If regeneration were adaptive, it would have arisen autonomously by natural selection from non-regenerative antecedents. Unless each episode coincidentally 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 was incompatible. In vertebrate evolution, these adaptations have included the transition from aquatic to terrestrial habitats and the modification of poikilothermic to homeothermic metabolism. The former advance rendered the regeneration of weight-bearing limbs impractical; the latter favored rapid wound healing and scar formation which effectively precluded blastema formation. If the latent capacity for regeneration persists in non-regenerative appendages, as would seem to be the case, then the restoration of its overt expression should be possible if the mechanisms of its inhibition could be discovered and eventually rendered ineffectual.