Studies on Amphibian Metamorphosis: XV. Direct Tympanic Membrane Formation from Dermal Plicae Integument Transplanted to the Ear Region

The dermal plicae folds or ridges of the metamorphosed anuran constitute striking integumentary structures. In most anurans two sets of dermal plicae are present, viz. a dorso-lateral set which run posterior from the dorso-posterior border of each eye to a point just lateral to the urostyle prominence, and a lateral pair originating near the external nares of either side and thence running posterior along the upper jaw ventral to the tympanic membrane to a point just dorsal to the base of the fore limb. The folds, as the term indicates, appear externally as definite elevated ridges of which the dorsal-lateral pair are usually the more prominent. They are more often dull white in appearance although frequently of a somewhat greyish tinge, the latter condition being more typical of the lateral than of the dorso-lateral pair. Histologically, the folds are seen to be due chiefly to much enlarged mucous and poison glands, usually closely approximating one another. This necessarily requires a corresponding thickening of the stratum spongiosum of the integument, in which layer the glands are located. In addition, the stratum compactum is also considerably hypertrophied, while the epidermis usually contains more cell layers as compared with that of ordinary back integument.

The various histological developments leading to dermal plicae formation have previously been dealt with by Massie (1894) and also by the writer (1931a). Similarly, the question of whether or not the mucous and poison glands have separate or common origins has also been adequately discussed by many workers. In this regard the reader is referred to the papers of Szczesny (1867), Engleman (1872), Calmels (1883), Schultz (1889), Seeck (1891), Heidenhain (1893), Nicoglu (1893), Junius (1896), Ancel (1902), Fano (1903), Esterly (1904), Nirenstein (1908), Weiss (1908 and 1915), Wenig (1913), Dawson (1920), and of the writer (1931a). It is sufficient to state here that at the beginning of larval involution in the anuran there are no histological elements which distinguish potential dermal plicae integument from ordinary back or side integument. The areas of larval integument which give rise to dermal plicae structures during metamorphosis are, however, determined prior to the onset of involution. This fact was first demonstrated by the writer (1931a) in the following manner. Reciprocal autoplastic transplants were made between integument of dermal plicae regions and integument derived from the back, belly, or side of non-metamorphosing larvae. During metamorphosis, the potential dermal plicae integument developed typical dermal plicae structures, while the reciprocal grafts placed in dermal plicae regions failed to show any marked histological changes. Thus it was demonstrated that the potentialities for dermal plicae development are localized and determined in certain integumentary areas at some time prior to the onset of larval involution.

More recently, the writer (1933) has investigated the stability of fully differentiated dermal plicae following homoplastic and heteroplastic transplantation to non-metamorphosing larvae. Dermal plicae transplants obtained from young R. palustris frogs were transplanted to larvae of the same species, while in others the transplantations were made to R. catesbeiana tadpoles. In the former case no histological regression of dermal plicae structures occurred within 3 weeks of the transplantation period. Conversely, however, the heteroplastic transplants showed complete regression of dermal plicae structures within the same time interval. When allowed to remain long enough on the host, the heteroplastic transplants underwent complete integumentary degeneration. It is evident, therefore, that the R. palustris host larvae possessed the necessary influences (hormonal or otherwise) for the preservation of dermal plicae structures which the R. catesbeiana larvae apparendy lacked.

The fact that complete regression of fully differentiated dermal plicae structures could be obtained following heteroplastic transplantation gave rise to several interesting questions. Did such integument, for example, after having undergone complete dermal plicae regression back to normal integument, still possess the potentialities for redeveloping dermal plicae elements following suitable retransplantation ? The answer to this question was partially given in a later paper in which the writer (1934 a) showed that the retransplantation of such grafts to metamorphosing R. clamitans larvae failed to redifferentiate typical dermal plicae structures. A second and more interesting possibility presented itself, however, in that the original transplant after having undergone dermal plicae regression back to normal integument might be induced, following suitable transplantation, to differentiate an entirely different set of histological structures. The writer (1928, 1931 b) had previously shown that the tympanic membrane of the anuran is developed under the influences of the annular tympanic cartilage and columella. Furthermore, any larval integument when transplanted over these cartilaginous components of the middle ear will transform into tympanic membrane. Hence, the possibility of tympanic membrane differentiation from dermal plicae integument suitably retransplanted to the ear region, following the total regression of dermal plicae structures, was suggested. The results of such experiments (Helff, 1934a) showed quite clearly that it is possible to obtain tympanic membrane development in this manner. Briefly, the transplantations and results were as follows. Dermal plicae integu mentary grafts were removed from young R. palustris frogs and transplanted to non-metamorphosing R. catesbeiana larvae. Within 5−6 weeks complete macroscopic and histological regression of dermal plicae structures had occurred. The grafts were then retransplanted to the ear regions of metamorphosing R. clamitans and R. catesbeiana larvae, where they developed fully differentiated tympanic membranes within 5−7 weeks.

The above-mentioned results served to raise several other interesting questions. Of these, however, the question of whether or not dermal plicae transformation into tympanic membrane can be obtained following direct autoplastic, homoplastic, and heteroplastic transplantation of fully differentiated dermal plicae integument to the ear region of metamorphosing larvae, was of most interest to the writer. The results of experiments designed to answer this question and to throw light on various correlated problems furnish the material for the present paper.

Three species of metamorphosing larvae were available for the work in hand, viz. R. palustris, R. clamitans, and R. catesbeiana. They were collected during July and August in the vicinity of Lake Winnisquam, New Hampshire. The larvae were usually rapidly undergoing involution when collected, and were isolated in individual bowls in the laboratory until they had attained the proper degree of metamorphosis for the particular donor or host stage desired. For the homoplastic and heteroplastic series, the donors used for dermal plicae transplantation were newly metamorphosed frogs in which all larval characteristics had disappeared with the exception of small black tail stumps which usually persist for a week or more. The tail stumps of the R. palustris donors varied from 5 to 10 mm. in length, while those of the R. clamitans donors measured from 2 to 4 mm. In all other characteristics, however, the donors were typically froglike, and presented well-developed dermal plicae folds. The host animals of the three species used were in all cases involuting larvae well advanced in metamorphosis, and at a stage when the annular tympanic cartilage has nearly completed its posterior migration to the site of tympanic membrane formation. The autoplastic transplantations necessitated the use of a somewhat different stage of metamorphosis which can be roughly characterized as half-way between the stages used for the homoplastic and heteroplastic donors and hosts. Although the dermal plicae of such individuals were admittedly not as fully differentiated as those of the homoplastic or heteroplastic donors, the annular tympanic cartilages to which the dermal plicae grafts were autoplastically transplanted were still in a stage during which they exert strong tympanic membrane inductive influences. Had autoplastic transplantations been carried out with newly metamorphosed frogs, the conditions of the experiment would have been more seriously affected, since the annular tympanic cartilage at this stage has begun to lose some of its developmental influence as regards tympanic membrane formation (see Helff, 1934b).

The transplantation technique employed has previously been described by the writer (1934a) in some detail. Briefly, it consisted in the removal of a square piece of integument so as to include a portion of the dorso-lateral dermal plicae fold, the latter running through the centre of the graft. This was then transplanted over the annular tympanic cartilage of the host animal or of the same animal in the case of the autoplastic series, after first removing a square of integument from that region. The host larvae were then placed in individual aquaria and allowed to complete their metamorphosis. The young frogs were then killed and placed in Bouin’s fixative at various intervals following the attainment of complete larval involution, for the purpose of later sectioning and study of the dermal plicae transplants. Representative individuals were also killed and preserved in alcohol for photographic purposes. It may be stated here that Bouin fixed animals do not usually photograph well, due to the yellow coloration imparted by the fixative and the consequent lack of contrast so desirable for photographic purposes.

Thirty-two autoplastic R. palustris transplantations were made of which twentynine lived until the period of fixation. The transplants healed rapidly, and superficial union with the adjacent integument of the head region was always established within 4 days. As is usual, however, full union of the various integumentary layers with those of the adjacent integument required a much longer period of time. In general, the coloration of the transplanted skin remained quite normal. This was characteristic not only of the dermal plicae portion but also of those parts of the transplant adjacent to the dermal fold. The patches of melanophores so characteristic of back skin during the later stages of R. palustris metamorphosis and which give rise to the pigmentation pattern of the adult, were also maintained and in some cases even intensified in the transplants. In a few instances, however, the pigment spots were less distinct, due to a generalized darkening of the transplanted integument as a whole.

Externally, the dermal plicae folds of the transplants remained normal in all respects. Not only was the coloration unchanged, but, of more importance, their general shape, width, and height remained the same as when originally transplanted. In a few cases, however, the folds became somewhat greyish in tint, which fact was no doubt correlated with the generalized darkening of the graft in question. It is also interesting to note that although the transplants invariably fused completely with the surrounding head integument, the dermal plicae folds in no instance grew beyond the borders of the transplant.

All host animals were killed on the twenty-sixth day following the initial transplantation and either fixed in Bouin’s fixative or preserved in alcohol. Pl. I, fig. G, illustrates a typical example. Subsequent sectioning and study corroborated the macroscopic findings in that no regressive changes in the dermal plicae folds had taken place. Moreover, there were no degenerative changes evident in any part of the transplants which might be construed as representing even early signs of tympanic membrane formation. Apparently, therefore, transformation of dermal plicae structure into that of tympanic membrane is not possible following autoplastic transplantation in R. palustris-, at least, not under the conditions of the present experiment and within the time interval which was available for such transformation.

The autoplastic R. clamitans transplantations were made on metamorphosing larvae in the same stage of involution as selected for the R. palustris series. In such larvae the normal tympanic membrane is approximately half-formed, while the annular tympanic cartilage is still active as regards its ability to induce membrane formation. In all, thirty-one transplantations were made of which twenty-seven survived for at least 20 days following the autoplastic grafting of the dermal plicae integument to the ear region. The animals were killed for fixation or preservation in alcohol between the twentieth and thirty-second day following transplantation. As in the case of the previous series, fusion of the transplant with the surrounding head integument was achieved, externally at least, within a few days.

In fifteen of the twenty-seven cases, no appreciable diminution in height of the dermal plicae folds was noticeable, while their width and general contour remained quite normal (Pl. I, fig. A). The integumentary portions of the transplants adjacent to the dermal folds, however, in many cases gave slight external evidence of tympanic membrane formation in that the coloration underwent a change. Histological sections, however, gave little evidence of degenerative changes essential and normally prerequisite to the later developmental stages of membrane formation, while the component parts of the dermal folds were likewise found to be still intact.

Eight cases exhibited definite external signs of dermal plicae regression. These varied from slight reduction in height of the fold to a condition where the latter was barely discernible to the eye. A characteristic and proportionate widening of the fold accompanied its reduction in height (Pl. I, fig. B), while the transplants in general usually assumed a darker shade, although in a few cases a light brown tinge was evident. The histological findings revealed the dermal plicae elements in various stages of degeneration which were correlated quite closely with the degree of macroscopic fold reduction.

The remaining four cases gave external evidence of complete dermal plicae regression. In one of these, early signs of tympanic membrane formation were detected, while in another a half-formed membrane actually developed occupying the greater part of the transplant (Pl. I, fig. C). There were no histological signs indicating the persistence of dermal plicae elements in any of these four cases, while in the latter two cited definite cellular reorganizations were apparent typical of early membrane formation.

The results of the autoplastic R. clamitans series, therefore, were not as uniform as those obtained in the R. palustris series. Aside from the fact that species specificity evidently played a part in determining the two groups of results obtained, the differences in results exhibited between the twenty-seven cases of R. clamitans transplantation must be attributed to individual variations in the reactions of the dermal plicae folds to regressive influence or to variations in the inductive influences of the annular tympanic cartilages. Quite probably both of these factors were operative. The length of time during which regressive and inductive influences functioned in each case must also be taken into account. These factors and others will be taken up more fully in the discussion section of the present paper.

Twenty-eight of the thirty-four homoplastic R. palustris transplantations lived for more than 3 weeks following the grafting of dermal plicae integument over the annular tympanic cartilage. They were killed for fixation or alcohol preservation on either the twenty-sixth or thirty-second day following the transplantation procedure. The hosts were by this time fully metamorphosed young frogs in which even the tail stumps had been completely atrophied for at least a week.

The transplants appeared to heal, at their borders with the surrounding host integument, in about the same time interval as in the case of the autoplastic series. This rapidity of healing might be considered somewhat unusual in that homoplastic hosts, theoretically at least, should not offer as favourable an environment as when the transplants were grafted autoplastically. It should be borne in mind, however, that the host larvae in the homoplastic series were in an earlier and more rapidly involuting stage as compared with the autoplastic animals. Hence, the more favourable growth (healing) influences present probably were sufficient to counterbalance any antagonism towards healing which may have existed between the transplant and its homoplastic host.

The pigmentation pattern of the grafts in those regions adjacent to the dermal plicae folds remained quite normal, although it was not as pronounced as in the case of the autoplastic transplants on the same species. This was probably the result of the generalized darkening of the transplants which occurred, and which was more intense than was true of the autoplastic series. Usually, the transplants gave little evidence of superficial disintegration, although in two cases slight epidermal erosion was noticed, while in only one case were coloration changes recorded indicative of tympanic membrane development. Histologically, however, most transplants presented disintegrative changes similar to the cellular reorganizations which precede normal tympanic membrane formation.

The dermal plicae portions of the transplants presented a striking uniformity in their reactions. In all cases, without exception, a definite reduction in height of the folds took place, while the ridges became much broader and diffuse in appearance (Pl. I, fig. H). The folds also became less distinct to the eye, due no doubt to a certain amount of pigmentation which developed. This change, resulting in the loss of the characteristic whiteness of fully differentiated dermal plicae, was no doubt due to the same influences responsible for the generalized darkening of the transplants as a whole. Histologically, it was quite clear that the various components of the folds were undergoing rapid disintegration and regression. In no instance, however, could the histological picture be interpreted as presenting developmental changes typical of tympanic membrane formation. The results, therefore, would seem to indicate that homoplastic transplantation in R. palustris is conducive to regressive changes in dermal plicae structure, although transformation into tympanic membrane is not possible under the conditions and time limits of the experiment.

Thirty homoplastic R. clamitans transplantations of dermal plicae integument were made to the developing ear region. Of these, twenty-seven lived for at least 28 days following the transplantation procedure. All host animals were finally killed for fixation or preservation within 3 days of this period. The early history of the transplants followed quite closely that as described for the homoplastic R. palustris transplantations, especially as regards the rapidity of healing and the generalized darkening which occurred. In many cases, however, this initial increased pigmentation proved to be but a temporary condition in that its disappearance usually followed the early stages of tympanic membrane formation.

The fate of the dermal plicae portions of the transplants was strikingly definite and uniform, for in all except two cases, complete external and histological regression occurred. Externally, the folds were seen to first lose their characteristic whiteness and to assume a dull greyish tinge. Accompanying this coloration change, the folds were seen to gradually reduce their height and increase in width, becoming in general much more diffuse. In this respect they resembled quite closely the parallel condition as described for the homoplastic R. palustris transplantations. The speed and extent of the regression was, however, of a greater magnitude as compared with the R. palustris series, and resulted finally, as stated above, in complete dermal plicae degeneration. This latter observation was further emphasized by the histological findings which showed that, in all except the two cases cited above, all cellular components of the dermal folds had undergone degeneration; following which the degenerated elements had either been removed or transformed into tympanic membrane structure. In the two exceptions mentioned, the dermal plicae had undergone pronounced degenerative changes but were still visible, externally, as flat, diffuse, greyish bands. The histological pictures of these two cases showed the various structures of the folds in well-advanced stages of degeneration. Had another week elapsed before the killing of the host animals, there can be little doubt but that complete histological regression would have taken place in these two cases also.

The rapidity of dermal plicae regression was no doubt the result of influences affecting to some extent, at least, the entire integumentary graft. In this regard it was observed that the rest of the transplant (those portions adjacent to the dermal plicae folds) usually became quite soft and delicate. In fact, in three cases rupturing of the epidermis occurred, giving the transplants a somewhat spotty appearance. The generalized softening of the transplants, which apparently was localized in the epidermis, disappeared in all except a few cases following the onset of the developmental phase of tympanic membrane formation.

In all except the two cases in which complete dermal plicae regression was not obtained, some degree of tympanic membrane formation occurred. The developments attained presented varying degrees of differentiation both as regards external and histological appearance. All stages of development occurred from early stages of membrane formation (Pl. I, fig. D) to completely formed membranes (Pl. I, fig. F) which resembled the tympanic membranes of the opposite or left sides of the host animals very closely. Pl. I, fig. E, represents a case in which the peripheral regions of the membrane are developing, although the central region still presents the dark, uniformly pigmented and somewhat delicate integument which is usually found immediately following complete regression of the dermal plicae. There can be no doubt but that complete membrane development would have occurred in this case had the host animal been allowed to five for 7 or io days longer. In general, therefore, it can be concluded that the conditions for dermal plicae transformation into tympanic membrane were much more favourable in this series as compared with either of the autoplastic series or with the homoplastic R. palustris series.

The results of the heteroplastic transplantation of R. palustris dermal plicae integument to the ear region of R. clamitans were quite similar to those as obtained in the homoplastic R. clamitans series. In general, partial dermal plicae regression occurred, although its onset was usually delayed for a considerable period following the transplantation procedure. Thus, in one case (Pl. II, fig. K) no apparent change had taken place in the dermal plicae portion of the transplant up to the twentieth day following transplantation. Usually, however, typical macroscopic and histological regressive changes were initiated within 10−14 days following transplantation. Externally, the regression was evidenced by a progressive broadening and diffusing of the fold resulting in a reduction in its height. This degree of regression (Pl. II, fig. L) was not surpassed by sixteen of the twenty-one transplants of this series. In five cases, however, regression continued in the normal manner, resulting in the complete disappearance of dermal plicae structures in three instances. Early signs of tympanic membrane formation were visible in two of these three cases, while in one a half-developed membrane was obtained.

It seems quite logical to assume that, had the host animals survived for a week or two longer, more examples of complete dermal plicae regression and tympanic membrane formation would have been obtained, since all individuals were killed for fixation within 29 days following transplantation. It may be parenthetically stated here that the various hosts were killed as soon as it became evident that death would have occurred within a day or two. This procedure was necessary in order that fixation of the transplants could be secured before post-mortem changes would have become evident. The development of tympanic membrane structures was quite probably also greatly inhibited by the pronounced degenerative changes which frequently affected all regions of the transplant. This rapidity of degeneration was evidenced by the pronounced generalized darkening of the transplant coupled with the frequent disintegration and sloughing of its epidermis. Pl. II, fig. M, for example, represents a case in which the dermal plicae fold of the transplant has nearly disappeared, while the entire graft itself is subject to rapid degeneration. In fact, it was difficult to escape the impression that, had the degenerative changes been less rapid, the conditions for tympanic membrane development would have been more favourable. In conclusion, however, it can be said that the conditions for dermal plicae regression and tympanic membrane development were somewhat more favourable in this series as compared with those of the homoplastic R. palustris series.

Thirty-three transplantations of R. palustris dermal plicae integument were made to the developing ear region of R. catesbeiana hosts. Twenty-nine of the host animals survived metamorphosis for at least 3 weeks and were killed on either the thirty-third or thirty-eighth day following the initial transplantation. Dermal plicae regression was obtained in all of these cases and proceeded in a normal manner until all external and histological signs of the folds had disappeared. The integument of the transplants adjacent to the dermal plicae folds usually underwent little if any degeneration, although in a few cases some epidermal disintegration occurred following tympanic membrane formation later on. This much delayed degeneration appeared to particularly affect the epidermis of the tympanic membrane (Pl. II, fig. N).

The development of tympanic membranes was typical of all twenty-nine cases in which complete dermal plicae regression had occurred. In a few instances, complete membrane formation was not obtained by the time the hosts were killed, but in at least twenty-four cases the membranes which developed were complete in all essential respects and presented only slight differences as compared with the normal left-side membranes on the same animals. These differences were chiefly concerned with coloration, although many examples were obtained in which the characteristic brownish tint and texture of fully differentiated normal membranes was attained. In a few cases, where complete membrane development did not occur, the peripheral regions of the membranes had differentiated, while the more central portions were rapidly undergoing transformation.

In general, it can be said that the results of this series were the most uniform and satisfactory as compared with those of the other three heteroplastic series. The results, moreover, furnished clear-cut evidence that dermal plicae transformation into tympanic membrane can be obtained following direct heteroplastic transplantation of fully differentiated dermal plicae integument to the developing ear region. It is apparently not necessary, therefore, first to obtain dermal plicae regression by transplantation to non-metamorphosing larval hosts.

Thirty-six transplantations of R. clamitans dermal plicae integument were made to R. palustris hosts, of which thirty-one lived for 25 days following the transplantation procedure, at which time they were killed for fixation or preservation. The fate of the dermal plicae portions of the transplants was similar in all respects to that as described for the heteroplastic transplantation of R. palustris dermal plicae to R. catesbeiana hosts, in that total regression eventually occurred in all cases. The onset and rate of regression was subject to considerable variation, however, due no doubt to individual differences in the susceptibility of the transplants coupled with possible differences between the host animals as regards the intensity of the regressive influences they exerted.

Approximately one-half of the transplants developed well-differentiated tympanic membranes (Pl. II, fig. O), while in the remaining half the membranes developed were at least 50 per cent, formed as determined by external structure and histological appearance. The membranes were rarely typical as regards coloration but were usually darker than normal membranes and of a greyish tint. Several excellent cases were obtained, however, in which a typical brownish tinge was evident, making the membranes indistinguishable from those on the opposite or left side of the same host animals. There were only two cases in which no signs of tympanic membrane formation could be detected following total regression of the dermal plicae folds. In these two instances the grafts had undergone considerable disintegrative processes, especially as regards the epidermis which had eroded away leaving white denuded areas in the centre of the transplants. Apparently these transplants were destined to undergo total resorption, while the rapidity of the degeneration had inhibited whatever inductive influences were present favouring tympanic membrane formation.

Somewhat more favourable results were obtained in the trasplantation of R. clamitans dermal plicae integument to R. catesbeiana hosts. Thirty-three transplantations were made in all. Thirty of the host animals lived for at least 4 weeks following the operation. These were killed for preservation or histological study of the transplants from 30 to 47 days following the original transplantation procedure.

The fate of the dermal plicae portions of the transplants closely resembled that as described for the preceding series, in that an orderly process of regression occurred resulting finally in the complete disappearance of all dermal plicae structures both externally and histologically. The onset and rate of this regression, however, was considerably more uniform as compared with the same series. Tympanic membrane development was also more uniform and resulted in the differentiation of well-formed membranes in approximately 75 per cent, of the cases (Pl. II, fig. J). The coloration of these membranes was usually quite normal and compared very favourably with those on the opposite side of the same animals. In the remaining cases various stages of development occurred, resulting usually in membranes which could be characterized as being at least half-formed (Pl. II, fig. I). The results of the two heteroplastic R. clamitans series are interesting in that they show that not only R. palustris but also R. clamitans dermal plicae are capable of undergoing complete transformation into tympanic membrane following suitable transplantation to the developing ear region.

The failure of dermal plicae to undergo regressive changes in most instances, when autoplastically transplanted, can be readily understood in that the transplants were still subjected to the same general body (hormonic) influences which were responsible for their development. The development of tympanic membrane structure, however, was quite probably dependent on the intensity of the inductive influences arising from the annular tympanic cartilages in conjunction with the sensitivity of the transplants towards transformation. Thus, in the case of the R. palustris transplants, the inductive influences were apparently too weak to cause degeneration of the dermal plicae as a necessary stage in the transformation of these structures into those of tympanic membrane. The failure of integumentary portions of the transplants adjacent to the dermal plicae to undergo tympanic membrane formation is more difficult to understand. It seems likely, however, that a certain degree of resistance may have been imparted to these areas of integument due to their close association with the developing dermal plicae.

Such acquired resistance may not have been true of the integumentary portions of the R. clamitans transplants, since in many instances slight signs of external tympanic membrane formation were recorded in cases where no dermal plicae regression occurred. As will be recalled, however, there were no histological signs of membrane formation in such grafts, and hence the external signs (coloration changes, etc.) must have represented merely superficial evidences of the action of inductive influences at work. In fact, definite histological signs of membrane formation were recorded in only two of the four cases in which complete dermal plicae regression had occurred. It seems logical to conclude, therefore, that the influence of the annular tympanic cartilage in the R. clamitans series was in most cases too weak to bring about either dermal plicae regression or tympanic membrane formation in the adjacent integument. Moreover, in most cases where the degenerative influence was sufficient to degenerate the dermal plicae either partially or wholly, the requisite degree of inductive influence for membrane formation was still lacking. In the two cases where membrane formation was initiated, the inductive influences must have been correspondingly more intense. It is, of course, always possible to postulate that the inductive influences were of the same intensity in all cases and to account for the variable reactions on the part of the transplants by assuming different degrees of sensitiveness for the latter.

The partial regression of dermal plicae in the homoplastic R. palustris transplants was quite probably the result of influences emanating from the annular tympanic cartilage and facilitated, possibly, by a certain degree of antagonism between host and transplant. That homoplastic transplantation cannot in itself induce regression has been shown by the writer (1933), in that dermal plicae integument from newly metamorphosed R. palustris frogs when transplanted to the backs of non-metamorphosing larvae of the same species failed to undergo degenerative changes. However, it is necessary to realize that two separate influences which might be ineffectual when working separately could bring about dermal plicae regression when acting together. Thus, while the annular tympanic cartilage may not have been quite powerful enough to induce regression, the added effect of slight homoplastic antagonism may have been sufficient to induce regressive changes. The fact that full regression of dermal plicae was never attained and no signs of tympanic membrane development became evident, would tend to support the theory that the influence or influences available were not of maximum intensity. It is possible, also, that the annular tympanic cartilage was responsible entirely for the degree of dermal plicae regression that occurred, since it will be remembered that the cartilages in the homoplastic and heteroplastic series were in a somewhat earlier and hence more active inductive stage as compared with those of the autoplastic series.

The results of the homoplastic R. clamitans series are subject to the same general explanations. It seems quite probable here, however, that a definite antagonism did exist between host and transplant in that the grafts were at first usually subjected to a rapid generalized degeneration. This condition was not only conducive to dermal plicae regression but favourable to tympanic membrane formation. The regression of dermal plicae was no doubt also facilitated by the presence of the annular tympanic cartilage which was also, of course, responsible for the varying degrees of membrane formation which occurred.

The conditions for dermal plicae regression and tympanic membrane development were apparently much more favourable in the heteroplastic than in the homoplastic series. Here, for the first time, the possibility of species specificity and antagonism was a factor and no doubt was responsible for much of the dermal plicae regression which occurred. In this regard the reader is referred to a previous work of the writer (1933) in which it was shown that very definite regression of well-developed R. palustris dermal plicae followed the latter’s transplantation to non-metamorphosing R. catesbeiana larvae. It might be argued that in this case regression was to be expected in that the non-metamorphosing larval hosts were exhibiting no dermal plicae developmental influences at the time, while, conversely, the hosts as used in the present heteroplastic series had developed and were maintaining their own normal dermal plicae folds. However, the possibility of specific intolerance to foreign dermal plicae is still quite probable, and the generalised disintegration of transplants which occurred in many instances lends support to this idea. Obviously, further work is necessary before this particular point can be adequately dealt with. It would be interesting in this regard to determine the fate of heteroplastic transplantations of dermal plicae to other regions than that of the annular tympanic cartilage, employing host animals in the same metamorphic stage as utilized for the present work.

The completeness of dermal plicae regression in most cases of heteroplastic transplantation was no doubt responsible in a large measure for the high degree of tympanic membrane development which usually occurred. As has been stated before, the regression of dermal plicae not only reduces a highly differentiated structure to a state closely resembling normal integument, but also brings about a slight disintegration of that integument which is similar in most respects to the early disintegrative phases of normal tympanic membrane formation. Thus, the annular tympanic cartilage is not called upon to induce the early degenerative steps of membrane formation, but can exercise its full force in the induction of the later developmental phases. A comparison, moreover, of the results of the various heteroplastic series would lead one to conclude that the annular tympanic cartilage of R. catesbeiana is more active and efficient as compared with that of R. palustris or R. clamitans in relation to membrane induction.

The failure of R. palustris dermal plicae to undergo complete regression in most instances, when transplanted to R. clamitans hosts, is difficult of explanation. Apparently the regressive influences at work here were of a less intensity as compared with those prevailing in the other host species used, since the onset of dermal plicae regression was usually much delayed. In spite of this, however, the integumentary portions of the transplants were frequently subjected to rapid disintegrative processes, and it is difficult to escape the impression that the abnormal degeneration present may have in this case tended to neutralize the inductive effect of the annular tympanic cartilage towards membrane development. However this may be, it is evident that the host rather than the transplant was mainly responsible for the retarded plicae regression and membrane development which occurred, since the reverse transplantation gave more favourable results, while both the R. clamitans and R. palustris series involving R. catesbeiana hosts resulted in complete plicae regression and membrane formation.

(All photographs approximately $213$⁠)

Fig. A. Autoplastic R. clamitans graft 31 days following transplantation. No regression of dermal plicae fold. The fold is seen traversing the centre of the graft.

Fig. B. Autoplastic R. clamitans graft 29 days following transplantation. Dermal plicae fold has undergone partial regression. Note increased width and general diffusion of fold.

Fig. C. Autoplastic R. clasmtasu graft 31 days following transplantation. Complete regression of dermal plicae fold and partial development of tympanic membrane.

Fig. D. Homoplastic R. clamitans graft 28 days following transplantation. Complete dermal plicae regression and early tympanic membrane formation.

Fig. E. Homoplastic R. clamitans graft 29 days following transplantation. Complete dermal plicae regression with peripheral regions of tympanic membrane forming.

Fig. F. Homoplastic R. clamitans graft 28 days following transplantation. Complete dermal plicae regression and fully formed tympanic membrane.

Fig. G. Autoplastic R. palustris graft 26 days following transplantation. No regression of dermal plicae fold. The fold is seen traversing the graft just posterior and dorsal to the angle of the jaw.

Fig. H. Homoplastic R. palustris graft 26 days following transplantation. Dermal plicae fold has undergone partial regression. Note increased width and general diffusion of fold.

Fig. I. R. clamitans graft on R. catesbeiana host 47 days following transplantation. Complete dermal plicae regression and partial tympanic membrane formation.

Fig. J. R. clamitans graft on R. catesbeiana host 47 days following transplantation. Complete dermal plicae regression and well-differentiated tympanic membrane.

Fig. K. R. palustris graft on R. clamitans host 20 days following transplantation. No regression of dermal plicae fold.

Fig. L. R. palustris graft on R. clamitans host 23 days following transplantation. Partial regression of dermal plicae fold. Note increased width and general diffusion of fold.

Fig. M. R. palustris graft on R. clamitans host 22 days following transplantation. Dermal plicae fold in dorsal part of graft has nearly disappeared while graft, in general, is degenerating.

Fig. N. R. palustris graft on R. catesbeiana host 38 days following transplantation. Complete dermal plicae regression and well-differentiated tympanic membrane. The epidermis of the membrane has undergone degeneration.

Fig. O. R. clamitans graft on R. palustris host 25 days following transplantation. Complete dermal plicae regression and well-differentiated tympanic membrane.