34 Michael E. Seidel 



range of P. f floridana (Mobile, Alabama) and a resemblance to P. 

 floridana was noted prior to discriminant analysis. This skull was re- 

 identified as P. f floridana for subsequent comparison. The misclassifi- 

 cation of a P. f hoyi skull as P. c. hieroglyphica is not surprising con- 

 sidering the proximity of means (H,Y) for these forms (Fig. 4). This 

 skull was reassigned to P. c. hieroglyphica in further analyses. Based on 

 the 28 cranial characters measured, analysis of variance indicated signif- 

 icant differences comparing New River Pseudemys to all taxa (P < 0.01) 

 except P. c. concinna (P > 0.05), P. c. suwanniensis ( > 0.01), and P. c. 

 mobilensis (P>0.01). 



To further analyze the relationships of New River Pseudemys to P. 

 floridana and P. concinna, discriminant analysis of skulls was applied 

 again, with the rubriventris series omitted. Results of this analysis are 

 similar to the first, but three clusters now become apparent (Fig. 5): a 

 western and Mississippi Valley group (P. c. texana, P. f hoyi, P. c. 

 hieroglyphica; TYH), an eastern P. concinna group (P. c. suwanniensis, 

 P. c. mobilensis, P. c. concinna, and New River Pseudemys; SMCB), 

 and an eastern P. floridana group (P. f floridana, P. f peninsularis; 

 FP). Noteworthy is the broad separation (on the first axis, Ki) of P. f 

 hoyi from other P. floridana skulls and its extensive overlap with P. 

 concinna (Fig. 5). The first two canonical axes account for 49% and 

 20% of the total dispersion, respectively. In order of increasing impor- 

 tance, maxillary alveolar width, anterior skull width, interorbital width 

 and dentary alveolar width are the most influential characters providing 

 separation on the first axis (Table lb). Temporal arch width, jugal- 

 quadratojugal length and lingual alveolar width contributed most to 

 separation on the second axis (Table lb). All individuals were classified 

 into their assigned groups. Results from analysis of variance show New 

 River skulls significantly different from all taxa (P < 0.05) except P. c. 

 concinna and P. c. suwanniensis (canonical means enclosed by broken 

 lines in Fig. 5). 



A third discriminant analysis of skulls, with P. f floridana and P.f 

 peninsularis removed, was applied to further clarify relationships with 

 the races of P. concinna (Fig. 6). The first two canonical axes collec- 

 tively account for 77% of the total dispersion. Characters most respon- 

 sible for separation on the first two axes are presented in Table lc. 

 Results from this comparison indicate a very close phenetic relationship 

 between the eastern river cooter, P. c. concinna (C) and New River 

 Pseudemys (B) (Fig. 6). 



Observations on reproduction of cooters recently collected from 

 Bluestone Reservoir, discovery of a previously overlooked museum 

 record, and examination of archeological material have also contributed 

 to a better understanding of Pseudemys in the New River. Three adult 

 females from Bluestone Reservoir, collected 15 May 1980 and X-rayed 7 

 June (following the method of Gibbons and Greene 1979) showed no 



