240 



Fishery Bulletin 94(2), 1996 



Fontaine et al. , 1993 ), because their temporal-spa- 

 tial distribution was determined in part by when 

 and where they were released (Manzella et al., 

 1988; Caillouet et al., 1995); 



• small, pelagic-stage turtles, because they were not 

 considered vulnerable to incidental capture in 

 trawls; 



• strandings for which likely or possible causes had 

 been assigned. 



In some cases, particular codes that defined records 

 to be deleted did not occur among the 2,445 records 

 examined (i.e. they had zero frequency, Table 1 ), but 

 this was not known a priori. The 1,555 records (937 

 from 1986-89 and 618 from 1990-93) retained for 

 analysis represented strandings of wild sea turtles 

 for which no known or likely cause of mortality had 

 been assigned. As a consequence of more stringent 

 deletion criteria, our re-analyses for 1986-89 were 

 based on 111 fewer sea turtle records than were the 

 analyses of Caillouet et al. ( 199 1 ). 



For each year, monthly sea turtle strandings (spe- 

 cies combined, including those identified to species 

 as well as those not identified to species) were 

 summed over subareas within the upper and lower 

 coasts. This produced 96 observations of monthly sea 

 turtle strandings (2 geographic zones x 4 years x 12 

 months). Each observation was standardized to S, 

 the monthly number of strandings per 100 km of ac- 



cessible shoreline bordering the Gulf, which was a 

 measure of sea turtle stranding rate. 



A set of offshore, monthly shrimp fishing effort 

 (days fished) data for subareas 17-21 during 1986- 

 93 was obtained from the NMFS Galveston Labora- 

 tory (Patella 1 ). It was not possible to partition shrimp 

 fishing effort within subarea 17 into portions east 

 and west of longitude 93"07'30 "W; therefore the east- 

 ern boundary of subarea 17 (Fig. 1) marked the east- 

 ern boundary of the upper coast in regard to fishing 

 effort. The total fishing effort at depths seaward of 

 30 fm (54.9 m) represented less than 8% of the total 

 fishing effort in 1986-89 and 1990-93 on the upper 

 and lower coasts (Table 2), but we included depth 

 intervals seaward of 30 fm for comparison with shal- 

 lower intervals. For 1986-89 and 1990-93, monthly 

 fishing effort was summed over subareas within the 

 upper and lower coasts, by depth interval (0-5, 5- 

 10, ... >45 fathoms; see Kutkuhn, 1962; Patella, 1975). 

 This produced 960 observations of monthly fishing 

 effort (2 geographic zones x 10 depth intervals x 4 

 years x 12 months). Each observation was standard- 

 ized to E, the monthly fishing effort per 100 km 2 , 

 which was a measure of fishing intensity. 



' Patella, F. 1994. Galveston Laboratory, Southeast Fish. Sci. 

 Cent., Natl. Mar. Fish. Serv.. NOAA, 4700 Ave. U, Galveston, 

 TX 77551-5997 Personal commun. 



Table 2 



Percentage of shrimp fishing effort (days fished), total shrimp fishing effort, total catch (pounds of shrimp tails), and pounds 

 caught per unit fishing effort on the upper and lower coasts of the northwestern Gulf of Mexico in 1986-89 and 1990-93, by depth 

 interval.' 



' Based on data provided by Frank Patella (see Footnote 3 in the text) 



2 Shrimp statistical subareas 17 and 18 (Fig. 1; see Kutkuhn, 1962) 



3 Shrimp statistical subareas 19-21 'Fig. 1; see Kutkuhn, 1962), 



