Variable extent of parallelism in respiratory, circulatory, and neurological traits across lake whitefish species pairs

Abstract

Parallel adaptive radiation events provide a powerful framework for investigations of ecology's contribution to phenotypic diversification. Ecologically driven divergence has been invoked to explain the repeated evolution of sympatric dwarf and normal lake whitefish (Coregonus clupeaformis) species in multiple lakes in eastern North America. Nevertheless, links between most putatively adaptive traits and ecological variation remain poorly defined within and among whitefish species pairs. Here, we examine four species pairs for variation in gill, heart, and brain size; three traits predicted to show strong phenotypic responses to ecological divergence. In each of the species pairs, normals exhibited larger body size standardized gills compared to dwarfs - a pattern that is suggestive of a common ecological driver of gill size divergence. Within lakes, the seasonal hypoxia experienced in the benthic environment is a likely factor leading to the requirement for larger gills in normals. Interestingly, the morphological pathways used to achieve larger gills varied between species pairs from Québec and Maine, which may imply subtle non-parallelism in gill size divergence related to differences in genetic background. There was also a non-significant trend toward larger hearts in dwarfs, the more active species of the two, whereas brain size varied exclusively among the lake populations. Taken together, our results suggest that the diversification of whitefish has been driven by parallel and non-parallel ecological conditions across lakes. Furthermore, the phenotypic response to ecological variation may depend on genetic background of each population.

Keywords: Adaptive radiation; Coregonus clupeaformis; brain; ecology; gill; heart ventricle; morphological divergence; speciation.

Figure 1

Adult dwarf (upper) and normal (lower) lake whitefish.

Figure 2

Estimating gill size in lake whitefish. The figure shows, as an example, one of the four gill arches measured per whitefish. For each hemibranch (i.e., each side of the arch), we took a photograph and used the photograph to digitally estimate the length of every fifth filament, the total filament number, the area of the hemibranch (filamentous portion), and the perimeter of the hemibranch (filamentous portion), as depicted.

Figure 3

Relationships between body mass and five gill size metrics in the lake whitefish. Points corresponding to dwarf and normal whitefish are shown in gray and black, respectively. The regression line is fitted to all points. The fit of the model is shown in each figure, as is the equation indicating how each of (A) gill perimeter (mm), (B) gill hemibranch area (mm²), (C) number of gill filaments, (D) total gill filament length (mm), and (E) average gill filament length (mm) scale with body mass in whitefish. The gill size metrics are plotted against body mass on bilogarithmic scales.

Figure 4

Relationships between body mass and each of heart ventricle mass (A) and brain mass (B) in the lake whitefish. Points corresponding to dwarf and normal whitefish are shown in gray and black, respectively. The regression line is fitted to all points. The fit of the model is shown in each figure, as is the equation indicating how ventricle mass and brain mass scale with body mass in whitefish. Ventricle and brain masses are plotted against body mass on bilogarithmic scales.

Figure 5

Variation in body mass or length standardized principal components (PC) of gill size in dwarf (D-gray circles) and normal (N-black circles) lake whitefish from Cliff Lake (CL) and Indian Pond (IP) in Maine, USA, and East Lake (EL), and Témiscouata Lake (TL) from Québec, Canada. Principal component 1 (PC1) exhibited positive loadings for all five gill size metrics (TGFL, AFL, TFN, THA, THP) when standardized by body mass (A) or length (B), whereas TFN is the primary metric explaining variation in principal component (PC 2) for both body mass (C) and body length (D) standardized metrics. Mean component scores are reported ±1 SE.

Figure 6

Variation in heart ventricle and brain mass in dwarf (D-gray points) and normal (N-black points) lake whitefish from Cliff Lake (CL) and Indian Pond (IP) in Maine, USA, and East Lake (EL), and Témiscouata Lake (TL) from Québec, Canada. Mean relative ventricle (A) and brain (B) masses are shown as a percent of body mass ± 1 SE. Also plotted are least squares means (LSM) of ventricle (C) and brain (D) masses ± 1 SE in the species in each of the four lakes. LSM scores were derived from ANCOVA examining the contribution of lake, species, their interactions, and body mass to variation in ventricle and brain mass.