Understanding the adaptive divergence of crossbills

One goal when I began research on crossbills was to figure out what seasons or resources were the most critical in the evolution of trophic structure. What characterizes the periods or resources that animals are adapted to exploit? Similarly, when should we focus our studies if we are to understand community structure? These were critical questions that were rarely addressed satisfactorily and they remain important questions in ecology and evolutionary biology. Based on prior field observations and laboratory experiments, I came up with three characteristics of resources necessary for crossbills to specialize and then quantitatively tested this hypothesis with captive representatives of four putative species of Red Crossbill (L. curvirostra complex). I measured their feeding performances on cones that I hypothesized each species was adapted to exploit given the cone characteristics as in the forestry literature. [Here's a VIDEO that shows two call types foraging on western hemlock (Tsuga heterophylla) cones. The first birds are the small-billed call type 3, which are specialized for foraging on hemlock, whereas the last crossbill in the video is a much larger billed crossbill, call type 2. Nearly every time they lift their heads they are husking seeds. It is easy to see that the smaller billed call type is a much faster forager on hemlock cones.] The results were remarkable (Benkman 1993, Benkman and Miller 1996, Benkman et al. 2001). Each of the four taxa of crossbills had either an average bill size (depth) or groove width in their palate where seeds are held, or both, that approximated the predicted optima for foraging on the respective conifers. This indicates that certain characteristics of resources are critical for specialization - not only for crossbills but also for other predators (Thompson, J. N. 1994, The Coevolutionary Process, Univ. Chicago Press).

I extended this work by converting estimates of performance into fitness based on the relationships between morphology and performance in captivity and between morphology and survival in the wild (Benkman 2003). The result was an extremely rugged adaptive surface, which implies that divergent selection for foraging on different resources crossbills specialize upon is driving the adaptive radiation of crossbills (PDF of write-up in Science). This result is of substantial interest for two reasons. First, it provides the only empirically derived three-dimensional adaptive surface for any group of organisms that provides critical insight into an adaptive radiation. Adaptive surfaces, first developed by Sewall Wright in 1932, are considered “the most heuristically valuable diagrams in all of evolutionary biology” (Provine 1986:316) and have been central to virtually all discussions of ecological speciation and adaptive radiation because they capture how natural selection contributes to population divergence, adaptive radiations and ultimately speciation. However, adaptive surfaces are extremely difficult to quantify in a meaningful manner. Second, it is the first study that links measures of performance to fitness in the wild and ultimately evolution. This linkage is central to theoretical arguments that performance can be equated to fitness and to our understanding of adaptations.

The role of coevolution in structuring and diversifying populations

Evolutionary biologists have devoted considerable effort trying to understand coevolutionary arms races and how they contribute to biodiversity. Despite this interest, few studies have demonstrated both reciprocal selection and reciprocal adaptations for coevolving systems in the wild. We recently (Benkman et al. 2003) provided empirical evidence demonstrating the forms of reciprocal selection between two predators and their prey, and, in turn, how this results in the evolution of reciprocal adaptations. Earlier (Benkman 1999, Benkman et al. 2001) we presented various lines of evidence that in sum demonstrate replicated reciprocal adaptations between crossbills and lodgepole pine. Red Crossbills are the principal seed predator and selective agent on Rocky Mountain lodgepole pine (Pinus contorta ssp. latifolia) cones in several isolated mountain ranges lacking red squirrels (Tamiasciurus hudsonicus). In these ranges, crossbills have experienced selection favoring larger bills and have coevolved in response to increases in structural seed cone defenses. In contrast, where red squirrels are present, squirrels are the main selective agent on seed defenses and cones evolve quite differently. The result is the evolutionary divergence of crossbill populations in response to selection for foraging on the different cones in areas with and without squirrels. This divergence between crossbill populations is possibly leading to a new bird species in an isolated range in Idaho (we collected the type specimens, which eventually we will formally describe). Few studies have been able to capture and thoroughly explain the forms of selection on the traits mediating and responding to selection in a coevolutionary arms race. Even fewer studies can directly link the interaction to the formation of new species. Speciation as the result of coevolution is the subject of one of our two main avenues of future research.

South Hills, Idaho where crossbills are coevolving with lodgepole pine.

Coevolution causing ecological speciation and contributing to an adaptive radiation

One of the central problems in evolutionary biology is to understand the processes that lead to new species. Our currently NSF-funded research will determine whether and how the coevolutionary arms race between Red Crossbills and lodgepole pine is leading to what appears to be a new species of Red Crossbill. First, we are using mark-recapture methods to determine whether natural selection because of coevolution has caused the observed morphological divergence between crossbill populations. Second, field studies and aviary experiments conducted by several graduate students, Julie Smith and Lisa Snowberg, are addressing how reproductive isolation (i.e., speciation) might have occurred as a by-product of divergent selection. Finally, one of my graduate students, Thomas Parchman, is using AFLP (Amplified Fragment Length Polymorphism) primers to determine the extent of genetic differentiation between different crossbill populations and to develop a phylogeny of nearly all New World crossbills. This work aims to link two of the most important evolutionary processes, namely coevolutionary arms races and speciation, to patterns of biodiversity and to how geographic patterns of diversity arise. The coevolutionary interactions that have been documented, with the real possibility that they are contributing to the origin of new species, is exciting and of fundamental importance yet has rarely been as well documented for any taxa let alone for birds.

A second focus of our research is to determine the extent to which coevolution between crossbills and conifers has contributed to the adaptive radiation of crossbills. A fundamental challenge of evolutionary biology is to understand the extent to which microevolutionary processes generate macroevolutionary patterns. One of the most prominent hypotheses to explain macroevolutionary patterns is that of a coevolutionary arms race between predators and their prey and other exploiter-victim relationships. Although the details differ between authors, antagonistic evolutionary interactions are thought to be responsible for patterns of diversification of many taxa. Nevertheless, the extent to which coevolution has played a role in the adaptive radiation of any group has not been well quantified.

Our previous research has provided strong evidence of coevolution between Red Crossbills and both lodgepole pine and black spruce (Picea mariana). What is further remarkable is that the form of coevolution between crossbills and lodgepole pine is replicated between crossbills and black spruce.Together these studies show that the presence and absence of a dominant preemptive competitor, red squirrels, determines the geographic selection mosaic and location of coevolutionary hotspots and coldspots for its competitor, the Red Crossbill. Moreover, the patterns of cone evolution in response to selection by crossbills and squirrels are repeated among populations of lodgepole pine, and between lodgepole pine and black spruce (Parchman and Benkman 2002), which enables predictions of coevolution in other systems. Although our understanding of the geographic mosaic of coevolution for crossbills that specialize on lodgepole pine and black spruce is excellent, these findings might not be general to other crossbills specialized on other conifers in North America (and the Old World). There are two reasons to suspect this. First, annual fluctuations in the availability of lodgepole pine and black spruce seed are much less variable and not representative of most conifers; greater variation in the annual availability of seed should limit the competitive and evolutionary impacts of territorial or at least relatively sedentary tree squirrels. Second, Sciurus rather than Tamiasciurus is the tree squirrel that more commonly co-occurs with crossbills, and Sciurus is a weaker competitor than Tamiasciurus because it harvests fewer cones.

Consequently, the main objectives of our study, recently funded by NSF, will be to test two hypotheses. First, whether crossbills coevolve with conifers whose seed availability fluctuates yearly. Second, whether the presence of tree squirrels (Tamiasciurus or Sciurus) does not impede coevolution between crossbills and conifers whose availability of seeds fluctuate. The results of these studies will determine (1) how resource characteristics and competitor type affect the form of the geographic mosaic for crossbills and conifers and (2) to what extent coevolution has contributed to the adaptive radiation of crossbills.