(A.1)
Appendix A. A description of the manner in which crab spawning at spring tide was assessed.
The embryonic incubation of crab larvae has a defined time span, during which the embryo passes through a series of development stages (egg stages). We classified eggs in 5 stages (see Methods of main text) from stage 1 (newly laid) to stage 5 (empty egg cases, larvae hatched). Stage 5 is transient and was not recorded in Kenya.
We determined whether populations spawned at spring tide by visual inspection of two types of time frequency diagrams: (1) an egg stage index (ESI) was plotted against sampling date to examine if egg ripeness peaked rhythmically at NM and FM over the year. ESI was calculated per sample day for each species as
|
|
(A.1) |
where i was the egg stage, ni the number of females in the i'th egg stage and N the total number of egg-carrying females in the sample. ESI values range between 1.0 (all females in stage 1) and 5.0 (all females in stage 5). (2) Egg stage frequencies were plotted against the fortnightly spring-neap tide time-cycle to examine if embryonic incubation progressed systematically in order to reach hatching point (stage 4–5) at spring tide. There are ~25 spring-neap tidal cycles in a 365 day calendar, each consisting of 15 days (14.77 d); we translated our sampling dates into their respective spring-neap tidal days (d 0 to d 14; d 0 were dates of max amplitude spring high tides) and determined the percentage frequency-distribution of each egg stage over these 15 days. Note: spring tide spawning is normally determined using circular statistics and/or time series analysis. These techniques were not applicable because of variable time-intervals between the capture of gravid females, and since our data was not larval release times, but egg stage frequencies. Instead, visual inspection of the data was used; this is an accepted and sometimes superior approach (Saigusa 2001).
The results showed that rhythmic peaks could be detected in the egg stage indices of all species, except Uca vocans (Fig. A1-A2), although rhythms were less conspicuous in the Zanzibar data, due to a lower sampling frequency. Egg stage indices peaked with lunar (N. meinerti, C. ortmanni, U. annulipes) or semilunar periodicity (P. guttatum, U. inversa), with maxima around NM and FM, suggesting that populations spawned at spring tide. Rhythmicity was not discernable in the egg stage index of U. vocans because of low sampling frequency (Fig A2).
|
| FIG. A1. Egg stage indices of 5 Kenyan crabs (brackets = samples), and Kenyan high tides (quick wave = amplitudes of daily high tides; slow wave = difference between amplitudes of new moon (NM) and full moon (FM) spring high tides). |
|
| FIG. A2.Egg stage indices of six crabs from Zanzibar, and Zanzibar high tides. Details are as in Fig. A1. |
Figures A3 and A4 confirm that all species were spring tide spawners: egg development progressed steadily over the spring-neap tidal period and culminated with larval release (i.e., last appearance of stage 4 and first appearance of stage 5) during spring tides (shaded areas of Fig. A3-A4).
 |
| FIG. A3. The frequency of egg stage 1, 2, 3, or 4 over the fortnightly spring-neap tidal cycle (days 0 and 15 were days of maximum amplitude high tides), in five crabs from Kenya. Shaded areas represent the spring tide period nearest to larval release. Brackets show the number of samples and the number of females per egg stage. Note, there are two spring-neap tidal cycles back-to-back in the x-axis because most of our species take >15 d to ripen their eggs; they will lay their eggs inside one cycle, but not hatch them until the end of the next. |
|
| FIG. A4. The frequency of egg stage 1 to 5 over the spring-neap tidal cycle, in six crabs from Zanzibar. Other details are as in Fig. A3. |
LITERATURE CITED
Saigusa, M. 2001. Daily rhythms of emergence of small invertebrates inhabiting shallow subtidal zones: a comparative investigation at four locations in Japan. Ecological Research 16:1–28.