Figure 3 Estimated somatic cell-type number (except nerve cells) required by the body plans of various taxa, plotted against the time of origin of those taxa indicated by the fossil record. The irregular curved line represents the upper bound of cell-type number generated by a stochastic model based on passive increase or decrease of cell-type number within each lineage during diversification of lineages from 1 to 2000; the line is an average of five runs. [Reproduced with permission, from Valentine et al., 1994 (copyright Paleobiology).]

Cambrian explosion, some metazoan bodies were as complex as primitive arthropods and other higher invertebrates. This rise in complexity is obscured in the fossil record. Perhaps the best practical index of body-plan complexity is cell-type number (Bonner, 1965; Sneath, 1964; Raff and Kaufman, 1983), which may be taken to have begun as two in metazoans and to have increased through time as more complex bodies evolved. Figure 3 depicts the estimated cell-phenotype numbers of the more complex body plans known at a given time during the history of the metazoans, plotted against the times of their first appearances as judged by the fossil record (Valentine et al., 1994). A curve embracing the points should approximate the upper bound of body-plan complexity. Complexity increases may be forced, perhaps by natural selection, or may be passive, resulting from random opportunities to become either more or less complex (Fisher, 1986; Gould, 1988; McShea, 1991,