Garter snake eating newtI strive to understand both the selective forces shaping biodiversity and the genetic processes that translate natural selection into evolutionary changes. My research focuses on interactions at different levels of biological organization that promote phenotypic and genetic integration. Interactions can occur at many levels, from epistasis between loci within individuals to ecological interactions between different species. The results of interaction can range from genetic coadaptation (the evolution of coadapted gene complexes), to developmental integration, to phenotypic coadaptation of species engaged in coevolutionary interactions. Many of the techniques I employ stem from quantitative genetics, but I also incorporate basic fieldwork, behavioral observations, manipulative experiments, and the comparative method. I also use theoretical modeling as a means of understanding some of the more complex sorts of interactions in nature.

Much of my work concentrates on the predatory and antipredator adaptations of reptiles and amphibians, especially behavior and color pattern and the interaction between them. These interests have led me to investigations of predator exploitative ability and the coevolutionary arms races between predators and prey. I also explore indirect genetic effects and social selection through a combination of theoretical work and empirical studies of primarily insect systems.

Predator-prey arms races

My ongoing collaborative work with E. D. Brodie, Jr. (Utah State University) investigates the coevolution of tetrodotoxin (TTX) toxicity in the newt Taricha granulosa and resistance to TTX by the garter snake Thamnophis sirtalis. Our work on this system has taken a number of different directions, but generally explores role of geographic structure in the coevolutionary dynamics of predator-prey arms races. Based on some recent discoveries that TTX resistance has evolved multiple times, we are exploring the process and pattern of parallel evolution of resistance within garter snakes from both a phenotypic and molecular genetic perspective.

Some recent highlights of the project include an investigation of the geographic pattern of TTX resistance in over 40 populations of garter snakes covering much of western North America (Brodie et al. 2002 Evolution). Patterns of phenotypic evolution indicate that there are two areas of coevolutionary hotspots (one in California and one in Oregon) surrounded by areas of less intense reciprocal selection. Predators and prey are closely phenotypically matched in resistance and toxicity levels throughout the northern part of their ranges. In collaboration with Fred Janzen (Iowa State University), we have developed a picture of the phylogeographic relationships among these populations that clearly indicates the two hotspots are independently evolved (Janzen et al. 2002 Molecular Ecology).

Using some of the same populations, Shana Geffeney (Utah State University) has been able to demonstrate that population differences in TTX resistance are in part due to differences in skeletal muscle sodium channel function. Individual variation also exists at this physiological level and correlates closely with whole-animal resistance. We are in the process of cloning the skeletal muscle sodium channel gene in order to determine the molecular basis of differences among populations. This information will allow us to investigate the phenotypic and molecular genetic patterns of parallel evolution.

The Evolution of G-matrices in Anolis lizards

In collaboration with Jonathan Losos (Harvard University), we are looking at the evolution of genetic covariance structure in repeated radiations of Caribbean Anolis. One of the major questions in evolutionary biology is to what extent evolutionary change is dictated by genetic constraints (as capture by the G-matrix), versus how much selection can modify genetic architecture. The repeated evolution of similar ecomorphs of Caribbean Anolis provide an unparalleled opportunity to test these alternatives in the framework of a well understand phylogenetic and adaptive context. The project is just underway, but will involve captive breeding of at least 8 species of Anolis representing 3 common ecomorphs from 3 different lineages.

Natural selection in Horned Lizards

One consistent effort in my lab is to understand the forms and sources of natural selection in the wild. One example of such studies is a collaboration with Kevin Young, a former student at Utah State University, in which we were able to show that predation be shrikes led to directional selection on the horn lengths of flattailed horned lizards in Arizona. I consider this study one of the rare cases wherein we can reveal not only the quantitative strength and shape of a selection function, but also attribute that function to a specific biological agent. The original paper stimulated some interesting conversations about definitions of adaptation and what is necessary to demonstrate adaptation.

Indirect genetics and parental care behavior in the burrower bug

Aneil Agrawal and I are using the burrower bug, Sehirus cinctus, as a model system to explore indirect genetic effects empirically. Burrower bugs exhibit maternal care; females lay clutches of approximately 40 to 150 eggs in shallow burrows in the soil. A female guards her clutch for about 10 days until the eggs hatch. At that time, she begins collecting small mint nutlets (e.g., Lamium spp.) that she deposits in the burrow to provision the offspring. Provisioning is not directed to individual offspring, but rather to the clutch as a whole, and continues through the end of the second stadium (approximately 10 days post-hatching). Care is obligate; unprovisioned clutches do not survive. Although specific offspring signals have not been identified, the influence of offspring on provisioning can be observed (see below). Because a female can produce multiple clutches within a season, it is possible that maternal care expended on one clutch reduces a female's residual reproductive value.

Parental and offspring influences on provisioning are difficult to disentangle because of covariances that are expected to exist between parents and their offspring. Cross-fostering eliminates genetic and phenotypic sources of covariance between parents and the offspring for which they care. Using cross-fostering designs, we have been able to demonstrate that mothers respond to signals from offspring and that maternal provisioning and offspring signaling are genetically correlated. Split clutches receive correlated amounts of provisioning from unrelated foster mothers, indicating genetic variation among offspring in signaling. Experimental manipulation of offspring condition produces changes in the amount of maternal provisioning. Thus both genetic factors and condition influence offspring signaling. Offspring signaling is negatively genetically correlated with maternal provisioning as predicted by maternal-offspring coadaptation theory.

Models of indirect genetic effects (IDEs)

Indirect genetic effects (IGEs) are environmental influences on the phenotype of one individual that are due to the expression of genes in a different, conspecific individual. These are most familiar in the form of maternal effects, where the environment provided by relatives (especially moms) can both be genetically determined AND influence the expression of traits in offspring. Other situations in nature produce similar interactions between unrelated individual (especially with social behaviors, e.g., aggression, dominance, courtship behavior). I am currently working to investigate the evolutionary consequences of IGEs, and interacting phenotypes in general with several collaborators including Aneil Agrawal, Mike Wade, Allen Moore, and Jason Wolf.

We have explored the application of IGEs in a variety of contexts, including behavioral interactions among non-relatives, the effect of IGEs on sexually selected traits, and their importance in generating trait integration between parents and offspring. We have also applied the IGE perspective to social selection, epigenetic models of development, subdivided populations and the implications for group vs individual selection, parental provisioning and the evolution of aposematic signals and mimicry.

Models suggest that interacting traits, such as social behaviors, my respond to selection differently than most traits. When the expression of a behavior in one individual depends on a trait in another individual, genetic covariances between the two traits can play an important role in determining the response to selection for both traits. In many cases, this can lead to dramatically faster evolution, and even the evolutionary response of characters that have no direct additive genetic variance.

Phenotypic integration of behavior and color pattern in garter snakes

I am generally interested in the problem of phenotypic and genetic integration, especially as it results from adaptive processes. Functional interaction may create suites of traits on which selection acts simultaneously. Correlational selection for particular combinations of traits may explain patterns of phenotypic and genetic integration both within and between populations. In some older work, I conducted mark-recapture work in a natural population of the garter snake Thamnophis ordinoides and detected correlational selection for combinations of color pattern and antipredator behavior: individuals with striped patterns that flee directly and those with spotted or unmarked patterns that perform evasive reversals during flight have a higher probability of survival than others (Brodie, 1992, Evolution). Color pattern and escape behavior are genetically correlated in some natural populations of garter snakes (Brodie, 1989, Nature; 1993, Evolution). Theoretical investigations suggest that correlational selection may promote both these genetic correlations and the high level of genetic variation observed in each of the traits separately (Brodie, 1993, Evolution).

Mimicry and coral snakes

Mimicry is a classic problem in evolution with ramifications for almost every major concept from coevolution to speciation to parallel evolution. Mimicry has classically been studied in invertebrate systems that include noxious models and a few very good mimics. I am interested in how dangerous or deadly models affect specific features of mimetic complexes. In the 1990's, I studied coral snake mimicry systems in the field in using soft plasticine replicas as a technique to assess avian predation on different color patterns (Brodie, 1993, Evolution; Brodie and Moore 1995. Animal Behaviour; Brodie and Janzen 1995. Functional Ecology). This methodology was being used to test some predictions about how mimicry complexes operate when models are extra-noxious (e.g., deadly) or when a variety of mimics exist. Other studies have investigated the generalized avoidance of patterns, the possibility that millipedes are involved in the mimicry complex, and the the importance of various stimulus components of the aposematic patterns (e.g., band width, color). I continue to conduct occasional experiments on this problem and am generally interested in questions related to mimicry and aposematism. Aneil Agrawal and I (Brodie and Agrawal 2001. PNAS) recently developed a model showing that the evolution of aposematism is easier if signals are inherited via maternal effects.