(3018) Proposal to reject the name Cheiranthus armeniacus (Erysimum armeniacum) (Cruciferae)

(3018) Cheiranthus armeniacus Sims in Bot. Mag.: t. 835. 1 Mai 1805 [Angiosp.: Cruc.], nom. rej. prop.

Lectotypus (hic designatus): [icon in] Bot. Mag.: t. 835. 1 Mai 1805.

The name Cheiranthus armeniacus Sims (in Bot. Mag.: t. 835. 1805), the basionym of Erysimum armeniacum (Sims) J. Gay (Erysim. Nov.: 8. 1842), is the earliest available for the group of 3–4 predominantly Caucasian, basically biennial species of Erysimum L. referred to as E. subsect. Iberica V.I. Dorof. (in Bot. Zhurn. (Moscow & Leningrad) 72: 1538. 1987). In spite of this, its use has been most discontinuous for over two centuries and its interpretation remains unstable and still seriously varying among authors.

Description of Cheiranthus armeniacus was based on cultivated plants grown from seeds collected on Mount Ararat in eastern Anatolia. No relevant herbarium material seems to be extant, as suggested by Davis & al. (Fl. Turkey 10: 55. 1988; the author's search in K and elsewhere also failed), and trichome morphology, critically important in the taxonomy of Erysimum, was not described in the protologue. However, the original illustration, designated above as lectotype, showing the characteristic habit along with some mentioned details (e.g., stout sulcate stem, big and profoundly bilobed stigmas, erect young fruits) allows for defining its affinity. Bieberstein (Fl. Taur.-Caucas. 2: 118. 1808) was the first to consider C. armeniacus conspecific with C. ibericus Adams (in Beitr. Naturk. 1: 61. 15 Nov 1805; see TL-2 No. 16,898), described from north-eastern Georgia (lectotypified by Polatschek in Ann. Naturhist. Mus. Wien, B, 111: 254. 2010: Georgia, Mtskheta-Mtianeti, Mt. Kaishauri, A.A. Musin-Pushkin [B-W No. 12086-03]), and was followed by Candolle, who adopted the name E. ibericum (Adams) DC. (Syst. Nat. 2: 498. 1821) for the species. This synonymy was included in early synoptic accounts (Sprengel, Syst. Veg. 2: 907. 1825; Steudel, Nomencl. Bot., ed. 2, 1: 346. 1840; Jackson in Index Kew. 1: 508, 892. 1893) and in a few floras (Ledebour, Fl. Ross. 1: 187. 1841 [“1842”]; Boissier, Fl. Orient. 1: 192. 1867; Busch in Kuznetsov & al., Fl. Cauc. Crit. 3(4): 526. 1909, in Komarov, Fl. SSSR 8: 101. 1939), in the latter two cases with respect to the illustration only. Recently this approach was revived by Zernov & al. (Opred. Sosud. Rast. Karachay-Cherkess. Resp.: [205]. 2015).

As an accepted species, Cheiranthus armeniacus was treated by Gay (l.c.), who transferred it to Erysimum, then, indirectly, by Ruprecht by exclusion of the name from synonymy of Erysimastrum ibericum (Adams) Rupr. (in Mém. Acad. Imp. Sci. Saint Pétersbourg, Sér. 7, 15(2) [Fl. Caucasi]: 75, 77. 1869) and the next time, after over 110 years, by Polatschek (in Willdenowia 13: 88. 1983), who mentioned that the species also occurs in the Caucasus apparently implementing, though not directly claiming, its conspecificity with Erysimum ibericum. Since that time, the name has returned to the taxonomic literature, in almost all cases (except for Zernov & al., l.c.) referring to an accepted species both in treatments (e.g., Greuter & al., Med-Checklist 3: 108. 1986; Dorofeyev, l.c. 1987: 1538, in Takhtajan, Konspekt Fl. Kavkaza 3(2): 402. 2012; Davis & al., l.c.; Avetisyan in Takhtajania 1: 84–87. 2011; Mutlu in Güner, Checkl. Fl. Turkey: 273. 2012 [though also given in synonymy of E. ibericum on next page]) and compendiums/databases (Czerepanov, Vasc. Pl. Russia Adjac. States: 138. 1995; Francis & al. in Biodivers. Data J. 9: e58733. 2021, https://ipt.pensoft.net/resource?r=aafc-brassicaceae-checklist&v=1.8; Plants of the World Online, https://powo.science.kew.org; Species 2000 & ITIS Catalogue of Life, https://www.catalogueoflife.org; World Flora Online, https://www.worldfloraonline.org; all accessed 8 Jan 2024), but with differing interpretations.

One of these interpretations was intended to overcome the purported discrepancy between the description reporting “shrubby divided stem” and the figure (corresponding to Erysimum ibericum or its closest relatives) in the protologue, stressed by Gay (l.c.), Ruprecht (l.c.), and Polatschek (in Ann. Naturhist. Mus. Wien, B, 112: 404. 2011). In order to fix the use of the name Cheiranthus armeniacus, as he supposed, in accordance with the protologue, the latter author proposed (though not achieved under Art. 7.11 of the Shenzhen Code; Turland & al. in Regnum Veg. 159. 2018) its epitypification by the only specimen of Erysimum known to him from Mt. Ararat. The relevant specimen (GOET 002577) is represented by two plants (one is quite damaged and lacking a root, the other is apparently perennial) with just a couple of first flowers, presumably with affinity to E. leptophyllum (M. Bieb.) Andrz. ex DC., which sharply differs from both the original illustration of C. armeniacus and the morphology of any species of the E. ibericum group. Such an approach, defining E. armeniacum as an endemic of Mt. Ararat, can hardly be accepted since it radically changes the usual concept of the species and disagrees with the protologue, not only with respect to the illustration and above-mentioned characters but also in the single feature (life form) that was seemingly used as an argument against treating E. armeniacum as conspecific with or at least closely related to E. ibericum: the intended epitype is a herb, not even caudex-forming, and not a subshrub. As for the mentioned discrepancy in the protologue, two things can be said. First, normally biennial E. ibericum and related species always form rosettes, sometimes numerous, and, judging from the herbarium material, may behave as short-lived (and at least bicarpic) perennials. In this case, the lower part of plants, considerably lignified, look, especially when situated above ground (e.g., on rocks), quite “subshrubby”. Second, it is a known phenomenon that various species of Cruciferae demonstrate an elongation of life span, becoming somewhat woody in culture. Noteworthy, exactly this explanation (change of morphology under cultivation) was suggested by Bieberstein (l.c.).

A second interpretation of the name Cheiranthus armeniacus was launched and promoted by Dorofeyev (l.c. 1987, in Turczaninowia 6(3) [Crucif. Russian Caucasus]: 78. 2003, l.c. 2012) and accepted by some other authors (Murtazaliev, Consp. Fl. Dagestan 1: 230. 2009; Ivanov, Consp. Fl. Cauc. Ross.: 113. 2021). According to this concept, the name Erysimum armeniacum refers to a species very close to E. ibericum with a habit just slightly and not always reliably differing, but sufficiently distinct in details of indumentum, confined, according to the treated collections, to the Greater Caucasus within Russia and Georgia. Polatschek agreed that relevant specimens specifically differ from E. ibericum, but he recognised them, based on fine details of indumentum and variable petal length, as three vicariant species, E. macropetalum Polatschek (in Ann. Naturhist. Mus. Wien, B, 109: 152. 2008), E. osseticum Polatschek (l.c. 2008: 153), and E. dagestanicum Polatschek (l.c. 2010: 253), all synonymised by Dorofeyev (1.c. 2012: 402) with E. armeniacum. I support the latter author in recognition of one species instead of three, though, because of biogeographic considerations, not in naming it this way. More specifically, based on revised material in BAK, ERE, LE, MW, TBI, TGM and other herbaria, I find it endemic to the northern Caucasus and cannot confirm its reports for western and south-western Transcaucasia by Dorofeyev (1.c. 2012). This means that the distribution of E. armeniacum sensu Dorofeyev includes a gap covering most of the southern slope of the Greater Caucasus, the whole Lesser Caucasus and depressions between them, a highly unlikely pattern, especially for the species abundantly collected and evidently common throughout its range.

A third viewpoint on the identity of Cheiranthus armeniacus is the above-mentioned and most traditional, initiated first by Bieberstein (l.c.), treating it as conspecific with Erysimum ibericum. The problem in this case, highlighted by Davis & al. (l.c.) and Avetisyan (in Fl. Rast. Rastit. Res. Armen. 17: 21. 2009), is the priority of C. armeniacus, published half a year earlier, over C. ibericus, although all authors who merged them used Candolle's combination based on the latter name. As for E. ibericum, this binomial, in contrast to E. armeniacum, has been widely applied for over two centuries for a predominantly Caucasian species of wallflower (e.g., Schmalhausen, Fl. Sredn. Jushn. Rossii 1: 71. 1895; Grossheim, Fl. Kavkaza, ed. 2, 4: 239. 1950; Karjagin, Fl. Azerb. 4: 307. 1953; Avetisyan in Takhtajan, Fl. Armenii 5: 113. 1966; Chinthibidze in Kutateladze, Fl. Georgiae, ed. 2, 5: 40. 1979; Zernov, Fl. Severo-Zapadn. Kavkaza: 288. 2006; Askerov, Azerb. Fl. Konsp.: 59. 2011, etc., including all above-cited authors dealing with this region). With very few exceptions of mistreatment (e.g., Polatschek & Rechinger in Rechinger, Fl. Iranica 57: 302. 1968), E. ibericum was used in a sense that included its type, even though most of the cited authors (actually all before Dorofeyev, l.c. 1987) did not taxonomically separate the northern Caucasian “E. armeniacum”.

It should be noted in addition that the occurrence of Erysimum ibericum in Turkey and details of its distribution are still questionable. The species was first reported for Turkey by Yıldırımlı (in Ot Sist. Bot. Derg. 15(2): 53. 2008) from the area adjacent to Georgia and Armenia and was also reported for the east of the country by Mutlu (l.c.). This information was not confirmed by Polatschek (l.c. 2010: 255, l.c. 2011: 485), who considered E. ibericum as confined to Russia and Georgia (despite previous reports also from Armenia and Azerbaijan). Furthermore, he assigned all specimens from Armenia to E. gabrielianiae Polatschek (l.c. 2008: 163), a very closely related species treated by him as an Armenian endemic (Polatschek, l.c. 2011: 469). Unlike Dorofeyev (l.c. 2012: 403), who synonymised E. gabrielianiae with E. ibericum, Avetisyan (l.c. 2009: 21) accepted the prior, but demonstrated that the latter also occurs throughout (in 6 floristic regions out of 12) the country. Hence, from the geographic viewpoint, should E. gabrielianiae and E. ibericum be treated separately, a hypothetical finding on Mt. Ararat or, in other words, description of either of them as Cheiranthus armeniacus is more or less equally probable. However, such morphological characters of E. gabrielianiae as ovate-lanceolate, subentire or serrulate cauline leaves do not fit the more or less lanceolate and acutely incised-dentate leaf pictured in the protologue of C. armeniacus and typical of E. ibericum s.str. Besides, judging from the studied material of E. gabrielianiae in ERE, M, W and other herbaria, including its type gathering, this species (if not a diploid race of infraspecific category of normally tetraploid E. ibericum, as found by Polatschek, l.c. 2011) is the only member of E. subsect. Iberica which is strictly biennial and never revealing any tendency of forming a woody caudex. This means that a theoretical attempt to “save E. ibericum” by applying the name E. armeniacum to what was later described as E. gabrielianiae has no solid ground, and questionable taxonomic status of the latter further stresses the problem.

Summing this all up and in view of the low probability of occurrence on Mt. Ararat of yet another species of the group, identical in habit with Erysimum ibericum, rejection of the name Cheiranthus armeniacus seems to be the best option. Alternatively, its use in a sense of Dorofeyev (l.c. 1987 and further on) via conservation with a conserved type (or a new epitypification) could be considered but with the relatively short history of application of the name in this sense and the availability of three names, all well-typified and characterised (of which two have equal priority and potentially might be applied to relevant species in a wide sense), such an act might be unwelcome. Therefore, in order to prevent nomenclatural instability that will inevitably be caused by the change, after over 200 years of wide, unambiguous and continuous use, of the name E. ibericum to the earlier, but long-neglected, discontinuously and confusingly applied E. armeniacum, implemented by the direct application of the Code, rejection of the name C. armeniacus is proposed here.