Intraspecific clinal variation in Plagiodontes patagonicus (Gastropoda: Orthalicidae, Odontostominae), an endemic species from Argentina

Plagiodontes patagonicus (d'Orbigny, 1835) is a quite variable species endemic to southern Buenos Aires province (Argentina). Its taxonomic characterization did not include any quantitative analysis of variability for an accurate discrimination, and the extremes of its variation were described as different subspecies or species. In this paper, shell measurements and angles, and quantitative data on the terminal male genitalia were studied by Principal Component Analysis and Varimax multivariate analysis. Typical Plagiodontes patagonicus and the largest form known as P. patagonicus magnus Hylton‐Scott, 1951 showed an almost continuous pattern of shell variation, which is positively correlated with the altitude gradient over their geographical range, which in turn is correlated with a rainfall gradient, i.e. they constitute a size‐form cline that does not allow objective delimitation of different morphospecies. Data from the genital system were also arranged as a geographical gradient within the P. patagonicus patagonicus–P. patagonicus magnus continuum. The variability of protoconch sculpture and apertural teeth also indicate recognition of them as a single taxon.

size differences with no attribution of a taxonomic level. Pilsbry and Vanatta (in Pilsbry 1898) described a small form of P. patagonicus as a separate species, P. iheringi, which was soon recognized as a synonym of typical P. patagonicus by Pilsbry (1902). Hylton-Scott (1951) described the largest shells from the Ventania Mountain System as a new subspecies, P. patagonicus magnus. Later, Cazzaniga and Fernández-Canigia (1985) considered them as two different species, separated by morphological and geographical gaps. Typical P. patagonicus have a smooth shell with a striated protoconch and weakly developed apertural teeth, the lack of a transverse lamella being its distinctive trait. Plagiodontes magnus was described as having colour polymorphism, smooth protoconch, some differences in its genital anatomy mostly related to size, and a transverse lamella often present. Quantitative variables allowing discrimination among species in this genus were not adequately analysed.
From a geomorphological viewpoint, the distribution area of these forms or species is included in the Subhumid South Pampean morphogenetic system, where the mountain systems and wide closed depressions interrupt the extensive grassland plains. González-Uriarte (2002) included this area in the geoenvironments called ''Sierra relief'' and ''South Ventania plain''. The main factor varying on a south-north gradient across the area is altitude, from sea level to about 1200 m on the highest peaks, with a correlative gradual change in humidity and soil composition. Mean rainfall is a linear function of the altitude, with a mean positive gradient of 3.73 mm km 21 (Paoloni et al. 1988). Coincidently, zonal soil distribution shows a succession of the aridic, ustic and udic edaphic regimes. In the north part of the range, there is a predominance of the Order Mollisol with different degrees of udic and ustic evolution (association of Hapludol and Argiudol in the northwest, and of Haplustol and Argiustol in the central west) (M. González-Uriarte, personal communication).
The discovery of new live material of Plagiodontes from places scattered between the previously known localities allows us to rectify the taxonomic status of these forms, after realizing that a continuous pattern of variation, correlated with a geographical gradient, is more realistic than splitting the edges of variation as different taxa.
The aim of this paper is to analyse qualitative and quantitative characters of the shell and genital anatomy relating to the identity of P. patagonicus and P. magnus, which are hereby interpreted as the extreme forms of a size-shape cline of a single morphospecies.

Material and methods
Shell variability was analysed on 342 adult shells from several localities in south-west Buenos Aires province (Table I). Adulthood was assumed for shells with a full development of the apertural teeth and presence of a reflected outer aperture lip. Seven linear and angular variables were measured on shell drawings that were made with a camera lucida device on a stereoscopic microscope: shell length (SL), shell width (SW), length of the last whorl (LWL), aperture length (AL) and aperture width (AW); major angle and spiral angle were also determined as described in Pizá and Cazzaniga (2003). Characteristics of the apertural teeth were determined through direct observation under a stereoscopic microscope. Special attention was paid to the presence of a transverse lamella, because this genus-diagnostic character was deemed to be absent in P. patagonicus.
Protoconch sculpture of juvenile shells and adult shells showing the effects of erosion and rebuilding was analysed with a stereoscopic and a scanning electronic microscope at CRIBABB (Centro Regional de Investigaciones Básicas y Aplicadas de Bahía Blanca, CONICET).
Genital quantitative data were taken from 80 living specimens from four localities (20 snails from each): General Daniel Cerri; National Road 33 km 11 (''Cueva de Los Leones''); Cerro Bahía Blanca (a mountain in the Provincial Park Ernesto Tornquist); and Cerro Curamalal ( Figure 1). The snails were sacrificed in hot water. The measured variables, taken with a micrometre eyepiece on a stereoscopic microscope, were: penis length, epiphallus length, length of the spermathecal stalk, and diameter of the spermatheca.
Quantitative data were analysed by multivariate Principal Component Analysis (PCA) using the SPSS statistical package. Scores on the first axis of a two-axis Varimax solution were plotted against altitude, and the Pearsonian correlation was calculated, to detect sample clinal distribution (Gould 1969). ANOVA, Bonferroni's and least significant difference tests were used to compare shell length means between pairs of samples.

Results
Table II summarizes the morphometric information on the studied shells arranged in three geographical intervals (mountains, intermediate locations, and vicinities of Bahía Blanca); some representative specimens are shown in Figure 2. PCA showed a wide overlap of the  Figure 3). Evidence for a cline is unambiguous, since correlation between Varimax scores on axis 1 and ground altitude is highly significant (P,0.001; Figure 4). Figure 5 shows the variation of mean shell size along the geographic gradient from the coastal plain (General Daniel Cerri, 10 m above sea level) to the highest sampling sites (Cerro Bahía Blanca, 900 m high; Cerro Destierro, 1100 m high). Even though a great variability was found in each region, there was a clear tendency to increasing shell size as land altitude increases; the smallest adult shells from the Ventania Mountains did not have a significant shell size difference from the largest shells from areas around Bahía Blanca ( Figure 5). Factors other than altitude may influence local shell size, e.g. shells in set 6, showing a relatively small size, were collected from Cerro Destierro, at 900 m above sea level, in an area exposed to strong winds and deprived of sheltering vegetation, while sample 5 showing the largest shell lengths was mostly gathered from rock fissures occupied  by mosses and ferns, gaining advantage of the retained humidity, at 600-675 m on Cerro Bahía Blanca. Shells were thick, uniformly greyish white at all localities along the plain (localities 9-18); occasionally some shells showed a slight pink hue. Shell colour was polymorphic at the mountains, where greyish specimens may co-occur with either pinkish, rufous brown or wine-coloured shells. Most live specimens were intensely wine-coloured at Cerro Bahía  Blanca (Ventania Mountains, localities 4 and 5). Coloured shells are slightly shiny, smoother and thinner than the greyish white ones.
Differences were also found in the protoconch and apertural teeth of snails from different locations. The protoconchs of specimen areas around Bahía Blanca (typical Plagiodontes patagonicus) showed diverse levels of striation, from irregularly waved striae crossed by tiny thread lines giving the aspect of a stair (Figure 6), to weakened striae on which spiral lines are seldom or not visible (Figure 7). Young animals from the Ventania Mountains (typical P. patagonicus magnus) showed the same striation pattern as P. patagonicus (Figure 8). However, young adult shells had weaker striae, and most full-sized snails showed a    completely smooth protoconch, with a spiral cord or thread bordering the suture (Figure 9). This spiral thread was never observed in typical P. patagonicus.
Upper palatal, columellar and parietal teeth were present in all specimens from any locality. Lower palatal teeth showed a variation from zero to four teeth, with a tendency to an increasing number of lower palatal teeth with larger shell size ( Figure 10). More than 70% of the small adults from Bahía Blanca and its vicinities, i.e. typical P. patagonicus, were devoid of lower palatal teeth; approximately 70% of the large, .26 mm long adults from the Ventania Mountains, i.e. typical P. patagonicus magnus, had one to three lower palatal teeth.
Most of the studied specimens from all sites lacked a transverse lamella. Sometimes a small denticle was present in its place; the transverse lamella was occasionally formed by two or three linearly arranged teeth, and even a well-developed lamella was also seen in some shells (Figure 2). The frequency of shells with a rudimentary or developed transverse lamella increases with size ( Figure 11). So, adult snails from Bahía Blanca, 18-24 mm long, lacked in most cases a transverse lamella, but some populations (e.g. General Cerri) that were typical P. patagonicus in every other respect, showed up to 17% of shells with this fold; conversely, 67% of the snails from Cerro Bahía Blanca (in the Ventania Mountains), reaching 30 mm long (i.e. typical P. magnus), showed a transverse lamella, although only one out of 33 shells from Cerro Destierro had this structure. Table III shows the results of genital morphometry. PCA of genital anatomy showed that no difference can be traced among populations other than a general size factor; the only remarkable arrangement occurred on the first principal component, which accounts for 73.4% of the variation (Figure 12).

Discussion
Land snails usually show shell polymorphism due to the effect of environmental changes, food and calcium availability, differential predation, and other factors. Habitat fragmentation, which isolates groups of populations, contributes to the inaccurate assessment of specific variability, since the morphologically different snails collected from remnant patches of native vegetation may look as if they were of different species. The pampas are the most modified eco-region in Argentina and, since Plagiodontes snails are mostly  Values are given as mean SD (minimum-maximum); n520 specimens from each locality. PPET, Provincial Park Ernesto Tornquist. associated with relatively pristine areas, they are at present scattered in places showing suboptimal conditions for agriculture, especially fields with scarps or wide outcrops . Therefore, intermediate shell forms ranging between the extremes of the size-form cline are infrequently seen. Shell variability of Argentine land snails was usually described in a narrative style, with hardly any morphometric data. Parodiz (1962) authored a pioneering work on the variation of Spixia doellojuradoi (Parodiz, 1941) (Odontostominae), which was not followed by further similar contributions. Pizá and Cazzaniga (2003) have recently analysed the variability of Plagiodontes dentatus (Wood, 1828) and its morphologic discrimination from its closer species, among which P. patagonicus is found.
Identifying the latter species has always been a simple task because it is the one that has the smaller number of apertural teeth within the genus and, especially, because it lacks the transverse lamella so characteristic of the genus. Although d'Orbigny (1840) and other authors considered it at first sight as a variety of P. dentatus, its specific status was definitively clarified by Doering (1881) and Parodiz (1939), who recognized P. patagonicus as a different, endemic species from southern Buenos Aires province. The reference to its presence in Uruguay (d'Orbigny 1840) was based upon a misidentification of P. dentatus, as was discussed by Klappenbach and Olazarri (1973). Pizá and Cazzaniga (2003) demonstrated that, beyond differences in the number of teeth and other binary characters, shells of P. patagonicus and P. dentatus of the same size also have significant differences in their morphometric proportions. Doering (1876) stated that attaining larger sizes in the mountains and smaller sizes in the plains was a general characteristic of species in the genus Plagiodontes. Hylton-Scott (1951) Figure 12. Scatter plot of the scores of the first (PC1) and second (PC2) principal components for genital anatomy data of Plagiodontes snails from southern Buenos Aires province. Specimens from Bahía Blanca and Daniel Cerri (10-60 m above sea level) were recognizable as typical Plagiodontes patagonicus, while those from Curamalal and Ventania Mountains fitted the description of P. magnus. portrayed this size segregation by describing the new subspecies P. patagonicus magnus from the Ventania Mountain System, which was later treated as a different species (Cazzaniga and Fernández-Canigia 1985). However, the study of samples from residual populations along a north-south geographical axis between the type localities for P. patagonicus and P. patagonicus magnus highlighted some remarkable facts that drove us to change earlier concepts, e.g. the presence of a transverse lamella in up to 17% of the members of a population of P. patagonicus from an area where all the previously studied Plagiodontes snails lacked such a structure (General Daniel Cerri, Bahía Blanca District); the absence of transverse lamella in the shells from Cerro Destierro, i.e. an area where this fold is fairly common, very closely located to Cerro Bahía Blanca, where the maximum frequency of transverse lamella presence was recorded (67%); a continuous pattern of shell variation from P. patagonicus patagonicus to P. patagonicus magnus over a geographic gradient. There is not a clear-cut morphometric differentiation between the two alleged taxa, which should be interpreted as the extreme forms of a cline rather than two different morphospecies. Morphometry of the genital system confirmed the existence of a continuous geographic gradient of variation too. Goodfriend (1986) stated that land snail populations tend to reflect geographical variation of the environment because of their low mobility. The main cause for shell size and form variation appears to be humidity; higher humidity provides more frequent and/or longer periods when snail activity is possible, therefore increasing growth rates. If age at maturation is not negatively correlated with growth rate, then faster-growing snails would attain a larger adult size. This seems to be the case with P. patagonicus. Climate is semi-arid at Bahía Blanca, with approximately 500 mm of annual rainfall, and 300 mm of water deficit; conditions get progressively wetter towards the north, rainfall amounting to 900 mm, with only 100 mm of water deficit, at the Ventania Mountains (Glave 1975). Moreover, the snails living in the mountains generally inhabit rock cracks where lichens and ferns provide shelter and suitable microclimatic conditions. Age at maturation was estimated at 4 years for P. patagonicus (Delhey and Cazzaniga 2002a), a reflection of the apertural outer lip marking the end of growth. Our hypothesis is that snails in the mountains are larger because they have a higher growth rate rather than because they are older; attaining a larger shell size would be the simple direct cause for a greater number of teeth, including a higher frequency of presence of transverse lamella, with no taxonomic consequence. It is worth noting that the specimens collected at Cerro Destierro, which did not have a transverse lamella, were also relatively small for the area.
It is not only altitude and humidity that influence snail size. The small adults collected from Cerro Destierro formed a very dense population, an inverse relationship between shell size and population density having already been described for another population of P. patagonicus (Delhey and Cazzaniga 2002b).
Colour polymorphism in land snails is usually explained as an adaptive response to selective predation on the less cryptic morphs (Sheppard 1951), and/or by means of thermoregulatory constraints on dark shells in open habitats (Burla and Gosteli 1993); the interaction between these two selective forces would account for maintenance of a balanced polymorphism (Jones et al. 1977;Goodhart 1987). However, Slotow et al. (1993) reported a case in which neither selective predation nor thermal advantages could explain the colour variation of a polymorphic desert snail; rather, inequality in individual access to calcium carbonate among habitats seems to result in the observed differences of shell-colour frequencies, with whiter, thick-shelled snails present at sites with greater availability of calcium carbonate, and weaker, dark-coloured ones collected in low calcium carbonate sites (Slotow and Ward 1997). This hypothesis should be tested for P. patagonicus on account of the evidence that all specimens are thick and white over the loess plains, related to geologically younger terrains (Tertiary and Quaternary), with calcareous tufa outcrops, while dark-coloured shells are frequent at the rocky habitats in the Ventania Mountains, where Palaeozoic rocks of metamorphic origin constitute the main substratum. The rocks are quartziferous sandstones and schists associated with scarce conglomerate argillaceous material (Harrington 1947).