Material and methods
Hermit crabs were obtained monthly from January to December 1998 from the infralittoral rocky shores of South, East, and Small Beaches and Wind Bay of Anchieta Island (23°33?S, 45°05?W), an area recently declared an ecological reserve of Sao Paulo State. Every month, at least one but no more than two of these sites were sampled depending on accessibility by boat because of the ocean (waves and currents) and weather (wind) conditions. The infralittoral area of the collection sites present an irregular surface with many large boulders, are nine meters in depth, and differ only in their location but not varying on bottom and abiotic conditions. This sample design was adopted echat giriÅŸ to avoid negative impact on this ecological reserve and affecting the population; and because no significant difference was detected between the population structures of P. tortugae from the areas mentioned above ( Mantelatto and Sousa, in press). The hermit crabs were collected mainly on the rocks covered by coral colonies (Palythoa) and/or on the transition area between rocky/sandy bottoms. Collections were made during the daytime by three persons using diving equipment over a period of 30 min over an area of about |$850 \ <\rm>^<2>$|? . 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, 2001). After collection the animals were frozen and transported to the laboratory where they were thawed just before analysis. Each ovigerous female was removed by carefully cracking each shell, weighed (Female Wet Weight, FWW), and measured on the basis of shield length (SL = from the tip of the rostrum to the V-shaped groove at the posterior edge of the shield). Measurements were made with a caliper rule (0.1 mm) or, in the case of the small individuals, by drawing with the aid of a camera lucida. To better visualize the results in terms of seasonal fecundity, we grouped the monthly samples into summer (January, February, and March), autumn (April, May, and June), winter (July, August, and ples.
Cover Data
Shell species were identified according to Rios (1994) and by Dr. Osmar Domaneschi from the Bioscience Institute of the University of Sao Paulo. The following measurements were made: Shell Dry Weight (SDW), Shell Aperture Width (SAW), and Shell Aperture Length (SAL). Shell internal volume (SIV) was measured by the method of Bertness (1981a), determining the volume of sand |$( <\rm>= 0.25-0.125 \ <\rm>)$| required to fill the empty shell. To determine correlations between fecundity and shell variables, regressions and correlation coefficients were computed. The ANOVA on Ranks test was used to compare the fecundity between the four most occupied shell species, and the fecundity of the ovigerous females occupying heavier shells |$(3.0 \lt <\rm>\lt 7.55 \ <\rm>)$| in relation to those ones occupying lighter shells |$(0.57 \lt <\rm>\lt 2.99 \ <\rm>)$|? . The level of significance was |$P \lt 0.05$|? .
Fecundity
Of a total of 840 ovigerous females captured, only 160 were used for fecundity analysis, i.e., we selected hermit crabs with embryos in developmental stages from late blastula to early gastrula (according to the methodology of Mantelatto and Garcia, 1999, modified from Boolootian et al., 1959: Initial Stage: stages 1 to 4; Intermediate Stage: stages 5 to 8, and Final Stage: stages 9 and 10) in order to avoid confounding effects of embryonic diameter swelling during the late stages of development and because of egg loss from the brood during incubation. The eggs were carefully removed from the pleopods and counted under a light stereomicroscope. Frequency distributions of SL and fecundity were calculated for the females whose eggs were quantified. After the frequency tabulation, mean egg number of brooded egg mass was determined for each class. The average volume (mm 3 ) of an egg was calculated according to Jones and Simons (1983) from the formula |$1/6\pi \ <\rm>^<3>$| (where I is the mean of two diameters, including the chorionic membrane tightly adhering to the embryonic surface) of a subsample of 15 eggs per female. The longest and shortest diameters of each egg (the eggs tended to be oval) were measured under a compound light microscope with a camera lucida. Data were analyzed statistically by regression of dimensions plots |$( <\rm>= <\rm>^<<\rm>>)$|? . The ANOVA on Ranks was used to compare the fecundity seasonally and between individuals of different sizes (small ones > 1.8 le; SL le; 3.0 mm; large ones > 3.1 le; SL ? 6.7 mm). The level of significance was |$P \lt 0.05$|? .
