A COMMENT ON GENE INTROGRESSION VERSUS EN MASSE CYCLE SWITCHING IN THE EVOLUTION OF 13-YEAR AND 17-YEAR LIFE CYCLES IN PERIODICAL CICADAS

Abstract

In companion papers, Simon et al. (2000) and Marshall and Cooley (2000) discuss geographic distributions of phenotypic characters and genetic markers of Magicicada decim spp. in the lower Mississippi valley and vicinity. Periodical cicadas (Magicicada spp.) of eastern North America have spurred many debates concerning the evolutionary origin of their unusual 13-year and 17-year life cycles, and instantaneous speciation through en masse cycle switching has been a recurring hypothesis. En masse switching both from 13to 17-year cycles and from 17to 13-year cycles are components of the cycle-switching model of Magicicada evolution (e.g., Lloyd and Dybas 1966a, 1966b; Lloyd and White 1976; Simon 1979, 1988; Martin and Simon 1988, 1990; Williams and Simon 1995). The new field data of Marshall and Cooley (2000) document character displacement between 13-year cicadas with genotypes similar to northern 17-year populations, and sympatric 13-year cicadas with genotypes similar to southern 13-year populations. Simon et al. (2000) interpret this character displacement and other new genetic data as support for the hypothesis of instantaneous speciation by en masse cycle switching from 17to 13-year cycles. However, it is our view that these authors do not adequately explore conventional gene flow explanations of the observed phenotypic and genetic variation that do not require en masse cycle switching. Evaluation of these competing hypotheses can benefit from an interdisciplinary perspective incorporating paleoecological data. We (Cox and Carlton 1988, 1991, 1998; Cox 1992) previously proposed an alternative selection-gene introgression model for Magicicada evolution incorporating paleoecological reconstructions that also explains these new genetic and phenotypic observations of Simon et al. (2000) and Marshall and Cooley (2000). Our hypothesis offers an important caveat about the use of mtDNA in certain circumstances to interpret the magnitude of gene flow between populations and evolutionary processes. We hypothesize unidirectional gene flow via dispersing male cicadas that transferred limited nuclear DNA and no mtDNA to the recipient population. In brief, our model presented in Cox and Carlton (1988) proposes that life-cycle timing differences between cicada populations arose not from en masse cycle switching, but from stochastic selection processes involving Magicicada temperature thresholds, Pleistocene climate characteristics, and hybridization frequencies between populations having a spectrum of life-cycle lengths. Heath (1967, 1968) reports that periodical cicadas emerge from their burrows when soil temperature is 64.3?F and that the air temperature must reach 68?F for flight and mating. In our quantitative model, during glacial maximum it was much more common than it is now for summers to remain too cold for ancestral periodical cicadas to emerge and/or mate, and the ancestors of periodical cicada populations must have failed to reproduce during those years. This seasonal capriciousness gave rise to periodicity (years without cicadas) and selection of long life cycles (conferring a reduced chance of emerging on a cool summer). By the parameters of our model, selected life cycles would be ? 12 years and would be longer in colder climates. This aspect of our model explains the climatically correlated zonation of Magicicada life cycles (Fig. 1). Once periodicity and long life cycles were established, satiation of predators by synchronized emergence became an important strategy for periodical cicadas. Predator satiation works best with precisely synchronized periodicity, and cycle synchronization is diminished by increased heterozygosity. Thus, cicadas that are hybrids of two populations with different cycle lengths will suffer greater predation losses, as many may emerge on years before or after the main population. Cicadas with prime-numbered cycles (13 years and 17 years) will hybridize significantly less frequently than cicadas with non-prime (composite) cycles and thus will have larger emergences and a greater advantage of predator satiation. See Cox and Carlton (1988) for a full discussion of this topic. Marshall and Cooley (2000) studied M. decim spp. of southern 13-year Brood XIX (each brood consists of all synchronized individuals) within the contact zone with more northern 17-year broods (M. septendecim). These authors document character displacement of call pitch between male 13year cicadas with abdominal coloration like cicadas of contiguous 17-year populations (new 13-year cicadas, M. neotredecim) and sympatric male 13-year cicadas with coloration like contiguous Brood XIX populations in the lower Mississippi valley (old 13-year cicadas, M. tredecim). Marshall and Cooley (2000, see their Figs. 4 and 9) argue that M. tredecim have never been present within the range M. neotredecim north of the contact zone and that these two species have hybridized only along the contact zone at the southern margin of the M. neotredecim range. The steep gradient in call pitch frequency along the southern margin of M. neo-