Shards, sequences, and shorelines: two new species of Bembidion from North America (Coleoptera, Carabidae)

Abstract Two new species of Bembidion are described from river shores in North America. One, Bembidion mimbressp. nov., from the Gila River watershed in the lands of the Mimbres culture in New Mexico and Arizona, is closely related to the widespread Bembidion levigatum. DNA sequences from several linkage groups and morphology provide evidence of the distinctiveness of B. mimbres. The second, Bembidion corgenomasp. nov., has been the subject of recent genomic and transcriptomic studies. It belongs in the Bembidion transversale subgroup, and occurs from California north to British Columbia, east to Montana and Nevada. The B. transversale subgroup as a whole is reviewed, and morphological characters that distinguish B. corgenoma from the similar and sympatric B. transversale and B. erosum are described and illustrated. DNA sequences of these three species show no consistent differences in 28S, COI, CAD, and Topoisomerase, and a coalescent species delimitation analysis reveals no notable structure within the complex. This is the first known trio of species within Bembidion for which those genes provide no clear signal of species boundaries. A neotype is designated for the one name in the group that lacks a primary type, Bembidium haplogonum Chaudoir. Chromosomes of the new species and their relatives are as is typical for Bembidion, with eleven pairs of autosomes and an XY/XX sex chromosome system.


Introduction
Bembidion is a very large genus of small beetles with more than 1,200 species worldwide (Lorenz 2005). Most species of these small predators live along the edges of bodies of water and can be abundant in their habitats.
In the course of an ongoing project revising the bembidiine carabids of America north of Mexico, a number of undescribed species have been discovered. Most of these will be described in due course within complete revisions of subgenera or species groups. However, two of these new species are or will be soon discussed in the scientific literature, and warrant description more quickly, in order to provide them with names. These two are also especially significant, as they have cultural connections to humans, implicit or explicit, of very different sorts.
One of them is a member of the subgenus Hydrium, a group of relatively large Bembidion that is widespread in the Northern Hemisphere. The new species (Fig. 1) is only known from the Gila River watershed of southeastern Arizona and southwestern New Mexico, where it lives along the banks of rivers and creeks, on the ground a few meters away from the shoreline (Fig. 2), most commonly under willows (Salix). The distribution of this new species is within that of the Mimbres culture, which flourished in that area one thousand years ago. This culture is perhaps best known for black-onwhite Mimbres pottery, the designs of which depicted people, cultural icons, and organisms (Hegmon et al. 2018). The people of the Mimbres culture were deeply aware of the arthropods in their environment, as indicated by the astonishing array of images on their pottery of insects, including among others geometrid larvae (https://core.tdar. org/image/383483/1452-style-iii-bowl-from-cameron-creek; https://doi.org/10.6067/ XCV8Z60P2N), Orthoptera (https://core.tdar.org/image/383111/2685-style-iii-bowlfrom-swarts; https://doi.org/10.6067/XCV80Z7364), dragonflies, and ant lions (Hegmon, et al. 2018). In honor of these peoples who were so connected to the small organisms in their midst, this elegant beetle species is given the name Bembidion mimbres.
The second species described here is connected to humans via modern biological research: it is becoming the first model species of Bembidion for genomic and transcriptomic studies. It is member of the B. transversale species group of the Ocydromus complex of Bembidion, containing some of the largest Bembidion in North America (Lindroth 1963;Maddison and Swanson 2010). The B. transversale group consists of two subgroups, the B. transversale subgroup and the B. mexicanum subgroup (Maddison 2012;Maddison and Swanson 2010). Maddison and Swanson (2010) considered the B. transversale subgroup to contain three species (B. transversale Dejean, B. perspicuum (LeConte), and B. sarpedon Casey), although they noted that "B. transversale" showed enough morphological variation to suggest that it may contain multiple species. Because of the especially complex pattern of variation, it only recently become clear that the genomic and transcriptomic model species was undescribed. The new species (Fig. 3) is common in Oregon and California, with some populations in neighboring regions, living along cobble and gravel shores of rivers and creeks (Fig. 4). It is the best sequenced Bembidion genomically and transcriptomically (Gustafson et al. 2019;Gustafson et al. 2020; Pflug   et al. 2020), and has been used as one of the models for developing a UCE probe set for adephagan beetles (Gustafson, et al. 2019;as B. haplogonum Chaudoir). It is a centerpiece of ongoing and future studies of genome size in carabids (e.g., Pflug, et al. 2020). In this paper, it is described as Bembidion corgenoma.
An important step enabling future research about these beetles is providing the species with stable names. Although we now know the two new species in many ways unimaginable to those who lived a thousand years ago, including detailed aspects of their  Casey, B. vandykei Blaisdell, B. curtulatum Casey, B. platynoides Hayward, and B incrematum LeConte. DNA and genomes, we know very little about these two species in nature. With the decreasing emphasis on natural history in modern biology, it is possible that a person of the Mimbres culture knew aspects of the daily life of B. mimbres much better than we ever will. However, perhaps this paper, in giving names to the two species and presenting identification tools allowing them and their near relatives to be distinguished, will inspire research about these beetles, including into their way of life along river shores.

Materials and methods
Members of Bembidion were examined from or will be deposited in the collections listed below. Each collection's listing begins with the code used in the text.

Morphological methods
General methods of specimen preparation for morphological work, and terms used, follow Maddison (1993;2008). Genitalia were prepared, after dissection from the body, by treatment in 10% KOH at 65 °C for 10 minutes followed by a series of multi-hour baths of distilled water, 5% glacial acetic acid, distilled water, and then ethanol. Male genitalia were then mounted in Euparal between two small coverslips attached to archival-quality heavyweight watercolor paper, and, once dried, pinned beneath the specimen. Male genitalia were examined for 30-60 specimens each of B. transversale, B. erosum, and B. corgenoma, and four or five specimens each of B. levigatum and B. mimbres.
Photographs of entire beetles and antennae were taken with a Leica M165C dissecting scope and a Sony NEX-7 camera, and of male genitalia with either a Leica Z6 Apo lens and DMC4500 camera or a Leica DM5500B compound microscope and DMC425C camera. A stack of images from different focal positions was then merged using the PMax procedure in Zerene Systems's Zerene Stacker; the final images thus potentially have some artifacts caused by the merging algorithm. Measurements were made using Leica Application Suite v4.9 from images acquired using these imaging systems.

Cytogenetic methods
Twenty-two males were examined for chromosome number and sex-chromosome system. Methods used were as outlined by Maddison (1985;2008 Maddison (1985) were re-examined and re-identified.

DNA extraction and sequencing
Genes studied, and abbreviations used in this paper, are: 28S: 28S ribosomal DNA (D1-D3 domains); 18S: 18S ribosomal DNA; COI: cytochrome c oxidase I; wg: wingless; CAD: carbamoyl phosphate synthetase domain of the rudimentary gene; ArgK: arginine kinase; Topo: topoisomerase I. DNA was extracted using a Qiagen DNeasy Blood and Tissue Kit. Fragments for the seven genes were amplified using the polymerase chain reaction (PCR) on an Eppendorf Mastercycler ProS Thermal Cycler, using TaKaRa Ex Taq and the basic protocols recommended by the manufacturers. Primers and details of the cycling reactions used are given in Maddison (2012) and Maddison and Cooper (2014). The amplified products were then cleaned, quantified, and sequenced at the University of Arizona's Genomic and Technology Core Facility using a 3730 XL Applied Biosystems automatic sequencer. Assembly of multiple chromatograms for each gene fragment and initial base calls were made with Phred (Green and Ewing 2002) and Phrap (Green 1999) as orchestrated by Mesquite's Chromaseq package (Maddison and Maddison 2018a, c) with subsequent modifications by Chromaseq and manual inspection. Multiple peaks at a single position in multiple reads were coded using IUPAC ambiguity codes.

Taxon sampling for DNA studies
For the phylogenetic study of Bembidion (Hydrium), 19 specimens of the subgenus Hydrium, as well as five species serving as outgroups (Table 1) were used. Of the 152 sequences examined, 97 were newly acquired, with 55 being from previous publications (Maddison 2012;Maddison and Cooper 2014;Maddison and Maruyama 2019;Maddison and Ober 2011). For the phylogenetic study of the B. transversale group, I sampled 56 specimens of the B. transversale subgroup, as well as three species of the B. mexicanum subgroup (Table 2). Of the 237 sequences examined, 109 were newly sequenced, with 128 from previous publications (Kanda et al. 2015;Maddison 2012;Maddison and Swanson 2010;Wild and Maddison 2008). Sequences of the two holotypes listed in Tables 1 and 2 are "genseq-1", of paratypes "genseq-2", and the remainder are all "genseq-4" (Chakrabarty et al. 2013). In addition, sequences of the newly designated neotype of Bembidium haplogonum Chaudoir were acquired and deposited in GenBank (accession numbers MW151478, MW151506, and MW151563), and are "genseq-1". Localities of the sequenced specimens of B. levigatum, B. mimbres, B. transversale, B. erosum, and B. corgenoma are given in Table 3.

Sequence alignment
Alignment was not difficult for any of the protein-coding genes. There were no insertion or deletions (indels) evident in the sampled CAD, ArgK, Topo, wg, or COI sequences. An alignment of 28S was conducted in MAFFT version 7.130b (Katoh and Standley 2013), using the L-INS-i search option and otherwise default parameter values.

Molecular phylogenetic analysis
Maximum likelihood analysis was conducted for each gene individually using IQ-TREE version 1.6.7.1 (Nguyen et al. 2015), as orchestrated by Mesquite's Zephyr package (Maddison and Maddison 2018b, c). The ModelFinder feature within IQ-TREE Table 1. Sampling of members of Bembidion (Hydrium) and related subgenera for DNA-based study. Four-digit numbers under "#" are D.R. Maddison DNA voucher numbers; the specimen marked with an * is the holotype of B. mimbres.
All other specimens listed of B. mimbres are paratypes. An abbreviation for state or province of capture is given under "Loc"; further information on B. levigatum and B. mimbres specimens is given in Table 3. The last eight columns contain GenBank accession numbers.  Table 2. Sampling of members of Bembidion transversale group for DNA-based study. Four-digit numbers under "#" are D.R. Maddison DNA voucher numbers. Under "T", the holotype of B. corgenoma is indicated by "H", and paratypes by "P". An abbreviation for state or province of capture is given under "Loc"; further information on specimens of B. transversale, B. erosum, and B. corgenoma is given in   (Kalyaanamoorthy et al. 2017) was used to find the optimal character evolution models. The MFP model option was used for 28S, and the TESTMERGE option for the proteincoding genes. The TESTMERGE option sought the optimal partition of sites, beginning with the codon positions in different parts. Twenty searches were conducted for the maximum-likelihood tree for each matrix. For the B. transversale group, a multi-species coalescent approach was conducted with the 28S, COI, CAD, and Topo data to provide an algorithmic analysis of species boundaries. STACEY version 1.2.5 (Jones 2017) was used as implemented in BEAST version 2.6.2 (Bouckaert et al. 2014), with the epsilon value set to 1*10-4, CollapseWeight parameters to 0.5 and 10, and with a Beta prior. I evaluated sampling sufficiency using ESS values in Tracer version 1.7.1 (Rambaut et al. 2018); after four independent runs of 1E9 generations each, all ESS values exceeded 200, except for mutationRate.s:Topo, whose ESS value was 191. As I saved trees every 100,000 generations, with the first 10% of the trees discarded as the burn-in period, this yielded a sample of 72,000 trees.

Data resources
Sequences have been deposited in GenBank with accession numbers MW151366-MW151563. Aligned data for each specimen as well as files containing inferred trees for each gene are available in Supplementary material S1 and S2, and have been deposited in the Dryad Digital Repository, https://doi.org/10.5061/dryad.18931zcw1.

Molecular results
In the analysis of DNA data for the subgenus Hydrium, B. levigatum and B. mimbres sp. nov. differed in all genes except COI (Fig. 5), providing evidence that they are two separate species. In all gene trees except 28S, B. nitidum was the sister group of B. levigatum +B. mimbres (Suppl. materials 1).
The majority of specimens of B. levigatum and B. mimbres were indistinguishable in COI, but there were four sequences of B. mimbres that formed a separate clade (for  Topo specimens 1220, 1267, as well as the second sequences of 1944 and 2117). These four sequences have 29 sites at which they differ from all other sampled B. levigatum + B. mimbres, at 20 of which these four sequences have the same base as in at least one other sampled Hydrium species. These four sequences might be nuclear copies or numts (Thalmann et al. 2004), or they could represent the effects of Wolbachia infections (Smith et al. 2012). Although it is possible these are the true mitochondrial copies of COI, and that the other sequences are numts, the evidence points to the four unusual sequences being something other than true mitochondrial copies: the chromatograms for these four unusual sequences have several double peaks, indicating polymorphism within the PCR products for non-synonymous differences. These four sequences have been deposited in GenBank as "COI-like" sequences. In each of the four genes studied in the B. transversale group, the maximum likelihood tree showed a monophyletic B. transversale subgroup (Suppl. materials 2), with B. perspicuum and B. transversale s. l. (= B. transversale + B. erosum + B. corgenoma) forming a clade, the sister of which is B. sarpedon. None of the three species within B. transversale s. l. form a clade in any of the four genes studied (Fig. 6). The multispecies coalescent STACEY tree also showed no distinction between these species based upon the combined analysis of 28S+COI+CAD+Topo (Fig. 7).

Cytogenetic results
All males examined are inferred to have 22 autosomes (i.e., 11 pairs of autosomes) and an XY/XX sex chromosome system (Table 4).
Re-examination of voucher specimens identified as B. transversale in Maddison (1985) showed that they belong to three species. The specimens from Alberta and near Fernie, BC, are B. transversale; the specimen from near Cache Creek, BC, is B. corgenoma; the specimen from Salmon Valley, BC, is B. erosum. The specimens called "B. sp nr transversale-Nr 1" are B. pernotum, and the specimens called "B. sp nr transversale-Nr 2" are B. lugubre. The specimens reported as belonging to "B. sp.nr. transversale" in Pflug et al. (2020) are all B. corgenoma. These new identifications are incorporated into the summary shown in Table 4.

Morphological results
Morphological results for Bembidion (Hydrium) are presented in the taxonomic section below.
Members of the B. transversale subgroup are very similar morphologically. DNA sequence data of 28S, COI, CAD, and Topo do not reveal any consistent phylogenetic structure within B. transversale s. l. (Figs 6, 7), suggesting that it is perhaps a single species. In all other bembidiines investigated to date (e.g., Maddison 2008;Maddison and Cooper 2014;Maddison and Sproul 2020;Sproul and Maddison 2017), every form judged by morphological evidence as a distinct species is revealed as a clade in the tree of at least one of these four genes. I detected no variation in chromosomes within the group (Table 4); although Bembidion species typically have similar chromosomes (Maddison 1985), some subgenera have species that differ by chromosome number (Maddison 2008). However, in spite of the lack of genetic evidence supporting multiple species, the morphological results reveal that this complex consists of at least three distinct but very similar species.
Examination of primary types (documented in the Taxonomic Treatment section, below) indicates that two of the species have names (Bembidion transversale and B. erosum), and the third is described here as B. corgenoma; these names will be used in advance of the Taxonomic Treatment to simplify the text.
The morphological evidence indicating that there are three species includes color (Figs 11A-C, 12), mentum shape ( Fig. 13), and male genitalic structure (Figs 14-17), with the patterns summarized below and graphically in Fig. 18.
The dark and pale western species (B. erosum and B. corgenoma) are broadly sympatric from southern California through British Columbia (Figs 19,20), and are found microsympatrically (on the same gravel bank) at Wilson Creek, Del Norte County, California (41.6051°N, 124.1005°W), as well as along Floras Creek, Curry County, Oregon (42.9132°N, 124.4251°W), and the Siletz River E of Kernville, Oregon (44.8720°N, 123.9223°W). They differ consistently in size of a sclerotized lobe of the internal sac ( Fig. 15B vs. C); the thickness of the tip of the flagellar sheath, with B. corgenoma having a somewhat triangular sclerotized region of the tip (arrow in Fig. 16E), as opposed to a thin dark line in B. erosum (Fig. 16C, D); and color, with B. erosum being generally darker (Fig. 11B) than B. corgenoma (Fig. 11C), especially the appendages (Fig. 12B vs C). In addition, most males of B. erosum have the ventral surface of the apex of the aedeagus more strongly curved downward (Fig. 14C, D). Although there are no universal distinctions between the two species in the genes sequenced, the six B. erosum and five B. corgenoma sequenced from the gravel bank shown in Fig. 4B consistently differ in one base in Topoisomerase, suggesting, combined with differences in genitalia and color, that there is no or extremely limited gene flow at that locality.
The ranges of the two generally paler species (B. transversale and B. corgenoma) overlap in Nevada, Washington, Idaho, and Montana (Fig. 19), and there are three localities Table 4. Chromosome numbers and sex chromosomes of Bembidion (Hydrium) and Bembidion transversale group males. The Sample column indicates the total number of specimens examined in this paper and in previous papers. "1" in Reference indicates Maddison (1985); "2" indicates Pflug et al. (2020).     at which they co-occur (16 mi W Lolo Pass, Ravalli Co., Montana; Walla Walla, Washington; Spokane, Washington; all in OSAC). There are subtle but consistent differences in the male genitalia, as well as striking differences in the mentum, and I am convinced any gene exchange in the region of overlap is minimal. In the overlap region there are a very few specimens of B. transversale with paler antennae (similar to typical B. corgenoma), and there is one population at Hayden Lake, Idaho (CAS), which contains B. transversale typical in all regards except for one specimen that has the anterior margin of the mentum somewhat intermediate between the two species (Fig. 13B). With these minor exceptions, the differences in mentum, genitalia, and color are consistent throughout the overlap range among the many males whose genitalia were dissected.
Bembidion erosum and B. transversale are the two most similar species within the trio, differing most notably in the anterior lateral region of the mentum: in B. erosum this region is large and triangular, similar to that standard in Bembidion (i.e., like those shown in Fig. 13A), in contrast to the modified mentum of B. transversale (Fig. 13B-D), in which the anterior lateral region is much reduced. The male genitalia are very similar, both having a larger lobe on the basal sclerite (Fig. 15A, B), and a thin, non-triangular apex to the flagellar sheath ( Fig. 16A-D). However, the flagellar sheath is more dorso-ventrally compressed in B. transversale (Fig. 16A, B), and the flagellar complex is thinner (arrow in Fig. 17A, B). With one exception, the known ranges of B. erosum and B. transversale do not overlap, with B. erosum in the United States being restricted to the Cascades and west, and B. transversale only known from east of the Cascades (Figs 19, 20). The one exception is in eastern Washington: there is a series of seven specimens of B. erosum labeled "Colbert, Wash / V 30 1937 / Dan Bonnell". The two males from this series have typical B. erosum genitalia, and the mentum has the anterior lateral region large and triangular. Colbert is less than 20 km NNE of Spokane, where both B. transversale and B. corgenoma have been collected. As the only record of B. erosum east of the Cascades, some doubt is cast upon its authenticity, but there is no reason otherwise to question the locality data.
In addition, there is a form in the Sierra Nevada of California that requires further examination. My limited study of it indicates that it has all of the morphological features of B. transversale, except that it has a normal mentum shape, with large and triangular anterior lateral regions.
The species key in Lindroth (1963) can be modified as follows to take into account the new species.

19
More than two setae on the clypeus, and at least one long seta on the front angle of the prothorax. Elytra with a row of long setae on all intervals; prothorax wide, sides very rounded (Fig. 8A)  Elytra with a row of long setae on most intervals, but lacking on intervals 2 and 4. Prothorax narrower, with straighter sides (Fig. 8B)   Type locality. USA: New Mexico: Grant Co., Gila River, Billings Vista, 32.8163°N, 108.6032°W.
Derivation of specific epithet. Bembidion mimbres is named in honor of the people of the Mimbres culture, who lived alongside this species, including at the type locality, and who illustrated the insects in their world on their pottery (Hegmon, et al. 2018). The name is to be treated as a noun in apposition.
Diagnosis and description. Adults of this species are relatively large Bembidion (5.3-6.3 mm in length), with a striking appearance because of the smooth and shiny dorsal surface with its metallic reflections (Figs 1, 8B). Body piceous, with an aeneous, green, or blue metallic reflection. Legs with tarsi and tibiae testaceous, femora infuscated. First three antennomeres testaceous, with the tip of the third infuscated in some specimens; fourth basally testaceous. Palps testaceous except for the penultimate maxillary article, which is infuscated. Mentum with anterior lateral regions large and triangular as typical for a Bembidion; mentum tooth incised at tip, and thus bifid (similar to B. levigatum). Prothorax with relatively straight sides, especially posteriorly (Fig. 8B); with distinct posterolateral carina close to the lateral margin; posterior region of pronotum smooth, impunctate. Lateral bead of elytra extending inside shoulder well toward the midline, sharply angulate. Striae consisting of rows of distinct punctures, without an associated groove; on stria 1 complete; striae 2-6 absent in the posterior third; stria 7 virtually absent, represented by at most minute punctulae. Dorsal surface lacking microsculpture, and thus very shiny. More than two setae on the clypeus, and extra setae on the frons, at least anteriorly; at least one long seta near the front angle of the prothorax; prosternum with at most two setae. Elytra without the typical pair of distinguishable setose punctures associated with interval 3, but with a row of long setae on all intervals except 2 and 4. Aedeagus (Fig. 9C,D) with ventral margin thinner, and with internal sac sclerites very similar to those of B. levigatum, but with slight differences, especially basally. Most easily distinguished from B. levigatum by the narrower prothorax with straighter sides (Fig. 8B), the lack of setae on elytral intervals 2 and 4, and having at most one or two setae on the prosternum.
Additional characteristics. Diploid chromosome number 24, with 11 pairs of autosomes and an XY/XX sex chromosome system. Geographic distribution. Known from the Gila River watershed in Arizona and New Mexico (Fig. 10).
Habitat. At the type locality, found at night on damp sandy soil about 2-4 meters from the river shore, in the shade of a large willow (Fig. 2). At a site a few meters away, 28 specimens were found at night on damp clay/sand soil among small Salix and Populus saplings 2-8 meters from the water's edge; in spite of extensive searching in the same area, only one specimen was found during the day. At other sites along the Gila River, found in habitats similar to those in which B. levigatum is found, on damp sand and silt on the steep upper bank of the river, mostly at night.

Bembidion levigatum Say
Bembidium levigatum Say, 1823: 84. Lectotype female in MNHN, designated by Lindroth and Freitag (1969). Type locality Missouri. Bembidion laevigatum delawarense Casey, 1924: 24. Holotype male in USNM (type number 36814), examined. Type locality Pennsylvania. Diagnosis and geographic distribution. Adults of this species are large and distinctively wide, with a broad, rounded prothorax (Fig. 8A), and with a very shiny dorsal surface, with a green or bluish metallic reflection. The clypeus, frons, anterior corner of the prothorax, and all elytral intervals have long setae in addition to the standard set in Bembidion; the prosternum has four or more setae. Aedeagus as in Fig. 9A

Bembidion transversale species group
The Bembidion transversale group contains large Bembidion found primarily on river shores of cobble, gravel, and sand from Canada to Guatemala. Members of the group are characterized by large size (5.8-8 mm); posterolateral carina of pronotum lacking or indistinct and somewhat oblique; lateral bead of elytra not prolonged onto shoulder; crista clavicularis absent; elytral striae distinct and mostly complete; elytral microsculpture transverse; two discal setae of elytra in or close to third stria.  Casey, 1918 There is a total of 20 species-group names that have been applied to members of the B. transversale group (for details beyond those provided below, see Maddison and Swanson (2010)). I have examined detailed photographs of the primary type of one (the holotype of Bembidium transversale Dejean, in the MNHN), and the primary types themselves of 18. The twentieth lacked a type series, and a neotype is designated below.
The species key in Lindroth (1963) can be modified as follows to take into account species in the B. transversale subgroup. Specimens from this group are not easy to identify.
Diagnosis and geographic distribution. Adults of this species (Fig. 11A) are characterized by the reduced anterior lateral regions of the mentum (Fig. 13B-D), and the antenna with at least the second and third antennomeres apically infuscated (Fig. 12A). The prothorax is more cordate than in other species. The aedeagus has its ventral surface relatively straight, with the tip not abruptly curved downward (Fig. 14A, B); the basal sclerotized lobe is large (Fig. 15A), and the apex of the flagellar sheath has a long, thin dark line (Fig. 16A, B). This is the easternmost species, found from Newfoundland and Nova Scotia west through Ontario to southeastern British Columbia, central Oregon, northeastern Nevada, northern Utah, and Colorado (western portion of distribution shown in Fig. 19).

Bembidion erosum (Motschulsky, 1850)
Peryphus erosus Motschulsky, 1850:10. Lectotype female, in ZMUM, examined, designated by Bousquet and Larochelle (1993) Diagnosis and geographic distribution. Most adults of this species are the darkest members of this group (Fig. 11B), with the second and third antennomeres infuscated (Fig. 12B), at least apically, and with dark femora, although the more northern populations (e.g., from mainland British Columbia) have paler legs and paler ground color of the body. Prothorax moderately cordate, with a smooth basal region with few punctures. Tip of aedeagus abruptly curved downward (Fig. 14C, D); basal sclerotized lobe large (Fig.  15B); apex of flagellar sheath with long, thin dark line (Fig. 16C, D). This species is coastal, occurring from Haida Gwaii in British Columbia south along the coast to Big Sur and neighboring areas of California, with only one record from east of the Cascades (Fig. 20).  Derivation of specific epithet. The specific epithet is formed from the Latin word cor, meaning heart, and genoma, a modification (for easier pronunciation) of the coined word "genome". Corgenoma refers to this species being the heart or current focus of genomic studies in small carabid beetles. Cor-also alludes to the type locality of Corvallis, Oregon, whose name is derived from Latin, and means "heart of the valley". It is to be treated as a noun in apposition.
Diagnosis and description. ). Relatively light in color compared to B. transversale and B. erosum, with legs and basal three antennomeres pale, testaceous or rufo-testaceous. Head and prothorax piceous, with metallic reflections, on pronotum green or aeneous, on head bluish or green. Elytra paler, with shoulders and most of the anterior half testaceous with an orange tint, bordered posteriad by a dark band (with intervals 1-3 in this region dark rufous), followed by a pale testaceous band that either extends to the apex or that is bounded posteriad by dark lateral spots which in the darkest individuals merge in the middle. Mentum with anterior lateral region as typical for a Bembidion: triangular, large, and with anterior margin significantly anteriad of the central tooth (Fig. 13A); central tooth trapezoidal, rounded. Prothorax cordate, with more sinuate margins than B. perspicuum, with a relatively smooth basal region, with minute punctures; pronotum without or with very faint, oblique posterolateral carina as in other members of this group. Elytral striae 1-5 complete; stria 6 distinct and strong through much of its length; stria 7 shallower, less impressed than 6, but distinct. Microsculpture of elytra very transverse, with little tendency to form meshes. Two discal setae on each elytron, close to third stria. Tip of aedeagus not abruptly curved downward (Fig. 14E, F); basal sclerotized lobe small (Fig. 15C); apex of flagellar sheath with dark area triangular (Fig. 16E, F).
Geographic distribution. This species occurs from central British Columbia south through the Willamette Valley of Oregon, the Central Valley of California, with some records in Nevada, Idaho, eastern Washington, and Montana (Fig. 19), thus overlapping with the range of B. transversale.
Geographic variation. The specimens on or close to the beaches of the Pacific Ocean appear on average slightly paler than more inland specimens.   Habitat. This species occurs on gravel or cobble shores of the rivers and creeks (Fig. 4), more often where the bank is relatively flat and has small amount of clay and silt mixed in with sand and gravel under the rocks. They also can be common under cobbles on the shores of small creeks on the upper portions of beaches of the Pacific Ocean.

Bembidion perspicuum (LeConte, 1848)
Ochthedromus perspicuus LeConte, 1848: 466. Holotype male, in MCZ (type number 5510), external structure and aedeagus examined. Type locality "Rocky Mountains".  Lindroth (1975), in USNM (type number 36916), examined. Type locality Jemez Springs, New Mexico (Lindroth 1975). Bembidion excursum Casey, 1918: 59. Holotype female, in USNM (type number 36915), examined. Type locality Tucson, Arizona. Bembidion tuolumne Casey, 1924:30. Lectotype male, designated by Lindroth (1975), in USNM (type number 36917), external structure and aedeagus examined. In Maddison and Swanson (2010), this was treated as tentatively a synonym of B. transversale. Further examination of the lectotype, including of its genitalia, reveal that this is a specimen of B. perspicuum. Type locality Tuolumne, California. Chaudoir's (1868) description of Bembidium haplogonum is detailed enough to make determination of the species he was describing clear. In his description, the large size (8 mm) and absence of a carina near the hind angle of the prothorax could only apply (within California, the type locality) to a member of the B. transversale group. The color pattern of the elytra ("sur les élytres, une petite tache au milieu du bord antérieur de chacune, et une bande transversale un peu arquée aux trois quarts, d'un jaune testacé pâle, peu distinctes, surtout la tache basale", which translates to "on the elytra, a small spot on the front edge of each, and a slightly curved transverse band at three-quarters, of a pale testaceous yellow, indistinct, especially the basal spot") can only apply to some specimens of B. erosum, B. lugubre, or the northern, dark form of B. perspicuum (Fig. 11E), as the other species in California have the entire basal third to half of the elytra pale. The description of the appendages as having the first 3.5 antennomeres, palps, and legs all pale yellow eliminates B. erosum, as the palps, femur, and antennomeres 2-11 are dark in California specimens of that species. I have seen no specimens of B. lugubre with an isolated small spot on the front edge of each elytron; the only specimens that have a pale area in the basal half of the elytra have the entire sides and basal regions of the elytra a pale rufous,   any specimens I have seen from California, and in only very few individuals elsewhere. The flatter prothorax with less rounded sides and a wider lateral gutter is also distinctive of B. perspicuum relative to all three other species from California (B. erosum, B. lugubre, and B. corgenoma), as is the distinctly punctured pronotal base ("distinctement ponctué tout le long de la base"). The large size (8 mm) is more characteristic of B. perspicuum; I have seen no specimens of B. corgenoma longer than 7.5 mm, but have seen specimens of B. perspicuum that are 7.9 mm in length, and specimens of B. perspicuum are, in general, larger than those of B. corgenoma. Finally, the virtual absence of the seventh elytral stria is characteristic of B. perspicuum relative to B. corgenoma. As Lorquin travelled extensively in areas where the dark form of B. perspicuum occurs (Grinnell 1904), it is certainly reasonable that a specimen of that form could have been seen by Chaudoir. My early interpretations of Chaudoir's descriptions were in error, and led me to believe that Chaudoir's specimen was a member of what I here call Bembidion corgenoma; that mistake led me to call the species studied in Gustafson et al. (2019) Bembidion haplogonum. As a correct reading of the original description shows that Bembidium haplogonum refers to the dark form of B. perspicuum, I have designated a specimen from northern California with a color pattern matching Chaudoir's description as the neotype (shown in Fig. 11E).

Designation of a neotype for
Diagnosis and geographic distribution. Adults of this species are large, and have a pronotum that is flatter than in other members of the group, with less rounded sides, and with the basal region more evidently punctate (Maddison and Swanson 2010: fig. 4). At least antennomeres 4-11 infuscated. Specimens from most areas are relatively pale (Fig. 11D), with the front half of the elytra pale, but specimens from northern California and Oregon are much darker (Fig. 11E), with only elytral apices being pale. Aedeagal characterс are described in Maddison and Swanson (2010). This species is known from Texas, Kansas, Colorado, New Mexico, Arizona, Utah, Nevada, California, and Oregon.
Diagnosis and geographic distribution. Adults of this species (Fig. 11F) are the palest members of this group, with legs entirely testaceous or rufo-testaceous, with antennae gradually becoming slightly darker toward the apex, and pronotum in most specimens dark rufous as opposed to the piceous or black of other species. The dorsal surface is shinier than in other species, especially the pronotum, because of the nearly effaced microsculpture. The prothorax is moderately cordate; the elytral intervals are flatter than in related species, with small punctures in the striae. The ventral margin of the aedeagus has a slight downward bulge, and the internal sac sclerite complex of male genitalia is narrow in lateral view, with a long and thin flagellar complex. Known from New Mexico and Colorado west to Arizona and Utah, and north to Mammoth Hot Springs, Wyoming (OSAC).

Concluding remarks
The pathways that led to the recognition of the two species described here were very different. When I encountered Bembidion mimbres for the first time, as pinned specimens at the University of Alberta's Strickland Museum in 1981-1982, I immediately recognized them as an undescribed species. They shared the large size, setose elytra, shiny surface, and striking color of the distinctive Bembidion levigatum, but did not share B. levigatum's unusual prothorax shape and width. In contrast, it took at least 12 years of study for me to become confident that B. corgenoma was a new species, and that the B. transversale subgroup consisted of at least five species (B. sarpedon, B. perspicuum, B. transversale, B. erosum, and B. corgenoma). The distinctiveness of B. sarpedon and B. perspicuum was recognized many years ago. The specimens that remained (B. transversale s. l.), however, were so complex in their variation patterns, so lacking in a differentiating signal in DNA sequences, and with such similar genitalia, that at times I thought there was just one species in Bembidion transversale s. l., and at other times more.
I had become so accustomed to the clarity provided by DNA sequences in my other taxonomic projects on bembidiines that I had become somewhat skeptical of the value of traditional taxonomic methods utilizing only patterns of morphological variation. Two events changed my mind, as they caused the patterns to become evident at last. The first event was Kip Will's collecting of both dark and light specimens from the shores of Wilson Creek in north coastal California. They were so obviously different in color that I expected them to have clearly different genitalia, and different DNA sequences. My cursory inspection revealed only the slightest difference in the overall shape of the aedeagus (I had not yet noticed the differences in the structures of the internal sac), not notable enough to be significant in itself. In addition, all six sequenced dark specimens from that gravel bar differed from all five sequenced light specimens in one base in Topoisomerase, but they did not differ in 28S, COI, and CAD. The correlation between color, aedeagal shape, and that single base in Topoisomerase convinced me that there were likely two species at that site in northern California, although if so they would be much more similar than are most other pairs of closely related, sympatric Bembidion species. Examination of Motschulsky's specimens eventually revealed that the dark species had a name, Bembidion erosum, but the pale species at Wilson Creek and elsewhere continued to trouble me: other than the normal mentum, I saw no consistent differences from the more eastern Bembidion transversale. The distinctiveness of the pale western form (here called B. corgenoma) did not become evident until the basic morphological work was done: thorough examination of the genitalia of 63 B. corgenoma males and 33 B. transversale males, focused on the area of geographic overlap, revealed the consistent differences shown in Fig. 18, especially the shape of the basal sclerotized lobe (Fig. 18D) and the tip of the flagellar sheath (Fig. 18E). This confirmed that sequences of four genes will not necessarily reveal the presence or absence of gene flow, and that even in Bembidion, a group in which DNA sequences often work very well for species delimitation, species boundaries are sometimes more quickly uncovered by traditional morphological methods.
The lack of observed differentiation in DNA sequences between Bembidion transversale, B. erosum, and B. corgenoma suggests that these are young, recently differentiated species. The contrast is striking between this trio and other bembidiines; in most bembidiines, sequences in at least one of the handful of standard genes provides a clear signal of lack of gene flow between species (e.g., Maddison 2008;Maddison and Cooper 2014;Maddison and Sproul 2020;Sproul and Maddison 2017). Why is the signal of species boundaries so clear in most bembidiine groups, but not B. transversale s. l.? Given genomic resources now available for this group, one fruitful and available avenue of future research would be comparison of coalescent patterns of thousands of regions of the genome within both this trio of Bembidion transversale group species and other groups of bembidiines with similar levels of morphological divergence. permission to collect B. levigatum from the Matheson Wetlands Preserve in Moab, Utah, and advice about the lands, I am grateful to Linda Whitham of the Nature Conservancy. For permission to collect at Wallowa State Park and Bandon Beach, I am thankful to Shawnae Stanton, Ben Fisher, Sara Griffith, and all of the other helpful people at the Oregon Parks and Recreation Department.