however, only very general conclusions can be drawn. In 
many important details the literature shows contradictions, while 
each theory advanced has been supported by investigators of 
recognized ability. 

The following work was undertaken at the suggestion of Professor Huber. I take this opportunity to express my hearty 
thanks to him for the use of his collection, which he placed at 
my disposal, and for his invaluable help. This work concerns 
itself chiefly with the origin of the endothelial tubes, the early 
development of the pericardial cavity, and the mode of confluence 
of the bilateral myocardial tubes, written with the hope of contributing something of interest to the early development of the 
heart. 

The material for this study was obtained from many uninterrupted series of embryos of the guinea pig, from the embryological collection in the Department of Anatomy of the University of 
Michigan, prepared by Professor Huber. The great majority of 
the embryos were fixed in Carnoy's fluid and all were cut into 
either cross or sagittal series at either 5^ or 7^ thickness, also a 
few at lOju thickness. 

In a study of the development of the endothehal and the myocardial tubes and of the pericardial cavity, to show their successive developmental changes and their relative topographical relations, a number of models were prepared according to Bom's 
method of wax-plate reconstruction. These were made at a magnification of 300 diameters, with the aid of the projection apparatus, by superimposing the drawings of every section of the 
cranial part of the embryo. The thickness of the wax plates was 
changed in proportion to that of the sections from 1.5 to 2.1 and 
3 mm., respectively. All the models described here are deposited in the Department of Anatomy at the University of 
Michigan. 

All the figures of the sections presented here were drawn at 
relatively high magnifications with the aid of the camera lucida 
and subsequently reduced to the proper size in reproduction. The 
figures of the models were prepared according to the method described by Doctor Atwell. I am grateful to Doctor Guild and to 
Mr. Smith for help in preparing the wax plates. 

LITERATURE REVIEW 

Before considering the material studied in this investigation, 
I wish to refer to some of the theories held by earlier writers. I 
do not wish to give a complete resume of the historical developments, but rather briefly to indicate the representative opinions 
directly concerning the subject. 

In the embryonic shield of mammals at first there are present 
two pairs of longitudinal vessels, one of them situated near the 
middle line on the entoderm on each side representing the supraintestinal longitudinal blood vessels, and the others are found primarily on the entoderm approaching the lateral margins of the 
embryonic shield and these representing the subintestinal longi


EARLY DEVELOPMENT OF HEART IN MAMMALIA 241 

tudinal blood vessels. In the cephalic portion of these latter the 
heart is formed, the walls of these vascular anlagen being peculiarly dilated and thickened. From this ontogenetical standpoint, 
the early development of the heart is merely a part of the development of the intra-embryonic vascular system. 

In the survey of the literature on the subject of the origin of the 
intra-embryonic blood vessels in mammals, the theories held by 
the earlier investigators, and still maintained, can be divided into 
three categories. 

The first theory was advanced by His, who made his observations of the flat embryonic shield of the chick. Hertwig, Kolliker, 
and others supported this theory. According to the Einwachsungslehre of His, with which the theory of the parablast was first connected, the early blood vessels of the embryonic shield are formed 
by a sprouting or ingrowth of the preexisting endothelial lining 
of the blood vessels, which had previously developed in the extra-embryonic vascular area. The ingrowth of the blood-vessel anlagen into the embryonic area takes place as solid, tenuous sprouts 
of the cells, which are primarily found to be anastomosed with 
each other, to form a close network in the area pellucida. 

These sprouts of cells enter into the embryonic area through the 
space between the splanchnopleure and entoderm, until they reach 
the somitic region, where they anastomose and become canahzed 
to form hollow vessels. Ultimately, this network of the endothelial lining forms the dorsal aortae, uniting in a longitudinal direction. The ingrowth of the endothelial sprouts is not limited to 
the somitic portion, but takes place also in the cranial part, entering from the lateral margin of the embryonic shield, through the 
space between the splanchnopleure and the entoderm, until they 
reach the heart anlage, where they develop the endocardium. 
The endothelial sprouts spread out, forming the ventral aortae 
and the blood vessels of the cranial region. The former blood 
vessels will be connected with the dorsal aortae which are prolonged cranial-ward by the sprouting and in the same way by turning over of the blind end of the pharynx ventral ward. 

Ttirsting studied the development of the aortae in the rabbit 
and confirmed the conclusion of His, namely, that these vessels are formed by a longitudinal anastomosis of the ingrown endothelial 
sprouts, derived from the extra embryonic area, through the space 
between the splanchnopleure and the entoderm. He noticed the 
early connection of the dorsal aortae with the vitelline plexus. 

Vialleton and Evans studied the development of the intra-embryonic blood vessels in birds and came to the conclusion that 
in birds the greater part of the descending aortae is developed by 
a conversion into a continuation of the aortae of the innermost 
strand of the capillary plexus, extended into the embryonic 
shield from the neighboring yolk sac. 

Lewis investigated the intra-embryonic blood vessels in rabbits 
from eight and one-half to thirteen days after insemination, and 
claims that from the network of vessels in the splanchnopleure 
of the yolk sac all intra-embryonic vessels are apparently derived as 
offshoots. The network ends mesially in two aortae. With the 
formation of the pharynx, this net is so folded as to produce 
dorsal and ventral aortae with the connecting first aortic arch. 

Bremer recently repeated the investigation of the same material 
and came to practically the same results. His summary is given 
as follows: ''In the rabbit, the dorsal aorta, the first aortic arch, 
the conus arteriosus, and the lateral heart are all parts of an original network of angioblast cords, derived from the extra-embryonic 
plexus" of blood vessels. Those portions of the network which are 
mechanically favored in their position persist, the other portions 
disappear. Although dealing in this paper with the development 
as found in rabbit embryos, I have examined various other species, 
as the chick, pig, sheep, etc., and feel satisfied that in all essential 
points the story of the development of these primary vessels in 
other vertebrates will be found similar to that here described." 

Many other investigators support this theory, while still others 
do not accept it. Ranvier seems to think that the Einwachsungslehre of His must be regarded as a simple hy]oothesis, and states: 
"Mais il est clair qu'aucum embryologiste n'a pu suivre ce developpement continu par bourgeonnement dans le corps neme de 
I'embryon; c'est laune simple hypotese. " 

The second theory of embryonic vasculogenesis was first advanced by Rabl, who asserts that the first aortic arch in amphibia is formed by the accretion and extension of the endothehal cells 
proliferated in the paired heart rudiments, when these endothelial 
tubes were developed. Moreover, he applies this possibility to 
other blood vessels of the embryo, in which they are formed by 
the extension of the preexisting endothehal cells, as, for instance, 
can be seen in the regeneration of the capillaries. His statement 
reads as follows : " Die Beobachtung, dass bei den Amphibien die 
ersten Aorten Bogen durch Auswachsen des Endothelsiickchen 
entstehen, legt uns aber noch die Frage nach, ob es vielleicht auch 
das Endothel aller anderen Gefiisse in letzter Instanz auf die 
Zellen des Endothel-sackchens zuruckzufiihren sei, mit anderen 
Worten ob nicht vielleicht all Gefasse in derselben oder in ahnlicher Weisse entstehen, wie die Capillaren. " 

Furthermore, in his later work,'' The Theory of the Mesoderm, " 
he repeated his assertion: ''Ich habe die erste Entwickelung der 
Gefasse namentlich an den Aorten wiederholt genau verfolgt, und 
ist mir kein Fall erinnerlich, der mich an der Ueberzeugung irre 
gemacht hatte, dass neue Endothelen immer nur aus bereits bestehenden ihren Ursprung nehmen." 

Riickert investigated the early development in the eggs of 
selachians, in which at first the subintestinal veins can be seen in the 
anterior embryonic shield. Here the endothehal cells are produced 
from the ventral angles of the lateral plates, detaching as free 
angioblasts, which subsequently assume forms of the cellular 
groups or chains between the splanchnopleure and entoderm. 
With regard to the anlage of the aortae in the anterior part of the 
embryonic shield, Riickert assumes still further that their origin 
is in loco and that here are anticipated the adjacent mesodermal 
somites, subordinately the dorsal wall of the gut comes under 
consideration. 

P. Mayer discovered the transverse blood vessels in the eggs of 
the torpedo, connecting the subintestinal veins with the aortae, 
and he attributed their origin to the emigrated cells from the 
ventral parts of the mesoderm, wandering along the gut wall dorsalward. Riickert agreed with the opinion of Mayer. In the vascular development of the embryo he claimed that the angioblasts 
still arise in loco, for example, the blood vessels of the pronephros from the visceral walls of the somite. The angioblasts of the 
mandibular blood vessels from the visceral plate of the second 
cranial somite as well as from the wall of the foregut. 

Raffaele, Emmert, and others have accepted the local formation 
of the embryonic blood vessels and are in agreement with the idea 
of Rlickert. After the first publication of his work in 1888, Riickert investigated all classes of vertebrates and confirms his claims 
to the local formation of the embryonic blood vessels. Concerning the development of the chick, he expressed himself as follows: 
''So findet also beim Huhn statt des Einwachsen der 
Gefassanlagen eine von der Peripherie des Keimwalles gegen den 
Embryo zu fortschreitende Differenzierung- derselben aus dem 
Mesoderm statt, was mit den Beobachtungen an Selachieren iibereinstimmt, Auch innerhalb des Embryo entstehen beim Huhn 
die Gefassanlagen durch locale Ausschaltung des Zellen Materials aus dem ]\Iesoblast, wo von ich mich bei der Herz so wie der 
Aortenbildung iiberzeugt habe. " 

Mollier reached the same conclusion, confirming Rlickert. 
After thoroughly studying the material of a wide scope, he concludes as follows: ''Es lasst sich also fiir die Genese der embryonalen Gefasse der Amnioten zur Zeit ein Urteil dahin fassen, dass 
die Lehre von der localen Entstehung der Gefasszellen auch hier 
Geltung besitzt und dass die von His und Vailleton gegebenen 
Fliichenbilder, ferner die Rekonstruktionsbilder von Turstig in 
dem Sinne zu deuten Sind, dass die im Embryo sichtbaren ersten 
Gefasszellenstriinge nicht als Sprossen ausserembryonaler Gefassanlagen entstanden sind, sondern vielmehr ihre Entstehung aus 
einzelnen, in loco entstandenen und netzformig vereinigten 
Gefasszellen nehmen." 

Sobatta not only supported the theory of Mayer and Rlickert, 
but he also emphasized that those blood vessels found on the walls 
of the yolk sac were derived from the intra-embryonic blood vessels as a result of their continuous outgrowth. 

New light was shed on this contradictory evidence by a number 
of investigators who employed the methods of experimental embryology and were able to show that the yolk-sac angioblasts may 
be kept out of communication with the intraembrj^onic blood 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 245 

vessels, thus lending evidence in favor of the local formation of 
the embryonal blood vessels. 

The mechanical separation of the vessels of these two portions 
was employed by Graper, Hahn, Miller and Mc\Miorter. These 
workers have obtained endothelium on both sides of the chick embryo, in which one side was severed from the extra-embryonic blastoderm. A further strengthening of the theory of the local formation is found in the recent experimental work by Reagan, who 
writes as follows: "In conclusion it is well to consider the following recently established facts which should share in defining 
our morphological interpretation. The yolk sac is not necessarily the site of formation of the earliest blood vessels. intra-embryonic blood vessels develop in situ when communication of the 
extra-embryonic vessels with intra-embryonic tissues is prevented 
by chemical or mechanical means." 

The method of exposing the developing embryo to diluted anesthetics was employed by Stockard, who claimed as a result of his 
work that in Fundulus embryos the heart endothelium and the 
aorta arise in loco within the embryo, as the blood vessels, even the 
mesoderm are absent in the yolk sac in the cranial portion. 

Among the investigators who have accepted the local formation 
of the intra-embryonic blood vessels, opinions are still quite divergent at present as to from what part of the embryonal blastoderm the angioblasts are differentiated. I shall not again thoroughly discuss this point, as I have already done so in my previous paper on this subject. But I shall add that there are many 
authors who consider the origin of the angioblasts in mammals as 
derived from the entoderm. Martin figures a cross-section of a 
cat embryo 2 mm. long and in his prominent " Lehrbuch der Anatomic der Haustiere" states: Wahrend man liber die erste Entwickelung der Blutgefasse im Embryonalleib noch nicht im Klaren 
ist, kennt man die Bildung der peripheren Gefasse genau. Ihrer 
Abstammung nach sind die Innenwand (endothelium) auch der 
embryonalen Gefasse wie die des Herzens entodermaler Natur, 
wahrend die iibrige Wand vom Mesoderm geliefert wird. Ich 
finde bei Katze die erste gefassbildenden Zellen und Zellgruppen 
eininnigsten Zusammenhand mit dem Entoderm desDarmes, ja so gar schon rundliche Spaltildungen in Entoderm selbst. Auch die 
ausserembryonalen Blutgefasse sind nach Rabl entodermalen Ursprungs. " In the figure depicted by Martin, angioblasts can be 
seen, connected with both the mesoderm and the entoderm. 

In the study of a perameles embryo of 6.08 mm. in length, 
Miss Parker says with regard to the origin of the endothehal cells : 
" The evidence of this stage does not justify any definite statement 
with regard to the origin of the endothehum of the heart. " And, 
moreover, she adds, that in an earher stage the appearance by no 
means excludes the possibility of the entodermal origin of the 
endothelium. 

After investigating the origin of the endothelial and blood cells 
in the embryo of the ferret, Wang comes to the following conclusion: ''The facts revealed by the study of the early stages in the 
development of the ferret point to the conclusion that, whilst 
blood cells and vascular endothelium are closely related to each 
other and are formed invariably between the mesoderm and entoderm, there is evidence to show that, in the ferret, the origins of 
these two vascular elements are separate and distinct — the blood 
cells arising from the entoderm and the vascular endothelium from 
the mesoderm." 

"If the biphyletic origin of the blood cells and vascular endothelium is to be accepted, two more points still remain to be solved, 
namely, how, when, and where the first blood cells enter the circulation. Unfortunately, the ferret embryos, at present worked 
upon, provide no definite evidence on this point, but it is quite 
clear that angioblast cells are formed outside the embryonic area, 
and that blood vessels are formed inside the embryonic area, and 
are at first devoid of blood corpuscles. " 

AVe shall now briefly review the data concerning the mode of the 
fusion of the bilateral heart anlagen; there are many divergent 
opinions. 

Since Hensen first declared that in mammals the heart anlage 
is bilateral, one on each side of the embryo not far from the midsagittal plane and on the ventral aspect of the pericardial cavity, 
it was long believed that these two lateral heart anlagen ultimately came in contact and fused together in the middle ventral surface of the embryonic shield, forming a single secondary heart 
tube. Even though the secondary fusion of the bilateral endothelial tubes is universally acknowledged, there are many contradictory theories concerning the formation of the single myocardial 
heart anlage together with the mode of fusion of the bilateral 
primitive pericardial cavity. In his work on comparative embryology, Balfour thus speaks: ''In mammals the two tubes out of 
which the heart is formed, appear at the sides of the cephalic 
plates, opposite the region of the mid and hind brain. They arise 
at a time when the lateral folds which form the ventral walls of 
the heart, are only just becoming visible. On the formation of 
the lateral folds of the splanchnic walls, the two halves of the 
heart become carried inwards and downwards, and eventually 
meet on the ventral side of the throat. For a short time they 
here remain distinct, but soon coalesce into a single tube. " 

Minot writes: "In mammals by the bending down of the layers and the expansion of the coelom the vorderdarm is shut off and 
the lateral heart anlagen are brought together in the median line 
below the vorderdarm, and there fuse into a single thick tubular 
wall around the double endothelial heart ; it is not long, however, 
before the endotheUal tubes also fuse into one. As in the chick 
the two mesothelia, when the median heart arises, form a membrane (mesocardium) by which the heart is attached to the tissue 
above and below; both mesocardial membranes break through, 
putting the two coelematic cavities into communication and leaving the tubular heart suspended by its ends. " 

The theory of the fusion of the lateral folds of the splanchnic 
walls, enclosed within the laterally placed pericardial cavities at 
the ventral side of the foregut in mammals, as in amphibia or 
birds, as many investigators have asserted and now believe, is 
supported by the following authorities: Balfour ('81), Strahl and 
Carius C89)^ Tandler ('12), Wilson ('14), Bryce ('08), Minot ('92) 
Bailey ('12), Schultze ('15), Dandy ('10), Martin ('02), and Arey 
(' 17) . From this it would seem that, as many of the above named 
authors claim, the heart must be provided, at least temporarily, 
with a ventral and a dorsal mesocardium. 


In a similar way in his valuable work on the first heart anlage. 
MoUier says: "Das Mesocard, ventral oder medial gelegen, ist 
dem dorsalen der Anamnier zu vergleichen. Ein dem ventralen 
entsprechendes kann erst nach dem zusammenstossen beider 
Pleuro-pericardialhohlenwande gebildet werden. Beim Meerschweinchen hingegen liegen die ersten Herzzellenstrange lateral 
von den Firsten der Darmfalten, und sie werden durch den Darmschluss gar nicht unter die ventrale Darmwand verlagert, wie beim 
Kaninchen, sondern riicken, zwischendorsalerand ventraler Mesocardfalte gelegen, einander naher, bis zur Beruhrung und endlich Verschmelzung. Doch erfolgt auch hier, wie aus der Figur 
ersichhch, der Durchbruch des ventralem Herzgekroses zuerst." 

Robinson has pointed out that the formation of the foregut is 
mainly attributed to the unproportionately rapid development 
of the embryo over the relatively stationary line between the embryonal and extraembryonal areas. If the idea were true, that 
lateral folds of the mesoderm converge ventrally until their entodermal covering has met together in the ventral middle line and 
both lateral pericardial cavities have fused together beneath the 
ventral walls of the foregut, then the heart is not only attached 
by the dorsal mesocardium to the foregut, but also by the ventral 
mesocardium to the ventral wall of the body. However, Robinson denies this generally accepted idea and the existence of the 
ventral mesocardium absolutely, at any time in the development 
of mammals, and he applies this fact to support his theory, that 
the separation of the foregut from the yolk sac is not due to the 
tenaciously held process of the tucking in of the margins of the 
embryonic area, but to the fact that the relatively slow-growing 
margin is demarcated between the embryonic and extra-embryonic portions of the wall of the ovum, which rapidly increase their 
their extent, expanding over the boundary margin. According 
to him, in mammals, the pericardial mesoderm is present in the 
pericardial portion of the embryonic area, and it is completely 
separated into somatic and splanchnic layers before the head fold 
appears. There is therefore a single pericardial cavity which extends from side to side along the anterior boundary of the embryonic shield. As the head fold is formed, the pericardial region 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 249 

is reversed and it is carried into the ventral wall of the foregut, 
where it is present as a U-shaped tube, which communicates with 
the general coelom at each end. The rudiments of the heart are 
formed in the splanchnic layer of the pericardial mesoderm. 
Therefore, after the reversal of the pericardial area, they He on 
the dorsal wall of the pericardial cavity, attached to the ventral 
wall of the foregut by the dorsal mesocardium. 

Prior to Robinson, Ravn pointed out the correlation between 
the formation of the foregut and the forward growth of the embryonic shield. He also thoroughly described the mode of reversal 
in the primitive pericardial cavity. 

On the other hand, many investigators believe that an active 
backward progression of the foregut opening occurs, in addition 
to the forward growth of the head fold. They deny the actual 
fusion of the lateral mesodermal folds in the ventral middle line. 
Rouviere agrees with Robinson as to the absence of the ventral 
mesocardium in mammals, while he does not dismiss the meaning 
of the forward growth of the head fold on the formation of the 
foregut. According to his account, in a rabbit embryo of 201 hours, 
both the pericardial cavities (les deux cavities parietales) show 
the separated bilateral canal on each side, which has grown forward around the anterior end of the head fold and become fused 
together from a single continuous cavity in the embryo of 207 
hours. The splanchnopleure forming the caudal wall of the pericardial cavity assumes a continuous fold, which Tourneux designated as the cardiac fold (repli cardiaque) and which he considered as growing backward automatically as a whole. As the free 
edge of the splanchnopleural fold has progressed always in advance of the primordial heart, no fusion of the splanchnopleure 
is involved and also no ventral mesocardium is formed. 

In a description of the growing processes in the developing chick 
embryo, having kept them under direct observation while still 
ahve, Graper asserts that there is considerable evidence in support 
of the view that the margin of the foregut (Darmpforte) moves 
caudally, concurrently, with the forward growth of the head fold. 
Moreover, he marked out diagrammatically the mode of the backward progress of the foregut opening and a quite different manner 
of the closure of the foregut than that of the medullary canal. 


Uskow also claims the automatic backward progress of the foregut opening, according to the increase of the pericardial cavity. 

In her study of the early stages in the development of marsupials, Miss Parker declares that the forward growth of the 
head fold doubtless plays an important part in the initiation 
of the formation of the foregut and that the actual backward 
growth of the foregut opening, but not the fusion of the lateral 
folds, brings about the lengthening of the foregut. 

In the study of the early development of the heart and cranial 
blood vessels in ferret embryos, Wang agrees with Miss Parker 
in the absence of the ventral mesocardium, there being no fusion 
of this part of the pleuropericardial wall, nor that any part of 
the gut closure is effected by the fusion of the lateral folds. 

OBSERVATIONS 

Stage I 

The material for this stage consists of many specimens removed 
from the uterus of the guinea pig thirteen days and twelve 
hours or fourteen days and eleven hours, respectively, after 
insemination. Some of them were cut longitudinally and the 
others transversely. 

A. The first embryonic shield which came under consideration, 
was removed from the uterus thirteen days and twelve hours 
after insemination and was in a cross-section having a 7 ^ thickness; 233 sections fell to the embryonic shield. The head fold 
had not begun to develop. The shallow neural groove was 
present on the surface of the embryo. The primitive streak 
was well developed and terminated caudally in a shallow notched 
groove. The mesoderm was thickened in the caudal part of 
the embryonic area and indicated the allantoic mesoderm. In 
the notochord there was present the chordal canal at its caudal 
end, but in other parts it was spread out to form a chordal plate 
through dehiscence of the ventral wall. The caudal end of 
the notochord was fused with an area of ectodermal proliferation at the cranial end of the primitive streak. 



EARLY DEVELOPMENT OF HEART IN ALAMMALIA 251 

In the mesoderm there was observed no evidence of mesodermic somites nor could the coelomic cavity be detected. The 
mesoderm consisted of two lateral wings on each side in the 
cross-section, separated completely in the middle line by the 
notochord, except in the cephalad end of the embryonic shield 
and in the region of the primitive streak. In the cranial end 
of the embryonic shield, that is, the part distal to the future 
pharyngeal membrane, in which the ectoderm and entoderm were 
coalesced, the lateral wings of the mesoderm were fused, continuing 
caudally into the lateral wings, but sharply terminated against 
the extra-embryonic area cranially and laterally. In this portion 
of the mesoderm, namely, in the pericephalic mesoderm, it formed 
a thinner layer than anywhere else. In the primitive streak 
a large mass of undifferentiated mesodermal cells was fused 
with both lateral wings, obUterating the demarcation between 
the mesoderm and ectoderm. In the mesoderm two layers of 
the 'cell band could not be distinguished; the dorsal surface 
of the mesoderm was, in general, compact, its outhne was clearcut; here the spindle-shaped nuclei had a relatively regular 
arrangement. 

Between the dorsal surface of the mesoderm and ectoderm, 
as well as between the ventral surface of the mesoderm and 
entoderm, there could be distinguished clear intervals, which 
could be attributed in large measm^e to shrinkage. The mesodermal cells were spread out in two or three layers and were 
spindle-shaped and connected wdth each other by short protoplasmic processes. In the ventral surface of the mesoderm 
a loosening of the cell band could be seen, characterized by an 
increased distance between the respective nuclei. Intercellular spaces became more distinct and wider, the spindle-shaped 
nuclei had no definite arrangement. In nearly all sections there 
could be demonstrated some free, isolated cells, detached from 
the cell band of the mesoderm, lying between the ventral surface 
of the mesoderm and entoderm, as shown in figure 1. According to His, these cells are identical with the angioblasts. In 
some sections the angioblasts are connected with the indented 
ventral margin of the mesoderm by broad protoplasmic bridges; in some other sections the protoplasmic bridges are narrow, 
the cells are distinctly pedunculated; in still other sections they 
are joined to the mesoderm by faint fibrils. Frequently a mitotic figure can be recognized in the mesodermal cells, adjacent 
to the angioblasts. These findings show that there can be no 
doubt of a distinct proUferative activity of the mesodermal 
cells; furthermore, every transitional feature of the migration 
or the detachment of the mesodermal cells, which apparently 
are destined to become anigioblasts, point out the fact that 
these angioblasts have originated in the ventral surface of the 
mesoderm. 

Figure 1 was reproduced from the forty-ninth section, 
counting from the cephalic border of the embryonic shield, and corresponds approximately to the future hindbrain region, in 
which region endotheUal tubes had been differentiated from the 
angioblasts in the next stage. As the contour of the entoderm 
is distinctl}^ demarcated from these cells in our specimens, as 
the figure shows, it may be excluded from direct participation 
in the formation of these cells, 






LEGEND LETTERS FOR ALL THE FIGURES 

A., atrium F.G.O., foregut opening 

Am., amnion I. M.S., intermesocardial space 

Aug., angioblast L.P.C., lateral pericardial cavity 

Ao., aorta M.C., myocardial cavity 

A.V.C., atrioventricular constriction Mes., mesoderm 

A.V.Ca., atrioventricular canal M.G., medullary groove 

B., bulbus cordis M.T., myocardial tube 

B.V.C., bulboventricular constriction A^., notochord 

C.M.L., craniomedian limb of the peri- P., pericardium 

cardial cavity P.C., pericardial cavity 

Co., coelom P.M., pharyngeal membrane 

D.M., dorsal mesocardium P. P. P., pleuropericardial passage 

D.W.P., dorsal wall of pericardium S.A.C., sino-atrial constriction 

Eel., ectoderm Som., somatopleure 

End., endothelium Spl., splanchnopleure 

Ent., entoderm S.R., septum ridge 

E.O., endothelial offshoot S.V., sinus venosus 

E.T., endothelial tube T.A., truncus arteriosus 

F.G., foregut V.W.F., ventral wall of the foregut 

Fig. 1 The 49th section of a series of 233 cross-sections of 7 m thickness of an 
embryonic shield of the guinea pig, removed 13 days 12 hours after insemination. 
The early stage of the formation of the angioblasts from the ventral surface of 
the mesoderm. X 150. 

Fig. 2 The 41st section of a series of 237 cross-sections having a 7 m thickness, removed from the uterus of a guinea pig 14 days 11 hours after insemination. 
The ventral surface of the mesoderm becomes loosened and angioblasts are 
separated. X 150, 


Fig. 3 The 40tli setion of a series of 162 sagittal sections having a 7 ^i thickness. This embryonic shield was removed from the uterus of a guinea pig 13 
days 12 hours after insemination. The angioblasts form the cell strands between 
the splanchopleura and entoderm. X 150. 

Fig. 4 The 63rd section of a series of 300 cross-section of 7 m thickness of an 
embryonic shield of a guinea pig, removed 13 days 11 hours after insemination. 
In the lateral plate of mesoderm the discontinuous coelom is present. X 150. 




B. The next embryonic shield selected for discussion was removed from the uterus of a guinea-pig fourteen days and eleven 
hours after insemination. It is cut in cross-sections, having 
a 7 M thickness, and 237 sections fall to the embryonic area. 
The flat incipient head fold has begun to develop, marking 
definitely the cranial margin of the neural plate. In this portion 
the ectoderm has been elevated above the surrounding embryonic 
shield. A broad and shallow neural groove, which gradually 
narrows cephalad, is present on the surface of the embryo. The 
primitive streak is well distinguished on the embryonic surface 
as of a transitional portion from the cranial neural groove into 
the caudal primitive groove. The notochord has acquired a 
tubular form at its caudal end in some more anteriorlj^ placed 
portions, but in the majority of the sections its ventral wall is 
opened into the yolk sac, to form the chordal plate. The caudal 
end of the notochord is fused to the cranial part of the primitive 
streak. No mesodermal somites are recognized nor can any 
indication of the coelomic cavity be detected. 

Figure 2 was reproduced from the forty-first section, counting 
from the cephalic amnion attachment. The mesodermal layer 
is thicker than that found in the preceding specimen; this is 
due partly to the numerical increase of its component cells and 
partly due to a loosening of the arrangement of the cells. In 
the mesoderm no layers can be distinguished, the spindle-shaped 
cells having no definite arrangement. The loosening of the cell 
is more readily demonstrable on the ventral surface of the mesoderm; its ventral outline forms a zigzag contour, due to the 
shorter or longer protoplasmic processes which are seen 
extended from the ventral row of the mesodermal cells toward 
the underlying space. Moreover, as can be pointed out in the 
figure, some cells are projected into the underlying space beyond 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 255 

their surrounding group of cells, while others present a mitotic 
figure and have their axis directed to the space under the mesoderm and are pedunculated into the underlying space, but 
remain connected with the mesodermal layer by means of their 
narrow protoplasmic bridge. There can be seen a few spindleshaped cells, which appear completely detached from the mesoderm and he scattered between the mesoderm and entoderm. 
The area of the distribution of these cells, which we regard as 
angioblasts, more numerous in this embryo, extends over a 
wider range than is observed in the previous embryo. A glance 
at the figure will prove that the origin of these cells is derived 
from the mesoderm of the splanchnopleure. 

In figure 3 there is presented a drawing of a sagittal section 
of an embryonic shield of approximately the same stage of 
development as that described under figure 2. 

This series belongs to an embryonic shield of a guinea pig, 
removed from the uterus thirteen days and twelve hours after 
insemination. It includes 162 sections, having a 7 m thickness. 

The figure was reproduced from a drawing of the fortieth 
section, counting from the lateral amnion attachment. This section passed through the flat head fold near its lateral margin. 
The fine of sectioning in this series was almost parallel to the 
mid axis of the embryonic shield. A study of the series shows 
that the developmental stage is just prior to the formation of 
the first mesodermal somite, which is indicated but not completely formed. The primitive streak extends approximately 
a third of the length of the embryonic shield. The general 
finding of the mesoderm is similar to that of the foregoing specimen. In the midsagittal plane the cranial end of the chordal 
plate terminates insensibly in the entoderm, where the pharyngeal membrane will be recognized. Cephalad to this membrane the pericephahc mesoderm is observed; its cranial limit 
terminates freely at the cranial amnion attachment. This pericephahc mesoderm is continued into both mesodermal wings 
caudolaterally. The ventral surface of the mesoderm is loosened 
and presents a coarse appearance. Between the ventral surface 
of the mesoderm and the continuous layer of ectoderm angioblasts can be seen forming cell strands, ranging one after another, 
approximately parallel to the long axis. In the cross-section these 
cell strands may be shown as single cut cells. The faintly stained 
protoplasmic processes or sUghtly rotated, tenuous protoplasmic fibrils are given off from the surface of the cell strands. 
Some of these anastomose with each other and others are connected with the cells of the adjacent mesoderm. In this fashion 
their protoplasmic fibrils form a kind of feltwork between the 
mesoderm and entoderm. In brief, it is only a repetition of 
the processes which produce the angioblasts from the mesoderm 
of the splanchnopleure, as observed in the preceding embryo, 
but the angioblasts are a step further differentiated. 

C. In figure 4 there is presented a cross-section drawing of 
an embryonic shield of a guinea pig, removed from the uterus 
thirteen days and eleven hours after insemination. This series 
includes 300 sections, having a 7 // thickness. In actuality 
this embryo presents only a slight advance in development 
over that discussed under figures 2 and 3. 

The first mesodermic somite is indicated, but not completely 
formed. The intra-embryonic coelomic space, which may be 
regarded as the future primitive pericardial cavity, considering 
its topographic position, shows simply a beginning of a very 
narrow cleavage in the lateral mesoderm of the cranial portion 
of the embryo. In some sections the two layers of the lateral 
plate of the mesoderm are separated from each other, and there 
can be seen narrow, discontinuous clefts between the mesodermic 
layers, while in other sections the whole mesoderm remains 
apparently solid. As a transition of these two extremes, in 
still other sections the coelomic space is shown as little more 
than a lineal cleavage. In brief, the coelomic cavity is forming 
from multiple foci and is not connected with the extra-embryonic 
coelom. In the pericephalic mesoderm no coelomic space can 
be seen. 

The figure reproduces a drawing from the sixty-third section, 
in which there can be seen an irregularly outlined splanchnopleure, from a relatively clear-cut contour of the somatopleure, 
separated by an incipient slip of the coelomic space. The 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 257 

splanchnopleure shows a slightly thicker layer of spindle-shaped 
cells; in some of them mitotic figures are present, indicating a 
proliferative activity. A number of angioblasts are scattered 
singly, while some others are grouped in a flat strand between 
the ventral surface of the mesoderm and entoderm. 

Practically the same stage of development as that described 
in figure 4 can be seen in an embryonic shield, removed from 
the uterus fourteen days and four hours after insemination. 

In figure five we see a drawing of the thirty-second section, 
counting from the lateral amnion attachment. This series, 
cut in the sagittal plane, includes 156 sections, having a 7 /x 
thickness. This section passes through the well upwardly projected head fold near its uphfted lateral margin. Under this 
head fold there can be recognized five discontinuous coelom 
spaces in the lateral plate of the mesoderm, each of which is 
interrupted by a substantial bridge. The mesodermal cell layer 
of the splanchnopleure is distinctly thickened and loosened. The 
spindle-shaped mesodermic cells assume a somewhat irregular 
arrangement. A number of them have disposed themselves in 
such a direction that their long axis is vertical to the ventral surface. A number of angioblasts are scattered under the mesoderm 
of the splanchnopleure and some of them are connected with this 
layer by their protoplasmic processes. Mitotic figures, seen in 
some of the mesodermal cells, show their proliferative activit3^ 
In some of the sections the discontinuous coelomic space can be 
seen in the pericephalic mesoderm, but it entirely disappears 
as it approaches the midsagittal plane, where the mesodermic 
layer has remained still in a solid condition, as can be found 
in figure 5B. In this respect this embryonic shield differs from 
that shown by Robinson and of several other workers. Robinson 
says that in mammals the mesodermic layer extends through 
the pericardial portion and is cleft into somatic and splanchnic 
layers before the head fold is formed. In our specimens the 
intra-embryonic coelomic space is present discontinuously in 
the cranial region of the embryonic shield and totally absent 
as it approaches the middle plane of the pericephalic mesoderm, 
even though the head fold is already formed. 


Stage II 

The material on which the following description of stage II 
is based consists of several embryos, certain of which are cut 
transversely and others longitudinally. 

A. This specimen was removed from the uterus of a guinea 
pig fourteen days and eleven hours after insemination. The 
series includes 307 cross-sections, having a 7 ^ thickness. A 
plastic reconstruction of the cranial portion was made with the 
Born method, and the whole shield, reconstructed for another 
purpose, was used for this study. This embryonic shield, as 
the model shows, presents a flat head fold, which is slightly 
more elevated than in the previous stage. The head fold can 
be divided into two primary parts. The cranial part is long and 
wide and projects over the cranial and lateral walls of the cranial body elevation. The caudal part is small and passes insensibly into the spinal portion. There is present a well developed medullary groove in the cranial portion of the embryonic 
shield and its caudal end becomes gradually shallower until 
it disappears at the primitive streak in the caudal fourth of 
the embryonic shield. Its cranial end is terminated near the 
cranial extremity of the head fold. The deepest portion of the 
medullary groove corresponds to the region of the hindbrain 
plate. A well-marked anlage of the trigeminus is present as 
a thickening of the ectoderm. There are present three pairs of mesodermic somites with a fourth pair forming. The foregut 
has begun to develop in the embryo; for the length of five sections its lumen is invaginated upward into the head fold. 





Fig. 5A The 32nd section of a series of 156 sagittal sections of 7 m thickness of 
an embryonic shield of a guinea pig, removed 14 days 4 hours after insemination. 
The mesoderm is thickened and loosened, a number of discontinuous coelomic 
spaces is present, angioblasts are being produced from the splanchonpleura. 
X 150. 

Fig. 5B The 77th section of the same series from which figure 5 was drawn, 
passing through practically parallel to the midsagittal line. The pericephalic 
mesoderm shows a relatively thinner layer than elsewhere and lies anterior to the 
primitive pharyngeal membrane, as the foregut has not yet developed. In 
accordance therewith, the reversal of the preumbilical region of the embrj'onic 
body does not occur. X 150. 

Fig. 6 The 87th section of a series of 307 cross-sections of 7 ^ thickness of 
an embryonic shield of a guinea pig, removed 14 days 11 hours after insemination. 
The pericardial cavity opens widely, under the thickened splanchnopleura the 
endothelial tube is first formed in this embryo. X 150. 



Fig. 7 The 27th section of the same series from which figure 6 was drawn 
embracing the nearly anterior margin of the head fold. The pericephalic mesoderm separates into two layers, by the lineal coelem space, which continues into 
the lateral pericardial cavity. X 150. 



Both the lateral primitive pericardial cavities are presented 
in the lateral plate of the mesoderm (fig. 6), which is completely 
separated into two lateral wings by the notochord, except in 
the cranial end of the embryonic shield, namely, beneath the 
cranial extremity of the head fold and in the region of the primitive streak. In these places two lateral wings come to fusion. 
In the mesoderm, which is produced by the fusion of both the 
lateral mesodermic wings in the middle line, beneath the cranial 
extremity of the head fold, in front of the pharyngeal membrane, 
a lineal cleavage can be recognized (fig. 7) by which the 
mesoderm is separated into two distinct layers. This coelemic 
space in the pericephalic mesoderm is formed by a forward 
extension of both the lateral primitive pericardial cavities into 
pericephalic mesoderm. These communicate with each other 
through this pericephalic coelomic space, which is now forming 
the craniomedian limb of the inverted U-shaped pleuropericardial cavity and may be accounted for as the essential future 
pericardial cavity. This cavity communicates freelj'^ with the 
future pleural cavity, which, in turn, passes into the peritoneal 
coelom. But this does not communicate with the extraembr}^onic coelom (fig. 8). 

The reconstruction of the whole shield shows that the pleuropericardial cavity corresponds to a vague sweUing presented 
by the ectodermal laj^er on the dorsal surface of the model 
along the lateral margins of the neural plate, and their cranial 
extremities are connected with each other directly beneath the 
the cephalic end of the head fold. Therefore, the cranial extremities of the primitive pericardial cavity and of the head 
fold fall practically in the same level. The caudal extremities 
of both lateral pericardial cavities gradually disappear at the 
level of the caudal termination of the neural plate. And this 
corresponds to the gradual diminution of the prominent ectodermic swelling on the surface of the model. In this fashion, 
therefore, the rhomboidal shaped head fold is surrounded by the horseshoe-formed primitive pericardial cavity both laterally 
and cephalad. 

In this stage the dimension of the craniomedian limb of the 
pericardial cavity is yet very narrow. Its ventrodorsal extent 
is not more than a Hneal cleavage (fig. 7), while its craniocaudal 
length extends throughout five sections. In tracing the lateral 
limbs from the craniomedian limb, however, the pericardial 
cavity gradualh^ increases in width until it reaches its maximum 
opposite to the hindbrain region (fig. 6), and then again a gradual reduction takes place behind this region until the coelomic 
cavity has completely disappeared in the region of the first 
mesodermic somite. 

On the ventral surface of the model there can be seen the 
crescentic gut-groove of the entoderm at the cranial portion 
of the embryonic shield its apex being directed cranial-ward and 
deepened gradually until it reaches the opening of the fore-gut, 
which is invaginated cranial-ward between the ectodermal head 
fold and the craniomedian Umb of the pericardial cavity, as a 
horizontal diverticulum of the yolk sac. The base of the crescentic gut-groove is directed caudalward, and gradually becomes 
shallower, until it has entirely disappeared at the level of the 
hindbrain. On both sides of the gut-groove and along the foregut opening there is a rounded ridge running from the laterocaudal to the craniomedian end; in both position and direction, 
this ridge corresponds to the horseshoe-shaped pericardial cavity. 

In this stage a number of angioblasts are scattered under 
the considerably thickened mesoderm of the splanchnopleure 
throughout the full extent of the pericardial cavity. A few 
angioblasts can be seen in the portion of the narrowly opened 
craniomedian limb of the pericardial space in which the mesoderm 
of the splanchopleure is in close contiguity with the underlying entoderm, and they increase in number toward the lateral 
limbs. In some sections in which the wide open pericardial 
cavity is present, endothelial tubes can be seen differentiated 
from the angioblasts lying under the mesoderm of the splanchnopleura, which is subsequently elevated from the entoderm 
and projected into the pericardial cavity as a prominent fold, 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 263 

as depicted in figure G. As the differentiation of the angioblasts into the endothelial tubes takes place irregularly, the 
distribution of both kinds of cells intermingles irregularly with 
reference to the level of sections; for example, even in the wide 
open pericardial region in some sections a tubular endothelium 
appears, while in the next succeeding section there are merely 
scattered angioblasts or cell strands of angioblasts. For this 
reason the invagination of the mesoderm of the splanchnopleura into the pericardial cavity as a prominent fold cannot 
be attributed simply to the development of the endothelial 
tubes, resulting in the increase of their volume. In this embryo, 
generally speaking, the angioblasts apparently predominate over 
the endothelium. 

The dorsal aortae can be seen developed in the cephalad 
portion of the embryonic shield, while in many sections they 
are present as incompletely formed endothelial tubes resting 
on the entoderm on both sides of the notochordal plate (fig. 6), 
in their caudal extent they remain as cell strands of angioblasts. 

In this embryo the foregut is present as a completely formed 
short entodermic tube, and the gut-groove, following caudally, 
is shown as a deep furrow. Not until a later stage is reached 
are the endothehal tubes completely developed. This corresponds apparently with the development of the human embrj^o, 
as described by Graf Spee. In the slightly younger embryonic 
shield the coelemic cavity in the pericephalic mesoderm is not 
yet formed in the median plane, where this is separated in many 
places by thin mesodermic bridges. But this condition is only 
temporary and both lateral pericardial cavities will communicate 
with each other, as is shown in this embryonic shield. 

B. In figure 9 is represented a drawing of a midsagittal section of an embryonic shield of a guinea pig, removed from the 
uterus fourteen days and twenty-three hours after insemination. 
This series includes 172 sections, having a 7 m thickness. The 
figure represents a drawing of the eighty-fifth section, and this 
passes through the embryo just parallel and just lateral to the 
midaxis. As reckoned by age, this embryonic shield is just 
slightly older than that discussed under figures 6, 7, and 8 but, judged by the stage of general development, the same findings 
can be noted. Here it is noticed that the craniomedian limb 
of the pericardial cavity is not more than a lineal cleavage in 
the pericephalic mesoderm, which extends in this plane from 
the more anteriorly situated extra-embryonic mesoderm up to 
the pharyngeal membrane, where the mesoderm is entirely absent 
and the ectoderm and entoderm are in close contiguity. 

In the peripheral two-thirds of the mesoderm in this section 
the cells have no definite arrangement, but are loosely and 
irregularly scattered. This portion can be seen as a continuation 
of the extra-embryonic mesoderm, and the same condition is 
shown in figure 5B in the younger stage. In the central third 
of its extent the two layers of the cell band can be distinguished, 
namely, the mesoderm of the splanchnopleura and the somatopleura, and between them lies a craniomedian limb of the 
paricardial cavity. Its craniocaudal extent is short and its 
cranial extremity is approximately on the same level with 
that of the head fold. Its sagittal long axis forms an obtuse 
angle with the sagittal long axis of the embryonic shield. 

¥\g. 9 The 8.5th section of a series of 172 sagittal sections, having a 7 m thickness. This embryonic shield was removed from the uterus of a guinea-pig 14 
days and 23 hours after insemination. The pericephalic mesoderm separates 
into two laj^ers by the lineal coelem space, through which the two lateral pericardial cavities communicate with each other. The foregut has just begun to 
develop. X 150. 

Fig. 10 The 83rd section of a series of 166 sagittal sections of a 7 m thickness 
of an embryonic shield of the guinea pig, removed 14 days 12 hours after insemination. The craniomedian limb of the pericardial cavity in the pericephalic 
mesoderm is wide open. The ventral wall of the pericardial cavity in figure 9 
forms the caudal wall of it in this figure, as the reversing process occurs in the 
preumbilical region of the embryonic body, in conjunction with the development of the foregut, which in this embryo shows a longer lumen than in figure 9. 
X 150. 

Fig. 11 The 4oth section of a series of 318 cross-sections, having a 7 /x thickness of an embryonic shield of a guinea pig, removed from the uterus 15 days 
14 hours after insemination. This section passes through the forebrain plate 
near its anterior margin. The craniomedian limb of the pericardial cavity is 
wide open. The splanchnopleural fold projects into the pericardial cavity, 
rising from the underlying entoderm. This is especially prominent on the left 
side of the figure. Between the splanchnopleura and entoderm a number of 
angioblasts are scattered. X 150. 




In the younger stage the pericephaHc mesoderm is situated 
anterior to the pharyngeal membrane, approximately in the 
same horizontal plane with the embryonic shield (fig. 5B). But 
in this stage it is brought ventrally to the pharyngeal membrane, 
as the foregut has begun to develop; the reversal of the preumbilical portion of the embryonic body accompanies this 
development. 

C. In figure 10 is presented a drawing of a midsagittal section 
of an embryonic shield of a guinea pig, removed from the uterus 
fourteen days and twelve hours after insemination. This series 
includes 166 sections, having a 7 /x thickness. The figure is 
reproduced from a drawing of the eighty third section, passing 
through almost exactly parallel to the midaxis of the embryonic 
shield. As measured by age, this embryonic shield is slightly 
younger than that discussed under figure 9, but the general 
findings of the development are slightly in advance of that of 
the latter. 

The craniomedian limb of the pericardial cavity, which lies 
under the foregut, extends dorsocaudally in a cranioventral 
direction, and has increased its dimensions in both the ventrodorsal and craniocaudal directions (fig. 10). Compared with 
the foregoing embryo (fig. 9), its sagittal long axis forms an 
angle a little more acute with the longitudinal axis of the 
embryonic body. Its caudal half presents the crescentic coelom 
cleavage, directing its convexity caudo ventral ward, while its 
cranial half still remains as a lineal slit. The cranial extremity 
of the craniomedian limb of the pericardial cavity is practically 
situated on the same level with that of the head fold (fig. 10). 
It may be demonstrated, when we compare figures 9 and 10, 
that the backward movement of the foregut opening from the 
cranial extremity of the head fold is greater than the rate of 
progress of the head fold forward from a certain fixed point. 
This actual backward progress of the foregut opening brings 
about the lengthening of the foregut. The foregut in this embryo is longer than in the preceding embryo. Its cranial extremity is slightly caudad to that of the craniomedian limb 
of the pericardial cavity. The ventral wall of the foregut 



EARLY DEVELOPMENT OF HEART IN MAMMALL\ 267 

coalesces with the ectoderm, indicating the pharyngeal membrane, while its dorsal wall corresponds to the cranial end of 
the notochord. The entoderm, which forms the dorsocaudal 
wall of the craniomedian limb of the pericardial cavity, is reflected from the foregut opening to the ventral wall of the craniomedian limb of the pericardial cavity, representing the so-called 
cardiac fold (repli cardiaque Tourneux). The cranial wall of 
the craniomedian limb of the pericardial cavity is formed by 
the ectoderm, which is reflected from the pharyngeal membrane 
to the proamnionic region. 

A few cell strands of angioblasts can be seen between the 
mesoderm of the splanchnopleura and the underlying ectoderm. 
The splanchnopleura is present, its convexity turned caudoventrally, in accordance with the entodermic cardiac fold. Its 
central part has begun to invaginate into the pericardial cavity, 
rising from the underlying entoderm. Between these two layers 
the angioblasts are scattered. 

Stage III 

The material on which the description of stage III is based 
consists of two embryos, one of which was cut transversely 
and the other longitudinally. 

A. This specimen was removed from the uterus of a guineapig fifteen days and fourteen hours after insemination. This 
series includes 318 sections, having a 7 /^ thickness. The plastic reconstruction of the cephalic portion of the embryo was 
made with wax plates, and the whole shield, reconstructed for 
another purpose, was used for this study. 

The neural groove extends from the cranial end to the caudal 
amnion attachment. It is wide and shallow in the caudal portion, while it is narrow and deepens toward the head fold. The 
head fold is divided into two primary vesicles; the cranial vesicle 
is wide and long, projecting laterally and cranially over the 
cranial and lateral walls of the cranial body elevation. The 
caudal vesicle is small and passes insensibly into the spinal 
portion. The anlagen of the trigeminal ganglia, as well as the 
rudiment of the otic ganglia, are to be seen. Four somites are completely segmented, besides in their cranial and caudal 
territory, a somite is in process of formation. The dorsal aortae 
and the first aortic arch are present, while the ventral aortae 
are not yet completely differentiated. In the region of their 
anlagen the angioblasts are irregularly distributed. The foregut 
extends throughout twenty-two sections, appearing first in the 
twenty-seventh section and continuing to the forty-ninth section, 
while the craniomedian limb of the pericardial cavity extends 
throughout sixteen sections, appearing in the twenty-third section and continuing to the thirty-ninth section. The cranial 
end of the head fold appears in the eighth section, on account 
of the forward progress of the head fold. This fact can be 
demonstrated in the dorsal surface of the whole reconstruction 
model, in which the prominent swelUng of the dorsal surface 
of the pericardial cavity is much more distinct than that of the 
foregoing model. The cranial extremity of the pericardial cavity 
disappears under the head fold slightly caudad to its cranial 
margin. 

In this embryo the mesoderm of the splanchnopleural projects into the pericardial cavity, forming prominent folds, 
already presented in the previous stage, but in this stage is well 
developed. These folds are converted on both sides into continuous myocardial tubes. Their dorsal surfaces approach the 
dorsal wall of the pericardial cavity, coming nearly into contact 
with it. These lateral myocardial tubes are best developed 
at the level of the hindbrain, where they are relativel}' dilated 
and contain the well-developed endothelial tubes. On tracing 
cranial ward, both myocardial tubes gradually diminish in height 
and width, until they finally disappear opposite to the foregut 
opening on its left side, while on the right side the myocardial 
tube continues into the caudad portion of the craniomedian 
limb of the pericardial cavity, converting it into the prominent 
rounded ridge of the mesoderm of the splanchnopleura. In 
this region the thickened mesoderm of the splanchnopleura, 
present as a somewhat flattened fold, is elevated above the 
underlying entoderm. In the space between these two layers 
a number of angioblasts can be seen (fig. 11 and 12). In tracing still farther cranial-ward, the relatively thin layer of mesoderm 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 



269 




Fig. 12 Dorsal view of the reconstruction of the same embryonic shield (stage 
III, A) from which figure 11 was drawn. Dorsal wall of the pericardium has 
been removed to show the pericardial cavity and myocardial tubes. At the 
caudal part of the craniomedian limb of the perciardial cavity the splanchnopleura projects into the pericardial cavity, forming the prominent fold, which is 
absent in front of the cranial extremity of the myocardial tube on the left side. 
E-D indicates plane of section of figure 11. X 100. 

Fig. 13 Dorsal view of the reconstruction of the same embryonic shield 
(stage III, A) from which figures 11 and 12 were drawn. Dorsal wall of the pericardium and of the myocardial tubes have been removed to show the underlying 
endothelial tubes and angioblasts. Angioblasts are scattered under the splanchnopleural fold at the cuadal part of the craniomedian limb of the pericardial 
cavity, almost connecting both cranial extremities of the lateral endothelial 
tubes. But angioblasts are absent in front of the cranial extremity of the lateral 
endothelial tube on the left side, where the splanchnopleura has not risen from 
the underlying entoderm. X 100. 


of the splanchnopleura remains attached to the underlying entoderm; between them no angioblasts can be seen (fig. 13). 

It can be ascertained that the formation of the mesodermal 
splanchnopleural folds occurs in loco and progresses cranial-ward, 
until both cranial extremities of the myocardial tubes will ultimately unite and communicate with each other at the craniomedian limb of the pericardial cavity. The craniomedian limb 
of the pericardial cavity increases its dimensions both in the 
ventrodorsal and in the craniocaudal directions, while the formation of the myocardial anlage in this portion remains in its 
primitive condition. In the hindbrain region the pericardial 
cavity reaches its maximum width in proportion to the myocardial and endothelial development. But it shows here a rather 
narrower space in the ventrodorsal direction, on account of the 
dorsal expansion of the myocardial tubes. In tracing still further caudalward, the pericardial cavity gradually narrows until 
it entirely disappears in the somitic region, parallel with the 
gradually diminishing splanchnopleural folds and endothelial 
tubes. 

The endothelial tubes are differentiated at great length, extending throughout nearly the whole extent of the lateral pericardial cavity. But in many places these tubes are irregularly 
interrupted, their continuity bridged by angioblast cords. The 
endothelial tubes terminate cranially opposite to the foregut 
opening on the right side and slightly caudad to it on the left 
side. In front of these terminations a number of angioblasts 
are scattered. Extending still farther ci'aniomedially, by means 
of these angioblasts, the cranial extremities of the endothelial 
tubes are connected with each other through the middle plane 
underneath the flat splanchnopleural folds of the craniomedian 
limb of the pericardial cavity (fig. 13). There is a distinct 
significance in the fact that these angioblasts are directly derived 
from the mesoderm of the splanchnopleural cells in loco. Moreover, in some other parts the angioblasts or endothelial cells 
are undoubtedly connected with the thickened and indented 
mesoderm of the splanchnopleura. Therefore, it is conceivable 
that the productive activity of these cells from the mesoderm 
of the splanchnopleura is still continued in this embryo. 



EARLY DEVELOPMENT OF HEABT IN M.IMAIALIA 271 

Stage IV 

The material on which the following description of stage IV 
is based consists of two embryos, one of which was cut transversely and the other longitudinally. 

A. This specimen was removed from the uterus of a guineapig fourteen days and eight hours after insemination. This 
series includes 566 sections, having a o ^ thickness, from the 
cranial margin of the head fold to the caudal end of the mesodermic thickening of the allantois. The plastic reconstruction 
of the cephahc portion of the embryonic body was made with 
wax plates. 

The head fold has progressed cranial- and dorsalward. Its 
cephahc extremity is represented in that of the embryonic shield. 
There are present seven pairs of mesodermic somites, the first 
and last being small and indistinctly segmented. The neural 
groove extends from the cranial end to the caudal amnion attachment. In the hindbrain region the neural groove shows 
very narrow and deep as both the neural plates approach each 
other. 

In the model it can be recognized that the craniomedian 
limb of the pericardial cavity increases its dimensions in the 
ventrodorsal direction, while its lateral and craniocaudal extent 
remains approximately unchanged in comparison with the previous stage III. The cranial extremity of this portion extends 
at its dorsal part into the mesodermic cavity of the mandibular 
region on both sides. While their outhne gradually approaches 
the horizontal plane, the caudal extremity of this portion is 
continued into both the lateral pericardial cavities, which gradually diminish in width caudalward. In the region of the second 
somite they entirely disappear. 

The formation of both the lateral myocardial tubes, which 
has been discussed in stage III, are considerably developed and 
have so far progressed cranial-ward, that their cranial portions 
have partially come into contact and been fused together. 
Through this portion the myocardial tubes communicate with 
each other. On the dorsal surface of this fused portion of the lateral myocardial tubes the myocardial \Yalls are reflected directly onto the dorsal wall of the pericardimn, thus forming 
the dorsal mesocardium on both sides. Between the lateral 
mesocardial layers there is present an irregular triangular space, 
which we purpose to designate as the intermesocardial space 
and through which the endothelial offshoots come out from the 
myocardial cavity onto the space between the mesocardial layers 
and the floor of the foregut. Its apex is directed craniaiward, 
where the lateral mesocardial layers come in contact, marking 
the cranial margin of the communicating myocardial cavity. 
Its basal portion is directed caudalward and corresponds to the 
foregut opening, by which the lateral mesocardium layers diverge from each other and continue farther caudalward along 
the lateral myocardial tubes (fig. 14). Between the abovementioned adherent cranial margin of the mesocardium and the 
foregut opening, the lateral myocardial cavities communicate 
with each other across the middle plane to the extent of eight 
sections. From this communicating myocardial cavity are sent 
out two short cranial diverticula on either side, separated by a 
septal wall in the middle plane, corresponding to the cranial 
extremities of the lateral myocradial tubes. These diverticula 
are present as the rounded myocardial horns, directed cranial-ward and separated from each other by their own inner walls. 
These inner walls are caudally converted into a wedge-shaped 
prominent ridge, which continues into the communicating portion of the myocardial cavit}^ and gradually diminishes caudalward (fig. 15). The communicating portions of the lateral mj'^ocardial tubes are directly continued into the lateral myocardial 
tubes caudolaterally on both sides and they are separated from 
each other by the foregut opening. The ventral wall of the 
communicating portion of the myocardial cavity is reflected 
onto the ventral wall of the pericardium and is recognized only 
in the caudal portion. The reflection points from the myocardium to the ventral pericardial wall are fused together at 
the cranial part, but at the caudal part the reflection points 
diverge from each other and a triangular space remains between 
them in just the same manner as can be seen in the dorsal wall. 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 



273 



This ventral triangular space is covered by the entoderm cephalad to the foregut opening, while the dorsal intermesocardial 
space is covered by the foregut floor. Ventrall}" to the communicating myocardial cavity, the pericardial cavity passes from 
side to side, because of the absence of the ventral mesocardium. 
Along the ventral mesodermic reflection the anlage of the septum 
transversum of His will be presented in the future development, 
and the mesodermic reflection may be erroneously taken for the ventral mesocardium, if a single section of this portion 
should be examined, as many workers claim the existence of 
the ventral mesocardium in mammals. 




Fig. 14 Dorsal view of the reconstruction of an embryonic shield (stage IV). 
Dorsal wall of the pericardium has been removed to show the pericardial cavity 
and myocardial tubes. The two lateral myocardial tubes are partially confluent, 
slightly cephalad to the foregut opening. The two short cranial horns of the 
myocardial tubes are directed toward the top of the page. X 100. 

Fig. 15 Dorsal view of the reconstruction of the same embryonic shield 
(stage IV) from which figure 14 was drawn. Dorsal wall of the pericardium 
and of the myocardial tubes have been removed to expose the underlying endothelial tubes. The two lateral endothelial tubes approach most closely to each other 
at the confluent myocardial portion, where independent endothelial cells are 
interposed between the two tubes. X 100. 






The endothelial tubes are well developed and are enclosed 
within the myocardial cavities on both sides. Their lumina 
are patent throughout their cranial extent, their cranial extremities terminate blindly opposite to the cranial extremities of 
the myocardial tubes, where the myocardial tubes are projected 
into the pericardial cavity, like the rounded lateral horns on 
either side, which contain the cranial extremities of the myocardial cavity. In tracing caudally, the endothelial tubes 
reduce their caUbers gradually and they are irregularly interrupted in their continuity by angioblasts. They entirely disappear in the somitic portion, where the lateral pericardial 
cavities assume a narrow and horizontal space and the splanchnopleural folds have entirely disappeared. The lateral endothelial tubes are most remarkably dilated and considerably approximated to each other at the communicating myocardial 
cavity, where the independent intermediate endothelial cells can 
be seen between the endothelial tubes. Throughout many sections in the communicating myocardial cavity, the endothelial 
tubes give off their endothelial offshoots from their dorsomedian 
aspects, coursing dorsolateralward, between the dorsal mesocardium and the floor of the foregut. These offshoots may 
be considered the future truncus arteriosus. 

Owing to the gradual transition from these dilated endothelial 
tubes into the portion of the vitelline veins caudalward, their 
demarcation cannot be pointed out on the endothelial tubes 
nor on the myocardial tubes. 

The dorsal aortae and the first aortic arch are completely 
developed, while the ventral aortae are incompletely differentiated. In their anlagen a number of angioblasts are distributed irregularly. 

Stage V 

The material on which the following description of stage V 
is based consists of two embryonic shields, which were cut trans


EARLY DEVELOPAIENT OF HEART IN MAMMALIA 275 

verselj^ The partial plastic reconstruction of the cephalic portion and the reconstruction of the whole embryonic shield, made 
for another purpose, were used for this study. 

A. This specimen was removed from the uterus of a guinea 
pig fourteen days and eight hours after insemination. The series 
includes 612 sections, having a 5 m thickness. As measured 
by age, this embryonic shield is shghtly younger than that 
discussed in stage IV. As judged by the stage of general development, it is slightly more advanced, indicated by the facts 
that eight somites are present and that the medullary groove 
is much deeper and narrower in the hindbrain region, so that 
to a great extent both neural plates are in contact; here it passes 
insensibly into the spinal region. The forebrain plate still remains wide open, projecting cranially and laterally over the 
cranial and lateral wall of the anterior body elevation. It is, 
moreover, bent considerably ventralward. In the model it can 
easily be recognized that, owing to the fact that this embryonic 
shield is considerably folded off from the yolk sac, it is in general 
thicker in the ventrodorsal diameter and narrower in the lateral 
diameter than that in stage IV. In accordance therewith, the 
cranial extremities of the lateral myocardial tubes are forming 
a more acute angle than that of the previous stage. The craniomedian hmb of the pericardial cavity increases its ventrodorsal 
and craniocaudal dimensions, while its lateral diameter diminishes on comparison with the embryo of stage IV, in proportion 
to the rounded outUne of this embryo. The craniomedian Umb 
of the pericardial cavity is elongated at its dorsal part into the 
mandibular mesoderm. In coursing caudalward, the lateral pericardial cavities graduallj^ diminish their width, and at the same 
time their outline approaches the horizontal plane as a whole. 
They disappear entirely opposite to the fourth somite. The 
direction of the lateral myocardial tubes tends to their nmning 
parallel to each other. The lateral myocardial tubes are considerably dilated at their cranial portion, where they abruptly 
become voluminous in comparison with their caudal portion. 
The transitional points of these two different portions are situated a little caudad to the foregut opening on both sides, where the myocardial tubes mark shght indentations. These 
indentations indicate the future atrioventricular constriction 

(fig. 16). 

The cranial extremities of the lateral myocardial tubes become 
more voluminous as compared with those of the previous stage 
and are projected cephalad into the pericardial cavity as large 
lateral rounded horns. They are separated from each other 
by their own inner walls, which fuse caudally into one septal 
wall and, as we trace still farther caudally, we find them converted into the prominent wedge-shaped ridge which projects 
into the communicating myocardial cavity. This ridge gradually diminishes in height caudalward until it has entirely disappeared in the middle of the communicating cavity (fig. 17). 

The dorsal wall of the fused myocardial tubes is reflected 
directly onto the dorsal wall of the pericardium, forming the 
dorsal mesocardium on both sides. These lateral mesocardial 
layers come to fusion in a region a little cephalad to the foregut 
opening, where it makes the cranial margin of the communicating 
myocardial cavity. On the dorsal surface of the fused myocardial portion an irregular intermesocardial space can be seen, 
covered by the floor of the foregut. Its apex is directed cephalad, corresponding to the point where the lateral mesocardium 
layers are fused together. Its base is directed caudad, corresponding to the foregut' opening. Its sides are formed by the 
lateral mesocardial layers. Corresponding to this intermesocardial space, the lateral myocardial tubes communicate with 
each other through the median plane throughout the extent 
of nine sections. In a similar way, the ventral wall of the fused 
myocardial tube is reflected onto the ventral wall of the pericardium, but only in its caudal portion. Between these two 
lines of the mesodermal reflection there remains a narrow space 
free from the mesoderm and covered directly by the entoderm. 
However, these lines of reflection on the ventral wall are disposed 
in a rather transverse direction and are located only for a short 
extent in the caudal part of the communicating myocardial 
tube, while on the dorsal surface the lines of reflection of the 
mesocardium are directed rather longitudinally and extend 



E.\IILY DEVELOPMENT OF HEART IN MAMMALIA 



277 



throughout the whole length of the eoniinunicating myocardinal 
tube. The communicating portion of the myocardial cavity 
terminates blindly in the cranial diverticulum cephalad, corresponding to the cranial extremities of the lateral myocardial 
tubes, while caudally it is elongated into the lateral myocardial 
tubes, which are diverged by the foregut opening, and their 
lumina are gradually diminished toward the somitic region. 




Fig. 16 Dorsal view of the reconstruction of an embryonic shield (stage V). 
Dorsal wall of the pericardium has been removed to show the pericardial cavity 
and myocardial tubes. The two lateral myocardial tubes become quite voluminous, especially at their cranial portions, which continue farther caudalward, 
gradually diminishing in size. The atrioventricular constriction is marked on 
the surface of the myocardial tubes, caudad to the foregut opening on both sides. 
The two cranial horns of the myocardial tubes become enlarged, and they are 
directed toward the top of the page. X 100. 

Fig. 17 Dorsal view of the reconstruction of the same embryonic shield 
(stage V) from which figure 16 was drawn. Dorsal wall of the pericardium and 
of the myocardial tubes have been removed to expose the underlying endothelial 
tubes, which are apparently enlarged at their cranial portions, where they most 
nearly approach each other. X 100. 




In general, the endothelial tubes are much more developed 
than those of the previous stage, since they have become deeper 
and wider. At the widest portion of the endothelial tubes, 
corresponding to the communicating myocardial cavity, the endothehal tubes approach each other so that they come nearly 
into contact. On these portions, throughout many sections, 
the endothelial tubes give off a number of endothehal offshoots 
from their dorsomedian surface into the space between the 
dorsal mesocardium and the foregut floor. From these portions they become gradually narrower, toward both the cranial 
and caudal directions. The cranial extremities of the endothelial tubes terminate blindly opposite to the cranial myocardial 
extremities, while caudally they continue into the portion of 
the vitelline veins. The endothelial tubes assume the distinctly 
narrow calibers opposite to the atrioventricular constriction. 
The endothehal tubes sprout out into innumerable tenuous 
fibrils, often forming a feltwork, which occupies the ^^■ide space 
between the myocardium and the endothehum. 

The dorsal aortae and the first aortic arch are developed, 
while the ventral aortae are not completely formed, as in their 
anlagen a number of angioblasts are scattered. 

Stage VI 

The material on which the following description of stage VI 
is based consists of one embryo, cut transversely. The plastic 
reconstruction of the cephalic portion of the embryo was made 
with wax plates. 

This specimen was removed from the uterus of a guinea pig 
fourteen days and eight hours after insemination. The series 
includes 582 sections, having a 5 // thickness. As measured by 
age, this embryonic shield is the same as that of the previous 
embryo. As reckoned by general development and special de


EARLY DEVELOPMENT OF HEART IN MAMMALIA 279 

velopment of the heart, it is considerably advanced over the 
preceding embryo. Eight well-segmented somites are present. 
The medullary groove, extending from the cranial end to the 
caudal amnion attachment, is as deep and narrow as in the previous embryo. The form of the embryonic shield is, in general, 
more rounded in comparison than with the foregoing embryo, 
as the ventrodorsal diameter of this embryo is apparently increased while its lateral diameter has remained unchanged. The 
first visceral pouch and the oral pit are developed ; in these places 
the entoderm coalesces intimately with the ectoderm. 

The reconstruction shows that at this stage of development 
the craniomedian limb of the pericardial cavity increases considerably in the craniocaudal dimension and in the ventrodorsal dimension. The craniomedian limb of the pericardial 
cavity communicates caudally with the lateral pericardial cavities. On coursing caudally, these become gradually narrower, 
until they disappear entirely opposite to the sixth somite. The 
craniomedian limb of the pericardial cavity is elongated cranially at its dorsal part into the mandibular portion, lying under 
the foregut floor. The caudal half of the ventral surface of 
the craniomedian limb of the pericardial cavity is covered by 
the yolk sac, while its cranial half and all other surfaces are 
covered by the amnion. 

In this stage the fused portion of the lateral myocardial tubes 
increases remarkably throughout its craniocaudal extent. In 
accordance therewith, the cranial bilateral myocardial horns, 
which correspond to the cranial extremities of the lateral myocardial tubes, and predominate in the craniomedian limb of the 
pericardial cavity in the previous stage, apparently diminish 
their dimensions in this embryo, and show only their rudiments. 
They assume only short and wide bilateral processes, divided 
by a shallow and wide intervening groove. In the previous 
stage this groove was present as a narrow and deep sulcus. 
Subsequently, the inner walls of these horns diverged markedly 
from each other (fig. 18). The wedge-shaped ridge which, in 
the previous stage, projected into the communicating myocardial cavity at its middle cranial wall, as a caudal continuation of the converted septum walls of the cranial bilateral myocardial horns, is considerably retired cranial-ward in this embryo. 
Therefore, the cranial wall of the communicating myocardial 
cavity approaches in such a manner toward the cranial wall 
of the pericardium as to come nearly into contact with it and, 
simultaneously, the communicating myocardial cavity is elongated cranial-ward. The fused portion of the myocardial tube 




Fig. 18 Dorsal view of the reconstruction of an embryonic shield (stage VI). 
Dorsal wall of the pericardium has been removed to show the pericardial cavity 
and myocardial tubes. The confluent portion of the two lateral myocardial tubes 
is considerably elongated in the craniocaudal direction. The two cranial horns of 
the myocardial tubes diminish to short rudiments, as their septal wall retires 
cranial-ward. They are directed toward the top of the page. X 100. 



becomes distinctly narrower and thicker in comparison with 
that of the previous stage. In two of the same magnified 
models ( X300) the widest lateral diameter of this portion is calculated as 14.5 cm. in this embryo, instead of 19 cm. of the 
previous embryo, while the ventrodorsal diameter of this portion 
presents 5.1 cm. in this embryo and 3 cm. in the former embryo. 
The fused myocardial tube is reflected directly onto the dorsal wall of the pericardium, ?,nd thus forms the dorsal mesocardium on both sides. Between the lateral mesocardial layers 
there can be seen alongrectangularintermesocardial space; its 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 281 

plane is approximately parallel with the horizontal. Its cranial 
margin is formed by the fused portion of the mesocardial 
layers in the middle line; its caudal margin corresponds to the 
foregut opening, while both lateral margins are represented by 
the lateral mesocardial layers, which continue farther caudalward, diverted by the foregut opening. In accordance with 
this mesocardial space, both myocardial tubes communicate 
freely with each other through the median plane, and thus form 




19 

Fig. 19 Dorsal view of the reconstruction of the same embryonic shield 
(stage VI) from which figure 18 was drawn. Dorsal wall of the pericardium and 
of the myocardial tubes have been removed to expose the endothelial tubes. 
The two lateral endothelial tubes have fused and communicate with each other 
at a middle third of the ventricle, where they most closely approach each other 
in figure 17. The lateral endothelial tubes apparently diminish their size opposite to the atrioventricular constriction. X 100. 

the craniomedian limb of the myocardial cavity. This limb of 
the myocardial cavity is bifurcated caudally into the lateral 
myocardial tubes, which are diverged from each other by the 
foregut opening and in which the vitlline veins are enclosed, 
leading cranially into the craniomedian limb of the myocardial 
cavity (fig. 19). In brief, the myocardial anlage presents cranially two short rudimentary horns, which terminate blindly as 
the cranial myocardial extremities, while caudally there are two 
lateral myocardial prolongations, into which the vitelUne veins 
enter. Between these four extremities the myocardial wall is 
relatively considerably expanded dorsal, lateral, ventral-ward and contains the widest portions of the endothehal tubes, which are 
united in the communicating cavity. This region corresponds 
to the future ventricle region. On the midsagittal hne of the 
ventral surface of this fused myocardial portion a shallow longitudinal groove can be seen. 

On the ventral aspect of the fused myocardial portion the 
ventral myocardial layer is reflected onto the ventral wall of 
the pericardium, but this is limited to a short length, extending 
only to the caudal part of this portion. 

The transition from the cranial expanding ventricle to the 
caudal myocardial prolongations is indicated by an annular constriction, which is produced by the infolding of the whole myocardial wall, a little deeper on the right side than on the left. 
This indicates the atrioventricular constriction and is situated 
at the level slightly cephalad to the foregut opening on both 
sides. 

Proceeding caudally from this constriction, the lateral myocardial tubes reduce their calibers abruptly and diverge from 
each other. On the surface of these caudal prolongations of 
the myocardial tubes there are present indefinite, shallow indentations at the level shghtly caudad to the foregut opening, 
and these constrictions have been regarded as the future sinoatrial construction. 

The lateral endothelial tubes are partially fused and their 
lumina communicate with each other, for their inner walls have 
been absorbed throughout seven sections. This portion is situated in the middle third of the bulging ventricle anlage, where 
in the previous stage both endothelial tubes were closely approximated and presented the greatest dilation and where in. 
this embryo also the endothehal tube is greatly expanded. 

From this fused portion of the endothelial tube the two cranial horns and two caudal prolongations are given off. The 
bilateral cranial horns are short and gradually diminish in size 
cranial-ward, until they terminate in a pointed apex, opposite to the cranial extremities of the myocardial horns. From 
the dorsomedian part of these cranial endothelial horns a number of endothelial branches are given off. These endothehal 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 283 

branches are connected with the ventral aortae through the 
intermesocardial space. 

Bilateral caudal prolongations are given off on both sides 
from the caudal aspect of the fused endothelial tube. Continuing caudalward, both endothelial tubes gradually diminish 
their cahbers, until the lumina have entirely disappeared at the 
atrioventricular constriction. Still further caudalward from this 
constriction, again they begin to dilate their calibers gradually 
and continue into the endothelial vitelline veins without any 
indication at their transitional point. At the atrioventricular 
constriction the endothelial tubes closely approach the infolding of the myocardial wall, while in the ventricle the intervening space between the myocardium and endothelium is 
relatively wide. 

From the caudal aspect of the fused endothelial tube another 
intermediate endothelial branch is given off caudall3^ This 
branch is situated between the endothehal prolongations and 
terminates at the atrioventricular constriction. 

The dorsal aortae and the first aortic arch are developed, 
while, in many sections, the ventral aortae are interrupted by 
angioblasts. 

Stage VII 

The material on which the following description of stage VII 
is based consists of one embiyonic shield, which is cut transversely. The plastic reconstruction of the cephalic portion of 
the embryo was made from wax plates. 

The specimen was removed from the uterus of a guinea-pig 
fourteen days and eleven hours after insemination. The series 
includes 418 sections, having a 7 ^ thickness, from the cephalic 
end of the head fold to the end of the mesodermic thickening 
of the allantois. Xine pairs of well-segmented somites were 
found, each somite showing a thick wall and enclosing a uniform 
cavity with a compact arrangement of cells, except two caudal 
somites, which contained no cavity nor presented the regular 
arrangement of the cells. 



The medullary canal is closed from the second somite to the 
last, but elsewhere the medullar}^ plates remain open. The 
notochordal plate is separated from the entoderm throughout 
from the second somite to the last, but elsewhere it is still connected with the entoderm. The first and second visceral pouches 
are developed, in which the ectoderm and entoderm have tightlycoalesced. The oral pit is formed and the pharyngeal membrane 
becomes quite thin. 

In this embryo the pericardial cavity is closed in all directions, 
forming a sac, except at the dorsal part of its caudal extremities, 
where the pleuropericardial passages are opened on each side. 
These passages are represented by the narrow coelomic space, 
which continues farther caudalward into the peritoneal cavity 
and they are situated dorsomedian to the myocardial coat of 
the vitelline veins. From the dorsal part of the cranial extremity 
of the pericardial cavity a slit-like space is elongated into the 
mandibular region. 

The surface of the pericardium is covered by the yolk sac 
ventrally, while laterally, cranially, and dorsally it is covered by 
the amnion. The pericardial cavity shows a wide space around 
the myocardium. Between the cranial extremity of the myocardium and the cranial wall of the pericardium there remains 
a wider interval than that of the embryo of stage VI. 

The myocardium presents cranially an undivided cranial extremity, which expands considerably in all directions and assumes 
a sac form, while caudally this myocardial sac is bifurcated into 
two rather slender myocardial prolongations, in which the endothelial vitelHne veins are enclosed on both sides. The transition 
from the cranial myocardial sac to the bilateral myocardial 
tubes is indicated by the deep atrioventricular construction, 
at the level slightly cephalad to the foregut opening. This 
constriction is produced by the infolding of the whole myocardial wall and shows apparently deeper on the right sidfe than 
on the left. On the ventral surface of the ventricle there can 
be seen a shallow groove in the midsagittal line at its caudal 
half, and in accordance therewith the whole thickness of the 
myocardial wall is slightly infolded into the myocardial cavity. 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 



285 



This superficial groove and the infolding of the myocardial wall 
are located in the caudal half of the ventricle, for they gradually disappear toward its cranial extremity, which is of conical form. The ventral myocardial layer is reflected onto the 
ventral wall of the pericardium, but is confined to the atrial 
region to a short length opposite to the foregut opening. 




D.M. 



— AV.C. 



Fig. 20 Dorsal view of the reconstruction of an embryonic shield (stage VII). 
Dorsal wall of the pericardium has been removed to show the pericardial cavity 
and the myocardial tubes. The myocardium shows cranially a single sac-formed 
ventricle, which bifurcates into two slender myocardial tubes caudalward. The 
transition between them is marked by the atrioventricular constriction cephalad 
to the foregut opening. A single cranial extremity of the ventricle is directed 
toward the top of the page. X 100. 

On the dorsal aspect the myocardial wall is reflected onto 
the dorsal wall of the pericardium and forms the dorsal mesocardium on both sides. The cranial extremity of the dorsal 
mesocardial attachment corresponds to a middle third of the 
ventricle, and from this point it continues farther caudalward. 
Consequently, the cranial half of the ventricle is free from the 
mesocardium (fig. 20). Between the lateral mesocardial layers 



286 TAXZO YOSHINAGA 

there can be seen an irregular triangular intermesocardial space. 
Its plane is directed caudodorsalward on account of the abrupt 
dorsal expansion of the ventricle. Its apex is, therefore, situated caudoventrally and is formed by the lateral mesocardial 
layers, approaching contact, opposite to the atrioventricular 
constriction, while its basal portion is directed craniodorsally 
and corresponds to the portion where the lateral mesocardial 
layers come to fusion and mark their cranial extremities. 

On the caudal surface of the ventricle the right half assumes 
an apparently wider dimension than the left. This is attributed partly to the exceeding expansion of the myocardial wall 
in the laterocaudal direction on the right half and partly to the 
deeper infolding of the atrioventricular constriction on the right 
side. On this bulging portion of the caudal extremity of the 
ventricle at the right half the caudal extremity of the right 
ventricle will be developed, and this is shown distinctly in the 
next stage. 

The lateral myocardial tubes of the atrium are diverged from 
each other by the foregut opening. The right atrium is practically beginning just opposite to the foregut opening, while 
the left one begins slightly cephalad to it, for in comparison 
with the right side, the left atrioventricular constriction is shallower and situated slightly cephalad. 

On .the myocardial tubes of the atrial portion indefinite indentations, indicated as the sino-atrial constriction, can be recognized. These are between the atrioventricular constriction 
and the level of the pleuropericardial passages. This is especially noticeable on the left side. 

The lateral endothelial tubes are fused together throughout 
the cranial two-thirds of the ventricle. Its cranial extremity 
terminates as a single conical apex opposite to the cranial extremity of the myocardial ventricle. At this fused portion the 
endothelial cavities communicate with each other and show considerable dilation. In tracing caudalward from this united portion, the tubes are separated from each other, even though they 
appear to approach each other. At the atrioventricular constriction the endothelial tubes present their smallest size and 



EARLY DEVELOPAIENT OF HEART IN MAMMALIA 



287 



simultaneously they approach closely to each other. Proceeding 
still farther caudally from this portion, they are diverged from 
each other by the foregut opening and again assume a gradual 
enlargement of their calibers. At the caudal part of the ventricle, where the endothelial tubes are separated, they present 
an asymmetrical size, for the right one is extraordinarily developed and expended in the lateral and caudal directions, forming a curvature whose convexity is turned laterocaudalward. 
At this endothelial portion the endothelial tube is elongated 
vertically dorsalward and comes out from the myocardial cavity 
onto the foregut floor through the intermesocardial space. The 
cranial part of this endothelial elongation is bifurcated into two 
later branches which connect it cranially with the corresponding ventral aortae. The asymmetrical development of the endotheUal ventricle corresponds to the myocardial asymmetry in the 
ventricle, which has been mentioned above. In this part of 
the ventrical the most important change will be noted in the 
next stage, here developing, namely, the right limb of the ventricle. And this change is initiated in this embryo as a considerable asymmetrical expansion of the caudal extremity of the 
ventricle on the right side. 



Fig. 21 Dorsal view of the reconstruction of the same embryonic shield 
(stage VII) from which figure 20 was drawn. Dorsal walls of the pericardium 
and of the myocardium have been removed to expose the endothelial tubes. The 
two lateral endothelial tubes have fused and communicate with each other 
throughout the cranial two-thirds of the ventricle. The ventricular endothelial 
tubes show a distinct asymmetrj^, due to the extraordinary enlargement of the 
right side, regardless of the fused or non-fused portion. The endothelial tube is 
elongated dorsalward from the dorsal surface of the enlarged right endothelial 
tube, passing through the intermesocardial space. This endothelial elongation 
is bifurcated into the two lateral branches, which are continuous into the ventral 
aortae. X 100. 



At the ventricle the endothelial tubes are separated from 
the myocardial wall by a wide intervening space, but they 
gradually approach each other in the direction of the atrioventricular constriction caudalward, as in the caudal part of 
the atrium no more intervening space can be pointed out between 
the myocardial wall and the endothelium (fig. 21). 

The transition from the atrial endothelial tubes into the sinus 
portions can be pointed out by the abrupt decrease of the endothelial caliber on the left side, corresponding to the relatively 
distinct sino-atrial constriction of the myocardium. But on the 
right side the atrial endothelial tube continues farther caudalward without any demarcation, in accordance with the relatively indistinct myocardial constriction. 

On both sides the ductus cuvieri can be seen opening into 
the sinus venosus. The dorsal aortae and first aortic arch are 
well developed, while in many places the ventral aortae still 
retain the plexus form. 

Stage VIII 

The material on which the following description of stage VIII 
is based consists of one embryonic shield, which is cut transversely. The plastic reconstruction of the cephalic portion of 
the embryo was made; the reconstruction of the whole embryonic 
shield, made for another purpose, was used for this study. 

This embryo was removed from the uterus of a guinea pig 
fourteen days and twelve hours after insemination. The series 
includes 408 sections, having a 10 m thickness, from the cephahc end of the head fold to near the caudal extremity of the 
allantoic mesodermic thickening. Nine pairs of the well-seg


EARLY DEVELOPMENT OF HEART IN MAMMALIA 289 

mented somites are present and the tenth is in process of formation; each somite shows the thick wall and encloses a uniform cavity. The neural canal is closed from the region of the 
hindbrain to the region of the last somite, though in the foreand midbrain region it still remains open. The cranial flexure 
is shown in the region of the midbrain and that of the forebrain 
is bent downward and forward, bringing it to a plane parallel 
with the long axis of the hindbrain. The foregut is closed to 
the first somitic region. The first visceral pouch is found in 
the region of the midbrain with the entoderm and ectoderm 
coalesced, while the second visceral pouch is in process of formation also in the region of the hindbrain, in which region 
between the entoderm and ectoderm a thinner layer of the mesoderm than elsewhere is found interposes. The oral pit is well 
formed, the pharyngeal membrane is present as a thin single 
layer of cells. On the ventral surface of the model the edge 
of the foregut opening is elevated by two prominent Umbs, 
which on each side are confluent into an extensive ventral bulging 
cranially to the foregut opening. In position and direction this 
corresponds to the pericardial cavity, containing the voluminous 
heart. 

The pericardial sac is closed except at the dorsomedian part 
of its caudal extremity, where the pleuropericardial passages 
are found. In the region of the sinus venosus each one of the 
bilateral pericardial cavities is divided into a median and a 
lateral part by the myocardial fold and in the region of the 
pleuropericardial passages the lateral parts of the bilateral pericardial cavities terminate blindly caudalward, so that only their 
median portions are continued caudally into the peritoneal cavity. Accordingly, these passages are represented merely by narrow, crescentic coelomic spaces, dorsomedian to the viteUine veins, 
proceeding caudally and mesially, crossing with the vitelline 
veins, which run cranially and mesially. 

On account of the considerable enlargement of the muscular 
heart, the pericardial space is, in general, proportionately reduced, especially in the well-developed ventricular portion a 
simple narrow space surrounds the muscular sac of the ventricle. 
 

In the region of the atria and the sinus venosus a relatively wide 
space intervenes between the rather flat muscular tubes and the 
pericardial wall (fig. 22). There are present two distinct constrictions on the tubular muscular heart, infolding the whole 
thickness of the myocardial wall, one of which represents the 
atrioventricular constriction and the other the sino-atrial constriction. 

The atrioventricular constriction is present asymmetrically 
on both sides; on the right side it is marked more deeply and 
situated sUghtly caudad, while on the left side it is shallower 
and lies a little cephalad. Therefore, this constriction forms 
an obhque angle with the long axis of the muscular heart. 

On the contrary, the sino-atrial constriction is marked more 
deeply on the left side and is situated slightly cephalad to the 
foregut opening, while on the right side it is less deeply constricted and is situated slightly caudad, just opposite to the 
foregut opening. 

The ventricle can be divided into two lateral limbs by a ventral and a dorsal longitudinal sulcus. On the dorsal surface 
it is marked along the attachment line of the dorsal mesocardium 
and terminates caudally opposite to the right margin of the 
atrioventricular constriction, while its cranial extremity gradually disappears at the portion where the bulbus cordis is 
differentiated from the dorsal wall of the ventricle. On the 
ventral surface the longitudinal sulcus extends to a caudal third 
of the ventricle. 

At the caudal part of the ventricle, for a short length, both 
lateral limbs are divided into two completely independent cavities by the septal wall. The caudal extremity of the right 
ventricle is shown as the conical process, projecting caudolaterally and terminating blindly, while at the caudal extremity 
of the left ventricle the atrioventricular canal opens, which 
is formed by the infolding of the muscular wall, corresponding 
to the atrioventricular constriction. 

The septal wall between the two lateral limbs at the caudal 
part of the ventricle is farther continued cranial-ward and is 
converted into wedge-shaped prominent ridges at the inner surface of the ventral and dorsal wall of the ventricle, in relation 
with longitudinal sulci on the external surface. These ridges 
have gradually disappeared within a caudal third of the ventricle (fig. 23). These prominent ridges show their anlage only on the ventral wall of the ventricle, near its caudal end, as seen 
in the previous stage. Consequently, there can be but little 
doubt that these folds cannot be regarded as the remnants of 
the primitive cardiac septum. 





Fig. 22 Dorsal view of the reconstruction of an embryonic shield (stage 
VIII). Dorsal wall of the pericardium has been removed to show the pericardial cavity and myocardium, which is subdivided into several individual portions (bulbus cordis, ventricle, atrium, sinus venosus, etc.) by the distinct bulboventricular atrioventricular, sino-atrial constrictions. X 100. 







The bulbus cordis is differentiated from the dorsal wall of 
the right ventricle near its cranial end, bulging out its wall cranial-dorsal and laterally. Its ventral wall is distinctly separated from the dorsal wall of the right ventricle at its cranial 
portion, projecting cranial-ward as an independent muscular sac, 
while in its caudal portion there can be noted no distinct demarcation between the wall of the bulbus cordis and that of 
the right ventricle. At the left and cranial sides of the bulboventricular junction, a deep external furrow can be seen, accompanied by a consequent infolding of the muscular w^all. 
On the right side the bulboventricular furrow is indefinitely 
marked only its cranial part, while in its caudal part it has 
disappeared entirely and insensibly continues into the dorsal 
wall of the right ventricle. On this account the bend of the 
heart tube at the bulboventricular junction is effected toward 
the right side, turning its concavity to the left side, beneath 
the left layer of the dorsal mesocardium. 

On the dorsal surface of the bulbus cordis there can be seen 
a triangular intermesocardial space, its plane is directed dorsally 
and slightly to the left. Its apex is directed caudally and at 
a lower level, where the demarcation between the bulbus cordis 
and ventricle wall show indefinitely. Its base is situated cranially and at a higher level, where the wall of the bulbus cordis 
is distinctly demarcated from that of the ventricle, and there 
marks the cranial termination of the dorsal mesocardium. Both 
at the apical and basal portions the lateral mesocardial layers 
come to fusion. 

A single myocardial tube of the atria begins at the atrioventricular constriction cranially and continues into the sinus 
venosus caudally, demarcated by the sino-atrial constriction. 
This muscular tube shows a marked asymmetry on both sides, 
for the left side, being decidedly expanded in all directions in 
comparison with the right side, just contrary to the ventricle, 
in which the right side is apparently more voluminous than 
the left side and bulges considerably caudolaterally. Moreover, this opposing asymmetry must be attributed partly to the 
normal obliciue direction of the atrioventricular constriction. 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 



293 



In consequence of this asymmetrical relation, the atrial tube 
forms a typical curvature with the ventricular limb at the atrioventricular junction, so that its convexity is directed toward 
the left side in the horizontal plane and ventralward in the 
vertical plane. 




Fig. 23 Dorsal view of the reconstruction of the same embryonic shield 
(stage VIII) from which figure 22 was drawn. Dorsal wall of the pericardium 
and of the myocardium have been removed to expose the endothelium, which is 
subdivided into several individual portions in conformity with the myocardial 
subdivision. The two lateral endothelial tubes have fused and now communicate with each other throughout a middle third of the ventricle, but elsewhere 
they are separated. X 100. 



The cranial extremity of the atria opens into the left ventricle through the atrioventricular canal, which is situated on the 
left side from the midsagittal line, for the right atrioventricular 
constriction is apparently more deeply infolded and proportionately the prominent ridge at the inner wall is strongly projected into the canal on the right side. The caudal extremity 
of the atria is continued into the sinus venosus, which diverge 
from each other into the two lateral myocardial tubes in relation with the foregut opening. The demarcation of these different portions is indicated by the sino-atrial constriction, of 
which on the left side the myocardial wall is more deeply infolded than on the right side. 

The two lateral layers of the dorsal mesocardium are fused 
together from a middle third of the ventricle to the cranial 
part of the atria; they are in close contiguity at the atrioventricular constriction, in which region the dorsal mesocardium is 
beginning to disappear in an embryo slightly older than that 
of this stage. But at the portion of the bulbus cordis and the 
caudal part of the atria, the layers of the dorsal mesocardium 
have not come in contact. The arterial opening is disposed 
nearly vertically, but slightly to the left side, through which 
the endothelial tube comes from the myocardial cavity, while 
the venous opening is disposed dorsocaudally and assumes an 
irregular triangular space. Its base is situated caudall}^ and 
\entrally opposite to the foregut opening, as a result of which 
the layers of the dorsal mesocardium divert together with the 
corresponding muscular tubes, which continue caudally. Its 
apex of the venous opening is directed cranial-ward and Ues at 
the higher horizontal level. Through this intermesocardial 
space the enclosed endothelial tube can be seen. 

On the ventral surface the myocardium is reflected onto 
the ventral wall of the pericardium at the sinus venosus. Here 
it can be observed that the mesodermal cells have proliferated 
to form an appreciable thickening around the endothelial tubes, 
indicating the future septum transversum. 

The lateral endothelial tubes are fused in the middle third 
of the ventricle to the extent of fifteen sections. In this part 
they communicate with each other and the endothelial cavity 
is considerably dilated (fig. 23). This craniomedian part of 
the endothelial tube is bifurcated into the two cranial horns and 
two caudal prolongations. The cranial horns extend symmetrically from the cranial wall for a short distance on both sides 



EARLY DEVELOPMENT OF HEART IN MAMMALL\ 295 

and terminate blindly opposite to the cranial myocardial extremity of the ventricle, which is of conical form. 

The right caudal prolongation is given off from the right 
side of its caudal wall and terminates as a short conical projection; its apical terminus is directed opposite to the caudal extremity of the right ventricle, conforming with it. The left 
caudal prolongation is given off from the left side of its caudal 
wall and caudally connects with the endothelial tubes of the 
atrium. These separate from each other and continue farther 
caudalward. From the origin of this left caudal prolongation 
caudally to the atrioventricular canal, that is, in the caudal 
part of the left ventricle, the two endothehal tubes show their 
smallest size and are very close in contact in some places, while 
in others the}- separate into two quite independent tubes with 
complete walls. Opposite to the atrioventricular canal the two 
endothelial tubes above mentioned begin definitely to separate 
into two lateral atrial endothehal tubes, increasing their calibers gradually caudalw^ard, while they lie approximately in a 
parallel direction. 

The caudal extremities of the atrial endothelial tubes continue 
immediately into the endothelial tubes of the sinus venosus 
and then into those of the vitelline veins. The transition from 
the atrial endothelial tubes into those of the sinus venosus 
is marked by a sudden diminution in their diameter together 
with the general decrease of their cahbers and the abrupt lateral divergence of their course, due to the intervention of the 
foregut opening. These transitional points correspond to the 
external groove of the sino-atrial constriction. 

In the ventricle the endothehal tubes are separated from the 
myocardial wall by a wide intervening space. In the atrium 
the intervening space becomes considerably narrower, and 
finally in the sinus venosus the endothehal tubes are enclosed 
intimately by their own independent myocardial wall, so that 
no appreciable space can be seen. 

The endothelial tube of the bulbus cordis extends from the 
dorsal surface of the right ventricular endothelium as its continuous prolongation. This endothelial tube proceeds at first dorsocranially and then slightly toward the left side. This is 
enclosed by the corresponding myocardial wall of the bulbus 
cordis, which is closed cranially, but caudally opens and communicates with the ventricular cavity, as already mentioned. 
The right ventricular endothelium, which gives off the endothelial 
tube of the bulbus cordis, is fused together with the left one, 
but the left part of the fused ventricular endothelium participates in no way with the bulbus cordis. 

The endothelium of the truncus arteriosus continues farther 
dorsally and slightly toward the left side from the bulbus cordis 
and passes through the above-mentioned arterial opening, and 
then bifurcates into lateral symmetrical branches, which are 
located between the foregut floor and the lateral dorsal mesocardial layers and continue farther cranially into the ventral 
aortae. 

The first aortic arch and the ventral aortae are completely 
formed and the dorsal aortae are considerably elongated caudalward. 

SUMMARY AND CONCLUSION 

In our observations the first sign of the formation of angioblasts is shown in stage I, embryo A, in which neither the head 
fold nor the anlage of the pericardial cavity has yet appeared. 

On the ventral surface of the mesoderm of the splanchnopleura of the cranial portion, cell bands first begin to separate, 
which separation is more advanced in the embryos B and C. 
These cell bands are regarded as angioblasts and they are frequently found to adhere to the indented and loosened mesoderm 
of the splanchnopleura by broader or narrower protoplasmic 
bridges. It has frequently been pointed -out that mitotic figures 
are found in the mesoderm of the splanchnopleura in the neighborhood of angioblasts. Furthermore, in many cases where the 
angioblasts are in close contact with the mesoderm of the splanchnopleura, it is impossible to discriminate the angioblasts from 
the mesodermal cells of the splanchnopleura, as concerns their 
sizes, forms, staining reaction, and the form of the nuclei, while 
a great difference can readily be recognized between the angio


EARLY DEVELOPMENT OF HEART IN MAMMALIA 297 

blasts and the adjacent entodermal cells. These findings show 
that in the genetic origin the angioblasts for the future endocardium are derived directly from the mesoderm of the splanchnopleura. The origin of the angioblasts from the mesodermal cells 
continues until a later stage, in which the greater part of the 
endotheUal tubes are already differentiated from the angioblasts 
in the anterior portion of the embryo, but the origin of the angioblasts can be recognized in the posterior part of the embryo, 
as is shown in the embryo of stage III. 

In their well-known work on pericardial development, Strahl 
and Carius found it impossible to decide whether the embryonic coelom in the guinea pig appears at first in the region of 
the heart anlage, proceeding forward into the pericephahc mesoderm, or whether it begins first in the pericephalic mesoderm 
and then spreads out caudall}^ They speak as follows: ''Doch 
konnen wir augenblicklich eine ganz sichere Entscheidung nicht 
geben." The cause of this ambiguity is that they began their 
investigation of the origin of the intra-embryonic coelom at too 
late a stage. 

For the dog, Bonnet states that, concerning the origin of the 
intra-embryonic coelomic space, that the lateral pleuropericardial cavities, having already distinctly appeared, anticipate the 
formation of the pericephahc space. To quote Bonnet directly: 
''Eine ebensolche Spaltung des Mesoderms fiihrt in VHIo 
gleichzeitig im Bereiche des Herzwulstes zur Bildung der PleuroPericardialhohle. Ihre zuerst paarig angelegten spalten vergrossern sich, vereinigen sich nach vorne und bilden so ein nachhinten offenes Hufeisen; den pericephalen und lateralen Teil der 
Pleuro-Pericardialhohle. ' ' 

In our specimens, embryo C, stage I, show^s the discontinuous 
formation of the intra-embryonic coelomic spaces in the cranial 
portion of the embryonic shield, as also in the pericephalic 
mesoderm, these spaces beginning as multiple foci. But they 
are primarily absent in the middle portion of the pericephalic 
mesoderm. 

In stage II, in which the head fold of the embryo begins 
to separate from the surrounding blastoderm and the foregut has just begun to develop, the intra-embryonic coeloniic space 
spreads out cranially into the pericephaUc mesoderm, cleaving 
the mesodermal layer in such a way, that the lateral primitive 
pericardial cavities communicate with each other. In a just 
sHghtly younger embryonic shield than this, each lateral pericardial cavity has progressed cranially into the pericephalic mesoderm, showing in this place a sht-hke space, which however, 
is divided by a thin mesodermic bridge in the middle line. 

The pericardial cavity, therefore, commences simultaneously 
in the multiple foci, separating irregularly by mesodermal 
bridges throughout the lateral plate in the cranial portion of the 
embryonic shield and in the pericephalic mesoderm. These multiple coelemic spaces become confluent to form a single pericardial cavity, having an inverted-U shape, when the mesodermal bridges at the middle line of the pericephalic mesoderm 
have ultimately disappeared and, in consequence, at this time 
the bilateral pericardial cavities, already widely confluent, communicate from side to side (stage II, A). In this embryonic 
shield relatively wide endothelial tubes are differentiated only 
in the region of the hindbrain plate, where the pericardial cavity 
is wide open and the mesoderm of the splanchnopleura is thickened, projecting into the pericardial cavity as a prominent fold. 
In the pericephahc portion of the mesoderm, however, the pericardial cavity is seen merely as a lineal cleavage. Here a few 
angioblasts are scattered between the slightly thicker mesoderm 
of the splanchnopleura and the underlying entoderm. 

In stage III the bilateral myocardial folds become quite prominent, so that in the region of the hindbrain plate they have 
been almost converted into the myocardial tubes, and enclose 
the well-developed endothelial tubes. 

The formation of these myocardial folds has progressed cranial-ward opposite to the foregut opening on the left side. At the 
same time the endothelial tube becomes gradually thinner cranial-ward and terminates slightly caudad to the foregut opening. 
On the right side the formation of the myocardial folds proceeds 
still farther cranially into the caudal part of the craniomedian 
Umb of the pericardial cavity, where the thicker mesoderm of 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 299 

the splanchnopleura is raised from the underl3'ing entoderm and 
in the space between them a number of angioblasts are 
distributed. On the right side the endothehal tube terminates 
cranially just opposite to the foregut opening. In front of the 
cranial termination of the lateral endothehal tubes a number of 
angioblasts are scattered, so that the cranial extremities of the 
endothehal tubes are nearly connected with each other through 
these angioblasts. 

The cranial extremities of the lateral myocardial folds have 
not yet come to complete confluence, as the mesoderm of the 
splanchnopleura, shghtly cephalad to the cranial extremity of 
the left myocardial fold, remains still in loose contact with the 
underlying entoderm. If this portion of the mesoderm of the 
splanchnopleura were completely raised from the underlying 
entoderm, forming the myocardial fold, then the myocardial 
anlagen would come into confluence. 

The most prevalent opinions with regard to the mode of 
the formation of the unilateral myocardial heart anlage from 
the bilateral myocardial tubes agree that, as above described, 
the bilateral myocardial tubes, at first independently, come to 
actual fusion with each other, and then the septal wall between 
them is absorbed secondarily, thus forming a single myocardial 
cavity. Our specimens show that the formation of the myocardial folds does not occur synchronously throughout the pericardial cavity, as in the region of the hindbrain plate they first 
appeared and developed considerably, while in the craniomedian 
limb of the pericardial cavity the formation of the myocardial 
folds was just starting and rising slightly from the underlying 
entoderm. However, the communication of the lateral myocardial tubes is accomplished when the formation of the myocardial folds is completed in the craniomedian limb of the pericardial cavity, in which region the formation of these folds 
occurs last. For this reason, the cranial prolongation of the 
lateral myocardial tubes has not been brought about by the 
direct extension of the first part of the myocardial tubes, but by 
the continuous progressive differentiation into the craniomedian 
limb of the pericardial cavity. Therefore, the confluence of the myocardial tubes into a single myocardial cavity is not' 
accomplished by the actual fusion of the bilateral myocardial 
tubes, followed by absorption of the septal walls. 

In stage IV the formation of the myocardial fold is completely 
accomplished in the craniomedian limb of the pericardial cavity, 
elevating the mesoderm of the splanchnopleura from the underlying entoderm and projecting into the pericardial cavity. Both 
lateral myocardial tubes communicate with each other in this 
region. 

Both the myocardial tubes are quite voluminous. They dilate in all directions, especially in their cranial portions, and they 
gradually reduce their dimensions caudal ward. On the surface 
of these transitions no distinct demarcation can be noted. 

The two lateral endothelial tubes are well developed, lying 
side by side together in the portion of the confluent myocardial 
cavity, and here independent, intermediate endothelial cells are 
scattered between the lateral endothelial tubes. 

In stage V the cranial portions of the two lateral myocardial 
tubes show^ much more dilatation and elongation, as their extent 
from the foregut opening to the cranial extremities considerably 
increases. But their confluent part still remains short. The 
transition from the cranial dilated myocardial tubes to their 
caudal slender portions is marked by a distinct annular atrioventricular constriction, which appears clearly first in this 
embryo. 

A number of workers declare that the lateral myocardial 
tubes are subdivided into many individual portions by the demarcations prior to the fusion of the lateral myocardial tubes; in 
the chick, Duval, '99; in the cat, IMartin, '02; in the rabbit, 
KoUiker, '84. Kolliker states, ''Ein Herz aus diesem Stadium 
ist sehr verschieden von dem primitiven Herzen eines Hlihnerembryo, was einfach darin begrundet ist, dass, wie bemerkt, 
bei Saugethieren schon vor der Verschmelzung der beiden Herzhiilften die drei Herzabschnitte angelegt sind." 

^lollier declares in similar language: 'Mn den beiden Herzrohren ist aber kurz vor ihrer Vereinigung schon eine Gliederung 
bemerkbar." 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 301 

In our specimens this embryo shows first the atrioventricular 
constriction, even though in the previous embryo two myocardial tubes were already confluent into a single myocardial 
cavity. 

The two lateral endotheUal tubes are considerably developed, 
especially in the confluent portion of the myocardial cavity, 
where they approach each other almost to contact. From the 
atrioventricular constriction caudalward the two endothelial 
tubes assume an abruptly narrow caliber, thus distinguishing 
the larger cranial ventricular from the smaller caudal atrial portion. In tracing farther caudally, no demarcation can be detected on them. 

In stage VI the confluent myocardial cavity increases especially in the craniocaudal extent. The cause of this may be 
attributed partly to the retirement of the wedge-shaped septal 
ridge cranial-ward, which had projected into the confluent myocardial cavity, as the conversion of the inner walls of the cranial 
lateral mj^ocardial extremities, and partly to the active backward 
progress of the foregut opening. Approximately, in the whole 
extent of the ventricle and in the cranial part of the atria the 
two lateral mj'ocardial cavities communicate with each other 
and these two myocardial portions are demarcated sharply by 
the atrioventricular constriction. In the middle third of the 
ventricle the two lateral endothelial tubes are actually fused 
and communicate with each other for a short distance. At 
the atrioventricular constriction the calibers of the two endothelial tubes are considerably reduced and, as we proceed still 
farther caudally, they are again gradually increased in diameter. 

In an embryo of Perameles nasuta having fifteen to sixteen 
somites, ]Miss Parker pointed out that the fusion of the lateral 
endothelial tubes first took place. The fused portion already 
extends throughout about eighteen sections. She states that 
from this portion the bulbus cordis is derived. 

Wang reports concerning a ferret embryo, having thirteen 
to fourteen somites, that the two endothelial tubes had united 
in a part of their extent. The fused portion, extending throughout about sixteen sections, appeared to be the ventricular part. 


In our specimens this embryo shows first the fusion of the two 
lateral endothehal tubes throughout only seven sections, having 
a 5 /x thickness, and this united part corresponds to a middle 
third of the ventricle and lies precisely on the midsagittal plane. 
In the guinea pig the fusion of the lateral endothelial tubes 
takes place at a relatively early stage of development — a stage 
in which in the myocardial and endothelial tubes there can be 
distinguished simply the ventricular and atrial portions. In 
the above-mentioned animals investigated by other authors, 
the fusion of the lateral endothelial tubes was first noted in the 
relatively older embryo, in which the different parts of the myocardial and endothehal tubes are already definitely subdivided. 
JMoreover, their embryos show that the fused portion is considerably extended in comparison with this embrj^onic shield. 

The factors which are generally accepted as the cause of 
the loop formation of the endothelial tubes depend on the fact 
that the rate of growth of the two endothelial tubes exceeds 
that of the pleuropericardial cavity. Bonnet depicts a dog embryo in which the primary subdivision of the endothelial tubes 
into sinus venosus, atrium, and ventricle has occurred before 
they have fused to form a single myocardial cavit}'. 

W^ang pointed out the loop formation with the subdivision 
of the heart (atrium, ventricle, bulbus, etc.) in the ferret embryo, 
before the endothelial tubes had become fused. 

But in our specimens there is no loop formation, nor can 
the subdivision of the heart be marked out on either of the 
endothelial tubes before they are fused together, even though 
the ventricle and atrium may be roughl}^ distinguished by their 
difference in size. 

Contrary to the above-mentioned assumption that the loop formation of the endothelial tubes has been brought about, the 
moment of fusion of the two lateral endothehal tubes shows 
quite other facts in the guinea pig. In the embryonic shield 
at this stage of development the confluent part of the myocardial tube grows excessively in the craniocaudal direction and 
decreases its lateral width in comparison with the embryo of 
stage V, as measured and compared on both reconstruction 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 303 

models which had been magnified to the same degree. Consequently, the two endothehal tubes are brought together in 
the median plane, where they come to fusion, by the extreme 
longitudinal stretching of that part of the myocardial tube in 
which the endothelial tubes are enclosed. Concurrently, the 
active dilatation of the two endothelial tubes plays a part in 
bringing about the fusion, which takes place first in the most 
dilated portions. 

In stage VII the myocardium presents cranially a single expanded craniomedian extremity, assuming a sac form, and here 
represents the ventricle, while caudally this myocardial sac is 
bifurcated into two rather slender myocardial prolongations. Their 
demarcation is indicated by the well-developed atrioventricular 
constriction sUghtly cephalad to the foregut opening. The transition from the atrium into the sinus venosus is marked by an 
indefinite indentation on the bilateral myocardial tubes caudad 
to the foregut opening and slightly more distinct on the left 
side. Corresponding to these myocardial constrictions, there 
can be pointed out a similar indentation on the endothelial 
tube on the left side. 

The two lateral endothehal tubes are fused together in the 
cranial two-thirds of the ventricle, and its cranial extremity 
is terminated as a single conical apex opposite to the cranial 
end of the myocardial ventricle. In tracing farther caudalward 
from this united portion, the two endothelial tubes are separated. 

In this stage of the development the myocardial and endothehal heart anlagen of the ventricle present a considerable asymmetry, due to the unequal growth of the individual parts, despite the fact that, in the embryonic shields prior to this stage 
of development, the heart anlagen are shown as a practically 
symmetrical development on both sides, even after the fusion 
of the lateral endothelial tubes has already been accomphshed. 

jMiss Parker describes the heart of the Perameles obesula 
stage V as follows: ''In the ventricular region of the heart, the 
right and left endothelial tubes are approximately equal in size, 
but where there is an inequality the right is the larger." 



In the ferret embryo Doctor Wang says: "It has been found 
that the two tubes, prior to fusion, appear to have been shifted 
as a whole toward the right side and that they remain in this 
position even after partial fusion has taken place.' ' 

In our specimens the myocardial asymmetry of the ventricle is attributed partly to the extraordinary bulging of its right 
wall dorsally, laterally, and slightly caudadly, and partly to 
the deeper infolding of the myocardial wall of the atrioventricular constriction of the right side. In the ventricle the endothehal tubes are much more dilated on the right side than on the 
left throughout its whole length, regardless of the fused or nonfused portions, and it expands considerably dorsally, laterally, 
and caudadly. Consequently, they deviate from the middle 
plane toward the right side and are situated more to the right 
side of the myocardial ventricle. The right endothelial tube 
is elongated dorsally at the caudal third of the ventricle, where 
the two lateral endotheUal tubes, being separated, pass through 
the intermesocardial space vertically. From this portion the 
bulbus cordis develops in a later stage; its termination is cranially bifurcated into two lateral endothelial branches, which 
continue farther cranial-ward into the ventral aortae. 

In stage VIII there are present three distinct constrictions 
on the tubular myocardial surface, infolding the whole thickness 
of the myocardial wall into the mj^oeardial csLvity. Consequently, the myocardium can be subdivided by very distinct 
demarcations into the bulbus cordis, bulboventricular constriction, ventricle, atrioventricular constriction, atrium, sino-atrial 
constriction, and sinus venosus in the craniocaudal succession. 

The bulbus cordis is demarcated from the dorsal wall of the 
expanded right ventricle at its cranial end by the horizontal 
bulboventricular constriction. The bulboventricular constriction shows considerable asymmetry, making a deeper furrow 
on the external surface of the mj^ocardium at the left and 
cranial sides, while at the right side it is present as a 
shallow depression on the external surface, diminishing imperceptibly caudalward, until it has entirely disappeared at 
the part of the right ventricle. Thus neither external furrow 



EARLY DEVELOPMENT OF HEART IN MAMMALIA 305 

nor infolding of the myocardium is shown along the caudal 
boundary of the bulboventricular junction, and here the wall 
of the bulbus is directly continuous into that of the right 
ventricle. Corresponding to the external view, the deep infolding of the myocardial wall as the inner prominent ridge is 
shown at the left and cranial sides of the bulboventricular canal, 
while on the right side it can be recognized only cranially 
and disappears insensibly caudalward. In this fashion the 
curvature of the myocardium at the bulboventricular junction is 
effected in such a vray that its convexity is turned toward the 
right side. On the dorsal aspect of the bulbus cordis there is 
a triangular intermesocardial space disposed vertically and 
sHghtly toward the left side, through which the truncus arteriosus 
passes out from the myocardial cavity up to the floor of the 
foregut. 

The atrioventricular canal is well marked and is disposed 
approximately in the vertical plane. It is produced by the infolding of the myocardial wall, which is deeper and more caudad 
on the right side than on the left. Consequently, the opening 
of this canal is situated on the left side of the middle plane. 
In this relation the myocardial tube forms a marked curvature 
at the atrioventricular junction, turning its convexity toward 
the left side and ventralward, with the result that the ventricular portion lies at the right side and slightl}^ ventrally to the 
atrial portion. This curvature is remarkably accentuated in 
the next stage of the development, in which, for a short extent, 
the dorsal mesocardium disappears at the atrioventricular junction and herewith the ventricle comes to the ventral surface 
of the atrium, being free from the restriction of the dorsal 
mesocardium. 

The ventricle can be divided incompletely into two hmbs 
by the ventral and dorsal longitudinal sulci. At the caudal 
part of the ventricle, for a short distance, the two hmbs are divided into two completely independent cavities by the septal 
wall. The caudal extremity of the right ventricle terminates 
bUndly as a conical process and it projects caudolaterally. 


In the literature I could not find a description of this. On 
first observation it seemed to me that the septal wall and its 
conversion into prominent ridges at the inner surface of the ventral and dorsal myocardial wall were produced by the actual 
fusion of the two lateral myocardial tubes. Therefore, this may 
account for the remnant of the primitive myocardial septum. 
But in the embryos of stages VI and VII there was present no 
septal wall nor prominent ridge similar to this in their single 
ventricle, in which they would be more distinctly present if 
they accounted for the production of the actual fusion of the 
lateral myocardial tubes and the remnant of the primitive cardiac septum. 

Accordingly, it appears to be due to the fact that the caudal 
surface of the myocardial ventricle on the right side is projected 
actively backward by the unequally excessive rate of growth 
in this portion, while in the middle plane a part of the myocardial 
wall does not proportionately accompany this active backward 
growth, but remains as the septal wall. 

LITERATURE CITED 

Arky 1917 A laboratory manual and textbook of embryology. Philadelphia and London. 

Atwell 1915 On the conversion of a photograph into a line drawing. Anat. Rec, vol. 10, p. 39. 

Balfour 1881 A treatise on comparative embryology, vol. 2, London. 

BiscHOFF 1852 Entwickelungsgcschichte des Meerschweinchens. Giessen. 

Bonnet 1884 Beitrage zur Embryologie der Wiederkauer, genommen am Shafei. Arch. f. Anat. u. Physiol. Anat. Ab. P. 1. Jahrg. 1889 a. P. 170 Jahrg. 

1901 Beitrage zur Embryologie des Hundes. Anat. Heft., Bd. 16, S. 233. 

1889 Beitrage zur Entwickelungsgeschichte des Saugethiereherzens. Arch. f. Mik. Anat., Bd. 33, S. 284. 

1900 Rekonstruktionsmethoden. Taschenbuch der Mikroskopischen Technik. Miinchen. 

Bremer 1912 The development of the aorta and aortic arches in rabbits. Am. Jour. Anat., vol. 13, p. 111. 

Dandy 1910 A human embryo with seven pairs of somites measuring about 2 mm. in length. Am. Jour. Anat., vol. 10, p. 86. 

Evans 1909 On the development of the arotae, cardinal and umbilical veins and the other blood vessels of vertebrate embryos from capillaries. Anat. Rec, vol. 3, p. 498. 


Graper 1912 Beobachtung von Wachstumvorgangen an Reiheaufnahmen lebender Hiihnerembryonen nebst Bemerkungen iiber vitale Farbung. Arch. f. Entwickel. d. Organ., Bd. 33, S. 330. 

Hahn 1909 Experimentelle Studien tiber die Entstehung des Blutes und der ersten Gefasse beim Hiihnchen. Arch. f. Entwickelungsmecha. d. Organ., Bd. 27, S. 337. 

Hensen 1875-1S76 Beobachtungen liber die Befruchtung und Entwickelung des Kaninchen und Meerschweinchens. Zeitschrift f. Anat. u. Entwickel. Bd. 1, S. 213-353. 

Hertwig, O. 1910 Lehrbuch der Entwickelungsgeschichte des Menschen und der Wirbelthiere. Jena. 

His 1880 Anatomie menschlicher Embryonen. Leipzig. 

1881 Mittheilung zur Embryologie der Saugethiere und des Menschen. Arch. f. Anat. u. Entwickel., S. 303. 

1886 Beitrage zur Anatomie des menschlichen Herzens. Leipzig. 

1901 Lecithoblast und Angioblast der Wirbelthiere. Abhandlungen der jMathematisch-physischen Classe der Konigslich Sachsischen Gesellschaft der Wissenschaften, Bd. 26, S. 173. Leipzig. 

HuBER 1918 On the anlage and morphogenesis of the chorda dorsalis in mammalia, in particular the guinea pig (Cavis cobaya). Anat. Rec, vol. 

14, p. 217. 
Keibel 1896 Studien zur Entwickelungsgeschichte des Schw^eines. Mor
phologische Arbeiten, Bd. 5, S. 1. 
KoLLiKER 1861-1879 Entwickelungsgeschichte des Menschen und der hoheren 

Thiere. Leipzig. 

1884 Grundriss der Entwickelungsgeschichte des Menschen und der hoheren Thiere. Leipzig. 

Lewis 1903 The intra-embryonic blood vessels of rabbits from 8i to 13 days. Proceedings of the Association of American Anatomists. Am. Jour. Anat., vol. 3, p. 12. 

{{Ref-Mall1912}}

Mall 1912 On the development of the human heart. Am. Jour. Anat., vol. 13, p. 249. 

Martin 1902 Die Ontogenie der Haustiere. Lehrbuch d. Anat. d. Haustiere. Bd. 1, S. 23. Stuttgart. 

Miller and McWhorter 1914 Experiments on the development of blood vessels in the area pellucida and embryonic body of the chick. Anat. Rec, vol. 8, p. 203. 

{{Ref-Minot1892}}

Minot 1892 Human embryology. New York. 

Parker 1915 The early development of the heart and anterior vessels in marsupials, with special reference to Perameles. Proceedings of the General Meetings for Scientific Business of the Zoological Society of London, p. 459. 

Rabl 1887 Ueber die Bildung der Herzens er Amphibien. Morphol. Jahrb., Bd. 12, S. 252. 

1889 Theorie des ^Mesoderms. Morphol. Jahrb., Bd. 15, S. 113. Ravn 1889 Ueber die Bildung der Scheidewand zwischen Brust und Bauchhohle in Saugethierembryonen. Arch. f. Anat. u. Entwickel., S. 123. 

1895 Ueber das Proamnion, besonders bei der Maus. .\rch. f . .\nat. u. Entwickel., S. 189. 

Reagan 1915 Vascularization phenomena in fragments of embryonic bodies completely isolated from yolk sac blastoderm. Anat. Rec, vol. 9, p. 328. 

Robinson 1892 Observation upon the development of the segmentation cavity, the archenteron, the germinal layers, and the amnion in mammals. Quarterly Journal of Microsc. Sci., vol. 33, Op. 369. 

1903 The early stages of the development of the pericardium. Journal of Anat. a. Physiol., vol. 37, p. 1. 
Rose 1889 Zur Entwickelungsgeschichte des Saugethierherzens. Morphol. Jahrb., Bd. 15, S. 436. 

1890 Beitrage zur vergleichenden Anatomie des Herzens der Wirbelthiere. Morphol. Jahrb., Bd. 16, S. 27. 

RouviERE 1904 Etude sur le developpement du pericarde chez le lapin. Jour. de I'Anatomie et de la Physiologie, p. 610.
 
RucKERT UND MoLLiER 1906 Die erste Entstehung der Gefasse und des Blutes bei Wirbeltieren. Handbuch der vergleichenden und experimentellen Entwickelungslehre der Wirbeltiere. Jena. 

ScHULTZE 1914 Early stages of vasculogehesis in the cat (Felis domestica) with special reference to the mesenchymal origin of the endothelium. Memoirs of the Wistar Institute of Anatomy and Biology, no. 3. 

:1915-16 The fusion of the bilateral anlagen of the heart and the formation of the bulbo-ventricular loop in embryo of the cat. Anat. Rec, vol. 10, p. 242. 

Selenka 1883-1892 Studien iiber die Entwickelungsgeschichte der Thiere, 1-5 H. Wiesbaden. 

Sobatta 1902 Ueber Entwickelung des Blutes, des Herzens und der grossen Gefassstilmme der Salmoniden nebst Mittcilungen liber die Ausbildung der Herzform. Anat. Heft., Bd. 19, S. 579. 

Spee 1889 Beobachtungen an einer menschlichen Keimscheibe mit offener Medullarrinne und Canalis neurcntericus. Arch. f. Anat. u. Entwickel., S. 159. 

Strahl und Carius 1S89 Beitrage zur Entwickelungsgeschichte des Herzens und der Korperhohlen. Arch. f. Anat. u. Entwickel., S. 231. 

UsKOW 1883 Ueber die Entwickelung des Zwerchfells, des Pericarduims und des Coeloms. Arch. f. mik. Anat., Bd. 22, S. 143. 

Wang 1917 The earliest stages of development of the blood vessels and of the heart of the ferret embryo. Jour, of Anat., vol. 3, p. 107. 

Wilson 1914 Observations upon young human embryos. Jour, of Anat. and Physiol., vol. 48, p. 315. 

Yeates 1915 Studies in the embryology of the ferret. Studies in anatomy from the Anatomical Department of the University of Birmingham, p. 71.
 
Yoshinaga 1920 Beitrage zur Streitfragen iiber den Ursprung der Gefass
zellen bei den Amphibien. Acta Scholae Medicinalis Universitatis Imperial is in Kioto, vol. 3.