Systematics of a widely distributed western North American springsnail, Pyrgulopsis micrococcus (Caenogastropoda, Hydrobiidae), with descriptions of three new congeners

Abstract We describe three new species of springsnails (genus Pyrgulopsis) from the Amargosa River basin, California and Nevada (P. licina sp. n., P. perforata sp. n., P. sanchezi sp. n.), each of which was previously considered to be part of P. micrococcus. We also restrict P. micrococcus to its type locality area (Oasis Valley) and redefine a regional congener, P. turbatrix, to include populations from the central Death Valley region and San Bernardino Mountains that had been previously identified as P. micrococcus. The five species treated herein form genetically distinct lineages that differ from each other by 4.2–12.6% for mtCOI and 5.2–13.6% for mtNDI (based on previously published and newly obtained data), and are diagnosable by shell and/or penial characters. The new molecular data presented herein confirm sympatry of P. licina and P. sanchezi in Ash Meadows (consistent with morphological evidence) and delineate an additional lineage of P. micrococcus (in the broad sense) that we do not treat taxonomically owing to the paucity of morphological material. Conservation measures are needed to ensure the long term persistence of populations of P. micrococcus and a genetically differentiated lineage of P. sanchezi which live in disturbed habitats on private lands.


introduction
The western North American hydrobiid gastropod genus Pyrgulopsis (commonly known as springsnails) is composed of 134 currently recognized species (Hershler and Liu 2012) which typically live in springs and have very narrow geographic ranges. This large radiation is characterized by a high degree of morphological conservatism and homoplasy, which has posed difficulties in delineating species limits and phylogenetic relationships (Hershler 1994, Liu andHershler 2005). Although the recent use of molecular tools has facilitated considerable progress in establishing a species level taxonomy that reflects the phylogenetic history of Pyrgulopsis (e.g., Hershler et al. 2003, Hershler and Liu 2004, Hershler and Liu 2009), a number of issues have yet to be addressed, including the unsettled status of several widely ranging congeners which have been shown to be composites of genetically divergent lineages (Hurt 2004. There is an urgent need to resolve these taxonomic problems to help identify conservation priorities for Pyrgulopsis, which is a current focus of attention of land managers owing to the increasing threats to its groundwater-dependent habitats (e.g., USFWS 2012a-c).
Pyrgulopsis micrococcus (Pilsbry in Stearns, 1893) was originally described based on shells from two localities in the Amargosa River basin and early treated as endemic to the upper portion of this watershed (Gregg andTaylor 1965, Taylor 1985: 317). This species was subsequently revised to include additional populations scattered within large portions of the Mojave Desert (southeastern California and southwestern Nevada) that resembled specimens from the type locality area in having a globose to ovate-conic shell and distally lobate penis with a terminal gland (sometimes reduced or absent) on the ventral surface (Hershler and Sada 1987, Hershler 1989, Hershler and Pratt 1990. A recent phylogenetic analysis resolved mtDNA sequences from 29 populations of P. micrococcus into five deeply divergent, allopatric clades-one of which also included morphologically similar and geographically proximate P. turbatrix Hershler, 1998-which were postulated to be distinct species (Liu et al. 2003, also see Hershler andLiu 2008). In this paper we detail previously unrecognized shell and penial differences supporting recognition of three of these lineages as new species which we describe herein while also clarifying the limits of P. micrococcus and P. turbatrix. We also present additional molecular data that confirm sympatry of two of these novelties at various sites in Ash Meadows (consistent with our morphological evidence) and delineate a new lineage of P. micrococcus (in the broad sense) in northern Death Valley that is not taxonomically treated owing to inadequate material.

Methods
The previous phylogeographic investigation ) was based on sampling across the entire broad range of P. micrococcus, including each of the drainage basins inhabited by this species. For the current study additional molecular sampling was done to confirm the apparent sympatry of two P. micrococcus lineages at various sites in Ash Meadows that was discovered during the course of this taxonomic study, and to evaluate the relationships of a distinctive morphotype (of P. micrococcus) in Grapevine Springs (northern Death Valley) that was not included in our previous analysis. Several previously analyzed populations-Grapevine Springs (M2), Purgatory Spring (M8), Tecopa Spring (M25), Shoshone Spring (M26)-were also additionally sampled to increase sample size and further evaluate their genetic distinctiveness. Newly collected material was preserved in 90% ethanol in the field. Genomic DNA was extracted from entire snails (2-10 specimens per sample) using a CTAB protocol (Bucklin 1992); each specimen was analyzed for mtDNA individually. LCO1490 and HCO2198 (Folmer et al. 1994) were used to amplify a 710 base pair (bp) fragment of COI, and ND43F and RND592F  were used to amplify a 550 bp fragment of NADH dehydrogenase subunit I (NDI). Amplification conditions and sequencing of amplified polymerase chain reaction product followed Liu et al. (2003). Sequences were determined for both strands and then edited and aligned using SeqUeNCHeRä version 5.0.1. The 51 newly sequenced specimens were analyzed together with our previously published P. micrococcus dataset Liu 2008). The new haplotypes from each sampling locality were deposited in GenBank (accession numbers KF559184-KF559202). Sample information and GenBank accession numbers are given in Appendix I. One example of each haplotype detected in a given sample was used in our analyses.
The partition homogeneity/incongruence length difference test (Farris et al. 1994) was used to determine whether the COI and NDI datasets were consistent and could be combined for the phylogenetic analysis. The test, which was conducted using parsimony-informative sites only and 1,000 replicates, indicated no significant incongruence (P=0.36) and consequently we combined the two datasets in our phylogenetic analysis. MRMODeLTeST 2.3 (Nylander 2004) was used to obtain an appropriate substitution model (using the Akaike Information Criterion) and parameter values for this analysis. Phylogenetic relationships were inferred by Bayesian analysis using MRBAyeS 3.1.2 (Huelsenbeck and Ronquist 2001). Metropolis-coupled Markov chain Monte Carlo simulations were run with four chains (using the model selected through MRMODeL-TeST) for 5,000,000 generations, and Markov chains were sampled at intervals of 10 generations to obtain 500,000 sample points. We used the default settings for the priors on topologies and the GTR + I + G model parameters selected by MRMODeLTeST as the best fit model. At the end of the analysis, the average standard deviation of split frequencies was less than 0.01 (0.0063) and the Potential Scale Reduction Factor (PSRF) was 1, indicating that the runs had reached convergence. The sampled trees with branch lengths were used to generate a 50% majority rule consensus tree with the first 25% of the samples removed to ensure that the chain sampled a stationary portion. Genetic distances (maximum composite likelihood) within and between species/lineages were calculated using MeGA5 (Tamura et al. 2011), with standard errors estimated by 1,000 bootstrap replications with pairwise deletion of missing data.
Types and other voucher material were deposited in the National Museum of Natural History (USNM) collection. Relevant material from the Academy of Natural Sciences of Philadelphia (ANSP), Bell Museum of Natural History (BellMNH) and the Santa Barbara Museum of Natural History (SBMNH) was also examined during the course of this study. Series of large adults (n=10) were used for shell measurements. Whorl counts refer to the entire shell. Sexual dimorphism in shells, which is occasionally observed in Pyrgulopsis (Taylor 1987), could not be quantified owing to small sample sizes. The total number of shell whorls was counted (WH) for each specimen; and the height and width of the entire shell (SH, SW), body whorl (HBW, WBW), and aperture (AH, AW) were measured from camera lucida outline drawings using a digitizing pad linked to a personal computer (see Hershler 1989). In addition, three ratios were generated from the raw data (SW/SH, HBW/SH, AH/SH). Descriptive statistics were generated using SySTAT FOR WINDOWS 11.00.01 (SSI 2004). Penial variation was described from series of adult specimens (typically n=30) that were relaxed with menthol crystals and fixed in dilute formalin prior to preservation in 70% ethanol. Descriptive penial terminology is from Taylor (1987) and Hershler (1994Hershler ( , 1998. Variation in the number of cusps on the radular teeth (n=5) was assessed using the method of Hershler et al. (2007).
We used a conservative, evolutionary lineage concept in describing new species only for those snails that are morphologically diagnosable as well as phylogenetically independent and substantially divergent genetically ). Inasmuch as the principal goal of our paper is to delimit species, we provide only brief taxonomic descriptions which focus on those aspects of morphology that have proven most useful in previous such studies of Pyrgulopsis (Taylor 1987, Hershler 1994, Hershler 1998.

Results
The alignment of COI and NDI sequences yielded 1188 bp. The five previously reported clades (A-e) were similarly recovered in the Bayesian analysis of this combined dataset (Fig. 1). The P. micrococcus morphotype in Grapevine Springs which was not included in our prior analysis formed an additional lineage (F) together with specimens from a spring in the Southern California coastal drainage. This clade is not formally treated herein owing to the paucity of morphological material. The additional molecular sampling conducted for this study also confirmed sympatry of morphologically distinctive clades C and e at three localities in Ash Meadows (M51-52, M53-54, M57-58) ( Fig. 1), providing additional support for recognizing these as separate species.
Clades A-F differed from each other by 4.2-12.6% for COI and 5.2-13.6% for NDI; variation within clades ranged from 0-2.5% for COI and 0-3.5% for NDI (Appendix II). The geographic distributions of these genetic lineages are shown in Fig. 2. Based on the genetic evidence of distinctiveness and diagnosable shell and/or penial characters (detailed below) we recognize three of these lineages as new species which are described below (clade B as P. perforata, clade C as P. licina, clade e as P. sanchezi), restrict P. micrococcus to its type locality area (Oasis Valley, clade A), and revise P. turbatrix to include populations from the central Death Valley region and San Bernardino Mountains that had been previously identified as P. micrococcus (clade D).    Diagnosis. A small congener (maximum shell height, 2.4 mm) having a narrowconic shell. Distinguished from similar regional species by its strongly curved penial filament and absence of glands on the penis. Further differentiated from frequently sympatric P. sanchezi (described below) by its highly convex, deeply incised teleoconch whorls and ovate shell aperture.
Operculum ( Fig. 3D-e) as for genus; edges of last 0.5 whorl frilled on outer side; muscle attachment margins variably thickened on inner side. Radula ( Fig. 3F-H) as for genus; dorsal edge of central teeth concave, lateral cusps four-six, basal cusp one. Lateral teeth having three-four cusps on both inner and outer sides. Inner marginal teeth with 20-25 cusps, outer marginal teeth with 24-31 cusps. Radula data are from USNM 850348. Penis ( Fig. 4A-B) medium-sized; filament medium length, narrow, weakly tapering, strongly curved (to outer side); lobe small, rectangular, horizontal or oblique; glands almost always absent (87/90 specimens), two specimens had a small, dot-like   Etymology. The epithet is an adjective derived from the New Latin licinus, meaning bent or turned upward, and refers to the distinctive shape of the penial filament in this species.
Remarks. The relationships of P. licina were not well resolved in the molecular phylogenetic analysis (Fig. 1). Haplotype variation within this clade was relatively small (Appendix II). Pyrgulopsis micrococcus. -Hershler 1989 (in part Diagnosis. A small to medium-sized congener (maximum shell height, 2.6 mm) having a broadly to ovate conic shell. Differentiated from similar regional species except P. micrococcus by its low-spired, broadly umbilicate shell. Differs from P. micrococcus in having a larger distal lobe and smaller gland on the penis.
Operculum (Fig. 6D-e) as for genus; outer side smooth; inner side smooth or weakly thickened along portions of the muscle attachment margin. Radula (Fig. 6F-H) as for genus; dorsal edge concave, lateral cusps four-eight, basal cusp one. Lateral  teeth having two-four cusps on inner sides and three-six cusps on outer sides. Inner marginal teeth with 14-24 cusps, outer marginal teeth with 18-31 cusps. Radula data are from USNM 857965. Penis (Fig. 4C-D) medium-sized; filament medium length, narrow, tapering, oblique; lobe medium-sized, rectangular, horizontal or slightly oblique; small (terminal) gland present on ventral edge of lobe (60/60 specimens), one specimen had an additional dot-like gland on the ventral surface of the lobe and one specimen had a similar glandular unit on the dorsal surface of the lobe. Penial data are from BellMNH 20891, USNM 883371. Distribution. Lower portion of Grapevine Canyon, and Grapevine Mountains, lower Amargosa River basin (M3, M4, M5, Fig. 2). The type locality (Fig. 5C-D) is the uppermost of a small series of springs to the east of Scotty's Castle.
Etymology. An adjective derived from the New Latin, perforare, meaning to pierce, and referring to the broad umbilicate shells of this species.
Remarks. The relationships of P. perforata were not well resolved in the molecular phylogenetic analysis (Fig. 1). Haplotype variation within this clade was relatively small (Appendix II). Diagnosis. A small to medium-sized congener (maximum shell height, 2.9 mm) having an ovate to narrow conic shell. Differentiated from similar regional species by its short, strongly tapering penial filament.
Operculum (Fig. 7e-F) as for genus; outer side smooth or with last 0.5 whorl weakly frilled; inner side smooth or slightly thickened along a small portion of the muscle attachment margin. Radula (Fig. 7G-I) as for genus; dorsal edge concave, lateral cusps three-six, basal cusp one. Lateral teeth having one-four cusps on inner sides and two-six cusps on outer sides. Inner marginal teeth with 10-26 cusps, outer marginal teeth with 12-33 cusps. Radula data are from BellMNH 21116, USNM 857963, USNM 883361, USNM 883365, USNM 883372. Penis (Fig. 4e-F) medium-sized; filament short, broad, strongly tapering, oblique; lobe short, rectangular, horizontal or slightly oblique; small (terminal) ovate gland almost always present on ventral surface of lobe (92/93 specimens), gland usually positioned horizontally, rarely borne on a raised swelling (one specimen), one specimen had a second, dot-like gland on the ventral surface of the lobe. Penial data are from BellMNH 2116, USNM 857963, USNM 857964, USNM 883361, USNM 883666.   (Fig. 2). The type locality (Fig. 5B) is a flowing well that was drilled into a small spring mound (Dudley and Larson 1976).
Etymology. This species is named for Peter G. Sanchez, who spearheaded early efforts to protect and conserve regional springsnails and their associated aquatic habitats while serving as a Resource Management Specialist in the Death Valley National Monument (now National Park) and Chair of the Desert Fishes Council (1978)(1979)(1980).
Remarks. Pyrgulopsis sanchezi was resolved as sister to P. arizonae (Gila River basin, Arizona) in the Bayesian analysis (Fig. 1). The five geographically separated groups of P. sanchezi populations are genetically differentiated-e.g., mean genetic distance is 1.5+/-0.3% (ranging from 1.3-2.3%) for COI and 2.1+/-0.6% (ranging from 1.8-3.2%) for NDI, however we have not found consistent morphological differences to support their recognition as distinct species.  Gregg and Taylor 1965: 109 (comb. n.). Pyrgulopsis micrococcus. -Hershler and Thompson 1987: 29 (new combination Revised diagnosis. A medium-sized congener (maximum shell height, 4.4 mm) having a broadly to elongate conic shell. Differentiated from similar regional species by the large size of the gland on the ventral surface of the penis.

Pyrgulopsis micrococcus
Description. Shell (Fig. 8A-D) broadly to narrow-conic, whorls 3.50-5.0. Teleoconch whorls weakly to strongly convex, sutures impressed. Aperture ovate, parietal lip complete, usually disjunct, last 0.25-0.5 whorl often loosened behind aperture, umbilicus small. Outer lip usually thin, orthocline. Sculpture of faint, irregular spiral striae. Operculum (Fig. 8e-F) as for genus; edges of last 0.5 whorl frilled on outer side; muscle attachment margins thickened on inner side. Radula (Fig. 8G-I) as for genus; dorsal edge of central radular teeth concave, lateral cusps five-eight, basal cusp one. Lateral teeth having three-four cusps on inner sides and four-five cusps on outer sides. Inner marginal teeth with 18-23 cusps, outer marginal teeth with 21-29 cusps. Radula data are from USNM 847246. Penis (Fig. 4G-H) medium-sized; filament short, narrow, tapering, slightly oblique; lobe small, tapering, horizontal; a large (terminal) gland (borne on a raised swelling) present on ventral surface of penis, extending from near mid-length almost to tip of lobe (90/90 specimens), one-two additional small glands sometimes present on ventral surface of lobe (8 specimens), one specimen had a glandular dot on the dorsal surface near the base of the filament. Penial data are from BellMNH 20663, BellMNH 20669, BellMNH 20744.
Remarks. Pilsbry (in Stearns 1893; also see Stearns 1901) listed a single "type" lot for P. micrococcus, [USNM] 123622, which is composed of six dry shells. Baker (1964) subsequently designated ANSP 67279a as the "type" without explaining his rationale for this action. ANSP 67279a (Fig. 8A) closely conforms to Pilsbry's description and figure and is also very similar to the USNM type material (Hershler and Sada 1987, fig. 8a). The labels associated with ANSP 67279a indicate that it was part of the original collection of P. micrococcus (made by C. Hart Merriam) and this lot was almost certainly known to Pilsbry, who was the curator of mollusks at the Academy of Natural Sciences during the time period when his description was prepared and published. Based on this evidence we conclude that ANSP 67279a is part of the type series and thus Baker's subsequent lectotype designation is valid.
Pyrgulopsis micrococcus was resolved in the Bayesian tree as sister to an undescribed species from the Amargosa Canyon, south of Tecopa (Fig. 1). Specimens assigned to P. micrococcus vary somewhat in size and shell shape, but are closely similar both genetically  and in penial morphology. Hershler, 1998 http://species-id.net/wiki/Pyrgulopsis_turbatrix Figs 4I-L, 9
Operculum ( Fig. 9e-F) as for genus; edges of last 0.5 whorl frilled on outer side; muscle attachment margins slightly thickened on inner side. Radula (Fig. 9G-I) as for genus; dorsal edge of central teeth moderately to highly concave, lateral cusps threeseven, basal cusps one-two. Lateral teeth having two-six cusps on inner sides and three-six cusps on outer sides. Inner marginal teeth with 14-31 cusps, outer marginal teeth with 15-33 cusps. Radula data are from USNM 857968, USNM 860450, USNM 860699, USNM 883373. Penis ( Fig. 4I-L) medium-sized; filament long, narrow, tapering, oblique; lobe medium-sized, tapering, slightly oblique; ventral surface of lobe having a small (terminal) gland (199/200 specimens) and rarely one or two additional glandular dots (3 specimens), dorsal surface sometimes having a small (penial) gland at base of filament (24/200 specimens) and rarely having an additional glandular dot (one specimen Remarks. The penial gland was not observed in >25% of the males in any of the seven samples that we studied and consequently has been removed from the diagnosis. The three geographically separated subunits of P. turbatrix are somewhat diverged genetically-e.g., mean sequence divergence is 0.9+/-0.2% (ranging from 1.1-1.5%) for COI and 0.9+/-0.2% (ranging from 1.1-1.3%) for NDI, but we have not found any consistent morphological differences among them.

Discussion
The three novelties described herein increase the number of Pyrgulopsis species in the Death Valley region to 17 (Hershler and Sada 2002) and add to the large body of evidence supporting recognition of this desert area as one of the most significant hotspots of rarity and richness in the United States (Chaplin et al. 2000). Our revision of P. micrococcus obviously is not yet complete as we have not treated clade F, which differs from the other lineages of P. micrococcus (in the broad sense) by 4.3-12.6% for COI and 7.4-13.3% for NDI (Appendix II). The two populations in clade F differ by only 0.8+/-0.3% for COI and 1.2+/-0.5% for NDI, suggesting that they may be conspecific. Additional studies are needed to clarify the taxonomy of this clade and to evaluate the biogeographic cause of its broadly disjunct distribution.
All of the new species described herein are endemic to the Amargosa River basin. Pyrgulopsis perforata and P. licina are distributed entirely within the confines of Death Valley National Park and Ash Meadows National Wildlife Refuge, respectively, and consequently are being afforded some measure of protection. The P. licina populations are also being monitored by The Nature Conservancy as part of their Nevada Springs Conservation Plan (Abele 2011). Three of the five genetically differentiated lineages of P. sanchezi are distributed in the land management areas mentioned above. The Tecopa lineage (M25) is distributed on private and public water resources lands and is being monitored as part of the Bureau of Land Management's Amargosa River Area of environmental Concern Implementation Plan (BLM 2006). The Shoshone lineage (M26) lives in a spring on private land that has a long history of disturbance which includes diversion of most of its flow and extensive modification of its outflow channel (USFWS 1984, Castleberry et al. 1990. This population appears to be restricted at present to a small area of leakage from a spring box (RH, HPL, CB, personal observation 15.XI.2011) and some measures will need to be taken to ensure its long term persistence. Our finding that P. micrococcus is restricted to its type locality area (Oasis Valley) suggests a need for additional conservation-related activities. The known populations of this species are in disturbed springs (per Maciolek 1983) on private land and some protection is needed to ensure their long term persistence. Surveys are also needed to evaluate the current status of several populations of this species that have not been sampled for several decades (e.g., that in Ute Springs) and to determine whether there may be previously unrecorded populations in Oasis Valley, which contains a large area of groundwater discharge (Reiner et al. 2002). Pyrgulopsis micrococcus (as currently constituted) was a candidate for addition to the Federal List of endangered and Threatened Wildlife and Plants (USFWS 1976) prior to its removal from this list in 1994 (USFWS 1994