Integrative taxonomic review of the genus Peschetius (Coleoptera, Dytiscidae, Hydroporinae) from India with description of two new species

The diving beetle genus Peschetius Guignot, 1942 (Coleoptera: Dytiscidae) in India is reviewed. Integrative taxonomic approach using morphology, multivariate morphometry and genetic analysis of cytochrome oxidase subunit 1 revealed the presence of four species, two of which are described here as new: Peschetius bistroemi sp. nov. from southern Western Ghats (Kerala) differs from all known congeners with distinctly broadened male antennomeres IV and V, shape of the prosternal process and the male genitalia; P. nilssoni sp. nov. from northern Western Ghats, Rajasthan and Madhya Pradesh is similar to the widespread Indian P. toxophorus Guignot, 1942, from which it differs in habitus, elytral colour pattern and the shape of the male genitalia. New records are presented for the remaining Indian species, namely P. quadricostatus (Aubé, 1838) and P. toxophorus. All species are diagnosed, illustrated and a key to their identification is provided.


Introduction
The dytiscid genus Peschetius Guignot, 1942 includes ten species, out of which seven occur in Africa-south of the Sahara, and three in Asia (Bistrӧm and Nilsson 2003; Bis trӧm and Bergsten 2015; Nilsson and Hájek 2021). The genus is represented by three species from the Indian sub continent: Peschetius toxophorus  is en demic to peninsular India while P. quadricostatus (Aubé, 1838) is widely distributed in India, and is also known from south-eastern Iran, Pakistan and Nepal (Hájek 2006;Ghosh and Nilsson 2012). The third species, P. taproba- nicus Biström and Bergsten, 2015, is endemic to Sri Lan ka (Biström and Bergsten 2015). Peschetius was proposed by Guignot (1935) to accom modate three previously described aberrant Hydroporus; later,  formally made the genus name available by designating P. nodieri (Régimbart, 1895) as its type species. Due to the aberrant morphology, the ge nus Peschetius was traditionally included in the tribe Hy droporini of the eponymous subfamily. Guignot (1935) and  both suggested its affinity with Australian genera Antiporus Sharp, 1882 and Necterosoma Macleay, 1871, currently classified within the subtribe Sternopriscina. However, Miller et al. (2006) re vised the classification of Hydroporinae and proposed the genus Peschetius as a sister group to the members of the tribe Bidessini, chiefly based on presence of a prominent spermathecal spine, and the five-lobed teeth in the pro ventriculus. Therefore, the authors formally transferred Peschetius to the tribe Bidessini which is now widely ac cepted (see, e.g., Miller and Bergsten 2014).
African Peschetius were reviewed by Omer-Cooper (1970), while the two Indian species were diagnosed by Vazirani (1970a). Vazirani (1977c) discussed elytral pattern variability of P. toxophorus. The comprehensive, morphology-based, revision of the genus by Bistrӧm and Nilsson (2003) provided detailed species diagnoses and the first cladistic analysis of the genus.
While studying the systematics and morphology of dy tiscid beetles from India, particularly Western Ghats, we have discovered four morphologically distinct species of Peschetius, for which species limits were also confirmed by a genetic analysis of mitochondrial cytochrome oxi dase subunit 1 and by the analysis of morphometric data in an integrative way. The importance of combining tradi tional taxonomy and modern tools like DNA sequencing to unveil cryptic species has been currently highlighted e.g. by Dayrat (2005), Will (2005), Padial et al. (2010) and Schlick-Steiner et al. (2010). The delimitation of taxa us ing such an integrative approach, including the description of two new species is the main aim of the present paper.

Study area
India is a major part of the Indian subcontinent which is flanked by the Himalayan mountains in the north, Arabi an Sea in the west, Indian Ocean in the south and Bay of Bengal in the east (Fig. 1). The country has several phys ical features, such as Himalayas, Indo-Gangetic plains, central and eastern highlands, Thar desert, Gondwanan peninsular plateau, Western and Eastern Ghats and coast al plains. The Satpura range of mountains lies north of the peninsular plateau and forms a chief biogeographical barrier. The plateau gradually slopes down in the north via Madhya Pradesh to Indo-Gangetic plains in Uttar Pradesh, and in the northwest to Thar desert of Rajasthan.
The western edge of this plateau is bordered by a chain of escarpments i.e. Western Ghats or Sahyadri range. The range passes through Gujarat, Maharashtra, Karnataka, Tamil Nadu and Kerala States of the country. The Ghats are interrupted by two biogeographic barriers, namely Palghat Gap and Shencottah Gap near Kerala (Fig. 1). The Palghat gap is flanked by Nilgiri range to the north and Annamalai hills to the south. The position of Hima layan orogen and afore-mentioned physical features of the Indian plate play a key role for the tropical monsoon as well as various climatic zones in India (Mani 1974).

Taxon sampling and specimen deposition
The beetles were captured using a pond net of mesh size 1 mm (EFE and GB Nets, Educational field equip ment UK Limited; now https://www.nhbs.com/telescop icpondnet) from the Western Ghats ( Fig. 1) and were preserved in absolute ethanol. The alcohol was changed in laboratory and specimens were stored at -20°C for mo lecular work (Table 1). This material is deposited in the following collections:

Morphological study and illustrations
Measurements were taken with an ocular micrometre.
The following abbreviations were used in the descrip tions: TL -total length of body, a single measurement of length from front of head to apex of elytra; TL-h -total length without head length, length of body from anterior margin of pronotum to apex of elytra; MW -maximum width of body. Miller and Nilsson (2003) was followed for the terminology to denote the orientation of the gen italia. Digital images of habitus and male genitalia were pre pared as described by Sheth et al. (2021). Additionally, the specimens were studied under Nikon SMZ800 and photographed under Nikon SMZ25 and Nikon SMZ1270, both with NIS elements D software (version 5.01.00 and version 5.20.00, respectively; Nikon Corporation; https:// www.nikon.com). For the study of female genitalia, the female specimens were treated using 10% KOH for 24 hours. The spermathecae were dissected out in a water drop under Nikon SMZ800 and photographed in glycer ine jelly using Olympus BX3+Olympus DP3+ Olympus U-CMAD3 T7 assembly with CellSens dimension soft ware (version 1.16; Olympus Corporation; https://www. olympus-lifescience.com/en). The photographs were stack ed using Helicon-Focus software (version 5.1.19; He licon Software Limited; https://www.heliconsoft.com). The photographs of habitats of new species were captured using Google Pixel phone (model 3a; Appendix 1).

Morphometry and morphometric analysis
Fifteen morphological characters were measured using a Lawrence and Mayo stereo zoom microscope fitted with an ocular micrometre for 58 adult beetles. The abbrevia tions and full names of characters are as follows ( Fig. 2; see also Ribera and Nilsson 1995): TL-h -body length, MW -maximum width, HLlength from clypeal border to posterior side between eyes, HW -maximum width across eyes, PL -median length of pronotum, PW -maximum width of pronotum, DW -distance between level of maximum width to tip of elytra, DM -distance between end of metacoxae to tip of elytra, FL -length of metafemur, FW -width of metafe mur, BL -length of metatibia, RL -length of metatarsus, EH -maximum length of elytra; lateral, MH -maximum height of body; lateral, and DH -distance between level of maximum height to tip of elytra.
Between-group Principal Component Analysis (bgPCA) on raw morphometric data was performed. To account for scale difference among characters, bgPCA on correlation matrix was performed. Since in bgPCA, the eigenanaly sis is carried out on the group means (Krzanowski 1979), it extracts fewer principal components that explain most of the variation in the data; as a result, low dimensional PCA plot is reliable for understanding most of the vari ation in high dimensional multivariate data. Because bg PCA can suffer from certain limitations (Cardini et al. 2019), the significant differences between groups were independently tested using Permutations Multivariate Analysis of Variance (PERMANOVA) (Anderson 2001). PERMANOVA tests the null hypothesis that the cen troids and dispersion of the groups are equivalent for all groups. PERMANOVA was performed using Euclidian distance and 9999 permutations. Overall PERMANOVA was performed to check whether at least one of the group centroids was different. If overall PERMANOVA was significant, then significant differences between pairs of groups were tested using pairwise PERMANOVA. Since multiple tests were performed on the same data, family wise error rate was controlled using sequential Bonfer roni correction. All statistical analysis was performed in the software PAST (version 4.02; Hammer et al. 2001).

Molecular analysis
The DNA was extracted from whole individuals using a QIAamp DNA Mini Kit (Qiagen, Hilden, Germany, Cat alog No. 51306) following the manufacturer's protocol. Partial sequence of mitochondrial cytochrome oxidase subunit 1 (cox1) was amplified using the primer pair Jerry (5'-CAA CAT TTA TTT TGA TTT TTT GG-3') and M70 (5'-TTC ATT GCA CTA ATC TGC CAT ATT A-3') with an annealing temperature of 57°C (Simon et al. 1994;Lunt et al. 1996). PCR amplification, PCR product pu rification and sequencing protocols were done according to Suranse et al. (2017). Molecular sequence data gener ated for the present work are deposited in the GenBank database. Please refer to Table 1 for details of sequences generated in this study and other sequences obtained from the GenBank database. Amarodytes sp. (KF575474) was used as an outgroup following its sister taxa relationship provided by Miller et al. (2013). Sequences were aligned in MEGA (version 7; Kumar et al. 2016) using MUSCLE (Edgar 2004). Pairwise raw genetic distances were estimated using MEGA 7 (Kumar et al. 2016). Data were partitioned by the three codon po sitions of the cox1 gene. Partition analysis (Chernomor et al. 2016) and ModelFinder (Kalyaanamoorthy et al. 2017) were used to find the optimal partitioning scheme with the best-fitting nucleotide substitution model for each partition selected by the minimum Bayesian Infor mation Criterion (BIC) (Schwarz 1978). A maximum likelihood (ML) analysis was conducted using IQ-TREE (version 1. 6.12;Nguyen et al. 2015) on the partitioned dataset using the proposed models with topological sup port inferred by 1000 iterations of ultrafast bootstrapping (Hoang et al. 2018). The resulting phylogenetic tree was edited in FigTree (version 1.4.2; Rambaut 2009).
We performed genetic species delimitation using two methods. Assemble species by automatic partitioning (ASAP) delimits species based on genetic gap analysis (Puillandre et al. 2021). ASAP was performed online (https://bioinfo.mnhn.fr/abi/public/asap/) using genetic uncorrected p distances. General mixed Yule-Coalescent (GMYC) method is a likelihood-based method for de limiting species by fitting within-and between-species branching models to reconstructed gene trees (Fujisawa and Barraclough 2013). GMYC was performed online (https://species.hits.org/gmyc/) using single threshold and ultrametric Bayesian tree as an input. The Bayes ian ultrametric tree was generated using Markov Chain Monte Carlo (MCMC) analysis implemented in BEAST v1.8.4 (Drummond et al. 2012) with strict clock and two runs of 10 million generations (sampling trees every 1,000 generations and first 10% trees were discarded as burnin).

Morphometric analysis
Between-group PCA extracted three components which explained all the variation in the data. The specimens of Indian Peschetius grouped under four separate clusters in the PCA (Fig. 3A). The null hypothesis that all the clus ters were the same was rejected (PERMANOVA, F = 33.93, p = .0001) (Fig. 3B), suggesting that the centroid of at least one of the clusters was significantly different. Pairwise comparison of clusters revealed that all the clus ters were significantly different from each other, even af ter sequential Bonferroni correction (Fig. 3C). Thus, this analysis clearly indicated that there are four Peschetius morphospecies occurring in India.

Molecular analysis
ModelFinder identified two partitions, one comprising the combined first and second codon positions, and other comprising third codon position of cox1 gene. Nucleotide substitution models for the partitions were TIM2+F+R2 and HKY+F+G4, respectively. The maximum likelihood analysis placed the specimens of Indian Peschetius into four well-supported clades (Fig. 4). Peschetius nilssoni sp. nov. was recovered convincingly as the sister species to P. toxophorus (ultrafast bootstrap support, UFB = 96). Peschetius bistroemi was recovered as being more dis tantly related to the other Indian species which are to gether placed in a clade, albeit with weak support (UFB = 68). Maximum intra-species raw genetic distance among Indian Peschetius species was 1.0 % while the minimum inter-species genetic divergence was 2.7 % (Table 2). Peschetius nilssoni sp. nov. differed from all its conge ners, for which the genetic data are available, with a raw genetic distance of 2.7-14.3%, while P. bistroemi sp. nov. differed from other congeners with a raw genetic distance of 12.7-14.3%.
Both species delimitation methods, ASAP and GMYC, indicated four distict species of Peschetius from Indian subcontinent (Fig. 4). The best partition of ASAP had the highest relative gap width metric W of 0.00372 and threshold distance of 1.8% and identified Peschetius nils-

Taxonomy
Peschetius    . 6Aa). Parameres with long setae in apical half, apex rounded (Fig. 6Ab). FEMALE. As male but antennomeres simple, not modi fied. Pro-and mesotarsi slender or less broadened. Sper matheca as in (Fig. 8A)-spermathecal spine long and slender. With nearly a flat prosternal process (i.e. without trans verse depression or longitudinal keel), the new species is similar and probably related to P. taprobanicus from Sri Lanka; however, it differs from the latter species in the shape of male genitalia: the apex of median lobe is not bent as in P. taprobanicus and the curvature of the median lobe of P. bistroemi sp. nov. is shallower. The parameres of P. bistroemi sp. nov. are abruptly narrowed and rounded at their apex while in P. taprobanicus those are gradually narrowed. The setae of the parameres are present in the apical half in P. bistroemi sp. nov. while in P. taprobanicus these are restricted to the apex. Further, the spermathecal spine in P. bistroemi sp. nov. is longer than the other three Indian species, and not curved like that in P. nilssoni sp. nov.

Measurements (N=10
Etymology. The species is named in the honour of Prof. Olof Bistrӧm (Helsinki, Finland) for his significant con tribution to the taxonomy of Dytiscidae, including the ge nus Peschetius. The name is a noun in the genitive case.
Collection circumstances. The specimens were found in slow flowing streamlets with rock and mud as substra tum, and decaying leaves.
Distribution. The species is so far known only from three close localities in Kottayam district, Kerala, southwestern India. The specimens listed in other material agree well with the type material of P. nilssoni but in absence of the male, we prefer not to designate them as paratypes.

Peschetius nilssoni
Description of male holotype. Habitus: Body elongate, oblong oval, widest before midlength of elytra; outline discontinuous with distinct angle between pronotum and elytra; elytral keels prominent; dorsal surface submatt (Fig. 5B). -Colouration: Head black except testaceous occipital part posterior to eyes, appendages testaceous. Pronotum testaceous with bilobed black band near pos terior margin extending to posterior corners. Elytron blackish with typical testaceous markings consisting of two subbasal spots, two premedian spots, postmedian transverse band and preapical spot. Ventral side overall ferruginous. Prosternum darker along anterior margin, prosternal process with black border. Posterior margins of abdominal ventrites dark. Legs testaceous. -Head: transverse, eyes slightly emarginate. Antennae with all antennomeres slender, club-shaped. Width across eyes 1.8X the width between eyes. Clypeus arcuate. Labrum deeply emarginate with series of setae on anterior mar gin. Punctation of head dense, distance between punc tures smaller than puncture diameter. Punctures fine on clypeus, becoming progressively coarser posteriorly on frons, occipital part posterior to eyes impunctate. Setif erous punctures present in well-developed fronto-clypeal depressions and as a row along inner margin of eyes. Reticulation consisting of polygonal, slightly transverse meshes present on clypeus and in anterior part on frons; posterior part of frons smooth. Impunctate occipital part coarsely microreticulate. -Pronotum: Transverse. An terior margin straight, sides evenly rounded, posterior margin gently sinuate; anterior corners acute, posterior angles obtuse. Pronotum with distinct depressions be tween disc and sides, mediolaterally between disc and posterior margin. Pronotal disc strongly vaulted. Punc tation dense, distances between punctures smaller than puncture diameter. Punctures setiferous, finer on disc, becoming coarser on margin and sides. Surface between punctures microreticulate with shallowly impressed, polygonal meshes. -Elytra: Widest before midlength, keels prominent. Punctation dense, distance between punctures approximately equal to puncture diameter. Punctures finer along suture, costae and lateral margin, coarser on disc. Surface between punctures microretic ulate, reticulation similar to that of pronotum. -Legs: Tibiae club-shaped, dorsally with long natatorial setae; pro-and mesotarsi broadened, dorsally with long nata torial setae, ventrally with adhesive setae; metatarsi with long natatorial setae on both sides. -Ventral side: Pros ternum sinuate on anterior margin, area between procox ae narrowed. Prosternal process elongate, flat basally, laterally compressed posteriorly, convex with short keel apically, apex tuberculate, posteriorly narrowed, without transverse depression (Fig. 7B). Mesoventrite bifurcated on anterior margin, posterior margin rounded. Metaven trite densely punctate with coarse punctures, distance between puncture approximately equal to puncture diam eter. Surface microreticulate with shallowly impressed polygonal meshes. Anterior border of metaventrite with two prominent depressions below mesocoxae. Metacoxal plate with punctation and reticulation similar to that of metaventrite. Metacoxal lines raised. Abdomen with five ventrites (V1 to V5); V1 with 8-10 macropunctures in one row and V2 with 8-10 macropunctures in two rows on either side; punctures on V2 prominent while those on V3 to V5 shallow; V3 to V5 with lateral depression shallow; V3 longitudinally obtusely keeled; reticulation of V2 to V5 consists of polygonal meshes. Punctures on ventral surface setiferous.-Male genitalia: Medi an lobe broad at base, narrowed towards apex, evenly curved or 'C' shaped, and with basal process (Fig. 6Ba). Parameres with extended setae in apical half, apex round ed (Fig. 6Bb). FEMALE. Identical to male in habitus, dorsal surface reticulation more impressed, thus beetles appearing matt. Apex of prosternal process non-tuberculate. Pro-and mesotarsi less broadened. Spermatheca as in (Fig. 8B)spermathecal spine curved.
Variability. The species slightly varies in body size and width. The shape of sub-basal yellow spot on elytra varies within species. Differential diagnosis. With the black head, and the pros ternal process convex with a short apical keel, and the general shape of the male genitalia, Peschetius nilssoni sp. nov. is very similar and undoubtedly closely related to P. toxophorus. This fact is confirmed also by the raw genetic distance as measured by the cox1 gene, which is 2.7-3.3%-the least differentiated within Indian Peschetius. The two species can be easily recognised based on the shape of the testaceous premedian transverse band on elytra, which is always interrupted between elytral costae in P. nilssoni sp. nov. forming lateral longitudinal spot (Fig. 5Ba) and discal transverse spot (Fig. 5Bb) while the band is always uninterrupted in P. toxophorus (Fig. 5Da). Additionally, the body shape of P. nilssoni sp. nov. is more elongate and narrower (Fig. 5B), while it is broader in P. toxophorus (Fig. 5D). These differences in body shape were also confirmed with the multivariate morphometric analysis (Fig. 3). Further, the median lobe of P. nilssoni sp. nov. is gently and evenly curved (Fig. 6Ba), while that of P. toxophorus is more strongly and unevenly curved (Fig. 6Da). Parameres are gradually narrowing to their apex in P. nilssoni sp. nov., (Fig. 6Bbx) but they are dis tinctly tapered subapically in P. toxophorus (Fig. 6Dbx). Finally, the spermathecal spine in P. nilssoni sp. nov. (Fig.  8Ba) is curved unlike compared to other Indian species.
Etymology. The new species is dedicated to Dr. Anders N. Nilsson (Mullsjö, Sweden) for his immense contri bution to aquatic Coleoptera. The name is a noun in the genitive case.
Collection circumstances. The species was collected in ponds with mud and rock as substratum. It was frequent ly found sympatrically with P. quadricostatus and some times with P. toxophorus.
Distribution. The distribution of the new species is con fined so far to north-western, central and western India, namely Rajasthan, Madhya Pradesh and Maharashtra States. Some of the previous records of P. toxophorus, especially those from northern half of India, may actual ly also represent P. nilssoni sp. nov. and their revision is necessary.

Figs 5C, 6C, 8C
Redescription. Total length 3.10-3.45 mm and maximum width 1.65-1.85 mm (N = 25). See also Supplementary file 4. -Head ferruginous with two dark fronto-lateral spots (Fig. 5C). Pronotum ferruginous with bilobed black band along posterior margin and medial black streak along anterior margin. Elytron blackish with typical testa ceous markings consisting of two subbasal spots, preme dian and postmedian transverse bands and preapical spot. Punctation of head dense, distance between punctures smaller than puncture diameter. Punctures fine on cly peus, become progressively larger posteriorly on frons, occipital part posterior to eyes impunctate. Setiferous punctures present in shallow but distinct fronto-clypeal depressions and as a row along inner margin of eyes. Reticulation consisting of shallowly impressed, polygo nal meshes on clypeus; posterior part of frons smooth. Impunctate occipital part posterior to eyes coarsely mi croreticulate. Pronotal disc with posterior depression less prominent but clearly distinguishable. Punctation dense, distances between punctures smaller than puncture diam eter. Punctures setiferous, finer on disc, becoming coarser on margin and sides. Surface between punctures microre ticulate with shallowly impressed, polygonal meshes. El ytra broadest at midlength, keels prominent. Punctation dense, distance between punctures approximately equal to puncture diameter. Punctures finer along suture, cos tae and lateral margin, coarser on disc. Surface between punctures microreticulate with well impressed polygonal meshes. Prosternal process elongate, narrowed at apex, apically keeled. Abdomen with five ventrites (V1 to V5); V1 with 6-9 macropunctures in one row while V2 with 3-5 macropunctures on each side, arranged randomly in two rows. Punctures on V2 to V5 setiferous; lateral de pression on V3 to V5 prominent; reticulation of V2 to V5 consists of polygonal meshes. Median lobe of aedeagus gradually curved, tapering apically, apex pointed; with a basal process (Fig. 6Ca). Parameres with short setae in apical half, apex blunt, inner margin not sinuate (Fig.  6Cb). Spermatheca as in (Fig. 8C)-spermathecal spine straight, short and broad.
Collection circumstances. The species was found in pools, ponds, tanks, reservoirs and slow flowing streams, frequently with P. nilssoni sp. nov. This species was also found in the same habitat as P. bistroemi sp. nov in Aim combu, Kerala.

Peschetius toxophorus Guignot, 1942
Collection circumstances. The species was found in habiting pools, ponds, tanks, reservoirs and slow flowing streams. In northern Maharashtra, the species was some times found sympatrically with P. nilssoni sp. nov.

Discussion
Peschetius bistroemi sp. nov. from Kerala is rather unique as it is the only known member of the genus with broad ened male antennomeres; its weakly supported distant placement compared to other Indian species is most like ly due to insufficient sampling. The diagnostically dis tinct prosternal processes of P. bistroemi sp. nov. and Sri Lankan endemic P. taprobanicus are similar, indicating a possible close relationship between these two species. However, P. bistroemi sp. nov. differs from the latter in the shape of its male genitalia. Therefore, more work in cluding a better sampling of African and the Sri Lankan species, and multigene phylogeny is definitely necessary to clarify the position of P. bistroemi sp. nov. Moreover, based on the preliminary data, P. bistroemi sp. nov. is described from the region between geologically ancient Palghat and Shencottah gaps in the Western Ghats (Fig.  1). Various studies have reported the role of these gaps as biogeographical barriers leading to genetic variation and speciation in the case of flora, and fauna including both vertebrates and invertebrates (e.g. Vidya et al. 2005;Ba hulikar et al. 2006;Joshi and Karanth 2013;Anoop et al. 2018). Additionally, many endemic and threatened fresh water fishes of Kerala, for example Travancoria elongata Pethiyagoda & Kottelat, 1994 are known to inhabit the River Chalakudy (Raghavan et al. 2008) that originates south of the Palghat Gap (Arunachalam 2000). There fore, extensive sampling of water beetles over a wide geographical range together with the afore-mentioned barriers is needed to understand biogeography of P. bistroemi sp. nov. On the other hand, the second newly described spe cies, Peschetius nilssoni sp. nov. is without any doubt closely related to P. toxophorus. Interestingly, at the be ginning of the 20 th Century, French specialist Maurice Régimbart correctly recognised two Peschetius mor phospecies with dark head within the material in BMNH and labelled them as two new species. However, he did not describe them, and Balfour-Browne (1946) mixed both taxa under his P. andrewesi. Subsequently, Vazira ni (1977c) mentioned the differences in elytral pattern of 'two forms of P. toxophorus' and predicted the presence of another undescribed Peschetius species in the Western Ghats. Yet, the species remained unrecognised for anoth er 40 years, until the present integrative approach of mor phological study, morphometry and molecular analysis confirmed its status and enabled us to describe the new species. Despite being found sympatrically, we did not encounter any specimen showing intermediate characters between P. quadricostatus or P. toxophorus and this new species. Further, molecular analysis has shown that the inter-specific genetic distance between P. nilssoni sp. nov. and its sister taxa P. toxophorus, is comparatively smaller (2.7-3.3%) than inter-specific genetic distances between the other Peschetius species studied here, and for the oth er species for which molecular data are available. Both the genetic methods of species delimitation, ASAP and GMYC, clearly identified P. nilssoni sp. nov. and P. toxophorus as distinct, reciprocally monophyletic species.
Low genetic distances among species have been previ ously reported for several insect taxa, for example, 2.2% inter-species divergence has been observed in certain Australian insects (Pons et al. 2006). The known genetic distance within Coleoptera using cox1 ranges from 2.0 to 4.0% (Hendrich et al. 2010;Ribera et al. 2010;Abellán et al. 2012). Similarly, in the predaceous diving beetle genus Antiporus, the known genetic divergence between species ranges from 3.5 to 6.6% (Hawlitschek et al. 2011). A low genetic divergence may suggest relatively recent specia tion event between the two species.
Our integrative taxonomic approach towards under standing the diversity of aquatic beetles not only unveiled two new species of Peschetius but also provided interest ing insights, albeit preliminary, into the ecology and evo lution of these species. Our study suggests that such an approach can provide better understanding of diversity of invertebrate taxa in the Western Ghats. Both the species of Peschetius described in this work belong to the West ern Ghats-Sri Lanka biodiversity hotspot (Myers et al. 2000), reemphasizing its importance as high biodiversity reserve also with respect to its invertebrate fauna. While insects play a vital role in ecosystem functioning, these have often been neglected compared to vertebrate taxa (Goulson 2019). Diversity of the invertebrate fauna in the Western Ghats is riddled with Linnean shortfall (Brown and Lomolino 1998) owing to limited taxonomic studies in this region. Despite the presence of unique habitats in the Western Ghats the studies on its invertebrate fauna are limited (Myers et al. 2000). Further, Short (2018) em phasized the need of thorough inventory work on water beetles with the possibility of discovery of novel species from southeast Asia including India. Given that Linnean shortfall compromises biodiversity knowledge essential for evolutionary, ecological and conservation research (Hortal et al. 2015) overcoming the shortfall is essential (Bini et al. 2006). Freshwater ecosystems are among the most threatened habitats in the anthropocene and dedi cated efforts to their conservation are essential (Dudgeon 2019).

Conclusion
The combined approach of morphology, geometric mor phometry and molecular analysis revealed the presence of four Peschetius species in India; two species from Western Ghats biodiversity hotspot are described as new to science. While one of those species was collected only recently, the second was known but remained unrec ognised for more than 100 years. Therefore, the integra tive taxonomic approach is considered important for the study of the biodiversity.

Authors' contributions
SDS performed fieldwork, museum study, morphological and molecular work, and data analysis. JH, HVG and SDS studied and identified mate rial. SDS and JH contributed to preparation of illustrations. ND verified and analyzed data; helped planning a part of the fieldwork; supervised both laboratory work and data analysis of SDS. HVG provided inputs for data analyses. JH corrected, revised and discussed the data.
All authors collaborated in the development of the research prob lem identified by HVG, discussed results and contributed for the man uscript preparation.

Competing interests
The authors have declared that no competing interests exist.