The trochophora concept and the literature on cleavage patterns and differentiation of ectodermal structures in annelids (''polychaetes'') and molluscs are reviewed. The early development shows some variation within both phyla, and the cephalopods have a highly modified development. Nevertheless, there are conspicuous similarities between the early development of the two phyla, related to the highly conserved spiral cleavage pattern. Apical and cerebral ganglia have almost identical origin in the two phyla, and the cell-lineage of the prototroch is identical, except for minor variations between species. The cell-lineage of the metatrochs is almost unknown, but the telotroch of annelids and the ''telotroch'' of the gastropod Patella originate from the 2d-cell, as does the gastrotroch in the few species which have been studied. The segmented annelid body, i.e. the region behind the peristome, develops through addition of new ectoderm from a ring of 2d-cells just in front of the telotroch. This whole region is thus derived from 2d-cells. Conversely, the mollusc body is covered by descendants of cells from both the C and D quadrants and a growth zone is not apparent. This supports the notion that the molluscs are not segmented like the annelids, and that the repeated structures seen in polyplacophorans and monplacophorans do not represent a segmentation homologous to that of the annelids.