Mollusks ( Gastropoda and Bivalvia ) of the Multiple-Use Reserve Martín García Island , Río de la Plata River : biodiversity and ecology

The Island of Martin Garcia is located in the Upper Río de la Plata, to the south of mouth the Uruguay River. The aim of the present study was to analyse the biodiversity of the island freshwater mollusks and their relationships to environmental variables. Twelve sampling sites were selected, five were along the littoral section of the island and seven were Inland ponds. Seven major environmental variables were measured: water and air temperature, percentage of oxygen saturation, dissolved oxygen, electrical conductivity, total dissolved solids and pH. Twenty-seven mollusk species were found, Antillorbis nordestensis, Biomphalaria tenagophila tenagophila, B. t. guaibensis, B. straminea, B. peregrina, Drepanotrema kermatoides, D. cimex, D. depressissimum, Chilina fluminea, C. rushii, C. megastoma, Uncancylus concentricus, Hebetancylus moricandi, Stenophysa marmorata, Heleobia piscium, H. parchappii, Potamolithus agapetus, P. buschii, P. lapidum, Pomacea canaliculata, P. megastoma, Asolene platae, Corbicula fluminea, Eupera platensis, Pisidium sterkianum, P. taraguyense and Limnoperna fortunei. UPGMA clustering of species based on their occurrence in different ecological conditions revealed two main species groups. The Canonical Correspondence Analysis suggests that the species distribution is related to the physico-chemical condition of water. Axis two of the ordination diagram displayed the approximately 95.6% of the correlation between species and environmental variables. Dissolved oxygen, conductivity, water temperature and pH showed the highest fluctuations during the sampling period. The species richness (S) showed relationships mainly with water temperature and conductivity. The biodiversity of the gastropods and bivalves from Martín García Island amounts to up to 26 species. Among the Gastropoda, the Planorbidae family made the most sizeable contribution. The Lithogliphidae P. agapetus (26.28%) and P. buschii (9.50%) showed the highest relative frequencies of occurrence within the littoral environments, while the Planorbidae D. cimex (23.83%) and D. kermatoides (11.59%) likewise did so in the inland ponds.


Introduction
The Island of Martín García is located in the Upper Río de La Plata River, to the South of the mouth of the Uruguay River (34° 11' 09" S and 58° 15' 09" W).This island is an outcropping of crystalline Brazilian basement and inland, contains both permanent and temporary ponds, some hollows of which having originated in the old basalt excavations and are located in several sectors around the plateau (centre and southern regions of the island).In addition, small temporary ponds are formed in natural depressions, within the marginal forests in regions close to the coasts or at the bottom of sand dunes.
The aim of the present study, conducted in the Multiple-Use Reserve Martín García Island, was to analyse the biodiversity of the island freshwater mollusks and their relationships to environmental variables.

Material and Methods
The sampling was conducted on eight occasions from 1995 through 1997.Twelve sampling sites were selected: five lay along the littoral section of the island and seven were inland ponds (Figure 1).Specimens were collected with 15 cm diameter and 0.14 mm-mesh-size sieves (captures per unit effort = specimens/30 minutes/person).The mollusks were taken in vivo for identification with a binocular stereoscopic microscope.In the laboratory, the animals were fixed in 70° alcohol and deposited in the Molluscan Collection of the La Plata Museum.
The following physicochemical variables were measured with portable digital sensors: water and air temperature (T °C), dissolved oxygen, percentage of oxygen saturation, conductivity (µs.cm -1 ), total concentration of dissolved solids (TDS mg.L -1 ), and the pH.
The relationships between the species and the environmental variables were examined by the Canonical Correspondence Analysis (CCA), upon consideration of the 19 most abundant and frequently occurring species along with 7 environmental variables.Because the temperature of the air and the concentration of dissolved solids contained redundant information (inflation factors >20), these parameters were not considered in the analysis (Ter Braak, 1986;Ter Braak and Verdonschot, 1995;Külköylüoğlu, 2003;2005;Külköylüoğlu and Dügel, 2004;Maltchik et al.,2010).The mean value of each environmental variable together with data on the species abundance from eight sampling stations were used in the CCA (Kovach, 1998).Associations among the species were analysed by species-clustering analysis (UPGMA) with a Jaccard Index (Crisci and López Armengol, 1983).The species analysed were the same as those used for CCA.Pearson correlation analysis was conducted and the statistical significance of the correlations tested by means of the Student t-test.

Results
During the sampling period 20 species and 2 subspecies of the Gastropoda were found, belonging to the families Planorbidae, Chilinidae, Physidae, Ancylidae, Cochliopidae, Lithogliphidae, and Ampullariidae along with 5 species of the Bivalvia of the families Corbiculidae, Sphaeridae, and Mytilidae.Table 1 shows the species found at the sites surveyed and their relative frequency.
With respect to the relative abundances of the mollusk species throughout the four seasons (Figure 2), the Planorbiidae family as represented by Drepanotrema cimex (Moricand, 1839) was the most frequent and abundant species, with a peak in autumn followed by another in summer and similar, but lower, values during winter and spring.D. kermatoides (D'Orbigny, 1835) was not found in the winter, but rather exhibited a peak of abundance in the autumn that declined towards the summer reaching a minimum value in the springs.D. depressissimum (Moricand, 1839) showed a peak of abundance in summer but exhibited poor values during the rest of the seasons of the year.Biomphalaria straminea (Dunker, 1848) (Lucena, 1954) exhibited a peak of abundance in autumn; the species became scarcely present in winter, but then recovered to give lower values in spring and summer.
The Chilinidae family was represented by three species: Chilina fluminea (Maton, 1809) was found throughout the four seasons, with a peak of abundance in summer, followed by high values in autumn and lower but similar values in winter and spring.C. rushi Pilsbry, 1911 andC. megastoma Hylton Scott, 1958 occurred sporadically, but always with low values of abundance.Both species were found mainly in autumn.
The Physidae family was represented by only one species, Stenophysa marmorata (Guilding, 1828) throughout the four seasons of the year, with peak values in autumn that gradually decreased from winter through spring to summer.
The Ancylidae family was represented by two species: Hebetancylus moricandi (D'Orbigny, 1837) was found the most frequently with peak values in spring, slightly lower levels in winter and summer, and a minimum representation in autumn.Uncancylus concentricus (D'Orbigny, 1835) evidenced a peak of abundance in summer, but gave minimum values in winter.
The Cochliopidae family was represented by Heleobia piscium (D'Orbigny, 1835) and H. parchappii (D'Orbigny, 1835).The former species showed a peak of  Finally the family Mytilidae, represented by an invading species, Limnoperna fortunei (Dunker, 1857), showed a peak of abundance in autumn but manifested very low values in winter, spring, and summer.
The results from the species-clustering analysis (UPGMA) revealed two main groups of species (Figure 3).The first consisted in S. marmorata, D. cimex, D. kermatoides and B. (t.) tenagophila.The second comprised a subgroup composed of P. canaliculata, D. depressissimum, H. parchappii, and P. megastoma along with another subgroup composed of P. agapetus, P. bushii, L. fortunei, H. piscium, H. moricandi, and U. concentricus-all of which groupings gave the maximum coefficient of Jaccard: Ch. fluminea (J = 0.68) was included in the last group with a lower coefficient along with a third subgroup composed of P. taraguyense, B. straminea, and A. nordestensis.
In the CCA (Figure 4), the arrows depict the environmental variables and designate the maximum variation of the parameter.The length of each arrow is proportional to the significance of the variable in the ordination diagram.DO conductivity water temperature, and pH exhibited the highest fluctuations during the sampling period.The analysis suggests that the distribution of species was associated with the physicochemical conditions of the water.Axis 2 of the ordination diagram accounted for approximately 96% of the correlation between the species distribution and the environmental variables (cf.Table 2).
The dissolved oxygen, pH, and water temperatures were highly correlated among themselves.The electrical conductivity, however, was negatively related to the other physicochemical variables, but displayed a high correlation with certain environments -e.g., Cantera Basural (Site 5); Cantera Grande (Site 9); and to a lesser extent Cantera Tanque (Site 7), the inland ponds on the island.
The water temperature also correlated closely with the relative abundances of the mollusks within the environments -e.g., Puerto Viejo (Site 11) and Cantera Tanque (Site 7) -in the ordination diagram.
The dissolved oxygen and the pH exhibited a high correlation with the environments near the ponds -e.g., Casa Bomba (Site 1), Juncal (Site 4), and Playa de Arena (Site 3) along with the coastal drainage channels of Playa de Arena (Site 2).
With respect to the distribution of the species relative to the environmental variables, none were located at the centre of the ordination diagram.For the pulmonate gastropods (Table 3), the ordination diagram shows that species such as D. cimex were positively correlated with high values of conductivity but negatively so with temperature (r = -0.57),DO (r = -0.48)and pH (r = -0.38).Furthermore, in the ordination diagram S. marmorata, displayed a satisfactory association with conductivity, water temperature, and DO (r = -0.34);while C. fluminea was distributed at medium DO values and pH.D. kermatoides exhibited a negative correlation with DO (r = -0.44)and pH (r = -0.63);B. tenagophila tenagophila, was negatively associated with DO (r = -0.70);B. straminea was negatively correlated with conductivity (r = -0.32),but positively correlated with DO (r = 0.45); while D. depressissimum showed an excellent correlation with temperature (r = 0.56); H. moricandi, displayed a close positive correlation with DO (r = 0.43), though a negative one with conductivity (r = -0.32);U. concentricus was more closely associated  The numbers indicate strong correlations (*p < 0.05, **p < 0.01) where n = 48 for all variables.The abbreviations are the same as in Table 1.
with temperature; and A. nordestensis exhibited a negative correlation with both DO (r = -0.37)and pH (r = -0.59).
For the Caenogastropoda, P. canaliculata was negatively correlated with pH (r = -0.51)and DO (r = -0.39),but positively correlated with temperature (r = 0.37); P. megastoma exhibited a negative association with pH (r = -0.49)and DO (r = -0.32)as well as with water temperature; H. parchappii, however, was well correlated with temperature, while H. piscium was negatively associated with dissolved oxygen, pH, and conductivity.P. agapetus was correlated mainly with dissolved oxygen and pH, though to a lesser extent with temperature and conductivity; while P. bushii was found more closely associated with conductivity, on the negative side, and with temperature.
Both the infaunal (C.fluminea) and the epifaunal (L.fortunei) bivalves were positively correlated with the pH and the DO variables.Conversely, the infaunal bivalve P. taraguyense was negatively associated with both the pH (r = -0.59)and the DO.
The total number of species (NS) indicated relationships mainly with temperature and conductivity, with the latter on the negative side (p < 0.05).

Discussion
In all of Argentina, 158 mollusks, gastropods, and bivalves have been recorded; but the greatest richness and diversity has been found in the Del Plata basin of the Brazílic Subregion (Tassara et al., 2001;Rumi et al., 2008).That the biodiversity of the gastropods and bivalves from Martín García Island amounts to up to 26 species, even though the island is only 184 ha in area, is therefore not particularly surprising.Among the Gastropoda, the Planorbidae family made the most sizeable contribution.The Lithogliphidae P. agapetus (26.28%) and P. buschii (9.50%) showed the highest relative frequencies of occurrence within the littoral environments, while the Planorbidae D. cimex (23.83%) and D. kermatoides (11.59%) likewise did so in the inland ponds.
The results of the species-clustering analyses revealed two main groups of mollusk species (Figure 2).The first consisted in S. marmorata, D. cimex, D. kermatoides, and B. tenagophila tenagophila; which species are typical of Cantera Grande (Site 9).In this regard, studies by Bonetto et al. (1982) along both banks of the middle Paraná River revealed the association between D. kermatoides and B. tenagophila tenagophila.The second group and one of its subgroups consisted in the typical species P. canaliculata, D. depressissimum, H. parchappii, and P. megastoma from coastal habitats such as Playa de Arena (Site 3).Another subgroup onsisted in P. agapetus, P. bushii, L. fortunei, H. piscium, H. moricandi, and U. concentricus with the highest clustering coefficient along with C. fluminea (J = 0.68).These species, typically of coastal environments, inhabit rocky substrates.By contrast, the infaunal bivalve C. fluminea-it restricted to sandy substrates-joined the latter group with a lower clustering coefficient.Accordingly, Da Silva Gama (2004) had reported that the distribution of this latter bivalve was more associated with those same sandy substrates.
The final subgroup, consisting in P. taraguyense, B. straminea, and A. nordestensis, contained typical species from coastal environments.P. taraguyense was also found in Cantera Basural (Site 5).Since this latter site is an aquatic environment very close to the coast, the introduction of the species very likely occurred during floods from the river.
The ordination diagram shows the distribution of the mollusks with respect to the environmental variables.Although none of the species was located at the centre of the diagram, P. canaliculata, P. megastoma and D. depressissimum are the species closest to the average environmental variables.S. marmorata, D. cimex, D. kermatoides and B. tenagophila tenagophila were the most closely associated with medium and high values of conductivity.The same pattern has been observed for B. (t) tenagophila by Bonetto et al. (1982).This species was found exclusively in the inland ponds (Sites 5, 7 and 9) and was associated with floating vegetation.
Species of gastropods and littoral bivalves such as H. piscium, H. parchappii, P. agapetus, P. bushii, U. concentricus and L. fortunei were most closely associated with the variables pH, DO and water temperature as well as with hard substrates (basaltic rocks).H. moricandi was found on the surfaces of coastal vegetation; while C. fluminea and P. taraguyense, infaunal species, inhabited sandy loam substrates of the littoral sites.B. straminea and A. nordestensis were associated with aquatic macrophytes, such as Ludwigia elegans, L. bonariensis and L. peploidesplants typically found along the coasts of Martín García Island.
Rodrigues Capítulo et al. (2001) reported the presence of P. canaliculata, D. kermatoides and S. marmorata in streams of the pampean tributaries of the Río de la Plata River, at conductivity readings ranging from 705 to 878 µS.cm -1 .From the present results we can conclude that P. canaliculata, the closest species to the centre of the CCA diagram, should be poorly sensitive to conductivity; while the other two species should prefer environments with medium-to-high conductivities.
According to De Francesco and Isla (2003), salinity would constitute the most influential parameter determining the distribution of the Heleobia species.Those authors observed that H. parchappii demonstrated a conductivitytolerance range of 4 to 34‰ and concluded that the species was highly resistant to the salinity of the estuarine and continental waters (Rodrigues Capítulo et al., 2001).This species is also abundant in the pampean region (Gaillard and Castellanos, 1976;Castellanos and Landoni, 1995).Ciocco and Scheibler (2008) recorded H. parchappii on the littoral benthos of a saline lake in Mendoza.Although Gaillard and Castellanos (1976) considered H. piscium a freshwater species, Darrigran (1995) stated that it was an eurihalin species because it was found in inland fluvial (salinities < 0.5‰) and intermediate-fluvial (salinities between 0.5 to 25‰) zones of the Río de la Plata River.
Our results agree with the observations of those authors regarding the distribution of both species on the Martín García Island, since the two are found in littoral environments associated with specific conditions of water temperature, dissolved oxygen and pH.
With respect to the species distribution of the Potamolithus genus on the island, the ordination diagram would indicate that P. bushii is negatively related to conductivity, it clearly being a freshwater species; while P. agapetus presents a higher tolerance to salinity; with that characteristic being in agreement with López Armengol and Darrigran (1998), who reported a distribution of the species up to as far as the Balneario Atalaya (the marinefluvial zone of the Río de la Plata River).H. moricandi, U. concentricus, Ch. fluminea and the bivalve C. fluminea -those species cited by Lanzer and Schafer (1985) in coastal ponds in the south of Brazil-were found by them under similar conditions of conductivity and pH to those recorded for the littoral zone of Martín García Island.
During the sampling period of 1995-1997, the invasive bivalve L. fortunei, exhibited low densities on the island with an initial density recorded in 1991 of 4 to 5 individuals m 2 (Pastorino et al., 1993).Damborenea and Penchaszadeh (2006) reported that the species manifested a high flexibility during its reproductive cycle and that its dispersion was favored by human activity (Darrigran and Ezcurra de Drago, 2000;Manzur et al., 2003;Avelar et al., 2004;Brugnoli et al., 2005;Belz et al., 2005).This invasive species exhibits a great plasticity with respect to environmental variables, and this characteristic has been confirmed from the results of the analysis of correlation by Pearson, which data indicate special relationships to the physicochemical parameters examined in this present study.
On the island, the Sphaeridae family is represented by P. sterkianum with 3 individuals at Site 11 as well as by P. taraguyense with a greater abundance and an occurrence in sites 2 and 11.Although the morphology, taxonomy, and distribution of the species have been studied; little information is available regarding their ecology.For this reason, a comparison of the results obtained in the present paper with other references of the species cannot be made.
evidenced the highest frequency of occurrence in autumn, lower values in summer and spring, and the lowest representation in winter.B. tenagophila tenagophila (D'Orbigny, 1835) displayed a peak of abundance in summer that gradually declined towards autumn and winter and finally became reduced to as maiores oscilações durante o período de amostragem.A riqueza de espécies (S) apresentou relações principalmente com a temperatura da água e a condutividade.A biodiversidade de gastrópodes e bivalves da Ilha Martín García é de 26 espécies.Entre os Gastropoda, a família Planorbidae exibe a contribuição mais importante.O Lithogliphidae P. agapetus.(26,28%) e P. buschii (9,50%) apresentaram as maiores frequências relativas de ocorrência dentro dos ambientes costeiros, enquanto o Planorbidae D. cimex (23,83%) e D. kermatoides (11,59%) também apresentaram frequências expressivas nas lagunas interiores.very low percentage in spring.B. tenagophila guaibensis Paraense, 1984 was scarcely found at all, with the highest peak occurring in spring but with minimum values during the rest of the year.B. peregrina (D'Orbigny, 1835) passed through a peak of abundance in spring but likewise declined to minimum values in the other seasons.Finally, although Antillorbis nordestensis a

Table 1 .
Mollusk species with their code and collection sites on Martín García Island during eight samplings.inspring,followedbysummer,though thereafter a decrease to slightly lower values; whereas no specimens were observed in winter.The latter species was recorded throughout the four seasons, with peak values in spring, minimum levels in summer, and similar intermediate abundances in autumn and winter.The Ampullariidae family was represented by three species: Pomacea canaliculata (Lamark, 1822) was present throughout the year, with maximum values in summer that became minimum in winter, but passed through similar intermediate levels in autumn and spring; P. megastoma (Sowerby, 1825) evidenced a higher abundance in summer; while Asolene platae (Maton, 1809) passed through a low frequency of occurrence throughout the four seasons of the year.The Lithoglyphidae family was represented by three species, Potamolithus.agapetusPilsbry,1911andP. buschii  (Fraunfeld, 1865); both with the highest values of relative abundance and presence throughout the sampling period, exhibiting a peak in summer though minimum values in winter.P. lapidum (D'Orbigny, 1835) displayed a peak of abundance in autumn, but had a low frequency of occurrence throughout the rest of the seasons.Within the Bivalvia, the Sphaeridae family was represented: Eupera platensis Doello Jurado, 1921, exhibited a peak of abundance in summer that decreased in autumn to pass through minimum values in winter and spring; while Pisidium sterkianum Pilsbry, 1897 and P. taraguyense Ituarte, 2000 both evinced peak values of abundance in summer.Of the family Corbiculidae, Corbicula fluminea (Müller, 1774) exhibited the highest abundances in summer.
Frequency (Freq.%)indicates percentage of occurrence of the species as a percentage of the total number of species registered for the entire sample.The asterisk (*) indicates the 19 most frequently occurring species, which group accounts for more than 98% of the total species recorded.Figure 2. Mollusk relative abundance by season.abundance

Table 3 .
Pearson correlations among four environmental variables, total number of species (NS), and the nineteen most abundant species analysed.