DISCUSSION AND CONCLUSIONS (continued)

4. What are the relative contributions of mutation and introgression to pattern evolution?

My answer to this question is obvious from the topic and content of this paper. With respect to mutations, I have cultivated Heliconius populations for 29 years and conducted long term mark-release-recapture studies in the field. I estimate that I have personally examined several tens of thousands of Heliconius in that time. Although I have found several interesting (but non-transmissible) homeotic alterations of the wing pattern, I have spotted only three apparent mutants, a white-eyed H. cydno, a white banded H. charitonius which appears to be an alteration of color but not pattern, and one H. cydno with a heritable modification of the FW shutter. Kapan (personal communication) has noted two dominant color mutants (yellow to white in H. eleuchia), at the same location in Western Ecuador. Interpreting this as one event with two offspring captured among his total sampling effort indicates an approximate 1:4000 probability for the occurrence of this mutation.

Virtually all novel and heritable alterations of wing scale pattern seen in my cultures over three decades have proved to be the results of introgression as discussed in this paper. In the field or in museum collections, and with the exception of non-heritable developmental anomalies, I have yet to see a phenotype of the MCS clade that is not clearly interpretable as a product of introgression between adjacent races or between sympatric species. Mallet et al (1999) have documented the fact that hybridization is common in Heliconius. Yet because mutations of genes affecting wing pattern do occur, why don't mutations constitute a significant source of adaptive variation in this genus?

To answer this question it is useful to consider the probable fate of pattern mutants in natural populations. Because of their long adult lifespan and sedentary behavior, Heliconius are capable of persisting as small, local demes with densities at less than one adult per hectare (Gilbert 1991). Benson's (1972) experimental studies of selection suggest that if dominant or epistatic mutants for pattern novelty arise in such small populations they would be removed rapidly by birds. However, during episodes of environmental change predator pressure may be relaxed (see Turner 1984). But even in this context, pattern mutants that are recessive are likely to be lost by genetic drift before they can support an episode of rapid pattern evolution. If a novel recessive pattern allele is fixed by drift in a small local population during an episode with no predator pressure, it is possible that the entire population would quickly be eliminated by birds which, having learned other aposematic patterns elsewhere, re-colonize the patch and begin to sample unrecognized patterns. >

Thus, one compelling feature of introgression in the system described here is that the variation it generates is renewable and continues generating macro-mutation-like variants in spite of genetic drift, strong selection by local predators, or sequential combinations of both. Moreover, these "macro-mutants" are likely to be compatible in any MCS "operating system." Clearly, introgressive recombination will be orders of magnitude more likely than basic mutational processes to generate the type of genetic variation necessary for rapid adaptive responses to shifts in adaptive peaks for predator protection. Hybridization as a source of novelty has been reviewed by Maynard Smith (1982) and of adaptive variation in insects has been suggested for Drosophila (Lewontin and Birch 1966) and for Heliconius (Linares 1989, 1997b).