Parallel evolution of character displacement driven by competitive selection in terrestrial salamanders

Abstract

Background: Parallel evolution can occur when common environmental factors exert similar selective forces on morphological variation in populations in different geographic localities. Competition can also generate morphological shifts, and if competing species co-occur in multiple geographic regions, then repeated instances of competitively-driven morphological divergence (character displacement) can occur. Despite the importance of character displacement for inferring the role of selection in morphological evolution however, replicated instances of sympatric morphological divergence are understudied.

Results: I tested the hypothesis that interspecific competition generated patterns of parallel morphological divergence in multiple geographic locations where two competing salamander species, Plethodon jordani and P. teyahalee, come into contact. I used geometric morphometrics to characterize head shape and found ecological character displacement in sympatric localities on each of three distinct mountains (geographic transects), where sympatric specimens displayed greater cranial differences and an increase in cranial robustness as compared to allopatric specimens. Using a recently developed analytical procedure, I also found that the observed morphological evolution within each species was consistent among transects; both in the total amount of morphological change as well as the direction of evolution in the morphological data space. This provided strong statistical evidence of parallel morphological evolution within species across replicate geographic transects.

Conclusions: The results presented here reveal that the morphological evolution of each species followed a common evolutionary path in each transect. Because dispersal between sympatric locations among transects is unlikely, these findings suggest that the repeated instances of character displacement have evolved in situ. They also suggest that selection from competitive interactions plays an important role in initiating sympatric morphological divergence in these species, and drives parallel sympatric morphological divergence between species.

Figure 1

Map of study region. Upper Left: Map of the eastern United States. Right Center: Map of the Great Smoky Mountain National Park, with collection localities from the USNM collections shown (red: P. jordani allopatric localities; blue: P. teyahalee allopatric localities; black: sympatric localities for both species). Lower Left: Map of the study region, with the locations of three geographic transects circled: left = Taywa Creek localities; center: Kephart Prong localities; right:, Heintooga Ridge localities. Allopatric and sympatric localities are designated with the same colors as in the central figure. The boundary of the Great Smoky Mountain National Park is shown, and the light line denotes 1,200 m in elevation.

Figure 2

Landmarks used to characterize head shape of salamanders. Positions of 12 anatomical landmarks used to quantify head shape in Plethodon (image from [44]).

Figure 3

Principal components plot of head shape variation. (A) Data for Plethodon jordani are emphasized, and (B) data for P. teyahalee are emphasized. The first two principal component axes explain 48.6% of the total variation in head shape. For each panel, individuals of the focal species are shown in large, gray symbols, the other species in light symbols. Also displayed are least-squares means for each population, and vectors of phenotypic evolution for each geographic transect. Geographic transects are shown as: Heintooga Ridge (solid lines and circles), Taywa Creek (dashed lines and squares), Kephart Prong (dotted lines and triangles). (C) Phenotypic evolution vectors for each transect for both species in this study. Phenotypic means for allopatric populations are shown in gray; phenotypic means for sympatric localities are shown in black. Thin-plate spline deformation grids depict typical head shapes for sympatric populations of P. jordani and P. teyahalee. A single deformation grid represents the allopatric populations for both species.