The many facets of diversity are an enduring source of fascination for biologists. In my dissertation, I present three case studies exploring how species interactions affect the generation, distribution, and maintenance of biological diversity.
My work centers on a guild of specialized herbivores that consume creosote bush (Larrea tridentata) in the warm deserts of North America. The nominal species L. tridentata comprises three insipient species that arose through autopolyploidy (whole genome duplication without hybridization). Speciation by polyploidy can have manifold consequences for plants. One common outcome is that cytotypes (plants of different ploidy level) geographically segregate. Indeed, the diploid, tetraploid, and hexaploid cytotypes of creosote bush are parapatrically distributed in the Chihuahuan, Sonoran, and Mojave Deserts, respectively. A second outcome is that the phenotypic changes accompanying polyploidy alter interactions with other species, including mutualists, competitors, and consumers. My first two chapters consider how the consequences of polyploidy affect the biogeography and genetic divergence of specialized herbivores.
In Chapter 1, I surveyed the diversity of gall midges in the genus Asphondylia across the full distribution of creosote bush (>2,300 km extent). I found that many gall species (6/17, including two new species) are specialized on either diploid or polyploid host plants. These species preferentially attack just one host cytotype even where diploid and tetraploid creosote bush commingle. The contact zone between creosote bush cytotypes is thus a biotically-mediated dispersal barrier, irrespective of abiotic factors or physical distance. Because more gall midges species specialize on polyploid host plants than on diploids, the distribution of creosote bush cytotypes structures a gradient in gall midge richness across the North American deserts.
I next addressed the hypothesis that adaptation to alternative host plant cytotypes drives reproductive isolation between herbivore populations. In Chapter 2, my investigation focused on two ecologically similar herbivores, the creosote bush grasshopper (Bootettix argentatus) and the creosote bush katydid (Insara covilleae). I first used RADcap sequencing to genotype hundreds of individuals from across the full distribution of each species. I then inferred range-wide population structure and modeled spatial variation in gene flow across the species’ range. To discern the relative role of ecology vs. geography in structuring insect populations, I then compared regions of reduced gene flow to physical barriers as well as creosote bush contact zones. My findings differed markedly between species. Bootettix comprises three major lineages, each associated with a different creosote bush cytotype. Two Bootettix lineages meet in a short hybrid zone that is coincident with the contact zone between tetraploid and hexaploid creosote bush, and gene flow was more restricted at this contact zone that across any physical barriers (e.g., mountain ranges, rivers). In contrast, Insara shows minimal population structure across most of its distribution. Although I found moderate subdivision across the Madrean Archipelago, evidence for host-associated genetic divergence was weak. This work shows that there are not necessarily uniform effects of host plant polyploidy even on ecologically similar herbivores.
In Chapter 3, I consider the forces that maintain adaptive genetic and phenotypic variation within populations. Two discrete color morphs of the desert clicker grasshopper (Ligurotettix coquilletti) coexist across the species’ range: a uniform morph with homogeneous color, and a banded morph with strong light-dark contrast. This patterning likely affects the efficacy of crypsis to avoid predation, raising the question of why polymorphism persists, and why the proportion of each morph varies across sites. One possibility lies in life history differences between sexes: while territorial males are exposed to predation only on creosote bush stems, females are most vulnerable while laying eggs on the ground. I hypothesized that selection may thus favor different cryptic phenotypes in each sex (sexually antagonistic selection), thereby maintaining crypsis polymorphism within populations. With quantitative analysis of crypsis and the visual environments used by male and female grasshoppers, I found that different phenotypes are favored in each sex. I then used RADcap sequencing and association mapping to identify a single autosomal locus strongly associated with pattern polymorphism. Surprisingly, the divergent alleles at this locus are also found in a sister species, suggesting that sexually antagonistic selection may preserve ancient genetic variation.