Gulf of Mexico (Table 1). This indicates that per 

 unit effort, 16 turtles were captured in the At- 

 lantic for every one turtle captured in the Gulf. 



An attempt was made to compare mean depth 

 and duration of tow for turtle captures with the 

 mean depth and duration of tow for all effort by 

 area with and without turtle captures. The mean 

 depth of fishing and mean length of tow were 

 computed from effort data for each statistical zone 

 and for tows in which loggerhead, Kemp's ridley, 

 or green turtles were captured. In most cases 

 (particularly the Gulf of Mexico) sample sizes 

 were small, and no patterns or consistency were 

 evident. We suggest that despite some apparent 

 statistical differences which we attribute to small 

 sample sizes, average depth and tow duration of 

 turtle captures were probably not different from 

 that of the effort. 



Summary information on observer effort, 

 CPUE, shrimping effort, estimated captures, and 

 estimated mortality in the Gulf of Mexico and 

 southern North Atlantic are presented for logger- 

 head, Kemp's ridley, and green turtles (Table 1). 

 Estimated CPUE for all turtles in the Gulf of 

 Mexico (zones 1-21) was 0.0031 ± 0.0008 turtles/ 

 net hour, and CPUE for the southern North At- 

 lantic (zones 24-33) was 0.0487 ± 0.0041 turtles/ 

 net hour. 



The calculation of estimated mortality used 



minutes fished as a means of estimating the per- 

 cent of the turtles captured that are killed. Based 

 on mean tow times from our effort data, the over- 

 all mortality rate for the Gulf of Mexico is 29%. 

 The eastern Gulf mortality rate is 34%, the cen- 

 tral Gulf rate is 22%, and the western Gulf rate is 

 38%. For the Atlantic coast, the rate is 21% re- 

 flecting the shorter average duration of trawl 

 tows on this coast. 



The mortality rates based on minutes fished do 

 not distinguish among species. This is because of 

 the small numbers of captures for species other 

 than loggerhead turtles. If there are differences 

 in the ability of the other turtle species to survive 

 long periods of immersion and the stress involved 

 in being captured in a trawl, the differences are 

 not measurable from these data. 



In using minutes fished to estimate mortality, 

 the data did not conform to expected models over 

 the range of our observations. In tows of <60-min 

 duration, mortality rates were <1% suggesting 

 that the logistic model might be most appropriate 

 to describe the relationship. However, of logistic, 

 2d and 3d order polynomial and linear models, 

 the best fit over the range of tow times observed 

 in these studies was provided by the linear model. 

 In tows of <60-min duration and in tows longer 

 than 360 minutes, the linear model is probably 

 inappropriate; mortality is negligible in very 



Table 1 . — Observer effort, turtle captures, CPUE, shrimping effort, estimated captures and estimated mortality of loggerhead, 

 Kemp's ridley, and green turtles in the Gulf of Mexico and the southern North Atlantic. 



'Gulf of Mexcio effort estimates provided by NMFS. Galveston Laboratories (E. Kiima text footnote 5) and soutfiern North Atlantic effort based 

 on estimates from Anonymous 1983 

 2Based on CPUE for ttie overall Gulf of Mexico. 



815 



