A HIERARCHICAL ANALYSIS OF GENETIC DIFFERENTIATION IN A MONTANE LEAF BEETLE CHRYSOMELA AENEICOLLIS (COLEOPTERA: CHRYSOMELIDAE)

Abstract

Abstract. -Herbivorous insects that use the same host plants as larvae and adults can have a subdivided population structure that corresponds to the distribution of their hosts, Having a subdivided population structure favors local adaptation of subpopulations to small-scale environmental differences and it may promote their genetic divergence. In this paper, I present the results of a hierarchical study of population structure in a montane willow leaf beetle, Chrysomela aeneicollis (Coleoptera: Chrysomelidae). This species spends its entire life associated with the larval host (Salix spp.), which occurs in patches along high-elevation streams and in montane bogs. I analyzed the genetic differentiation of C. aeneicollis populations along three drainages in the Sierra Nevada mountains of California at five enzyme loci: ak-1, idh-2, mpi-L, pgi-1, and pgm-L, using recent modifications of Wright's F-statistics. My results demonstrated significant differentiation (FST = 0.043) among drainages that are less than 40 kilometers apart. One locus, pgi-1, showed much greater differentiation than the other four (FST = 0.412), suggesting that it is under natural selection. C. aeneicollis populations were also subdivided within drainages, with significant differentiation I) among patches ofwillows (spanning less than three kilometers) and 2) in some cases, among trees within a willow patch. My results demonstrate that this species has the capacity to adapt to local environmental variation at small spatial scales.Key words.-Allozymes, Chrysomela aeneicollis, F-statistics, gene flow, genetic differentiation, insect population structure, natural selection, phosphoglucose isomerase. Received February 4, 1991. Accepted November 20, 1991 All species consist ofpopulations that vary in size and degree of isolation from other populations. Local population sizes, migration rates, and the patterns ofmatings within a population all affect a species' population structure (Hartl and Clark, 1989;McCauley and Eanes, 1987;Slatkin, 1987). In many cases, local environmental variation causes natural selection to operate differently among local populations, and populations may differ genetically in response to this natural selection. Even in the absence ofselection, genetic drift among small-sized populations increases their likelihood ofbecoming genetically differentiated. On the other hand, gene flow tends to make populations more genetically homogeneous (Slatkin, 1987). To measure the genetic structure ofpopulations, Wright (1978) proposed that studies should use a hierarchical sampling design that corresponds to spatial subdivisions ofnatural populations. For ex-1 Present address: Institute fur terrestrische okologie, ETH-Ziirich, Grabenstr. 3, CH-8952 Schlieren, ample, if a herbivorous insect lives on trees that grow in distinct patches separated by geographical barriers, a study ofthat insect's population structure should include samples from individual trees from several host-tree patches. The advantage of a hierarchical study is that one...

Natural Selection as a Population-Level Causal Process

“…Perhaps this is because, as he states at the outset: 'Even in the absence of selection, genetic drift among smallsized populations increases their likelihood of becoming genetically differentiated' (Rank [1992], p. 1097).…”

“…Rank's results could be due to local drift, local microhabitat specialization, or a combination, subject to further study-though genespecific variation in the extent of frequency differences led him to postulate local selection, at least on the PGI gene. In [Rank 1992 and a similar study by Costa and Ross 1994], use of allozymes to probe population structure has ruled out some possibilities, while leaving the need for further specific work to distinguish among other causal scenarios -as the authors have been careful to state' (Watt [1995], p. 878; emphasis added).…”

“…Watt urges biologists to seek the causal basis for claims about selection, and in this way provide a more solid basis than the statistical analysis of population genetics alone can provide. In this context, he discusses Rank's ([1992]) study of the montane leaf beetle in particular: 11 I use the terms 'process' and 'outcome' rather than the more common 'process' and 'product' to avoid the teleological implications of the term 'product.' 'Rank (1992)…found substantial local allozyme frequency differentiation, even among adjacent drainages in the same mountain range.…”

“…In this context, he discusses Rank's ([1992]) study of the montane leaf beetle in particular: 11 I use the terms 'process' and 'outcome' rather than the more common 'process' and 'product' to avoid the teleological implications of the term 'product.' 'Rank (1992)…found substantial local allozyme frequency differentiation, even among adjacent drainages in the same mountain range. Rank's results could be due to local drift, local microhabitat specialization, or a combination, subject to further study-though genespecific variation in the extent of frequency differences led him to postulate local selection, at least on the PGI gene.…”

“…The montane willow leaf beetle is indigenous to western North America from the Yukon to California, and it can be found on willow shrubs or trees that commonly grow in separated patches of boggy or moist habitats (Rank [1992]). It can fly, but it rarely does so; thus, there is limited migration between localities.…”