Specimens of Provanna were obtained from one of six sites at the Pacific Costa Rica Margin (CRM) during research expeditions conducted from 2017 to 2019 (Fig. 1, Table 1). Provanna were sampled by the human-operated vehicle (HOV) ‘Alvin’ and the remotely operated vehicle (ROV) ‘Subastian’ using various sampling tools attached to the HOV or ROV such as the manipulator arm or suction hose. The locations of sampling events were recorded for all specimens collected. Upon arrival to the surface, specimens were kept cold before being promptly preserved in > 95% ethanol. Specimens were then stored long term in ethanol at room temperature (20–25 °C).

Figure 1. 

Map of the Costa Rica Margin A overview of the Costa Rica continental shelf with hydrocarbon seep sites labelled B close-up view of the sites Mound 12 and Mound 11, which appear overlapping in the larger map. Bottom bathymetry is demarcated by black lines every 250 m.

All morphological characters and measurements are defined in Fig. 2 and Table 2. To begin identifying our specimens and constructing the key, original taxonomic descriptions of all extant Provanna species were obtained. The following shell characters were annotated for each species: The number of axial and spiral ribs on the body whorl, the relative strength of the body whorl sculptures, sculptural elements formed, how far down the body whorl the axial ribs extend, the presence of basal ribs, the average roundness (width / length) of the original holotype and paratype shells, the depth of the shell suture, and the descriptive shape of the aperture (Table 3). The following radular characters were also annotated: The relative width of the central teeth, the descriptive shape of the central teeth cusps, the roundedness of the central teeth’s anterior ridges, the number of denticles on the first lateral teeth, which denticle on the first lateral teeth is the most longest (“major” denticle), the descriptive shape of the first lateral major denticles, the angle of the first lateral posterior buttress, and the number of marginal tooth denticles (Table 3).

Figure 2. 

Visual definitions of morphological characters and measurements used in the study A shell morphological characters B informative measurements assessed for our own specimens C radular terminology and morphological characters. Abbreviations: SW: Shell Width, SL: Truncated Shell Length, GL: Maximum Granule Length, AL: Aperture Length, AW: Aperture Width.

To begin identifying the specimens from the CRM, they were first sorted into distinct morphotypes. Representatives from the full geographic, temporal, and size range of each morphotype were then selected for detailed morphological assessment. The following characters were measured for each specimen: Shell width (mm), truncated shell length (measured from the right, posterior tip of the penultimate whorl to the lowest point of the aperture (mm)), aperture length (mm), aperture width (mm), number of basal ribs (counted), relative shell texture (maximum granule length on the body whorl / truncated shell length), aperture roundness (width / length), and shell roundness (width / length). Truncated shell length was used as most Provanna lack any whorl past the penultimate whorl. Measurements of characters were taken from photographs captured by a mounted AmScope microscope adapter camera attached to a standard dissection microscope (Leica S6D, Leica Microsystems GmbH). A standardized 1-mm marker was present in every photo to allow for standardized measurements. Specimens were kept submerged in > 95% ethanol while images were taken. The line measurement tool within AmScope was used to measure morphological characters. To identify any potential collinearity among shell morphological characters, Pearson correlation coefficients (PCC) were calculated using the package Ggally (Schloerke et al. 2021) in R (v. 4.2.3; R Core Team 2022). One-way ANOVAs were then conducted in base R to identify significant differences in shell morphology among the morphotypes sampled from Costa Rica.

To characterize the radulae of each morphotype, we performed the following protocol. First, the body whorl of the shell was punctured using a sharp probe. The whole animal was then incubated in a 1.5 mL microcentrifuge tube containing a 10% solution of proteinase-k for 5–15 min at 56 °C. Incubation was monitored and terminated once tissue was visibly loose and degraded, but not fully digested. The microcentrifuge tube was removed from the heat source, pulse-vortexed three times, and then its contents were rinsed into a clean glass petri dish using deionized (DI) water. Under a dissection scope, the radular ribbon was identified, extricated from any remaining soft tissue, and moved to another clean glass petri dish containing DI water to further dilute the proteinase-k solution and prevent further breakdown of the radular ribbon. Silicon wafer chips cut into ~ 1 cm³ squares were used as mounting substrate for scanning electron microscopy. To mount the radula, a very small droplet of DI water was placed onto a chip. The radula was then placed into this water droplet and manipulated under a light microscope into a flat, teeth-up position using forceps or a sharp probe. Manipulation was most successful when the radula was wet but not submerged. The radula’s position was monitored and adjusted under a light microscope while the water was allowed to evaporate. Once dry, radulae naturally adhered to the chip’s surface and were then stored dry until imaging. Scanning electron microscopy was undertaken using a QuantaTM 450 FEG scanning electron microscope (FEI 2012) in its low-vacuum setting at Temple University College of Engineering’s Nano Instrumentation Center. High-quality images were obtained without sputter coating. Tentative morphological identities were then ascribed to our specimens.

To confirm the morphological identifications, the cytochrome oxidase 1 (CO1) mitochondrial gene and the histone 3 (H3) nuclear gene were sequenced. Tissue was obtained by pulling aside the operculum and pinching off a small sample of tissue from the foot (approximately 1 mm³). This tissue was then digested and its DNA extracted using a Qiagen Blood and Tissue DNA Extraction kit (QIAGEN, Valencia, CA). Extracted DNA was quantitated using a Nanodrop 2000 spectrophotometer. DNA was kept frozen at -20 °C following extraction. A 710 base pair (bp) section of the CO1 gene was targeted for sequencing using the primers LCO1490/HCO2198 and polymerase chain reaction protocol put forth by Folmer et al. (1994). A 274 bp section of the H3 gene was targeted for sequencing using the H3F/H3R primers put forth by Colgan et al. (2000) and the following PCR protocol: 94 °C for 1 min, 40 cycles of 95 °C for 30 seconds, 55 °C for 30 seconds, and 72 °C for 30 seconds, followed by a final extension period of 72 °C for 7 min. Forward and reverse reads were obtained through GeneWiz (Azenta Life Sciences, South Plainfield, NJ). Each sequence was quality-assured, trimmed, and reverse reads were reverse-complemented using the BioEdit desktop software (v. 7.2.5; Hall 1999). Forward and reverse reads were then used to create one consensus sequence per individual.

For all phylogenetic analyses, sequences were input and aligned using ClustalW embedded within the MEGA-X environment (v. 10.0.1; Kumar et al. 2018). The Model Finder embedded within MEGA-X was used to find the best-fit substitution model based on the lowest Bayesian Information Criterion. All base positions with less than 95% site coverage were excluded from analyses. Bootstrap confidences of branch points were assessed using 10,000 bootstrap replicates within MEGA-X. Bayesian topologies and Bayesian posterior probabilities (BPP) of branch points were computed using the joint programs BEAUti (v. 1.10.4) and BEAST (v. 1.10.4) (Suchard et al. 2018). The maximum clade credibility tree was then selected from the BEAST output using TreeAnnotator (v. 1.10.4). The resulting figures were cleaned and finalized using FigTree (v. 1.4.4) (Rambaut 2018) and Adobe Illustrator (v. 27.3.1). All published gene sequences were downloaded from the National Center for Biotechnology Information (NCBI) nucleotide database. All alignments are freely available on Github (Repository: melissajbetters/CRM_Provanna).

To verify inclusion within the genus Provanna, we assessed our novel sequences in relation to other Abyssochrysoids including species in the genera Abyssochrysos (Tomlin, 1927), Cordesia (Warén & Bouchet, 2009), Rubyspira (Johnson et al., 2010), Desbruyeresia (Warén & Bouchet, 1993), Alviniconcha (Okutani & Ohta, 1988), and Ifremeria (Bouchet & Warén, 1991). The Vetigastropods Caymanabyssia solis (Kano et al., 2016) and Notocrater pustulosus (Thiele, 1925) were used as the outgroup for investigations of CO1. The Vetigastropods Lepetodrilus pustulosus (McLean, 1988) and Pyropelta sp. (McLean & Haszprunar, 1987) were used as outgroup for investigations of H3. To verify the specific identity of our specimens, we assessed our novel CO1 sequences in relation to all other Provanna species available on NCBI. While CO1 sequences exist for P. annae (Nekhaev, 2023), these were amplified using a primer set that targeted a different region of the CO1 gene from our novel sequences, thus precluding comparison. To account for intraspecific variation, a maximum of three sequences per species (chosen at random) were included in the tree. Additionally, gene sequences with tentative or unknown identities were also included in case our novel sequences matched these. Desbruyeresia melanioides (Warén & Bouchet, 1993) was used as the outgroup.

To assess the robustness of current species delimitations within the genus, we calculated the average pairwise sequence divergence (APD) across CO1 sequences for Provanna. All sequences with verified species identities were included; Sequences with tentative or unknown species identities were excluded. Our novel sequences were assigned to their hypothesized species identities. All sequences were aligned using ClustalW embedded within MEGA-X and assessed using a Tamura 3-parameter substitution model (Tamura 1992), the pairwise deletion option (threshold = 95%), and 5,000 bootstrap replicates within MEGA-X. We then tested the number of species partitions supported within this dataset using the hierarchical clustering program ASAP (Assemble Species by Automatic Partitioning) (Puillandre et al. 2021).

Using the conclusions drawn from the preceding sections, a taxonomic key for all genetically supported, extant species of Provanna was constructed. A polytomous key was chosen as the format to capture the natural variation found in Provanna shells (Fig. 3). Of the morphological characters annotated, we prioritized sorting shells based on aspects of shell sculpturing, as these characters are easily recognized and do not require additional processing to observe. Radular characteristics were only utilized within the key when no other shell character could discern between species. Several morphological characters were excluded from the key either because we determined that they would introduce too much subjectivity in responses (precluding consistent utility), overlapped among species, overlapped with other characters, or varied too much within species. Incorporating our own morphological results, the presence and number of basal ribs and penultimate whorl morphological characters were excluded. It is noted in the key where there is uncertainty in a species hypothesis which is then addressed in the Discussion.

Figure 3 

. Examples of Provanna shell morphological variety A P. kuroshimensis, no sculpturing, growth lines present, flattened suture B–F constricted suture: B P. cooki, spiral sculpture only, no sculptural elements C P. chevalieri, axial sculpture only, no sculptural elements D–F both axial and spiral sculpturing: D P. fenestrata, sculptures about equal in strength, no sculptural elements E P. clathrata, axial sculpture stronger than spiral, blunt, sloping nodules F P. reticulata, spiral sculpture stronger than axial, minor spines.

Superfamily Abyssochrysoidea

Family Provannidae

Genus Provanna (Dall, 1918)

In total, 1,817 Provanna specimens were sampled from six sites at the CRM (see Table 1 for details). All specimens were sorted into one of four distinct morphotypes and subsequently assigned the following species identities: Provanna laevis (n = 1624) sampled from Jaco Summit, The Thumb, and Mound 12, Provanna ios (n = 180) sampled from Jaco Scar, Provanna pacifica (n = 9) sampled from Quepos Seep and Mound 11, and Provanna cf. lomana (n = 2) sampled from Mound 11 (Fig. 4). Provanna laevis was identified as it is the only smooth-shelled species from the Eastern Pacific. Our specimens of P. ios fit most closely the morphological description of P. goniata and were originally designated as such (Warén and Bouchet 1986; Betters et al. 2023) However, given our genetic results (detailed in the Results section Genetic Analysis), we amend this original identification to P. ios and address the validity of P. goniata as a distinct species in the Discussion. Specimens of P. pacifica were identified based on their sculpturing, their relatively small size, and the fact that this species was originally described from a soft-bottom, low-productivity seep in the Gulf of Panama very similar to its habitat at Costa Rica (Warén and Bouchet 1986). Provanna cf. lomana was tentatively identified by its unique feature of having only axial sculpturing on its body whorl. Of these specimens, a total of 158 representative Provanna covering the full geographic, temporal, and size range of each morphotype were measured (P. laevis, n = 96; P. ios, n = 52; P. pacifica, n = 8; and P. cf. lomana, n = 2). Representative radulae were successfully extracted and imaged for all species except P. cf. lomana (Fig. 5).

Figure 4. 

Provanna morphotypes sampled from the CRM A, B P. laevis from mussel shells, Mound 12, AD4917, 965 m C, D P. ios from unknown substrate, Jaco Scar, SD214, 1803 m E Specimens of P. pacifica from sunken wood, Mound 11, AD4988, 1017 m F Specimen of P. cf. lomana from mussel shells, Quepos Seep, AD4924, 1413 m. Both the dorsal and ventral view of each shell is shown. Scale bars: 1 mm.

Across all Costa Rican specimens, measurements of shell length, shell width, aperture length, and aperture width showed significant collinearity (PCC > +0.95, all pairs). Because the whorls past the body whorl showed variable levels of degradation, shell size was represented in analyses by aperture length alone, as we had more confidence in this measurement. All species sampled were comparable in size, with Provanna ios being the largest and P. pacifica being the smallest (Fig. 6B). The number of basal ribs, despite being commonly used to describe Provanna species, showed variation across all morphotypes (Fig. 6C). The difference in relative texture between P. ios and P. pacifica was significant (p < 0.001), supporting the utility of their sculptural elements in delineating species (Fig. 6D). In general, shells with granules longer than 3% of the shell length (major spines) were reliably P. ios. Provanna ios had the most oblique shell shape overall (Fig. 6E) and P. laevis had the most oblique aperture shape overall (Fig. 6F).

Figure 5. 

Representative radulae of Costa Rican Provanna species A P. laevis from mussel shells B P. laevis from unknown substrate C P. ios from mussel shells D P. ios from unknown substrate E, F P. pacifica from wood. Scale bars: 20 µm (F); 30 µm (E); 40 µm (C); 50 µm (A, B, D).

Figure 6. 

A the four morphotypes sampled B–F comparison of morphological traits among each morphotype. The number of individuals represented on the graph and included in one-way ANOVAs are denoted by “n =” above or below each bar. Resultant p-values from one-way ANOVAs are denoted above each graph (p-value: 0 < *** < 0.001 < ** < 0.01 < * 0.05). Provanna cf. lomana was excluded from all ANOVAs due to the small number of individuals, but is included here for graphical comparison. Note that the graphs of Shell Roundness E and Aperture Roundness F have y-axes that do not start at zero. Scale bars: 1 mm.

CO1 sequences were obtained from our specimens of P. laevis (n = 4), P. ios (n = 2), and P. pacifica (n = 2). All efforts to amplify CO1 for specimens of P. cf. lomana were unsuccessful. CO1 sequences generated were uploaded to GenBank and assigned accession numbers (OM914402–OM914408 & OP577954). H3 sequences were obtained from our specimens of P. laevis (n = 1), P. ios (n = 2), P. pacifica (n = 2), and P. cf. lomana (n = 1). H3 sequences generated were uploaded to GenBank and assigned accession numbers (OR687645–OR687650).

Phylogenetic analyses support the inclusion of our specimens in the genus Provanna with high confidence for CO1 (Bayesian Posterior Probability (BPP) = 100, ML = 90) (Fig. 7A) and H3 (BPP = 100, ML = 93) (Fig. 7B). Species-level investigations showed that sequences from the same Provanna species nested together and away from others on the tree (Fig. 8). Our specimens of P. laevis nested among P. laevis and P. glabra with high confidence and little to no distinction (BPP = 100, Bootstrap = 92). Despite our samples matching the physical description of P. goniata, these grouped together with sequences of P. ios and P. aff. ios with moderate confidence (BPP = 85). These were, however, all delineated as the sister group to P. variabilis (BPP = 100, Bootstrap = 80). Our specimens of P. pacifica grouped together and away from all other species on the tree (BPP = 100, Bootstrap = 99), though they were most closely related to P. aff. pacifica. However, given that P. pacifica has never been barcoded before, and that our specimens closely match the physical description and geographic distribution of the species proper, we assert that our specimens are indeed P. pacifica.

Figure 7. 

Bayesian topology of Abyssochrysoid gastropod mollusks A topology based on a 449 bp region of the mitochondrial CO1 gene. Topology was inferred using the HKY+G+I substitution model B topology based on a 266 bp region of the nuclear H3 gene. Novel sequences are bolded and highlighted in yellow. Numbers above branch nodes represent Bayesian posterior probabilities. Numbers below branch nodes represent the proportion of replicate trees in which the associated taxa clustered together in the bootstrap test (10,000 replicates). Only values above 50% are shown. The tree is drawn to scale, with branch lengths representing the number of base substitutions accumulated over time.

Figure 8. 

Bayesian phylogenetic tree of Provanna based on a 452 bp region of the mitochondrial CO1 gene. Novel sequences are bolded and highlighted in yellow. Topology was inferred using the HKY+G+I substitution model. Numbers above branch nodes represent Bayesian posterior probabilities. Numbers below branch nodes represent the proportion of replicate trees in which the associated taxa clustered together in the bootstrap test (10,000 replicates). Only values above 50 are shown. The tree is drawn to scale, with branch lengths representing the number of base substitutions accumulated over time.

Sequences of unknown identity (Provanna sp. 1 (GQ290577) and Provanna sp. 2 (GQ290578)) did not group together with any known species on the tree. Sequences from the Manus Basin that were previously identified as P. clathrata (Poitrimol et al. 2022) nested with high confidence among sequences of P. clathrata from the Okinawa Trough (Sasaki et al. 2016) (BPP = 100, Bootstrap = 99), confirming this identification and the range expansion for this species (see Table 4). As previously found, sequences from the Woodlark and Lau Basins group apart from all other known species on tree, as well as each other (Poitrimol et al. 2022) (BPP = 100, Bootstrap = 99). However, more detailed morphological investigations are needed to address whether these represent P. buccinoides, P. segonzaci, or one or more new species.

Average pairwise sequence divergences (APD) were computed across CO1 sequences (n = 236) (Table 5). APD calculations confirmed Provanna species as being more closely related to each other than to the outgroup (ingroup < 0.2 < outgroup). Almost all APD calculations fell between 0.05–0.13, confirming robust species distinctions overall within this genus (Hebert et al. 2003; Johnson et al. 2008). One exception to this was P. glabra and P. laevis, which showed very low sequence divergence (APD = 0.01, SE = 0.00). Provanna variabilis also appeared closely related to P. ios (APD = 0.04, SE = 0.01), supporting the topology from Fig. 8 designating them as sister clades. Comparisons within species were limited to those with more than one representative CO1 sequence, thus excluding P. cingulata, P. cooki, P. exquisita, P. fenestrata, P. lomana, P. sculpta, P. shinkaiae, and P. stephanos. Of the remaining species, average within-species APD’s fell below 0.03, confirming that intraspecific divergence was lower than interspecific divergence across Provanna species.

Hierarchical clustering performed by ASAP yielded 14 discreet subsets from an input of 19 hypothesized species (n = 236 sequences, p < 0.0001): (1) P. glabra-P. laevis, (2) P. variabilis-P. ios, (3) P. pacifica, (4) P. fenestrata, (5) P. cooki, (6), P. beebi, (7), P. cingulata, (8) P. exquisita-P. stephanos-P. clathrata, (9) P. sculpta, (10) P. lomana, (11) P. macleani, (12) P. subglabra, (13) P. kuroshimensis, (14) P. lucida. The threshold distance (Dt) used to partition the samples into species was 0.0496 (p < 0.0001) and the most common genetic distance between sequence pairs fell between 0.09–0.1.

Several Provanna species show little to no morphological distinction. For example, certain shell morphotypes of P. clathrata and P. ios have no discernable differences from one another besides the number of denticles on their outer marginal teeth (P. clathrata have about ten while P. ios have about 20) (Table 3). Certain shell morphotypes of P. ios and P. shinkaiae may also resemble one another. Yet, the lobate major denticles of P. shinkaiae’s first lateral teeth distinguish it from P. ios. Depending on morphotype, P. cooki and P. variabilis may also display criticism. Both may have no axial ribs, three spiral ribs on the body whorl, and no sculptural elements. However, they may be distinguished by the shape of the major denticles of their first lateral teeth (Table 3). Finally, P. annae and P. lucida both have unsculptured, translucent shells with constricted sutures. Their distinguishing feature is the shape of the major denticles of their first lateral teeth (P. annae are lobate, P. lucida are triangular). Regardless of the reliability of these radular characteristics, each of these species pairs may also be readily distinguished through CO1 barcoding (Table 5) (Nekhaev 2023).

The smooth-shelled species P. laevis, P. glabra, P. kuroshimensis, and P. subglabra are also morphologically indistinguishable based on shell characters (Table 3). All species have no sculpturing, slender shells, and a flattened suture. Provanna subglabra may be distinguished from the other three by having long, sharp major denticles on its first lateral teeth. Similarly, P. laevis should be distinguishable by having truncated, lobate denticles on its first lateral teeth and its location in the Eastern Pacific. Contrary to expectations, the radulae of our specimens from the CRM closely resembled that of P. glabra or P. kuroshimensis with long, lobate major denticles (Fig. 5A, B). Genetic characterization could not distinguish P. laevis from P. glabra, but could readily distinguish P. kuroshimensis (Fig. 8, Table 5).

Specimens collected from the Costa Rica Margin revealed that not all shell characters are useful in delineating species. Despite its widespread use in taxonomic descriptions, the number of basal ribs showed notable variation within species. Furthermore, as basal ribs are often very weak and difficult to count consistently, these were not used as a taxonomically informative characters in the key, nor do we recommend their use in distinguishing species in the future.

Our genetic investigations supported most current taxonomic delimitations, finding robust genetic distances among the 19 species from which CO1 sequences exist. Nonetheless, automatic partitioning based on CO1 supported the consolidation of several species. Provanna variabilis and P. ios, for example, were not partitioned. However, as these species are distinguished in both our phylogenetic analyses as well as by their morphological characteristics, more data are needed to verify this genetic similarity before taxonomic revision is undertaken. Similarly, the species P. exquisita, P. stephanos, and P. clathrata were also not partitioned. However, as these are also distinguished in our phylogenetic analyses and by their shell and radular characteristics, we believe more data are needed to warrant collapse.

Provanna laevis from the Eastern Pacific and P. glabra from the Western Pacific exhibited significant genetic overlap in our species-level phylogeny (Fig. 8), our distance matrix (Table 5) and were not distinguished during automatic partitioning. This similarity has been noted by previous studies (Sasaki et al. 2016; Linse et al. 2019). Given that CO1 seems informative for the rest of this genus, this similarity is noteworthy. Both P. laevis and P. glabra inhabit similar depth ranges, a variety of chemosynthesis-based environments, and are morphologically indistinct in all but radular morphology (Warén and Bouchet 1986; Okutani et al. 1992), which may not be sufficient to indicate complete lineage sorting. They have typically been distinguished by their distributions. Nevertheless, there may be numerous, undiscovered chemosynthetic sites, including large biomass falls (e.g., whale falls) that could provide the stepping stones necessary for connectivity across the Pacific. Furthermore, Pacific northern equatorial and subsurface counter currents may transport upper and lower layers of water, respectively, from west to east along the 7°N latitude (Kessler 2006), coinciding neatly with the region of study in Costa Rica (~ 8–9°N). Conversely, water upwelled at the Costa Rica Dome may flow east across the Pacific via the Northern Equatorial Current (Kessler 2006) and into the North Pacific Gyre, where it may realistically encounter the Eastern shores of Japan and even the Western shores of North America. Recent work supports the highly adaptable nature of P. laevis, which may explain a very broad distribution (Betters et al. 2023). Given these morphological, genetic, and biogeographic data, collected over several independent studies (Sasaki et al. 2016; Linse et al. 2019; Betters et al. 2023) and supported once again here, it is thus recommended that P. glabra and P. laevis be considered one species. As P. glabra is the younger name, we here synonymize it with P. laevis, as per the International Code of Zoological Nomenclature Article 23.1. These species are thus treated as synonymous hereafter.

Specimens of P. ios from the CRM were originally identified as P. goniata, given that their shells are decorated with major spines rather than minor spines (Fig. 4C, D) and that they are found at hydrocarbon seeps, rather than that at hydrothermal vents (Betters et al. 2023). We first reconsidered this identification when genetic barcoding of both CO1 and H3 could not reliably distinguish these specimens from P. ios (Figs 7, 8, Table 5). Furthermore, recent work has found that when specimens at the CRM were sampled from higher concentrations of hydrocarbons and sulfides, they tended to have more slender, thinner, and larger shells (Betters et al. 2023). This means that as they inhabit more vent-like conditions, they resemble more closely the vent species P. ios (Table 3). Thus, it is highly likely that these two species are actually ecotypes of a single molecular taxonomic unit, where P. ios is the vent ecotype and P. goniata is the seep ecotype. Additionally, both species are found at similar depths in the Eastern Tropical Pacific. Given that this study presents one of the most extensive collections of the morphospecies P. goniata known to date and is the first to genetically characterize them, we suggest that P. goniata and P. ios are one species. As P. goniata is the younger name, we here synonymize it with P. ios, as per the International Code of Zoological Nomenclature Article 23.1. These species are thus treated as synonymous hereafter.

Provanna are currently found in nearly every oceanic basin (Table 4). While many original descriptions distinguish species based on oceanic basin, the effect of geographic distance on population divergence of these gastropods from chemosynthetic habitats remains unclear (Distel et al. 2000; Vrijenhoek 2010; Breusing et al. 2023). For instance, while P. laevis (inclusive of P. glabra) spans the entire perimeter of the Northern Pacific Gyre with little genetic distinction, P. laevis, P. kuroshimensis, P. lucida, and P. subglabra all have overlapping biogeographic ranges at the Okinawa Trough yet display marked genetic divergence. Provanna are also notably adaptable across habitats, with six species currently known from more than one chemosynthesis-based ecosystem. These results indicate that more work is still needed to understand the drivers of genetic variation and isolation within this genus across a variety of contexts.

Finally, this study amends the biogeographic distribution of P. muricata. This species is listed as present in the North Fiji and Lau Basins in several secondary sources (Sasaki et al. 2016; Linse et al. 2019) based on Desbruyeres et al. (2006). This resource, however, does not present new records of occurrence, and instead summarizes known occurrence records. However, no primary literature nor museum specimens exist that place this species there. Therefore, until specimens are collected from the Western Pacific Basins and positively identified as P. muricata, this study proposes an amendment to their published biogeographic range, limiting it to the Eastern Pacific vents from which they were first found and described (Table 4).

The authors have declared that no competing interests exist.

No ethical statement was reported.

This research was supported by the National Science Foundation (OCE 1635219), Temple Uni-versity, and the American Malacological Society.

Conceptualization: EEC, MJB. Data curation: MJB. Formal analysis: MJB. Funding acquisition: EEC. Investigation: MJB. Methodology: MJB. Supervision: EEC. Validation: EEC. Visualization: MJB. Writing – original draft: MJB. Writing – review and editing: EEC.

Melissa J. Betters https://orcid.org/0000-0002-8975-257X

Erik E. Cordes https://orcid.org/0000-0002-6989-2348

Gene sequences generated in this study are accessible on GenBank under the accession numbers OM914402–OM914408, OP577954, and OR687645–OR687650. Morphological and genetic data from novel specimens are available on Github (Repository: melissajbetters/CRM_Provanna) or through correspondence with the lead author. The data underpinning the analysis reported in this paper are deposited at GBIF, the Global Biodiversity Information Facility, and are available at [https://doi.org/10.15468/4w9oc7] and [https://ipt.pensoft.net/resource?r=crm_provanna].