From morphology to molecules: a combined source approach to untangle the taxonomy of Clessinia (Gastropoda, Odontostomidae), endemic land snails from the Dry Chaco ecoregion

Background Land gastropods of the Dry Chaco merit special attention because they comprise a highly diverse but barely studied group. Clessinia Doering, 1875 are typical inhabitants of this ecoregion. The inclusion of their distribution areas into Spixia range, their shell shape similarities, and a former molecular study raised doubts on the monophyly of this genus. The present study review the species of Clessinia, under a morphological, geometric morphometrics, and molecular combined approach. Methods Adults were collected, photographed, measured, and dissected for anatomical studies. Shell ultrastructure was studied with scanning electron microscope. Geometric morphometric analyses on shells were performed testing if they gave complementary information to anatomy. Two mitochondrial genes, and a nuclear region were studied. Phylogenetic reconstructions to explore the relationships of DNA sequences here obtained to those of Clessinia and Spixia species from GenBank were performed. Results Species description on shell, periostracal ornamentation and anatomy is provided. We raised former Clessinia cordovana striata to species rank, naming it as Clessinia tulumbensis sp. nov. The periostracum, consisting of hairs and lamellae, has taxonomic importance for species identification. Shell morphometric analyses, inner sculpture of penis and proportion of the epiphallus and penis, were useful tools to species identification. Nuclear markers do not exhibit enough genetic variation to determine species relationships. Based on the mitochondrial markers, genetic distances among Clessinia species were greater than 10%, and while C. cordovana, C. nattkemperi, and C. pagoda were recognized as distinct evolutionary genetic species, the distinction between C. stelzneri and C. tulumbensis sp. nov. was not evident. Clessinia and Spixia were paraphyletic in the molecular phylogenetic analyses. Species of Clessinia here treated have narrow distributional areas and are endemic to the Chaco Serrano subecoregion, restricted to small patches within the Dry Chaco. Clessinia and Spixia are synonymous, and the valid name of the taxon should be Clessinia Doering, 1875 which has priority over Spixia Pilsbry & Vanatta, 1894. Discussion Our results support the composition of C. cordovana complex by three species, C. cordovana, C. stelzneri, and C. tulumbensis sp. nov. The low genetic divergence between C. stelzneri and C. tulumbensis sp. nov. suggests that they have evolved relatively recently. The former Spixia and Clessinia are externally distinguished because Clessinia has a detached aperture from the body whorl forming a cornet, periostracal microsculpture extended over dorsal portion of the peristome, five inner teeth on the shell aperture instead of three–four found in Spixia. Morphological similarities exists between both genera in shell shape, type of periostracum microsculpture, reproductive anatomy, besides the overlap in geographic ranges.


INTRODUCTION
Taxonomy is a crucial discipline in biology if practiced within an evolutionary framework (Dubois, 2017). The taxonomic and biodiversity crisis requires a strong acceleration of the work of exploration, study, description, and naming of the species of the globe (Wheeler, Raven & Wilson, 2004;Dubois, 2007Dubois, , 2010. However, there is a tendency towards a strong decrease in morphological and anatomical studies while "replacing" them with molecular analyses which are unable, if are used alone, to provide the wealth of diverse information on organisms which morphology, anatomy, and other biological studies offer (Dubois, 2017). Species are hypotheses, and as such it is required that they make predictions (that more data of approximately the same quality will support such groupings) and are thereby testable (that more data of approximately the same quality do not suggest alternative groupings) (Wheeler, 2004;Valdecasas, Williams & Wheeler, 2008). Then, identification of species utilized in a study impacts all subsequent comparisons or any further studies on species-specific traits or attributes.
The combination of morphological and ecological information with different molecular markers can be a good method of species identification, because it can provide an accurate perspective on evolutionary history of an organism and its taxonomic relationships (Davison, Blackie & Scothern, 2009). In this way, new methods do not replace, but complement the traditional, tested methods, and procedures (Wheeler, Raven & Wilson, 2004). Geometric morphometrics is a useful tool to accurately analyze shell variability decomposing shell form into size and shape in each species (Carvajal-Rodriguez, Conde-Padin & Rolan-Alvarez, 2005;Cruz, Pante & Rohlf, 2012;Greve et al., 2012). When multiple sources are used for analyzing a taxonomic problem, agreement among them is expected, but differences between them can also be rich revealing different aspects of a same problem and contributing to interpretation of the evolutionary patterns.

Distribution
Species distribution is based on point records (geographical coordinates) of species occurrences obtained through field work in the Córdoba and Catamarca provinces, Argentina between 2006 and 2017. All specimens collected were deposited in the Instituto de Biodiversidad Neotropical (IBN), Instituto-Fundación Miguel Lillo (IFML-MOLL), Tucumán, Argentina and the malacological collection at the Instituto de Biología Subtropical, Misiones, Argentina (IBS-Ma). Additionally, we examined other specimens and obtained records from the following collections: IBN; IFML-Moll; MACN-In, Museo Argentino de Ciencias Naturales "Bernardino Rivadavia," Buenos Aires, Argentina (Table S1). This information was used to digitize geographical range maps and depict the extent of occurrence of each species by using QGis 2.18 (Quantum GIS Development Team, 2009; http://qgis.osgeo.org). Shapefiles layers corresponding to administrative areas of Argentina were obtained from DIVA resources (http://www.diva-gis.org/gdata) and Instituto Geográfico Nacional (http://www.ign.gob.ar/sig). Classification of Argentinean ecoregions follows Olson et al. (2001), but ecoregions and subecoregions shapefiles were obtained from ProYungas Fundation (http://siga.proyungas.org.ar/recursos).

Morphological studies
The different zones in which the shell aperture is divided: basal, palatal (divided in upper and lower zones for internal teeth) and parietal, are the same as used by Solem (1966). Differences in terminology between a tooth and a lamella follow Cuezzo (2003). Anatomical information was obtained by dissecting 10 adult specimens per species under a Leica MZ6 stereoscope; dissected parts were illustrated with the aid of a camera lucida. Terminology for anatomical descriptions follows Tompa (1984). Terms proximal and distal refers to the position of an organ or part of an organ in relation to the gamete flow from ovotestis (proximal) to genital pore (distal), as in previous works (Cuezzo, 1997(Cuezzo, , 2006. The limit between epiphallus and penis is based on the sculpture of their inner wall. Radula, jaw and shell were observed and photographed with a SEM Zeiss Supra 55VP at the Integral Center of Electron Microscopy (CIME) of the National University of Tucumán, Argentina. The terms Diagnosis and Definition for the species description are used as established in the glossary of the International Code of Zoological Nomenclature (http://www.iczn.org).

Morphometrics
Traditional linear shell measurements were taken from specimens of each species according to availability (Fig. 1). The number of whorls was calculated following Kerney & Cameron (1979). Descriptors of measurements and proportions (mean, standard deviation, and range) were also calculated in each case. Measurements of type material of each species is recorded in the species description with the following arrange: maximum-minimum (mean) of each measurement.

Geometric morphometrics
This study was performed to quantitatively analyze the relationship between shape and size of the species of Clessinia, testing if they gave complementary information to our anatomical observations or differed from them. On these grounds, the geometric morphometric analysis was performed with 15-61 specimens per species according to their availability, totalizing 144 specimens used (Table 1). Specimens were taken from different populations ranging the whole species distribution area. Images of shell in ventral view of adult specimens were converted to TPS format with TpsUtil 1.68 (Rohlf, 2016a). Shell landmarks, discrete anatomical loci that are homologous in all individuals in the analysis, expressed by coordinates, were chosen in each case. Landmarks were located on the same shell whorl number so that comparisons among them were possible, even when shells have different whorl numbers. A total of 14 landmarks from ventral view were digitized by means of the TpsDig2 2.26 program (Rohlf, 2016b;Figs. 1E-1F). Landmarks selected in ventral view represent the general shell shape features such as body whorl, spire and aperture. A second analysis was performed using only those species of the cordovana-group from nine geographic localities to enhance the possible differences among them. Finally, another morphometric analysis was performed using six landmarks in lateral shell side (Fig. 1F) to test if the degree of detach of the aperture was significative for species delimitation. The morphometrics analyses were performed with MorphoJ 1.06d (Klingenberg, 2011). The shape symmetric components associated with position, rotation, translation, and size were removed using the Procrustes fit. A multivariate regression of the Procrustes coordinates against logarithm of centroid size, defined as square root of the sum of the squared distances of each landmark to the centroid of the landmark configuration (Bookstein, 1991), was performed to asses allometric effects (i.e., if shell shape variation is correlated with size). A permutation test was also performed with 10,000 rounds to evaluate the independence among the variables. Variation in the shell shape was examined using canonical variate analysis (CVA).
DNA extraction, polymerase chain reaction amplification, and DNA sequencing Total DNA was extracted from three mm 3 samples of foot muscle of ethanol-preserved specimens by means of a cetyltrimethylammonium bromide protocol (Beltramino et al., 2018). We selected 16 samples belonging to Clessinia and Spixia species and the outgroup species Plagiodontes daedaleus. Collection information and GenBank accession numbers for the samples analyzed are presented in Table 2. Partial sequences of the mitochondrial 16S-rRNA and the cytochrome oxidase subunit I (COI) genes, and a nuclear region including the 3′ end of the 5.8S-rRNA gene, the complete ITS-2 region, and the 5′ end of 28S-rRNA gene (hereafter referred to as ITS-2) were amplified by means of the primers 16SF-104 (5′-GAC TGT GCT AAG GTA GCA TAA T-3′) and 16SR-472 (5′-TCG  -GGT CAA CAA ATC ATA AAG ATA TTG G-3′) and HCO2198  (5′-TAA ACT TCA GGG TGA CCA AAA AAT CA-3′) for COI (Folmer et al., 1994), and LSU-1 (5′-CTA GCT GCG AGA ATT AAT GTG A-3′) and LSU-3 (5′-ACT TTC CCT CAC GGT ACT TG-3′) for the ITS-2 (Wade & Mordan, 2000). The amplification of the 16S-rRNA gene was performed as in Rumi, Vogler & Beltramino (2017) in a T21 thermocycler (Ivema Desarrollos). The amplification of the COI gene was conducted following Vogler et al. (2014) and run on a T18 thermocycler (Ivema Desarrollos). The amplification of the ITS-2 region was performed in a total volume of 50 ml containing 30-50 ng of template DNA, each primer at 0.25 mM, 1X reaction buffer, 0.2 mM dNTPs, 2.5 mM MgCl 2 , and 2 U Taq Pegasus DNA polymerase (Productos Bio-Lógicos, Bernal, Argentina). The amplification was conducted in a T18 thermocycler as follows: after an initial denaturing for 3 min at 94 C; 35 cycles of 1 min at 94 C, 1 min at 50 C, 1 min at 72 C were performed; followed by a final extension at 72 C for 5 min. The success of polymerase chain reactions (PCRs) was verified by agarose gel electrophoresis. The PCR products were purified by means of an AccuPrep PCR Purification Kit (Bioneer, Daejeon, Korea). Following purification, both DNA strands for each gene were then directly cycle-sequenced (Macrogen Inc., Seoul, South Korea). The resulting sequences were trimmed to remove the primers, and the consensus sequences between forward and reverse sequencing were obtained by means of the BioEdit 7.2.5 software (Hall, 1999). For S. minor, the repeated attempts to amplify the COI and ITS-2 regions were unsuccessful, and for this species only the 16S-rRNA was included in further analyses.

Sequence data, phylogenetic analyses, and molecular species delimitation
The sequence alignment of the 16S-rRNA gene was performed with MATFF 7 via the MATFF web-server (https://mafft.cbrc.jp/alignment/server/; Katoh, Rozewicki & Yamada, 2017); the COI and ITS-2 alignments were performed with Clustal X 2.1 (Larkin et al., 2007). Genetic distances among the Clessinia and Spixia species were investigated in MEGA X software (Kumar et al., 2018) using the number of differences (p) and the Kimura's two-parameter (K2P) substitution model. Phylogenetic analyses were performed using maximum likelihood (ML), and Bayesian inference (BI). For both analyses, the COI and 16S-rRNA datasets were concatenated to improve the resolution of phylogenetic reconstructions. The total length of the analyzed matrix was 992 bp. In addition, COI-based phylogenetic reconstructions were performed to explore the phylogenetic relationships of the DNA sequences here obtained to those of other Clessinia and Spixia species from various locations available in GenBank ( Table 2). The total length of this matrix was 655 bp. We also obtained phylogenetic trees for the nuclear region as an independent marker based on an 832 bp matrix. In all phylogenetic reconstructions, Plagiodontes daedaleus was used as outgroup species, with Cerion incanum (Leidy, 1851) used as an additional outgroup for the mitochondrial DNA sequence data. Sequences of C. incanum were extracted from the complete mitochondrial genome for the species (KM365085; González et al., 2016).
The ML analysis was conducted with PhyML 3.0 (Guindon et al., 2010) available via the ATGC bioinformatics platform (http://www.atgc-montpellier.fr/) with the Nearest-Neighbor Interchange branch swapping algorithm. Substitution models were selected using the SMS program (Lefort, Longueville & Gascuel, 2017) according to Akaike Information Criterion: GTR+I+G model for the concatenated dataset and the COI alignment that included GenBank sequences, and GTR+G for the ITS-2 sequences. Nodal support values were computed by bootstrapping with 1,000 replicates (Felsenstein, 1985). The BI was conducted in MrBayes 3.2.6 (Ronquist et al., 2012) with the same substitution models used in the ML analyses, as identified in jModelTest 2.1.7 (Darriba et al., 2012) by means of the corrected Akaike Information Criterion. Two runs were performed simultaneously with four Markov chains for 2 million generations, sampling every 200 generations. The first 1,001 samples of each run were discarded as burn-in, and the remaining 18,000 trees were used to estimate posterior probabilities.
The Automatic Barcode Gap Discovery (ABGD) method, which clusters sequences in putative species based on differences between intraspecific and interspecific distance variation (Puillandre et al., 2012) was used to explore species boundaries in the concatenated dataset, and the larger COI dataset including sequences from GenBank. These aligned datasets (excluding the outgroups) were analyzed via the ABGD web-server (http://wwwabi.snv.jussieu.fr/public/abgd/) using the K2P model (Vogler et al., 2016). The minimum relative gap width was set to 0.5, and the default range of prior values for maximum divergence of intraspecific diversity (p) from 0.001 to 0.1 was used. In addition, the K/h method was used to assess the status of Clessinia species under the evolutionary genetic species concept (EGSC) (Birky et al., 2010;Birky, 2013). This method is based on basic coalescent theory and requires a phylogenetic tree as well as distance matrices to estimate the mean genetic differences within (h) and between clades (K), in order to identify clades that are diverged enough to be considered separate species (Birky, 2013;Restrepo et al., 2014;Fontaneto, Flot & Tang, 2015). The K/h method was performed on the concatenated dataset following Schön et al. (2012) and Birky (2013). Those clades with K/h ratios 4 were considered to represent sequences that come from different evolutionary species with probability 0.95 (Birky, 2013 and references therein). Mean pairwise differences between clades were estimated in MEGA X.

Nomenclatural acts
The electronic version of this article in portable document format will represent a published work according to the International Commission on Zoological Nomenclature (ICZN), and hence the new names contained in the electronic version are effectively published under that Code from the electronic edition alone. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix http://zoobank.org/. The LSID for this publication is: urn:lsid:zoobank. org:pub:8DB0CC34-AE26-44BA-B7F8-A5F17254BD13. The online version of this work is archived and available from the following digital repositories: PeerJ, PubMed Central, and CLOCKSS.

Periostracal ornamentation
Periostracal structures are particularly well developed in Clessinia, consisting in hairs of different lengths and densities, spines and rounded to quadrate lamellae. These structures were useful tools for species recognition. Both C. cordovana and C. stelzneri show periostracal hairs on the teleoconch surface, being notably longer among C. cordovana specimens, with more prominent hairs in specimens from Sierra de Pocho area in Córdoba. Periostracal hairs are shorter and more densely arranged in C. stelzneri. Teleoconch surface is traversed by periostracal spiral rows in the three species of the cordovana-group, with a greater number of minor spiral rows between the major hair bearing rows in C. stelzneri. In C. tulumbensis sp. nov. the periostracal hairs are absent and the spiral rows are more scatter. In Clessinia pagoda periostracal structures consist of spiral rows bearing rounded to quadrate lamellae slightly imposed over each other. In Clessinia nattkemperi lamellae are spine-shaped with wider bases almost as a triangle. All Clessinia species have an interesting pattern of periostracal microsculpture in the space between spiral rows which is traversed by axial irregular microfolds cut by spiral or diagonal microribs forming an irregular net.

Anatomy
Anatomical information obtained on the pallial, digestive and reproductive systems of each species is described in the taxonomic section.

Morphometrics
We extracted meaningful measurement differences among taxa and present these in a summary table (Table 1). In the geometric morphometric analysis performed to evaluate shell shape differences in ventral view among all Clessinia species ( Fig. 2A), allometric relationships between shape and size was registered (4.94% of the total amount of shape variation; p < 0.001). The shell shape variation among the five taxa considered was successfully discriminated using CVA of the residuals from the regression of shape on centroid size. On the canonical axis 1 (CV1) (captures 74.41% of the total shell shape variation), the main changes in shell shape are associated with the expansion of the base of spire and body whorl. Specimens of C. cordovana with high scores on CV1 have thinner whorls, whereas specimens of C. nattkemperi and C. pagoda with low scores, have both spire and body whorl more expanded. It also indicates that when shells are thinner they are also taller while shells more expanded are less tall. On CV2 axis (captures 11.56% of the total shell shape variation) the main shell shape variation referred to the shape of the aperture and the degree of inclination of the suture before the aperture (landmarks 6 and 10, Fig. 1E). High scores in specimens of C. nattkemperi indicate a marked expansion in the central portion of the aperture. Specimens of C. pagoda showing low scores exhibit an oval shaped aperture, while C. tulumbensis sp. nov., C. stelzneri, and C. cordovana have intermediate forms of aperture between C. pagoda and C. nattkemperi with a higher inclination of the suture. A second analysis was performed to evaluate shell shape differences in ventral view using specimens of the cordovana species-group alone (Fig. 2B). As a result, allometric relationships between shape and size was registered (12.64% of the total amount of shape variation; p < 0.0001). Residuals from the regression of shape on centroid size were used in the analysis. On the CV1 (captures 57.49% of the total shell shape variation), the main changes in shell shape are associated with the expansion of the base of spire and body whorl. Specimens of C. cordovana from Cerro de la Cruz and the area surrounding San Marcos Sierras, plus specimens of C. tulumbensis sp. nov. from Virgen de Fatima, Route 16 and Cerro Colorado have high scores on CV1, showing thinner whorls. Specimens of C. stelzneri from Cerro San Vicente, C. tulumbensis sp. nov. from route between Dean Funes and Tulumba, Tulumba and San José de la Dormida and specimens of C. cordovana from Sierra de Pocho have low scores, showing whorls of the spire and body whorl more expanded. On CV2 axis (captures 23.66% of the total shell shape variation) the main shell shape variation is related with the shape of the aperture and the expansion of the first whorl of the shell. High scores in specimens of C. tulumbensis sp. nov. from all localities considered, except Cerro Colorado, indicate wider first whorls of the shell and a thinner aperture. Specimens of C. stelzneri and C. cordovana, with low scores, exhibit thinner fist whorl of the shell and more expanded central portion of the aperture. C. tulumbensis sp. nov. from Cerro Colorado have intermediate forms between both previous described groups. Analysis using landmarks in shell lateral views did not show significative differences among the species (Fig. S1).

Molecular analyses
Sequence data, phylogenetic analyses, and molecular species delimitation We successfully amplified both mitochondrial loci and the nuclear region in the majority of Clessinia and Spixia specimens, except for S. minor in which amplification of the COI and ITS-2 markers was not possible. Partial 16S-rRNA sequences ranged between 287 and 295 bp, COI sequences consisted of 655 bp, and ITS-2 sequences were of 822 bp in length for all individuals. The ITS-2 region showed no sequence variation within each species and exhibited little genetic differentiation among species (Tables 3 and 4). Phylogenetic reconstructions obtained with the nuclear marker were unresolved (Fig. S2). For the mitochondrial markers, ML and BI results revealed congruent topologies; consequently, we reported only the BI tree. From the analyses of the concatenated dataset, Clessinia stelzneri clustered with C. tulumbensis sp. nov.; similarly, S. cuezzoae clustered with specimens of Clessinia pagoda, and this group clustered with S. holmbergi. C. cordovana clustered with the group formed by these three species (Fig. 3). Therefore, these trees did not support the monophyly of Clessinia (Fig. 3). The phylogenetic trees inferred from the larger COI dataset including sequences from GenBank congruently identify roughly the same major groups, with both genera being paraphyletic  due to association of Clessinia and Spixia specimens in well-supported arrangements, as shown by the relationships between C. nattkemperi and S. tucumanensis (Parodiz, 1941) or S. cuezzoae and C. pagoda (Fig. 4). Sequence divergence for the mitochondrial loci amongst the species are presented in Tables 5 and 6. By using the concatenated dataset, the ABGD approach recovered six candidate species based on the distribution of the pairwise genetic distances with a maximum prior of Table 4 Genetic distances in ITS-2 sequences among Clessinia and Spixia species.

Species
GenBank No. * ID 1 2 3 4 5 6 7 8 9 10 11 12 13 14 C. nattkemperi  intraspecific divergence of 0.035938 (Fig. 3). The same as the ABGD, the K/h method provided support for six of the morphospecies to be considered different evolutionary genetic species (Table 7), except for C. stelzneri and C. tulumbensis sp. nov. which were not supported as distinct genetic species by either method. Based on the COI dataset, the ABGD analysis clustered sequences into 10 stable putative species based on the distribution of the pairwise genetic distances with a maximum prior of intraspecific divergence of 0.035938. The species C. stelzneri, C. tulumbensis sp. nov., and S. popana (Doering, [1877a) were clustered within the same group. All the remaining species, that is, C. nattkemperi, S. tucumanensis, S. cuezzoae, C. pagoda, S. philippii , S. holmbergi, S. pervarians (Haas, 1936), C. cordovana, and C. gracilis Hylton Scott, 1966 were assigned to different candidate species (Fig. 4). As a result of the anatomical studies performed, shell periostracum observations, shell geometric morphometrics and genetic analyses, and based on previous findings (Breure & Romero, 2012), we here synonymized the genus Clessinia and Spixia and according to the principle of priority (ICZN Code, Art.23.1) the valid name of the taxon should be Clessinia Doering, 1875 which has priority over Spixia Pilsbry & Vanatta, 1894. In the following, we provide the taxonomic description and new systematic arrangement of the treated species.  Type species. Clessinia stelzneri .
Definition. Shell fusiform to turritelliform. Protoconch with delicate axial ribs and spiral bands delimited by thin grooves. Shell with periostracal complex structures consisting on spiral rows bearing "hairs" or triangular, rectangular to quadrate lamellae. Few species lacking periostracal ornamentation. Last portion of body whorl with aperture with slightly reflexed peristome, some species forming a cornet detached from rest of shell body whorl with peristome thin and expanded. Body whorl microsculpture complex, consisting

Notes:
Uncorrected (below the diagonal) and corrected (K2P; above the diagonal) distances are shown.
in microfolds forming an irregular net, which in some cases is expanded dorsally over shell cornet. Fourth to five inner apertural teeth forming a complex apertural barrier, except for one species with three teeth. Dorsal portion of shell body whorl with a medial marked notch corresponding to the basal lamella. Columellar lamella undulating, in some species L-shaped. Presence of a short penial sheath overlapping distal portion of the penis. Insertion of penial muscle at proximal penis or distal epiphallus. Vas deferens thin, running freely along penis and attached to penial retractor muscle.
Diagnosis. Clessinia is one of the odontostomid groups showing most complex apertural teeth arrangements. Number of apertural teeth/lamellae ranges from three to five. Together with Plagiodontes (type species = Helix dentata Wood, 1828) has a similar protoconch sculpture consisting on axial ribs and transversal grooves between ribs. It differs from Plagiodontes in general shell shape, showing thinner and taller spires and larger numbers of whorls plus strong differences in number of apertural teeth/lamellae. Species here redescribed plus the new species show a shell aperture detached from the body whorl and this character is not observed in the remaining species of Clessinia (former Spixia) and in no other odontostomid genus. Shell aperture shape varies from subcircular to subquadrate. Clessinia is found in dry habitats with its distribution area ranging from Argentina, Uruguay to Bolivia and Paraguay. Clessinia differs from Cyclodontina (type species = Clausilia pupoides Spix, 1827) (Cowie, Cazzaniga & Glaubrecht, 2004) in having more apertural teeth/lamellae, some species of Cyclodontina are even toothless. Shells of Cyclodontina are basally wider, sometimes glossy, without any particular ultrastructural shell sculpture described for this genus. The shell aperture in Cyclodontina is not detached from the body whorl as in some Clessinia species. Cyclodontina is distributed in Bolivia, Paraguay, Argentina, Uruguay, and Brazil. Clessinia differs from Pilsbrylia (type species = Pilsbrylia paradoxa Hylton Scott, 1952), in the general shell shape because Pilsbrylia species have a fusiform, broader shell shape, and the presence of only two apertural teeth. On the contrary to Clessinia, Pilsbrylia species inhabit in humid forest.
Habitat preferences. Species of Clessinia inhabit dry areas where rocky formations are frequently found among low xerophytic vegetation. Few species occur in Yungas ecoregion, but in transition zones with dryer forests. They usually live below rocks in contact to the ground, in rock crevices, or buried in soil under shrubs. Some species can be found glued to leaves in bushes. Clessinia nattkemperi is usually found attached to the surface of cactuses or under dead cactus in contact with soil.

Species distribution (Figs. 5A-5D)
Clessinia species here treated are distributed in the Pampean Sierras of Central Argentina, in the portion corresponding to the provinces of Córdoba and Catamarca (Fig. 5A). These Sierras form a mountain complex of about 300 km 2 in extent with a direction of north to south and consist in a series of parallel mountain ranges. An extended depression of salty surface called Salinas Grandes, located between northern Córdoba, southeastern Catamarca and La Rioja provinces, and Salinas de Ambargasta between southern Santiago del Estero and northwestern Córdoba, subdivide the Pampean Sierras forming a real ecological barrier for land snail dispersion (Figs. 5B and 5D). Main mountain systems in Córdoba are the Sierra Chica, Sierra Grande and Sierra de Comechingones, this last is extended to San Luis province. Clessinia is mainly distributed around and to the north of the Sierra Chica, including minor mountains such as Sierras de Ischilin, Higuerita, Copacabana, and Massa in Córdoba. Also scatter occurrences have been registered to the southwest in Sierra de Pocho. In Catamarca, occurrences are registered in the Ambato and Esquiu departments, both also corresponding to the

Description
External features (Figs. 6A-6C): Body dark to light grey with two blackish pigmented, longitudinal bands extending from the mantle collar to the tentacles. Black tentacles. Foot short, light gray, with a blunt end.
Shell (Figs. 6A-6L and 7): Turritelliform to subfusiform, comprising 8 ½ to 9 ½ slightly convex whorls. Coloration pale to dark brown, uniform (Figs. 6A-6I). Protoconch with axial, regularly arranged strength ribs, and thin spiral parallel bands delimited by spiral grooves between ribs (Figs. 7A and 7B). Teleoconch with axial, oblique, shallow thin costules separated by regular spaces. Surface of the teleoconch traversed by spiral rows bearing two types of periostracal hairs (Figs. 7C-7F). Spiral rows bearing long hairs of 200-300 mm (mean = 227, n = 8) intercalated with two to three spiral rows, one of each bearing hairs wider at base and less tall (Fig. 7E). Departing from each spiral row, interconnected axial microfolds giving the appearance of an irregular net. Suture deeply impressed. Distal portion of body whorl detached from rest of the shell forming a cornet (Figs. 6D, 6E, 6G, 6H and 7C). Aperture suboval, round to square with thin, continuous, expanded peristome. Five inner lamellae in the aperture not connecting to the peristome (Figs. 6J-6L). Upper columellar lamella long, straight, spirally following the columellar axis. Lower columellar lamella running parallel, slightly undulating, spirally following columellar axis (Fig. 6M). Basal teeth straight, short, to the left of the aperture producing a groove on dorsal side of the shell (Fig. 6F). Some specimens with dorsal groove not marked (Fig. 6N). Upper palatal teeth small, generally triangular shaped. Lower palatal teeth short. Both palatal teeth perpendicular to columellar axis, deeply located inside cornet. Dorsal side of the aperture with an inner marked groove (Figs. 6J-6L). Umbilicus narrow. Shell measurements represented in Table 1. Jaw (Fig. 8A): Wide horseshoe shaped. Ten plaques with a triangular central one subdivided into three subplaques, the middle one more triangular-shaped. Five lateral quadrangular to rectangular shaped plaques at both sides of the central one. Lateral plaques slightly increasing their size toward the tip of the horseshoe. Each plaque traversed by several thin transversal grooves.
Pallial system: Pulmonary roof thin and long traversed by few veins mostly concentrated on distal portion. Kidney triangular, short, of a quarter of the total length of the pulmonary roof. Secondary ureter closed over most of its length, opening slightly before rectum. Pallial gland thin, parallel to mantle collar. Afferent vein parallel to main pulmonary vein.
Reproductive system (Figs. 9A-9E): Ovotestis embedded into digestive gland within the fourth and fifth shell whorls. Hermaphroditic duct inserting in distal portion of albumen gland (Fig. 9A). Seminal receptacle swollen. Fertilization pouch-spermathecal complex long, digitiform broaden at its base. Bursa copulatrix with sac rounded, longer than spermoviduct reaching the albumen gland. External limits between epiphallus and penis not evident (Figs. 9A and 9B). Penis cylindrical, long, with a short penis sheath overlapping in part distal penis. Inner morphology of the penis divided into three areas marked by differential pattern of sculpture (Figs. 9C and 9D). Proximal portion with same diameter than resting portions, inner sculpture with tightly appressed pustules (Fig. 9E). Penial papilla absent. Penis medial sector long, cylindrical, inner wall with rhomboidal to hexagonal pustules covering the surface, without pilasters (Figs. 9D and 9E). Distal penis short, inner sculpture consisting in three to four longitudinal, straight, thin pilasters, parallel to each other (Figs. 9C and 9D). Penial retractor muscle short and thick, inserting in penis proximal portion. Epiphallus ¼ of penial length. Flagellum thinner than epiphallus and ½ epiphallus length. Vas deferens thin, running freely along penis, attached to penial retractor muscle, then free along epiphallus and inserting between flagellum and epiphallus. Vagina cylindrical, with a distal portion thinner in diameter than the proximal, inner wall with longitudinal pilasters. Vagina about double in length of the distal portion of penis. Punilla, Pocho, and Tulumba departments. Clessinia gracilis was described in 1966 from a single shell found in La Puerta, Ambato department, Catamarca province, and was synonymized to C. cordovana (Cuezzo, Miranda & Ovando, 2013) because the holotype has same size and shape as C. cordovana. However, during different collecting trips to the area of La Puerta carried out during summer in different years, specimens were not found. This species inhabits the Dry Chaco ecoregion, Chaco Serrano subecoregion.    (2), SMF 325584 (4).

Description
External features (Fig. 10A): Body light brown. Foot short with a blunt end. Some specimens with sole lighter than dorsal body coloration.
Shell (Figs. 10B-10F and 11A-11D): Fusiform, comprising 8 ½ to 9 slightly convex whorls. Coloration pale to dark brown, sometimes with longitudinal strips clearer in color, other shells with uniform coloration (Figs. 10B-10D). Protoconch with axial, regularly arranged strength ribs, and thin spiral parallel bands delimited by spiral grooves (Fig. 11A). Teleoconch with shallow axial costules separated by regular spaces. Surface of the teleoconch traversed by densely arranged spiral rows bearing two types of periostracal hairs (Figs. 11B-11D). Rows of tall hairs intercalated with three to five spiral rows some of which not bearing hairs while other bearing hairs triangular shaped, less tall, usually in touch with each other through their bases (Figs. 11C and 11D). Ultrastructural ornamentation of body whorl extending over dorsal portion of the cornet. Last portion of the body whorl detached ending into a cornet. Aperture subcircular with five inner teeth and lamellae not connecting to the peristome (Figs. 10B and 10E). Upper columellar lamella long, straight, spirally following the columellar axis. Lower columellar lamella running parallel, slightly undulating, spirally following columellar axis (Fig. 10F). Basal lamella straight, short. Dorsally the body whorl shows a deep groove produced by the basal lamella. Shell measurements represented in Table 1. Jaw (Fig. 8B): Markedly horseshoe shaped. Nine plaques with a triangular central one subdivided into three triangular subplaques. Four lateral rectangular shaped plaques at both sides of the central one. Lateral plaques strongly increasing their size toward the tip of the horseshoe. Each plaque traversed by several transversal grooves.
Radula (Figs. 8F-8I): Radular teeth transversally arranged on a straight line. Central tooth tricuspid, with mesocone triangular to rhomboidal. Lateral tooth bicuspids with a high mesocone and a short ectocone in an opposite position to the central tooth.
Marginal tooth tricuspid to multicuspids, broader than laterals. Pallial system: same as in C. cordovana.
Reproductive system (Figs. 12A and 12B): Bursa copulatrix with sac rounded, usually longer than spermoviduct, in some specimens longer than spermoviduct plus albumen gland (Fig. 12A). External limits between epiphallus and penis not evident. Penis cylindrical, long, without penis sheath. Inner morphology of the penis divided into  three areas marked by differential pattern of sculpture. Proximal portion globose with higher diameter than restant portions, inner sculpture with pustules. Penial papilla absent. Penis medial sector cylindrical, inner wall with a longitudinal, thick, well-delimited pilaster running from proximal to distal end of the medial zone. Inner sculpture of distal penis portion consisting in three to four longitudinal straight thin pilasters, parallel to each other. Penial sheath absent. Penial retractor muscle short and thick, inserting in penis proximal portion. Epiphallus ⅓ of penial length. Flagellum thinner than epiphallus and ⅔ epiphallus length. Vas deferens thin, running freely along penis, attached to penial retractor muscle (Fig. 12B), then free along epiphallus and inserting between flagellum and epiphallus. Vagina cylindrical, with inner wall with thick, longitudinal pilasters, as long as distal portion of penis.
Habitat: calcareous rocky outcrops on mountain slope, under and between roots of woody shrubs.
Definition. Shell with marked wide axial ribs narrowly separated at regular spaces, traversed by major periostracal spiral rows with shallow continuous lamellae. Periostracal hairs absent. Lower columellar lamella in shell aperture deeply undulating. Proximal penis inner sculpture with undulating thin folds in a reticulated disposition and a short thick, medium pilaster. Medial portion with a pilaster of ⅔ the length of penis. Vagina almost the same length of the distal portion of penis.

Description
External features (Figs. 10A and 10G): Soft body, including the sole with dark brownish homogeneous coloration. Foot short with blunt end.
Shell (Figs. 10G-10O and 11E-11F): Turritelliform to subfusiform comprising 9-10 slightly convex whorls (Figs. 10H-10M). Coloration light brown, with whitish ribs. Protoconch with axial strength ribs regularly arranged, with thin spiral parallel bands between ribs (Fig. 11E). Teleoconch with marked oblique, wide axial ribs narrowly separated at regular spaces (Figs. 10H-10M, 11E and 11F). Surface of the teleoconch traversed by major periostracal spiral rows with shallow continuous lamellae, parallel to each other. Between these major spiral rows, two to three minor spiral, undulating shallow lines (Figs. 11E and 11F). Periostracal hairs absent. At ultrastructural level, departing from each major spiral row, growth axial ramified microfolds traversed by spiral lines, giving the appearance of an irregular net (Fig. 11F). Suture deeply impressed. Last portion of body whorl detached from rest of the shell ending into a cornet (Figs. 10I and 10L). Aperture suboval with parietal side slightly excavated (Figs. 10H, 10K and 10N). Peristome simple, thin, slightly expanded and reflexed. A dorsal groove present in upper parietal-palatal side of aperture formed by the detached suture of body whorl. Peristomal ultrastructural sculpture of body whorl extending over dorsal portion of the cornet. Five lamellae present in the interior of the aperture, not connecting to the peristome. Upper columellar lamella long, straight, spirally following the columellar axis. Lower columellar lamella running parallel, deeply undulating, spirally following columellar axis (Fig. 10O). Basal lamella short, located to the right side of the cornet and making a deep indentation on dorsal side of the shell wall (Figs. 10J and 10M). Umbilicus narrow. Shell measurements represented in Table 2.
Jaw (Fig. 8C): Horseshoe shaped, less open than in C. cordovana. A total of 12 plaques with a triangular central one subdivided into three subplaques. Six lateral narrow rectangular shaped plaques at both sides of the central one. Lateral plaques slightly increasing their size toward the tip of the horseshoe. Each plaque traversed by several thin transversal grooves.
Reproductive system (Figs. 12C-12F): Ovotestis embedded into digestive gland into the fourth or fifth spire whorls. Hermaphroditic duct inserting at distal portion of the albumen gland. Seminal receptacle swollen. Fertilization pouch-spermathecal complex long, digitiform broaden at its base. Albumen gland spread within the sixth whorl. Bursa copulatrix with sac rounded and folded over proximal section of the duct (Fig. 12C). Bursa copulatrix duct, longer than spermoviduct, surrounding the spermoviduct and running toward basal portion of the albumen gland. External limits between epiphallus and penis not evident, only differentiated by its inner sculpture. Penis cylindrical, long, with a short, thin penis sheath overlapping distal penis (Figs. 12C and 12D). Inner morphology of the penis divided into three areas marked by differential pattern of sculpture. Proximal portion slightly swollen than remaining portions, inner sculpture with undulating thin folds in a reticulated disposition and a short thick, medium pilaster (Figs. 12D and 12E). Penial papilla absent. Penis medial portion long, cylindrical, inner wall with rhomboidal to hexagonal pustules on the surface, with a pilaster of ⅔ the length of the medial penis portion. Distal penis short, thinner than medial portion with inner sculpture consisting in three to four longitudinal, straight, thin pilasters, parallel to each other. Penial retractor muscle thin and long, inserting in penis proximal portion (Fig. 12F). Epiphallus ¼ of penial length, with an inner constriction cutting the longitudinal, thin folds. Flagellum thinner than epiphallus and ½ epiphallus length. Vas deferens thin, running freely along penis, attached to penial retractor muscle, then free along epiphallus and inserting between flagellum and epiphallus (Fig. 12F). Vagina cylindrical, with a distal portion thinner in diameter than the proximal, inner wall with longitudinal pilasters. Vagina almost the same length of the distal portion of penis. Atrium short.
Habitat (Fig. 10P): Living in rocky outcrops, on and under shrubs. Always dry environments.
Distribution (Fig. 5D): This species has a small area of distribution located in northwestern Córdoba within Tulumba, and Totoral departments, Argentina. Western limit for C. tulumbensis sp. nov. area of distribution is in Sierra de Macha, extending to the east toward Villa Tulumba, and to the north toward Cerro Colorado. Localities of occurrences are all below 700 m, in Chaco Serrano subecoregion.
DNA sequence data. Partial sequences of mitochondrial COI and 16S-rRNA genes, and the nuclear ITS-2 region from two paratypes (IBN 883, specimens 1 and 2) have been deposited in GenBank with accession numbers: MG963436 and MG963437 for COI; MG963462 and MG963463 for 16S-rRNA, and MH789460 and MH789461 for ITS-2.
Remarks. The new species, Clessinia tulumbensis sp. nov. include Clessinia cordovana striata (Parodiz, 1939). The name striata has not been used here to avoid homonymy with Pupa striata Spix, 1827, the type species of Spixia, since in the present study the genera Clessinia and Spixia are proposed as synonymous. The new species with its own holotype and paratypes is defined based on live-collected material from which DNA sequences were obtained and the anatomy described. In this sense, although the Parodiz name is preoccupied, we are not replacing the name proposed by him in 1939 but creating a new species with its own type series. C. tulumbensis sp. nov. has clear differences with the remaining species of the cordovana-group showing shell axial ribs more marked than in C. cordovana and C. stelzneri. The periostracum lacks spines or hairs, only major periostracal spiral rows with shallow continuous lamellae are present. C. tulumbensis sp. nov. has a clear delimited pilaster in the medial penis portion, also present in C. stelzneri, but absent in C. cordovana. The penis shows a short, thin penial sheath overlapping part of the penial distal portion and the vagina is short as in C. stelzneri. C. tulumbensis sp. nov. has a narrow distribution area in northern Córdoba, occurring in sympatry with C. stelzneri in some localities within eastern portion of Tulumba department. Clessinia pagoda Hylton Scott, 1967: 98;-Fernández, 1973: 144;-Breure, 1974: 120;-Neubert & Janssen, 2004: 221, pl. 19, fig. 247;-Cuezzo, Miranda & Ovando, 2013: 163, fig. 2A. Type locality. Argentina, Córdoba, Cruz del Eje department, Sierra Chica de Córdoba, Quilpo.

Description
External features (Figs. 13A and 13B): Animal dark brown to black, with light brown ocular tentacles. Homogeneous coloration over the cephalopedial region, same specimens with sole light cream. Usually shells of live snails covered with a coat of sand granules (Fig. 13A).
Shell (Figs. 13C-13J and 14A-14F): Subpyriform with conic spire, solid. Seven to eight whorls with median keel in each spire whorl (Figs. 13C-13E). Body whorl with convex contour. Homogeneous light brown when periostracum is present (Figs. 13A,. Protoconch with axial strength ribs regularly arranged, space between axial ribs with thin spiral parallel bands (Fig. 14A). Teleoconch with oblique shallow ribs slightly marked (Fig. 14B). Each spire whorl with an equatorial and/or lower spiral row bearing rounded lamella of 300-400 mm tall and 200-250 mm wide 14B,14C and 14E). Each lamella superimposed with the following in a row (Fig. 14C). Several major and minor spiral rows parallel to the lamellae medial row. Suture between whorls also with a row of lamellae (Fig. 14E). Lamellae are lost in abraded specimens without periostracum (Figs. 13C-13E). Minor, shallower spiral rows regularly spaced between major rows with lamellae. Space between minor rows showing microaxial folds with the appearance of an irregular net (Fig. 14D). Sculpture of body whorl consisting in at least five major spiral rows of smaller lamellae than former described for spire. Variable number of minor spiral rows between major rows of lamellae (Fig. 14E). Aperture detached from body whorl forming a cornet, peristome expanded (Figs. 13D and 13G). Microsculpture of body whorl prolonged dorsally over peristome (Fig. 14F). Suture of body whorl when detached forming a marked keel that produce a marked dorsal angle of the aperture (Figs. 13C and 13D). Aperture rounded to oval (Figs. 13I and 13J). Dorsally, last portion of body whorl with a marked groove (Fig. 14F). Five lamellae obliterating the aperture. Umbilicus narrow. Jaw (Fig. 8D): Horseshoe shaped. Eleven plaques with a triangular central one subdivided into three longitudinal subplaques. Lateral plaques quadrangular to rectangular shaped, increasing their size toward the tip of the horseshoe. Each plaque traversed by several transversal grooves.
Pallial System (Figs. 15A and 15B): Pulmonary roof thin, traversed by few veins mostly concentrated on distal portion. Kidney triangular, short, ¼ the length of the pulmonary roof. Kidney with several longitudinal folds in its interior (Fig. 15B). Secondary ureter closed over most of its length, opening slightly before rectum. Pallial gland thin, parallel to mantle collar. Afferent vein parallel to main pulmonary vein. Mantle collar deeply marked by shell lamellae. swollen. Fertilization pouch-spermathecal complex long, digitiform broaden at its base (Figs. 15C and 15D). Bursa copulatrix with sac rounded and its duct slightly swollen at its base. Bursa copulatrix sac level with distal portion of albumen gland, its duct surrounding the spermoviduct, longer than spermoviduct in total length (Fig. 15C). Phallic complex formed by flagellum, epiphallus, and penis. External limits between epiphallus and penis not evident, only differentiated by its inner sculpture. Flagellum tapering toward its tip, thinner than epiphallus and about as long as epiphallus length (Fig. 15E). Epiphallus about the same length than penis, slightly increasing its diameter toward distal portion, with an inner constriction cutting the inner surface into two portions (Fig. 15E). Proximal portion with thick, pronounced, longitudinal pilasters, while distal portion with thin, scatter folds more separated between each other. Penis cylindrical, long, with a short, thin penis sheath overlapping its distal portion (Figs. 15E and 15F). Penis also separated from epiphallus by a thin inner constriction. Inner surface of penis  wall divided into three areas marked by differential pattern of sculpture. Proximal portion externally slightly swollen than resting portions, with inner sculpture formed by tightly appressed thin folds arranged in a reticular shape. Penial papilla absent. Penis medial portion long, cylindrical, inner wall traversed by thin folds that toward distal penis became diagonally arranged (Figs. 15E and 15F). Distal penis short, thinner than medial portion with inner sculpture consisting in three to six longitudinal, straight, thin folds, parallel to each other. Penial retractor muscle thick and short inserting in penis proximal portion. Vas deferens thin, running under penis sheath and then freely along penis, attached to penial retractor muscle, then free along epiphallus and inserting between flagellum and epiphallus. Vagina cylindrical, even in diameter, inner wall smooth or with shallow longitudinal pilasters. Vagina longer than distal portion of penis (Fig. 15C). Atrium short.
Habitat. Found in mountains with xerophytic vegetation usually under rocks or in crevices in rocks.
Distribution (Fig. 5D): Clessinia pagoda is only known from the localities of Quilpo and San Marcos Sierras in Córdoba province, Cruz del Eje department, in the Chaco Serrano. C. pagoda is a narrow range endemic species from northwestern Córdoba. The Cerro de la Cruz is close to Quilpo and to San Marcos Sierras, mountain where the species is easily found from 600 to 900 m of altitude.
Remarks. Clessinia pagoda paratypes specimens are completely worn out  and therefore all the miscrosculpture of the periostracum is lost. Living snails are found under clay or granite rocks. They are usually camouflaged with sand grains of the substrate that adhere to the periostracum (Fig. 13A), but the fragile lamellae and complex structures of the periostracum are not perceived until the shell is clean (Figs. 13F-13H). Strikingly, other carinated, rare land snail species, Plagiodontes weyenberghii (Doering, [1877a) of the family Odontostomidae (Pizá & Cazzaniga, 2012) also occurred in scatter areas through Córdoba in the Chaco Serrano subecoregion.
Clessinia nattkemperi (Parodiz, 1944)  Protoconch consisting of the first two whorls, with strength axial ribs, space between axial ribs with thin spiral parallel bands. Teleoconch with axial ribs and conspicuous periostracum ornamentation when present (Figs. 16F-16H). Periostracal sculpture consists of 10-20 thin spiral rows bearing triangular spines (Figs. 17A-17E) separated at regular spaces. Spines with wide base (50-60 mm) and about 100 mm tall (Figs. 17B and 17C). Space between spiral rows traversed by axial irregular microfolds cut by spiral or diagonal microribs forming an irregular net. Only the major type of spiral row is present, some of them without spines intercalated with the ones bearing spines (Fig. 17B). Aperture slightly detached from body whorl forming a shallow cornet (Figs. 16G, 17D and 17E), subovate to subquadrate, with a marked dorsal groove in upper portion of the aperture. Peristome expanded. Microsculpture of body whorl prolonged dorsally over peristome (Fig. 17D). Five inner teeth or lamellae not touching the peristome (Figs. 16I-16K). Upper palatal tooth triangular, not present in some specimens (Fig. 16J). Lower columellar lamella with rounded or rectangular shape when viewing from outside (Figs. 16I-16K). Jaw (Fig. 8E): Wide horseshoe shaped formed by 15 plaques, medial one triangular in shape and subdivided into three narrower plaques. Lateral plaques very narrow and about same size. The three last plaques on each side of the jaw broader than more central plaques.
Pallial system: Pulmonary roof thin and long traversed by few veins mostly concentrated on distal portion. Kidney triangular, short, a quarter of the total length of the pulmonary roof. Secondary ureter closed over most of its length, opening slightly before rectum. Pallial gland thin, parallel to mantle collar. Afferent vein parallel to main pulmonary vein.
Full-size  DOI: 10.7717/peerj.5986/ fig-18 length. Bursa copulatrix sac level with distal portion of albumen gland, its duct surrounding the spermoviduct, longer than spermoviduct in total length. Phallic complex formed by flagellum, epiphallus, and penis. External limits between epiphallus and penis not evident, only differentiated by its inner sculpture. Flagellum thin, tapering toward its tip, thinner than epiphallus and shorter in length (Figs. 18A and 18B). Epiphallus slightly shorter than penis, increasing its diameter toward distal portion, with an inner constriction cutting the inner surface into two portions. Proximal portion with thin, straight longitudinal pilasters, distal portion shorter with longitudinal folds less separated between each other and scalloped outline. Both portions internally separated by an inner constriction (Fig. 18C). A cylindrical papilla of the epiphallus with distal digitiform extensions is present. Penis mostly cylindrical, long, with a short, thin, transparent penis sheath overlapping its distal portion. Proximal portion more swollen than remaining portions, globular in some specimens, with inner sculpture formed by tightly appressed thin folds arranged in a reticular shape with a central, short pilaster (Figs. 18B and 18D).
Penial papilla absent. Penis medial portion long, cylindrical, inner wall with thin, parallel folds, with marked festooned outline (Fig. 18D). Distal penis short, thinner than medial portion with inner sculpture consisting in three to six longitudinal, straight, thin folds, parallel to each other. Penial retractor muscle thick and short inserting in penis proximal portion (Fig. 18A). Vas deferens thin, running under penis sheath and then freely along penis, attached to penial retractor muscle, then running parallel and attached to epiphallus by thin tissue and inserting between flagellum and epiphallus. Vagina cylindrical, short, even in diameter, inner wall smooth, or with shallow longitudinal pilasters. Vagina shorter than distal portion of penis. Atrium short.
Habitat (Figs. 16L-16N): Clessinia nattkemperi is found in close association with xerophytic plants, mainly cactuses, in patches of Chaco Serrano subecoregion. Specimens usually are hiding between long spines of cactuses or below dead cactuses branches lying over the ground. Found in sandy, dry substrate. Not found in rock crevices as other species of the genus.
Distribution (Fig. 5B). This species is endemic to the Sierra de Graciana mountain system and was only collected around the locality of Pomancillo, in Catamarca province, Northwestern Argentina. Dry Chaco ecoregion, Chaco Serrano subecoregion.
Remarks. C. nattkemperi is the species of the genus more similar to the former Spixia in general shell shape morphology. Its shell aperture is only slightly detached from the body whorl. It also shows triangular periostracal lamellae that are similar to the ones present in C. martensii (Doering, 1874) and C. tucumanensis (Parodiz, 1941).

DISCUSSION
Traditional used characters for taxonomic diagnosis in Odontostomidae, such as the shell morphology, provide important information for species identification, but due to their intraspecific variability they should be carefully considered. When Doering described Clessinia in 1875, he mentioned that the number of plaques of the jaw would be the best character to differentiate the new created genus from Odontostomus, Plagiodontes, and Spixia. However, the number of plaques in Clessinia overlaps with the ones present in species of other genera, so by itself this character is not enough for genera differentiation.
Our combined morphological and molecular study allows us to propose that the so called cordovana-group is formed by three species, C. cordovana, C. stelzneri, and C. tulumbensis sp. nov. The shell in C. tulumbensis sp. nov. has prominent, well-marked ribs, more raised than in C. cordovana and C. stelzneri. C. cordovana and C. tulumbensis sp. nov. have thinner shells than the other species. Among the cordovana species complex, the shape of the lower columellar lamella is deeply undulating around the columellar axis in C. tulumbensis sp. nov. and straighter in C. cordovana and C. stelzneri, thus showing to be a good character for species identification. The periostracum is of special taxonomic interest because it bears distinct microscale architectures. However, where and how these structures are formed is yet unknown for the majority of the species (Allgaier, 2011). Haired shells occur in several species of Stylommatophora, as for example, in families Polygyridae, Helicidae, Hygromiidae, Clausiliidae, Vertiginidae, and Solaropsidae. These families are distantly related suggesting that these features have evolved several times independently. In some cases, the ornamentation can be a response to structural demands from the environment, including camouflage and defense against predators and parasites. These ornaments can be the support for sand granules, dirt, or any other component of the natural habitat that can camouflage the shell. Pfenninger et al. (2005) suggest that hairs on shells of Trochulus Chemnitz, 1786 confer a selective advantage in humid habitats only and that are lost in drier habitats. Strikingly, in Argentina drier habitats hold the species with longer hairs such as C. cordovana from central western mountains in Córdoba. Clessinia pagoda is frequently found covered by a thick layer of sand grains overlapping the shell spiral rows but not completely the lamellae. The coloring of the shells covered by sand is perfectly camouflaged with the rocks where they live. In Clessinia, the type and diversity of periostracal ornamentation needs to be further investigated taking into account not only their ecological niche differentiation but also species reproductive behavior. Anatomical studies show that the reproductive system holds the most useful characters for taxonomic identification as in other stylommatophoran snails. Inner sculpture of the penis in the five species studied show particular characters. The presence of a penial pilaster and a sheath also contributes to species differentiation in the cordovana-species group.
The geometric morphometric analyses performed in this study confirmed the distinctiveness among all the species here treated. While C. pagoda and C. nattkemperi are clearly different, there is some degree of overlap in shell shape among the species of the cordovana-group. While C. cordovana has a slim body whorl with a suboval aperture, C. stelzneri has a body whorl more expanded and voluminous with subcircular aperture and C. tulumbensis sp. nov. has an intermediate shape of body whorl with expansion of the central portion of the aperture. When the cordovana-species complex was analyzed alone, shell differences are more evident but still a degree of overlap exists. Clessinia cordovana from Cerro del la Cruz and surrounding San Marcos Sierras areas shows the typical shapes according to the species description, while specimens from Sierra de Pocho are more closely related in shape to C. stelzneri. In the case of C. tulumbensis sp. nov., Cerro Colorado is the locality that shows specimens with smaller aperture and spire first whorls more expanded. Even when geometric morphometrics is useful for taxonomic identification, we support the necessity of a comparative analysis using also anatomical studies for a correct taxonomic identification of specimens within the cordovana-group.
Molecular studies performed with the nuclear marker, revealed that the ITS-2 region contains much less polymorphism than the mitochondrial genes and does not exhibit enough genetic variation to determine species relationships among Clessinia and Spixia species, which are shown as a polytomy in our phylogenetic reconstructions. These results conform well to Breure & Romero (2012), who obtained poor resolution at lower taxonomical levels within phylogenetic trees of Orthaliciodea by using the same nuclear region. Because this nuclear region seems to be too conserved to depict specific relationships, further studies with more nuclear markers are required in order to reconstruct fully resolved phylogenetic trees in the genus Clessinia. On the other hand, our analyses performed with 16S-rRNA showed that C. cordovana differs from C. stelzneri by genetic divergences ranging between 17% (p distance) and 19.3% (K2P distance) and from C. tulumbensis sp. nov. by distances ranging between 17% (p distance) and 19.8% (K2P distance), thus suggesting that C. cordovana is a different species from C. stelzneri and C. tulumbensis sp. nov. Regarding pairwise interspecific divergence between C. stelzneri and C. tulumbensis sp. nov., the greatest genetic distances we found are less than 2% (1.1-1.8%). Similarly, based on the COI locus we found that C. cordovana differs from C. stelzneri by genetic distances ranging between 14.3% (p distance) and 17% (K2P distance), and from C. tulumbensis sp. nov. by divergences ranging between 14.9% (p distance) and 19% (K2P distance), while distances between C. stelzneri and C. tulumbensis sp. nov. ranged between 1.3% and 4.1%. As with the ribosomal marker, these COI-based values suggest that C. cordovana represent a different species from C. stelzneri and C. tulumbensis sp. nov.
Our analyses with two different methods to test if morphological species of Clessinia satisfied the criteria to be considered different evolutionary genetic species, allowed us to recognize C. cordovana, C. nattkemperi, and C. pagoda as distinct evolutionary genetic species. However, the results from ABGD and K/h approaches failed to recognize the morphological diversity between C. stelzneri and C. tulumbensis sp. nov., and the phylogenetic trees did not support the presence of separate species for both morphological species. The trees showed C. stelzneri as paraphyletic with respect to C. tulumbensis sp. nov., with DNA sequences of C. tulumbensis sp. nov. nested among those of C. stelzneri. Furthermore, the distance values obtained between C. stelzneri and C. tulumbensis sp. nov. are relatively low. This low genetic divergence suggest that the two species have evolved relatively recently, with C. tulumbensis sp. nov. having evolved within C. stelzneri, and would explain the reduced resolution of species boundaries detection using genetic data, as both ABGD and K/h approaches are known to fail in cases of very recent speciation (Puillandre et al., 2012;Birky, 2013). Thus, based on the molecular markers 5. Based on our current analyses, and on previous findings (Breure & Romero, 2012) Clessinia and Spixia are synonymous, and according to the principle of priority (ICZN Code, Art.23.1) the valid name of the taxon should be Clessinia Doering, 1875 which has priority over Spixia Pilsbry & Vanatta, 1894.
Roberto Eugenio Vogler conceived and designed the experiments, performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, prepared figures and/or tables, authored or reviewed drafts of the paper, approved the final draft, wrote the paper. Ariel Anibal Beltramino conceived and designed the experiments, performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, prepared figures and/or tables, authored or reviewed drafts of the paper, approved the final draft, wrote the paper.

Field Study Permissions
The following information was supplied relating to field study approvals (i.e., approving body and any reference numbers): Field and collection permits for this study were issued by National Parks Administration (DRNOA 126/17) and Environment and sustainable development secretary, Catamarca province (SEAyDS 208/16).

DNA Deposition
The following information was supplied regarding the deposition of DNA sequences: GenBank accession numbers: MG963434-MG963464 and GenBank MH789452-MH789466.

Data Availability
The following information was supplied regarding data availability: The raw data is available in Supplemental Files.

Supplemental Information
Supplemental information for this article can be found online at http://dx.doi.org/10.7717/ peerj.5986#supplemental-information.