246 



Fishery Bulletin 94(2), 1996 



depth intervals suggested that monthly fishing in- 

 tensities in contiguous depth intervals increase or 

 decrease in the same direction. This pattern also 

 suggests that monthly fishing intensities are high 

 in deeper waters when they are low in shallower 

 waters, and vice versa, probably in association with 

 the seasonal migrations of shrimp and the fishing 

 fleet (Tables 5-8). Thus, when fishing intensities in 

 depth intervals between and 20 fin were positively 

 associated with strandings, the fishing intensities 

 seaward of 20 fm were low, producing inverse asso- 

 ciations with stranding rates. 



The pattern of correlations between the \n(E + 1) 

 of the various depth intervals may have influenced 

 the correlations between ln(S + 1 ) and ln(i? + 1 ). Some 

 of the significant positive correlations detected be- 

 tween ln(S + 1 ) and ln(£ + 1 ) may have resulted from 

 correspondences between stranding rates and fish- 

 ing intensities, whereas others may have been coin- 

 cidental. Our analyses determined only that there 

 were significant positive correlations between 



ln(S + 1) and \n(E +1), but they cannot prove there 

 was a cause and effect relationship between fishing 

 intensities and stranding rates. On the other hand, 

 our results would be consistent with a cause and ef- 

 fect relationship. Although the positive correlations 

 we detected were not significantly heterogeneous, 

 they did tend to decrease in strength with depth, 

 which may suggest a decreasing degree of statistical 

 association between stranding rates and fishing in- 

 tensities as depth increases. This tendency also could 

 occur if sea turtles injured or killed in deeper waters 

 were less likely to wash ashore than those injured or 

 killed closer to the shoreline (Caillouet et al., 1991 ). 

 The significant positive correlations we detected 

 between ln(S + 1) on ln(E + 1) were circumstantial 

 evidence of a relation between sea turtle stranding 

 rates and shrimp fishing intensities, but they did not 

 demonstrate that shrimping caused the strandings. 

 Nevertheless, they were consistent with previous 

 findings that sea turtles are caught and killed inci- 

 dentally in shrimping (National Research Council, 



