Epperly and Teas; Escape openings in turtle excluder devices 



467 



Hi-i^ht: 1(1 in (25.40 cm) CJiill Ot Mexico 

 12 in (30.48 cm) Allunlic 



Width: 32 in (SI. 28 cm) Gulf of Mexico 

 3.'i in (88.90 cm) Atlantic 



Figure 1 



Minimum dimensions (height and width) specified m current U.S. federal regulations 

 (Federal Register, 1992c) for e.scape openings in single-grid hard turtle excluder devices. 

 The copyrighted figure of the turtle was kindly provided by Jamie Serino of Flying Turtle 

 Productions. 



TEDs are equipped with a trap door that allows sea 

 turtles to escape from trawl nets (Seidel and McVea, 1982) 

 and may be either rigid or soft in design ( Federal Register, 

 1992c). To be certified by NMFS, a TED design must be 

 97% effective, in comparison with a control net, in exclud- 

 ing sea turtles (Federal Register 1987, 1992c). Since 1990, 

 turtles used for TED trials have been small loggerhead 

 turtles, mostly 2->T-olds that have averaged 34.4 cm SCL 

 (SD=4.1, 7! = 1730, NMFS-^). Regardless of design, certain 

 parameters of the TED architecture are regulated. Most 

 important to this discussion are the requirements of the 

 height and width dimensions of the opening in the net 

 through which turtles escape. Along the Atlantic Coast 

 these requirements are width >35 in (88.90 cm ) and height 

 >12 in (30.48 cm) (Federal Register, 1992c). In the Gulf of 

 Mexico these measurements are >32 inch (81.28 cm) and 

 >10 inch (25.40 cm), respectively. Height is measured si- 

 multaneously with width and is measured at the midpoint 

 of the straight-line distance of width (i.e. the width and 

 height of a taut triangle is measured. Fig. 1). 



The purpose of our study was to compare the sizes of 

 stranded sea turtles with the size of the TED openings. 

 This evaluation was prompted by the need, identified 

 by the NMFS Turtle Expert Working Group (TEWG), 

 to reduce the strandings of mature loggerhead sea 

 turtles [Caretta caretta) from the northern subpopula- 

 tion (TEWG, 1998). We compared the sizes of stranded 

 loggerhead, green (Chelonia mydas), and Kemp's ridley 

 iLepidochelys kempii) sea turtles, the three species most 

 commonly found stranded, to the minimum widths and 

 heights of TED openings. 



3 NMFS (National Marine Fisheries Service). 2001. Unpubl. 

 data. Galveston Laboratory, National Marine Fisheries Ser- 

 vice, 4700 Ave. U, Galveston, TX 77551. 



Materials and methods 



To compare the sizes of stranded turtles with the min- 

 imum size of TED openings we first constructed a pre- 

 dictor of carapace width and body depth. Thus, a morpho- 

 metric analysis for each species was conducted, generally 

 with data from reared, live captured, or nesting turtles 

 and not with data from strandings. The predictive regres- 

 sion for carapace width was then applied to the strand- 

 ings data when this measurement was not recorded for a 

 given turtle; the predictive regression for body depth was 

 applied to all turtles in the database because body depth 

 was rarely measured at stranding sites. We then analyzed 

 the entire strandings database to compare turtle sizes 

 with the minimum size of TED openings. 



Morphometric analyses 



The species-specific relationship between both body depth 

 and carapace width with carapace length was explored 

 through regression analysis and predictive regression 

 equations were developed. Regressions of untransformed 

 data were compared with regressions of log^-transformed 

 data by comparing goodness-of-fit values. 



Morphometric data (straight line carapace length, notch- 

 to-tip ISCL], straight line carapace width [SOW], and body 

 depth [BD] ) were recorded by a number of researchers 

 throughout the southeast United States and at the Cay- 

 man Turtle Farm, Cayman Islands. Data for loggerhead 

 turtles were concentrated in the 20-30 cm and 30-40 cm 

 SCL size classes and were censored (randomized selection 

 of ^=37 in each of the two size classes) to create a more uni- 

 form distribution for the analysis (Table 1). Green turtle 

 data were more uniformly distributed across size classes 

 and were not censored (Table 1). Data for Kemp's ridley 

 turtles were concentrated in the 1-10 cm and 10-20 cm 



