Talk:Paper - The early development of the sheep heart (1946)

From Embryology

[ 75 ]


By E. J, FIELD, Anatomy Department, St Bartholomew’s Hospital Medical College, London

Since about 1880 an extensive literature has accumulated relating to the development of the heart in a variety of animals representative of the main groups; and on the basis of this knowledge certain generalizations have been established regarding the sequence of events whereby the unpaired heart rudiment arises in these different groups (Mollier, 1906). Amongst the mammals most attention has been devoted to the cat (Martin, 1902; Schulte, 1914; Watson, 1924), the ferret (Wang, 1917) and the guinea-pig (Yoshinaga, 1921), though valuable observations have also been recorded for the rabbit (Hensen, 1876; Rouviére, 1904), the sheep (Bonnet, 1884, 1889) and the dog (Bonnet, 1901) amongst others. For the human, the earliest stages are to be found in the Glaevecke embryo of von Spee and in the Sternberg and Payne embryos.

Apart from Kuhlemann (1754), whose remarkable work is nowadays largely of historical interest only, Bonnet appears to be the sole author who has recorded observations on the early development of the sheep heart in material of known age. His work is primarily concerned, however, with the evolution of the general form of the embryo, and such references as he does make to the heart are purely incidental. The writer has searched in vain for the further exposition of his findings promised by Bonnet at the end of his 1889 paper (‘wird fortgesetzt’). In view of the incompleteness of these observations, especially as they indicate a mode of origin of the cardiac rudiments at variance with that established for other forms, it was thought desirable to seek confirmation of Bonnet’s findings and to extend them as far as might be possible with the material at hand.

Much controversy exists in the literature regarding the relative roles played by the mesoderm and the entoderm in vasculogenesis, both inside and outside the embryo. Certain conclusions have been reached from the study outlined below, and how far these findings and those recorded by other authors are compatible with the classical theory of germ layers is briefly discussed. In this description reference is made to extra-embryonic vasculogenesis in so far only as it throws light on the endothelial problem in general.

The subject-matter of the present work thus falls under two heads:

(1) the origin of the angioblastic cells, and

(2) the sequence of events which culminate in the establishment of an unpaired heart rudiment.

The further changes whereby this rudiment attains its definitive form are not here under review.


The writer was fortunate in having access to serial sections of some ‘timed’ embryos in the possession of Profs. W. J. Hamilton and J. D. Boyd, to whom he would express his gratitude. Further, it was possible through the kind offices of Mr John Hammond, Jr., to acquire additional ‘timed’ embryos of the ages required for further study. Altogether three embryos of 15 days, and four of 16 days were found to show heart formation at the required stages. Earlier specimens showed no cardiac anlage.

The sheep were killed at the requisite number of days from the time of mating and the uteri excised immediately. The embryos with their membranes were washed out of the uterine horns with Locke’s solution. Sometimes prolonged search was necessary before the embryo could be identified amidst the voluminous entanglement of membranes. After being photographed in saline the embryos were fixed, some in Bouin-Allen and some in Zenkerformol. The smaller embryos were embedded in agar and blocked in paraffin according to the method described by Samuel (1944), whilst the larger were treated in the ordinary way by Peterfi’s double embedding method. Serial sectioning was carried out in. all cases, sometimes at 7, but mostly at 5y. Staining was by a modified Harris’s haematoxylin or by Heidenhain’s method. Eosin or orange G were employed alternatively as counterstains.

DESCRIPTION Extra-embryonic vasculogenesis

The earliest stages of this process can be seen in the 15-day embryos B15 and B 47L possessing four and six pairs of somites respectively. Here and there in the yolk sac wall spaces are present between the entoderm and the mesoderm (PI. 1, fig. 1), and when followed in serial sections these spaces are found to be of spherical or spheroidal form, They are little more than well-defined and localized intervals between the two germ layers. That they are not mere artefacts is suggested by their discrete arrangement amidst tissue which shows but little evidence of shrinkage. Very soon, however, these vascular spaces become walled entirely by mesodermal cells (Pl. 1, fig. 2) which are at first plump, but assume a flattened and characteristically endothelial form as fluid accumulates within the vesicles (Pl. 1, fig. 8). Evidences of mesodermal proliferation are often to be found near such vesicles. 76

Entodermal participation in the wall of these primary angiocysts is thus a transient feature only, and the later angiocysts appear located entirely within the mesoderm; these vesicles may be called secondary angiocysts. The appearances are such that it seems unlikely that the entoderm makes any contribution to the definitive vessel wall; nor is blood-cell formation to be found in these early stages. ,

Intra-embryonic vasculogenesis and cardiogenesis

The general arrangement of each embryo will be described in so far as it is necessary for an appreciation of the ontogeny of the earliest vessels and of the heart.


The wax-model reconstruction (Text-fig. 1) shows that this part of the coelom has resulted from the coalescence of multiple, discrete, slit-like spaces in the mesoderm between the yolk-sac entoderm ventrally and the bulging anterior part of the brain plate dorsally (not shown in the figure). Many of these small spaces are as yet unconnected with the coelom or connected with it only by narrow and tortuous channels. The antero-posterior measurement of the uninterrupted transverse part of the coelom is about 32. On either side it is continued into well-defined lateral channels, though here, too, there is evidence of origin from multiple spaces as indicated by the double channel seen on the left side. In effect there is in this embryo a horseshoe

Text-fig. 1. A wax-plate reconstruction of the foregut diverticulum of embryo B 15 (15 days; four somites) together with the entoderm of the adjoining part of the yolk sac. The coelomic spaces are dark and appear as two lateral channels joined by a cranio-median limb below and in front of the foregut pouch. This cranio-median coelom is in process of formation and is represented in large part by more or less discrete mesodermal clefts.. On the left side the lateral channel is double. On the right side there is a gap where slight damage to a few sections prevented the precise outline of the channel from being modelled. The small black rectangular dots indicate the position of angioblasts. The lines AB and CD are the planes of the sections shown in Pl. 1, figs. 4 and 5 respectively. The model is viewed from above and slightly behind. Magnification x 250.

B15. This embryo, of 15 days and possessing four pairs of somites, is the youngest in which vascular anlagen are to be found. In the anterior part of the embryonic plate changes have occurred resulting in the appearance of a shallow foregut diverticulum some 32, in depth. The floor of this diverticulum in its posterior part projects ventrally as a gutter, just anterior to the anterior intestinal portal. It presents many mitotic figures, perhaps indicating an active ‘knitting together’ of the lateral coelomic folds by whose apposition, the appearances suggest, the gut has come into being. Below and in front of the anterior end of the foregut pouch is the cranio-median limb of the coelom.

shaped cavity whose anterior part has just become established. Between the entoderm and mesoderm which form the floor of the distinctly patent part of the cranio-median coelom (pericardial cavity) are a few cells—some half-dozen in all—which, for reasons to be considered later, can be called angioblasts. Indeed, mitotic figures are no more numerous in the floor of the pericardium than elsewhere in this rapidly growing embryo; nor is there a heaping up of the splanchnopleure to be seen in the floor of the pericardial cavity. A section along the line AB in Text-fig. 1, just caudal to the pericardial cavity, is shown in Pl. 1, fig. 4. The lateral coelomic channels leading back from the lateral Development of the sheep heart

extremities of the pericardial cavity are seen on either side, and four well-defined angioblastic cells can be seen ‘pavementing’ the entoderm. Similar cells are dusted about between the entoderm of the

yolk sac and the mesoderm over the whole anterior |

part of the embryonic plate. Many are in relation to the coelomic clefts mentioned above, but many also are not so situated (Text-fig. 1). A section taken a little more caudally, along the line CD in Text-fig. 1, shows the foregut diverticulum and also some angioblastic cells in relation to it (Pl. 1, fig. 5).

Behind the anterior intestinal portal the coelomic folds stand apart, and it is in this region, where vitelline veins will come into existence at a slightly: later stage, that vasculogenic cells are most in evidence. Pl. 1, fig. 6 shows such cells ‘pavementing’ the entoderm in a more or less continuous strand, and since similar appearances are manifest

in many sections at about this level the name

‘gefassblatt’ employed by Bonnet is not inappropriate for the vasofactive sheet.

The chief criteria by which these cells are considered angioblastic are, first, their location close up against the entoderm—where the next embryo to be considered will show the process of their conversion into angiocysts; and, secondly, their characteristic orientation—their long axes being arranged parallel to the entoderm layer (Pl. 2, fig. '7). This alinement of the angioblasts is all the more distinctive in that the cells of the splanchnopleure forming the medial wall and floor of the lateral coelomic channel also show a definitive arrangement, their long axes being disposed radially, i.e. at right angles to the angioblasts and so producing a ‘palisade’ effect. No difference in nuclear or cytoplasmic staining is to be noted between the angioblasts and either the entoderm or mesoderm cells, though the angioblasts do tend to have rather scanty cytoplasm. An isolated angioblast is usually spindle-shaped.

As to the origin of these angioblasts or vasofactive cells, there is evidence of a genetic relationship with both the splanchnopleure and the entoderm. Between these two layers there is throughout a well-marked space which is no doubt attributable in large measure to shrinkage. The cells of the splanchnopleure~bordering this space show signs of ‘loosening’, and many gradations can be seen as individual cells appear to detach themselves from the main layer and take up the orientation and position described above as characteristic of the angioblasts. It seems certain that many (probably the greater number) of the angioblasts are of such mesodermal origin. On the other hand, at many points the most intimate contact between angioblasts and entoderm exists, and an entodermal origin of some of the former seems very probable (Pl. 2, fig. 8). Certainly protoplasmic continuity


between them can be established especially with the stereoscopic binocular microscope.

Traversing the ‘shrinkage space’ mentioned above are long or short protoplasmic fibres, some plump and others very tenuous, derived from the splanchnopleure cells adjoining the space. Some of these fibres may make contact with the entoderm cells, but most tail off and are lost in the space (PI. 2, fig. 7). To what extent these fibres exist in the actual living condition must necessarily remain uncertain. This matter will be discussed later.

This embryo does not show a heaping up of the splanchnopleure in the floor of the lateral coelomic channel nor do the angioblasts exhibit that propensity to the. formation of ring-like structures or angiocysts both of which are such prominent features in the next specimen.

_B47L. This embryo, though also of 15 days, is

_ further developed than B 15 and has six pairs of -somites. The foregut has deepened to about 175,

and stretching across beneath all but the caudal 60, of its extent is the cranio-median limb of the coelom. This, too, is 175, in antero-posterior measurement and so projects forwards about 60, in front of the foregut diverticulum. Evidently many of the rudimentary coelomic clefts seen in B 15 have been taken up into the pericardial cavity resulting in the considerable increase in its anteroposterior extent, and at the same time the pericardium as a whole has undergone a caudal migration either relative or absolute, which has resulted in its new position partly underlying the foregut. The guttering of the pharyngeal floor noted in B15 is prominent especially in its more caudal part just in front of the anterior intestinal portal, and the splanchnopleure covering the ventral surface of the gut here bulges down into the pericardium. Scattered here and there in the closest relation to the steep entodermal side walls of the pharyngeal gutter are small vesicles and’ groups of angioblasts. One or two angiocysts occupy a median position beneath the foregut floor and represent the earliest median vascular rudiments. Such a median angiocyst is seen in Pl. 2, fig. 9, just caudal to the emergence of the coelomic channels from the lateral limits of the craniomedian coelom, and vasofactive strands are seen on both sides close up against the entoderm. The vasofactive strand consists of a chain of minute vesicles joined by solid nucleated protoplasmic bridges of angioblasts. A high-power view of a vasofactive strand is shown in Pl, 2, fig. 11. The angioblast cells are plump and as yet unflattened by accumulation of fluid. The close relation to the entoderm is well shown, and participation of this latter in vasculogenesis is almost certain. A strand of angioblasts continuous with entoderm and presumably derived from it is shown in PI. 2, fig. 8. 78

Caudal to the anterior intestinal portal where the coelomic folds stand apart, vesicle formation is much more in evidence than it is in relation to the pharyngeal floor (Pl. 2, fig. 10). This region is that

in which the vitelline veins later develop and corre- . is traced forwards its form is no longer seen to

sponds to that in which angioblasts were found to be most numerous in the preceding specimen. The vitelline anlagen are thus the most advanced in their representation at this stage.

B 52B (16 days). In this specimen, of six somites, the foregut has a depth of about 182y, and the transverse coelom stretches across beneath its anterior 105y. The pericardium extends forwards as far as the anterior end of the foregut diverticulum. It does not, as in the previous specimen, project



In its more caudal part this median vascular rudiment is projected ventrally in the midline in adaptation to the guttering of the pharynx and is uniformly covered with splanchnopleure. .

As this broad confluent collection of angiocysts

follow slavishly the contour of the gut floor, for, even where the latter shows but little guttering, the endocardium covered with its layer of splanchnopleure projects as a ‘keel’ into the pericardial cavity (Pl. 2, fig. 12).

From the front end of the subpharyngeal blood lake there pass forwards two more or less distinct vessels beneath the most anterior part of the foregut; these must be ventral aortae. They, too, are

Text-fig. 2. A wax-plate reconstruction of the foregut diverticulum of embryo B 52B seen from below and the right. The vesicular cardiac rudiment is seen besetting the pharyngeal floor. Caudally two vasoformative strands diverge to either side of the anterior intestinal portal. The strand A may be the precursor of a branchial arch artery.

Magnification x 250.

beyond it and is 105, in antero-posterior measurement. The pericardial cavity thus appears to have undergone a further caudal shift in relation to the pharyngeal floor. The caudal part of the foregut floors shows ventral guttering as in the preceding embryo, and the general conformation of the embryo and the coelom does not differ materially from B47L. The vesicular vascular primordia which were noted in this latter, however, are considerably more advanced in their development. A model of the foregut and the angiocysts related to it is shown in Text-fig. 2. The whole ventral aspect of

- the foregut is beset with bleb-like angiocysts, many of them discrete but many, on the other hand, confluent to form what, may be termed a ‘subpharyngeal lake’.

in close apposition to the entoderm. No welldefined branchial arch vessels are discernible though perhaps the angiocystic tracks indicated at A in Text-fig. 2 may be such rudiments.

B51L (15 days). Although a day younger than the preceding embryo this specimen is more advanced. Unfortunately, it sustained damage in preparation. There is in it, however, a definite capillary U-loop beneath the foregut.

B52A (16 days). This embryo has the same general conformation as has B 47L and like it has six pairs of somites. The foregut diverticulum has a depth of about 255, and the pericardial cavity extends beneath it for 120. Caudal to the anterior intestinal portal where the coelomic folds stand apart the vitelline veins are seen. They appear as Development of the sheep heart

confluent angiocystic chains, and the calibre of the endocardial tube so formed varies considerably at different levels. This disparity can also be seen on the two sides (Pl. 3, fig. 18). Over the endothelial tube the splanchnopleure bulges into the lateral channel, and it is of interest to observe that this bulging of the medial coelomic wall may occur independently of the presence of a subjacent tube as is seen on the left side of the figure. The ‘vitelline ridge’, as it may be called, cannot, then, be due to the mere mechanical presence of the endothelial tube.

There are two other features relevant to cardiac development to be noticed in this specimen. The first is the presence of a distinct reticular formation or mesostroma around the endothelial tube connecting it with the overlying splanchnopleure. This meshwork is much more distinct than the faint mesostroma which was seen partially to invade the ‘shrinkage space’ between mesoderm and entoderm in B 15, and is apparently a further development of it. Further, it is continuous with the fibrillary stroma which is related to the median heart rudiment. The second feature is that the vitelline vein at many points stands away a little from the entoderm, as is seen on the right in Pl. 3, fig. 18, and does not bear the same close and constant relation to this layer that the cardiac endothelium does (Pl. 3, figs. 14, 15).

As the vitelline veins are traced cranialwards they assume the form of a series of contiguous and partially continuous angiocysts applied to almost the whole medial wall of the coelomic channel. Thus when communication between the vessels of each side takes place cranial to the anterior intestinal portal, a broad space is formed showing traces of its origin by the fusion of initially discrete angiocysts. In cross-section this space has the appearance of a capillary U-loop subjacent to the foregut entoderm (PI. 3, fig. 14). This is an elaboration’ of the conditions obtaining in the embryo B 52B, and indicated in the model (Text-fig. 2). This broad vascular rudiment underlying the pharynx is to be looked upon as the heart primordium. How closely the endocardium is related to the entoderm is well shown in PI. 8, fig. 15; and here, too, the character of the fibre meshwork related to the non-entodermal aspect of the endocardium is apparent. This meshwork, together with the enclosing splanchnopleure, is the early myoepicardial mantle. This latter follows faithfully the contour of the endocardium and shows no independent variations in surface form, though its compact outer layer does exhibit localized differences in thickness. As the endothelial heart rudiment is traced forwards it is seen to conform closely to the shape of the foregut floor. Even in its anterior part, where in B 52B (slightly younger) the endocardium was seen to free itself from the influence


of the foregut dorsal to it, the endocardium in the present specimen continues to accommodate itself in this way. Primitive vessels may be recognized

‘passing dorsally from the forepart of this lake-like

heart round the foregut wall.

A section taken 40, in front of the anterior intestinal portal shows that as the lateral coelomic channels approach each other to break through beneath the rudimentary heart, there is left between them a septum of splanchnopleure (PI. 3, fig. 16) which is looked upon by the writer as a ventral mesocardium ; it will be further considered below.

From the conditions described above it will be appreciated that there can be no question of the existence of a dorsal mesocardium at this stage, for the broad lake-like heart is applied in a sessile manner to the foregut entoderm and is in no wise ‘suspended’ from it.

B16R (16 days). In this specimen the foregut extends forwards through 368. The pericardial cavity underlies it from a point 72, in front of the anterior intestinal portal to a point about 96 behind the anterior extremity, so that the cavity itself is 200, in antero-posterior measurement.

Caudal to the anterior intestinal portal the vitelline veins are found on either side as well-formed endothelial tubes, each being surmounted by a cushion of splanchnopleure which projects into the lateral coelomic channel (Pl. 3, fig. 17). The fibre meshwork between endothelium and splanchnopleure is present though it is not so well marked as in the preceding specimen. The. two venous channels are separated from the entoderm by distinct intervals occupied by the same mesostroma. Two small vessels visible more dorsally are found, by tracing them up through the series of sections, to become continuous with the dorsal extremities of the U-shaped heart. The few scattered cells © between the coelomic folds in the lower part of the section indicate the imminent fusion of these folds which occurs, indeed, a few microns more anteriorly. .

Anterior to the anterior intestinal portal fusion of the two vitelline veins takes place producing an endothelial channel beneath the foregut floor. This confluence of the vitelline veins is regarded as the caudal part of the heart, and the broad band of splanchnopleure below is looked upon as a ventral mesocardium. These appearances are shown in Pl. 4, fig. 18, taken 40, in front of the portal. The myo-epicardial mantle in this and the succeeding figure is seen to exhibit changes in contour independent of those of the endocardial tube. Pl. 4, fig. 19 is of a section taken 120, in front of that shown in the previous figure and portrays the heart U-formation at its highest development in this embryo. The close apposition of the endocardium to the foregut entoderm is well shown in Pl. 4, fig. 20, a high-power view of the same section. In it 80

the flattened endothelial cells are clearly seen ‘pavementing’ the low columnar entodermal cells. Incomplete septa derived from the contiguous walls

of the original angiocysts project here and there

into the lumen like baffle-plates and account for the cellular material seen scattered within the lumen. This median fusion of the two vitelline veins extends forwards for some 120, gradually narrowing anteriorly, and throughout this extent the blood space reaches up the side wall of the gut to about the same level as in Pl. 4, fig. 19. The name ‘subpharyngeal lake’ is therefore well merited. From the sides of the anterior part of this space channels stretch round the gut wall and are recognized as aortic arches.

Text-fig. 3. A wax-plate reconstruction of the foregut pouch of embryo B16R viewed from below and the right. The position of the foregut floor is indicated by the dotted line. The heart rudiment has begun to bulge ventrally, and the wide extent of its applicatiqn to the entoderm is illustrated. Caudally the vitelline veins are seen skirting the anterior intestinal portal. Magnification x 250, reduced to x 140 for reproduction.

The myo-epicardial mantle shows localized thickenings in its wall. In addition, it shows for the first time a surface contour not entirely dependent on that of the enclosed endecardium; this is seen in Pl. 4, fig. 19 (left side), where a distinct groove is found on the upper part of the heart surface.

B16L (16 days). The foregut diverticulum in ‘this embryo of. eight somites extends forwards some 280u—considerably less than in the twin embryo (B 16R) just described. The general form is similar to that of the latter, but the present embryo is definitely more advanced. The pericardium begins 80, anterior to the anterior intestinal portal and has an antero-posterior measure E. J.


ment of 200, so that it underlies the foregut floor save in its caudal 80,, and does not project beyond it distally. The heart region of this embryo has been reconstructed in wax. Init the ventral mesocardium remarked upon in the last specimen is quite distinct in felation to the caudal -part of the heart. It is wedge-shaped with the base opposite the anterior intestinal portal. A section just behind the apex of the wedge, 72,,.anterior to the intestinal portal, besides showing the mesocardium presents other features relevant to the heart development (PI. 4, fig. 21). On the left surface of the myo-epicardium a groove runs fgrwards from the bay above the ‘vitelline ridge’, and this groove is the beginning of the demarcation of the heart into venous and arterial segments, or in other words indicates the beginning of a cardiac loop with convexity directed forwards, ventrally and to the right; and secondly, the ‘arch’ artery leaving the left part of the median heart is considerably larger than that from the right part. A branch from the left arch passes across ventral to the pharynx immediately beneath its entoderm and joins the right outflow. The origin of the right branch is, moreover, rather caudal to that of the left branch. These points are being further investigated. Thirdly, the loose cellular part of the wall of the left outflow. is in marked contrast to the dense cellularity of the ventral. portion of the cardia¢ loop. ©

The stages which have been described may be here recapitulated. Strands of angioblasts make their appearance at the 15th day at a time when the foregut diverticulum has just come | into existence. The cranio-median limb of the coelom (‘pericardial cavity’) is yet incompletely formed and there is no evidence of cardiogenesis in its floor. The foregut elongates and the pericardium comes to underlie it. The strands of angioblasts meet in front of the anterior intestinal portal and in their course angiocysts develop. The vasculogenic strands become progressively more cystie, the change being more advanced in the more caudal parts—where vitelline veins will be recognized later—and in this way a median vascular rudiment closely applied to the foregut floor entoderm, and formed of contiguous and partially continuous blebs, comes into existence. An extensive subpharyngeal blood lake is formed by continued confluence of these vesicles, and in its posterior part it possesses a ventral mesocardium. No dorsal mesocardium is present in these early stages.

DISCUSSION Extra-embryonic vasculogenesis The description given above of early vessel formation in the yolk-sac wall of the 15- and 16-day-old

sheep embryo is in full agreement with the findings of Bonnet (1884). He, too, was of the opinion that Development of the sheep heart

participation of the entoderm was only a transient phenomenon. In later stages, he says, the mesoderm cells bend right round and form tubes (‘durch starkeres Einbiegen zu Rohren geschlossen’) and the entoderm is thus excluded from the vessels and is seen as a well-defined and regular layer. A precisely similar sequence of events is recorded by van der Stricht (1899) in the bat; and for the pig, too, Keibel (1894) established a like process. Recently, Gladstone & Hamilton (1941) have described an analogous mechanism of vasculogenesis in the yolk-sac wall of the Shaw embryo. It appears, therefore, that the process described above is one which is common to man and a variety of mammals. Duval (1897), on the other hand, _ claims that in the bat those cells which are destined to form blood vessels (and blood cells) can be seen to bud off from the entoderm of the yolk-sac wall. His figures 107, 108 and 109, Planche X, are certainly very suggestive of this. A re-examination of the present sheep material in the light of these figures did not, however, reveal any comparable appearances.

Intra-embryonic vasculogenesis

Some of the earliest researches on this problem were carried out by His on the chick embryo, and his definitive viewpoint is embodied in what is usually called the ‘Einwachsungslehre’. According to this concept vessels which are first formed in the yolk-sac wall bud off nucleated protoplasmic sprouts which grow inwards and invade the embryo itself between the entoderm and splanchnopleure. By the joining up and subsequent canalization of these ‘Gefass-sprossen’ the early intra-embryonic vessels come into being. From these capillaries formed round the cephalic end of the embryonic plate the heart takes origin. Although widely accepted this doctrine evoked a certain scepticism. Thus Ranvier expressed himself: ‘Mais il est clair qu’aucun embryologiste n’a pu suivre ce développement continu par bourgeonnement dans le corps de l’embryon; c’est la une simple hypothése.’ As will’ be seen, recent work has fully justified this criticism.

Impressed by his demonstration that the first aortic arches in amphibians developed as outgrowths from the heart endothelial sacs, Rabl propounded the theory that the endothelium of all blood vessels is in the last resort derived from the original heart. endothelium. Rabl’s belief that ‘Endothel stammt nur von Endothel’ is in accordance with the His teaching, Rabl, however, postulating an intra-embryonic vasculogenic focus. As to the precise origin of the initial vasofactive cells he is silent. This theory, too, claimed many adherents, and Sobotta, from his work on bony fishes, even went so far as to evoke a His theory in reverse—for he described the yolk-sac vessels as


outgrowths from the embryonic vessels; and .this he did—‘ ohne irgendwelche Beweise fiir die Richtigkeit derselben zu erbringen und ohne gegenteilige Angaben zu entkraften’ (Mollier, 1906).

The third theory concerning the origin of vascular endothelium is associated with the name of Riickert. He observed the apparent in loco formation of vasculogenic cells in selachian embryos as a delamination from the ventral part of the mesoderm. Riickert also made the very important observation that the endothelium of the aortae in these fishes had a dual origin, being derived in part from the local mesoderm and ‘in geringerer Anzahl aus der dorsalen Darmwand’. Here then is an admission of a possible duplicity of source for the vascular endothelium, and it stands in sharp contrast to the restrictive teaching of both His and Rabl. Riickert’s assertion that entodermal as well as mesodermal cells may contribute to endothelium formation was based not only on the development of the aortae, but also on his findings in relation to the vessels of the mandibular arches; for here, too, entoderm of the gut was found to be implicated. Many authors gave their support to Riickert’s theory of the local origin of vascular endothelium, though differences of opinion were expressed regarding the participation of the entoderm. (For a summary of the literature up to 1906 see Mollier.)

In a study in which observations are so much a matter of opinion, fresh advances can come only with the introduction of a new technique. In 1914, | Miller & McWhorter published their since classical work in which it was clearly demonstrated that intra-embryonic vessel formation in the chick was an in loco phenomenon and took place even when all continuity with the adjacent extra-embryonic region had been severed. These studies ushered in a phase of experimental work (typified by that of Stockard (1916) on Fundulus) which has declared itself in favour of the local origin of endothelium. That this does occur may now be taken as an established fact.

The earliest vascular anlagen in the embryos which form the basis of this paper take the form of strands of cells or even isolated cells to which the name angioblasts or vasofactive cells has been applied. The appearances they present and their characteristic disposition between the entoderm and the splanchnopleure have been set out above (Pl. 2, figs. 7, 8). They do not differ in essential characters from those described for all other animal forms investigated. Thus, von Schulte (1914), who has gone into these features at great length in his discussion of early vasculogenesis in the cat, describes these cells as ‘dorso-ventrally flattened, in groups of two or three’ (p. 44). His statement that they take on a slightly paler stain (p. 43) is not borne out in the present material in which no difference from the adjacent mesoderm 82

can be seen; a finding in accordance with that of Yoshinaga (1921) in the guinea-pig. Sternberg (1927), on the other hand, claims that the cells of the vascular anlage possess characteristic histological features. Their nuclei are described as being smaller and more deeply stained by haematoxylin, and their rather scanty cytoplasm is said to contain very fine fibrils which cause it to appear rather darker than that of the neighbouring cells. These features are not apparent in the sheep material. The characteristic orientation of the angioblasts von Schulte seeks to attribute to a current of fluid directed ‘ventromesad’ between the mesoderm and the entoderm, and presumably the cells are supposed to present themselves ‘end on’ to the stream. To most workers, however, such a mechanism will, no doubt, appear improbable, especially when the syncytial nature of the tissue concerned is recalled. In the embryos of the present series and in those pictured by Bonnet and by von Schulte himself, as well as those shown by other authors, there exists a considerable space or series of spaces between the ventral aspect of the mesoderm and the entoderm. This space is not quite so large in Yoshinaga’s figures (1921), though he, too, refers to it in his text. It is evidence of considerable shrinkage having occurred during the preparation of the material. Obviously disruptive forces of considerable violence have been at work, and any ‘Entwicklungsmechanik’ based upon the detailed and precise appearances of individual cells under such circumstances must necessarily be treated with caution. Probably the region between the splanchnopleure and the entoderm is indeed one rich in fluid, and this fluid may be contained within a protoplasmic meshwork which is continuous with and acts as a bridge between the two layers. The changes induced by fixation (cf. Tellyesniczky, 1898-1905), as well as those incidental to paraffin embedding, may well account for the ‘interdermal cytodesmata’ of Studnitka (1911) or the ‘fibres’ of von Szily (1903) so accurately and painstakingly described. The forcible tearing apart of the germ layers might likewise be responsible for the appearances described as ‘loosening’ of the deepest part of the splanchnopleure, and ‘detachment’ of splanchnopleural cells to form angioblasts. Probably also the ‘palisade’ arrangement of the mesoderm cells bounding the lateral coelomic channel is in part due to the operation of these same forces which would tend to orient the cells in the manner observed. On the other hand, the angioblasts are in the closest possible relation to the entoderm (Pl. 2, fig. 8). In many places, indeed, direct protoplasmic continuity between entoderm and syncytial strands of angioblasts can be made out. Perhaps the truth of the matter may be that we are dealing here with a syncytium comprising mesoderm and entoderm as a single entity. It is interesting to


observe in this connexion that Martin’s (1902) figure does not show cell outlines in many places. If this interpretation be correct, then angioblastic cells arise from both the mesodermal and entodermal components of the syncytium, and the antithesis between views so hotly maintained in the past is, in large part, resolved. There can be no doubt that our ideas of the individuality of germ layers born in the heyday of early mechanistic embryology have been much modified at the hands of modern experimentalists and tissue culture workers,

The plasticity retained by cells, even in the fullgrown organism, has been much underestimated (cf. Clarke (1914, 1916), and more especially Levander (1945)). Certainly the ‘determination’ of germ layers in the early embryo is much less rigid than has been supposed. The genealogy of a particular cell in the embryonic mesenchyme complex is of less importance in determining its fate than are its position and its relations. In the particular instance of angiocyst formation it is probable that a ‘focus’ appears which induces in the surrounding cells the tendency to enclose it. The characteristic orienta‘tion of the angioblasts and the disposition of the angiocysts with their long axes for the most part parallel to the entoderm suggests that. powerful directing forces are at work. Some of these cells may later escape beyond the influence of the ‘directing focus’ and may then revert to their non-specific form. Such retrogression has long been known to occur. Huntington (1911) described such changes in connexion with the development of the perivenous plexuses—‘their endothelial lining appears to revert to the indifferent type of the embryonic mesodermal cell’ (cf. von Schulte, 1914, pp. 10 ff.).

Origin of the heart

Our knowledge of the development of the heart in a number of vertebrate groups has been summarized, up to 1906, by Mollier in O. Hertwig’s Handbuch der Vergleichenden und Experimentellen Entwicklungslehre der Wirbeltiere, vol. 1. In order to assess the position of the sheep in this respect it is necessary to outline in brief the essential findings in reptiles, birds and other mammals including man.

In the reptilian embryo (Lacerta) paired cardiac anlagen unite whilst they are yet in the form of strands of angioblastic cells, amongst which, however, fine spaces are beginning to appear. At this stage of commencing union the pleuro-pericardial channels are quite separate. Only when union has been established is a ‘loosening’ of the cells of the median rudiment to be observed, and this leads to a ‘mesenchymatous’ or ‘meshwork’ condition. The further development of this type of heart is complicated by a characteristic lateral bending of the embryo and does not concern us here. Development of the sheep heart

In the chick embryo the cardiac anlagen take the form of angioblastic strands which run together cranial to the anterior intestinal portal, at a time when the pleuro-peritoneal channels are quite independent. In the process of union the heart strands take on a meshwork or ‘mesenchymatous’ appearance, but no definitive endothelial tubes come into existence before fusion is complete. The coelomic channels then extend ventrally and medially towards each other beneath the median cardiac rudiment, and in this way the ventral mesocardium comes into being as the septum between them. This mesocardium is only transient and gives way very soon. A myo-epicardial mantle of splanchnopleure clothes the ventral and lateral aspects of the fused anlagen, whilst the dorsal aspect is in broad contact with the entoderm of the foregut floor. It is important to note that in the chick the fusion of the cardiac anlagen occurs at a time when each is in the form of a meshwork of cells showing here and there a tendency to form a short segment of endothelial tube. Fusion at such a stage has been described only for the chick.

For the mammalian embryo it has been generally accepted, since the pioneer work of Hensen (1876), that the cardiac primordia are bilateral and that fusion occurs after they have taken on the form of distinct endothelial tubes with well-defined nonloculated lumina. The mechanism of approximation of the tubes is not the same in various animals, and their spatial relationships are dependent upon two factors: (1) the form of the foregut at its inception, and (2) the position occupied by the endothelial heart tubes in the floor of the lateral coelomic channels. The operation of these factors is seen in an examination of the conditions obtaining in the rabbit on the one hand and in the guinea-pig on the other. In the rabbit, the foregut when newly formed is very broad. Further, the endothelial heart tubes appear high up in the medial wall of the lateral coelomic channel. When closure of the foregut takes place the endothelial heart tubes are thus situated far apart, each being related to the lateral part of the gut floor. The fusion of the coelomic folds leads to a septal formation below the foregut. This cannot, however, be regarded as a ventral mesocardium, since the endothelial tubes are placed well on either side of the midline and preserve their independence in respect of both endocardial and myo-epicardial layers. Secondarily this septum ‘gives way’ and the coelomic channels communicate across the midline beneath the foregut to constitute a pericardial cavity. Into this pericardial cavity the endothelial heart tubes project from the lateral part of the pharyngeal floor. There is no indication of a dorsal mesocardium at this stage. Only secondarily do the cardiac tubes approach each other, fuse, and become suspended by a dorsal mesocardium.


The guinea-pig (Strahl & Carius, 1889), on the other hand, manifests certain differences owing to the fact that the endocardial tubes make their appearance in the floor of the lateral coelomic channel rather than in its medial wall. As a result, when these walls come together a septum is formed uniting the floor of the foregut to the ‘middle cardiac plate’ which intervenes between the floors of the coelomic channels of the two sides. Narrowing of the middle cardiac plate occurs as the endocardial tubes approach one another and come into apposition. Secondarily, the coelomic spaces break through from side to side beneath the pairing tubes. From this account it appears that it is not quite accurate to say that the enclosure of the foregut is the cause of the union of the heart anlagen, for this closure may yet leave a considerable interval between the two tubes.

Yoshinaga (1921) has reinvestigated the process of cardiogenesis in the guinea-pig and draws attention to the resemblances between the cardiac development seen in an embryo with three pairs of somites and that found in the Glaevecke embryo of von Spee (1889). In the former, however, the angioblasts are already beginning to group themselves into endothelial tubes at certain points, and these groupings are producing localized bulgings of the covering splanchnopleure into the lateral coelomic channels. Although endothelial tubes are by no means fully developed in this embryo, and certainly much less so than in most mammals at this stage of short foregut diverticulum, yet they are much in advance of the conditions found in the Glaevecke specimen. Vasculogenesis in this latter is, indeed, so early that to von Spee himself it seemed virtually absent: ‘Gefassendothelréhren fanden sich nirgends in der Keimscheibe. Auch an der Stelle, wo man die primitiven Herzanlagen erwarten wiirde, fanden sich ganze vereinzelte Zellen, die aus dem festeren Continuitaétsverbande der Keimblatter (vielleicht in Folge einer Verletzung) ausgetreten zwischen Mesoderm und Entodermlagen.’ H. M. Evans (1912), however, interpreted these scattered cells as cardiac anlagen. The condition observed in the embryo ‘Glae’ are much more like those seen in the sheep specimen B 15 described above. As in the embryo ‘Glae’, B15 nowhere shows endothelial tube (or even vesicle) formation, vessel rudiments being represented merely by scattered angioblasts. Pl. 1, fig. 5 is quite comparable with Evans’s fig. 404, in Keibel and Mall. At a rather more advanced stage, also, the sheep resembles the human more closely than does the guinea-pig. For in a guinea-pig embryo with four pairs of somites (Yoshinaga, p. 270) union of the paired cardiac rudiments in front of the anterior intestinal portal is by angioblasts ‘ underneath the flat splanchnopleural folds of the cranio-median limb of the pericardial cavity’, whilst ‘the endothelial tubes are differentiated at great length, 84

extending throughout nearly the whole extent of the lateral pericardial cavity’. Although ‘these tubes are irregularly interrupted, their continuity (is) bridged by angioblast cords’. Yoshinaga pictures (fig. 13) these endothelial tubes in a considerably more advanced stage of continuity than his text indicates, and certainly more so than is to be found in the corresponding sheep embryo B 47L. in which the first union is occurring. Here fusion is occurring between vesicular strands (Pl. 2, fig. 9), and not simple angioblast chains as in the guinea-pig. However, the resemblances between the guinea-pig and sheep cardiac primordia at the time of fusion are notable, especially when comparison is made with the processes described in other mammals such as the rabbit, cat, dog and ferret, for in the latter group well-defined endothelial tubes are formed, and it is between these that fusion takes place.

Perhaps the most striking comparisons, however, are those between the sheep and human hearts round about the time of fusion of the paired primordia. Cardiogenesis in the Sternberg (1927) embryo (four pairs of somites) appears to be comparable in many respects with that in the sheep specimens B47L and B52A. In the Sternberg embryo (p. 161) the heart is represented by ‘einem Netz von regellos angeordneten Strangen, in welchen zahlreiche Hohlraume ausgebildet sind... und liegt an der Ventralseite...des...Vorderdarmes, zwischen diesem und der... Perikardialhéhle’. From the lateral parts of this cardiac network vessel anlagen pass backwards on either side of the anterior intestinal portal—vitelline veins—and Sternberg suggests that these will be progressively taken up in the formation of the endocardial tube as the intestinal portal moves caudally. From the cranial end of the subpharyngeal network the anlage of the first aortic arches can be traced forwards in much the same way apparently as in the model (Text-fig. 2). The actual character of the ‘Herzplexus’ is very similar to that of the sheep specimen, being made up of contiguous and partially continuous angiocysts closely applied to the foregut entoderm (cf. Abb. 16, p. 163).

Davis (1927), in his review of the early stages of development of the human heart, describes conditions in embryos of two and four somites which show close resemblances to those in the sheep. However, there are one or two differences to be noted. Thus Davis describes (p. 257) the endocardium as beginning ‘caudally. the extremity of the atrium, as a blind endothelial channel, which abuts against, but apparently is not connected with, the primordial umbilico-vitelline channel extending caudally therefrom’. No such remarkable discontinuity has been found in the sheep material, and would in any case be very difficult to establish with certainty. Further, at this early stage when


the heart plexus is yet ‘a flattened sheet of endothelial strands’ with ‘vesicles, or angiocysts, more or less fused with one another’, there are already demarcations in the myo-epicardial mantle giving some indication of the future atrial, ventricular and bulbar regions. In the sheep these early divisions are not to be found. In the Ingalls twosomite embryo the conditions appear to be very similar to those obtaining in the sheep specimen (Ingalls, 1920, Pl. 3 and present Text-fig. 2). However, as Davis says (p. 261), the Ingalls embryo shows a ‘general tubular endocardium’ at the two extremities of the heart whilst the intervening part retains the ‘primitive plexiform arrangement’. This state of affairs cannot be substantiated for the sheep. ‘

It appears then that in the guinea-pig, man and sheep, fusion of the bilateral cardiac anlagen occurs at a time when their condition has not progressed beyond the stage of vesiculated strands. The causal factor may be the early closure of the foregut with consequent approximation of the rudiments; and it is possible that this early closure may itself be referable to the comparatively small size of the yolk sac (Bonnet, 1884).

The presence of a ventral mesocardium (Pl. 4, fig. 21) in relation to the caudal part of the heart at once raises the question of the mechanism of closure of the foregut; for it was on the alleged absence of such a structure that Robinson (1902)

_in large part based his denial of the fusion of the

lateral coelomic folds.

Bonnet-Peter (1929) shows a section through a sheep embryo with eight pairs of somites resembling precisely the above specimen B 16L, and he, too, indicates a mesocardium ventrale (fig. 319). In his text he suggests that it is formed by the apposition of the lateral coelomic folds. Without longitudinally sectioned specimens it is not possible to arrive at a definite opinion on this question, but it may be said that the appearances found in the embryos described above are strongly indicative of such fusion. So far as the human embryo is concerned the existence of a ventral mesocardium is strongly denied by Davis (1927). Opinions to the contrary are brushed aside by both Davis and Wang (1917), and the latter offers a reinterpretation of Wilson’s (1914) finding of a small ventral mesocardium in a human embryo of probably three somites. Wang further records appearances identical with those shown by Wilson in Johnstone’s second embryo. However, he prefers not to apply the name ‘ventral mesocardium’ to the band of tissue binding down the heart to the septum transversum, but merely to call it part of the septum transversum because the vitelline veins pass through it. Wang’s fig. 85, nevertheless, shows a distinct endothelial tube surrounded by a muscular heart tube, and it is difficult to see why the band of tissue in question Development of the sheep heart

should not be looked upon as a ventral mesocardium. In the sheep material described above the band of splanchnopleure shown in PI. 4, fig. 21 connects a well-defined unpaired heart rudiment with the developing septum transversum and appears to be a true mesocardium. .

The validity of such minute distinctions as Wang seems bent upon making is further called into question by the difficulty of demarcating precisely the limits between vitelline vein behind and ‘heart’ in front. Thus Watson (1924) points out that Wang uses the term ‘vitelline vein’ for even the anterior portion of the structures commonly called lateral heart tubes ‘...and also applies the term to structures which I described as lateral heart tubes in Dasyurus’.

The existence of a ventral mesocardium would seem to be established for the sheep but still doubtful in the case of other mammals including man. However, the general resemblances in cardiac development in the sheep and in man are so striking that it would be surprising if further study did not establish at least a small and transient ventral mesocardium in the latter.


1. An account is given of the mechanism of extra-embryonic vasculogenesis in the sheep.


2. The process of cardiogenesis is traced from its first stages up to the formation of an unpaired heart rudiment, and is found to show differences from that described for many mammals (such as the rabbit, cat and dog) but striking similarity to that which takes place in the guinea-pig and more especially in man.

8. The origin of vasoformative cells is discussed with reference to their ‘germ-layer’ origin.

4. The existence of a ventral mesocardium is established, and its significance indicated.

It is with pleasure that the author records his gratitude to Profs. W. J. Hamilton and J. D. Boyd for access to the sheep material made available to them by the help of a research grant from the University of London. To Prof. Hamilton he would also express his indebtedness for constant encouragement and helpful criticism, to Mr A. E. Westwood for his skill and patience in preparing the photographs, and to Mr E. J. Park for his share in the preparation of the sections. The expenses of microphotography have been met by a grant from the Thomas Smythe Hughes Medical Research Fund of the University of London, for which the author expresses his thanks.


Bonnet, R. (1884). Arch. Anat. Physiol. (Anat. Abt.), p. 170.

Bonnet, R. (1889). Arch. Anat. Physiol. (Anat. Abt.), p.1

Bonnet, R. (1901). Anat. Hefte, 16, 233.

Bonnet-PErsER (1929). Lehrbuch der Entwicklungsgeschichte, 5th ed. Berlin.

CuaRKkE, W. C. (1914). Anat. Rec. 8, 95.

CiaRKke, W. C. (1916). Anat. Rec. 10, 301.

Davis, C. L. (1927). Contr. Embryol. Carneg. Instn, 19, 245.

Duvat, M. (1897). J. Anat., Paris, 38, 1.

Evans, H. M. (1912). Manual of Human Embryology (Keibel and Mall), 2, 498. Philadelphia and London. Guapstonz, R. J. & Hamiton, W. J. (1941). J. Anat.,

Lond., 76, 9. Hensen, V. (1876). Arch. Anat. Physiol. (Anat. Abt.),

p. 1. Hountineton, G. 8. (1911). Mem. Wistar Inst., no. 1.

Incas, N. W. (1920). Contr. Embryol. Carneg. Instn, 11, 61.

Kerset, F, (1894). Morph. Arb. 3, 1. : Kuuuemann, J. C. (1754). Observationes quaedam circa negotium generationis in ovibus factae, 2nd ed. Lipsiae.

LEVANDER, G. (1945). Nature, Lond., 155, 148. Marri, P. (1902). Lehrbuch d. Anatomie d. Haustiere, 1. Stuttgart. ,

Minter, A. M. & McWuorter, J. E. (1914). Anat. Rec. 8, 203.

Mo.urer, S. (1906). Handb. d. vergl. u. exp. Entwick. Wirbeltiere (O. Hertwig), 1, 1020 and 1261.

Rosryson, A. (1902). J. Anat. Physiol. 37, 1.

Rovuvibre, H. (1904). J. Anat., Paris, 40, 610.

Samuzt, D. M. (1944). J. Anat., Lond., 78, 173. .

Scuutts, H. von W. (1914). Mem. Wistar Inst., no. 3.

Srze, F. von (1889). Arch. Anat. Physiol. (Anat. Abt.), p. 159.

STEENBERG, H. (1927). Z. ges. Anat. 1.Z. Anat. EntwGesch. 82, 142.

StooKarp, C. R. (1916). Amer. J. Anat. 18, 227.

Srrant, H. & Cartus, F. (1889). Arch. Anat. Physiol. (Anat. Abt.), p. 231. '

Stricut, O. van D. (1899). Bull. Acad. Méd. Belg. 18, 336.

SrupniéKa, F. K. (1911). Anat. Anz. 40, 33.

Sziny, A. von (1903). Anat. Anz. 24, 417. .

TELLYESNICZEY, K. (1898). Arch. mikr. Anat. 52, 202.

TELLYESNIGZEY, K. (1902). Arch. mikr. Anat. 60, 681.

TELLYESNICZEY, K. (1905). Arch. mikr. Anat. 66, 367.

Wana, C. C. (1917). J. Anat., Lond., 52, 107.

Watson, K. M. (1924). J. Anat., Lond., 58, 105.

Wuson, J. T. (1914). J. Anat. Physiol. 48, 315.

Yosninaaa, T. (1921). Anat. Rec. 21, 239. 86



Puate 1

Fig. 1. A photomicrograph of a well-defined vesicular space or primary angiocyst situated between the mesoderm above and the entoderm below. Embryo B15, 15 days old. Magnification x 1300, reduced to x 980 for reproduction.

Fig. 2. A photomicrograph of a well-defined secondary angiocyst, the wall of which is formed entirely from mesoderm. To the left of the vesicle a mesoderm cell is in mitosis. The cells of the vesicle wall are plump and have not yet become flattened by the accumulation of fluid within the angiocyst. The entoderm is seen as the lower layer and is excluded from the angiocyst wall. Embryo B15, 15 days old. Magnification x 1300, reduced to x 980 for reproduction.

Fig. 3. A photomicrograph of an angiocyst at a later stage of development. The increase in fluid within the vesicle has resulted in the cells of the vesicle wall assuming a typical flattened endothelial form. The entodermal layer of cells is seen below. Embryo B47L, 15 days old. Magnification x 1300, reduced to x 980 for reproduction.

Fig. 4. A section taken along the line AB in Text-fig. 1 (embryo B 15, 15 days old). The lateral coelomic channels are seen on either side just caudal to their emergence from the lateral parts of the cranio-median limb of the coelom. Four angioblasts are seen in relation to the entoderm. There is here no evidence of a ‘cardiogenic plate’. Magnification x 200, reduced to x150 for reproduction.

Fig. 5. A section taken along the line CD in Text-fig. 1. The foregut diverticulum is seen in the midline apparently formed by the fusion of the lateral coelomic folds. Several mitotic figures are to be found here. In relation to the lower part of the foregut floor there are a few. scattered angioblasts on either side, and further laterally some are also found below the lateral coelomic channels. Embryo B15, 15 days old. Magnification x 200, reduced to x 150 for reproduction.

Fig. 6. A section through embryo B 15 (15 days old) taken caudal to the anterior intestinal portal. On either side a string of angioblasts is to be seen ‘pavementing’ the entoderm. Between the latter and the splanchnopleure of the lateral coelomic channels there is a wide shrinkage space. Magnification x200, reduced to x150 for reproduction. :

PLatE 2

Fig. 7. A section through embryo B 15 to show the intimate relationship of angioblasts to entoderm. A strand of angioblasts is seen ‘pavementing’ the entoderm in the lower part of the figure. The orientation of the angioblasts parallel to the entoderm is shown, as is also the characteristically ‘radial’ disposition of the splanchnopleure cells forming the floor of the lateral coelomic channel. The shrinkage space between entoderm and

splanchnopleure is partially invaded by protoplasmic processes from the cells of the latter. Magnification x 520, reduced to x 390 for reproduction.

Fig. 8. A photomicrograph illustrating the intimate relationship between the entoderm and a strand of angioblasts. Protoplasmic continuity seems very probable. Embryo B47L, 15 days old. Magnification x 1300, reduced to x 980 for reproduction.

Fig. 9. A section through embryo B 47L (15 days) just anterior to the anterior intestinal portal. Angiocystic strands are seen in relation to the entoderm of the pharyngeal floor, and in the midline below there is a well-defined vesicle. Such are the first median cardiac rudiments. Magnification x 200, reduced to x 150 for reproduction.

Fig. 10. A section through embryo B 47L a short distance caudal to the anterior intestinal portal. Angiocystic strands are seen on either side adjacent to the entoderm (cf. Pl. 1, fig. 6). Magnification x 200, reduced to x 150 for reproduction.

Fig. 11. Part of the angiocystic strands seen in fig. 10 at a higher magnification. The character of the strand is shown. The vesicles are joined by angioblasts, and the whole chain is closely related to the entoderm. Magnification x 520, reduced to x 390 for reproduction.

Fig. 12. A section through the anterior part of embryo B 52B (16 days). Below the foregut diverticulum is a ‘keel-like’ heart projecting down into the underlying pericardial cavity. Magnification 200, reduced to x 150 for reproduction. ,


Fig. 13. A section through embryo B 52A (16 days) just caudal to the anterior intestinal portal. The vitelline vein formations are seen on either side, the overlying splanchnopleure being heaped up into the lateral coelomic channel in each case. The mesostroma between the endothelial channels and the splanchnopleure is well defined. Magnification x200, reduced to x150 for reproduction. ,

Fig. 14. A section through embryo B 52A (16 days) in front of the anterior intestinal portal. The vesicular ‘subpharyngeal lake’ is seen in relation to the pharyngeal entoderm. The mesostroma of the myo-epicardial mantle is prominent. Magnification x 200, reduced to x 150 for reproduction.

Fig. 15. A high-power view of the heart rudiment shown in fig. 14. The intimate relationship of the endocardium to the entoderm (on the right) is shown, as is also the character of the mesostroma. Magnification x 520, reduced to x 390 for reproduction.

Fig. 16. A section through embryo B 52A (16 days) just in front of the anterior intestinal portal. The two lateral coelomic channels are separated by a’ band of tissue—a ventral mesocardium. Magnification x 200, reduced to

x 150 for reproduction. Journal of Anatomy, Vol. 80, Part 2 Plate 1

FIELD—THE EARLY DEVELOPMENT OF THE SHEEP HEART Plate 2 Journal of Anatomy, Vol. 80, Part 2

FIELD—THE EARLY DEVELOPMENT OF THE SHEEP HEART Journal of Anatomy, Vol. 80, Part 2 Plate 3

FIELD—THE EARLY DEVELOPMENT OF THE SHEEP HEART Plate 4 Journal of Anatomy, Vol. 80, Part 2

FIELD—THE EARLY DEVELOPMENT OF THE SHEEP HEART Development of the sheep heart 87

Fig..17. A section through embryo B 16R (16 days) taken immediately caudal to the anterior intestinal portal. The few cells between the coelomic folds in the lower part of the field indicate commencing fusion. Welldefined vitelline veins are present on either side, projecting into the lateral coelomic channels. Magnification x 200, reduced to x 150 for reproduction.

Puate 4

Fig. 18. A section through embryo B 16R (16 days) taken just in front of the anterior intestinal portal. The vitelline veins are in the process of fusing, and below there is a ventral mesocardium. Magnification x 200, reduced to x 150 for reproduction.

Anatomy 80

Fig. 19. A section through embryo B16R showing the

subpharyngeal heart rudiment at its fullest development in this specimen. Note the wide extent of the ‘lake’ (cf. Text-fig. 3). Magnification x 200, reduced to x 150 for reproduction. *

Fig. 20. A high-power view of part of fig. 19 showing the

endocardium in close relation to the foregut floor (above). The remains of contiguous walls of angiocysts project into the lumen of the endocardium like baffle-plates. Magnification x 520, reduced to x 390 for reproduction.

Fig. 21. A section through embryo B16L (16 days)

showing a well-defined ventral mesocardium connecting the heart to the septum transversum. Changes in surface contour of the heart are beginning. Magnification x 200, reduced to x 150 for reproduction.