French Mediterranean islands as a refuge of relic earthworm species: Cataladrilus porquerollensis sp. nov. and Scherotheca portcrosana sp. nov. (Crassiclitellata, Lumbricidae)

. The area comprising the pyrenees, Northeast Spain, Southern France and Corsica-Sardinia supports a large part of the diversity of Lumbricidae earthworms, including most species of the endemic genera Prosellodrilus , Cataladrilus and Scherotheca . In this region, the probability of encountering new species for science is significant, especially in scarcely sampled localities. In this study, we describe two unidentified species recently collected in the Hyères Archipelago (France), which we assigned to the genera Cataladrilus and Scherotheca based on morphological characters and molecular phylogenetic analyses. other species of Scherotheca from Montpellier (including the type species of the genus, Sc. gigas gigas ) were included in the analysis to clarify their conflicting systematics. A reduced molecular marker set ( COI , 16S, 28S and ND1 ) proved as successful as larger marker sets for identifying phylogenetic relationships within the Lumbricidae. remarkable disjunctions between both Cataladrilus porquerollensis Marchán & Decäens sp. nov., Scherotheca portcrosana Marchán & Decäens sp. nov. and their most closely related relatives, suggesting a strong influence of paleogeographic events on the earthworm fauna of the area and a possible specific status as well as the retention of other subspecies, highlighting the importance of testing for such delimitation with molecular methods. & French Mediterranean islands as a refuge of relic earthworm species: Cataladrilus porquerollensis sp. nov. and Scherotheca portcrosana sp. nov. (Crassiclitellata, Lumbricidae). European Journal of Taxonomy 701: 1–22.

The genus Scherotheca includes 41 species and subspecies (Qiu & Bouché 1998c), half of which occur in occitanie, provence-Alpes-Côte d'Azur and Corsica. Scherotheca gigas (Dugès, 1828), the type species of the genus, includes several subspecies; most of them were established as species by Qiu & Bouché (1998d). Domínguez et al. (2015) included 5 representatives of Scherotheca in their phylogeny of Lumbricidae, two of which were unidentified species. In addition, the specimens identified as Sc. gigas were collected in Navarra (Spain), which is far removed from the type locality of the species (Montpellier) raising doubts about their identification. Thus, the systematics of this highly diverse genus would benefit from the corroboration provided by a comprehensive molecular phylogenetic analysis, which has not yet been performed.
Despite the intensive research on earthworm diversity in southern France, some endemic species in relatively remote, comparatively weakly explored areas may remain undescribed. one example of such an area is the Hyères Archipelago, which faces the coast of Provence. These continental islands remained connected to Corsica and Sardinia until Late Chattian-Aquitanian -ca 24 Ma -when the first marine ingression preceded the rotation and drifting of the Corso-Sardinian microplate (oudet et al. 2010). Due to their proximity to the coast, the islands were intermittently connected to the French mainland during the pleistocene until 11 000-12 000 BC, when they became permanently isolated (Médail et al. 2013). The Port-Cros National Park was created in 1963 to protect the Hyères Archipelago, and two of its three main islands (Porquerolles and Port-Cros) constitute its core areas.
Two undescribed earthworm species were collected during a sampling survey of the islands of Porquerolles and Port-Cros (Decaëns et al. 2020). Further detailed morphological diagnoses and molecular phylogenetic analyses confirmed the novelty of both taxa, but also revealed that they should be assigned to the genera Scherotheca and Cataladrilus respectively. The inclusion of closely related congeneric species in the molecular phylogenetic reconstruction provided further insight into the systematics of these genera. In this work we relate the results of both of these morphological and molecular analyses, and propose the formal description of the two new species under the names Scherotheca portcrosana Marchán & Decäens sp. nov. and Cataladrilus porquerollensis Marchán & Decäens sp. nov.

Specimens, sampling and morphological description
Specimens described in this work were collected in a sampling survey carried out in the port-Cros National park (France) in March 2018 (Decaëns et al. 2020). The rest of the specimens were collected in Montpellier (France) and its vicinities at different times between 2015 and 2019. The list of species and the localities where they were collected is shown in Table 1. Earthworms were obtained by soil digging and hand-sorting, rinsed with water and fixed in 100% ethanol to enable further molecular analyses. Species classification and morphological diagnoses were carried using the set of external and internal morphological characters used by Qiu & Bouché (1998a, 1998b, 1998c, 1998d, and following the format established by Domínguez et al. (2018). Main external morphological characters were: average length, average number of segments, average weight, pigmentation, type of prostomium, setal arrangement, position of papillae, position of first dorsal pore, nephridial pore arrangement, position and development of male pores, position and development of female pores, position of spermathecal pores, position of clitellum, position of tubercula pubertatis. Main internal anatomical characters were: position of oesophageal hearts, position and morphology of calciferous glands, position of crop, position of gizzard, type of typhlosole, shape of nephridial bladders, number and position of seminal vesicles, number and position of spermathecae.

Institutional acronyms
Earthworm holotypes and paratypes were deposited in the following institutions: CEFE = Center of Functional and Evolutionary Ecology, Montpellier, France uCMLT = Earthworm Collection of universidad Complutense de Madrid, Spain
DNA sequences obtained in this study (including the two new species and the above mentioned Scherotheca spp), as well as their associated meta-data and genBank accession numbers are all available in the public dataset "DS-EWSppCNp" on the BoLD bioinformatics platform (https://doi.org/10.5883/DS-EWSppCNp).
Sequences reported by Domínguez et al. (2015), pérez-Losada et al. (2009, paoletti et al. (2016) and De Sosa et al. (2019), including representatives from most of the Lumbricidae genera and two members of the closest families (hormogastridae Vejdovsky, 1884 and Criodrilidae Michaelsen, 1900) were downloaded from genBank and used as a reference dataset. Included species are shown in Appendix 1.
Sequences were aligned with MAFFT ver. 7 (Katoh & Standley 2013) with default settings and concatenated with BioEdit (Hall 1999), resulting in a matrix of 3348 bp. The best fitting evolutionary  1978). GTR+I+G was selected as best-fitting evolutionary model for COI, 28S and ND1, and hKY+I+g was selected for 16S.
Maximum Likelihood analysis was performed with raxML-hpC ver. 8 (Stamatakis 2014) as implemented in the CIprES Science gateway ver. 3.3 (Miller et al. 2010), using gTr+I+g for each data partition with 10 alternative runs and estimating the support for the resulting topologies by 1000 rapid bootstrap replicates. Bayesian Inference of the phylogeny was estimated with MrBayes ver. 3.1.2 (Ronquist & huelsenbeck 2003) as implemented in the CIprES Science gateway ver. 3.3. parameters were set to 50 million generations sampling every 5000 th generation (10 000 trees). Two independent runs with four chains each were performed and 20% of the trees were discarded as burn-in. The remaining trees were combined and summarized on a 50% majority-rule consensus tree. Clade support (Bootstrap and Posterior probability) values over 70% and 90% respectively were considered as high (see Marchán et al. 2018;De Sosa et al. 2019).
uncorrected average pairwise distances between the newly described species and their closest relatives for the molecular markers COI and 16S were calculated in MEgA X (Kumar et al. 2018) in order to support their status as separate species.

Etymology
The species name is derived from Porquerolles, the island where this species was found.

Morphological description
External morphology Body pigmentation absent in live specimens. White-beige homogeneous color in fixed specimens (Fig. 2).

Distribution and ecology
Cataladrilus porquerollensis Marchán & Decäens sp. nov. is known from the island of Porquerolles in the Hyères Archipelago, France. This species has been found in meadows, vineyards and olive groves, thus appears to inhabit moderately to highly anthropized habitats.

Etymology
The species name is derived from port-Cros, the island inhabited by this species.

External morphology
Body pigmentation very faint brown-grey. White-beige with dorsal brownish mid-segment brown bands in fixed specimens (Fig. 3).

Distribution and ecology
Scherotheca portcrosana Marchán & Decäens sp. nov. is known from the island of port-Cros in the Hyères Archipelago, France. This species has been found in meadows, pine and evergreen oak forests, thus appears to have a preference for natural habitats.

Systematic implications
Molecular phylogenetic analyses and genetic distances clearly support the morphology-based inclusion of Ca. porquerollensis Marchán & Decäens sp. nov. within the genus Cataladrilus. It is worth noting that the relationship between Cataladrilus and Prosellodrilus is currently not fully resolved based on molecular information, however the morphological characters clearly separate both genera (Table 3) according to the number and position of spermathecae: two pairs in intersegments 9/10, 10/11 for Cataladrilus (except for Ca. multithecus Qiu & Bouché, 1998) instead of two to three pairs (sometimes multiple) in intersegments (12/13), 13/14, 14/15 for Prosellodrilus (Qiu & Boché 1998a, 1998b. Several species from Cataladrilus and Prosellodrilus are yet to be included in a detailed molecular phylogenetic analysis, with emphasis on including representatives from the different subgenera (possibly genera for Prosellodrilus sensu lato according to Blakemore (2012). This will be necessary to resolve the reciprocal monophyly of both genera and the status of the subgenera/genera included within them. Marchán & Decäens sp. nov. resembles Cataladrilus (Latisinella) mrsici, the most in the position of their tubercula pubertatis (segments 29-31) and their clitellum (segments (19)20-32, 33(34) and 22-32 respectively), but they both differ in the total extension of the clitellum, body size, disposition of their chaetae (closely paired vs separate) and number of seminal vesicles (two pairs in segments 11, 12 vs 3 pairs in segments 9, 11, 12). Cataladrilus porquerollensis Marchán & Decäens sp. nov shows a remarkable external morphological similarity with Allolobophora festae rosa, 1892 (sometimes attributed to genus Prosellodrilus), however, they can be easily distinguished by the position of their spermathecae (9/10, 10/11 vs 12/13, 13/14).  Qiu & Bouché, 1998 according to their similar position of clitellum and tubercula pubertatis (Table 4). however, it differs from both species by the exact position of clitellum and tubercula pubertatis, as well as being slightly smaller and less pigmented. Scherotheca portcrosana Marchán & Decäens sp. nov. also shows a remarkable resemblance to Eumenescolex simplex zicsi, 1981 (initially described as Sc. corsicana simplex then transferred due to its single pair of spermathecae), but they clearly differ in the number and position of their spermathecae (in intersegments 12/13, 13/14 vs 13/14). This similarity highlights the importance of including representatives of Eumenescolex Qiu & Bouché, 1998 in a molecular phylogenetic framework to test the close relationship among both genera that was previously suggested by Qiu & Bouché (1998d). Qiu & Bouché (1998d) elevated 9 subspecies of Scherotheca to species while maintaining or creating others. genetic distances within the clade comprising Sc. gigas gigas from Montpellier and Sc. gigas heraultensis and Sc. gigas mifuga are way below the usual threshold established for divergence between conspecific taxa (less than 10% for COI, Chang et al. 2009), supporting the subspecific status of these taxa; while Sc. rhodana (described as Sc. gigas rhodana) appears to be as divergent from Sc. gigas gigas as from Sc. monspessulensis idica or other species of the genus (with genetic distances above the aforementioned threshold). Thus, the morphological criteria followed by Qiu & Bouché (1998d) appear supported by molecular data. However, larger sample sizes would be necessary to confirm this statement. These findings suggest that molecular markers are suitable for delimiting species and subspecies within Scherotheca. If this approach performs consistently across different animal taxa, it may be advisable to implement it widely in order to avoid the dismissal of unrecognized species-level taxa in ecological studies and biodiversity assessment (for different examples see rutgers et al. 2016).

Cataladrilus porquerollensis
Scherotheca cf. gigas (from Domínguez et al. 2015) collected from Navarra appear to be unrelated to Sc. gigas gigas from Montpellier, but closely related to Sc. savignyi. These earthworms may belong to one of the former subspecies of Sc. gigas such as Scherotheca aquitania Bouché, 1972, which is relatively common in Northeastern Spain (Qiu & Bouché 1998d). preliminary re-inspection of these specimens showed external morphological characters compatible with Sc. aquitania. Thus, the inclusion of French specimens of Sc. aquitania (ideally from the type locality) would enable confirmation of this hypothesis. This is just one of the many points highlighting the need for a comprehensive molecular phylogeny of the genus Scherotheca. robust systematics of the many species of this genus in northeastern Spainsouthern France-Corsica-Italy would contribute to the following: a) a better understanding of the impact on ecosystems of these large, deep burrowing anecic earthworms, which have been found to be important in both agricultural and unmanaged habitats (gavinelli et al. 2018), b) assessment of the conservation status of the genus, with several taxa being narrowly restricted and potentially vulnerable to human impacts (rida & Bouché 1995) and c) insight into the historical biogeography of native earthworm families in this geologically complex area.

Phylogeographic implications for Cataladrilus and Scherotheca
The discovery of a species of Cataladrilus in the Hyères Archipelago is puzzling from a phylogeographic point of view. The main range of Cataladrilus is restricted to Catalonia (Spain) and Andorra, with two species in neighboring locations in France separated from the Hyères Archipelago by 340 km. This disjunct distribution could be explained by a failure to find closely related species in the geographic gap, which would suggest that southeastern France could still harbor other undescribed species. This, however, seems quite unlikely if considering the intense sampling already done by Marcel Bouché at the end of the past century (Bouché 1972). Alternatively, Ca. porquerollensis Marchán & Decäens sp. nov. may be a relict from a formerly wider distribution. This would be consistent with paleogeographic events in the area: the Eastern pyrenees and provence were connected through an exposed gulf of Lion up to Late oligocene-Early Miocene (around 24 Ma) when a wide marine ingression separated them up to the present (except for the duration of the Messinian Salinity Crisis) (Sissingh 2006). The possibility of Cataladrilus and Prosellodrilus being synonyms (as the lack of reciprocal monophyly may suggest) would not change the observed disjunction, as the closest species of Prosellodrilus (Pr. tikalus Qiu & Bouché, 1998) is found 225 km to the west of Porquerolles.
The closer relationship between Sc. portcrosana Marchán & Decäens sp. nov. and the Corsican Scherotheca than with French mainland Scherotheca is also unexpected. however, several species of plants, arthropods and vertebrates are common to the Hyères Archipelago and Corsica (Médail et al. 2013). Scherotheca portcrosana Marchán & Decäens sp. nov. may have diverged from its relatives before or during the rifting and drifting of the Corso-Sardinian microplate around 24 Ma (oudet et al. 2010). Corsica remained connected to the French mainland through the rotation to its current position (Sissingh 2006), which would have allowed later migration to the Hyères Archipelago of a common ancestor of Sc. portcrosana Marchán & Decäens sp. nov. and Corsican Scherotheca. however, under this second scenario, one would expect to find other relatives in eastern Provence (where the land bridge was supposed to join both areas).
The area of provence has been strongly disturbed by human activities, especially in the coastal area, while the small islands like Porquerolles and Port-Cros remained relatively well preserved. It is possible that these islands acted as refugia while continental relatives got extinct or relegated to relict habitats. This phenomenon has already been suggested to explain the distribution of the gecko Euleptes europaea gené, 1839, which is frequent in Corsica and Sardinia and rare in the coastal islands of Provence and Liguria, probably after a range retraction in the mainland (Delauguerre et al. 2011). Further survey focusing on relic natural ecosystems in the littoral area (Cap Lardier) or further into the mainland (i.e., Sainte Baume, plaine des Maures, Massif de l'Esterel) could be performed to test this hypothesis.
Whatever explicative hypothesis is retained, these disjunct distributions, as previously observed in other earthworm genera (pérez Losada et al. 2011;Domínguez et al. 2018;Marchán et al. 2018), highlight the strong connection between paleogeographic events and earthworm evolution and divergence. A more robust approach to time-calibrated phylogenies in earthworms (hindered by the lack of body fossils) and the integration of different paleogeographic reconstructions could illuminate both the origin of palearctic earthworms (hormogastridae, Lumbricidae) and the geological history of the Western Mediterranean terranes from the Late Cretaceous to the Neogene.