The evolution of host preference in allopatric vs. parapatric populations of Timema cristinae walking-sticks.

Divergent habitat preferences can contribute to speciation, as has been observed for host-plant preferences in phytophagous insects. Geographic variation in host preference can provide insight into the causes of preference evolution. For example, selection against maladaptive host-switching occurs only when multiple hosts are available in the local environment and can result in greater divergence in regions with multiple vs. a single host. Conversely, costs of finding a suitable host can select for preference even in populations using a single host. Some populations of Timema cristinae occur in regions with only one host-plant species present (in allopatry, surrounded by unsuitable hosts) whereas others occur in regions with two host-plant species adjacent to one another (in parapatry). Here, we use host choice and reciprocal-rearing experiments to document genetic divergence in host preference among 33 populations of T. cristinae. Populations feeding on Ceanothus exhibited a stronger preference for Ceanothus than did populations feeding on Adenostoma. Both allopatric and parapatric pairs of populations using the different hosts exhibited divergent host preferences, but the degree of divergence tended to be greater between allopatric pairs. Thus, gene flow between parapatric populations apparently constrains divergence. Host preferences led to levels of premating isolation between populations using alternate hosts that were comparable in magnitude to previously documented premating isolation caused by natural and sexual selection against migrants between hosts. Our findings demonstrate how gene flow and different forms of selection interact to determine the magnitude of reproductive isolation observed in nature.

Migration and the genetic covariance between habitat preference and performance.

Studies of the genetic covariance between habitat preference and performance have reported conflicting outcomes ranging from no covariance to strong covariance. The causes of this variability remain unclear. Here we show that variation in the magnitude of genetic covariance can result from variability in migration regimes. Using data from walking stick insects and a mathematical model, we find that genetic covariance within populations between host plant preference and a trait affecting performance on different hosts (cryptic color pattern) varies in magnitude predictably among populations according to migration regimes. Specifically, genetic covariance within populations is high in heterogeneous habitats where migration between populations locally adapted to different host plants generates nonrandom associations (i.e., linkage disequilibrium) between alleles at color pattern and host preference loci. Conversely, genetic covariance is low in homogeneous habitats where a single host exists and migration between hosts does not occur. Our results show that habitat structure and patterns of migration can strongly affect the evolution and variability of genetic covariance within populations.

Counteracting selective regimes and host preference evolution in ecotypes of two species of walking-sticks.

The evolution of ecological specialization has been a central topic in ecology because specialized adaptations to divergent environments can result in reproductive isolation and facilitate speciation. However, the order in which different aspects of habitat adaptation and habitat preference evolve is unclear. Timema walking-stick insects feed and mate on the host plants on which they rest. Previous studies of T. cristinae ecotypes have documented divergent, host-specific selection from visual predators and the evolution of divergent host and mate preferences between populations using different host-plant species (Ceanothus or Adenostoma). Here we present new data that show that T. podura, a nonsister species of T. cristinae, has also formed ecotypes on these host genera and that in both species these ecotypes exhibit adaptive divergence in color-pattern and host preference. Color-pattern morphs exhibit survival trade-offs on different hosts due to differential predation. In contrast, fecundity trade-offs on different hosts do not occur in either species. Thus, host preference in both species has evolved before divergent physiological adaptation but in concert with morphological adaptations. Our results shed light onto which traits are involved in the initial stages of ecological specialization and ecologically based reproductive isolation.

Reproductive isolation driven by the combined effects of ecological adaptation and reinforcement.

Recent years have seen a resurgence of interest in the process of speciation but few studies have elucidated the mechanisms either driving or constraining the evolution of reproductive isolation. In theory, the direct effects of reinforcing selection for increased mating discrimination where interbreeding produces hybrid offspring with low fitness and the indirect effects of adaptation to different environments can both promote speciation. Conversely, high levels of homogenizing gene flow can counteract the forces of selection. We demonstrate the opposing effects of reinforcing selection and gene flow in Timema cristinae walking-stick insects. The magnitude of female mating discrimination against males from other populations is greatest when migration rates between populations adapted to alternate host plants are high enough to allow the evolution of reinforcement, but low enough to prevent gene flow from eroding adaptive divergence in mate choice. Moreover, reproductive isolation is strongest under the combined effects of reinforcement and adaptation to alternate host plants. Our findings demonstrate the joint effects of reinforcement, ecological adaptation and gene flow on progress towards speciation in the wild.