117 



long as its jiostcro-clorsal poiimii (I ii'. lli, 

 c-j). Only ill ('isloclciiiniys, liic anlcio- 

 doisal porlion was ahoul Iwo limes longer 

 llian tlic poslcio-clorsal poilioii (I ig. Ic, cl). 

 Wc coded this character as: (()) as long as 

 its postero-dorsal portion; (I) two times 

 longer than its postero-dorsal portion. 



35. S/iii/ll Jo/uiiic/i oil Yciiirnl iintxilUi 



Rliinoclcnunys Jinicrca aiul A'. />iinciii- 

 /(//■/(/ usually had a \'er\ small loramen im 

 ilic \ ciil romedial poi lion ol ihe maxilla. 

 Such a loiamcn was abscni m ihc icmainiiig 

 batagui ids. \\ c cmlcd ihis chai acier as: (0) 

 absent; (!) usuall> present. 



Results 



All chaiacters examined were iiiloi niai i\ e 

 I'm I he mierence of relationships anumg 

 ihe 26 geoemydine OTUs. 



1 hree ec|nall\ [larsimoiiious trees (tree 

 length 107; consistency inde.x (CI)=:0.449; 

 retention index=().762; rescaled consistency 

 index=0.342) were delected. Maitremys 

 japonica, ihe \/. Icprosa species-group, 

 and the M. Linihiniciisis-M . imilica clade 

 were trifurcaled in all trees. These trees 

 di tiered IKim each oilier in topology 

 iinoUing Khiiiuclcniniys unniilcila, R. 

 nihlda, and the Geoemyda-Cistoclemmys- 

 Pyxiclca clade. In the strict consensus tree 

 (I'ig. 3), therefore, their relationships were 

 also expressed as trifurcated. 



The consistency inde.x lor these data was 

 Kn\ (CI =0.449). However, this inde.x is 

 known U) he highly negatively correlated 

 with numbers ot OTUs included in an anal- 

 ysis (Sanderson and Donoghue, 19.S9; Klas- 

 sen el al., 1991). Our analysis in\oI\ed a 

 relatively large number of OTUs (28), and 

 the value, while being similar to that in a 



Fill. I. Mcdi.il Slews ol pcbu girdles. H.il.inuim.ic: (a) i hini/nw rfwfsii. ibi 

 crassicollis; Cicociuydinac: (c) Cistoi U'inmys fhivonwr^maia /liivonutrginala. (U) Ci. - ) 



Pyxidvii mouholii. (1) Maiiremyx Japonu-a, (g) M. multca mutica, (h) Cuora irtJiiMiHti. {%) 

 Rhinodvmmyx piiU fwrnina monni, (j) Gt'wmytia Ja/wnica. 



