Pattern of shell utilization by the hermit crab Paguristes tortugae ( Diogenidae ) from Anchieta Island , southern Brazil *

The objective of this study was to characterize the pattern of shell utilization by Paguristes tortugae, from infralittoral areas of Anchieta Island (Ubatuba, São Paulo State, Brazil) using the percentage of the different shell types occupied and the morphometric relationship between hermit crabs and occupied shells. Specimens were collected monthly (January to December 1998) by SCUBA diving. The animals and the shells were measured and weighed. A total of 2,429 hermit crabs were captured occupying 21 species of gastropod shells. The shells more occupied were Pisania auritula (35.49%), Cerithium atratum (27.83%), Morula nodulosa (12.70%) and Leucozonia nassa (6.83%), respectively. There was differential shell occupation between sexes: males and ovigerous females occupied in higher percentage P. auritula shells while the non-ovigerous females occupied C. atratum. Regression analysis showed best correlation between the size of the crabs (shield length) and shell dry weight. This study showed that shell occupation by P. tortugae involves interand intraspecific competition as well as shell type availability in the field.


INTRODUCTION
According to Conover (1978), the ability of hermit crabs to occupy and carry gastropod shells is one of the main reasons for their evolutionary success.Hermit crabs need increasingly larger shells during their life cycle, a fact that keeps them in constant activity searching for suitable shells (Bertness, 1981).Some species are adapted to occupying the shells more easily encountered in the field (Mesce, 1993), while others select shells according to their preference for type, size, shape and/or shell covering (Grant and Ulmer, 1974;García and Mantelatto, 2001).Therefore, hermits should modify their preference for shell size based on the availability of empty shells (Hazlett, 1992).At the same time, the gastropod shells can be a limiting resource to hermit crab populations (Vance, 1972).
Hermit crabs are entirely dependent on the processes that make suitable shells available for their use, including snail mortality and inter-and intraspecific shell exchanges (Hazlett, 1981).The hypothesis more commonly used to explain that shells are in short supply are that natural populations of crabs are often in smaller shells than preferred in laboratory experiments (Vance, 1972;Scully, 1979;Bertness, 1980), empty shells are often rare in the habitats that support hermit crab populations (Hazlett, 1970;Bertness, 1980) and, hermit crabs often have ritualized shell-exchange behavior that may be beneficial when shell resources are scarce (Carlon and Ebersole, 1995).Several authors have studied shell occupation by hermit crabs in the field (Walters and Griffiths, 1987;Gherardi and Vannini, 1989;Gherardi, 1991;Negreiros-Fransozo and Fransozo, 1992;Manjón-Cabeza and García Raso, 1999;Martinelli et al., 1999;Bertini and Fransozo, 2000), and others have studied hermit crabs under laboratory conditions to verify the preference pattern by shells utilized (Lively, 1988;Hazlett, 1992;García and Mantelatto, 2001).Despite this background and considering the high number of hermit crab species described worldwide, few studies have evaluated comparatively shell occupation in the field and in laboratory conditions (see Mantelatto and García, 2000 for review).
Paguristes tortugae Schmitt, 1933 is a species relatively abundant, from Florida (U.S.A) to Brazil (Hebling and Rieger, 1986), and a reasonable number of studies have been conducted with this species: including larval development (Hebling and Negreiros-Fransozo, 1983), population aspects (Negreiros-Fransozo and Fransozo, 1992;Mantelatto and Sousa, 2000), fecundity and reproductive potential (Negreiros-Fransozo et al., 1992;Mantelatto et al., 2002).The objective of the present study was to characterize the pattern of shell utilization in the field by P. tortugae inhabiting the infralittoral area of Anchieta Island, Brazil investigating the factors that influenced the occupation of gastropodshell species by each sex and the morphometric relationships between hermit crabs and their shells assessed.

Study Area
Anchieta Island (23°33'S, 45°05'W) (Fig. 1) has a total area of about 10 km 2 and was recently declared an ecological reserve of São Paulo State because the natural environment has been affected by anthropogenic factors such as tourism and fishery activities in coastal areas.This island is located landwards, separated from the coast by a 300 m long and 35 m deep canal.However, the decapod crustacean fauna from this area is poorly studied, and their scientific information scarce.

Sampling and analysis
The hermit crabs were obtained monthly from January to December 1998 from four infralittoral rocky/sandy areas of Anchieta Island (East and South Beaches, Prainha and Saco do Vento), whose surface is irregular, with many huge boulders.Every month, at least one of these sites was sampled depending on ocean and weather conditions.Specimens were captured during the daytime by three SCUBA divers over a period of 30 min over an area of about 850 m 2 .This methodology provided large amounts of material from this area of irregular surface and thus was considered efficient for the hermit crab population study (Mantelatto and Garcia, 2002).Animals were frozen and transported to the laboratory for analysis where they were carefully removed from their shells, weighed (WW = Wet Weight) and measured for shield length (SL), shield width (SW), propodus length (PL) and propodus height (PH).Shell species were identified by Rios (1994), confirmed by a specialist (Dr.Osmar Domaneschi, University of São Paulo), measured (SAW = shell aperture width; SAL = shell aperture length), and their dry weight (SDW) was determined.Measurements were made with a caliper (0.1 mm) or by drawing with the aid of a camera lucida.Shell internal volume (SIV) was measured on the basis of the volume of fine sand (Ø = 0.25-0.105mm 3 ) required to fill the empty shell modified by Mantelatto and Garcia (2000).
The distribution frequency was tested by the Kolmogorov-Smirnov normality test (KS), and the Pearson Correlation was utilized to evaluate the relationship between the variables (hermit crab versus shell dimensions) (Zar, 1996).The best models obtained from Pearson analysis were checked by linear regression and by correlation coefficients.The shell type preference was estimated by the percentage of occupation of the chosen species and the chisquare test (χ 2 ) was used to compare occupancy rates of shell species.

DISCUSSION
According to Mantelatto and García (2002) The taxocoenosis in this area was controlled by P. tortugae, and for this reason a long-term effort has been devoted to this species during the last four years.Recent investigations revealed that P. tortugae presented continuous reproduction, a high reproductive potential (fecundity) influenced by the shell species characteristics, and this reproduction strategy also included the precocious size of sexual maturity (SL = 1.5 mm) (Mantelatto and Sousa, 2000;Mantelatto et al., 2002).
The population dynamics of the hermit crabs is closely tied to the species of gastropods that are present in their habitat (Mantelatto and García, 2000).Hermit crabs utilize different gastropod shell species at different rates, a fact that can be attributed to the preference for a certain shell architecture type and/or to the higher abundance of some species in the field (Reese, 1962).The availability of shells may be a limiting factor for some species of hermit crabs and can be affected by their abundance in a particular habitat and by inter-and intraspecific competition for shells (Reese, 1969).
Paguristes tortugae was found occupying 21 gastropod shell species, demonstrating plasticity in shell type utilization, as well as the shell stock potential in this area.In the studied area, the population of live gastropods consists of 32 species (unpublished data), with Cerithium atratum, Pisania auritula and Morula nodulosa the most abundant (74.9% of the total shells collected), and Leucozonia nassa showing reduced occurrence.Pisania auritula, C. atratum, M. nodulosa and L. nassa (those most often occupied in the survey during the present work) have a similar architecture.Despite this similarity, some differences in hermit size occupation were detected.Conversely, Negreiros-Fransozo and Fransozo (1992) observed P. tortugae occupying Stramonita shells only in a different area (São Sebastião).Therefore, we can infer that results might have been related to the wide availability of shells in the survey.
The size of the occupied shells usually is well correlated with hermit crab size (Abrams, 1980) when the availability of shells in the field is high.In this study, it was noted that shell weight showed the best correlation with P. tortugae size.Alternative hypotheses concern the advantages of occupying heavy shells are postulated (see Mantelatto and García, 2000 for review).Particularly, we agree with Reese (1969), that the occupation of heavy shells by hermit crabs protects them against predators, water currents and the impact and abrasion of waves.This is true for P. tortugae, which occurs in rocks and corals and therefore may be better protected by heavy shells in these habitats, with greater stability against currents and a frequent downfall escape movement (hermits enter the shell and drop from the rocks/corals falling to the crevices when they detect the presence of collector/predatorpers. obs.).According to Borjesson and Szelistowski (1989), small shells are more easily damaged or destroyed by shell-crushing predators than larger shells.In this study, the smaller shells were available in high quantity while large shells were available from more species.However, the shells most frequently occupied (P.auritula) showed low adequacy (r) for the hermit crab dimensions, except for L. nassa shells, indicating that P. tortugae might search for more adequate ones.In the present study the shell occupation pattern of P. tortugae was probably influenced by inter-and intraspecific competition, or by sympatry with three other hermit species with similar size in the studied area (Mantelatto and García, 2002).
It was noted that C. atratum and M. nodulosa shells were occupied by the smaller hermit crabs, while specimens of all size classes occupied L. nassa shells, reflecting wide size range available for this shell species.A clear occupation of P. auritula was observed for P. tortugae from 3.5 to 7.0 mm SL size range.This pattern characterizes a clear competitive ability between the different size individuals in the population.
The differences in shell utilization between sexes can be explained by intraspecific competition, intimately related with size of the individuals or by energy expenditures, related to reproductive activity and growth (Abrams, 1988).The highest correlations between hermit crabs and shell dimension were observed in ovigerous females of P. tortugae.The females allocate energy to growth, reproduction and egg incubation, resulting in a decrease in their somatic growth (Bertness, 1981;Abrams, 1988), occupying smaller shells (C.atratum) while the males allocate energy to growth and gonad development, occupying larger shells.The ovigerous females fit larger shells (P.auritula), perhaps for best space availability and protection of their eggs.Thus shell occupation showed an inclination of larger individuals to fit larger shells.Natural selection may favor males in relation to females, because they have the advantage of size over females in regarding competition for shells, such as reported for males of Calcinus tibicen by Mantelatto and García (2000).
According to Conover (1978), hermit crabs do not enter shells at random, but select them.Despite the low coefficients obtained in the regression analysis of data concerning their shells, we verified that C. atratum (large internal volume) was more adequate (considering the correlation coefficients) than M. nodulosa.The dimensions of L. nassa shells that have a straight aperture and an elongated shape (Rios, 1994) are those that best fit the dimensions of P. tortugae.Pisania auritula has a stronger shell with a straight aperture.The males, being larger and heavier, fit large shells, but they are not always available, occupying P. auritula probably because its favorable characteristics and greater availability.The ovigerous females of the studied species may have had available shells that they favoured as it was showen that three shells (P.auritula, C. atratum and M. nodulosa) had significant higher occupation rates by ovigerous females.The highest fecundity of P. tortugae in Anchieta Island was observed for ovigerous females occupying larger shells, i.e.P. auritula (Mantelatto et al., 2002).According to this work, the reproductive aspects of P. tortugae seem to be related to strategies developed to compensate for interspecific competition, i.e., high and continuous reproductive effort, early maturity, low fecundity and larger eggs produced.Almost all these strategies were correlated with the fact that the specimens attain maturity quickly at smaller size (Mantelatto and Sousa, 2000).
In this respect, we may infer that the shell utilization pattern of P. tortugae follows the pattern of shell-type availability that is related to the size and weight of the shells and varies between sexes.The better adequacy of L. nassa for the hermit dimensions in this survey lead us to conclude that this species is the one that best fits P. tortugae.On the other hand, the individuals fit large shells, demonstrating that P. tortugae do not always have the most adequate shells available.
FIG. 4. -P.tortugae.The regressions between hermit shield length (SL) and shell dry weigth (SDW), shell aperture width (SAW), shell aperture length (SAL) and shell internal volume (SIV) for the individuals as a whole.
268 F.L.M. MANTELATTO and L.C.C. DOMINCIANO FIG. 3. -P.tortugae.Hermit crab size frequency as a function of the four shells more occupied.

TABLE 2 .
-P. tortugae.Regression equations in relation to the sex of the individuals that demonstrated the highest correlation coefficients (N = number of individuals; r = coefficient of correlation; SL = Shield length; SDW = Shell dry weight; SIV = Shell Internal Volume; * = significant correlation [P < 0.05]).