Editing Crocus (section)

==== Molecular phylogeny ====

The availability of [[molecular phylogeny]] methods revealed problems with the traditional systems based on [[Morphology (biology)|morphology]] alone. The first analysis of the complete genus was carried out by Mathew and colleagues in 2008 using [[nucleotide sequences]] from [[plastid]] regions. In particular, the DNA data suggest there are no grounds for isolating ''C. banaticus'' in its own subgenus ''Crociris'', though it is a unique species in the genus. Because it has a prophyll at the base of the pedicel, it therefore would fall within section ''Crocus'', although its exact relationship to the rest of the subgenus remains unclear.{{sfn|Petersen et al|2008}}

Of the 15 series in the Mathew scheme, only seven were [[monophyletic]], and in particular the largest series, ''Biflori'' and ''Reticulati'', which include a third of all species, were non-monophyletic. Another anomalous species, ''C. baytopiorum'', should now be placed in a series of its own, series ''Baytopi''. ''C. gargaricus'' subsp. ''herbertii'' has been raised to species status, as ''C. herbertii''. The autumn-flowering ''C. longiflorus'', the [[type species]] of series ''Longiflori'' (long regarded by Mathew as "a disparate assemblage"), appeared to lie within series ''Verni''. In addition, the position of ''C. malyi'' was currently unclear.{{sfn|Petersen et al|2008}}

DNA analysis and morphological studies suggest further that series ''Reticulati'', ''Biflori'' and ''Speciosi'' are "probably inseparable", ''C. adanensis'' and ''C. caspius'' should probably be removed from ''Biflori'', ''C. adanensis'' falls in a clade with ''C. paschei'' as a sister group to the species of series ''Flavi'' and ''C. caspius'' appears to be sister to the species of series ''Orientales''.{{sfn|Petersen et al|2008}}{{sfn|Harpke et al|2013}}

The study showed "no support for a system of sections as currently defined", although, despite the many inconsistencies between Mathew's 1982 classification and the current hypothesis, "the main assignment of species to the sections and series of that system is actually supported". The authors state, "further studies are required before any firm decisions about a hierarchical system of classification can be considered" and conclude "future re-classification is likely to involve all infrageneric levels, subgenera, sections and series".{{sfn|Mathew et al|2009}} A further study, using the [[internal transcribed spacer]] region (ITS) of the nuclear ribosomal DNA (rDNA), together with a [[chloroplast]] marker, broadly confirmed these findings.{{sfn|Harpke et al|2013}}

''Crocus'' forms a monophyletic [[clade]], with a basal [[polytomy]] of four subclades. The first clade (A) corresponding to section ''Crocus'', but including ''C. sieberi'' and several closely related species (originally included in section ''Nudiscapus'' series ''Reticulati''). The remaining three clades (B-D) include all the remaining species of section ''Nudiscapus''. Of these, B and C are small, corresponding to series ''Orientales'' and ''Carpetani'' respectively, with all remaining series in the large D clade. The exception is ''C. caspius'', originally in series ''Biflori'', which [[Segregate (taxonomy)|segregates]] in clade B. Thus, although division of the genus into two sections is well supported, no single morphological character defines these two groups. The ''C. sieberi'' group are assumed to have lost their prophyll secondarily. Of the series, eight could be shown to be monophyletic; ''Crocus'', ''Kotschyani'' and ''Scardici'' (section ''Crocus'') and ''Aleppici'', ''Carpetani'', ''Laevigati'', ''Orientalis'' and ''Speciosi'' (section ''Nudiscapus''). Flowering season did not correspond to molecular groupings and nor did any of the previously used morphological characteristics, indicating a high degree of [[homoplasy]], in which traits are gained or lost independently in different lineages. The remainder of the series could not be supported as natural groupings. Mathew's concept of subspecies status within ''C. biflorus'' could not be supported, each being considered a separate species, resulting in the genus having at least 150 species.{{sfn|Harpke et al|2013}}

A more detailed molecular and morphological study of series ''Verni'' (section ''Crocus'') allowed it to be better characterised and circumscribed, as well as the closely related series ''Longiflori''. Series ''Verni'' ''sensu'' Mathew was found to consist of two groups, the first being ''C. vernus'' ''sensu'' Mathew and the other consisting of ''C. etruscus'', ''ilvensis'', ''kosaninii'' and ''longiflorus''. The taxonomic status of ''C. vernus'' had been uncertain for some time, given the observation that the name was more properly applied to ''C. albiflorus'',{{sfn|Peruzzi et al|2013}} requiring a new designation of ''C. neapolitanus'' for those previously known as ''C. vernus''. Subsequently ''C. vernus'' was split into 5 separate species. The incorporation of ''C. longiflorus'' into series ''Verni'' resulted in making series ''Longiflori'' no longer a legitimate taxonomic unit.{{sfn|Harpke et al|2015}}

In section ''Nudiscapus'', series ''Reticulati'' was polyphyletic with species intermingled with series ''Biflori'' and ''Speciosi'', requiring a recircumscription, confining ''Reticulati'' to 8 species, to obtain monophyly.{{sfn|Harpke et al|2014}} Among the thereby displaced species, are a number of very closely related taxa, referred to as the ''Crocus sieberi'' aggregate, which has been proposed as a new series ''Sieberi''.{{sfn|Kerndorff et al|2016}} Other new series, such as ''Isauri'' and ''Lyciotauri'', continue to be created out of the ''Biflori'' series.{{sfn|Kerndorff et al|2014}}{{sfn|Kerndorff et al|2013}}

Mathew's circumscription of ''Crocus'' introduced the rank of subspecies, of which the largest number (14) were those of ''Crocus biflorus'' <small>[[Philip Miller|Miller]]</small>, the type species of series ''Biflori'', a number which continued to grow. Molecular methods identified these as a [[polyphyletic]] assemblage rather than closely related subordinate infraspecific taxa. This necessitated a complete taxonomic revision of series ''Biflori'', elevating each subspecies to species status.{{sfn|Harpke et al|2016}} A similar issue occurs with ''C. reticulatus'' ''sensu'' Mathew, who created two subspecies, resulting in 9 newly defined species.{{sfn|Harpke et al|2014}}