Go north, young salamander

Shenandoah salamanders are a case study in restricted distributions, known only from three mountainsides in Shenandoah National Park, in the Appalachian Mountains of Virginia. What’s keeping them in such a restricted range? A new paper in the journal Ecology and Evolution aims to answer that question using population genetics.

The genus Plethodon, which includes the Shenandoah salamander, has been described as a case of "non-adaptive radiation" — the rapid formation of species, not because they’re all evolving to cope with different habitats and lifestyles, but simply because their populations are prone to becoming fragmented and isolated. Plethodontid salamanders lack lungs, and respire through their moist skin; they’re restricted to hospitably wet habitats, and have likely speciated mostly through habitat isolation and vicariance events as Appalachian woodlands shifted with glacial cycles over the last several million years.

The authors of the new study, led by Kevin Mulder at the Smithsonian Conservation Biology Institute, were interested in the specific habitat features that restrict the movement of Shenandoah salamanders. They collected mitochondrial genome sequences and sequence-captured hundreds of nuclear loci in samples from all three known populations of the salamanders. A good old-fashioned mitochondrial haplotype network and clustering analysis using nuclear markers strongly indicated that population on the southernmost mountain was isolated from the other two; it formed a separate cluster in both analyses.

It wasn’t immediately obvious why that should be; the southern mountain is about as distant from the central mountain as the central mountain is from the northern one — at least, as the crow flies. To determine what might make that landscape look different to a (non-flying) salamander, Mulder et al. used a Circuitscape analysis to identify habitat variables that were consistent with lower rates of gene flow, as seen in their genetic data. Circuitscape uses genetic data and spatially explicit habitat information to estimate a "resistance" to movement associated with different values of a habitat measure. The coauthors estimated resistance associated with eight different variables, including an estimate of solar radiation intensity and an index of vegetation type; but most of these only showed that the salamanders were unlikely to travel to higher elevations than they usually occupy.

The variable that did explain the isolation of the southern population was topographic — more north-exposed slopes had lower resistance, and while northern-exposed slopes link the northern and central populations, the orientation of the southern mountain positions a south-exposed slope between the southern population and the rest of the Shenandoah salamander’s range.

What, specifically, it is about south-facing slopes that stymies salamanders is unclear. None of the more obvious habitat variables examined in the paper produce the same kind of resistance-surface signal as northness; though Mulder et al. note that southern-exposed slopes are warmer and drier in that part of the Appalachian Mountains, and northern exposure is a feature of the salamander’s usual habitat. It’s possible, they suggest, that genetic connections among the three salamander populations are determined by sufficiently rare dispersal events that the broad-scale landscape features they examined aren’t particularly informative. I also wonder whether the distribution of sampling across the three populations — three relatively tight clusters of sample locations separated by large gaps where the salamanders don’t occur — is an ideal case for the Circuitscape method.

Nevertheless, the scale and thoroughness of the data reported here is a dramatic improvement over the past state-of-the-art for understanding the Shenandoah salamander’s diversity and population structure, and that can only help efforts to protect the endangered amphibian.