The influence of shell species and size on the shell selection pattern of Paguristes tortugae ( Decapoda , Diogenidae ) from Anchieta Island ( Ubatuba , Brazil )

Shell selection by the hermit crab Paguristes tortugae Schmitt, 1933 from Anchieta Island (Brazil) was analyzed using the six most frequently occupied shell species in the field and taking into account the sexual condition of the individuals, the shell size and the shell species. The experiments were conducted under laboratory conditions and the shell species preference was estimated on the basis of the frequency that each species was chosen by the individuals. The preferred shell species and size were determined by regression analysis. The highest correlation coefficients were obtained for the relations between the hermit dimensions and shell dry weight. The ovigerous females preferred shells with larger internal volume: Leucozonia nassa (Gmelin, 1791) and Cerithium atratum (Born, 1778). In the experiment of shell size, males preferred heavier shells whereas females selected the shape characteristics of the shell, such as the aperture and the internal volume, which are probably related to the growth and offspring guarantee, respectively. In general, and independent of sex condition, P. tortugae showed significant selection among all shells utilized. The results suggest that shell selection by P. tortugae involves sexual and reproductive condition preferences.


INTRODUCTION
According to REESE (1962), the hermit crabs may distinguish shells of various species and also recognize a prior occupation.They can discriminate the weight of shells of the same species, as well as their form, cover and dimension, possibly selecting adequate shells with proper weight (MARTINELLI & MANTELATTO, 1999).
During the last three decades, shell selection process has been well investigated (GARCIA & MANTELATTO, 2001), especially with regard to the interesting and sometimes unclear mechanism developed by hermit crabs.Considering the large number (46) of hermit crab species recorded in Brazilian waters (MELO, 1999;MANTELATTO et al., 2001), studies on shell selection behavior involving field and laboratory comparative observations have not been developed, except for Calcinus tibicen (Herbst, 1791) in Ubatuba region, São Paulo State (GARCIA & MANTELATTO, 2001).The genus Paguristes Dana, 1852, represented by more than 100 species distributed in all tropical and subtropical oceans (PROVENZANO, 1978), has been a particularly difficult group to study (MANTELATTO et al., 2002).
Paguristes tortugae Schmitt, 1933 has been found occupying 21 gastropod shell species in Ubatuba region, demonstrating plasticity in shell species utilization and a clear competitive ability among individuals of different sizes in the population (MANTELATTO & DOMINCIANO, 2002).Considering that previous studies have suggested that hermit crab shell preferences are important in mediating the shell utilization patterns found in natural populations (FOTHERINGHAM, 1976;BERTNESS, 1980;ELWOOD & NEIL, 1992;SIU & LEE, 1992;OHMORI et al., 1995;GARCIA & MANTELATTO, 2001), the present work aimed to evaluate the size and shell species preferences of males and females of P. tortugae from Anchieta Island as part of a long-term study undertaken to identify the most important parameters of shell occupation and shell selection by coexisting hermit crab species in a subtropical rocky intertidal area, through both laboratory and field observations.
Specimens were collected between February 2000 and March 2001 from infralittoral rocky/sandy areas of Anchieta Island, Ubatuba, Southern Brazil by SCUBA.Details of the methods of capture and study area can be found in MANTELATTO & DOMINCIANO (2002) and MANTELATTO & GARCIA (2002).All shell-selection tests were conducted in the laboratory.The size of crabs tested was checked and then experiments were performed using a set of shells that involved appropriate sizes for crabs.The hermit crabs were placed naked with a larger number of shells of appropriate sizes inside a glass aquarium (30 x 30 x 40 cm) with aerated seawater, and bottom compound DOMINCIANO & MANTELATTO of crashed shells and medium sand.Each animal was used only once in order to avoid any acquired behavior.After the experiments, some animals were returned to the field and others were deposited as vouchers at the Coleção de Crustáceos do Departamento de Biologia da Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto (FFCLRP), Universidade de São Paulo (USP).The animals that died during the experiments were not included in the analysis.
After 72 h (a period of time established on the basis of previous experiments, when more than 80% of the individuals analyzed stopped changing shells), the hermit crabs were removed from their chosen shells by heating the shells in water at ~100ºC (GARCIA & MANTELATTO, 2001).Each hermit crab was weighed (WW = Wet Weight) and measured for shield length (SL), and had its sex checked.The species of the chosen shells were identified and their aperture width (SAW), internal volume (SIV) and dry weight (SDW) were determined.
The shell species preference was estimated on the basis of the frequency with which each species was chosen by the individuals (x 2 -chi-square test; ZAR, 1996) and the morphometric relationships were established by regression analysis and correlation coefficients (r) (P < 0.05) performed with SigmaStat Statistical Software v. 2.0.
The shells of C. atratum and M. nodulosa were not compared with other species during this experiment because these shells are smaller than the others and were occupied only by small individuals (1.1 -4.7 mm SL).
The experiments were not performed separately in relation to the sex of individuals because we assumed that the presence of males and females in each test would be important to avoid effects related to hierarchy behavior, i.e., males are hierarchically superior to females and their absence could lead the females to some behavior different from that observed if males were present and, besides both sexes are found grouped in the wild (GARCIA & MANTELATTO, 2001).
Shell size preference.The experiments were conducted for each of the six shell species.In each test, 15 hermit crabs were placed naked in the aquarium with about 150 empty shells (10 shells for each hermit crab) of the same species and of adequate size.Four replicates of each test were done.After 72h, the hermit crabs were removed from their chosen shells by the same method as above and measured as described for the shell species experiments.
Results of the tests were analyzed using the following regression for each shell type: log Y = a + b log X (where Y = shell measurements: SAW, SIV, SDW and X = hermit crab measurements: SL and WW).It was established an equation that reflected the species and size of shells more adequate to the hermit crabs under study, established by significance analysis of correlation coefficients (P < 0.05; ZAR, 1996).

RESULTS AND DISCUSSION
A total of 403 animals were utilized (176 males, 137 non-ovigerous females and 90 ovigerous females) in the experiments of shell species preference.Males and females of P. tortugae showed different preferences for one of the two shell species offered during pair-wise experiments (tab.I).The order of shell species preference was L. nassa > P. auritula > S. haemastoma > T. viridula for larger individuals and C. atratum > M. nodulosa for smaller ones.
On the experiments of shell size preference 340 animals were utilized (145 males, 140 non-ovigerous females and 55 ovigerous females).Shell size experiments revealed that the hermit crabs' preference was significant in all relations analyzed to L. nassa and S. haemastoma.The hermit crabs showed strong association with shell weight and aperture width of L. nassa and S. haemastoma.In general the best correlations occurred between weight of hermit crab and shell weight and internal volume.The other shell species showed lower coefficients (P < 0.05) (tab.II).
The shell choice of males, ovigerous females and nonovigerous females showed different patterns for each shell species.Males that preferred M. nodulosa showed significant correlation with the internal volume and aperture width of shells, while non-ovigerous females showed significant correlation only with shell internal volume.Males presented significant correlations in all morphometric relations evaluated to Leucozonia nassa and S. haemastoma.In contrast, males and non-ovigerous females showed significant correlations only in few relations with P. auritula.Only non-ovigerous females presented significant correlation with the dimensions of C. atratum.In T. viridula, only males showed significant correlation.Non-ovigerous and ovigerous females showed different patterns between the dimensions of S. haemastoma.
The smaller individuals of P. tortugae showed a clear preference for C. atratum.This shell species has a larger internal volume compared to M. nodulosa, which is smaller and has spiral nodose ribs and strait aperture (RIOS, 1994).This choice may be related to the fact that small hermit crabs occupy larger shells to allow their future growth.According to REDDY & BISEWAR (1993), the aperture size, weight and shell internal volume are important correlates of the choice of shells.
In relation to larger size shells, the species more available and more occupied in the field (P.auritula) showed better correlation with the hermit crabs and also were preferred in laboratory by P. tortugae.Leucozonia nassa has angular returns (RIOS, 1994) and its architecture provides lightness, larger internal volume and larger aperture size when compared to P. auritula (spiral well-developed, heavier, with minor internal volume and aperture).The larger and heavier males (MANTELATTO & SOUSA, 2000) preferred L. nassa because of its architecture.Females (ovigerous and nonovigerous) did not show this significant preference, probably because the shells of both L. nassa and P. auritula are adequate to females' occupation.
According to BLACKSTONE (1984), hermit crabs show preference for shells previously occupied in the field, and the utilization and preference is a function of the shell availability.This pattern presents variability according to the adaptations of hermit crabs in the field.GARCIA & MANTELATTO (2001) showed that Calcinus tibicen did not present significant preference for shells under laboratory conditions when compared to shells utilized in the field.Paguristes tortugae showed a clear preference for shell species occupied in the field, compared to the high plasticity of shell occupation in Anchieta Island (MANTELATTO & DOMINCIANO, 2002).A similar pattern of shell selection according to field occupancy and differences in shell species selection between sexes was found in Pagurus brevidactylus (Stimpson, 1859) in the same area (Andrea Meireles, pers. comm.)In relation to sex, the ovigerous females preferred P. auritula and L. nassa instead of S. haemastoma, what was probably related to egg protection.According to

Table I .
Shell species preference of Paguristes tortugae among the six most frequently occupied shells in the infralittoral rocky/sandy areas of Anchieta Island, Ubatuba, São Paulo (c 2 -test value, in brackets; n/n, number of occurrence of the shell species chosen by the individuals; ns, not significant;*, significant value, P < 0.05).The influence of shell species and size on the shell selection pattern of Paguristes tortugae...