One promising approach for assessing adaptive landscape topography is the analysis of fitness distributions of hybrids derived from crosses between recently diverged lineages (Wright 1978; Geiger 1988; Whitlock et al. 1995a). The main goals of such analyses should be focused on establishing characteristics of the topographic features of regions separating pairs of divergent populations. For example, the mean fitness of early hybrid generations (i.e., Fj, F2, and backcross hybrids) would indicate whether the average fitness elevations were lower than the presumed fitness peaks. The range of fitnesses among individual hybrid genotypes would help establish the depth of a valley or hole, occupied by the lowest fitness genotypes, or the elevation of a putative ridge, occupied by the highest fitness genotypes. Low fitnesses for all second-generation hybrid genotypes (F2 and backcross) would suggest that a true valley exists between the divergent lineages and the local topography is rugged, or that divergence has proceeded to the point where all hybrid genotypes fall into regions of low fitness (Fig. 1). If divergence had not proceeded too far, the holey landscape model would predict both high and low fitness hybrids (Gavrilets 19976), which would effectively indicate the presence of both a fitness ridge and a hole separating the two divergent lineages.
Hybrid fitness distributions could provide valuable information on local adaptive landscape topography, but there are several cautions that should be considered before proceeding with such experiments. First, this kind of analysis would be best conducted with recently diverged lineages. Focusing on closely related taxa would improve the chance that the information obtained pertains to a specific portion of the local topography and would avoid instances where divergence had proceeded too far to detect intervening ridges (Fig. 1). Secondly, the taxa being compared need to be from similar environments, and fitness assessments need to be conducted in their native habitats. Controlling for environmental variation as much as possible will help ensure that valid comparisons are being made across a single landscape and circumvent complications due to additive genetic variation for adaptation to different habitats. In this paper we provide an example of an analysis of fitness distributions of hybrids between two closely related and ecologically similar groups of populations. Our analyses endeavor to establish the basic characteristics of the topographical features separating two putative adaptive peaks and to identify the primary modes of adaptive evolution leading to diversification in this group.
Was this article helpful?