New data on the distribution of the two mole species Talpa aquitania Nicolas, Matinez-Vargas & Hugot, 2017 and T. europaea Linnaeus, 1758 in France based on museum and newly collected specimens

ABSTRACT A new species of mole, Talpa aquitania Nicolas, Martínez-Vargas & Hugot, 2017, was recently described from France. Based on the genetic identification of 270 individuals it was hypothesized that T. aquitania and T. europaea Linnaeus, 1758 are allopatric, being distributed on opposite sides of the Loire River, with the exception of a small area of sympatry in eastern Pyrénées mountains. The aim of the present study is to get a better understanding of the distribution of these two species in France based on the re-identification of museum specimens and extensive field sampling through a citizen science study. 1099 specimens were identified at the specific level (574 T. europaea and 525 T. aquitania) based on external characteristics, molar tooth morphology and/or genetic data. Our results confirm that T. aquitania is mainly distributed south and west of the Loire River, while the reverse is true for T. europaea. However, the Loire River is not a strict biogeographic barrier between these two species: several specimens of T. aquitania were recorded north and east of the Loire River in the Loire-Atlantique, Loiret and Nièvre departments. Moreover, two specimens were captured in the Var department. Several specimens of T. europaea were captured south and west of the Loire River in the Indre-et-Loire, Loir-et-Cher, Loiret and Loire departments. Furthermore, both species were found in sympatry or in close proximity in the Pyrénées Mountains. The role of climate, soil type, food resources abundance (earthworms) and historical factors in explaining the actual distribution of these species is discussed.


INTRODUCTION
Mammalian systematics is dynamic, with a long-term global rate of 25 new species recognized per year (Burgin et al. 2018). In the last 10 years numerous progress were made regarding our knowledge of the species diversity, systematic relationships and geographical distribution of mole species of the genus Talpa Linnaeus, 1758. This genus is endemic to the western Palaearctic region and is well known to the public through its characteristic heaps of soil, i.e., the molehills. The number of species within this genus was heavily underestimated for a long time because of their cryptic morphology linked to their semi-fossorial lifestyle. However between 2005 and 2020, the number of recognized species increased of 56%, from nine (Hutterer 2005) to 14 (Kryštufek et al. 2017;Kryštufek & Motokawa 2018;Demırtaş et al. 2020). This spectacular increase was possible due to the development of molecular tools in alpha systematics combined with a wide taxonomic sampling.
In 2017, Nicolas et al. (2017a) described a new species from France, Talpa aquitania Nicolas, Martínez-Vargas & Hugot, 2017, based on molecular (mitochondrial gene sequencing) and morphological data (eyelids fused together or not, and form of the mesostyle of the first, second and third upper molar). Nuclear data based on the HDAC2 gene confirmed the distinctiveness of this species (Nicolas et al. 2017b). Based on genetic identification of 270 individuals from France it was hypothesized that T. aquitania and T. europaea Linnaeus, 1758 may be allopatric, the first one being distributed south and west, and the second one north and east of the Loire River. However, two specimens of T. europaea were captured in the locality of Mosset in the Pyrenees mountains, inside the expected geographical range of T. aquitania (Nicolas et al. 2017b). More recently Rosoux & Lemarchand (2020) mentioned the capture of three specimens of T. aquitania north of the Loire River in the locality of Bray-Saint-Aignan (Loiret department), and Poitevin & Quéré (2021) mentioned it east of the Loire River in the Ardèche and Loire departments. Species distribution data are useful not only for population monitoring, biodiversity mapping and conservation management (Reese et al. 2005), but also in evolutionary studies aiming to understand the processes that lead to species distribution. Ultimately, spatial distributions of species are limited by three broad groups of constraints (Stewart et al. 2015): 1) tolerance of physical environmental conditions (temperature, pH, humidity, etc.), including both insurmountable physical barriers, such as rivers, oceans or mountains, but also spatial gradients in, for example, climatic variables (temperature, precipitation), or soil and water chemistry; 2) the availability and spatial distribution of resources (food, nest sites, shelter, etc.); and 3) interactions with other species, including predation, competition, mutualism, and parasitism/disease. History may also have an influence (e.g. where the location of an introduction or of refugia during previous range shifts determines the pattern of subsequent geographical spread). Although cases may exist where a single factor limits the distribution of a species, it is undoubtedly more likely that combinations of factors act synergistically, antagonistically, or independently of one another in limiting the expansion of species beyond their current range limits (Mott 2010).
Based on the re-identification of museum specimens and an extensive field sampling through a citizen science study, the aim of the present work is to get a better understanding of the distribution of the two mole species in France and the factors explaining it.

MATERIAL AND METHODS
This study includes only specimens whose species identification has been confirmed by the authors and does not includes bibliographical data only.
Specimens from the following museums were included in this study: MNHN (France) In order to include as many specimens as possible a citizen science study was developed. Numerous professional mole trappers and French naturalists sent us new records (either mole bodies in alcohol, skulls, tissues in ethanol or pictures of the eyes or the molars). Species records validated by genetic data and present in the Genbank database were also included.
1099 specimens from France were identified at the specific level (574 T. europaea and 525 T. aquitania; Appendices 1 and 2) based on external characteristics (open eyes in T. europaea; eyelids fused together in T. aquitania); molar tooth morphology (in T. europaea the mesostyles of the upper second molar (M2) and the third upper molar (M3) are divided into two cusps of subequal size, and they are aligned on a plane that extends parallel to the parastyle and the metastyle; T. aquitania specimens have either a simple mesostyle in M2 or an additional minute cusp, but unlike T. europaea this cusp is much smaller than the main cusp of the mesostyle and is located in the crest that connects the mesostyle to the metacone of M2 that is in a more lingual position than the mesostyle itself; in some T. aquitania specimens the mesostyle of M3 is composed on a main anterior cusp and a slightly smaller posterior cusp, while in other specimens this posterior cusp is not clearly discernible because its posterior border is fused to the crest that runs from the mesostyle to the metacone of M3) and/or genetic data following Nicolas et al. (Nicolas et al. 2017a, b). For specimens identified based on molar tooth morphology only adult specimens with unworn molars were considered (the pattern on the mesostyle cannot be determined on worn molars).
The locality of collect was known for 1058 specimens. Species records were mapped using DIVA-GIS 7.5.0.0 software. When available, exact GPS coordinates of the capture site were used. For museum specimens, when precise GPS coordinates were not available, they were estimated based on locality names and the use of public database such as mapawi. com or google earth.

RESULTS
Our results confirm that T. aquitania is mainly distributed south and west of the Loire River, while T. europaea is mainly distributed north and east of it (Fig. 1). However, the Loire River is not a strict biogeographic barrier between these two species: several specimens of T. aquitania were recorded north and east of the Loire River in the Loire-Atlantique (locality of Pontchâteau), Loiret (Bray-Saint-Aignan) and Nièvre (Gouloux and Montsauche-les-Settons) departments. Moreover two specimens, identified based on molars morphology, were captured in southern France in the Var department (Saint-Raphaël). Several specimens of T. europaea were captured south and west of the Loire River in the Indre-et-Loire (Tours, la Ville-aux-Dames), Loir-et-Cher (Saint-Aignan, Sassay, Neuvy, Sologne region), Loiret (Isde) and Loire (Saint-Haon-le-Châtel) departments. Furthermore, in the Pyrénées Mountains both species are found in sympatry or close proximity in the Pyrénées-Orientales (Mosset) and in south of the Haute-Garonne (Juzet-de-Luchon and Luchon). The two specimens of T. europaea found in Mosset were identified based on molecular analyses (both mitochondrial and nuclear data), and the specimen from Juzet-de-Luchon was identified based on molar morphology. All records were double checked to be sure that there is no mislabelling or taxonomic misidentification. One could argue that morphological traits are not always clear-cuts and that taxonomic misidentification is possible when it is based on only one morphological character. Even if it can never be completely ruled out, we are confident in our species identification because we had a perfect match (100% identity) between our identifications based on molecular data and morphological characters for the 351 specimens identified based on the two structures (121 specimens identified based on both molecular data and eyes morphology, and 230 specimens identified based on both molecular data and molars morphology).

DISCUSSION
This study shows that while T. aquitania is mainly distributed south and west of the Loire River, and T. europaea north and east of it, the Loire River is not a strict biogeographic barrier between the two species. This is not surprising given that moles are able to swim for 30-50 minutes, during which they can cover distances of over 1 km (Gorman & Stone 1990). Other big rivers like the Seine, the Garonne or the Rhine do not constitute effective barriers to geographical dispersal of French moles. Nicolas V. et al. 50 km According to Nicolas et al. (2017b), who modelled the climatically favourable areas for the two mole species on the basis of environmental niche models, climatic factors only partly explain the actual distribution of these species. In fact, the actual distribution of both species, and particularly those of T. aquitania, is much more reduced that expected based on climatic variables alone. Thus, the factors explaining the current distribution of these two species are probably not only climatic.
Species interactions (e.g. mutual competitive exclusion, predation, parasitism, mutualism) are also important in determining species distribution (Pearman et al. 2008;Sinclair et al. 2010;Giannini et al. 2013). Mutual competitive exclusion between the two French mole species may explain their geographical distribution. Sympatric tandems of congeneric moles occur in North America (genus Scapanus Pomel, 1848), the Mediterranean area (genus Talpa) and eastern Asia (genus Mogera Pomel, 1848) (Kryštufek & Motokawa 2018). Species in such tandems differ in size, and the larger mole is normally more abundant and widespread and occupies a broadest niche. It was shown that despite their widely corresponding distributions and their non-overlapping sizes, the co-occurring moles are only exceptionally syntopic, and can co-occur only if they occupy different habitat patches in a habitat mosaic (Kryštufek & Motokawa 2018). For example, in the Japanese moles, the dominant mole species, Mogera wogura (Temminck, 1842), is progressively expanding its range northwards, dis-placing the small inferior species M. imaizumii (Kuroda, 1957) (Abe 2001). Soil hardness is an important factor affecting the geographical distribution of these species and it allows their coexistence only under very specific circumstances. In other subterranean mammals, like pocket gophers, interspecific differences in body size and digging strategy have been shown to confer competitive dominance of one species over another depending on soil characteristics (Marcy et al. 2013). In regions where divergent soil types co-occur, the ranges of different pocket gopher species can overlap (Thaeler 1968). Significant body size differences between T. aquitania and T. europaea were found (Nicolas et al. 2017b). Studies are underway to test how these differences in body size, and potentially other morphological differences in anterior limb and hand morphology, affect the digging behaviour of these moles and their competitive dominance. It is interesting to note that areas of contact between the two French mole species correspond to highly heterogeneous lithological areas: areas of sympatry in the Pyrénées-Orientales, Haute-Garonne, Loire-Atlantique, Nièvre and Var departments correspond mostly to a mixture of Acid Plutonic Rocks and Metamorphic Rocks, and areas of sympatry in the Indre-et-Loire and Loiret departments correspond mostly to a mixture of Siliciclastic Sedimentary Rocks and Unconsolidated Sediments (Hartmann & Moosdorf 2012). More data on the ability of moles to dig in different soil types are harshly needed to test whether interspecific differences in body size and digging strategy explain the mostly allopatric distribution of these two species and their possible coexistence in few highly heterogeneous lithological areas.
Due to its small stomach capacity, fast digestion and low fat storage in tissues moles can endure only short periods of fasting, and are constantly in search of food during most of their active time (Kryštufek & Motokawa 2018). Several studies suggested that food is an important limiting factor affecting the spatial distribution and abundance of moles (Beolchini et al. 1996;Loy & Capanna 1998). Moles are mainly earthworm feeders, even if there may be seasonal and regional differences in the importance of earthworms in their diet (Gorman & Stone 1990;Kryštufek & Motokawa 2018). Earthworm abundance is not evenly distributed in France, being strongly affected by land use, climate-related parameters (e.g. soil moisture), vegetation, soil texture, soil organic matter and soil pH (Rutgers et al. 2016;Phillips et al. 2019). It is much lower is the south-western part of France (less than 50 individuals/m 2 ) than in the rest of the country (mostly 100-500 individuals/m 2 ), and this area of low earthworm abundance matches pretty well the distribution of T. aquitania. Additional data are needed on the diet of the two mole species in France and their ability to catch and consume different kind of prey, and on the competitive behavior. Unpublished geometric morphometric analyses showed significant differences in both size and shape of the skulls and mandibles of the two species suggesting different bite force and olfactory capacities (Martínez-Vargas et al. comm pers.) Finally, historical processes are also important in explaining species present-day distribution. The time of divergence between the two mole species is dated back to the Pliocene-Pleistocene boundary (Feuda et al. 2015;Nicolas et al. 2017b). Climatic oscillations leading to extreme drought or freezing, which dramatically increase hardness of soils and lower the availability of feeding resources represented by the foil fauna, have a significant impact on moles (Gorman & Stone 1990). During the Quaternary Period, which was dominated by Ice Ages and involved repeated global cooling, moles were frequently isolated in several glacial refuges, their populations expanding again during interglacial periods (Feuda et al. 2015;Nicolas et al. 2017b). Their current distribution would thus depend on their capacity to disperse from these refuges. The populations of T. aquitania in the Var department and of T. europaea in the Pyrénées Mountains may correspond to relict populations dating form glacial periods: climatic niche modelling analyses based on the Model for Interdisciplinary Research on Climate highlighted a high probability of presence of these species during the Last Glacial Maximum in these areas (Nicolas et al. 2017b).