A phylogeny of the genus Crocus (Iridaceae) based on sequence data from five plastid regions

TLDR: The phylogenetic hypothesis is in conflict with the primary classification of Mathew into subgenera and sections, most notably by placing the highly morphologically deviant C. banaticus not as sister group to subgenus Crocus but imbedded within it.

Abstract: For the first time a phylogenetic analysis of the genus Crocus is presented. The phylogeny includes all but two of the currently 88 recognized species of the genus. The analysis is based on a total of 222 phylogenetically informative characters derived from nucleotide sequence data from three protein-coding (ndhF, accD, rpoC1) and two non-coding (trnH-pshA, rpl36-rps8) plastid regions. The phylogenetic hypothesis is in conflict with the primary classification of Mathew into subgenera and sections, most notably by placing the highly morphologically deviant C. banaticus (subgenus Crociris) not as sister group to subgenus Crocus but imbedded within it. The grouping of the taxa into series is better supported, though not entirely. Of the fifteen series recognized by Mathew, eight are confirmed as being monophyletic, monophyly of one more remains a possibility, and monophyly of a further two series is only violated by one taxon. Not unexpectedly, the two most species-rich groups, series Reticulati and series Biflori encompassing almost 1/3 of all species, are clearly non-monophyletic, but the present data do not satisfactorily solve their phylogenetic relationships.