Web Topic 21.1 Variations on the Themes of Amphibian Metamorphosis

Many amphibians have altered their life cycle by modifying the duration of their larval stage. This phenomenon, whereby animals change the relative time of appearance and rate of development of characters present in their ancestors, is called heterochrony. Here we will describe three extreme types of heterochrony:

  • Neoteny refers to the retention of the juvenile form as a result of retarded body development relative to the development of the germ cells and gonads (which achieve maturity at the normal time).

  • Progenesis also involves the retention of the juvenile form, but in this case the gonads and germ cells develop at a faster rate than normal, becoming sexually mature while the rest of the body is still in a juvenile phase.

  • In direct development, the embryo abandons the stages of larval development entirely and proceeds to construct a small adult.


In certain salamanders, the reproductive system and germ cells mature while the rest of the body retains its juvenile form throughout life. In most such species, metamorphosis fails to occur and sexual maturity takes place in a “larval” body.

The Mexican axolotl (Ambystoma mexicanum) does not undergo metamorphosis in nature because its pituitary gland does not release the thyrotropin (thyroid-stimulating hormone) that would activate T4 synthesis (Prahlad and DeLanney 1965; Norris et al. 1973; Taurog et al. 1974). The axolotl does synthesize functional thyroid hormone receptors, however, and when investigators administered either thyroid hormones or thyrotropin, they found that the salamander metamorphosed into an adult form not seen in nature (Figure 1; Huxley 1920; Safi et al. 2004).

Figure 1

Figure 1   Metamorphosis in Ambystoma. (A) Normal adult Ambystoma, with prominent gills and broad tail. (B) Metamorphosed Ambystoma not seen in natural populations. This individual was grown in water supplemented with thyroxine. Its gills have regressed, and its skin has changed significantly. (A © Mark Boulton/Alamy; B courtesy of K. Crawford.)

Other species of Ambystoma, such as A. tigrinum, metamorphose only in response to cues from the environment. In parts of its range, A. tigrinum is neotenic: its gonads and germ cells mature and the salamander mates successfully while the rest of the body retains its aquatic larval form. However, in other regions of its range, the larval form is transitory, leading to the land-dwelling adult tiger salamander. The ability to remain aquatic is highly adaptive in locations where the terrestrial environment is too dry to sustain the adult form of this salamander (Duellman and Trueb 1986).

Some salamanders are permanently neotenic, even in the laboratory. Whereas T3 is able to produce the long-lost adult form of A. mexicanum, the neotenic species of Necturus and Siren remain unresponsive to thyroid hormones (Frieden 1981). Strangely, Necturus was recently found to have functional thyroid hormone receptors. It appears that these receptors do not bind to those genes that initiate and promote metamorphosis (Safi et al. 2006; Vlaeminck-Guillem et al. 2006).

De Beer (1940) and Gould (1977) have speculated that neoteny is a major factor in the evolution of more complex taxa. By retarding the development of somatic tissues, neoteny may give natural selection a flexible substrate. According to Gould (1977, p. 283), neoteny may “provide an escape from specialization. Animals can relinquish their highly specialized adult forms, return to the lability of youth, and prepare themselves for new evolutionary directions.”


In progenesis, gonadal maturation is accelerated while the rest of the body develops normally to a certain stage. Progenesis has enabled some salamander species to find new ecological niches. Bolitoglossa occidentalis is a tropical salamander that, unlike other members of its genus, lives in trees. This salamander’s webbed feet and small body size suit it for arboreal existence, the webbed feet producing suction for climbing and the small body making such traction efficient. Alberch and Alberch (1981) showed that B. occidentalis resembles juveniles of the related species B. subpalmata and B. rostrata (whose young are small, with digits that have not yet grown past their webbing). B. occidentalis reaches sexual maturity at a much smaller size than its relatives, and this appears to have given it a phenotype that made tree-dwelling possible.

Direct development

While some animals have extended their larval life stage, others have “accelerated” their development by abandoning their larval form for direct development. Thus, there are frog species that lack tadpoles and sea urchins that have no pluteus larvae.

Elinson and his colleagues (del Pino and Elinson 1983; Elinson 1987) have studied Eleutherodactylus coqui, a small frog that is one of the most abundant vertebrates on the island of Puerto Rico. Unlike the eggs of Rana and Xenopus, the eggs of E. coqui are fertilized while they are still in the female’s body. Each egg is about 3.5 mm in diameter (roughly 20 times the volume of a Xenopus egg). After the eggs are laid, the male gently sits on the developing embryos, protecting them from predators and desiccation (Taigen et al. 1984).

Early E. coqui development is like that of most frogs. Cleavage is holoblastic, gastrulation is initiated at a subequatorial position, and the neural folds become elevated from the surface. However, shortly after the neural tube closes, limb buds appear on the surface (Figure 2A,B). This early emergence of limb buds is the first indication that this animal will not pass through the usual limbless tadpole stage. Moreover, the development of E. coqui is modified such that the modeling of most of its features—including its limbs—does not depend on thyroid hormones. Its thyroid gland does develop, however, and thyroid hormones appear to be critical for the eventual resorption of the tail (which is used as a respiratory rather than a locomotor organ), the differentiation of the skin, and the remodeling of the kidney and musculature (Lynn and Peadon 1955; Callery and Elinson 2000). It appears that the thyroid-dependent phase has been pushed back into embryonic growth (Hanken et al. 1992; Callery et al. 2001). What emerges from the egg jelly 3 weeks after fertilization is not a tadpole but a tiny frog (Figure 2C).

Figure 2

Figure 2   Direct development of the frog Eleutherodactylus coqui. (A) Limb buds are seen as the embryo develops on the yolk. (B) As the yolk is used up, the limb buds are easily seen. (C) Three weeks after fertilization, tiny froglets hatch. They are seen here in a petri dish and on a Canadian dime. (Courtesy of R. P. Elinson.)

Direct-developing frogs do not need ponds for their larval stages and can therefore colonize habitats that are inaccessible to other frogs. Direct development also occurs in other phyla, in which it is also correlated with a large egg. It seems that if nutrition can be provided in the egg, the life cycle need not have a food-gathering larval stage.

Tadpole-rearing behaviors

Most temperate-zone frogs do not invest time or energy in providing for their tadpoles. However, among tropical frogs, there are numerous species in which adult frogs take painstaking care of their tadpoles. An example is the poison dart frog Dendrobates, found in the rain forests of Central and South America. Most of the time, these highly toxic frogs live in the leaf litter of the forest floor. After the eggs are laid in a damp leaf, a parent (sometimes the male, sometimes the female, according to the species) stands guard over the eggs. If the ground gets too dry, the frog will urinate on the eggs to keep them moist. When the eggs mature into tadpoles, the guarding parent allows them to wriggle onto its back (Figure 3A). The parent then climbs into the canopy until it finds a bromeliad plant with a small pool of water in its leaf base. Here it deposits one of its tadpoles, then goes back for another, and so on until the entire brood has been placed into numerous small pools. The female returns each day to these pools and deposits a small number of unfertilized eggs into them, thus replenishing the tadpoles’ food supply until they complete metamorphosis (Mitchell 1988; van Wijngaarden and Bolanos 1992; Brust 1993). It is not known how the female frog remembers—or is informed about—where the tadpoles have been deposited.

Figure 3   Parental care of tadpoles. (A) Tadpoles of the poison dart frog Dendrobates are carried on their parent’s back to small pools of water in the Peruvian rain forest canopy. (B) This female Rheobatrachus of Australia brooded more than a dozen tadpoles in her stomach. They emerged after completing metamorphosis. Unfortunately, the last time any- one saw a Rheobatrachus frog alive was in 1985. (A © Michael Doolittle/Alamy; B courtesy of M. Tyler.)

Brooding frogs carry their developing eggs in depressions in their skin. Some species brood their tadpoles in their mouth and spit out their progeny when their tadpoles undergo metamorphosis. Even more impressive, the gastric-brooding frogs of Australia, Rheobatrachus silus and R. vitellinus, eat their eggs. The eggs develop into larvae, and the larvae undergo metamorphosis in the mother’s stomach. About 8 weeks after being swallowed alive, about two dozen small frogs emerge from the female’s mouth (Figure 3B; Corben et al. 1974; Tyler 1983). What stops the Rheobatrachus eggs from being digested or excreted? It appears that the eggs secrete prostaglandins that stop acid secretion and prevent peristaltic contractions in the stomach (Tyler et al. 1983). During this time, the stomach is fundamentally a uterus, and the frog does not eat. After the oral birth, the parent’s stomach morphology and function return to normal. Unfortunately, both of these remarkable frog species are now believed to be extinct. No member of either Rheobatrachus species has been seen since the mid-1980s.

Literature Cited

Alberch, P. and J. Alberch. 1981. Heterochronic mechanisms of morphological diversification and evolutionary change in the neotropical salamander Bolitoglossa occidentalis (Amphibia: Plethodontidae). J. Morphol. 167: 249–264.

Brust, D. G. 1993. Maternal brood care by Dendrobates pumilio: A frog that feeds its young. J. Herpetol. 26: 102–105.

Callery, E. M. and R. P. Elinson. 2000. Thyroid hormone-dependent metamorphosis in a direct developing frog. Proc. Natl. Acad. Sci. USA 97: 2615-20.
PubMed Link

Callery. E. M., H. Fang and R. P. Elinson. 2001. Frogs without polliwogs: Evolution of anuran direct development. BioEssays 23: 233–241.
PubMed Link

Corben, C. J., M. J. Ingram and M. J. Tyler. 1974. Gastric brooding: Unique form of parental care in an Australian frog. Science 186: 946–947.
PubMed Link

De Beer, G. 1940. Embryos and Ancestors. Clarendon Press, Oxford.

del Pino, E. M. and R. P. Elinson. 1983. A novel development pattern for frogs: Gastrulation produces an embryonic disk. Nature 306: 589–591.

Duellman, W. E. and L. Trueb. 1986. Biology of Amphibians. McGraw-Hill, New York.

Elinson, R. P. 1987. Change in developmental patterns: Embryos of amphibians with large eggs. In R. A. Raff and E. C. Raff (eds.), Development as an Evolutionary Process. Alan R. Liss, New York, pp. 1–21.

Frieden, E. 1981. The dual role of thyroid hormones in vertebrate development and calorigenesis. In L. I. Gilbert and E. Frieden (eds.), Metamorphosis: A Problem in Developmental Biology. Plenum, New York, pp. 545–564.

Gould, S. J. 1977. Ontogeny and Phylogeny. Harvard University Press, Cambridge, MA.

Hanken, J., M. W. Klymkowsky, C. H. Summers, D. W. Seufert and N. Ingebrigtsen. 1992. Cranial ontogeny in the direct-developing frog, Eleutherodactylus coqui (Anura: Leptodactylidae), analyzed using whole-mount immunohistochemistry. J. Morphol. 211: 95–118.
PubMed Link

Huxley, J. 1920. Metamorphosis of axolotl caused by thyroid feeding. Nature 104: 436.

Lynn, W. G. and A. M. Peadon. 1955. The role of the thyroid gland in direct development of the anuran Eleutherodactylus martinicenis. Growth 19: 263–286.
PubMed Link

Mitchell, A. W. 1988. The Enchanted Canopy. Macmillan, New York.

Norris, D. O., R. E. Jones and B. B. Criley. 1973. Pituitary prolactin levels in larval, neotenic, and metamorphosed salamanders (Ambystoma tigrinum). Gen. Comp. Endocrinol. 20: 437–442.
PubMed Link

Prahlad, K. V. and L. E. DeLanney. 1965. A study of induced metamorphosis in the axolotl. J. Exp. Zool. 160: 137–146.
PubMed Link

Safi, R. and 8 others. 2004. The axolotl (Ambystoma mexicanum), a neotenic amphibian, expresses functional thyroid hormone receptors. Endocrinol. 145: 760–762.
PubMed Link

Safi, R. and 10 others. 2006. Pedomorphosis revisited: Thyroid hormone receptors are functional in Necturus maculosus. Evol. Dev. 8: 284–292.
PubMed Link

Taigen, T. L., F. H. Plough and M. M. Stewart. 1984. Water balance of terrestrial anuran (Eleutherodactylus coqui) eggs: Importance of paternal care. Ecology 65: 248–255.

Taurog, A., C. Oliver, R. L. Eskay, J. C. Porter and J. M. McKenzie. 1974. The role of TRH in the neoteny of the Mexican axolotl (Ambystoma mexicanum). Gen. Comp. Endocrinol. 24: 267–279.
PubMed Link

Tyler, M. J. 1983. The Gastric Brooding Frog. Croom Helm, London.

Tyler, M. J., D. J. Shearman, R. Franco, P. O’Brien, R. F. Seamark and R. Kelly. 1983. Inhibition of gastric acid secretion in the gastric brooding frog, Rheobatrachus silus. Science 220: 607–610.
PubMed Link

van Wijngaarden, R. and F. Bolanos. 1992. Parental care in Dendrobates granuliferus (Anura, Dendrobatidae) with a description of the tadpole. J. Herpetol. 26: 102–105.

Vlaeminck-Guillem, V., R. Safi, P. Guillem, E. Leteurtre, M. Duterque-Coquillaud and V. Laudet. 2006. Thyroid hormone receptor expression in the obligatory paedomorphic salamander Necturus maculosus. Int. J. Dev. Biol. 50: 553–560.
PubMed Link