analysis (these axes explain more than 95 percent of the total information). The snails are defined by seven ratio measures of shape; shell size does not enter the analysis at all. The three groups include: tiny juvenile shells (Y) of a nonpaedomorphic ancestor, adult shells (A) from the same nonpaedomorphic sample, and adult shells (P) of a paedomorphic descendant (same size as the nonpaedomorphic adults). Ancestral juveniles and adult paedomorphs group together; adult nonpaedomorphs are separate. The paedomorphs are truly scaled up replicas—in shell form, shell thickness, and color—of ancestral juveniles.

This iterative paedomorphosis is the only evolutionary event in the recorded history of P. bermudensis, and it happens again and again. Moreover, its adaptive trigger can be identified in the almost totally lime-free soils that served as substrate for the most paedomorphic forms. (Most nonpaedomorphs lived in carbonate dunes.) Land snails build their shells from ingested lime and the correlation of abundant snails with limestone substrates is among the best documented phenomena of animal ecology. When faced with a great scarcity of lime, land snails either die, move, or grow very thin shells. The juvenile shell of Poecilozonites is paper thin, the adult quite thick. Paedomorphosis must have been the easiest path to the large and rapid reduction of shell thickness that changing environments required.

Frequency of Paedomorphosis in the Origin of Higher Taxa

It is unfortunate that most literature on paedomorphosis is cast in the same mold that bolstered recapitulation during the previous half century—speculative phylogeny of higher taxa. There is scarcely a major group of animals that has not inspired a paedomorphic theory for its origin. And when one considers that insects (de Beer, 1958, derives them from myriapod larvae), copepods (Gurney, 1942, pp. 22–26; Noodt, 1971; from a larva like the decapod protozoea), and vertebrates (Garstang, 1928; Berrill, 1955; Whitear, 1957; Bone, 1960; from hemichordate larvae, tunicate tadpoles, and echinoderm larvae) have been among the most popular candidates, the number of modern paedomorphic derivates could be overwhelming. All phyla have been implicated, from protists (Hadzi, 1952) to higher plants (Carlquist, 1962; Takhtajan, 1969). General supporters of paedomorphosis have supplied long lists: de Beer (1958) favors a paedomorphic origin for, among others: ctenophores, siphonophores, cladocerans, copepods, pteropods, insects, chordates, appendicularian tunicates, ratite birds, hominids, hexacorals, proparian trilobites, and grapto-