116 



Current Her pet ol. 20(2) 2001 



buttress was long, expanded, and 

 strongly connected to the costals, and 

 the connected portion was distinctly 

 concave. In Hieremys, Cistoclemmys, 

 Cuora, and Pyxidea, the dorsal portion 

 was very short, and the connected por- 

 tion was not concave. In the other 

 batagurines, the buttress was strongly 

 developed, almost touching the rib. We 

 coded this character in the same way as 

 Character 28. 



30. Internal choanae 



Parsons (1960) examined the structure of 

 the choanae for most batagurids. We fol- 

 lowed his terminology for the structure. 

 Our examination revealed that Chinemys, 

 Hieremys, Ocadio, Orlitia, and Siebenrock- 

 iella of batagurines, and most geoemydines 

 had the internal choanae with a flap or a 

 ridge. Some specimens of Mauremys 

 mutica, Rhinoclemmys pulcherrima, and 

 Cistoclemmys flavomarginata had the 

 internal choanae with a very weak flap, 

 resembling a ridge. In Mauremys rivulata, 

 M. mutica, Melanochelys, and Rhino- 

 clemmys pulcherrima, some specimens had 

 a flap and others had a ridge. The internal 

 choanae had a small papilla in Callagur, 

 Hardella, Kachuga, Morenia, 2/4 of 

 Notochelys, Cyclemys, Rhinoclemmys 

 funerea, and R. punctularia. The internal 

 choanae lacked a papilla, flap, or ridge in 

 Malayemys and Heosemys. We could not 

 examine this character state in Batagur and 

 Geoclemys. We coded this character as: 

 (0) without papilla, flap, or ridge; (1) with 

 flap or ridge; (2) usually with papilla. 



31. Skin of posterior head 



The skin of the posterior head was 

 smooth in most geoemydines, but subdi- 

 vided into small scales in Cyclemys, Geo- 

 emyda silvatica, Heosemys, Melanochelys, 

 Notochelys, and Pyxidea of the geoemy- 

 dines, as well as in the batagurines. We 

 coded this character as: (0) subdivided into 

 small scales; (1) smooth. 



32. Radiated inarkings on plastron 

 Radiated markings were observed on 



each plastral scute in Cyclemys and Heos- 

 emys. We coded this character as: (0) 

 absent; (1) present. 



33. Entoplastron 



In turtle taxonomy, many authors paid 

 attention to whether the entoplastron was 

 intersected by the humero-pectoral seam 

 (e.g., Boulenger, 1889; Smith, 1931; Bour- 

 ret, 1941, Pritchard, 1979; Ernst and 

 Barbour, 1989). Our examination revealed 

 that among batagurines, the state of this 

 character was variable, seemingly reflecting 

 the phylogeny (Hirayama, 1984). On the 

 other hand, the entoplastron was consis- 

 tently intersected by the seam in all the 

 geoemydines but Geoemyda silvatica. 

 Therefore, the absence or presence of the 

 intersection was not informative for the 

 inference of geoemydine phylogeny. In the 

 specimens of G. silvatica examined by us, 

 the entoplastron was intersected by the 

 seam near the posterior rim (4/5) or over- 

 lapped by the seam at the posterior rim 

 (1/5). We thus considered that the absence 

 of the intersection represented rather minor 

 individual variation in this species, 

 although Moll et al. (1986) reported the 

 absence of intersection of entoplastron by 

 the humero-pectoral seam in this species. 



We also examined whether the entoplas- 

 tron was intersected by the gulo-humeral 

 seam. In most batagurid species, the 

 entoplastron was intersected by this 

 seam or barely separated from it (1/10 

 of M. mutica and 2/9 of Sacalia), but in 

 Geoemyda the entoplastron was in a 

 separate location posterior to the gulo- 

 humeral seam. We coded the character as: 

 (0) entoplastron mostly intersected by gulo- 

 humeral seam; (1) entoplastron separated 

 from gulo-humeral seam. 



34. Antero-dorsal portion of iliac blade 

 In most batagurids, the antero-dorsal 



portion of the iliac blade was about as 



