Divergence time and historical demography
The divergence time between the two M. virgata cox1lineages was estimated using a strict clock model and the HKY+I model in BEAST v2.4 (Drummond & Bouckaert, 2015). The mtDNA cox1molecular clock for this species have not been empirically assessed and we applied two different fossil-based divergence rates (a minimum rate of 0.7% and a maximum rate of 2.4%/Mya) for mtDNA cox1estimated from two different molluscan taxa (0.7%-1.2%/Mya [the ark clam; Marko, 2002]) and 2.4%/Mya; [teguline gastropod; Hellberg & Vacquier, 1999]). To estimate the timing of divergence between the twoM. virgata lineages, two independent runs with an MCMC chain length of 2 x 108 were performed with sampling every 1,000 generations. Log and tree files were combined with LogCombiner v2.4.8 (Rambaut & Drummond, 2014a) with a burn-in of 25%. The effective sample size (ESS) values of all parameters above 1,000 were obtained in Tracer v1.6 (Rambaut, Suchard, Xie, & Drummond, 2014). Trees were used to build a maximum clade credibility tree in TreeAnnotator v2.4.8 (Rambaut & Drummond, 2014b) and visualized in FigTree v1.4.3 (Rambaut, 2016). For testing historical demographic events, Tajima’s D and Fu’s FS statistics were first calculated to test for neutrality using ARLEQUIN. We then used the pairwise mismatch distribution to test population expansion using DnaSP.