SUMMARY AND CONCLUSION

Genes which regulate differentiation of scales and patterns of scale types within developmental compartments or lineage restriction zones of Heliconius wings account for major adaptive changes in aposematic pattern between races and species in at least one clade of these neotropical mimetic insects. Addition of early expressing regulatory genes, possibly even homeotics, to the tool box of pattern genetics may be one way that bold, qualitative steps in morphology and behavior evolve in this genus. There is no evidence that mutation has played a role recently in providing the variation needed for mimetic evolution in the MCS clade. Given known rules of the pattern tool box described above and assuming interracial and interspecific introgression, it is possible to account for all the wing patterns so far seen in this group, including many that have not been seen in nature.

In Heliconius, cycles of population differentiation, diversification, hybridization, and introgressive recombination constitute positive feedback systems in which diversity of genotypes promotes diversity of races and species and vice versa (see Figure 10). Yet at any instant, only so many patterns are displayed by a given species or clade in spite of the fact that the feedback system described has been at work through the much of the history of the genus. Where are all the past patterns generated by this process? They may have been lost, like alleles that have been lost through drift or by selection. Another possibility is that extinct patterns are present but suppressed, and recovered only as recombinants in hybrid zones. Thus we expect to, and do, recover patterns that may have been part of past mimicry systems as suggested by Linares (1997a). But it is also possible that the pattern tool box, at least as described for the MCS clade, has the potential to generate virtually any known pattern from the few basic components described: three scale types, displayed as components of alternative patterns which compete to varying degree with ancestral nymphaline pattern elements on compartments of the wing. All that is required to generate many basic patterns of this clade is the presence of several distinct species capable of occasionally exchanging genes, and that's what we usually have.

Finally, it seems likely that similar developmental genetic innovations to those which promote novelty on Heliconius wings, but which instead affect more cryptic attributes of behavior or morphology, may underlie intra-generic adaptive radiation in many groups of organisms. Increasing the number of coexisting species per genus is a major way that species diversity increases towards tropical latitudes. Therefore, the genus Heliconius is an ideal model system for helping us understand how such planet-level patterns come about, and, as Figure 10 suggests, provides an opportunity to develop an integrative theory of biodiversity.