Unmasking cryptic diversity in grasses (Poaceae)
Abstract
Grasses (Poaceae, about 11000–12000 species) are among the taxonomically most difficult flowering plants. Reduced florets, convergent inflorescences, phenotypic plasticity, polyploidy, hybridisation and apomixis blur species boundaries, and lineages that are evolutionarily distinct yet morphologically near identical have been documented in Pooideae, Panicoideae, Chloridoideae, Arundinoideae and Oryzoideae, with consequences for conservation, crop wild relative management and weed and invasive species control. We review macromorphological, micromorphological (epidermis, caryopsis, phytoliths, pollen), cytogenetic, molecular and ecological evidence for cryptic grass lineages, compile eleven documented cases and show that they are concentrated in one family, a pattern that mirrors research effort more than biology. We compare methods by resolving power and by the expertise and time they consume and argue that trained agrostologists rather than instruments limit most workflows. We show how polyploidy, hybridisation and apomixis violate the assumptions of widely used delimitation models, drawing on apomictic grasses such as Poa pratensis L., Cenchrus ciliaris L., Paspalum notatum Flüggé and Urochloa P.Beauv. Finally, we ask when a cryptic lineage should be named, address the nomenclatural consequences, and propose a three-stage Delimit-Diagnose-Decide framework with graded treatments (species, subspecies, vouchered informal lineage, aggregate), tiered reporting standards and worked applications. Detecting a lineage, recognising it and naming it are separate acts, and sound naming requires concordant independent evidence together with a tested route to identification.