Shell preference of the hermit crab Pagurus exilis ( Anomura : Paguridae ) from Brazil and Argentina : a comparative study

Shell preferences, as shown by laboratory choice experiments, are important determinants of shell utilization under natural conditions. Size and shell species preferences of the hermit crab Pagurus exilis were determined and compared for the most occupied shell types (Olivancillaria urceus, Natica isabelleana and Buccinanops gradatum) in the Caraguatatuba region, Brazil, and for the two most occupied species (B. gradatum and N. isabelleana) in Mar del Plata, Argentina. All experiments were conducted under laboratory conditions, using glass aquaria where the hermit crabs were placed naked with a large number of shells of appropriate sizes. After 36 h the crabs and chosen shells were measured. The relationship between preferred shell type, shell size and shell volume was determined by regression analysis. Pagurus exilis showed significant choices amongst shell types. Specimens from Argentina preferred N. isabelleana and those from Brazil chose B. gradatum > N. isabelleana > O. urceus. Preferences were characterized by shell weight and internal volume. The present data lead us to conclude that shell selection by hermit crabs involves individual preferences related to the shell features that best provide protection and survival. Rev. Biol. Trop. 55 (Suppl. 1): 153-162. Epub 2007 June, 29.

Animals often must compromise among different sets of demands when selecting resources (Yoshino et al. 1999).In hermit crabs particularly, shell selection is not by chance but based on adequacy and availability of resources (Reese 1962, Conover 1978), and is affected by both shell size and species (Abrams 1978, Conover 1978, Blackstone 1985, Lively 1988, Siu and Lee 1992, Ohmori et al. 1995, Hahn 1998, Rodrigues et al. 2000, Mantelatto and Dominciano 2002, Mantelatto and Meireles 2004).
Moreover, the fitness of a particular shell may differ between hermit crab species, reflecting several selection pressures, which associated with different habitats, act in different ways on each crab species (Bertness 1981) in different areas.Although the shell selection process has been well investigated (Vance 1972, Elwood et al. 1979, Bertness 1980, Wilber 1990, Hazlett 1992, 1996, Ohmori et al. 1995, Wada et al. 1997, Garcia and Mantelatto 2001, Meireles and Mantelatto 2005), there is a scarcity of studies about the shell preferences of two different populations of the same species from different locations distantly removed from one another.Comparative studies of species known to occur over wide geographical ranges can provide valuable information on the development of intraspecific adaptations to different environmental conditions.Unfortunately, the number of studies conducted on this subject is too limited to permit an understanding of the life cycle strategies of the high number of described decapod crustaceans worldwide (FL Mantelatto pers. obs.).
According to Hendrickx and Harvey (1999), three groups of hermit crabs in particular (i.e., the genera Pagurus, Paguristes, and the "Pylopagurus-like" species) have been difficult to study because there are many stillundescribed species.Many species are known from one or a few localities, and many stillundescribed species are known to occur worldwide.The genus Pagurus Fabricius, 1775 is a heterogeneous pagurid group consisting of more than 170 species worldwide (Lemaitre and Cruz Castaño 2004), and ecological information for this genus is still scarce despite the easily accessible habitats it inhabits and its high abundance (Goshima et al. 1996).
Despite of a number of publications about the distribution of Pagurus exilis (Benedict, 1892) (Melo 1999), studies on the species have been limited to the post-embryonic development under laboratory conditions (Scelzo and Boschi 1969) and on distributional aspects and population structure in Brazilian waters (Terossi et al. 2006, Mantelatto et al. in press, respectively).Here we evaluated the shell size and species preference of two different biogeographical populations (Brazil and Argentina) of P. exilis under laboratory conditions, comparing the puzzling mechanisms of shell selection.Laboratory experiments followed the methodology adopted by Garcia and Mantelatto (2001) in almost all aspects; each animal was utilized only once to avoid any acquired behavior and returned to the field in a different area from previous capture (except those who died in laboratory or used as voucher).In all experiments, hermit crabs were removed from their chosen shells by heating the apex of the shell in hot water.The data of animals that died during the experiments were not included in the analysis.

Hermit
To determine correlations between characteristics of hermit crabs and their preferred shells, regression analyses were computed.The chi-square test (χ 2 ) was used to compare occupancy of shell species, and morphometric relationships were established by regression analysis and correlation coefficients (r) (Zar 1996).Shell size preference data were analyzed using the multiple linear regression: lnY= a + blnX, (where Y = shell measurements: SAL = shell aperture length; SAW = shell aperture width; SDW = shell dry weight; SIV = shell internal volume, and X = hermit crab measurements: SL = shield length, and W = hermit weight).
The experiments were not divided by sex because we assumed that the presence of the three groups (males, non-ovigerous and ovigerous females) in each test would be important to avoid disturbance related to hierarchy behavior.Furthermore, the absence of one or other sexual group could lead the individuals to some behavior different from that observed in nature (Meireles and Mantelatto 2005), where they were found grouped.
Shell-species preference.All experiments were conducted in a pair-wise fashion.In these tests, 15 naked hermit crabs were placed in the aquarium with 150 empty shells of two different species (75 of each species), various and adequate sizes for each locality.
Following 36 h in this free-choice situation, hermit crabs were removed from their chosen shells, weighed (HW) and measured for SL.Shell species were identified and the SDW, SAW and SIV were determined.Measurements were made with a Vernier caliper (0.1 mm) or by measuring the drawings of the specimen taken under a light stereomicroscope equipped a camera lucida.Tests were replicated three times.Some voucher hermit crabs were deposited in the Crustacean Collection of the Biology Department of FFCLRP (CCDB), University of São Paulo (catalogue numbers from 1201 to 1204).
Shell-size preference.The experiments were conducted for each of the two shell species separately (Buccinanops gradatum and Olivancillaria urceus).In each test, 15 naked hermit crabs were placed in aquarium with a minimum of 150 empty shells (10 shells/hermit crab) of various and adequate sizes.After 36 h, the hermit crabs were removed from their preferred shells, and both were measured as described above for the shell-species experiments.Each test was repeated three times.

RESULTS
Shell-species preference.A total of 137 animals were utilized.P. exilis showed a significant choice amongst the gastropod shell types.Specimens from Argentina preferred N. isabelleana, and those from Brazil chose: B. gradatum > N. isabelleana > O. urceus (p < 0.01) (Table 1).In general, the morphometric relations that best described the association between Brazilian and Argentinean hermit crabs and their selected shells species were those involving shell dimensions and hermit weight, independent of the shell species (Table 2).
Shell-size preference.During shell-size experiments, a total of 172 animals were utilized.Experiments with specimens from Brazil revealed that the hermit crab's size preference was strongly associated with the shell type.Low correlation coefficients were observed in the less preferred shell type (O.urceus), while in B. gradatum a strong association between hermit crab size and the shell dimensions was found.Experiments with specimens from Argentina revealed a different pattern.In spite of N. isabelleana being the most preferred, the correlation coefficients of hermit crabs and B. gradatum to the shell size experiments were higher (Table 3).

DISCUSSION
Recent investigations with molecular analysis by 16S mtDNA made by one of us (FL Mantelatto) have showed that no marked genetic variability occurs among southern populations of P. exilis from both regions (Brazil and Argentina) in the southwestern Atlantic.Therefore, the general results were not influenced by genetic variability.Despite of this, some preliminary results have detected ecological differences in distribution and size of specimens between the above mentioned populations (FL Mantelatto and MA Scelzo, unpubl. data).The need to explain this evolutionary relationship certainly inspired the present study.
Animals often must compromise among different sets of demands when selecting resources (Yoshino et al. 1999).Hermit crabs do not necessarily live in shells they prefer, and the availability of different shell types and contact with competitors for empty shells influences shell occupation (Bertness 1980).P. exilis showed significant choices amongst the gastropod shell species in the two studied regions.Specimens from Brazil preferred B. gradatum over N. isabelleana and O. urceus shells, while those from Argentina preferred N. isabelleana over B. gradatum.Considering that in the Ubatuba region O. urceus is one of the most available shells (AL Meireles pers.obs.) and it is also occupied by other hermit species, its availability to P. exilis not necessarily determines this shell species occupation.In Argentina, there is a lack of studies on gastropod availability but according to Scelzo et al. (2004), Buccinanops shells are the most occupied by P. exilis in the field followed by Natica.
Previous studies conducted by Abrams (1978), Elwood et al. (1979), Siu and Lee (1992), Ohmori et al. (1995), Hahn (1998), Dominciano and Mantelatto (2004), Biagi et al. (2006) found that the hermit crabs in the laboratory preferred those shell species most occupied in the field.A similar pattern of shell selection also was found in the present study.
The shell size experiments revealed that O. urceus shells provided a low specimen adequacy to P. exilis.This fact, associated with the preference for B. gradatum in Brazil, suggested that O. urceus shells are the morphometrically less adequate within the most occupied shell species in the field (Terossi et al. 2006).According to Yoshino et al. (1999), there is a trade-off between shell size and species preference, and that less preferred shell species are actively chosen when the more preferred shell species the crabs encounter frequently in the field are of a less suitable size.Considering the great availability (in size and number) of O. urceus shells in the field (AL Meireles pers.obs.), we may infer that the assertion proposed by Yoshino et al. (1999) may not be applied to P. exilis and that it may not have exhibited preference for O. urceus because of its higher weight when compared to the other shells.
Specimens from Argentina demonstrated no particular preference among the shell dimensions studied.However, higher correlation coefficients were obtained for Buccinanops complex shells despite the fact that N. isabelleana were the most preferred in the shell type experiments.In contrast, P. exilis showed different patterns of occupation in the field observations carried out in Argentina, where Buccinanops gradatum complex species were the most occupied shells (66 %) compared with N. isabelleana (19 %) (Scelzo et al. 2004).As demonstrated for Loxopagurus loxochelis (Moreira, 1901) by Biagi et al. (2006), populations of the same species inhabiting different and distant areas may present distinct shell selection behaviors.P. exilis specimens from Argentina presented no preference among the tested shells under laboratory conditions; in contrast, Brazilian ones preferred one shell instead of the other tested.Furthermore, according to the latter authors this discrepancy is a function of the hermit crab species, the origin of the population as well as the availability of the resources in the area of occurrence.
In the present investigation, P. exilis from different areas exhibited different preferences among shell species which lead us to conclude that some particular differences in shell preferences may be attributed both to the morphology of hermit crabs and to environmental conditions, with the animals looking for protection against the factors to which they are exposed (predation, energy expenditure, osmotic stress and wave action) (Garcia and Mantelatto 2001).
The P. exilis populations from Brazil and Argentina present differences in the mean size of the specimens, with Argentinean hermit crabs reaching larger sizes (5.46±0.59mm SL) than the Brazilian ones (5.01±0.86mm SL).This pattern is in accordance with the one proposed by Abele (1982), who postulated that the size of crustaceans decreases with decreasing latitude.Furthermore, according to Ohmori et al. (1995) shell utilization patterns would change with different crab sizes, even within the same species.This might explain the differences found in the patterns of shell species preference between both hermit crabs populations.
We are in agreement with Lancaster (1988) that "hermit crabs are particularly good at solving the problems of limited resources and are efficient at both exploiting what they have and contesting what they have not".
In the present study, P. exilis preferred lighter shells that promote less protection but with greater internal volume (B.gradatum and N. isabelleana).We believe that this preference is strongly related to energy savings where males would invest in fast growth as they can enjoy high levels of mating when occupying adequate shells (Harvey 1990) and females in reproductive activities.Also, considering the high mobility and fragile body structure exhibited by P. exilis and its great abundance (the second most abundant species) in the nonconsolidate infralittoral area in São Paulo coast (Meireles et al. in press), it would be expected that the occupation of a shell would facilitate its behavior the occupation of a shell that would facilitate its behavior and guarantee survivorship.However, further studies on reproductive and growth aspects may contribute to a better understanding of the role of the shell utilization in the population dynamics of P. exilis along the western Atlantic as well as the relationship of shell occupation in relation to the coexistent species L. loxochelis.
crabs were collected in Caraguatatuba Bay (23º 47' S and 45º 08' W) (Brazil) during 2002 and in Mar del Plata and Mar Chiquita (38º 02' S and 57º 31' W) (Argentina) during 2003 by a fishing boat equipped with otter-trawl nets.Both study localities are approximately 5.000 km separated from each other.The animals collected in Brazil were transported to the Laboratory of Bioecology and Crustacean Systematics at the Faculty of Philosophy, Sciences and Letters (FFCLRP), University of São Paulo (USP), Ribeirão Preto, São Paulo, Brazil; those collected in Argentina were taken to the Department of Marine Sciences, Faculty of Exact and Natural Sciences, Mar del Plata National University (UNMdP), Mar del Plata, Argentina, where experiments were conducted using glass aquaria provided with flowing sea water.

TABLE 1
Pagurus exilis.Shell species preference among the shells offered during the experiments Regression equations for Brazilian and Argentinean chosen shell species r = correlation coefficient; SL = shield length; W = hermit crab weight; SAW = shell aperture width; SAL = shell aperture length; SDW = shell dry weight; SIV = shell internal volume.* significant correlation, p< 0.05; ns = not significant