Variation in reproductive isolation across a species range

Abstract

Reproductive isolation is often variable within species, a phenomenon that while largely ignored by speciation studies, can be leveraged to gain insight into the potential mechanisms driving the evolution of genetic incompatibilities. We used experimental greenhouse crosses to characterize patterns of reproductive isolation among three divergent genetic lineages of Campanulastrum americanum that occur in close geographic proximity in the Appalachian Mountains. Substantial, asymmetrical reproductive isolation for survival due to cytonuclear incompatibility was found among the lineages (up to 94% reduction). Moderate reductions in pollen viability, as well as cytoplasmic male sterility, were also found between some Mountain populations. We then compared these results to previously established patterns of reproductive isolation between these Mountain lineages and a fourth, widespread Western lineage to fully characterize reproductive isolation across the complete geographic and genetic range of C. americanum. Reproductive isolation for survival and pollen viability was consistent across studies, indicating the evolution of the underlying genetic incompatibilities is primarily determined by intrinsic factors. In contrast, reproductive isolation for germination was only found when crossing Mountain populations with the Western lineage, suggesting the underlying genetic incompatibility is likely influenced by environmental or demographic differences between the two lineages. Cytoplasmic male sterility was also limited in occurrence, being restricted to a handful of Mountain populations in a narrow geographic range. These findings illustrate the complexity of speciation by demonstrating multiple, independent genetic incompatibilities that lead to a mosaic of genetic divergence and reproductive isolation across a species range.

Keywords: Campanula; cytonuclear incompatibility; genetic distance; reproductive isolation; speciation.

Figure 1

(a) Approximate distribution of each genetic lineage of C. americanum in the eastern United States. (b) Location of C. americanum populations used for evaluating RI; populations shaded according to genetic clade. (c) Crossing schematic for obtaining parental (P), within‐clade (W), and between‐clade (B) seed, with reciprocal crosses marked as 1 and 2. The shaded boxes correspond to the two main crossing groups

Figure 2

Performance of C. americanum hybrids relative to parental populations calculated as the percent reduction in hybrid values relative to the parents, such that negative numbers represent RI. Mean hybrid germination (a), survival (b), and cumulative fitness (c) for each cross graphed against chloroplast‐genetic distance. r² values obtained from regression analysis of relative between‐clade hybrid values against chloroplast‐genetic distance. Smaller gray triangles depict relative hybrid performance for between‐clade crosses from the earlier range‐wide study.^† p < .07; **p < .01

Figure 3

Survival (a) and cumulative fitness (b) of between‐clade C. americanum hybrids relative to parental populations for each crossing direction graphed against chloroplast‐genetic distance. Negative values represent RI. r² values obtained from regression analysis of survival and cumulative fitness for each crossing direction against chloroplast‐genetic distance

Figure 4

Patterns of male sterility in four C. americanum between‐clade crosses. For each cross, the number of hybrid individuals is shown that were male sterile (no viable pollen) or male fertile (viable pollen). Data are only shown for the direction of the cross where the Smoky populations were maternal, as male sterility did not occur in the other crossing direction. Population IDs are listed for each cross, as well as the type of between‐clade cross to which they belong

Figure 5

Example of a hybrid variegated C. americanum rosette resulting from a cross between genetic lineages