Troglobiont micro snails of the genus, Iberozospeum Jochum, Kneubühler, Prieto & Neubert, 2022 are rarely found alive in caves of their known range encompassing the Cantabrian-Pyrenean region of northern Spain and the Pyrénées-Atlantiques in southwestern France (Fig. 1). The species, I. bellesi (Gittenberger, 1973), is considered uniquely Pyrenean and differentiates morphologically from its eight known Basque-Cantabrian congeners including I. schaufussi (von Frauenfeld, 1862); I. biscaiense (Gómez & Prieto, 1983); I. vasconicum (Prieto, De Winter, Weigand, Gómez & Jochum, 2015); I. zaldivarae (Prieto, De Winter, Weigand, Gómez & Jochum, 2015); I. percostulatum (Alonso, Prieto, Quiñonero-Salgado & Rolán, 2018); I. gittenbergeri (Jochum, Prieto & De Winter, 2019); I. praetermissum (Jochum, Prieto & De Winter, 2019), I. costulatum Prieto & Jochum, 2022 and the considered fossil, I. goliath Prieto & Gómez, 2025. In contrast to its eastern Alpine-Dinaric sister genus Zospeum Bourguignat, 1856 multiple species sympatry is not unusual for the genus (Kneubühler et al. 2022).

Although empty shells of I. bellesi (Gittenberger, 1973) have frequently been found in southwestern French caves (Dupré 1991, 1994; Tricoche and Origny 2002; Armand et al. 2012; Prieto and Zuazu 2018; Bertrand 2020; Chauvin et al. 2025), only two live populations had been recorded so far (Gargonomy 2018; Kneubühler et al. 2022). Enhanced exploration in April 2025 recently contributed four more to the tally (Chauvin et al. 2025). Of these, only one (Lezea cave) is known for which DNA sequence data exists from the French side of the Franco-Spanish border region (Kneubühler et al. 2022). In the most current phylogenetic study of the genus, only two populations from two proximate non-contiguous cave systems: one from Lexotoa cave (Spain, 100 m eastwards from the Franco-Spanish border) and Lezea cave (in Pyrénées-Atlantiques, less than 1 km from Lexotoa cave) (Fig. 1) were assigned by Kneubühler et al. (2022) to the I. bellesi clade. This clade differentiated genetically from the biscaiense-zaldivarae clade, but did not however, differentiate genetically from other species using the selected genetic markers (COI, 16S, H3, 5.8 S rRNA +ITS2). Kneubühler et al. (2022) hence suggested that I. bellesi likely constitutes a young radiation and that other genetic markers might nonetheless reveal genetic differentiation in I. bellesi.

Owing to the scarcity of data from caves on the French side of the Pyrenean chain and the paucity of DNA sequence data for I. bellesi in Kneubühler et al. (2022), new biospeleological expeditions (Chauvin et al. 2025) were conducted in March 2024 and April 2025 in karst caves of the Sare-Aïnhoa sector of the Franco-Spanish border region of the Pyrenees and its foothills in the department Pyrénées-Atlantiques. The Sare-Aïnhoa karst region lies about fifteen kilometers from the Atlantic Ocean and spans only a few square kilometers into both France and Spain (Navarra) (Vanara and Besson 2010).

In this work, three new populations sampled in March 2024, are assessed using DNA sequence data in sync with Kneubühler et al. (2022), whereby the sequences from these populations are added to the existing DNA sequence dataset. These new data are integrated into Kneubühler et al. (2022) phylogenetic reconstruction and presented here in context with its iberozospeid congeners (Fig. 2). Hereby, we provide a fundamental integrative presentation of extant members of the I. bellesi clade, incorporating image data of type material of I. bellesi (RMNH.MOL.234137) in conjunction with that of shells from the new French populations using 3D X-ray microscopy. The new morphological data present an enhanced description of the species, Iberozospeum bellesi (Gittenberger, 1973). The geological context, topography and ecology of the caves in the Franco-Spanish border region of the Pyrénées-Atlantiques are described. Live prevalence of the species is confirmed in multiple caves for the first time in France.

Our ML analysis revealed that the French populations cluster in the iberozospeid lineage I. bellesi (Fig. 2). Here, the only animal delivering the complete dataset is integrated in the cladogram shown by Kneubühler et al. (2022, Fig. 3). Remarkably, the other six specimens with an incomplete marker set always clustered in a polytomy including the Spanish I. bellesi populations. For this reason, the species identification is considered reliable. Addition of the new data set resulted in a new arrangement of the lineages. In the analysis ofKneubühler et al. (2022, Fig. 3), the subclade “zaldivarae + biscaiense” represented the highly supported sister group to all Iberozospeum species investigated. In the current analysis, the I. bellesi and the “I. zaldivarae + I. biscaiense” subclades are united (although only weakly supported) taking over the sister position now (support value > 92). This shows that the current state of knowledge is still rather incomplete and that more precise phylogenetic relationships can be expected when applying other methods and markers.

Measurements of shell characters reveal that these shells are not only minute but demonstrate a narrow range in shell size (sh: 1.17–1.60 mm) and shell width (0.85–1.02 mm) (Table 3). Despite their tininess, I. bellesi shells show considerable variation in shell shape and proportions. Since the relationship between shell height (sh) and shell width (sw) defines the slenderness and stoutness of the shell, the ratio of these measurements shows a variation ranging from 1.28 to 1.63, indicating the presence of both relatively squat and markedly elongate shell forms. The shell (NMBE-MOLL-0590059-2) from cave ES102 for example, demonstrates the stoutest shell of this study, which is most comparable to that of the paratype material (RMNH.MOL.234137) from Cueva [del Molino] de Aso, Huesca, Spain (Figs 3, 4, 5). In contrast, the individual from Urio gaina (NMBE-MOLL-0590058-1) presents a much slenderer shell form. Aperture dimensions are consistent with aperture height (ah) ranging from 0.45–0.58 mm while aperture width (aw) ranges from 0.45–0.54 mm. The aperture shape ratio (ah/aw) ranges from 0.83 to 1.16 with a ratio ah/aw frequently close to 1, indicating that the aperture scales proportionally with shell size and that the aperture is roughly oval to rounded in shape. The height of the last whorl (hlw) constitutes the major portion of total shell height in all specimens. The hlw/sh ratio ranges from 0.54 to 0.65, indicating that the last whorl consistently makes up more than half the total shell height, reflecting a compact spire and a shell form dominated by the body whorl. Of all the measurements in Table 3, the spire angle (SA) presents the greatest degree of variation, ranging from 51.84° to 70.93°. It clearly defines the key shape in these shells whereby a low SA (ca. 52°) reflects a tall, narrow and pointed spire and a high SA (ca. 71°) defines a short, wide and blunt spire. Of the 9 shells comprising the morphological part of this study, shell NMBE-MOLL-0590057-1 (Burgosa cave) bears the lowest spire angle (very pointed) measuring at 51.84° while that of cave ES102 (NMBE-MOLL-0590059-2) demonstrates the highest at 70.93° (very broad). We remark an apparent negative correlation between shell slenderness (sh/sw) and spire angle (SA), with more slender shells tending to exhibit lower spire angles.

In conjunction with the measurement data, our Micro-CT image data (Table 4, Figs 6, 7, 8, 9) reveal minimal variation in morphological characters such as teleoconch sculpture, peristome thickness, columellar configuration, protoconch and whorl configuration, and peristome rim alignment to the penultimate whorl (aperture facing left). On the other hand, there is noticeable differentiation in aperture shape from ovate to roundish-ovate to round with the aperture alignment either oblique to the shell axis or vertically aligned to it. The configuration of the parietal callus is also quite variable in degree of thickness and the extent to which it extends along and onto the body whorl (frontal view). The shells of this study also vary in the configuration of the last ¼ coiling of the body whorl in relationship to the umbilical zone (ventral view). For example, in the paratype shell (RMNH.MOL.234137), the end of the final whorl is very short and compactly formed, aligning straight next to a narrow umbilical depression (Fig. 5F). Conversely, one shell from Turziloa cave (NMBE-MOLL-0590056-1) shows this part not being compact and straight but rather, extending to the right side of the body whorl, forming a large umbilical depression (ventral view) (Fig. 6F). The second shell from Turziloa cave (NMBE-MOLL-0590056-2) bears rather, a shorter, more compact end part of the final whorl (Fig. 6L), which forms a larger umbilical depression than that present in the paratype shell. A similar contrasting situation is observable with the two shells from Burgosa cave (Fig. 7F, L) and from cave ES102 (Fig. 9F, L) while the ventral sides of the two shells from Urio gaina (NMBE-MOLL-0590058-1–2) (Fig. 8F, L) largely match the same ventral morphological configuration as that of the I. bellesi paratype (RMNH.MOL.234137) from Cueva [del Molino] de Aso, Huesca, Spain. We remark that these contrasting differentiating conditions in shells collected in the same cavities of the same caves may well reflect different generations of I. bellesi impacted by different environmental influences at different times. Considering that 5 from 9 shells in our morphological study encountered some degree of breakage during the usually non-destructive Micro-CT scanning process, is an indication that these shells are extremely thin and may possibly constitute fossil material. Moreover, Gittenberger (1980) remarked that the shells collected by JGM Raven (in 1978) from Navarra were less fragile than those he described from the type locality in Huesca. Hereby, environmental influences may largely factor in the degree of shell thickness of I. bellesi in these specific caves of the Franco-Spanish border region.

Ecologically, living Iberozospeum were found in areas typically characterized by moist to moderately humid walls, which occasionally exhibit signs of carbonated efflorescence. The runoff from these walls fosters the regular accumulation of organic mud where I. bellesi were found in a variety of locations, including cave walls, rock crevices, and the base of walls above the interface to the ground. These sites are characterized by their relative protection from air currents, serving as microhabitats that remain unexposed and prone to continuous, consistent environmental conditions over time. In such habitats, several live individuals were found in areas measuring approximately 3 to 4 square meters (e.g. in ES102). We remark that though ample, empty iberozospeid shell material was collected from other caves in the Pyrénées-Atlantiques during the speleological sampling for this investigation, we assessed only those shells from the live populations within the context of this study.

Iberozospeum bellesi is a strictly Pyrenean species still relatively underrepresented in the only molecular investigation of the genus (Kneubühler et al. 2022) but well documented in morphological and distributional assessments of the Basque Country (Dupré 1991; Prieto and Zuazu 2018). In our molecular study, only the population from ES102 (NMBE-MOLL-0585639) provided a complete DNA sequence data set, resulting in a new alignment of the lineages defining the I. bellesi clade (“blue clade”) in Kneubühler et al. 2022.

To our knowledge, the MHNG houses the only known I. bellesi shells from multiple caves in the Pyrénées-Atlantiques, collected by the Swiss speleographer Reno Bernasconi between 1979 and 1985. Dupré (1991) collected shells from 9 French cavities and later (Dupré 1994) cited his collecting “transparent” shells (“ident. BERNUSCONI”) from cavities in France, referring to the same material he and Bernasconi collected in 1985. The latter report suggests the shells represented recent mortalities at the time and thus, we infer that the respective caves harboured thriving I. bellesi populations. The shells at the NHMG are likely those reported by Dupré (1991).

Known diagnostic characters of Z. bellesi include the conical spire, oval aperture with thick parietal callus, spiral sculpture and a columella lacking lamellae (Gittenberger 1973; Prieto and Zuazu 2018). In their study of I. bellesi, Prieto and Zuazu (2018) remarked no significant differences in the holotype nor in shells they documented from the French Basque Country. The taxonomic examinations in this study corroborate these authors’ observations while underscoring the significance of the columella as a key diagnostic character in iberozospeid species. Its continuity in shape (cylindrical), moderate thickness (in aperture facing right view), and characteristic tortional form (irregular bulging under penultimate whorl and above the umbilical depression) is reliably conform in all four populations of I. bellesi in this study as well as in the paratype material (RMNH.MOL.234137) from Cueva [del Molino] de Aso, Huesca, Spain.

Autecologically, while empty I. bellesi shells are frequently found in large numbers in various parts of cavities and often result from displacement due to flooding events (Lukeneder et al. 2024), living populations remain relatively localized. These findings not only corroborate those reported by Jochum et al. (2012) in caves of the Spanish Basque Country and in Burgos province, but they also show that neither the size, the altitude nor the structure of the karst network appear to be determining factors for the presence of living Iberozospeum in the western Pyrenees. Studies are in progress to better characterize the ecological parameters (humidity, temperature, air circulation) that determine the distribution and habitat requirements of this troglobiont species.

This work confirms that I. bellesi demonstrates a genetically reliable but ecophenotypically variable species in multiple caves in the Pyrénées-Atlantiques of France.

Our analyses corroborate earlier investigations of I. bellesi in the Basque Country (Prieto and Zuazu 2018). Our additional investigation of shell measurements reveals a broad range of spire angles, influencing shell sizes and shape across populations from different caves. For example, cave ES102 harbours smaller shells, high spire angles and stouter morphs. Urio gaina harbours taller shells with lower spire angles and more elongated morphs. Turziloa and Burgosa caves appear to harbour individuals with intermediate shell forms. These differences likely reflect influences of environmental factors (i.e. slender shells living in narrow crevices and squat shells for stability on steep, slippery cave walls), instances of genetic divergence (Schilthuizen et al. 2006), or locally adaptive processes (Proćków et al. 2022).

Our investigations underscore the reliability and continuity of the columella as a species-specific diagnostic character while whorl number and size measurements remain too variable and reflect rather, factors conducive to the development of ecophenotypes. The multifarious topography and hydrogeologically diverse subterranean systems of the Pyrénées-Atlantiques harbour ample microhabitats, driving ecophenotypic variation in I. bellesi.

AJ, SG, QW, JC conceptualized the research. The authors, SG, JC, and QW collected and documented the material. JK and EN performed the molecular analyses. PM and BR conducted the 3-D X-ray tomographic imaging and rendering. AJ wrote the first draft of the manuscript. AJ, JC, QW, JK, EN and BR wrote parts of the manuscript; SG, JC and QW funding acquisition; EN provided infrastructure for the research; all authors reviewed and approved the final version of the manuscript.

Laboratory support was provided by the NMBE.

Open Access funding provided by the project, “Les troglotins des Pyrénées françaises: sciences participatives” obtained by JC, SG and QW from INPN (PatriNat MNHN-OFB), Fondation Biotope, Comité Départemental de Spéléologie des Pyrénées-Atlantiques, Comité Spéléologique Régional de Nouvelle-Aquitaine and the support of MIFENEC (Maison d’Initiation à la Faune et aux Espaces Naturels – Etudes et Conseils).

The authors have no conflicting interests.

AI assistance was used to format Table 4.

All data is published in the manuscript. Sequences are deposited in GenBank.