seaturtle.org : MTN : ARCHIVES : Sign In

Morphometric data were collected from nesting loggerheads at Cape Island, Cape Romain National Wildlife Refuge, in 2000 and Pritchards Island, Beaufort County from 2001-2003 (Fig 1). Curved carapace length (CCL) and width (CCW), straight carapace length (SCL) and width (SCW), and body depth (BD) were measured. For both CCL and SCL, measurements came from the nuchal notch to the tip of the longest posterior marginal. Curved measurements were taken with a flexible vinyl measuring tape and straight measurements were taken with tree calipers. Body depth measurements were taken with tree calipers. If the turtle was found nesting, sand was removed under its plastron during the egg deposition phase allowing body depth measurements to be taken during the nesting process. If the turtle was found on its descent from the beach, two parallel metal bars were placed under the front flippers lifting the anterior end of the turtle off the sand allowing for a measurement to be taken. All measurements were rounded to the nearest 0.1cm. Using regression analyses, four conversion formulae were derived. The sample sizes were n = 125 (formulae 1-3 below) and n = 136 (formula 4 below). This difference is due to additional data available for BD.

Figure 1. Sampling locations.

Table 1. Mean, standard deviation, and range of morphometric measurements taken at each study site (cm).*n1 is the sample size of turtles used for the SCL, SCW, CCL, and CCW data. **n2 is the sample size of turtles used for the BD data.

Conversion Formulae
1. The ?rst formula converts SCL to CCL using linear regression. This is compared to the formula derived from stranded loggerheads, primarily juveniles, throughout the Gulf of Mexico and southeastern Atlantic regions, USA (Teas, 1993).
2. The second formula converts CCL to SCL using Bartlett’s Three Group Method for Model II regression (Sokal & Rohlf 1981). This is compared to the formula derived from adult loggerheads at Kennedy Space Center, Florida, USA (Frazer & Ehrhart 1983).
3. The third formula converts SCL to SCW using linear regression. This is compared to the formula derived from data, primarily from juveniles, collected in Galveston, Texas and the southeast USA (Epperly & Teas 1999; Epperly & Teas 2002). Data were transformed using the natural logarithm (Ln) for direct comparisons.
4. The fourth formula converts SCL to BD using linear regression. This is compared to the formula derived from data, primarily from juveniles, collected in Galveston, Texas and the southeast USA (Epperly & Teas 1999; Epperly & Teas 2002). Data were transformed using the natural logarithm (Ln) for direct
comparisons.
Formulae derived from our data and from published literature were
compared in two ways:

Slope Analyses
The null hypothesis that the slopes generated from the Cape/Pritchards Island data and published slopes (formulae 1, 3, and 4) were not significantly different was tested using model utility t-tests (Devore 2000). Similarly, the null hypothesis that the Cape/Pritchards Island slopes and published slopes (formula 2) were not significantly different was tested using a bootstrap method. (R. Norton & A. Strand, pers. comm.).

Calculated vs. Measured Data Analyses
Values of SCL, CCL, and SCW and BD for the Cape/Pritchards Island data were calculated using the Teas (1993), Frazer and Ehrhart (1983), and Epperly and Teas (2002) conversion formulae, respectively. The derived morphometric values were compared to measured values for the Cape/Pritchards Island data. The null
hypothesis that the measured and calculated values were not significantly different was tested using one sample t-tests (Sokal & Rohlf 1981).

All statistical analyses were done using R (R Development Core Team 2003).

Mean, standard deviation, and range of morphometric measurements
for Cape and Pritchards Island turtles are in Table 1.

Table 2. Slope and calculated vs. measured analyses.

Figure 2. a) Linear Regression of SCL on CCL. b) Bartlett’s 3 Group Method Regression of CCL on SCL. c) Linear Regression of LnSCW on Ln SCL. d) Linear regression of LnBD on Ln SCL.

Conversion Formula Analyses
Regression analyses produced four morphometric conversion formulae for the Cape/Pritchards Island data.
1. Linear regression: SCL = 3.61 + 0.892(CCL), p = < 0.001; R2 = 0.90; n=125 (Figure 2a)
2. Bartlett’s Three Group Method for Model II regression: CCL = 6.580 + 1.009(SCL), p = < 0.001; R = 0.949; n = 125 (Figure 2b)
3. Linear regression: Ln(SCW) = 0.878 + 0.747[Ln(SCL)], p = < 0.001; R2 = 0.44; n = 125 (Figure 2c)
4. Linear regression: Ln(BD) = 0.171 + 0.744[Ln(SCL)], p = < 0.001; R2 = 0.14; n = 136 (Figure 2d)

Slope Analyses
The slopes of the conversion formulae for SCL to CCL and for SCL to SCW derived from the Cape/Pritchards Island data were significantly different than slopes from previous studies (Epperly & Teas 2002; Teas 1993; Figures 2a and 2c). The slopes of the conversion formulae for CCL to SCL and for SCL to BD derived from the Cape/Pritchards Island data were not signi.cantly different than previous studies (Epperly & Teas 2002; Frazer & Ehrhart 1983; Figures 2b and 2d).

Calculated vs. Measured Analyses
The Teas (1993) and Epperly and Teas (2002) formulae for CCL to SCL and for SCL to SCW predicted values larger than measured values for the Cape/Pritchards Island turtles (Figures 3a and 3c). Although the differences seen in BD were not signi.cant, the same trend was seen as in the other two measurements (Figure 3d). The Frazer and Ehrhart (1983) conversion formula predicted values smaller than the measured values for the Cape/Pritchards Island turtles (Figure 3b).

The differences between conversion formulae by Teas (1993), Frazer and Ehrhart (1983), and Epperly and Teas (2002) and the Cape/Pritchards Island formulae (Table 2) may be due to variations in turtles sampled. Teas (1993) used morphometric data from strandings, occurring on the Gulf of Mexico and southeastern Atlantic coasts, USA from 1985 to 1991. Approximately 74.3% of the data consisted of juvenile loggerheads (loggerheads under 80 cm SCL were assumed to be juveniles.). Frazer and Ehrhart (1983) used adult loggerheads from the south Florida subpopulation at Kennedy Space Center, Florida, USA. Epperly and Teas (1999 and 2002) used morphometric measurements generally from reared, live captured, or nesting turtles in Galveston, Texas and the southeast
USA. Approximately 73.6% were juveniles in the censored data used to derive the formulae. Differences between Cape/Pritchards Island formulae and the published formulae (Teas 1993; Teas & Epperly 2002) may re.ect non-allometric growth in adult loggerheads. Differences in the formulae from Cape/Pritchards and Frazer and Ehrhart (1983) may be due to turtle conformational differences by geographic regions or subpopulations. Stoneburner (1980) showed that body depth decreases on a north to south gradient in loggerheads in the western Atlantic population. A greater body depth may equal a larger CCL, indicating that Frazer and Ehrhart’s formula, derived from “.atter” Florida turtles, would predict smaller measurements for turtles in South Carolina.

Figure 3. a) Measured vs calculated SCL. b) Measured vs calculated CCL.
c) Measured vs calculated LnSCW. d) Measured vs calculated LnBD.

Some variation in the formulae may be due to differences in data collection methods. The Teas (1993) data was collected by participants in the Sea Turtle Stranding and Salvage Network throughout the USA Gulf of Mexico and southeastern USA Atlantic regions. By comparing standard deviations of SCL for the Teas (1993) data and the Cape and Pritchards Island data, more variation was seen in the Teas (1993) data. This may be due in part to the variability in people taking measurements and the differences in sample size. The Epperly and Teas (1999; 2002) data was collected by a number of researchers throughout Galveston, Texas and the southeast USA. It can by hypothesised that the Epperly and Teas (2002) data may also have differences in variation for the same reasons. However, standard deviations were not available for comparison.

It is important to note that individual measurements used to calculate published conversion formulae were not available. If the data were available, the methodology for comparing the formulae would change. Differences in conversion formulae shown in this study indicate that the same formula may not be applicable to all life stages, geographic regions, and subpopulations. In the absence of other data these formulae may serve as decent approximations for
morphometric measurements. However, when applying formulae for management decisions, such as regulating the size and dimensions of turtle excluder devices (TEDs), it is important to remember differences do exist and to try and account for these differences when implementing regulations.

Acknowledgements: The authors would like to thank Dr. Robert Norton and Dr. Allan Strand for their statistical assistance. We would also like to thank all of the volunteers and employees who helped collect field data on Cape and Pritchards Island, SC, USA. A huge thanks is given to DuBose Griffin for her patience, time, and insightful editorial help.

DEVORE, J. 2000. Probability and Statistics for Engineering and the Sciences. Duxbury Thomson Learning, California. 775pp.

EPPERLY, S.P. & W.G. TEAS. 2002. Turtle Excluder Devices – are the escape openings large enough? Fisheries Bulletin 100:466-474.

EPPERLY, S.P. & W.G. TEAS. 1999. Evaluation of TED opening dimensions relative to size of turtles stranding in the Western North Atlantic. U.S. Dep. Commerce, National Marine Fisheries Service SEFSC Contribution PRD-98/99-08, 31p.

FRAZER, N.B. & L.M. EHRHART. 1983. Relating straight-line to over-the-curve measurements for loggerheads. Marine Turtle Newsletter 24:4-5.

R DEVELOPMENT CORE TEAM (2003). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-00-3, URL http://www.R-project.org.

SOKAL, R.R. & F.J. ROLF. 1981. Biometry. W.H. Freeman and Company, New York. 887pp.

STONEBURNER, D.L. 1980. Body depth: an indicator of morphological variation among nesting groups of adult loggerhead sea turtles (Caretta caretta). Journal of Herpetology 14(2):205-206.

TEAS, W.G. 1993. Species composition and size class distribution of marine turtle strandings on the Gulf of Mexico and southeast United States coasts, 1985-1991. NOAA Technical Memorandum NMFSSEFSC-315, 43pp.