Hamel et al. (2008) reported that veliger larvae of the nudibranch gastropod Palio dubia settled 1 to 3 days after hatching and that metamorphosis commenced shortly thereafter. This is an anomalously short larval period for a nudibranch described as having planktotrophic development. I examined the embryonic development and hatching larvae of P. dubia collected intertidally from Maine, USA. Veliger larvae with shells 120 μm long and lacking eyespots and propodia developed in 7 days at 20°C from eggs averaging 69 μm in diameter. Their size and morphology were typical of planktotrophic nudibranchs known to have minimum larval periods of weeks to months, and the available evidence does not suggest any other mode of development exists in P. dubia. The larval period of P. dubia is likely an order of magnitude longer than reported by Hamel et al. (2008).

The ‘origin of larvae’ has been widely discussed over the years, almost invariably with the tacit understanding that larvae are secondary specializations of early stages in a holobenthic life cycle.

Considerations of the origin and early radiation of the metazoan phyla have led to the conclusion that the ancestral animal (=metazoan) was a holopelagic organism, and that pelago-benthic life cycles evolved when adult stages of holopelagic ancestors became benthic, thereby changing their life style, including their feeding biology.

The literature on the larval development and phylogeny of animal phyla is reviewed in an attempt to infer the ancestral life cycles of the major animal groups. The quite detailed understanding of larval evolution in some echinoderms indicates that ciliary filter-feeding was ancestral within the phylum, and that planktotrophy has been lost in many clades. Similarly, recent studies of the developmental biology of ascidians have demonstrated that a larval structure, such as the tail of the tadpole larva, can easily be lost, viz. through a change in only one gene. Conversely, the evolution of complex structures, such as the ciliary bands of trochophore larvae, must involve numerous genes and numerous adaptations.

The following steps of early metazoan evolution have been inferred from the review.

The holopelagic ancestor, blastaea, probably consisted mainly of choanocytes, which were the feeding organs of the organism. Sponges may have evolved when blastaea-like organisms settled and became reorganized with the choanocytes in collar chambers.

The eumetazoan ancestor was probably the gastraea, as suggested previously by Haeckel. It was holopelagic and digestion of captured particles took place in the archenteron. Cnidarians and ctenophores are living representatives of this type of organization. The cnidarians have become pelago-benthic with the addition of a sessile, adult polyp stage; the pelagic gastraea-like planula larva is retained in almost all major groups, but only anthozoans have feeding larvae.

Within the Bilateria, two major lines of evolution can be recognized: Protostomia and Deuterostomia. In protostomes, trochophores or similar types are found in most spiralian phyla; trochophore-like ciliary bands are found in some rotifers, whereas all other aschelminths lack ciliated larvae. It seems probable that the trochophore was the larval type of the ancestral, pelago-benthic spiralian and possible that it was ancestral in all protostomes. Most of the non-chordate deuterostome phyla have ciliary filter-feeding larvae of the dipleurula type, and this strongly indicates that the ancestral deuterostome had this type of larva.