18 Hans Spemann and Hilde Mangold cells is intercalated in its floor, in sharp contrast to the adjacent regions. This white strip is part of the cristatusimplant that was clearly recognizable from the outside in the living embryo before the neural folds closed(Fig 3). The anterior end of this strip is approximately at the point where the thickness of the neural tube decreases ratherabruptly; it opens to the outside shortly thereafter The strip is wedge shaped, with the pointed edge toward the outside; as a result, only the tapering ends of the cells reach the surface of the embryo(figs. 5 and 6)or the central canal at the short stretch where they border it Fig. 5. Um 8b. Crass sec- tion through middle third of the embryo (df Figs. 2 and 3). pr. Med., pI anday neural tube. The E 3 pr Med implant (light)isin the sec- At its posterior end, the cristatus strip reaches the blastopore, and it is continuous with a mass of cristatuscells that is located between the secondary neural tube and the mesoderm on one side, and the endoderm on the other(Fig. 6). Because of their position one would be inclined to consider these cells as endoderm; but in size they resemble more the mesoderm of the taeniatusembryo, with which they are associated. At any rate, this cell mass, which extends a bit farther rostrad, has reached its position by invagination around the blastoporal lip. There is yet another mass of cristatus cells still farther rostrad. It has the form of a thin plate underlying the anterior part of the induced neural tube, as far s it is closed; at its anterior end and at its sides, it coincides approximately with the edge of the tube. and at its posterior end, it extends to the ectodermal strip of the implant. This plate is incorporated in the normal taeniatus mesoderm(Fig. 4). It is not differentiated further into notochord or somites Fig 6 Um 8b. Cross sectionin the region ofthe blastopore(Bl )(d Figs. 2 and 3). pr: Med, primary meuraltube sec Med. secondary neura/tube. The implant (light) has severalcells in the secondary neural tube, with its main mass in the mesoderm (sec. Mes. crist ).100r Altogether, a rather substantial part of the implant remained in the ectoderm. This portion was greatly stretched in length; as a result, the circular white disk that was implanted has become a long narrow strip that turns inwards around the blastoporal lip. Shifting of cells in the surrounding epidermis may have played a role in these form changes; the extent to which this occurs would have to be tested by implantation of a marker of indifferent material. a piece from a region near the upper lip of the blastopore could handily be considered as suitable for this purpose. We know from earlier experiments(Spemann 1918, 1921)that convergence and stretching of the cell material occurs at the posterior part of the neural plate. It is improbable that the cells of the neural plate are entirely passive in this process; rather, they may have aninherent tendency to shift that perhaps has been, togetherwith other characteristics, induced by the underlying endo-mesoderm. This tendency would be retained by
18 Hans Spemann and Hilde Mangold Fig. 6. Um 8b.Cross section in the region of the blastopore (Bl.) (cf. Figs. 2 and 3). pr. Med., primary neural tube; sec. Med., secondary neural tube. The implant (light) has several cells in the secondary neural tube, with its main mass in the mesoderm (sec. Mes. crist.). 100X. At its posterior end, the cristatus strip reaches the blastopore, and it is continuous with a mass of cristatus cells that is located between the secondary neural tube and the mesoderm on one side, and the endoderm on the other (Fig. 6). Because of their position one would be inclined to consider these cells as endoderm; but in size they resemble more the mesoderm of the taeniatus embryo, with which they are associated. At any rate, this cell mass, which extends a bit farther rostrad, has reached its position by invagination around the blastoporal lip. There is yet another mass of cristatus cells still farther rostrad. It has the form of a thin plate underlying the anterior part of the induced neural tube, as far as it is closed; at its anterior end and at its sides, it coincides approximately with the edge of the tube, and at its posterior end, it extends to the ectodermal strip of the implant. This plate is incorporated in the normal taeniatus mesoderm (Fig. 4). It is not differentiated further into notochord or somites. Altogether, a rather substantial part of the implant remained in the ectoderm. This portion was greatly stretched in length; as a result, the circular white disk that was implanted has become a long narrow strip that turns inwards around the blastoporal lip. Shifting of cells in the surrounding epidermis may have played a role in these form changes; the extent to which this occurs would have to be tested by implantation of a marker of indifferent material. A piece from a region near the upper lip of the blastopore could handily be considered as suitable for this purpose. We know from earlier experiments (Spemann 1918, 1921) that convergence and stretching of the cell material occurs at the posterior part of the neural plate. It is improbable that the cells of the neural plate are entirely passive in this process; rather, they may have an inherent tendency to shift that perhaps has been, together with other characteristics, induced by the underlying endo-mesoderm. This tendency would be retained by Fig. 5. Um 8b. Cross section through middle third of the embryo (cf. Figs. 2 and 3). pr. Med., primary neural tube; sec. Med., secondary neural tube. The implant (light) is in the secondary neural tube. cells is intercalated in its floor, in sharp contrast to the adjacent regions. This white strip is part of the cristatus implant that was clearly recognizable from the outside in the living embryo before the neural folds closed (Fig. 3). The anterior end of this strip is approximately at the point where the thickness of the neural tube decreases rather abruptly; it opens to the outside shortly thereafter. The strip is wedgeshaped, with the pointed edge toward the outside; as a result, only the tapering ends of the cells reach the surface of the embryo (Figs. 5 and 6) or the central canal at the short stretch where they border it
Induction ofembryonic primordia by implantation oforganizers from a different species 19 the piece in the foreignenvironment. Inthis way wemight also explainthe fact that the piece gains contact with the invaginating region of the normal blastoporal lip, although it was originally far distant fromit. Once it has arrived there by active stretching, it could be carried along, at least in part, by local cell shiftings Whereas this posterior cell mass is continuous with the cell strip that has remained on the surface it is separated from the more anterior cristatus cell plate by taeniatus mesoderm. Therefore, this anterior plate that underlies the neural tube cannot have arrived at its position by invagination around the upper blastoporal lip; it must have been located in the deeper position from the beginning Undoubtedly it derives from the inner layer of the implant; hence it was originally just under the cristatuscells, some of which are now formed partly in the neural plate as a narrow strip, and others of which had migrated inside around the blastoporal lip. These displacements carried it along and brought it forward to such an extent that now its posterior margin is approximately level with the anterior end of the cristatus cell strip in the neural tube a Although a piece of presumptive neural plate taken from a region a little anterior to the actual nsplant would have become epidermis after transplantation to presumptive epidermis, this implant has resisted the determinative influences of the surroundings and has developed essentially according to its place of origin. Its ectodermal part has become part of the neural plate and the endo-mesodermal part has placed itself beneath it. Furthermore, not only did the implant assert itself, but it made the indifferent surroundings subservient to it and it has supplementeditself from these surroundings. The host embryo has develop a second neural plate out ofits own material, that is continuous with the small strip of cristatuscells and underlain by two cell plates of cristatusorigin. This secondary plate would not have arisen at all without he implant, hence it must have been caused, or induced, by it. There seems to be no possible doubt about this. However, the question to the way nwhich the induction has taken place. In the present case it seems to be particularly plausible toassume a direct influence on the part of the transplant. But even under this assumption, there are still two possibilities open. The ectodermal component of the transplant could have self-differentiated into the strip of neural plate, and could have caused the differentiation of ectoderm anterior and lateral to it progressively to formneural tissue Orthe determination could have emanated fromthesubjacent parts ofthe endo-mesodermand have influenced both the cristatusand taeniatuscomponents of the overlying ectoderm in the same way. And finally, it is conceivable that the subjacent layer is necessary only for the first determination, which thereafter can spread in the ectoderm alone. A decision between these possibilities could be made if it were possible to transplant successfully pure ectoderm, and pure endo- mesoderm from the region ofthe upper lipofthe blastopore, and, finally, suchectoderm which had been underlain by the endo-mesoderm. In such experiments, heteroplastic transplantation offers again the inestimable advantage that one can establish afterwards with absolute certainty whether the intended isolation was successful In our case, such a separation of the factors under consideration has not been accomplished Nevertheless it seems noteworthy that the induced neural plate is poorly developed in its posterior part where it is in closest and most extensive contact with the ectodermal part of the transplant; and, in contrast, that it is well developed at its anterior end where it is remote from the cristatuscell strip, but underlain by the broad cristatus cell plate We shall discuss later a second possibility of a fundamentally different nature thatis particularly applicable tomore completely formedsecondary mbryonic primordia A second experiment, similar to the first, confirms it in all essential oints. They both have in common that the implant remains estodermal to a considerable extent, and therefore later forms part of the neural tube The situation is different in the following experiment Experiment Triton 1922, Um 25b. A median piece of the upper blastoporal lip was taken from a cristatus embryo at the beginning of gastrulation (sickle-shaped blastopore). It came from directly above the Fig, 7. U m 25b. hetaenijatus margin of invagination and was implanted into a taeniatusgastrula of the On the is the primary opore. Twenty-two hours later, when the taeniatusembryo had completed tube. alf she secondarymeural same stage in the ventral midline at some dista nce from the future blast
Induction of embryonic primordia by implantation of organizers from a different species 19 the piece in the foreign environment. In this way we might also explain the fact that the piece gains contact with the invaginating region of the normal blastoporal lip, although it was originally far distant from it. Once it has arrived there by active stretching, it could be carried along, at least in part, by local cell shiftings. Whereas this posterior cell mass is continuous with the cell strip that has remained on the surface, it is separated from the more anterior cristatus cell plate by taeniatus mesoderm. Therefore, this anterior plate that underlies the neural tube cannot have arrived at its position by invagination around the upper blastoporal lip; it must have been located in the deeper position from the beginning. Undoubtedly it derives from the inner layer of the implant; hence it was originally just under the cristatus cells, some of which are now formed partly in the neural plate as a narrow strip, and others of which had migrated inside around the blastoporal lip. These displacements carried it along and brought it forward to such an extent that now its posterior margin is approximately level with the anterior end of the cristatus cell strip in the neural tube. Although a piece of presumptive neural plate taken from a region a little anterior to the actual transplant would have become epidermis after transplantation to presumptive epidermis, this implant has resisted the determinative influences of the surroundings and has developed essentially according to its place of origin. Its ectodermal part has become part of the neural plate and the endo-mesodermal part has placed itself beneath it. Furthermore, not only did the implant assert itself, but it made the indifferent surroundings subservient to it and it has supplemented itself from these surroundings. The host embryo has developed a second neural plate out of its own material, that is continuous with the small strip of cristatus cells and underlain by two cell plates of cristatus origin. This secondary plate would not have arisen at all without the implant, hence it must have been caused, or induced, by it. There seems to be no possible doubt about this. However, the question remains open as to the way in which the induction has taken place. In the present case it seems to be particularly plausible to assume a direct influence on the part of the transplant. But even under this assumption, there are still two possibilities open. The ectodermal component of the transplant could have self-differentiated into the strip of neural plate, and could have caused the differentiation of ectoderm anterior and lateral to it progressively to form neural tissue. Or the determination could have emanated from the subjacent parts of the endo-mesoderm and have influenced both the cristatus and taeniatus components of the overlying ectoderm in the same way. And finally, it is conceivable that the subjacent layer is necessary only for the first determination, which thereafter can spread in the ectoderm alone. A decision between these possibilities could be made if it were possible to transplant successfully pure ectoderm, and pure endomesoderm from the region of the upper lip of the blastopore, and, finally, such ectoderm which had been underlain by the endo-mesoderm. In such experiments, heteroplastic transplantation offers again the inestimable advantage that one can establish afterwards with absolute certainty whether the intended isolation was successful. In our case, such a separation of the factors under consideration has not been accomplished. Nevertheless it seems noteworthy that the induced neural plate is poorly developed in its posterior part where it is in closest and most extensive contact with the ectodermal part of the transplant; and, in contrast, that it is well developed at its anterior end where it is remote from the cristatus cell strip, but underlain by the broad cristatus cell plate. Fig. 7. Um 25b. The taeniatus embryo at the neurula stage. On the right is the primary and on the left the secondary neural tube. 20X. We shall discuss later a second possibility of a fundamentally different nature that is particularly applicable to more completely formed secondary embryonic primordia. A second experiment, similar to the first, confirms it in all essential points. They both have in common that the implant remains estodermal to a considerable extent, and therefore later forms part of the neural tube. The situation is different in the following experiment. Experiment Triton 1922, Um 25b. A median piece of the upper blastoporal lip was taken from a cristatus embryo at the beginning of gastrulation (sickle-shaped blastopore). It came from directly above the margin of invagination and was implanted into a taeniatus gastrula of the same stage in the ventral midline at some distance from the future blastopore. Twenty-two hours later, when the taeniatus embryo had completed
20 Hans Spemann and Hilde Mangold its gastrulation, the implant had disappeared from the surface, which looked completely smooth and normal. Another 24 hours later, the embryo had two neural plates whose folds were about to close. secondary neural plate starts from the same blastopore as the ry one; at first it runs parallel to he primary plate, adjacent to its left side, and then it bends sharply to the left (Fig. 7). Shortly thereafter, the embryo was fixed; the sections were cut perpendicular to the posterior part of the axial The primary neural tube is completely closed and separated from the epidermis; its optic vesicles are protruding. The notochord is separate down to its posterior end which becomes lost in the indifferer zone. Seven or eight somites are formed. ce. Ucs. crist. Fig. 8. Um 25b. Cross section in the middle third ofthe embryo (cf Fig. 7). In the figure the secondary neural tube is seen to the right ofthe primary tube. The implant (light)is in the right primary mesoderm(sec. Mes. crist. ) 100r The secondary neural tube is also closed and separated from the epidermis; anteriorly its walls are broad and its lumen is transverse (probably an indication of optic vesicles). It decreases in thickness posteriorly. In its anterior one-third, it is bent sharply to the left and is therefore at some distance from the primary neural tube: but more posteriorly, at its posterior two-thirds, it approaches the latter and eventually fuses with it. However, the lumina, as far as they are present, remain separate. This secondary neural tube is formed completely by taeniatuscells, that is, by material supplied by the host embryo. Cristatus cells, that is, material of the organizer, do not participate in its formation The implant has moved completely below the surface. Its most voluminous, anterior part is a rather typical mass located directly under the secondary neural tube(Fig. 8), between it and the large yolk cells of the intestine. Separate somites cannot be seen, but the contourofa notochord can be delineated n the anterior sections, where the axial organs curve outward, itis cut longitudinally, but transversely in the more posterior ones( Fig 8). Toward its posterior end, the implant tapers off; it forms only the notochord and a few cells that merge with the endoderm(Fig 9). Thereafter, the notochord disappears also, and the implant liesentirely in theendodermand forms the uppercoveringofasecondaryintestinal lumen that extends over a few sections. In its entire posterior part, the implant is separated from the secondary neural tube by interposed mesoderm of the taeniatus embryo(Fig. 9). The neural tube extends considerably farther caudal than the implant. Fig. 9. Um 25b. Cross section in the posterior third of the embryo (cf Fig. 7. The secondary neural tube is attached the left side(right in the figure)of the primary tube. The implant (ight) forms secondary notochord (ec Ch.). 100r
20 Hans Spemann and Hilde Mangold Fig. 8. Um 25b. Cross section in the middle third of the embryo (cf. Fig. 7). In the figure the secondary neural tube is seen to the right of the primary tube. The implant (light) is in the right primary mesoderm (sec. Mes. crist.). 100X. Fig. 9. Um 25b. Cross section in the posterior third of the embryo (cf. Fig. 7). The secondary neural tube is attached to the left side (right in the figure) of the primary tube. The implant (light) forms secondary notochord (sec. Ch.). 100X. its gastrulation, the implant had disappeared from the surface, which looked completely smooth and normal. Another 24 hours later, the embryo had two neural plates whose folds were about to close. The secondary neural plate starts from the same blastopore as the primary one; at first it runs parallel to the primary plate, adjacent to its left side, and then it bends sharply to the left (Fig. 7). Shortly thereafter, the embryo was fixed; the sections were cut perpendicular to the posterior part of the axial organs. The primary neural tube is completely closed and separated from the epidermis; its optic vesicles are protruding. The notochord is separate down to its posterior end which becomes lost in the indifferent zone. Seven or eight somites are formed. The secondary neural tube is also closed and separated from the epidermis; anteriorly its walls are broad and its lumen is transverse (probably an indication of optic vesicles). It decreases in thickness posteriorly. In its anterior one-third, it is bent sharply to the left and is therefore at some distance from the primary neural tube: but more posteriorly, at its posterior two-thirds, it approaches the latter and eventually fuses with it. However, the lumina, as far as they are present, remain separate. This secondary neural tube is formed completely by taeniatus cells, that is, by material supplied by the host embryo. Cristatus cells, that is, material of the organizer, do not participate in its formation. The implant has moved completely below the surface. Its most voluminous, anterior part is a rather atypical mass located directly under the secondary neural tube (Fig. 8), between it and the large yolk cells of the intestine. Separate somites cannot be seen, but the contour of a notochord can be delineated; in the anterior sections, where the axial organs curve outward, it is cut longitudinally, but transversely in the more posterior ones (Fig. 8). Toward its posterior end, the implant tapers off; it forms only the notochord and a few cells that merge with the endoderm (Fig. 9). Thereafter, the notochord disappears also, and the implant lies entirely in the endoderm and forms the upper covering of a secondary intestinal lumen that extends over a few sections. In its entire posterior part, the implant is separated from the secondary neural tube by interposed mesoderm of the taeniatus embryo (Fig. 9). The neural tube extends considerably farther caudal than the implant
Induction ofembryonic primordia by implantation oforganizers from a different species 21 In contrast to the first experiment, the implant in the present case forms a uniform mass; it is not separated into two sections by intervening mesoderm. This must have something to do with the way in which it was shifted to below the surface However nothing definite can be ascertained concerning this point. The fact that the two embryonic anlagen share the remainder of the blastopore proves that the implant has been invaginated in the normal way around the blastopore. However, itis doubtful whether the implant was entirely passive in this process. It comes from region whose cells normally participate actively in invagination; and in other instances they have retained this capacity after transplantation For this reason, the situation becomes complicated The implant has formed the entire notochord, the greater part of the mesoderm, which however is not typically segmented, and a small part of the intestinal primordium. Itis not clear in the present case whether it has also exerted an inductive effect on the adjacent mesoderm. However it has certainly evoked the formation of the entire secondary neural tube; but in which way this has occurred remains undecided. a direct influence would be possible in the anterior region where the implant lies directly under the neural tube(Fig 8). However this explanation is improbable farther back where the implant is displaced by host mesoderm(Fig. 9)or is entirely missing. One would have to assume that this mesoderm has been altered by the organizer and has, in turn, initiated the formation oftheneural plate in the overlyingectoderm However, it could be that the organizer hadexertedits entireeffect on the ectoderm before it had moved to the interior Insummary, itis characteristicofthis case thatimplant cellsare completely absent in the secondary neural tube, and that the notochord is formed completely by cells of the implant. The same thing is shown, perhaps even more beautifully, in another case (Triton 1922, Um 214), in which the notochord formed by the implant, and also the induced neural tube, extend almost over the entire length of the host embryo, and are both near the normal Fig. 10. Um 131a. The cristatus axial organs. But this case again fails to indicate whether the bryoat the neurulastag implant can form somites or induce them in host mesoderm. The plant(dark), in the si next case gives information on that point. gle with umegua/sides, lies in the paste rior dorsa/ half 20n Experiment Triton 1922, Um 131b. The exchange of material was done inadvanced gastrulae, after formation of the yolk plug. A large piece of cristatus, derived from the median line directly above the blastopore, was interchanged with a piece of taeniatuswhose origin could not be definitely determined The taeniatusimplant has not participated in invagination in the cristatusembryo; it has caused a peculiar fission(Fig. 10). The neural tube is closed anteriorly; at the point where it meets the taeniatus piece, it dividesinto two halves, one to the left and one to therigh At this point, a bit of endoderm comes to the surface, perhaps as the result of incomplete healing or of a later injury. The cross- sections show a neural tube and notochord in the anterior part back to the point of bifurcation The two divisions of the neural tube are still distinct for a few sections, but then they become indistinguishable from the surrounding tissue. The same is true to a greater degree, of the The taeniatusembryo has reached the neurula stage 20 hours later. The implant is located on the right side, somewhat behind the middle and next to the right neural fold. Its original anterior halfis still on the surface and stronglyelevated over the surround- ings; its original posterior half is invaginated and appears as a Fig.11.Um131b. Thetaeniatus embryo light area underneath the darker cells of the taeniatusembry the neurula stage. The neuralfolds The piece is stretched lengthwise and directed from posteriorly, dasing.The implant ( h ) in the middle and somewhat above, to anteriorly and somewhat downward nd pasterior third, to the right of the Invagination still continues; a half-hour later, astrip of cristatus dorsal median plane, is visible throug the surface layer, and continues into the cells is visible only at the outer margin of invagination. Twenty five hours later, the neural folds are almost closed; the implant is
Induction of embryonic primordia by implantation of organizers from a different species 21 In contrast to the first experiment, the implant in the present case forms a uniform mass; it is not separated into two sections by intervening mesoderm. This must have something to do with the way in which it was shifted to below the surface. However nothing definite can be ascertained concerning this point. The fact that the two embryonic anlagen share the remainder of the blastopore proves that the implant has been invaginated in the normal way around the blastopore. However, it is doubtful whether the implant was entirely passive in this process. It comes from a region whose cells normally participate actively in invagination; and in other instances they have retained this capacity after transplantation. For this reason, the situation becomes complicated. The implant has formed the entire notochord, the greater part of the mesoderm, which however is not typically segmented, and a small part of the intestinal primordium. It is not clear in the present case whether it has also exerted an inductive effect on the adjacent mesoderm. However, it has certainly evoked the formation of the entire secondary neural tube; but in which way this has occurred remains undecided. A direct influence would be possible in the anterior region where the implant lies directly under the neural tube (Fig. 8). However this explanation is improbable farther back where the implant is displaced by host mesoderm (Fig. 9) or is entirely missing. One would have to assume that this Fig. 10. Um 131a. The cristatus embryo at the neurula stage. The taeniatus implant (dark), in the shape of a triangle with unequal sides, lies in the posterior dorsal half. 20X. mesoderm has been altered by the organizer and has, in turn, initiated the formation of the neural plate in the overlying ectoderm. However, it could be that the organizer had exerted its entire effect on the ectoderm before it had moved to the interior. In summary, it is characteristic of this case that implant cells are completely absent in the secondary neural tube, and that the notochord is formed completely by cells of the implant. The same thing is shown, perhaps even more beautifully, in another case (Triton 1922, Um 214), in which the notochord formed by the implant, and also the induced neural tube, extend almost over the entire length of the host embryo, and are both near the normal axial organs. But this case again fails to indicate whether the implant can form somites or induce them in host mesoderm. The next case gives information on that point. Experiment Triton 1922, Um 131b.The exchange of material was done in advanced gastrulae, after formation of the yolk plug. A large piece of cristatus, derived from the median line directly above the blastopore, was interchanged with a piece of taeniatus whose origin could not be definitely determined. The taeniatus implant has not participated in invagination in the cristatus embryo; it has caused a peculiar fission (Fig. 10). The neural tube is closed anteriorly; at the point where it meets the taeniatus Fig. 11. Um 131b.The taeniatus embryo at the neurula stage. The neural folds are closing. The implant (light), in the middle and posterior third, to the right of the dorsal median plane, is visible through the surface layer, and continues into the protuberance. piece, it divides into two halves, one to the left and one to the right. At this point, a bit of endoderm comes to the surface, perhaps as the result of incomplete healing or of a later injury. The crosssections show a neural tube and notochord in the anterior part back to the point of bifurcation. The two divisions of the neural tube are still distinct for a few sections, but then they become indistinguishable from the surrounding tissue. The same is true, to a greater degree, of the notochord. The taeniatus embryo has reached the neurula stage 20 hours later. The implant is located on the right side, somewhat behind the middle, and next to the right neural fold. Its original anterior half is still on the surface and strongly elevated over the surroundings; its original posterior half is invaginated and appears as a light area underneath the darker cells of the taeniatus embryo. The piece is stretched lengthwise and directed from posteriorly, and somewhat above, to anteriorly and somewhat downward. Invagination still continues; a half-hour later, a strip of cristatus cells is visible only at the outer margin of invagination. Twentyfive hours later, the neural folds are almost closed; the implant is