What is Allium achaium Boiss. & Orph.? Disentangling the taxonomy of a Greek mountain species

Allium oreohellenicum Tzanoud., Tsakiri & Raus (Amaryllidaceae), endemic of the Greek mountains, is described and illustrated as a species new to science. Information regarding its cytology, geographical distribution and taxonomic relationships is also provided. Material belonging to A. oreohellenicum was formerly known/classified as A. achaium Boiss & Orph., a taxon described in 1882 based on a gathering by Orphanides from Mt Klokos (N Peloponnisos, Greece). A detailed study of the Orphanides gathering concerned revealed that it essentially belongs to A. frigidum Boiss. & Heldr., a taxon described 28 years earlier, and so the need for a new name.


Introduction
Allium L. (Amaryllidaceae) is one of the most speciesrich genera of the Greek flora. Presently it is known to be represented in the country by 103 species with more than 50 % of them (54) endemics. Considering that continental Greece including some of the larger islands is characterized by dozens of mountain massifs approaching or exceeding 2000 m in altitude, it is worth noting that among the Greek Allium endemics only three are considered as "mountain species", i.e. adapted to and exclusively occurring in high-altitude habitats (>1600 m). These species are currently named A. achaium Boiss. & Orph., A. frigidum Boiss. & Heldr. and A. parnassi cum (Boiss.) Halácsy (Strid & Tan 1991;Dimopoulos & al. 2013). They have some morphological characters in common, viz. an infundibular-campanulate to infundibular perianth, the perianth segments at least 6 mm long, the stamens included in the perianth, and the ovary much longer than wide.
Twenty years later, Bogdanović & al. (2011) completely changed the taxonomic treatment of Allium achaium. Indeed they continued to quote the voucher of Orphanides,  as the holotype of A. achaium, but confusingly described and illustrated under this designation an Allium from Mt Klokos characterized by morphological characters that do not at all match either the specimens of the Orphanides gathering, mentioned above, or any other specimen belonging to the species group of A. achaium sensu Andersson (1991) and previous botanists, A. frigidum and A. parnassicum (Fig. 1). In fact, the characters given by Bogdanović & al. (2011), viz. exserted stamens, cup-shaped perianth and globose ovary, characterize the A. flavum/A. stamineum group, which is quite different from those groups in which A. achaium, A. frigidum and A. parnassicum belong (with the stamens included in an infundibular-campanulate to infundibular perianth  Bogdanović & al (2011) as the "true A. achaium". Willdenowia 49 -2019 and the ovary much longer than wide; Fig. 1 & 2). The present paper aims at disentangling these contradicting taxonomic points of view.

Material and methods
Mt Klokos, the locus classicus of Allium achaium Boiss. & Orph., had been visited and explored by the first author, accompanied by varying staff of the Botanical Institute of Patras, several times from 2013 to 2017 in the framework of monitoring regional "Natura 2000" sites, and all Allium taxa occurring in the area have been completely registered. In parallel, herbarium specimens of A. achaium sensu Andersson (1991), A. frigidum and A. parnassicum from all mountains of Peloponnisos, Sterea Ellas and the Pindos range, which are available in the Botanical Museum of the University of Patras (UPA), have been comparatively studied, among them an iso-type sheet of the original gathering by Orphanides from Mt Klokos (Orpha nides,Fl. Gr. Exsicc. no. 427), on which the description of A. achaium was based (Fig. 5). In addition, the type material of the three species A. achaium, A. frigidum and A. parnassicum deposited in the Geneva herbarium (G, G-BOISS) was investigated. Herbarium specimens of these three species examined by Andersson (1991) and deposited in the herbarium of Lund (LD) were also examined. No relevant historical exsiccata could be traced in the herbarium of Berlin (B), very likely destroyed there in 1943.

Results and Discussion
Mt Klokos (1774 m) is situated SW of the city of Egion (previously Vostitsa) and NNW of the much higher mountain massifs of Chelmos (2355 m) and Erimanthos (2224 m) in northern Peloponnisos. Based on the information given by local people, the locality "Pente Vryses" (i.e. five springs), which is given on the label of Orphanides, Fl. Gr. Exsicc. no. 427, is located a few kilometres SW of Pteri, a village on the NE-facing slope of the mountain at an altitude of c. 1300 m. The altitude given by Orphanides for this locality is almost the same (4000 ft). In the framework of our floristic exploration, we visited the slopes above the locality "Pente Vryses" as well as the opposite, SSE-facing slopes from an altitude of 1400 m up to the summit area. During our field work, three Allium species were found in the area: (1) The first, and most common one, was a representative of Allium sect. Codonoprasum (Rchb.) Endl. characterized by spathe valves longer than pedicels, the perianth 4 -5 mm long with evidently exserted stamens and a globose, more or less stipitate ovary (Fig. 1D &  2D). It was found on the slopes just above the locality "Pente Vryses" as well as on the opposite-facing slopes and in the summit area. Without any doubt this is the plant that Bogdanović & al. (2011) described and considered as the "true A. achaium", although none of the plants of the Orphanides gathering from Mt Klokos (G!, UPA!) has any of the characters provided by Bogdanović & al. (2011). However, plants with the mentioned characters do occur in the mountains of Greece, where they are more or less common and widespread; they represent A. fla vum subsp. tauricum (Besser ex Rchb.) K. Richt.
(2) The second taxon met with on Mt Klokos was an Allium characterized by hairy sheaths and leaves, very long spathe valves and a pinkish brown perianth with included stamens. Because of these characters it belongs to A. rhodopeum Velen. subsp. rhodopeum, a taxon also known to occur at moderate altitudes on other mountains next to Mt Klokos in N Peloponnisos.
(3) The third representative of the genus on Mt Klokos was an Allium characterized by a pendulous inflorescence, spathe valves shorter than pedicels, whitepinkish to brownish perianth, included stamens and an ovoid to ellipsoid ovary. Only a few individuals were found, confined to the summit area. When comparing this plant with material of A. frigidum from other mountains of Peloponnisos as well as with the plants of the Orphanides gathering from Mt Klokos (G!, UPA!) designated as the type material of A. achaium, we found no differences. Hence Halácsy (1904) was completely correct in pointing out that there is no difference between these two taxa ("inter A. frigidum et A. achaium differentiam videre nequeo"). They are conspecific, and the name A. frigidum has priority over A. achaium, hence the latter is correctly treated as a synonym of the former by Halácsy (1904), Hayek (1932Hayek ( -1933, Zahariadi (1975aZahariadi ( , 1975b, Stearn (1978Stearn ( , 1980Stearn ( , 1981 and Kollmann (1984).
There is no doubt that in high-altitude habitats of the mountains of N Peloponnisos one more taxon of the Al lium paniculatum group is present, sometimes growing side by side with A. frigidum. It is characterized by having outer bulb tunics often longitudinally splitting into parallel fibres, inflorescence erect (versus pendulous in A. frigidum), spathe valves longer than the pedicels, perianth segments pinkish white in the living state, to 6(-6.5) mm long, stamens with white-yellowish anthers included in the perianth, and ovary much longer than wide, narrowed at the base and truncate at the apex (Fig.  1A, 2A & 4). Plants with a similar morphology have also been collected in the high mountains of Sterea Ellas and in the Pindos range of NW Greece (Thessalia/ Ipiros) and were treated by Tzanoudakis & Vosa (1988) under the misnomer "A. parnassicum". These plants were erroneously classified by Andersson (1991) as A. achaium. We agree with Andersson, who recognized a third Allium species in the mountains of C and NW continental Greece in addition to A. frigidum and A. par nassicum. However, Andersson misapplied the name A. achaium and its type, chosen by him, to this third species, whereas in fact they belong taxonomically to A. frigidum. This third species therefore needs a new name, a new description and a new type, which are provided here. Diagnosis -Species ad sectionem Codonoprasum Allii generis pertinens, ab Allio parnassico simili non modo caule robustiore, floribus compluribus, periantho roseoalbo ad 6(-6.5) mm tantum longo antherisque alboflavidis, sed etiam chromosomatum numero plerumque triploideo (2n = 3x = 24) satis diversa.
Karyology -Material of Allium oreohellenicum from the mountains of Peloponnisos and Sterea Ellas was investigated cytologically by Tzanoudakis & Vosa (1988, under "A. parnassicum"). The triploid chromosome number 2n = 3x = 24 was reported from all examined populations, but the diploid chromosome number, 2n = 2x = 16, was additionally found in some individuals from Mts Iti and Velouchi in Sterea Ellas. The chromosome morphology and karyograms from the populations concerned were given by Tzanoudakis (1992: fig. 3 & 6), who pointed out that in both diploid and triploid plants the haploid complement consists mainly of metacentric and submetacentric chromosomes. The SATchromosomes observed (m A , Sm A ) belong to the "panicu latum" type sensu Tzanoudakis (1983). In the present study, plants from the type gathering were also cytologically investigated and likewise turned out to be triploid (2n = 3x = 24, Fig. 3).
Ecology -Allium oreohellenicum grows preferentially at 1800 -2400 m in high-mountain meadows and rocky places, flowering from mid-July to August, depending on altitude and exposition. In the type locality, it is accompanied by several endemic and critical taxa such as Allium frigidum  Etymology -The specific epithet is a compound of oreo (mountain) and hellenicum (Greek), reflecting the ecological and chorological traits of the species, which is confined to high-mountain habitats of continental Greece. The taxon was hitherto concealed by misapplication of the name Allium achaium (Andersson 1991) or by misidentification of A. flavum subsp. tauricum (Bogdanović & al. 2011).