Estimations of ploidy
The differences in genomic histories between M. spicata andM. × rotundifolia can potentially be explained by differences in ploidy (Chambers & Hummer, 1994; Harley & Brighton, 1977). The ploidy estimates using HMMploidy were variable and somewhat consistent with previously reported chromosome counts. Although our estimates are uncertain, they suggest that specimens morphologically assigned as M. spicata are polyploids and those assigned asM. suaveolens are diploids (Table 1). The estimates forM.longifolia and M. × rotundifolia are variable (Table 1). Generally, specimens morphologically assigned as M. longifoliaand not inferred as admixed were estimated to be diploids whereas specimens morphologically assigned as M. × rotundifolia were estimated to be “likely polyploids” or “polyploids” (Table 1).