BUTLER ET AL.: ESTIMATING LOGGERHEAD TURTLES BY TRAWL SURVEY 



N = CIP 



where P = probability of capture 



C = number of animals captured. 



If more than one sample_tow was made in a 

 substation, the mean catch (C ) was substituted in 

 the above formula. To estimate the number of 

 loggerhead turtles in a channel substation, sta- 

 tion, or the entire channel, thejnean number cap- 

 tured per substation sample (C) times the nuni- 

 ber of substations (s) was substituted: N = sCIP. 

 The estimated variance of this estimate is (Mood 

 et al. 1974): 



Table 1 . — Estimated probability of loggerhead turtle capture in a 

 Port Canaveral ship channel substation using an 18 m fish trawl. 



^ Data set discarded. 



Var (iV) = (sIP)'^ [Var (C) + (CIP)'^ Var (P)] . 



RESULTS 



Estimates of the probability of capture and as- 

 sociated standard error estimates from nine 

 repetitive trawl experiments are presented in 

 Table 1. Estimated probability of capture within 

 a substation based on six experiments ranged 

 from 0.21 to 0.31 (P = 0.28; 95% confidence inter- 

 val = ±0.05; estimated variance = 5.18 x 10""^). 

 Three experiments were excluded from the analy- 

 ses: two were discarded because the catch failed to 

 decline due to low population levels, and a third 

 was eliminated because of problems with the 

 sampling trawl. 



Estimates of loggerhead turtle abundance by 

 survey for the Port Canaveral ship channel 

 ranged from 701 ± 291 turtles in late February 

 1982 to a low value of 38 ± 26 turtles in late Au- 

 gust 1982 (Table 2). Port Canaveral channel sta- 

 tions 9 through 11 (Fig. 2) exhibited the highest 

 loggerhead turtle abundance during all seasons of 

 the year. Mean catch for all samples in the chan- 

 nel was 2.55 turtles/tow and 0.50 turtles/tow for 

 control samples, supporting the hypothesis that 

 loggerhead turtles congregate in the Port Canav- 

 eral ship channel. 



Loggerhead turtle abundance estimates for the 

 remaining four survey sites were low during all 

 seasons of the year (Table 3). Over the study pe- 

 riod, a total of 18 loggerhead turtles was cap- 

 tured: 2 at St. Mary's entrance, 6 at Ponce de Leon 

 Inlet, 3 at Fort Pierce Inlet, and 7 at St. Lucie 

 Inlet. 



DISCUSSION 



Our estimates of the probability of capture 



Table 2. — Estimated number of loggerhead turtles (A/) at Port 

 Canaveral ship channel by station and survey period (1981-1982). 



■I Station not sampled 



2statlon incompletely sampled. 



3|ncludes 4 Kemp's ridley turtles, Lepidochelys kempi. 



4Estimate is for stations 9-14, others are for 7-14, 



Table 3. — Estimated loggerhead turtle abundance dunng quarterly 

 surveys of St. Mary's entrance — King's Bay, Ponce de Leon Inlet, 

 Ft. Pierce Inlet and St. Lucie Inlet. 



Date 



St. Marys 

 King's Bay 



Ponce de 

 Leon Inlet 



Fort Pierce 

 Inlet 



St. Lucie 

 Inlet 



11/81 

 2/82 

 5/82 

 8/82 



9± 18 

 

 

 







11 ± 15 













 4±7 

 4±8 











4±7 



11 ± 11 



4± 7 



were based on the supposition that catch-per-tow 

 in a given substation will decrease as loggerhead 

 turtles are removed. The regression of cumulative 

 loggerhead turtle catch on catch per sample can 

 then be used to estimate the orginal population 

 size in the substation (Brownlee 1965) and using 

 this estimate, the probability of capture can be 

 computed. Assumptions associated with this pro- 

 cedure are a closed population, the trawl fishes 

 only within the defined bounds of the substation, 

 each tow is an equal unit of effort and the proba- 

 bility of capture remains constant. 



451 



