Experimental investigation of the endoderm in early stages of development shows that the endodermal organ rudiments, like those derived from other germinal layers, are not initially determined, but that the endodermal cells destined to participate in the formation of the various organ rudiments are no more determined for their respective fates than are the cells of the other germinal layers.

In the earlier stages of development, their fate is a function of the position that each cell or group of cells occupies in the embryo as a whole. This can be proved by isolating parts of the presumptive endoderm and cultivating them apart from the rest of the embryo or by transplanting them into an abnormal position.

In an extensive series of experiments, pieces of presumptive endoderm of young gastrulae were cultivated in the “Holtfreter solution.” Various tissues were observed to differentiate from such isolated pieces; some conformed to the normal destiny of the isolated parts, and some did not.

The range of differentiations included not only orobranchial epithelium, stomach epithelium, liver, pancreas, and intestine, but also notochord and muscle, which should not have developed from the presumptive en­doderm if it had kept its prospective significance.


Thus, the fate of the endoderm is not established finally in the early gastrula stage. Much greater deviations from the prospective significance of the various endodermal parts could be observed when these parts were placed in surroundings which, unlike the saline solution, could actively influence the differentiation of these parts.

In the early neurula stage, it is possible to separate the entire endoderm of a newt embryo from the ectoderm and mesoderm. The endoderm is removed as a whole through a slit on the ventral side of the embryo, leaving the ectoderm and mesoderm as an empty shell.

The isolated endoderm can then be inserted again into the ectomesodermal shell of the same embryo or of another embryo of the same species, or even into the ectomesodermal shell of an embryo of another species. The endoderm of the small Triturus taeniatus has been successfully implanted into the ectomesodermal shell of the larger Triturus alpestris.

The implantation may be carried out so that the orientation of the endoderm is in harmony with the orientation of the ectomesoderm, or the endoderm may be implanted in an inverted position. In the first case, a completely normal larva has been observed to develop.


A normal embryo also developed if the endoderm was implanted with its dorsoventral orientation reversed. This result shows that the determination of the dorsal and ventral parts in the endoderm is not fixed in the endoderm itself but is imposed on the endoderm by the surrounding ectomesoderm.

As we already known that transplantation of the dorsal lip of the blastopore (primary organizer) the endoderm was often observed to develop a secondary lumen of the mid-gut, just underneath the notochord developed from the transplanted organizer. This secondary lumen was, of course, part of the dorsal differentiation of the endoderm.

However, if the anteroposterior axis of the endoderm was inverted with respect to the axis of the ectomesodermal shell, the development was highly abnormal, thus showing that the differentiation of the endoderm along the anteroposterior axis cannot be dominated by the ectomesoderm.

Small pieces of endoderm taken from a late gastrula or nearly neurula stage were implanted in various positions into another embryo. When the pieces of endoderm were taken from embryos in the gastrula stage, the grafts were often smoothly incorporated into the endoderm of the host.


The use of heteroplastic transplantation made it possible to distinguish the grafted cells from the host cells (by differences in cell size in grafts between Triturus taeniatus and Ambystoma mexicanum) and thus to make sure that the graft was not destroyed but had fitted into the construc­tion of local tissues. Thus, presumptive orobranchial endoderm was found to be able to develop into intestinal epithelium and vice versa.

Stomach epithelium was developed from endoderm having a different prospective significance. Occasionally, however, grafts differentiated out of harmony with their surroundings, and the later the stage of the embryo from which the graft was taken, the more often this occurred. After the end of neurulation the grafts differentiated, in the main, according to their prospective significance.

Similar results were obtained when different parts of the neurula endoderm were transplanted into parts of the ectomesodermal shell, either from the anterior half of the neurula or from the posterior half. The endoderm taken for this experiment was either part of the foregut endoderm, mainly destined to become pharynx, or endoderm from the mid-gut, normally differentiating as stomach and intestine. It was found that mid-gut endoderm grafted into the anterior ec­tomesoderm produced pharynx (in addition to other parts).

Foregut endoderm sur­rounded by posterior ectomesoderm was in part differentiated as intestine. In both cases, endoderm produced parts which were not in accord with the prospective significance of the endodermal cells, and it seems plausible that these differentiations were induced by the adjoining mesoderm. Again we find that the endoderm, as well as the ectoderm, is dependent on the mesoderm in its differentiation.

There is, as yet, very little information concerning the earliest determination of endodermal organs in higher vertebrates. In birds, some information has been derived from experiments in which parts of the chick blastoderm were grafted to the chorioallan­toic membrane of another chick embryo.

Various endodermal organs were observed to differentiate from the grafts, namely, pharyngeal epithelium, thyroid, lung, liver, and large and small intestine. However, these tissues developed without a very definite relationship to the origin of the grafts.

It has been concluded that in itself the endoderm has a very low power of differentiation. Liver and thyroid usually differentiate in explants which also show the presence of the heart, and intestine is accompanied by mesoderm forming coelomic spaces. This broadly corresponds to what has been found in the amphibian embryo.

The epiblast alone, without the hypoblast, when cultivated on the chorioallantois, produces various endodermal tissues, such as thyroid, liver, pancreas, and intestine, almost in the same way as a whole blastoderm consisting of both epiblast and hypoblast. This is in agreement with the origin of definitive endoderm from the epiblast.

Only when the endodermal gut becomes separated from the yolk sac during the second day of incubation does the differentiation of endodermal explants correspond to their prospective significance.


Earlier work on mammalian embryos, employing the method of explantation of parts of the blastoderm, did not yield very clear results. To date, there is no overall picture of the early determination of the endoderm as a whole in mammalian embryos. Some more recent experiments are concerned with the development of particular parts of the alimentary tract.

In mouse and rat embryos the determination of the pancreas was studied by the method of explantation in a culture medium. Normally the (dorsal) pancreas appears as an evagination on the ninth day of gestation when the gut is already closed into a tube and has become segregated from the yolk sac.

A piece of gut, including the presumptive material of the pancreas, if explanted on the eighth day, will differentiate, producing pancreatic tissue, provided that the adjoining mesodermal tissue is explanted together with the endoder­mal epithelium. Pieces of gut from younger embryos, in their seventh day of gestation (embryos with five pairs of somites, cannot develop pancreatic tissue, although they produce liver, lung, stomach, and intestinal tissues.

Presumptive pancreatic endoderm explanted without mesoderm fails to differ­entiate, however, even if taken from 11-day embryos in which the pancreas had already attained the stage of a pouch-like evagination. The mesoderm is thus necessary to promote the differentiation of pancreatic tissue.