New Approach to the Systematics of the Section Psammiris (Iris, Iridaceae): What Does Chloroplast DNA Sequence Tell Us?

Iris sect. Psammiris comprises rhizomatous perennials distributed in the north temperate zone of Eurasia. The systematics of the section are currently based on morphology, and the phylogenetic relationships within it still remain unclear. In the framework of Iris systematics, we conducted molecular and morphological analyses of the currently recognized I. sect. Psammiris species to elucidate the taxonomic composition and relationships within the section. The phylogenetic reconstructions based on sequence variation of four noncoding chloroplast DNA regions support the monophyly of I. sect. Psammiris, which includes I. tigridia, while I. potaninii var. ionantha belongs to I. sect. Pseudoregelia. The proposed novel classification of I. sect. Psammiris recognizes three series: an autonymic series with I. humilis, I. bloudowii, and I. vorobievii and two unispecific series (I. ser. Potaninia with I. potaninii and I. ser. Tigridiae with I. tigridia). In addition, the taxonomic statuses of I. arenaria, I. ivanovae, I. kamelinii, I. mandshurica, I. pineticola, I. psammocola, and I. schmakovii are clarified herein. We provide a revised taxonomic treatment for I. sect. Psammiris, including notes on the types; updated information on species synonymy, distributions, habitats, and chromosome numbers; and a new identification key to the species. Three lectotypes are designated here.

Another species, I. mandshurica, was described based on plants collected near the Razdolnaya River (Primorsky Krai, Russia). It was often considered an ally of I. flavissima or as an intermediate between I. flavissima and I. bloudowii [5,24,26,31]. In the Iris treatments of the Russian Far East flora, I. mandshurica was indicated as a synonym of I. humilis [32,33]. Nevertheless, a preliminary molecular analysis shows that I. mandshurica merits to be recognized as a separate species [34].
While revising the Far Eastern Iridaceae, Nonna Pavlova [32] came to the conclusion that the plants from southern Khasansky District (Primorsky Krai, Russia) collected near the borders with China and North Korea were a separate species, I. vorobievii (Figure 1e). It is a little-known species that was referred to as I. mandshurica by Georgi Rodionenko [1,12]. He believed that I. vorobievii is unrelated to I. humilis and should be transferred to I. sect. Pseudoregelia Dykes, which has not been confirmed by a preliminary molecular study [34].
Iris psammocola, a little-known species, was described by Yu-Tang Zhao on the basis of a single specimen collected in the vicinity of Baijiatan, Ningxia Hui Autonomous Region, China [53]. It has been accepted in the botany databases [42,54]. In 2005, I. psammocola was reported, along with I. potaninii, from the Tsugeer-Els area, a sand cluster of the Ubsunorskaya Kotlovina Biosphere Reserve, southeastern Republic of Tuva, Russia [55]. In addition, I. potaninii var. arenaria Doronkin, described from Kyakhtinsky District, southern Republic of Buryatia, Russia [11], was synonymized with I. psammocola [55]. Subsequently, I. psammocola was reported from the Altan-Els sand dune region of the Borig-Del-Els sandy areas, the Mongolian part of the Uvs Lake Basin [56]. It was asserted that I. psammocola occurs only on sandy arrays and has a disjunctive distribution range in Central Asia [55,56]. It is a relative of I. potaninii and has a chromosome number of 2n = 22 [55]. The same chromosome number has been reported for I. potaninii from eleven localities in the Altai Republic [57,58], from the Republic of Buryatia [57,59], from the Republic of Tuva, and the Zabaykalsky Krai [60]. In our opinion, the Russian populations of I. psammocola remain in question and require further studies, since their identity to I. potaninii is probable.
Iris kamelinii, another relative of I. potaninii, was described by Nina Alexeeva on the basis of plants collected near Verkhniye Boguty Lake on the western side of the Chikhachev Ridge, Southeast Altai Mountains, Russia [61]. According to the author, I. kamelinii in the type locality occurs together with I. potaninii [15,62] and has the same chromosome number, 2n = 22 [58,61]. Moreover, I. kamelinii shows the nearest affinity with I. potaninii in flowering [63] (p. 51) and in seed morphology [64]. Meanwhile, I. kamelinii, treated as endemic to the Altai Mountain Country, also grows in Mongolia and China [17,65].
In view of all the above facts, a molecular study would be a great contribution to understanding the taxonomic composition and phylogenetic relationships among the I. sect. Psammiris species. A few studies based on cpDNA data have examined the relationships between different taxa within Iris, including psammirises [14,50,51,66,67,69,72], and elucidated the I. sect. Psammiris systematics, although only to a limited extent. A combination of trnH-psbA and trnL-trnF was previously proposed as the core barcode for plants [79]. As in our previous publications, in the present study, we focused on nucleotide sequences of four cpDNA noncoding regions (trnH-psbA, rps4-trnS GGA , trnS-trnG, and trnL-trnF) that proved to be useful as phylogenetic markers [34] and that we widely applied to assess interspecific relationships in I. subgen. Limniris (Tausch) Spach [80][81][82]. In the framework of the taxonomic research carried out on Iris, the objectives of the present study are as follows: (1) clarify the phylogenetic relationships of I. sect. Psammiris and I. potaninii var. ionantha with I. tigridia using four cpDNA regions; (2) elucidate the phylogenetic relationships within I. sect. Psammiris and determine the taxonomic statuses of I. arenaria, I. ivanovae, I. kamelinii, I. mandshurica, I. pineticola, I. psammocola, and I. schmakovii; (3) study the morphological characters of the I. sect. Psammiris species; and (4) compare the results of molecular and morphological studies in order to resolve the systematics of I. sect. Psammiris.

Taxa Used
We attempted to provide an extensive taxon sampling as possible and ensure that all accessions were fully verified. One of us (E.V. Boltenkov) undertook two botanical expeditions to southern Siberia (Russia): to the Altay Republic in 2020 and to Transbaikalia (Republic of Buryatia and Zabaykalsky Krai) in 2021. In addition, we collected plant material in Primorsky Krai, Russia, in 2020-2021. The complete list of the examined taxa, including information on samples, is provided in Table 1. The collected samples approximately represent the distribution range of the I. sect. Psammiris species (Figure 2).   Table 1.
The taxon samples for the present study are as follows: I. psammocola from the Republic of Tuva, Russia, including the sample TTL specified in reference [55] (four accessions); I. potaninii from the Altai Republic, Republic of Buryatia, as well as Zabaykalsky Krai, Russia, and Mongolia (25 accessions); I. kamelinii from the type locality (ABL) and two of the three Mongolian specimens specified in reference [17] (three accessions); I. bloudowii from Kyrgyzstan, Kazakhstan, and the Altai Republic, Russia, including AUY, a sample closest to the type locality (10 accessions); I. pineticola from a pine forest in Ukraine west of the type locality (two accessions); I. humilis from Belgorod Oblast, Altai Krai, as well as the republics of Altai, Tuva, and Buryatia, Zabaykalsky Krai, and Amur Oblast, Russia, including four samples from the type locality of I. flavissima (17 accessions); Hungarian samples of I. arenaria from the location where the species was described (three accessions); I. schmakovii from the type locality (two accessions); I. mandshurica from Primorsky Krai, Russia, including two samples (GSS and SRS) from the type locality (four accessions); I. vorobievii from Primorsky Krai, Russia, including a sample (KKR) from the type locality (two accessions); I. tigridia from the Altai Republic, including a sample (ACR) from the type locality (four accessions); I. ivanovae from the Republic of Buryatia and Zabaykalsky Krai, including a sample (ZKV) from the type locality (seven accessions); I. goniocarpa Baker from Sichuan and Gansu provinces, China (two accessions); and I. potaninii var. ionantha from Qinghai Province, China (one accession). The sampling localities for each species under study (except the I. sect. Pseudoregelia species) are shown in Figure 2. Two samples (CQM and CSJ) for which accurate species identification by morphological features was impossible were labeled as unidentified Iris samples.
During the fieldwork in the type localities, I. kamelinii was collected on 6 June 2020 from the northern slope opposite the northern bank of Verkhniye Boguty Lake, where it was found in flowering on soddy soils of mountainous steppes [84] on the hill and in fruiting opposite the hill at the base of the mountain slope (Figure 3a,b). Iris ivanovae was collected on 5 June 2021 at the end of flowering from a chestnut soil in a dry steppe heated at noon, where it was found growing, along with Stipa krylovii Roshev., on a sunlit lower part of the northern slope (Figure 3c,d). The type locality of I. potaninii var. arenaria was inspected twice; however, these plants were not found (although I. tigridia was abundant there), and the taxon is therefore not included in the analysis. No samples of I. psammocola were available from the type locality.

Plant Samples, DNA Extraction, and Sequencing
For genetic analysis, leaf samples were collected across the distribution range of the I. sect. Psammiris species. Total genomic DNA was isolated from the leaf samples collected during the fieldwork and dried in silica gel or taken from the herbarium specimens deposited at ALTB, E, KW, LE, MHA, MW, NENU, and UUH (herbarium codes according to Index Herbariorum [83]). The methods for DNA extraction, amplification, and direct sequencing of four cpDNA noncoding regions (trnH-psbA, rps4-trnS GGA , trnS-trnG, and trnL-trnF) were described previously [34,85]. The cycle sequencing reactions were performed on both strands, and fragments were separated on an ABI 3130 genetic analyzer (Applied Biosystems, Bedford, MA, USA) at the Joint Center of Biotechnology and Gene Engineering, the Federal Scientific Center of the East Asia Terrestrial Biodiversity, Far Eastern Branch, Russian Academy of Sciences (Vladivostok, Russia). Forward and reverse sequences for each region were assembled using the Staden Package, version 1.4 [86]. In a preliminary study, no polymorphism in the cpDNA regions was found in the sample of five individuals from the localities of I. arenaria, I. humilis, I. kamelinii, and I. potaninii; therefore, one specimen from each locality was used for further analysis. The sequences of the four cpDNA regions obtained for 88 accessions representing 14 taxa were deposited in the GenBank database. The accession numbers for all the sequences used are listed in Table 1.

Sequence Alignment and Phylogenetic Analysis
The sequences of each cpDNA region were aligned manually in SeaView version 4 [87] using the CLUSTAL algorithm, manually edited when necessary, and concatenated for each specimen. We included indels and length variation in mononucleotide repeats in the dataset because the repeatability tests allowed for exclusion of PCR errors. In the dataset, we also included the sequences for the most frequent haplotypes identified previously [34] in the localities of I. humilis (ALT-03), I. mandshurica (NAKH-01 and NAKH-07), and I. vorobievii (KRAS-01, KRAS-03, and KRAS-04). The haplotypes were identified on the basis of combined DNA sequences using DnaSP version 5 [88]. A network of haplotypes was constructed using Network version 4.6 [89], with each deletion/insertion treated as a single mutational event, regardless of size, and using the MJ method with default settings.
Phylogenetic relationships among the I. ser. Psammiris species were assessed using the MP and ML methods as implemented in PAUP version 4.0 b10 [90], as well as the BI method in MrBayes version 3.2.2 [91] via the CIPRES portal [92]. The dataset for the phylogenetic analysis included haplotypes obtained previously [80][81][82] for I. dichotoma of I. subgen. Pardanthopsis (Hance) Baker and for 14 species representing four series of I. subgen. Limniris as outgroups. For the MP method, optimal trees were found using a heuristic search with 1000 random addition sequence replicates, starting trees obtained via stepwise addition, TBR branch swapping, and the MulTrees option in effect. For the ML and BI methods, the GTR + I + G model was selected according to the Akaike information criterion using Modeltest version 3.6 [93]. ML heuristic searches were performed using the resulting model settings, 100 replicates of random sequence addition, TBR branch swapping, and the MULTrees option. In BI, using the default prior settings, two parallel MCMC runs were carried out for 10 million generations, with sampling every 1000 generations for a total of 10,000 samples. Convergence of the two chains was assessed, and PP was calculated from the trees sampled during the stationary phase. The robustness of nodes in the ML and MP trees was tested using bootstrap with 1000 replicates.
Degrees of divergence between the species were calculated based on nucleotide substitutions using DnaSP. Pairwise F ST among them were determined by AMOVA as implemented in Arlequin version 3.5 [94]. Significance of genetic distances was tested using 1000 random permutations.

Morphological Data
To compile a morphological key to the accepted species of I. ser. Psammiris in the present study, 22 characters were selected for comparison: (1) rhizome shape (creeping, forming branches like stolons or shortened or nodose, slowly creeping (compact)), (2) rhizome diameter, (3) root shape (adventitious roots gradually tapering to the apex, not thickening (equal); fleshy at the proximal part, resembling a cone (obconical); or evenly thickened at the proximal part with wrinkled transverse patterns (contractile)), (4) root diameter (measured at the proximal end), (5) leaf shape (straight, sword-shaped rosette leaves with more or less parallel margins (ensiform) or one slightly convex margin falcate at the distal part and margins abruptly apically narrowed (subfalcate)), (6) leaf apex (rosette leaves apex straight or slightly incurved, gradually narrowed (narrowly acute), or abruptly narrowed (acute)), (7) leaf texture (rosette leaves noticeably tough or less tough and flexible (thin); the surfaces finely ribbed (smooth); or with discrete central veins (ribbed)), (8) leaf length (measured from the base to the apex of the longest rosette leaf), (9) leaf width (measured at the broadest part of the widest rosette leaf), (10) stem height (measured from the base of the flowering stem to the base of the outer bract), (11) stem branching (classified as simple, bearing only the terminal cluster (designated as 0), or branched, with 1-2 lateral one-flowered cluster(s)), (12) number of flowers (flowers per stem), (13) cauline leaf length (measured from the base to the apex of the upper cauline leaf), (14) number of bracteoles (secondary bracts, i.e., bracteoles, per terminal cluster of the inflorescence), (15) bract length (measured from the base to the apex of the outer bract of the terminal cluster), (16) bract texture (coriaceus, pliable but thin when dry bracts (tough) or membranous and somewhat translucent (thin)), (17)  The scores of the characters for each species were obtained from living specimens collected from wild localities; from our own observations of herbarium specimens at AA, ALTB, BM, E, IRK, K, LE, MHA, MW, NENU, NS, NSK, TK, UUH, VBGI, and VLA, including the original material for the names studied; and from the relevant species descriptions available in the literature [32,45,[95][96][97]. The rhizome and root diameter were measured in the dry state with a digital Vernier caliper Series 532 (Mitutoyo, Aurora, IL, USA). Because I. psammocola is not represented in the Chinese botany databases [54,98,99], its taxonomy is based on a comprehensive study of the protologue.
In the Taxonomic Treatment section (see below), we gathered information on the distribution of the accepted species from our own field data, the herbarium specimens, and relevant literature [47] and critically assessed the collection points a priori from social networks [98,99,101-103].

Genetic Divergence and Phylogenetic Relationships within Iris Sect. Psammiris
Four cpDNA regions were sequenced from 72 accessions of 10 I. sect. Psammiris species, 4 accessions of I. tigridia, 7 accessions of I. ivanovae, 2 accessions of I. goniocarpa, and 1 accession of I. potaninii var. ionantha, as well as from 2 samples of unknown species. A total of 18 haplotypes were identified among the samples from 10 species of I. sect. Psammiris, I. tigridia, and I. ivanovae based on polymorphic sites found at 3783 aligned positions of a combined dataset. The distribution of these haplotypes among the studied species is shown in Figure 2.
A total of 6 haplotypes (H1-H4, H6, and H7) were identified in 25 localities of I. potaninii, 3 haplotypes occurred in 17 localities of I. humilis (H9-H11), and 3 haplotypes occurred in 3 localities of I. kamelinii (H1, H3, and H8). Two haplotypes were found in I. psammocola (H1 and H5), I. arenaria (H12 and H13), I. schmakovii (H9 and H11), and I. ivanovae (H17 and H18); the following species showed one haplotype each: I. pineticola (H9), I. bloudowii (H15), I. mandshurica (H14), I. vorobievii (H16), and I. tigridia (H17). Of the six haplotypes found in I. potaninii, haplotype H7 was shared by the accessions from ZTLW, ZTLN, and ZAC; haplotypes H2 and H4 were shared by the accessions from two localities; and the accessions from the other 16 localities of I. potaninii shared a single common haplotype, i.e., H1. The latter was found to be common to three species: I. potaninii, I. psammocola, and I. kamelinii. Another haplotype (H3) was common to I. potaninii and I. kamelinii (localities MKS and MAK, respectively). Of the three haplotypes found in I. humilis, haplotype H9 proved to be the most frequent: it was shared by the accessions from 11 out of 17 localities. Moreover, this haplotype was also found in both studied localities of I. pineticola, while I. schmakovii shared two haplotypes (H9 and H11) with I. humilis.
The genealogical relationships between the haplotypes of the studied species are shown in Figure 4. All the haplotypes, including NAKH-01, NAKH-04, NAKH-07, KRAS-01, KRAS-04, KRAS-07, and ALT-03 retrieved from reference [34] and the haplotypes of I. goniocarpa and I. dichotoma, were connected in a single network. All of them, except for the haplotypes of I. goniocarpa, were closely related and originated from the same unsampled or extinct ancestral haplotype connected via many mutational steps with the haplotype of I. dichotoma. Three haplogroups were detected in the network, separated from each other by several mutational steps (six to eight). Haplogroup I included closely related haplotypes H1-H8 arranged into a star-like pattern around haplotype H1, which was common to I. potaninii, I. kamelinii, and I. psammocola. The pairwise F ST values between these species were not significant (p > 0.1), no nucleotide substitutions or indels differentiating these species were revealed, and K S between them varied from 0.00006 to 0.00015, indicating a lack of genetic differences between these species.  Table 1.
The other two haplogroups (II and III) descended from a haplotype that may be either extinct or missing from the current sampling. The pairwise F ST value between these groups was 0.782 (p = 0.00001), and the K S between them was 0.00126. Four nucleotide substitutions and 8 bp insertion distinguished the species from these groups. Haplogroup II included two haplotypes found in I. tigridia and I. ivanovae, of which one was common (H17) and the other (H18) was found only in samples from the I. ivanovae localities, differing from H17 by an insertion of 25 bp within the trnH-psbA spacer. The low and non-significant F ST value (p > 0.05) between I. tigridia and I. ivanovae and the absence of sequence divergence between them (K S = 0.0000) may indicate that they belong to the same species.
Haplogroup III included 14 closely related haplotypes found in 7 species: I. arenaria, I. bloudowii, I. humilis, I. mandshurica, I. pineticola, I. schmakovii, and I. vorobievii. Haplotypes in this haplogroup differed from the neighboring haplotypes by one or two mutational steps. The most frequently occurring haplotype (H9) occupied a central position in this haplogroup and was common to most I. humilis accessions from different parts of the range, as well as to the accessions from the two I. pineticola localities and to one I. schmakovii accession (Figures 2 and 4). Many haplotypes of haplogroup III were connected with H9 via one (H12 of I. arenaria, H14 of I. mandshurica, and H15 of I. bloudowii) or two mutational steps (H10, H11, and ALT-03 of I. humilis), forming a star-like structure. The haplotypes of I. mandshurica, which were interconnected via one or two mutational steps, were also closely related to the haplotypes of I. humilis. Alternative connections (loops in the network) between some haplotypes, including the most common haplotype (H9), indicated a homoplasy that hampered unambiguous identification of genetic relationships between the haplotypes of I. mandshurica and I. humilis. The haplotypes of I. vorobievii formed a group with a single haplotype of I. bloudowii (H15), which differed from the most common haplotype (H9) by a single substitution.
Trees with nearly identical topologies and with slight differences in statistical supports of some nodes were inferred by the MP, ML, and BI methods based on the cpDNA dataset ( Figure 5). In these trees, all the accessions were distributed with a robust support (PP 1.0, BP > 90%) in accordance with their affiliation to the corresponding sections of the genus Iris. The sister-group relationship between I. sect. Pseudoregelia and I. sect. Psammiris was strongly supported (PP 1.00, BP 93, and 94%). The position of the CQM and CSJ accessions in the I. sect. Pseudoregelia group, together with I. goniocarpa and I. potaninii var. ionantha, was also strongly supported (PP 1.00, BP 99, and 100%).
The species of I. sect. Psammiris formed a monophyletic clade (PP 1.00, BP 99, and 100%), with I. tigridia and I. ivanovae nested within it. This clade was divided into two sister subclades, with the nucleotide divergence (K S ) between them being 0.00182. Subclade I corresponded to haplogroup I revealed by the MJ methods ( Figure 4) and included all the samples of I. potaninii, I. psammocola, and I. kamelinii, with moderate support (PP 0.94, BP 77, and 76%). Subclade II combined I. tigridia, I. ivanovae, and all other species recognized in I. sect. Psammiris (PP 1.00, BP 86, and 87%). This subclade, in turn, was divided into two well-supported clusters, of which one (cluster 2, support values of PP 1.0, BP 84, and 87%) contained the samples of I. tigridia and I. ivanovae, while the other (cluster 3, PP 1.0, BP 86, and 88%) included haplotypes of seven species (I. arenaria, I. bloudowii, I. humilis, I. mandshurica, I. pineticola, I. schmakovii, and I. vorobievii). These two clusters corresponded to haplogroups II and III revealed by the MJ methods. A low nucleotide divergence was observed between the species within cluster 3 (K S ranged from 0.00001 to 0.00088), and the relationships between I. arenaria, I. humilis, I. mandshurica, I. pineticola, and I. schmakovii remained unresolved. Only the haplotypes of I. bloudowii and I. vorobievii formed a group that received weak support from the MP and ML methods (BP 58, 53%) and high support only from the BI method (PP 0.96).  Table 1.
Thus, the results of this study based on sequencing of the cpDNA regions of 12 taxa show that I. sect. Psammiris includes five species (I. bloudowii, I. humilis, I. potaninii, I. tigridia, and I. vorobievii) and is divided into three groups.

Morphological Comparison of the Iris Sect. Psammiris Species
A detailed morphological comparison among the I. sect. Psammiris species accepted in the present study is listed in Table 2 (also see Table S1). They can be easily distinguished by the number of flowers and bracteoles, as well as by the stem height and perianth tube length. Iris bloudowii, I. humilis, and I. vorobievii, forming a group of related species, are distinguished as having non-contractile, adventitious roots; thin and broad rosette leaves, usually with more than one flower and with bracteoles; tough green bracts; long pedicels; and a short perianth tube. Among them, I. humilis is similar to I. bloudowii but differs in its habit (less robust), number of bracteoles (from their absence to three; Figure 6a-c), and number of flowers (up to three; Figure 6b,c); moreover, it occasionally occurs in clumps (Figure 1c). In addition, a white-flowered form of I. humilis, a rare feature in this section, was found in Partizansky (S. Prokopenko, pers. comm.) and Khankaysky (A. Malyk, pers. comm.; Figure 1b) districts, Primorsky Krai, Russia. Iris vorobievii is characterized by a very short rhizome (up to 2 cm in length), storage-like roots spreading almost horizontally (Figure 6d,e), and often branched stems (e.g., VBGI79851; see http://botsad.ru/herbarium/, accessed on 20 December 2022). Iris potaninii is similar to I. tigridia in the characters of roots (contractile), rosette leaves (ensiform, acute, or narrowly acute at the apex; tough; narrow; 0.1-0.6 cm wide), stem (with non-curled remains of leaves at the base, simple, 1-flowered, without bracteoles), bracts (lanceolate, whitish, and thin), pedicel (extremely short, less than 0.7 cm), and elliptical fruit. However, I. potaninii is distinguished from I. tigridia by its shortened, branching rhizomes; by having fibrous remains of leaves (vs. the rhizome surface glabrous); often forming large colonies or clumps (Figure 1g,h); by longer (up to 50 cm long), slightly less thickened roots; by a much shorter stem (to 2.5 cm long) not emerging above ground, resulting in fruit always being borne at the soil surface (Figure 6f,g); by having an extremely short internode between the upper cauline leaf and bracts (barely 0.2 cm long); and by its even longer perianth tube (usually more than two times as long as that of other species in the section). After opening, the color of I. potaninii flowers is bright yellow, subsequently turning into pale yellow (Figures 1g and 6h, respectively). The species is variable in the color intensity of broken lines (brownish against yellow background) of the fall blade (Figure 6i-k), the shape (obovate or elliptic) of the inner perianth segments (or standards), in terms of whether they gradually or abruptly narrow into the claw (Figure 6i,j), and in terms of whether falls and standards have a notch at the apex (emarginated) (Figure 6i-k). Variability in these characters can be observed within the same locality or even clump.
Iris tigridia is clearly distinguished by its flower color, varying from pale blue to dark blue and purple or, rarely, white (Figures 1i-m and 6l,m); however, it is never yellow as others in I. sect. Psammiris. Generally, it is variable (within a locality) in leaf length and width, stem height, and cauline leaf length (Table 2), as well as in floral diameter (3.5-6 cm).

Discussion
This study presents the most comprehensive phylogenetic analysis for I. sect. Psammiris of those carried out to date. The reported results provide new insights into the taxonomic composition and classification of this section. The samples represent almost all known taxa (a total of 12) in I. sect. Psammiris. Specimens of the species closely related to I. humilis (I. arenaria, I. pineticola, and I. schmakovii) and I. potaninii (I. kamelinii and I. psammocola) were included in the phylogenetic analysis for the first time. Our sampling of all the currently recognized species of I. sect. Psammiris from different parts of the ranges and type localities made it possible to clarify the genetic relationships between them, as well as with I. tigridia and I. ivanovae, which are now considered as representatives of I. sect. Pseudoregelia [1,11,67,68].
The monophyly I. sect. Psammiris has been questioned by other authors [14,50,51,66], since I. potaninii var. ionantha was shown to be related to I. sect. Pseudoregelia. The findings of this study clearly show that all the specimens of I. tigridia and its close relative, I. ivanovae, belong to the clade of I. sect. Psammiris (Figures 4 and 5), which is monophyletic and sister to the clade formed by the taxa of I. sect. Pseudoregelia, including I. potaninii var. ionantha. The phylogenetic placement of I. potaninii var. ionantha is fully congruent with the tree topologies inferred in recent phylogenetic studies [14,50,51].
Within the I. sect. Psammiris clade, we revealed three well-supported monophyletic groups (haplogroups in the MJ network and clusters in the phylogenetic tree) treated by us at the series level, two of which we consider unispecific. The first group includes I. kamelinii, I. potaninii, and I. psammocola. These species have common or closely related haplotypes and demonstrate the lack of clear differentiation from each other (Figures 4 and 5). Therefore, the first group can be considered unispecific. We choose I. sect. Psammiris ser. Potaninia Doronkin to represent the group, as it has the same type (I. potaninii) as Doronkin's original group [11]. A thorough revision of the morphological characters previously proposed to distinguish between these species confirmed the lack of clear differences.
The following diagnostic features were used to distinguish I. kamelinii from I. potaninii: rhizomes bear membranous remains of leaf bases (vs. fibrous remains); standards roundedelliptic, emarginated at apex, abruptly narrowed into a linear claw (vs. obovate, gradually narrowed into a claw); and ornamentation of falls with a dense pattern of purple veins (vs. veins poorly visible) [61]. In the present study, we clearly showed that these features of I. kamelinii are identical or slightly differ from those of I. potaninii (see Section 3.2); thus, I. kamelinii does not have any diagnostic features that clearly distinguish it as a distinct species. It has long been noted that the standards in I. potaninii are usually emarginated at the apex [95][96][97]. Our data confirm that the standards in I. potaninii can be emarginated (or not) at the apex, even within the same plant from the type locality of I. kamelinii (Figure 6k); fall ornamentation is a variable character in I. potaninii (Figure 6i-k). As a consequence, we regard I. kamelinii as a synonym of I. potaninii.
The plants of I. psammocola from the Republic of Tuva, Russia [55], and the plants from the type locality of I. kamelinii [61] are found growing together with I. potaninii and all have the same chromosome number, i.e., 2n = 22 (see Introduction). In China (where I. psammocola was described), this species is known to date only from the protologue consisting of a diagnosis, a description, and an illustration [53]; from a single specimen deposited at NENU (NENU00014009!; Figure 7), which is a holotype of the name; and from reference [46]. The holotype of I. psammocola is represented by a small herb plant in flowering collected in early April. This specimen has a rhizome (broken) of about 0.5 cm in diameter; its adventitious roots are yellowish white, thickened at the proximal part, and gradually tapering to the apex, up to 20 cm long; the rosette leaves are ensiform, narrowly acute at the apex, tough, and finely ribbed, up to 18.5 cm long and 0.2-0.4 cm wide; the flowering stem is very short, not emerging above ground, probably not more than 2 cm tall (the height was impossible to measure), simple, bearing one terminal flower, and without bracteole; the stem and rhizome bear erect (non-curled) fibrous remains of leaves; two bracts are lanceolate and membranous; the pedicel is very short; the perianth tube is filiform, about 5 cm long; the outer perianth segments have a distinct beard, about 4.5 cm long. According to references [46,53], the rhizome of I. psammocola is short and non-stoloniferous, the bracts are 3.5-4 cm in length, and the flowers are yellow. After a critical examination of the I. psammocola protologue, we found that the features of the rhizome, roots, rosette leaves, flowering stem, bracts, and flowers are identical to those of I. potaninii (Table 2). Our analysis of cpDNA variability indicate the lack of genetic differences between the specimens of I. potaninii and the specimens from the Republic of Tuva, Russia, including the specimen (TTL) treated as I. psammocola (Figures 4 and 5).
The phylogenetic analysis reported above (Figures 4 and 5) confirmed that the plants from the type locality of I. tigridia and the plants from the Republic of Buryatia and Zabaykalsky Krai, here named as I. ivanovae, belong to the same species, I. tigridia, which is nested in I. sect. Psammiris and comprises a separate unispecific series. To describe I. ivanovae based on plants from Zabaykalsky Krai, Russia, the following diagnostic features were used to distinguish it from I. tigridia: flowers 2.5-3.5 cm in diameter (vs. flowers 4.0-6.0 cm in diameter); falls abruptly narrowed into a long, filiform claw (vs. falls gradually narrowed into a thin claw); bracts narrowly lanceolate, gradually acuminate (vs. bracts oblong-elliptical and short-pointed at the apex); and leaves gradually acuminate, 0.1-0.2 cm wide (vs. leaves shortly acuminate and 0.4-0.5 cm wide) [73]. However, our field study at the type locality of I. ivanovae did not confirm some of these features. We found that the flowers were mainly 3.5-6.0 cm in diameter, the falls were gradually (not abruptly) narrowed into a thin claw, 0.1 cm wide at the base, and the leaves were 0.1-0.3 cm wide (Figure 6l). The diameter of one of the ten flowers that we found was only 2.5 cm due to the underdeveloped blades of falls at the apex (Figure 6m). This plant can be considered merely an aberrant, which can be explained by the climatic conditions of the locality where it grew. In addition, we found that the lanceolate, gradually acuminate bracts are characteristic of all the plants of I. tigridia from Siberia, as well as the plants with leaves gradually narrowed to the apex. However, to the best of our knowledge, the latter dominate the Transbaikalian steppes and Mongolian habitats due to the rather xerophytic conditions. Furthermore, leaves in I. tigridia were previously characterized as gradually narrowed to the apex [45,95,96]. Thus, we did not find any differences between the plants from the type locality of I. ivanovae and the plants from the I. tigridia distribution range, which confirmed Gubanov's opinion [76] that I. ivanovae is a synonym of I. tigridia. The third group, revealed by the MJ methods and phylogenetic analyses of I. sect. Psammiris, comprises seven species: I. arenaria, I. bloudowii, I. humilis, I. mandshurica, I. pineticola, I. schmakovii, and I. vorobievii (Figures 4 and 5). The isolated position of I. bloudowii and I. vorobievii in this group, along with the data on their morphology presented here (Table 2), are consistent with the results of previous studies [15,27,34,[46][47][48] that showed them as separate species. Phylogenetic relationships between I. arenaria, I. humilis, I. mandshurica, I. pineticola, and I. schmakovii from Hungary, Ukraine, Mongolia, and Russia (Belgorod Oblast, Altai Krai, Altai Republic, Republic of Tuva, Republic of Buryatia, Zabaykalsky Krai, Amur Oblast, and Primorsky Krai) remain unresolved, and our results (the shared haplotypes and the star-like pattern) indicate a lack of clear genetic differentiation between them, which suggests that they belong to a single species, I. humilis.
Iris humilis var. umbrosa was described based on plants collected on the right bank of Lake Khuvsgul, Khuvsgul Aimag, Mongolia. As follows from the brief description, it is a plant with a height of 20-30 cm; green linear-lanceolate leaves 3-8 mm wide ; yellow flowers with purple veins ; bracts coriaceus ; wide, swollen, acuminate, and fruit elliptical, tapering at the apex [70]. Unfortunately, the taxon I. humilis var. umbrosa was published without a diagnosis, and in our opinion, it is still unclear what distinguishes it from the autonymic variety. Moreover, all the features indicated in the protologue of I. humilis var. umbrosa and in reference [17] are identical to those of I. humilis ( Table 2). All experts on the Mongolian flora listed I. humilis for the Khuvsgul phytogeographical region [27,[74][75][76][77]. Despite the considerations mentioned above, Alexeeva referred to a "more detailed comparative morphological analysis of characters" (that she, however, never presented) and came to the conclusion that I. humilis var. umbrosa is actually a new species, I. schmakovii [17]. In accordance with the molecular data presented here, I. schmakovii belongs to I. humilis (Figures 4 and 5).

Taxonomic Treatment
In the present study, we propose I. sect. Psammiris to be divided into three series consisting of five species. In particular, we confirm that I. tigridia, the type species of I. sect. Pseudoregelia ser. Tigridiae [11], is nested in I. sect. Psammiris. Therefore, we suggest excluding I. ser. Tigridiae from I. sect. Pseudoregelia, as originally published in [11], and transferring it to I. sect. Psammiris. In addition, I. ser. Humiles Doronkin and I. ser. Vorobievia Alexeeva are synonymized here for the first time with the autonymic series of I. sect. Psammiris.
Moreover, as found in the present study, there are some problems related to the type citation in I. sect. Psammiris; therefore, the following issues should be addressed: (i) Taylor indicated I. humilis as the type species of I. sect. Psammiris [9]. Since then, this approach has been accepted [11,12,14,15,71,113]. However, I. subgen. Psammiris was actually published by Spach as a monotypic taxon based on I. arenaria, although he noted this group to apparently also include I. flavissima and I. bloudowii as follows: "Huc referendae etiam videntur Iris flavissima, Jacq., et Iris Bloudowii, Ledeb." [3]. We tend to interpret this phrase as non-inclusion of I. flavissima and I. bloudowii in I. subgen. Psammiris by Spach. He did not include these species in I. subgen. Psammiris either in the following study published four months later [114], which can be considered an indirect argument in favor of our opinion. The type of I. arenaria, not I. humilis, is therefore, the type of I. sect. Psammiris (see Art. 10.3 of the ICN).
(ii) The herbarium sheet at MW (MW0021793!) with a label handwritten by Georgi ("Iris pumila ad Baical, 1772"), which is the current lectotype of I. humilis [115], consists of four plants representing two species: I. pumila L. and I. humilis (as currently applied). When a type contains parts belonging to more than one taxon, the initial choice is superseded (see Art. 9.19 of the ICN), and the name must remain attached to the part that corresponds most nearly with the original description or diagnosis (Art. 9.14 of the ICN). Hence, because MW0021793 proved to be mixed and belong to more than one taxon, it cannot be accepted as a type of I. humilis, as previously proposed [115]. One of us (E.V. Boltenkov) numbered the plants belonging to I. humilis from the Lake Baikal area as 1 and 2, which was noted by Alexeeva [15], and the plants belonging to I. pumila of unknown origin as 3 and 4. However, Alexeeva [15] did not achieve the type designation because the typification statement did not include the phrase "designated here" or an equivalent (see Art. 7.11 of the ICN).
(iii) Contrary to the Alexeeva's statements [116], the type of I. tigridia was not indicated by Grubov [27] (see Art. 40 Note 2). Similarly, the type designation of this name was not effectively published by Alexeeva [15,116], as required by Art. 7.11 of the ICN.
(iv) Iris pineticola was published [39] (p. 407) as a replacement name for I. flavissima subsp. stolonifera f. orientalis Ugr. Hence, the latter name is its replaced synonym (Art. 6.11 of the ICN) that has the same type as that of the replacement name (see Art. 7.4 of the ICN). Klokov indicated the specimen deposited at KW as the "typus speciei" of I. pineticola as follows: "RSS Ucr., dit. Charcoviensis, in pineto prope pag. Choroshevo, 5-6 V. 1855. Legit B.M. Czernjajev; in Herbario Instituti Botanici Ac. Sc. RSS Ucr. conservatur" [39]. While preparing his publication [22], Ugrinsky used the Vassilii Czernajew's herbarium; therefore, Klokov's indication could have been accepted as the lectotype for I. flavissima subsp. stolonifera f. orientalis, satisfying the requirements of Art. 7.11 of the ICN. However, Czernajew's specimen cited by Klokov [39] was lost or destroyed, and for this reason, a neotype of I. pineticola (KW000114271) was selected (see Art. 9.16 of the ICN) [41]. Unfortunately, the authors of the latter paper did not consider all the original material in the context of the protologue of I. flavissima subsp. stolonifera f. orientalis (see Arts. 9.4 and 9.13 of the ICN), which contains an illustration [22] (p. 307). The same illustration was provided by Klokov in reference [39] (p. 293). In accordance with Art. 9.19 of the ICN, the choice of the neotype [41] should be superseded since the original material (illustration) was found to exist and can serve as lectotype.
As a consequence, lectotypes are designated here for I. flavissima subsp. stolonifera f. orientalis, I. humilis, and I. tigridia.

List of Taxa
Below is a list of the accepted species (highlighted in bold italics) that contains information on their synonyms and nomenclatural types, as well as on their distributions, habitats, and chromosome numbers. Republic of Khakassia, Republic of Tuva, southern Krasnoyarsk Krai), and northwestern China (Xinjiang). It grows on grassy subalpine and alpine meadows, among shrubs and in shady places, on hillsides or at forest edges, and along mountain streams at elevations of 850-2200 m.
Chromosome number: 2n = 16 [58,60].   Distribution and habitat: This species is distributed in southern Siberia, Russia (southern Krasnoyarsk Krai and Republic of Khakassia, southeastern Altai Krai, Altai Republic, Republic of Tuva, southern Republic of Buryatia, and Zabaykalsky Krai), eastern Kazakhstan (East Kazakhstan Region), northern Mongolia, and China (Shanxi, Hebei, Jilin, and Liaoning provinces, Beijing, and Inner Mongolia). It grows in gravelly, stony, or sandy places in steppes among grasses, as well as on dunes, rocky slopes, and often on hilltops at elevations of 400-1200 m.

Conclusions
Although many specialists have carried out extensive studies of Iris sect. Psammiris, a number of taxonomic problems in this section remain unresolved. Here, we present the first comprehensive molecular phylogeny of the section, with a large set of samples covering most of the distribution ranges and type localities of the species and almost all of its previously recognized taxa. The results obtained in the present study confirm that all previous data, based solely on morphological characters, do not fully clarify the taxonomic composition and phylogenetic relationships among the I. sect. Psammiris species. Our results based on cpDNA data provide a number of novel insights. The important finding is that the phylogenetic results strongly support the monophyly of I. sect. Psammiris and the placement of I. potaninii var. ionantha in the I. sect. Pseudoregelia clade, which is sister to I. sect. Psammiris. It should also be emphasized that the taxonomy of I. potaninii var. ionantha requires further research. Furthermore, the molecular studies confirm the placement of I. tigridia in I. sect. Psammiris rather than in I. sect. Pseudoregelia.
Other our results are, in general, as follows: (1) five species (I. arenaria, I. humilis, I. mandshurica, I. pineticola, and I. schmakovii) should be treated as a single species, i.e., I. humilis; (2) the specimen listed in reference [55] as I. psammocola from Russia and other studied samples from the Tsugeer-Els area (Republic of Tuva, Russia), also referred to as I. psammocola, belong to I. potaninii; (3) a critical evaluation of the original material and literature showed that I. psammocola and I. potaninii are the same taxon; (4) the specimens of I. kamelinii from the type locality and from Mongolia [17] also belong to I. potaninii; (5) the molecular data and a critical examination of the type material and living plants from the type locality confirm that I. ivanovae, which has been recognized on the basis of morphology, is a synonym of I. tigridia. In view of the findings reported above, we provide an updated classification of I. sect. Psammiris. The section is unambiguously subdivided into an autonymic series with three species (the most widespread bearded iris I. humilis, I. bloudowii, and I. vorobievii) and two unispecific series: I. ser. Potaninia with I. potaninii and I. ser. Tigridiae with I. tigridia. Thus, here, we present a new taxonomic treatment for I. sect. Psammiris and an identification key for all of its species. The members of this section are distributed from southeastern Europe through southern Siberia, northern Kazakhstan, China, and Mongolia to the Russian Far East. The results presented herein will undoubtedly contribute to our understanding of the phylogenetic relationships within Iris s.l. and the taxonomic composition of the genus in Russia and adjacent areas.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/plants12061254/s1, Table S1: Raw data of the morphological analysis (the codes of the characters are provided in Abbreviations). Funding: This research received no external funding.

Data Availability Statement:
The sequences resulting from this study are available in the NCBI database (https://www.ncbi.nlm.nih.gov/, accessed on 15 January 2023) with GenBank accession numbers ON569443-ON569794.