Fig 1.
Islands/eco-islands of this study.
A: Dongsha, B: Siziwan, C: Xiaoliuqiu, D: Lanyu, E: Howan, F: Tongpanyu, G: Huayu, H: Giang-jun-ou-yu, and I: Yuanbeiyu.
Fig 2.
Coenobita rugosus. Identification of shell condition, with examples of (a) disadvantageous, upper row, and (b) advantageous shells, lower row.
Fig 3.
Coenobita rugosus. Comparison of body sizes (shield lengths) at different sites.
Whiskers indicate the entire ranges, boxes indicate the 25th and 75th percentiles of the groups, and lines indicate the median values. A: Dongsha, B: Siziwan, C: Xiaoliuqiu, D: Lanyu, E: Howan, and H: Giang-jun-ou-yu.
Fig 4.
Coenobita rugosus. Comparison of the 95th percentile sizes of shield lengths from various islands.
Error bars indicate 95% confidence intervals estimated using the resampling method. A: Dongsha, B: Siziwan, C: Xiaoliuqiu, D: Lanyu, E: Howan, and H: Giang-jun-ou-yu.
Table 1.
Coenobita rugosus. Comparison of shell conditions at six sites.
All comparisons are against Dongsha.
Table 2.
Coenobita rugosus. Cohort structures at two sites according to FISAT II estimation of cohort composition.
Fig 5.
Coenobita rugosus. Size structures and cohort analyses for Dongsha in 2013.
Black dots indicate the means and standard deviations of each cohort, as estimated using FISAT II.
Fig 6.
Coenobita rugosus. Size structures and cohort analyses for Dongsha in 2014.
See Fig 5 for other legend.
Fig 7.
Coenobita rugosus. Size structures and cohort analyses at Siziwan in 2013.
See Fig 5 for other legend.
Fig 8.
Coenobita rugosus. Size structures and cohort analyses at Siziwan in 2014.
See Fig 5 for other legend.
Fig 9.
Coenobita rugosus. Comparison of size difference (in shield lengths) between cohorts 1 and 2.
Fig 10.
Coenobita rugosus. Comparison of annual increment of the first cohort of 2013 with the second cohort in 2014 for Dongsha and Siziwan.
Fig 11.
Coenobita rugosus. Comparison of the condition indices of males on Dongsha (n = 5) with those at Siziwan (n = 8).
(P < 0.01, Mann–Whitney U test). Data points may overlap in the figure.
Fig 12.
Coenobita rugosus. Comparison of the condition indices of Dongsha and Siziwan males.
The slopes of the two populations are not significantly different (P = 0.69, ANCOVA); the intercepts, however, are significantly different (P = 0.03, ANCOVA), with the Dongsha samples exhibiting a greater intercept.
Fig 13.
Coenobita rugosus. Results of food preference test of hermit crabs collected from Dongsha and Siziwan.
A significant difference was found in the Dongsha crabs but not in the Siziwan crabs (P = 0.015 and P = 0.83, respectively, Wilcoxon signed-rank test). The diagonal line indicates a 1:1 dietary consumption.
Fig 14.
Coenobita rugosus. Comparison of size increment of crabs from Dongsha in the growth experiment.
P < 0.01 for Dongsha crabs, P = 0.27 for Siziwan crabs, Mann-Whitney U Tests.
Table 3.
P values of Mann–Whitney U tests comparing crabs according to diet and location.
Underlined categories indicate greater increments in cases of significant difference.
Table 4.
Horizontal distribution of Coenobita species at the shore.
Table 5.
Altitudinal segregation of Coenobita spp.
Table 6.
Coenobita rugosus. Number of individuals with and without congeners in the same traps.
Fig 15.
Coenobita rugosus. Comparison of density index (count/trap) at each study site (P = 0.65 with empty traps omitted or 0.22 when including all traps, Kruskal–Waillis tests).
A: Dongsha, B: Siziwan, C: Xiaoliuqiu, D: Lanyu, E: Howan, and H: Giang-jun-ou-yu.