Is Holosteum glutinosum (M. Bieb.) Fisch. et C. A. Mey. (Caryophyllaceae: Alsinoideae) just a subtaxon of H. umbellatum L. or a distinct species?

The genus Holosteum L. distributed mostly in temperate Eurasia has very confused taxonomy. Modern estimation of the species richness varies from one species worldwide to seven species occurring just in Eastern Europe. One of the most problematic species of the genus is H. glutinosum(M. Bieb.) Fisch. et C. A. Mey. usually treated as a subspecies or a variety of H. umbellatum L. On the base of extensive study of the protologues, type material, a rich variety of other herbarium specimens, photos of all relevant taxa, and living plants in the field we recognize H. glutinosum as a distinct species. This species has three characters unique in the genus Holosteum: pure green color of leaves and sepals, completely herbaceous bracts, and lanceolate leaves; additional diagnostic characters are 10 stamens, dense glandular pilosity of leaves, sepals, bracts, stem, and pedicels, and entire petals longer than sepals. The species also has statistically significant differences from H. umbellatum (longer leaves, longer and wider petals, and longer capsule) while occurring in sympatry with it. Two subspecies are recognized within H. glutinosum: H. glutinosum subsp. glutinosumdistributed in Caspian Depression, Transcaucasia, Western and Middle Asia, and H. glutinosum subsp. liniflorum Fateryga, nom. et stat. nov. (≡ H. liniflorum Steven ex Fisch. et C. A. Mey. 1840 [non 1837], nom. illeg.), presumably endemic to the Crimea. The lectotypes are designated for the latter name and also for H. glandulosum Bertol., H. umbellatum var. pleiandrum Fenzl, H. imberbe J. Gay, nom. illeg. superfl., H. imberbe var. brachypetalum J. Gay, and H. imberbe var. macropetalum J. Gay. Reports of H. glutinosum from most European countries (other than Russia) are considered to be erroneous due to numerous misidentifications in herbaria and incorrect synonymy (e. g., with H. heuffelii Wierzb., which is actually conspecific with H. umbellatum). A key to the species and subspecies is provided; a total of six taxa are recognized in the genus: H. breistrofferi Greuter et Charpin, H. glutinosum subsp. glutinosum, H. glutinosum subsp. liniflorum, H. marginatum C. A. Mey., H. tenerrimum Boiss., and H. umbellatum.


Introduction
Holosteum L. (Caryophyllaceae: Alsinoideae) is a small genus of annual herbs native to temperate Eurasia (Hernández-Ledesma et al., 2015) and North Africa (Greuter et al., 1984) with the type species H. umbellatum L. that has additionally naturalized in the United States of America, Canada, Argentina, and the Republic of South Africa (Rabeler, Hartman, 2005). Although the genus is not species-rich, its taxonomy is very confused and requires a revision (Fateryga et al., 2017). Modern estimations on the number of species in the genus vary. World Flora Online (2020) accepts the one species H. umbellatum with four subspecies: H. umbellatum subsp. glutinosum (M. Bieb.) Nyman, H. umbellatum subsp. hirsutum (Mutel) Breistr., H. umbellatum subsp. viscosissimum (Čelak.) Dvořák, and the type subspecies, while many other names in the genus Holosteum remain unresolved in this database. K. Marhold (2011) in Euro+Med PlantBase lists four species occurring in Europe and the Mediterranean Is Holosteum glutinosum just a subtaxon of H. umbellatum or a distinct species? countries. Among them, H. umbellatum has five subspecies: those accepted in World Flora Online plus H. umbellatum subsp. tenerrimum (Boiss.) Greuter et Burdet. Three other species in Euro+Med PlantBase are H. marginatum C. A. Mey., which is a well recognized species (Fateryga et al., 2017), H. sperguloides Lehm., which actually belongs to the genus Spergularia (Pers.) J. Presl et C. Presl (World Flora Online, 2020), and H. subglutinosum Klokov. P. Hernández-Ledesma et al. (2015) reported the genus Holosteum to include three to four species worldwide without listing their names. A. V. Fateryga et al. (2017) supposed that there were four species: H. glutinosum (M. Bieb.) Fisch. et C. A. Mey., H. tenerrimum Boiss., H. umbellatum, and H. marginatum. One of the most "splitting" estimations was provided by N. N. Tzvelev (2012) who recognized six species within the territory of Eastern Europe. Among them, there were H. marginatum, H. glutinosum, H. umbellatum, H. klopotovii (Tzvelev) Tzvelev, H. syvaschicum Kleopow, and H. subglutinosum; in our opinion, however, the latter three represent different pilosity variations of H. umbellatum and do not merit the rank of species or even subspecies.
A. S. Zernov and I. V. Sokolov (2004) Lazkov (2006) reported that H. glutinosum had entire petals while H. umbellatum had petals with dentate apical margin. These characters were overlooked by A. S. Zernov and I. V. Sokolov (2004), as well as by earlier European authors (Coode, 1967;Walters, Akeroyd, 1993). Such a confusion with the diagnostic characters of these two taxa led us to an assumption that some of the authors erroneously assigned the most pubescent plants of H. umbellatum s. str. to H. (umbellatum subsp./var.) glutinosum and such a mistake could be partial reason for reducing the latter species to a subtaxon of the former one.
At the same time, we found in 2013 a peculiar population of H. glutinosum in the Crimea, which occurred in the same habitat with H. umbellatum and showed remarkably distinct characters from the latter. These plants had characters corresponding to H. liniflorum Steven ex Fisch. et C. A. Mey. 1840 [non 1837], nom. illeg. (Fischer et al., 1840;see also Steven, 1856), which is currently treated as a synonym of H. umbellatum var. glutinosum (Zernov, Sokolov, 2004). Thus, the purpose of the present contribution is to ascertain the taxonomic status of H. glutinosum, as well as other taxa hitherto assigned to its synonyms, and to provide an identification key to all taxa known in the genus Holosteum.

Material and methods
Our study was based on traditional taxonomic methods. All protologues of the species of interest, as well as additional relevant references, were studied. The type material of most taxa and a large amount of additional specimens were examined in nine herbaria: CSAU, G (CHG, 2020), K (Kew Herbarium, 2020), KW, LE, MW (Seregin, 2020), P (MNHN, 2020), PHEO, and YALT, as well as in various digital herbaria via the Global Plants webservice (https://plants.jstor.org/). Photos of living plants were examined on the Plantarium web-site (Plantarium, 2007(Plantarium, -2020 and on various other web-sites via the Google search (https://www. google.com/). We used the morphological method and examined the main important characters of Holosteum plants: leaf shape; leaf, bract, and sepal coloration; bract size and structure; petal size and 53 Turczaninowia 23, 4: 50-64 (2020) shape; capsule size; stem, pedicel, bract, and sepal pilosity. Distribution of the species was studied on the base of critically reviewed literature data, as well as the studied herbarium and photographic materials.
Living plants of H. glutinosum were studied in April 2016 in Zuya, Belogorsk District of the Republic of Crimea (45°03′39″N, 34°20′06″E), where they occurred in the same habitat with H. umbellatum. Thirty plants of each species were randomly selected and measured with a vernier caliper. The measured parameters were the shoot length, the upper leaf length and width, a petal length and width, and an opened capsule length. Then, the minimum, maximum, and mean values were calculated for each parameter. Confidence intervals of the mean values were calculated for 95 % confidence level (p = 0.05). Statistical significance of the differences between the mean values of two species was checked with Student's t-test (Lakin, 1990). PCA analysis of the same plants was made with Statistica 7 software.
To ascertain what taxonomic rank is to be applied to H. glutinosum, we assumed that a subspecies is a population (or several populations) of the species with morphological differences recognized by taxonomists (Mayr, 1974). This assumption means that two subspecies cannot co-occur in the same place. At the same time, the assumption does not mean that any morphological difference between allopatric populations is the reason to recognize one of them as a separate subspecies (many such differences are unstable and associated with ecological conditions).

What is Holosteum glutinosum?
Studying of all available materials has revealed that H. glutinosum is a well recognized species clearly distinct from four other well recognized species in the genus (see the key below) by three unique characters: pure green color of leaves and sepals, completely herbaceous bracts, and lanceolate leaves. These differences, however, are generally overlooked by most authors except N. N. Tzvelev (2004). Four other species have leaves with glaucous tincture, bracts either with membranous margins or nearly/almost completely membranous, and ovate leaves. Additional diagnostic characters of H. glutinosum, which are not unique, are 10 stamens, dense glandular pilosity of leaves, sepals, bracts, stem, and pedicels, and entire petals longer than sepals.
Holosteum glutinosum is quite variable across its geographical range (Figs 1-2). The plants from the Volga Region, close to the type locality of the species (Fig. 1A), and the Middle Asia ( Fig. 1B-D) usually have rather erect stem and relatively small petals. Plants from Asia Minor (Fig. 1E) and Transcaucasia ( Fig. 1F-G) are rather ascending while their petals are also not very large. The most remarkable plants can be found in the Crimea where they always have erect, sometimes very tall stem and peculiarly large petals ( Fig. 2A-F). All diagnostic characters of H. glutinosum listed above, however, are constant across its range.
The species has statistically significant differences from H. umbellatum when occurring in the same habitat with it. These differences are longer leaves, longer and wider petals, and longer capsule (Table). Samples of the two species studied in the Crimea also form two well recognized clouds in PCA analysis (Fig. 3). Thus, H. glutinosum cannot be treated as a subspecies or a variety of H. umbellatum. There is also a difference in phenology, which is well visible in Fig. 2B where these two species are shown together: H. glutinosum in flower and H. umbellatum in fruit. Another question, however, arises: is the name H. glutinosum correctly applied to the "green" species (i. e., the species with pure green color of leaves and sepals, completely herbaceous bracts, and lanceolate leaves)?
Holosteum glutinosum was originally described by F. A. Marschall von Bieberstein (1808: 344) as Arenaria glutinosa M. Bieb. The author did not specify the color of leaves and the structure of bracts (herbaceous or membranous) of his species. He, however, provided the data on the leaf shape: "foliis oblongis obtusiusculis" that corresponded rather to the "green" species. F. A. Marschall von Bieberstein (1808) also placed his description to the chapter "Decandria trigynia" that meant that the plant had 10 stamens, i. e., also corresponded to H. glutinosum in its current treatment (H. umbellatum has usually 5 stamens). Early authors, however, placed A. glutinosa to either synonyms or subtaxa of H. umbellatum ("glaucous" species). For example, E. Fenzl (1842)     ] pleiandrum/glutinosum was up to 10 ("6-10" and "7-10", respectively). Thus, their varieties may probably represent an unsatisfactory mixture of the traits of various specimens (of both "glaucous" and "green" species).
Soviet authors (Muravyeva, 1936;Grossheim, 1945;Schischkin, 1947;Czerepanov, 1995) were different in their taxonomic point of view on these two taxa. They clearly separated H. glutinosum (plants with 10 stamens) from H. umbellatum (plants with 3-5 stamens) but only N. N. Tzvelev (2004) completed their diagnoses with bract structure and leaf shape. The lectotype of A. glutinosa was designated by G. A. Lazkov and A. S. Zernov (see Zernov, Sokolov, 2004). The plants of the lectotype (Fig. 4) definitely represent the "green" species and clearly correspond to both the original description (Marschall von Bieberstein, 1808) and the modern Soviet/Russian treatment of H. glutinosum, although A. S. Zernov and I. V. Sokolov (2004) themselves did not accept it as a full species (see above).
As it was stated in the Introduction, modern taxonomic databases such as Euro+Med PlantBase (Marhold, 2011) or World Flora Online (2020) treat H. glutinosum as a subspecies of H. umbellatum that is not congruent with their co-occurring in sympatry (see above). One more possible reason of such an incorrect treatment, besides previous incorrect treatments by the early authors, is commonly occurring misidentifications of H. umbellatum, or sometimes also H. tenerrimum or even H. marginatum, as H. glutinosum in most European herbaria (with the exception of some Russian ones such as MW, see Seregin, 2020). For example, the plants on only one of nine sheets stored as H. glutinosum in P (MNHN, 2020) are correctly identified, namely a gathering from Armenia (P04975719). Among eight other sheets, six ones (P04975720, P04975802, P05020761-P05020763, and P05439558) represent typical H. umbellatum collected by Russian botanist A. K. Becker in Sarepta (currently Volgograd, Russia), close to the This synonymy has probably come from the "Atlas Florae Europaeae" (1983) and it was also repeated in the "Flora Europaea" (Walters, Akeroyd, 1993). We studied the original material on H. heuffelii from Romania on the Global Plants web-service (HAL0118102 and HAL0118103) and can state without doubts that the plants on both sheets belong to H. umbellatum instead of H. glutinosum. Moreover, we failed to find any specimens of H. glutinosum from any European country except Russia (see also Greimler, 2001). Thus, the authors who treat H. glutinosum as a subtaxon of H. umbellatum may actually be not familiar with true H. glutinosum.

What is Holosteum liniflorum?
The name H. liniflorum Steven ex Fisch. et C. A. Mey. was originally published in F. E. L. Fischer et al. (1837). The authors applied this name to the "glaucous" species with 10 stamens occurred in the Crimea and the area adjacent to the Caspian Sea. Thus, H. liniflorum was listed there with a halfwrong description, which contained some characters of both "glaucous" and "green" species. Arenaria glutinosa was cited as a synonym of H. liniflorum (Fischer et al., 1837). Thus, the latter name is illegitimate according to the Art. 52.1, 52.2(e) of ICN (Turland et al., 2018).
Three years later the same authors published a new combination H. glutinosum applied to the "glaucous" species ("H. glaucum; <…> petalis oblongis subintegerrimis longitudine latitudineque sepala vix excedentibus") (Fischer et al., 1840: 52) instead of the previously published H. liniflorum. At the same time, they provided a new diagnosis for H. liniflorum Steven ex Fisch. et C. A. Mey. 1840[non 1837: "H. viride; <…> petalis obovatis integerrimis latitudine longitudineque sepala multo superantibus" (Fischer et al., 1840: 52). The latter name was applied to the plants from the Crimea, which clearly belonged to the "green" species, although this name was again illegitimate according to the Art. 53.1 of ICN (Turland et al., 2018) (Marhold, 2011) or World Flora Online (2020) usually provide only the "earlier" H. liniflorum (published in 1837) among the synonyms of H. umbellatum subsp. glutinosum but miss the name published in 1840. Only A. S. Zernov and I. V. Sokolov (2004) listed both the "earlier" and the "later" H. liniflorum, although they also did not take into account their different diagnoses and treated both names as synonyms of H. umbellatum var. glutinosum, nom. superfl.
We can confirm that the Crimean H. glutinosum (Fig. 2) has very peculiar large petals, resembling those of a flax ("liniflorum"), and merits its own name, however at the rank of subspecies. After extensive study of the literature, we are not aware that any legitimate name at subspecies rank has been previously published for the taxon that used to be named "H.