Forelimb Spike Regeneration in Xenopus laevis: Testing for Adaptiveness

Transcription

1 JOURNAL OF EXPERIMENTAL ZOOLOGY 301A: (2004) Forelimb Spike Regeneration in Xenopus laevis: Testing for Adaptiveness ROY A. TASSAVA Department of Molecular Genetics, The Ohio State University, Columbus, Ohio ABSTRACT Experiments were designed to test adaptability of forelimb spike regenerates in Xenopus laevis froglets. The results show that when amputation is at the radius/ulna level, regeneration occurs in 100% of the cases and a single spike of cartilage is the result. The spike regenerates originating from radius/ulna level amputations can be used for feeding and froglet growth is only minimally compromised by the spike. The spike grows in length as the froglet body grows and thus is in homeostasis with the body. The spike develops nuptial pad tissue in reproductively mature males and is occasionally molted, indicating responsiveness to gonadal and thyroid hormones. Finally, and most important, the spike can be used for amplexus and successful mating. In contrast, spikes originating from humerus level amputations were considerably shorter and regeneration from that limb level was less frequent. When amputation was at the body wall regeneration did not occur. J. Exp. Zool. 301A: , r 2004 Wiley-Liss, Inc. INTRODUCTION The South African Clawed frog, Xenopus laevis, uses its forelimbs mainly for feeding and amplexus. The forelimbs are shaped and anatomically positioned so that immediately upon metamorphosis and then throughout life the frog can use them to scoop food into its mouth. Thus the autopod is perpendicular to the substrate rather than flat as seen with most anurans. The male frog is about two-thirds the size of the female but nevertheless can grasp the female firmly just anterior to the hind limbs around the base of the abdomen during amplexus. It thus is positioned ideally to deposit sperm on the newly released oocytes. This is accomplished in part by the geometry of the forelimb and also by the irregular and highly cornified epidermis ( black sticky hairs, Deuchar, 75; nuptial excrescences, Duellman and Trueb, 86) that develop exclusively on the inner surfaces of the forelimb autopod and zeugopod (digits, palms, and forearms) in the sexually mature male. Even though the female skin is covered by mucus and is very slippery, the male holds the amplexus position until mating is completed. While swimming, the frog is propelled almost exclusively by the hindlegs; the autopod is again geometrically positioned to get the most forward thrust from the push backwards. The hindlimb digits also contain keratinized claws which are often used to tear food into small pieces which the forelimbs then push into the mouth (Deuchar, 75). Thus, both fore- and hindlimbs are arguably highly adaptive for Xenopus in its aquatic environment. Xenopus, as is true for most anuran amphibians, can regenerate complete fore- and hindlimb patterns during early stages of metamorphosis but in later stages the pattern of the regenerated structure becomes more and more hypomorphic (Dent, 62). Nevertheless, froglets and adult frogs retain the ability to regenerate a muscle-deficient, cartilagenous spike. Based on histological analysis, Korneluk and Liversage ( 84a) suggested that spike regeneration in Xenopus represents a dominant tissue regeneration response in contrast to epimorphic regeneration seen in newts. Epimorphic regeneration involves release of differentiated cells from their tissues and acquisition of pluripotency in developmental potential, i.e. dedifferentiation, and their proliferation and accumulation to form a blastema. Tissue regeneration, in contrast, involves the proliferation of cells with limited developmental potential, such as fibroblasts, without dedifferentiation but with some limited tissue replacement (reviewed in n Correspondence to: Roy A. Tassava, Department of Molecular Genetics, 484 W. 12th Ave., Columbus, OH tassava.1@osu.edu Received 22 April 2003; Accepted 1 October 2003 Published online in Wiley InterScience ( com). DOI: /jez.a r 2004 WILEY-LISS, INC.

2 XENOPUS REGENERATION IS ADAPTIVE 151 Korneluk and Liversage, 84a). That the Xenopus spike regenerate represents epimorphic regeneration has been argued based on both grafting (Goss and Holt, 92) and molecular (Endo et al., 2000) data. Whether or not limb regeneration is evolutionarily adaptive or inherent has been discussed by Goss ( 92) and Thouveny and Tassava ( 98). Goss ( 92) states, If regeneration were adaptive, it would have arisen autonomously by natural selection from non-regenerative antecedents. Unless each episode coincidently reinvented the same method of regeneration independently, one would expect the various lineages to differ basically from each other, which they do not. On the other hand, if regeneration were inherent to metazoan life, a derivative of embryogenesis, its various expressions should be as much like each other as they resemble the development of embryonic appendage buds, which they do. It follows that the uneven distribution of regeneration must have been due to its extinction here and there, not as a negative adaptation by natural selection but as a pleiotropic epiphenomenon linked to more useful adaptations with which it became incompatible. Goss ( 92) goes on to argue that these latter adaptations included the transition from water to land habitats and the physiological change from poikilothermy to homeothermy. Later he states, Only by being subjected frequently enough to the acid test of natural selection can a structure or process be preserved in the course of evolution. Goss ( 92) favored the inherent idea, at least as regards vertebrate limb regeneration, a process restricted largely to urodele amphibians. The issue of inherent vs adaptive is discussed by Thouveny and Tassava ( 98) who point out that one interpretation of the adaptive view regarding epimorphic regeneration would be that early vertebrates initially did not have the capacity to regenerate and that the process evolved in some species by natural selection. They further suggest that one selective pressure for urodele limb regeneration ability might have been sibling chewing due to decreasing pond size and increasing population density, a phenomenon readily demonstrated in laboratory populations of urodeles. Furthermore, when urodele larvae (Ambystoma tigrinum) are reared in high density, they are likely to become cannibalistic (Pfennig and Collins, 93). Wagner and Misof ( 92) suggest multiple independent origins of impaired regeneration in teleosts and amphibians. The present investigation was designed to test the hypothesis that the pattern deficient spike regenerates in Xenopus are adaptive. Measurements were done to assess whether froglets with one or both forelimb spikes could eat adequately and grow normally to adult stages. Measurements assessed whether the initial regenerated spikes subsequently grew as the frog grew and thus could be said to be in homeostasis with the body. Related to the test of adaptiveness, the extent of regeneration was compared at the stylopodium vs zeugopodium vs body wall levels. Finally, tests were carried out to assess whether frogs that had grown and matured sexually with one or two forelimb spikes could undergo amplexus and reproduce. MATERIALS AND METHODS Xenopus froglets and adults were raised from tadpoles fed strained baby food (peas and beans). Hindlimbs were present by 15 days after hatching (stage 48; Nieuwkoop and Faber, 56) and the first froglets metamorphosed by seven weeks (see also Brown, 70). Newly metamorphosed froglets were randomized by size and housed at a density of froglets per clear, plastic container in 2 liters of water at a depth of 1.5 inches and were fed small pieces of beef liver once daily. Liver was introduced at a rate of 4 5 pieces per froglet per day but some froglets showed greater assertiveness and consumed 6 7 pieces/day. Water was changed daily. Within one or two days of completion of metamorphosis, assessed by the complete loss of the tail (stage 66), froglets were lightly anesthetized in % MS 222 (methane sulfonate; Sigma) and forelimbs were amputated either through the radius/ulna (R/U; zeugopodium), humerus (H; stylopodium), or at the body wall. Amputations were randomly made at the R/U and H levels such that the plane of amputation varied from more distal, to middle, to more proximal levels of the respective skeletal elements. In each of two containers with 12 frogs each, two frogs had both forelimbs amputated at the R/U level, six frogs had only one forelimb amputated at the R/U level, and four frogs served as unamputated controls. In a third container of 12 frogs, four frogs had both forelimbs amputated at the H level, four frogs had one forelimb amputated at the mid-h level, and four frogs served as unamputated controls. In a fourth container of 12 frogs, four frogs had one forelimb amputated at the body wall, four frogs

3 152 R. A. TASSAVA had both forelimbs amputated at the body wall, and four frogs served as unamputated controls. It was not possible to determine the sex of the frogs until reproductive maturity, when the forelimbs of the males developed nuptial pad tissue and the size difference between males and females became evident. The frequency and extent of regeneration were assessed through time at the three amputation levels. Frog body growth was assessed and compared between the unamputated, single forelimb amputated at the R/U level, and dual forelimbs amputated at the R/U level. The mean of the body lengths of the control, unamputated frogs was determined and the body lengths of the two experimental groups were expressed as the percent of the mean of the control, unamputated frog body lengths. Length of the regenerate was determined by measuring from the mid-elbow to the tip of the regenerate and thus included a short portion of the stump. The mid-elbow was utilized as a baseline point because the amputation level became less clear as regeneration progressed. Growth of the regenerates was examined in proportion to the growth of the frog, means were determined, and the data were expressed as percent of the control limb length from the midelbow to the tip of the longest digit. A total of seven unamputated frogs (four males and three females) and 10 frogs with a single spike regenerating at the R/U level (four males and six females) were raised to adulthood. When the frogs reached reproductive maturity, at months of age (see Brown, 70; Deuchar, 75), three control, unamputated males and three males with one forelimb spike were paired with either unamputated females or single-spike females. The appropriate human chorionic gonadotropic hormones (Sigma) were administered by injection into the dorsal lymph sac (Brown, 70; Sive et al., 2000) and observations were made of the occurrence or absence of amplexus and whether fertile eggs were obtained, as assessed by observations of embryonic development. After observations on mating, some limbs were sampled for histological analysis. Limbs were fixed in Bouin s fixative, dehydrated and cleared, embedded in paraffin, sectioned serially, and deparaffinized sections were stained with hematoxylin and eosin. Unamputated forelimbs and single spike regenerates were examined in males to evaluate the presence and extent of nuptial pad tissue. Spike regenerates were sectioned longitudinally to evaluate the melding of the regenerate cartilage to the stump bones. Also, a single spike regenerate that was floppy and not melded to the stump bones was sectioned longitudinally and examined by histology. RESULTS Blastemas were present by 10 days after amputation at which time it could be ascertained whether a limb would or would not regenerate. Every limb regenerated when amputation was at the R/U level; single spike regenerates were invariably produced, sometimes with slight bends (Fig. 1a). In one case a distally bifurcated regenerate was seen (not shown). A lower proportion of froglets regenerated when amputation was through the H (8 of 10 limbs regenerated) and the spikes were considerably shorter than those at the R/U level (Fig. 1b). In one case an elbow-like bend was seen (Fig. 1b, 4th frog from the left). Of the 12 forelimbs amputated at the body wall, none regenerated (Fig. 1c). Froglets completed forelimb spike regeneration by five weeks after amputation, far before body growth was completed. The spike regenerate then grew in relatively good proportion to the body. At one month after amputation, the lower arm (zeugopodium stump plus regenerated spike) represented on average 38% of the length of the body. At the same age, with unamputated froglets, the length of the zeugopodium plus the autopodium represented on average 40% of the body length. Also, the average length of the zeugopodium stump plus the regenerated spike in proportion to that of the zeugopodium plus autopodium of intact forelimbs was 98%. At 10 months, when froglets had nearly reached adult size, spike regenerates were again measured in proportion to the opposite unamputated forelimb. The proportion on average was now only somewhat less, being 90%. The above results are based on average spike and body lengths of at least five froglets per measurement. It can be seen in Fig. 1a that the forelimbs with the spikes regenerating from the R/U level are very nearly equal in length to the contralateral, unamputated forelimbs. Variation in lengths of the regenerates was possibly related to inexact levels of amputation but an insufficient number of cases was available to carefully examine this possibility. Nevertheless, forelimbs amputated through the proximal portion of the R/U seemed to regenerate shorter spikes than forelimbs amputated through the distal R/U. For the proportional comparisons

4 XENOPUS REGENERATION IS ADAPTIVE 153 Fig. 1. Photos of froglets 5 weeks after amputation through the radius/ulna (1a), the humerus (1b) and at the body wall (1c). Note in 1a that the spike regenerates are nearly the same length as the opposite, unamputated forelimb. Arrow points to the amputation level of the froglet at the left in the figure. In 1b it can be seen that the spike regenerates are much shorter when amputation is through the humerus. The arrow indicates the level of amputation of the frog at the left in the figure. The frogs in the figure represent 4 of the 8 limbs that regenerated (out of a total of 10) when amputation was at the humerus level. The 4 frogs shown in Fig. 1c are typical of the non-regeneration seen when amputation is at the body wall (see arrow pointing to level of amputation of 2nd frog from the left). Bar in 1a (equivalent in 1b, 1c)¼10 mm. Photos represent actual size of frogs.

5 154 R. A. TASSAVA below, froglet growth was assessed by determining the means of the body lengths for 4 5 randomly chosen froglets previously amputated at the R/U level. Froglets with two intact forelimbs showed the most rapid body growth with mean body weights increasing from 0.55 g on day 0 (day of amputation) to g at one month after amputation, to 10.2 g at four months after amputation. Froglets with one leg amputated increased from 0.54 g at day 0 to 0.90 g at one month and 9.4 g at four months. Froglets with both forelimbs amputated showed mean body weights increased from 0.53 g at day 0 to g at one month to 8.8 g at four months. Thus, froglets with both forelimbs intact exhibited the most growth; froglets with both forelimbs amputated exhibited the least growth; and froglets with only one forelimb amputated exhibited intermediate growth. Expressed as percent of control, froglet mean body weight was less at one month for the froglets with both limbs amputated (79% of control) than at four months (86% of control) whereas for the singly amputated frogs, the percent of control was 94% at one month and 92% at four months. When forelimb regeneration was completed and spike regenerates reached maximum lengths, froglets could utilize the spikes for feeding, and body growth rate was increased but was still not equivalent to that of froglets with both forelimbs intact. Frog feeding was hindered more by having two spikes than by having one spike. The single-spike frogs that were raised to adulthood (seven months and beyond) grew to approximately the same size and weight as the unamputated frogs. Average weights were not determined because the sex of individual frogs began to influence the weight and size, with the weights of females ultimately ranging from grams and the weights of males ranging from grams (Brown, 70). Growth data were not collected on froglets with limbs amputated at the body level but it became obvious early on that these froglets struggled to obtain food, particularly if both forelimbs had been amputated. Froglets amputated at the body wall were not raised to adulthood. Upon reaching sexual maturity, Xenopus males developed typical nuptial pad tissue on the inside surface of the autopodium and zeugopodium (Fig. 2) and this nuptial pad tissue became more pronounced after the hormonal injections (not shown; see Brown, 70). None of the female frogs developed nuptial pad tissue (see Fig. 6). The spikes that developed from R/U level amputations also developed nuptial pad tissue, whether one or Fig. 2. Ventral view of a male frog exhibiting breeding plumage. Nuptial pad tissue is apparent on the inside of the left, unamputated limb and inside of the right, amputated limb, extending from the stylopodium (arrow) to the zeugopodium and autopodium of the unamputated limb and to the tip of the regenerated spike. The black claws at the tips of the hindlimb toes can be seen. Bar¼1 cm.

6 XENOPUS REGENERATION IS ADAPTIVE 155 both forelimbs were amputated, as seen with the unaided eye (Figs. 2, 3, and 5), the dissecting microscope, and by histology (Fig. 4a,b). The epidermis and the nuptial pad tissue was occasionally molted (Fig. 4a). Amplexus behavior in males is reflexive and a male will clasp onto a researcher s finger (Deuchar, 75). The stickiness of the intact forelimb nuptial pad vs. the nuptial pad tissue on the spikes seemed equivalent, as tested with a finger by two independent testers. Also, the density of the black, nuptial excrescences appeared similar between unamputated forelimbs and regenerated spikes but quantitation was not attempted. Each of the black nuptial excrescences appeared to be associated with a specialized group of epidermal cells immediately below it (Fig. 4b). All three male frogs with intact forelimbs exhibited amplexus behavior. Amplexus is lumbar in Xenopus as opposed to axial as in Rana (Brown, 70). All three of the females produced fertile eggs with percent fertilities of 78%, 75%, and 71%. Likewise, all three male frogs with one intact and one forelimb spike that had regenerated from the mid-radius/ulna level exhibited attempts at amplexus behavior but only two of the frogs were able to amplex successfully (Fig. 6) and mate. These two matings resulted in eggs with percent fertilities of 74% and 72%. It was estimated that in each of the latter two cases the female released over 200 eggs. The single spike frog that did not successfully amplex was subsequently noted to have a floppy spike that was not melded to the cut bones. Analysis of serial sections of this forelimb after H&E staining confirmed this observation (not shown). The cartilagenous spike regenerate in Xenopus is normally continuous with and firmly melded into the bones of the stump, as seen by histology (not shown; see Korneluk and Liversage, 84a). The frogs amputated at the H level were not raised to adulthood. It seems likely that these spikes would have been too short to enable the frogs to undergo amplexus with the respective female, even if one forelimb was intact. It is not known if the short spikes regenerated from the mid-humerus level would have developed nuptial pad tissue. DISCUSSION The results are consistent with the view that Xenopus forelimb spike regenerates originating from amputations at the R/U level are adaptive. First, froglets were able to utilize the spikes for feeding under laboratory conditions. Second, once regenerated, the spikes grew proportionally to the growth of the body, suggesting that they were in homeostasis with the body. Third, in males, Fig. 3. Ventral view of a male frog with regenerated spikes on both forelimbs. Both spikes developed nuptial pad tissue. Bar¼1 cm.

7 156 R. A. TASSAVA Fig. 4. Micrographs of histological sections showing the black, horn-like nuptial excrescences on the inner portion of the regenerated spikes in male frogs. In 4a it can be seen that the outer epidermal layer of the nuptial pad tissue is occasionally molted (arrow). g¼gland. c¼cartilage of the spike extending across the figure. Distal is to the right in both 4a and 4b. bar¼0.3 mm. 4b is a higher magnification showing the black, horn-like nuptial excrescences angled to maximize grasping of the female. Note that each excrescence seems to originate from a specialized group of epidermal cells immediately below it (arrow). g¼gland. bar¼0.1 mm. nuptial pad tissue developed along the inner limb surface which was occasionally molted, thus indicating that the skin was responsive to the relevant testicular and thyroid hormones (Duellman and Trueb, 86). If any of the above results were not seen, one could argue that the spike regenerates were not adaptive. However, these observations alone would not indicate adaptiveness if frogs with one or more forelimb spike regenerates were unable to breed. Thus, the fourth observation, that male frogs were able to utilize the spikes in amplexus and breeding, taken together with observations 1 3, lends strong support to the hypothesis that forelimb spike regeneration in Xenopus is adaptive. The adaptiveness of Xenopus regenerates is consistent with the views of Goss and Holt ( 92) and Endo et al. (2000) that this spike comes about by epimorphic regeneration. Goss and Holt ( 92) reasoned that the key feature of epimorphic regeneration is blastema formation and, furthermore, that blastema formation requires a functional wound epithelium (see Mescher, 76). Thus wound epithelium formation was prevented by inserting freshly amputated limbs into the body cavity. Such limbs did not form a blastema, leading Goss and Holt ( 92) to conclude that Xenopus spike regeneration is epimorphic. Endo et al. (2000) utilized probes for molecular markers of urodele blastema formation and observed the same markers in Xenopus blastemas, again concluding that Xenopus regeneration is epimorphic. It remains to be determined what patterning genes are not expressed during Xenopus regeneration and why selection was not rigorous enough to result in anterior/posterior regenerate patterning in this frog. It is of interest that when amputations occurred at the H level, regeneration either did not occur or was very poor. The spikes that did regenerate were less than half the length of the spikes regenerating at the R/U level. Whether or not H level spikes would have developed nuptial pad tissue upon sexual maturity was not determined. Finally, it seems likely that frogs with the short H level spikes, even with nuptial pad tissue, would be unable to amplex with female frogs. If a spike regenerated from the H level, even if it was the length of the missing limb portion, amplexus would not be possible without an elbow joint or without a bend in the medial direction. It follows that there would be no adaptive value to regenerate from the humerus level, consistent with the present results. Certainly spikes emanating from the body wall, if regeneration occurred from that level, could not be used in amplexus and would not be adaptive. Predictably, as shown here, regeneration does not occur from the body wall level (see also Gallien and Beetschen, 51). It should be noted that adult Xenopus also regenerate better from distal levels and there is no regeneration from very proximal levels (Gallien and Beetschen, 51), as seen here with froglets. The spike originating at the R/U level is tightly melded into the cut ends of the bones, as noted previously (Korneluk and Liversage, 84a). In the single case here, wherein the spike was not melded

8 XENOPUS REGENERATION IS ADAPTIVE 157 Fig. 5. A dorsal view of a male frog showing nuptial pad tissue visible on the autopod of the unamputated left forelimb and on the spike regenerate of the right forelimb. This frog is shown in amplexus with a female in Fig. 6. bar¼1 cm. Fig. 6. A dorsal view of the male frog shown in Fig. 5 in amplexus with a female. The photograph is somewhat reduced in size compared to Fig. 5. Note that the right forelimb with the spike regenerate has a good grasp of the female. This pair into the stump bones, amplexus was not possible. In this regard, it should be noted that spike regeneration occurs even if the stump bones are of frogs had a successful mating. The female has a spike regenerate on the left but neither forelimb exhibits nuptial pad tissue. Note the larger size of the female. Bar¼1.5 cm. not present at the level of amputation (Korneluk and Liversage, 84b) but such spikes probably could not be used in amplexus. Some spikes

9 158 R. A. TASSAVA (see Figs. 1 and 2) showed bending to various degrees, indicative of joint formation. However, in a preliminary histological study of Xenopus froglet regenerates that showed bending, only one of six had what appeared to be a joint; the other five were composed of solid cartilage (Tassava, unpublished). If a froglet regenerate did have a joint, then the musculature necessary for maintaining the distal segment in the amplexus position would necessarily have to be regenerated. But, Xenopus regenerates are noted for their absence of muscle (Korneluk and Liversage, 84a). If forelimb spike regeneration is adaptive, the question arises as to what factor or factors caused limb loss (and thus adaptive pressure for regeneration) in evolutionary history. Predation is perhaps the most likely. Limbs of late tadpoles and young froglets are somewhat fragile, certainly at least as fragile as limbs of Ambystoma larvae. The latter are often chewed off to varying degrees by siblings. However, Xenopus froglets do not chew legs of siblings in the laboratory, at least under conditions of ad libitum feeding, as in the present experiment. It would be of interest to examine natural populations of Xenopus to determine if froglet limbs are chewed by siblings or by predators. It is possible that froglet limbs would be chewed by adult frogs if sharing the same pond or even that larger froglets might chew the limbs of smaller froglets if in dense conditions. Note that adults will eat tadpoles (Deuchar, 75). Finally, it might seem to be intuitively obvious that hind limbs would be lost more often than forelimbs. Hindlimbs are present for a good part of tadpole life, i.e. during all of pro-metamorphosis, and are long and perhaps more available to be grasped by predators. To be adaptive, the hindlimb regenerative spike would have to be useful for swimming, either to find food or escape predation. The present results are interesting with regard to the view favored by Goss ( 92) that limb regeneration is inherent, as opposed to being adaptive. In attempting to apply this view to Xenopus spike regeneration, it would seem that initially this frog species would have been capable of regenerating a complete pattern and then for some reason the ability to regenerate a pattern was lost. The two main reasons for loss of inherent regenerative ability in non-regenerating vertebrates are acquisition of a terrestrial habitat and warm-bloodedness (Goss, 92). Since Xenopus is poikilothermic and aquatic, there seems be no intuitive reason as to why this species should not have retained the ability to regenerate a complete pattern. On the other hand, as noted above and a view favored by the present results, selective pressure may have been sufficient to result in Xenopus spike regeneration but with selection pressure not strong enough for pattern regeneration. Alternatively, the rigorous foreleg and hindleg movements in Xenopus possibly precluded the regeneration of a pattern due to the fragile nature of the intermediate blastema stages. Xenopus is the most primitive of the three genera in the family Pipidae. It is difficult to hypothesize about gain or loss of regeneration in Xenopus without knowing if this species was terrestrial at one time and then returned to water, or never evolved the ability to live on land (see Deuchar, 75). The initial regenerates of the froglets did not show dorsal/ventral or anterior/posterior patterning but when the male froglets matured, nuptial pad tissue always formed on the ventral surface, indicative of a dorsal/ventral pattern. Whether this pattern is the result of patterning mechanisms or merely due to contributions of pad-potential cells from the stump skin to the regenerate remains to be determined. In attempts to stimulate pattern regeneration in Xenopus, factors could be utilized that are known to be involved in urodele regeneration, wherein a complete pattern is the norm. One such factor, retinoic acid, resulted in cartilage condensations with complex cartilage distal subdivisions (Crawford and Liversage, 92) or had very little effect (Tassava, unpublished) when applied to Xenopus limb stumps. From the results of Endo et al. (2000) it would seem logical to express a sonic hedgehog transgene alone or in combination with dorsal-ventral patterning genes in the Xenopus limb stump. Another approach would be to remove presumed inhibitions (Goss, 80). Finally, the developmental program necessary to regenerate a spike is itself likely to be complex and some effort would be necessary to elucidate all of its complexities in order to modify it, in favor of a patterning program (Thouveny and Tassava, 98). ACKNOWLEDGEMENTS The author is grateful to Anthony L. Mescher for many helpful suggestions on the manuscript.

10 XENOPUS REGENERATION IS ADAPTIVE 159 LITERATURE CITED Brown AL The African clawed toad, Xenopus laevis. A guide for Laboratory Practical Work. London, UK: Butterworth and Company. p Crawford M, Liversage RA Pattern perturbation in Xenopus laevis forelimb regenerates following treatment with retinoic acid and carrier media. In: Taban CH, Boilly B, editors. Keys for regeneration. Monogr Dev Biol Basel, Switzerland: Karger. p Dent JN Limb regeneration in larvae and metamorphosing individuals of the South African clawed toad. J Morph 110: Deuchar EM Xenopus: the South African clawed frog. New York: John Wiley & Sons. P Duellman WE, Trueb L Biology of the amphibians. Baltimore, Maryland: Johns Hopkins Press. p Endo T, Tamura K, Ide H Analysis of gene expressions during Xenopus forelimb regeneration. Dev Biol 220: Gallien L, Beetschen J Extension et limites du pouvoir regenerateur des membres chez Xenopus laevis. C R Soc Biol 145: Goss RJ Prospects for regeneration in man. Clin Orthop 151: Goss RJ, Holt R Epimorphic vs tissue regeneration in Xenopus forelimbs. J Exp Zool 261: Goss RJ The evolution of regeneration: adaptive or inherent? J Theor Biol 159: Korneluk RG, Liversage RA. 1984a. Tissue regeneration in the amputated forelimb of Xenopus laevis froglets. Can J Zool 62: Korneluk RG, Liversage RA. 1984b. Effects of radius-ulna removal on forelimb regeneration in Xenopus laevis froglets. J Emb Exp Morph 82:9 24. Mescher AL Effects on adult newt limb regeneration of partial and complete skin flaps over the amputation surface. J Exp Zool 195: Nieuwkoop PD, Faber J Normal table of Xenopus laevis (Daudin). North-Holland, Amsterdam. p Pfennig DW, Collins JP Kinship affects morphogenesis in cannibalistic salamanders. Nature 362: Sive HL, Grainger RM, Harland RM Early development of Xenopus laevis: A laboratory manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. p 100. Thouveny Y, Tassava R Regeneration through phylogenesis. In: Ferretti P, Geraudie J, editors. Cellular and molecular basis for regeneration. New York: John Wiley & Sons. p Wagner GP, Misof BY Evolutionary modification of regenerative ability in vertebrates: a comparative study on teleost pectoral fin regeneration. J Exp Zool 261:62 78.