Widespread coexistence of genetically distinct morphotypes in the Satyrium longicauda complex (Orchidaceae)

Species-level taxonomy is traditionally based on herbarium collections that typically include few, or even single, representatives per site. This can lead to underestimation of diversity when there are sympatric populations of superficially similar plants belonging to different lineages. Satyrium longicauda (Orchidaceae) represents a taxonomic challenge for the delimitation of species boundaries due to the high degree of morphological variation detected within and among populations. Currently, just two varieties are accepted based mainly on length differences of the lateral sepal and nectar spur. However, there is extensive morphological variation within South African populations and evidence for several pollination ecotypes, indicating that this taxon represents an actively diverging species complex. Here, we evaluate intraspecific morphological variation through uni- and multivariate morphometrics and analyse internal transcribed spacer sequences for individuals sampled from 36 sites, including 14 sites where divergent morphotypes occur sympatrically. Morphometric analyses of 1802 individuals revealed the presence of eight morphotypes based on vegetative and floral characters. Up to six morphologically and genetically distinct morphotypes can coexist in sympatry. Morphological and genetic distances among populations were significantly correlated. Phylogenetic analyses of 120 accessions indicated that neither of the two varieties nor S. longicauda as a species is monophyletic, and provided evidence for the monophyly of some of the morphotypes including the newly described S. cernuiflorum. The presence of distinct morphological and genetic sympatric variants, which in several cases scale up to distinct evolutionary lineages, is consistent with the existence of different taxa according to morphological and biological species concepts. Our results therefore confirm that taxonomy based mainly on herbarium collections can grossly under-estimate actual diversity of disparate lineages, although further work is required to finalize taxonomic decisions. These findings have implications for efforts to estimate species diversity in groups that are in the process of diversifying and for conservation practice.