Stoner and Sandt: Habitat quality for Strombus gigas in seagrass meadows 



697 



high-biomass site (L2) than at any other site (P<0.05), 

 and zero at the sand site (C3) (Fig. 2). The same pat- 

 tern occurred with living seagrass biomass, except that 

 site CI had significantly higher biomass than the other 

 moderate-biomass sites, but less than that found at the 

 high-biomass location (L2) (ANOVA, F 58.12, P< 

 0.001; Neuman-Keuls test, P<0.05). 



For analysis of seagrass and macrodetritus standing 

 crops at the beginning and end of the experiment, 

 paired-comparison analysis of variance was conducted, 

 where differences between cages and the effect of time 

 were examined. Significant difference between cages 

 occurred only at Nl where enclosure 2 had a higher 

 biomass of living seagrass than enclosure 1 (F 7.46, 

 P<0.05). Paired cages had similar amounts of macro- 

 detritus at all sites (P>0.05). Significant differences 

 with time were found at LI where detritus decreased 

 (F 13.22, P< 0.005), while living seagrass increased 

 with time (F 20.15, P<0.001). Seagrass increased also 

 at C2 (F 23.12, P<0.001) and at L3 (F 34.89, P< 

 0.001). No other significant time effects occurred for 

 either macrodetritus or seagrass biomass. 



Mortality 



Mortality at experimental sites with natural popula- 

 tions of conch was low, ranging from zero at site CI 

 to a total of only four individuals during the last three 

 weeks of the experiment at site Nl (Fig. 3). Total mor- 

 tality at the sites with moderate seagrass density and 

 no resident conch was also low: nine individuals at sta- 

 tion C2 and 19 at LI. At both of these sites, mortality 

 rates decreased with time. Mortality was high at all 

 other sites, with either a constant (sites L2 and L3) or 

 accelerating rate (C3 and Wl). By the end of the ex- 

 periment, 61 animals had died at the high-biomass site 

 (L2) and 61 had died at one of the low-biomass sites 

 (L3). At the sand site (C3), the transplant experiment 

 was terminated after 95 days because of rapidly ac- 

 celerating mortality. 



Kruskal-Wallis test was insufficiently powerful to 

 distinguish mortality differences among the sites (H adj 

 13.37, P 0.064) despite mortality rates ranging from 

 to 129% of the original number of conch transplanted. 

 This is related to the low number of degrees of freedom 

 in the design (2). 



Growth 



Growth rates decreased over the three experimental 

 periods at most of the stations, but the decrease was 

 variable among stations, and most extreme at Wl (Fig. 

 4). Because of station-time interaction in growth rates, 

 Kruskal-Wallis analyses were performed for each 

 growth period to test for station effects. In all cases 



Figure 3 



Total mortality of juvenile queen conch shown as a function 

 of time at eight transplant sites near Lee Stocking I. Values 

 are direct counts where 48 conch were transplanted to each 

 site at the beginning of the experiment and replaced upon 

 death. There was zero mortality at station CI (not shown). 

 Conch were resident at stations CI and Nl. 



the effect of site was highly significant (H adj 164, P 

 <0.001 for period 1; H adj . 234, P<0.001 for period 2; 

 and H adj . 209, P< 0.001 for period 3). 



During the first growth period, there were no signifi- 

 cant differences between growth rates at low-biomass 

 site Wl and moderate biomass site C2 (Mann- Whitney 

 U-test, U 993, P 0.441), between moderate-biomass 

 sites CI and LI (U 858, P 0.297), or between sites Nl 

 and L2 (U 825, P 0.572). Growth rates at Wl and C2 

 were higher than all other sites (P<0.05). Rates at sites 

 L3 and C3 were different from all other sites (P< 

 0.05). During this period there were no significant 

 differences in growth rates between paired cages 

 (P>0.05). 



During period 2, growth rates at CI, C2, and Nl 

 were higher than those at the five other stations (P 

 <0.05) and different from one another (P<0.05). 

 Despite negative growth (shell erosion) at L2 and L3, 

 growth rates at L2, L3, C3, and Wl were not different 

 (P>0.05). A significant difference in growth rates be- 

 tween cages at C3 (U 371, P 0.001) was the result of 

 new animals being introduced in the rebuilt cage 1 dur- 

 ing this period; however, growth rates were low in both 

 cages. 



As in period 2, highest growth rate during period 3 

 was observed at the moderate-biomass site C2 where 

 there was no resident conch population. Rates at Wl, 



