﻿A new species of Atriplex (Amaranthaceae) from the Indian subcontinent

﻿Abstract A new subshrubby C4-species from the lowlands and foothills of India, Pakistan and SE Afghanistan, Atriplexpseudotatarica, is described and illustrated. Previously, it was incorrectly identified as A.crassifolia auct. non C.A.Mey. belonging to a distant C3-group of the genus. A phylogenetic analysis based on nrITS and nrETS revealed its position as sister to A.schugnanica (sect. Obionopsis). Both species share aphyllous inflorescence and smooth bract-like cover, but differ in life form, leaves, seed colour, and geographical distribution. We revised native Indian Atriplex species and excluded some of them from the flora of the country. An improved checklist of the native Atriplex species in India with their corrected synonymy and nomenclature is given, and a new diagnostic key is provided.


Introduction
Atriplex L. is the largest genus in the Amaranthaceae clade encompassing ca. 260 species distributed mostly in arid regions of the world (Žerdoner Čalasan et al. 2022). There is a relatively limited number of Atriplex species in the Indian subcontinent. The latest treatments for the flora of Pakistan (Ali and Qaiser 2001) and India (Paul 2012) counted 12 and seven species, respectively. Sukhorukov et al. (2019) revised Atriplex in the Himalayas and Tibet, and provided many taxonomic changes for the genus in the Indian Himalaya compared to the previous checklists and floras. All Atriplex species native to the Himalayas are represented by the annual C 4 species, but they have different origins. Two of them, A. pamirica Iljin and A. centralasiatica Iljin, are typical Central Asian elements, whereas A. schugnanica Iljin originated in the eastern Irano-Turanian region (Žerdoner Čalasan et al. 2022). Subsequently, the first two species are classified within A. sect. Obione (Gaertn.) Reichenb., and A. schugnanica is a member of A. sect. Obionopsis (Lange) Dumort. . In comparison to the species distributed in the Himalayas, the species growing in lowlands and foothills of the Indian subcontinent are still undercollected and poorly known, because the classical authors preferred to stay in the mountains rather than in the plains during the summer time due to harsh climatic conditions in the latter region, and had a preference for species-rich plant diversity in the mountains.
Unusual Atriplex plants were noted in the year 2021 in Haryana State (India) by one of the authors (NS) of the present paper. Further in situ studies have confirmed an assumption that the specimens cannot be assigned to any known species or their synonyms, and should be described as a new species.

Material investigated
Field studies were carried out in the Haryana State (India). Taxonomic revision of the herbarium material was undertaken in the herbaria BM, CAL (examined as digital images), DD, K, LE, MHA, and MW. Distribution map is based on the specimens cited in the text and was prepared using SimpleMappr online tool (http://www.simplemappr.net).

Sampling of the study, DNA extractions, amplification, and sequencing
Thirty-seven accession numbers were included in the phylogenetic analyses representing Atriplex species, and two accession numbers were taken as outgroups from Amaranthaceae. The samples are listed in Table 1. We have reconstructed a part of the global phylogenetic tree originally published by Žerdoner Čalasan et al. (2022) and indicated the position of the new species among its close relatives.
Among 16 species analyzed in A. sect. Obionopsis and close relatives (A. flabellum Bunge, A. moneta Bunge), 15 accessions were represented by ITS and ETS loci (see below) (Table 1). We included only ETS sequences for one species (A. kalafganica). Following Kadereit et al. (2010), we selected Halimione pedunculata (L.) Aellen as an outgroup for ITS-and ETS-based molecular phylogenetic analyses. In short, we analyzed 37 ITS and ETS sequences of 19 taxa (Table 1). We obtained two of these sequences (one of ITS and one of the ETS regions of rDNA) as a part of this study (Table 1) and took the remaining ones from the study of Žerdoner Čalasan et al. (2022).
The DNA from a sample of A. pseudotatarica collected in the state of Haryana, India (see also the Results section) was extracted from 5-10 mg of dried leaves employing the DNeasy Plant Mini Kit (Qiagen, the city of Valencia, CA, USA), as described in the manual.
PCRs were carried out in Thermal Cycler T100 (Bio-Rad, USA) using the primers and cycling protocols summarized in Table 2.
PCR products were purified with the Cleanup Mini BC023S Kit (Evrogen, Russia) following the manufactured instructions. Sanger sequencing was performed at Evrogen JSC (Moscow, Russia) employing PCR primers (Table 2).

Alignment and phylogenetic analyses
The L-INS-i alignment strategy with default settings of MAFFT version 7.0 (Katoh et al. 2017) was used to align sequences from both datasets (ITS and ETS). Two obtained alignments were manually edited and concatenated in program PhyDe version 0.9971 (Müller et al. 2010). The combined dataset (ITS and ETS) comprises 1032 bp (593 in ITS and 439 in ETS alignment) and 19 taxa. We reconstructed the ITS plus ETS phylogeny of Atriplex sect. Obionopsis and two close relatives (A. flabellum, A. moneta) using the Maximum Likelihood approach (ML; Felsenstein 1973Felsenstein , 1983 and Bayesian Inference (BI; Rannala and Yang 1996). Gaps were treated as missing data. A variant of the General Time Reversal nucleotide substitution model (Tavaré 1986) (GTR + G+ I) was automatically selected by jModelTest v.2.0 (Darriba et al. 2012) for each partition (ITS and ETS) following the Akaike Information Criterion (AIC; Akaike 1974). For the ML analyses of concatenated alignment, we employed RAxML v.8 (Stamatakis 2014). ML Bootstrap analysis was conducted with 2500 replicates by the same program.
BI was performed in BEAST v.2.6.7 (Drummond et al. 2002;Bouckaert et al. 2014). Two runs with four chains each were run for 20 million generations for the combined dataset; both chains were sampled every 20.000 generations with a default parameter. Output log files were analysed using TRACER v.1.6  to assess all parameters' convergence and effective sample size (ESS). Ten percent of the samples were removed as burn-in. A maximum clade credibility tree was generated using TREE ANNOTATOR v.2.4.5 (Drummond and Rambaut 2007). Description. Monoecious subshrub up to 1.5 m high, branched in upper half; leaves alternate, shortly petiolate; petioles up to 1.0 cm long; blades greyish-silvery on both sides, 1.0-4.0 × 0.5-1.0 cm (much smaller towards inflorescence), oblong or narrowly oblong, entire or shallowly sinuate, with Kranz-anatomy; inflorescences branched, up to 15 cm long, with pseudopposite bracts or with a few small leaves forming pseudowhorls (after fruiting turning into small alternate leaves in younger shoots), aphyllous in other parts; glomerules condensed or slightly interrupted, of both male and female flowers, the latter are also located at the axils of uppermost leaves below the main inflorescence; male flowers stipitate at base, with 5 free perianth segments, anthers 0.25 mm long; bract-like cover of female flowers ( Fig. 2A, B) rhombic, entire or scarcely dentate, with or without lateral angles, smooth at the back or rarely with 1-2 very short outgrowths, valves connate to the halfway, sometimes to one third of their length, with indistinct veins, heteromorphic in some other characters: (1) bract-like cover of female flowers located in leaf axils ± indurated in lower half Table 2. Primers and cycling protocols.  Etymology. The specific epithet is chosen due to the resemblance of the new species to A. tatarica L., which also has long aphyllous inflorescences.

Marker
Conservation status. Although there is currently a limited number of collected specimens of Atriplex pseudotatarica, this species is clearly more overlooked than rare. Given that it is often found in disturbed habitats, produces a large number of seeds and is naturally adapted to saline substrates, we propose that the species should be assigned to the IUCN Red List category "Least Concern" (IUCN Standards and Petitions Committee 2022). Distribution (Fig. 3) Notes. All examined herbarium specimens of A. pseudotatarica are represented by upper twigs of the plants. They were mostly incorrectly identified as A. crassifolia, or rarely left without identification, as Atriplex sp. To date, the name A. crassifolia may be found misapplied to some other species attributable to different groups of the genus. Atriplex crassifolia is an annual C 3 -species belonging to A. sect. Teutliopsis Dumort. (Moser 1934;Iljin 1936;Sukhorukov 2006;Žerdoner Čalasan et al. 2022) with a restricted distribution range in semideserts of Kazakhstan and South-West Siberia, Russia (Iljin 1936;Sukhorukov 2006), penetrating into western China (Sukhorukov in Nobis et al. 2016). All other records of A. crassifolia reported from Europe are erroneous (Sukhorukov 2006;Sukhorukov et al. 2019). Aellen (1939), Ivanov (1989) and Medvedeva (1996) erroneously applied this name to A. patens (Litv.) Iljin, another species from A. sect. Teutliopsis (Sukhorukov 2006). The specimens from the Mediterranean area (GBIF Sekretariat 2022; re-identifications in BM!, K!, LE!) belong to A. tornabenei Tineo (C 4 -clade, A. sect. Obionopsis (Lange) Dumort.: Sukhorukov et al. 2022). The name A. crassifolia has also been widely used for the plants growing in lowlands of India and Pakistan (e.g., Hooker (1890), Bamber (1916)), and it is still erroneously applied in recent floras, checklists and ecological studies (Puri et al. 1964;Shetty and Singh 1991;Hussain and Mirza 1993;Jain et al. 2000;Kumar 2001;Paul 2012;Kumar and Singh 2013;Ibrahim 2019  Some of the plants growing in the lowlands of Pakistan also belong to A. pseudotatarica, of which some specimens were misidentified as A. leucoclada Boiss. Hedge (1997) noted that this species is highly polymorphic in the area covered in "Flora Iranica", with the extreme forms having smooth, apically trilobate bract-like cover. Unfortunately, he did not indicate where such specimens were collected, but such characters are typical of A. pseudotatarica.
Phylogenetic relationships (Fig. 4). Based on the combined nrITS and nrETS analyses, A. pseudotatarica was found sister to A. schugnanica, and both form a subclade within the Eurasian clade, A. sect. Obionopsis (Lange) Dumort., which comprises ~ 15 C 4 -species predominantly distributed in Irano-Turanian floristic region Žerdoner Čalasan et al. 2022). Atriplex pseudotatarica and A. schugnanica share some characters typical of the members of A. sect. Obionopsis (aphyllous inflorescences, sclerified bract-like cover with the valves connate up to the half of their length, presence of the female flowers in leaf axils and both female and male flowers in the inflorescence), but have several conspicuous morphological differences (Table 3). Additionally, A. pseudotatarica is distributed in the lowlands and pre-Himalayan foot- hills, whereas A. schugnanica is a typical montane plant growing in the West Himalayas, Karakoram and Pamir at the altitudes of (2000) 2600-4800 m a.s.l. (Iljin 1936;Sukhorukov et al. 2019). In Table 3, we also included other similarly looking C 4 Atriplex species; three of them (except A. tatarica) are present in the lowlands of Pakistan, and only one (A. pseudotatarica) is reaching India. Atriplex tatarica, widely distributed in many parts of the "Flora Iranica" area, as well as A. kalafganica Aellen (Aellen in Podlech 1975) are also added here because of their morphological resemblance with A. pseudotatarica.
The most conspicuous characters of A. pseudotatarica are subshrubby life form, pseudopposite leaves or leaves in pseudowhorls below the inflorescence seen at fruiting, and presence of black seeds.

Genus Atriplex in India
A recent revision of the genus in the Himalayan area (Jammu and Kashmir, Himachal Pradesh, Uttarakhand States) has been provided by Sukhorukov et al. (2019, with references therein), and some species (A. crassifolia, A. laciniata L., A. rosea L., A. sagittata Borkh. [previously known as A. nitens Schkuhr: Paul (2012)] were excluded from this area. The Chenopodiaceae of lowland India are still poorly studied and have not been included in any detailed morphological and chorological studies. Additionally, the plant material from India is old and quite scarce in the European herbaria. All these factors influenced a poor knowledge of some genera like Atriplex. Below we provide an improved taxonomic survey of Atriplex in the tropical part of India, with some notes on alien species of the genus.
Several alien species of Australian and North American origin were reported from India: A. amnicola Paul G.Wilson, A. nummularia Lindl., A. lentiformis (Torr.) S.Watson (Rani et al. 2013;Kumar et al. 2021). As stated by Singh (2005), many areas in India, especially influenced by a monsoon, are unfavourable for (semi)desert plants such as Atriplex. The Rajasthan State and some adjacent areas are of particular interest because they lie in the desert zone that is suitable for Atriplex species. We did not find any Atriplex specimens in Rajasthan, but several species were reported from this region including subshrubby American A. lentiformis (Gupta and Arya 1995), European A. hortensis (Bole and Pathak 1988) and two native species, subshrubby A. stocksii (Wight) Boiss. (Puri et al. 1964;Bole and Pathak 1988;Arya and Lohara 2016) and annual A. "crassifolia" (Puri et al. 1964;Shetty and Singh 1991). Atriplex crassifolia sensu Puri et al. (1964) was reported from the vicinity of Jodhpur and described in the diagnostic key as "an annual, male flower clusters axillary or in short leafy spikes", but elsewhere these authors provided a different diagnosis ("male flower clusters in slender leafless interrupted spikes"). We were unable to trace which species should be recognised instead of A. crassifolia because these contradictory diagnoses cannot be applied to any species. Shetty and Singh (1991) described it as an annual species with interrupted inflorescences and inflated bract-like covers. These two reproductive characters are also found in A. pseudotatarica, but the life form is different in the latter species. Nevertheless, we presume that A. pseudotatarica may be present in both Rajasthan and Gujarat due to the records in Haryana State. This species is also restricted to Jammu and Kashmir State (Sukhorukov et al. 2019). A varietal rank of this taxon cannot be accepted, because A. tatarica and A. pamirica occupy distant positions on the molecular tree (Žer-doner Čalasan et al. 2022) and belong to different sections, A. sect. Obionopsis (Lange) Dumort. and A. sect. Obione (Gaertn.) Reichenb., respectively ). Note. The name Obione nummularia Moq. was validly published on the basis of the only specimen (holotype) originating from India, J.P. Rottler s.n. collected in the late 18 th century and received by A. de Candolle under the name "Atriplex cristata Koenig" from M.N. Puerari (now at G-DC). Zhu et al. (2003) and Zhu and Sanderson (2017) reported the presence of A. repens on Hainan Island (southern China); however, the latter species is distributed in the coastal areas in southern India (Karnataka, Kerala, Tamil Nadu, and Andhra Pradesh) and Sri Lanka. The correct name for the plants growing in Hainan and other parts of tropical China as well as in southern Japan is A. maximowicziana Makino.  (1859) re-described the species as Atriplex stocksii based on two collections by Stocks from Pakistan and one from southern Iran, Aucher-Eloy 5268. In the protologue of A. stocksii, Boissier cited the type collection used by Wight but employed the same species epithet. For this reason, the protologue of his species name included the type of Wight's species but Boissier's species name cannot be treated as illegitimate. Boissier explicitly described his species as new and validly published its name in its own right, without any presumed basionym; for this reason, this species name has no basionym even though the potential basionym exists. As Boissier's species name is not superfluous, it cannot be automatically typified by the type of Wight's species name. Wight (1852) used a single specimen to describe his new species, now stored at K, which is, however, not the holotype due to the availability of another element, i.e. an illustration published in the protologue. Boissier (1859) used three specimens, hence lectotypification is also needed. Hedge (1997: 84) indicated that the type of A. stocksii Boiss. is the specimen kept at G-DC, which belongs to the same gathering as the type of O. stocksii Wight. The later type designation with the specimen at K, which was made by Omer (2001), has no standing.

Atriplex schugnanica
The species is reported from Gujarat and Tamil Nadu States (Rao 1986;Paul 2012), but in the latter state its presence is dubious. Also present in Rajasthan State.