NOTE Caillouet et al.: Sea turtle strandings and shrimp fishing effort in northwestern Gulf of Mexico 715 



tions between 48 pairs (12 months x 4 years) of In 

 (S + l) and ln(E + l) were determined. 



Results 



Scatter plots for geographic zones and depth intervals 

 for which ln(S + l) was significantly (P<0.05) cor- 

 related with ln(E + 1) are shown in Figure 2. On the 

 upper coast, the three correlation coefficients, r, that 

 differed significantly from zero were positive and 

 occurred with fishing effort in the 0-5, 5-10, and 10-15 

 fm intervals. On the lower coast, r was significantly 

 different from zero and positive only with fishing ef- 

 fort in the 5-10 and 10-15fm intervals. These correla- 

 tions indicated that turtle strandings increased as 

 fishing effort increased in waters landward of 15 fm. 

 Correlation coefficients for fishing effort in other depth 

 intervals within the two zones did not differ significant- 

 ly from zero. 



The five significant correlations (P<0.05) were 

 detected despite the relatively coarse temporal-spatial 

 scale of the data sets (Fig. 2). Although they were of 

 moderate strength, ranging from 0.327 to 0.512, even 

 the lowest among them had a very small probability 

 (P 0.0232) of occurring due to chance alone (Fig. 2). 

 P was even smaller for the other four significant cor- 



relations. There was no significant (P>0.05) heterog- 

 eneity among the five correlation coefficients. 



The means of the transformed strandings for the 

 upper and lower coasts did not differ significantly 

 (Table 3). However, the upper coast exhibited more 

 months in which there were no strandings than did the 

 lower coast (Fig. 2). The mean of transformed fishing 

 effort on the lower coast was significantly higher than 

 that for the upper coast, indicating a higher average 

 fishing intensity on the lower than upper coast in 

 waters landward of 30 fm. 



Loggerheads and Kemp's Ridleys occurred most 

 frequently in the strandings, followed by hawksbills 

 Eretmochelys imbricata, greens Chelonia mydas, and 

 leatherbacks Dermochelys coriacea (Table 4). Turtle 

 strandings occurred year-round with peaks in April and 

 May, and with a secondary peak in August. Annual 

 strandings declined over the years covered by the 

 study, with 417, 259, 188, and 183 strandings reported 

 in 1986-89, respectively. 



Discussion 



The distributions of sea turtles and shrimp trawling 

 must overlap to some degree because it is well docu- 

 mented that sea turtles are caught in shrimp trawls 



