DNA was extracted from a fragment of the foot using the DNeasy Tissue kit according to the manufacturer’s specifications (QIAGEN, California, USA). A 600-bp fragment of the mitochondrial cytochrome oxidase subunit 1 (COI) gene was amplified by PCR using the universal primers LCO1490 and HCO2198 (Folmer et al. 1994). PCR amplifications were performed with a thermocycler (Techne TC-5000) on 25 μl reaction volume containing 2.5 μl of 10X PCR buffer, 2.5 mM MgCl₂ 10 mM dNTPs, 13 pmol of primers, 1U Taq DNA polymerase (Invitrogen) and 2 μl of template DNA. Cycling conditions were an initial 5 min denaturation step at 94 °C, followed by 35 cycles: 94 °C for 30 s, 50 °C for 30 s, and 72 °C for 45 s; and a final step at 72 °C for 7 min. The PCR products were purified using the GFXTM PCR DNA and GelBand Purification Kit (GE Healthcare, Buckinghamshire, UK) and sequenced by Macrogen, Korea.

The nucleotide sequences of the COI fragment were compared with sequences deposited in GenBank (http://www.ncbi.nlm.nih.gov) using BLASTn. Sequences were aligned and manually edited using CHROMAS v. 2.33 (http://www.technelysium.com.au/chromas.html) and SEAVIEW v. 4.0 (Gouy et al. 2010) software programs. After editing the sequences, a 450-bp fragment was used for phylogenetic analysis.

Phylogenetic analysis was performed by Bayesian inference with BEAST v. 2.7.3 (Drummond and Rambaut 2007) under the HKY substitution model. Tree runs were performed independently with four Markov chains that went for 10,000,000 generations, resampling every 10,000 states and a Burn-in of 10%. The new sequences were deposited in GenBank under the accession numbers PP431570, PP431571, PP431572, PP431573, PP431574, PP658222, PP658223, PP658224, and PP942456.

We examined shell shape variation with a two-dimensional geometric morphometric analysis, under the hypothesis that O. draparnaudi, O. cellarius, and O. alliarius represent three shell morphotypes that were defined by their genital anatomy. Shell shape variation was analyzed both in apertural and apical views, quantified through landmarks and semi-landmarks placed on photographs of the individual shells (Bookstein 1982; Zelditch et al. 2004). Shell images were captured using a camera attached to a stereomicroscope Nikon SMZ800, ensuring that the suture of the spire (in apical view) or the shell aperture (in apertural view) were observed in the same plane in all shells of either series.

A total of 22 apertural view photographs (eight O. cellarius, seven O. alliarius, and seven O. draparnaudi) and 39 apical view photographs (five O. cellarius, 17 O. alliarius, and 17 O. draparnaudi) were obtained. For the apical view, 50 semi-landmarks were used, while for the apertural view, 30 semi-landmarks were used (Fig. 1a, b). The x, y coordinates of the landmarks and semi-landmarks were digitized in the program TPSDIG2 v. 2.32 (Rohlf 2004), using the ‘draw background curves’ function, placing equidistant semi-landmarks along curves. In the apical view, two curves were used, one corresponding to the outer curvature of the shell starting at the distal point of the shell aperture and ending at the proximal point of the same aperture, and the other corresponding to the spire sutures (Fig. 2a); for the apertural view, only one curve that runs along the entire shell aperture, passing through the body whorl to the embryonic spire, was used (Fig. 2b).

Figure 1. 

Landmarsks sensu stricto (orange dots) and semi-landmarks placed on Oxychilus shells a apical view b apertural view.

Figure 2. 

Superimposition of landmark coordinates for the three morphotypes hypothesis a apical view b apertural view.

To minimize the tangential variation of the semi-landmarks, a coordinate adjustment was performed using the SEMILAND6 program from the IMP suite (Sheets 2014). To superimpose the specimen coordinates and extract relevant shape data for comparison, a Generalized Procrustes Analysis (GPA) was performed using the SEMILAND6 program from the IMP suite, in which all sets of landmarks were translated around a common origin to remove the effect of position, scale, and orientation (Masonick and Weirauch 2020).

Shell shape variation was evaluated using a Principal Coordinates Analysis (PCoA) in the PAST program v. 4.08 (Hammer and Harper 2006), based on the covariance matrix of individuals, obtaining a scatterplot with the two components that accounted for the highest percentage of variation. To represent the overall shape modification of each component, thin-plate spline deformation grids were calculated from the variability of the average shape along each PCoA.

Species distribution data were obtained from (1) the information on the labels of the specimens deposited in the IB-UNAM CNMO, (2) the Global Biodiversity Information Facility database (Hammer and Harper 2006), and (3) iNaturalistMX (https://mexico.inaturalist.org/). The citizen observations in iNaturalistMX were reviewed by VAG, considering only those that matched the characteristics of the genus. Blurry photos and erroneous identifications (two records involved Succinea sp., two records to family Polygyridae, one to family Euconulidae, and one was an insect larva) were removed. These corrections were documented on the iNaturalistMX observation platform. With these records, a data table was built with the following attributes: 1) latitude, 2) longitude, 3) source from which the record was obtained and 4) locality (Suppl. material 1: table S2). Records with incomplete locality data that did not allow locating the collection site were excluded from the analysis.

The model was created utilizing an R implementation (R Core Team 2024) of Maxent from the Maxnet library (Phillips 2021) and SDMtune (Vignali et al. 2022) for the training and evaluation of distribution models. Maxent employs the principle of maximum entropy to estimate the probability distribution of a species using presence and absence data. The species habitat modeling was done with 142 Oxychilus spp. occurrence locations in Mexico: 26 from our collections and the CNMO records, one from GBIF (a specimen in the Carnegie Museum of Natural History) (GBIFa 2022; GBIFb 2023) and 115 from iNaturalistMX. We employed 19 bioclimatic variables downloaded from the WorldClim v. 2.1 (http://worldclim.org/version2) with a spatial resolution of 30 s. The model establishes a statistical connection between the environmental factors found at the locations where a species is observed and data indicating the species presence. It is commonly used for habitat modeling, even when there is limited information about the species occurrences (Phillips et al. 2006). The model randomly selects background points across the study area, representing locations where the species is considered to be either truly absent or likely absent (pseudo-absent). The relative significance of the environmental variable was calculated using Jackknife analysis inbuilt in the model. Using jackknife and Spearman correlation analysis we included the variables which were not highly correlated (r < 0.7) (Fig. 3). The model was evaluated with the help of AUC (Area under curve) of the ROC (Receiver operating characteristics) plot and true skill statistics (TSS) (Allouche et al. 2006). AUC in percentages measures model performance and varies from random to perfect discrimination (Swets 1988; Fawcett 2006).

Figure 3. 

Spearman pairwise correlation coefficients between predictive variables. Variables selected for the final model were not highly correlated (r < 0.7).

We analyzed 50 specimens (Suppl. material 1: table S1) and 74 shells from four states of Mexico in two biogeographical provinces (Fig. 12). The live snails were found under logs and damp wood in (1) Tlaxcala (Atlihuetzia) where they lived in sympatry with Deroceras laeve (O.F. Müller), 1774, D. reticulatum (O.F. Müller 1774), Arion intermedius Normand, 1852, and Boettgerilla pallens (Simroth, 1912), (2) Querétaro where they were found in sympatry with D. laeve, (3) CDMX (Bosque de Tlalpan) where they co-occurred with D. laeve and B. pallens, and (4) Puebla (Teopancingo) where they were found in sympatry with B. pallens and Pallifera sp. The newly collected specimens were deposited at the CNMO (Suppl. material 1: table S1).

Superfamily Gastrodontoidea Tryon, 1866

Family Oxychilidae Hesse, 1927

Subfamily Oxychilinae Hesse, 1927

 Oxychilus draparnaudi (H. Beck, 1837)

Fig. 4

Worldwide distribution

Originally described from France, probably in the Montpellier area (Giusti and Manganelli 1997). Its native range includes western Europe and the western Mediterranean region (Barker 1999). It has spread to other parts of the world, including North America (Forsyth 2004) Russia, North and South Africa, Asia, Australia, and New Zealand (Barker 1999), Madeira (Seddon 2008), and Argentina (Virgillito and Miquel 2013).

Distribution in Mexico

Querétaro (Cadereyta de Montes), Tlaxcala (Atlihuetzia), State of Mexico (Tlalnepantla), CDMX (Álvaro Obregón, Benito Juárez, Tlalpan). According to Morrone (2019), localities belong to the Sierra Madre Oriental and Transmexican Volcanic Belt Province.

Diagnostic features

15 specimens were dissected. Anatomically, all of them displayed the genital features typical of O. draparnaudi as described by Giusti and Manganelli (1997): a penis with a very slender proximal portion and a wider distal portion, both separated by a “bottle-neck” (= BS, i.e., a twisted duct or constriction) covered by a thin translucent sheath (Fig. 4D, G). Internally the proximal penis shows prominent papillae (Fig. 4F, I), while the distal penis shows four to five thin longitudinal internal folds (Fig. 4E, H).

Figure 4. 

Oxychilus draparnaudi (CNMO 8470 Cadereyta, Querétaro). A–C shell in apical, ventral, and apertural view D, G genitalia of O. draparnaudi F, I internal ornamentation of the proximal penis E, H internal ornamentation of distal penis J, K live specimens. Abbreviations: BS bottleneck region, BC bursa copulatrix, DBC bursa duct, E epiphalus, F flagellum, P penis, POS prostatic portion of ovospermiduct, PR penial retractor, PS penial sheath, UOS uterine portion of ovospermiduct, V vagina, VD vas deferens, VG vaginal gland.

Shell discoidal (Fig. 4B, C) with a depressed spire (Burch 1962), thin, yellowish, shiny; shell surface generally smooth with fine growth lines most evident near the suture and few very fine spiral lines (Fig. 4A). Shell whorls: 5 ½ to 6. Umbilicus moderately wide, 1⁄6 of maximum shell diameter. Shell diameter 9–13 mm, shell height 3–6 mm (Fig. 7). Aperture width: 3.00–4.95 mm.

Radula (n = 5) composed of 30 rows with ~ 25 teeth/row (Fig. 8A). Radular formula: C/3 + 2–3 L/3 + 0–1 LM/2 + 9–12 M/1.

 Oxychilus alliarius (J.S. Miller, 1822)

Fig. 5

Worldwide distribution

Originally described from the “Environs of Bristol”, England (Miller 1822). Its native distribution includes Iceland, Greenland (Roth and Sadeghian 2003), and Central and Western Europe (Horáčková and Juřičková 2009). It has been reported as introduced in Austria, Belgium, Denmark, Estonia, Finland, France, Germany, Great Britain, Greece, Czech Republic, Spain, and Portugal (Horáčková and Juřičková 2009; Borges et al. 2010; De Oliveira and Altonaga 2010; IUCN Red List 2017), and spread to Hawaii (Cowie 1997), Tasmania (Kershaw 1991), South Atlantic (Preece 2001), North America (Forsyth 2004), Colorado, the Pacific coast states (Roth and Sadeghian 2003), South America (Hausdorf 2002), Chile (Cádiz et al. 2013), Sri Lanka (Naggs et al. 2003), Madeira (Seddon 2008), and the Philippines (Tañan and Sumaya 2024).

Distribution in Mexico

Puebla (Teopancingo), Transmexican Volcanic Belt Province.

Diagnostic features

Two specimens were dissected. The proximal part of the penis is initially wider and then gradually becomes distally narrower as described by Giusti and Manganelli (2002) (Fig. 5D, G). Internally the proximal (Fig. 5E, H) and distal penis show 2–4, thin, slightly undulating, longitudinal folds that seem to be a continuation of one another, but never showing lateral branching or assuming a papilla-like shape (Fig. 5F, I).

Figure 5. 

Oxychilus alliarius (CNMO 8464 Teopancingo, Puebla). A–C shell in apical, ventral and apertural view D, G genitalia of O. alliarius E, H internal ornamentation of the proximal penis F, I internal ornamentation of the distal penis. Abbreviations: BC bursa copulatrix, DBC bursa duct, E epiphalus, F flagellum, P penis, POS prostatic portion of the ovospermiduct, PR penial retractor muscle, PS penial sheath, UOS uterine portion of the ovospermiduct, VG vaginal gland, VD vas deferens.

Shell discoidal, depressed, slightly convex above, compressed below, thin, semitransparent, variably shiny, yellowish to yellowish-brown, opalescent below (Fig. 5B, C); surface fairly smooth, with fainter growth lines more pronounced at sutures and very fine, wavy, spiral lines (Fig. 5A); aperture oval, oblique. Shell whorls: 4–4½. Umbilicus is rather broad, ~ 1⁄6 of maximum shell diameter. Shell dimensions: 6–9 mm diameter, 3–4 mm height (Fig. 7). Aperture width: 2–4 mm.

Radula (n = 3) composed of~ 35 rows with 25–31 teeth/row (Fig. 8B). Radular formula: C/3 + 2–3 L/3 + 0–1 LM/2 + 9–13 M/1.

 Oxychilus cellarius (O.F. Müller, 1774)

Fig. 6

Worldwide distribution

Originally described from a wine cellar in Copenhagen (Müller 1774). Its native range includes northern and western Europe (Horáčková and Juřičková 2009), Asia Minor, and North Africa (Roth and Sadeghian 2003). It has been introduced into Tasmania (Kershaw 1991), Greenland, North America (Forsyth 2004), St Helena, South Africa, Chile (Stuardo and Vega 1985), Hawaii (Cowie 1997), Australia and New Zealand (Barker 1999).

Distribution in Mexico

CDMX (Tlalpan).

Diagnostic features

Two specimens were dissected. The penis is cylindrical with a relatively constant width in the middle portion (Fig. 6D, G). Internally the proximal penis shows small, evenly distributed papillae that tend to be joined together in rows which occasionally form undulating folds (Fig. 6F, I). The distal penis presents four, thin, internal folds (Fig. 6E, H).

Figure 6. 

Oxychilus cellarius (CNMO 3858 Tlalpan, CDMX). A–C shell in apical, ventral and apertural view D, G genitalia of O. cellarius E, H internal ornamentation of the distal penis F, I internal ornamentation of proximal penis. Abbreviations: BC bursa copulatrix, DBC bursa duct, E epiphalus, F flagellum, P penis, POS prostatic portion of the ovospermiduct, PR penial retractor muscle, PS penial sheath, UOS uterine portion of the ovospermiduct, V vagina, VD vas deferens, VG vaginal glands.

Shell discoidal, flattened, convex (Fig. 6C), light yellowish in color with fine lines of radial growth (Fig. 6A). Shell whorls: 4½–5. Umbilicus slightly flared, ~ 1/12 of maximum shell diameter (Fig. 6B). Shell dimensions: 8–10 mm diameter, 4 mm height (Fig. 7). Aperture width 3–4 mm.

Figure 7. 

Shells of Mexican (from left to right) Oxychilus alliarius (CNMO 8464 Teopancingo, Puebla, O. cellarius (CNMO 3858 Tlalpan, CDMX), and O. draparnaudi (CNMO 8470 Cadereyta, Querétaro). A apical view B ventral view C apertural view.

Radula (n = 2) composed of 35 rows with ~ 25 teeth/row (Fig. 8C). Radular formula: C/3+ 2–3 L/3+ 0–1 LM/2 + 9–14 M/1.

Figure 8. 

Radulae of Mexican. A O. draparnaudi (CNMO 8470 Cadereyta, Querétaro), B O. alliarius (CNMO 8464 Teopancingo, Puebla), and C. O. cellarius (CNMO 3858 Tlalpan, CDMX). Abbreviations: DC central tooth, DL lateral teeth, DM marginal teeth.

The COI tree recovered three clades corresponding to the three species studied in this study (Fig. 9), each with a posterior probability of 0.99–1.00. The sequence of the specimen from Teopancingo Puebla (TEN) was associated with GenBank sequences of O. alliarius from the United Kingdom and New Zealand, whose morphology agrees with that of the Mexican specimens. The other eight sequences from Querétaro, Tlaxcala, CDMX and Edo. Mex. (Suppl. material 1: table S3) formed a group with the GenBank sequences of O. draparnaudi from France, Canada, and the USA. Attempts to amplify the COI gene fragment for O. cellarius were unsuccessful due to insufficient DNA quality. However, the sequences reported in GenBank for this species were included in the analysis and formed a distinct group.

Figure 9. 

Bayesian phylogenetic tree of Mexican Oxychilus alliarius, O. cellarius, and O. draparnaudi based on the COI gene fragment (450 bp).

The superimposition of shell shape configuration in the apertural view for each individual did not exhibit regions with differences among the species. However, in the apical view, slight variations in the width of the shell aperture were observed (Fig. 9). The first two principal coordinates of the PCoA in the apertural view accounted for 70.96% of the variation (PCo1 = 53.39%, PCo2 = 17.57%), while in the apical view, they accounted for 60.80% of the variation (PCo1 = 42.13%, PCo2 = 18.67%). The scatterplots of the first two PCoA coordinates did not separate the species in either of the two evaluated views (Figs 10, 11).

Figure 10. 

Geometric morphometric variation among shells of Mexican Oxychilus alliarius, O. cellarius, and O. draparnaudi in apical view.

Figure 11. 

Geometric morphometric variation among shells of Mexican Oxychilus alliarius, O. cellarius, and O. draparnaudi in apertural view.

A map was obtained using data from Mexican specimens of O. alliarius, O. cellarius, and O. draparnaudi, showing that O. draparnaudi is the most widespread of the three species. Most records are in the Transmexican Volcanic Belt Province, including the states of Querétaro, Puebla, CDMX, and the State of Mexico (Fig. 12). The potential distribution map for the genus Oxychilus shows a tendency towards locations in the Volcanic Axis, a part of the Eastern Sierra Madre, and a part of the Southern Sierra Madre (Fig. 13). The variables that contributed most for predicting potential areas according to correlation and jackknife analysis (Suppl. material 1: fig. S1) were BIO_5 (maximum temperature of warmest month), BIO_14 (precipitation of driest month (mm), BIO_11 (mean temperature of coldest quarter), BIO15 (precipitation seasonality (coefficient of variation)), BIO_3 (isothermality), BIO_7 (temperature annual range) and BIO_2 (mean diurnal range). The max temperature of the warmest month (BIO_5) of potential areas was 20–25 °C, and the precipitation during the driest month (Bio_ 14) was 5–15 mm (Suppl. material 1: fig. S2). The model performed well with an AUC of 0.958 (Suppl. material 1: fig. S3) and a TSS of 0.86.

Figure 12. 

Actual distributions of O. draparnaudi (circle), O. cellarius (square), and O. alliarius (triangle) in Mexico. Biogeographic provinces (Morrone et al. 2017): T. Tamaulipas Province, ChD. Chihuahuan Desert Province SMO. Sierra Madre Oriental Province, TVB. Transmexican Volcanic Belt Province, BB. Balsas Basin Province, HCh. Chiapas Highlands Province, YP. Yucatan Peninsula Province.

Figure 13. 

Potential distribution of the genus Oxychilus in Mexico. Green areas indicate high suitability areas.

The authors have declared that no competing interests exist.

No ethical statement was reported.

This work formed part of the bachelor dissertation of Ali Trujillo, who received a scholarship from Instituto Politecnico Nacional (BEIFI-IPN).

Conceptualization: VAG. Data curation: VAG, ENG. Formal analysis: AGTD, VAG. Funding acquisition: GZ, VAG. Investigation: AGTD. Methodology: JGR, JLHD. Project administration: GZ, VAG. Resources: GZ. Software: JLHD, JGR. Supervision: VAG. Validation: ENG. Writing – original draft: AGTD. Writing – review and editing: VAG, ENG.

Ali Gabrielle Trujillo-Díaz https://orcid.org/0009-0002-2120-6414

Victoria Araiza-Gómez https://orcid.org/0000-0001-5583-1918

Jazmín García-Román https://orcid.org/0000-0003-0044-339X

Gerardo Zúñiga https://orcid.org/0000-0002-5041-4258

Edna Naranjo-García https://orcid.org/0000-0001-9015-1758

All of the data that support the findings of this study are available in the main text or Supplementary Information.