Is salamander arboreality limited by broad-scale climatic conditions?

Abstract

Identifying the historical processes that drive microhabitat transitions across deep time is of great interest to evolutionary biologists. Morphological variation can often reveal such mechanisms, but in clades with high microhabitat diversity and no concomitant morphological specialization, the factors influencing animal transitions across microhabitats are more difficult to identify. Lungless salamanders (family: Plethodontidae) have transitioned into and out of the arboreal microhabitat many times throughout their evolutionary history without substantial morphological specialization. In this study, we explore the relationship between microhabitat use and broad-scale climatic patterns across species' ranges to test the role of climate in determining the availability of the arboreal microhabitat. Using phylogenetic comparative methods, we reveal that arboreal species live in warmer, lower elevation regions than terrestrial species. We also employ ecological niche modeling as a complementary approach, quantifying species-level pairwise comparisons of niche overlap. The results of this approach demonstrate that arboreal species on average display more niche overlap with other arboreal species than with terrestrial species after accounting for non-independence of niche model pairs caused by geographic and phylogenetic distances. Our results suggest that occupation of the arboreal microhabitat by salamanders may only be possible in sufficiently warm, low elevation conditions. More broadly, this study indicates that the impact of micro-environmental conditions on temporary microhabitat use, as demonstrated by small-scale ecological studies, may scale up dramatically to shape macroevolutionary patterns.

Fig 1. PCA of climate variables.

All species are colored by microhabitat type under the 6-M classification scheme. Small, unfilled points indicate species’ means, and larger solid points indicate microhabitat-wide means. Polygons encapsulating all species within a microhabitat group are also included. Arboreal species on average display higher PC1 values, which corresponds to warmer temperatures, more precipitation, and lower elevations (see S2 Table for full loading data). This plot also suggests that although the means differ between arboreal and terrestrial species, these groups do not occupy fully distinct climate spaces. Rather, arboreal species appear to occupy a subset of the climates occupied by the terrestrial species.

Fig 2. Climatic variables extracted from terrestrial (black) and arboreal (green) species distributions.

Specific climatic variables include: Min Temperature of Coldest Month (BIO6), Annual Mean Temperature (BIO1), Max Temperature of Warmest Month (BIO5), Precipitation of Driest Quarter (BIO17), Annual Precipitation (BIO12), Precipitation of Wettest Quarter (BIO16), annual variation in Potential Evapotranspiration (Driest Quarter, Annual Mean, and Wettest Quarter), Cloud Cover, Elevation, and Climatic Moisture. White dots indicate pgls coefficients for each group. Asterisks indicate significant relationships across species means after accounting for phylogeny.

Fig 3. Schoener’s D and Warren’s I values by comparison type.

Horizontal lines indicate group means, all of which differ significantly from one another at an alpha of 0.05. Arboreal species show higher degrees of niche overlap with other arboreal species than with terrestrial species, indicating that arboreality may be constrained by climatic niche.