196 Report S.A.A. Advancement of Science. 



of the forms with the above characters must remain open. Such a 

 position shows how relatively uncertain is the standing of a specific 

 type. Taken by themselves, the members of geometrica would be 

 accorded specific distinction by every systematist, but could we secure 

 all the intermediate forms leading from the less specialized forms 

 up to these extreme stages — stages which must certainly at one 

 time have existed — they would be reduced to the rank of variety. It 

 is hoped that a larger number of individuals from intermediate 

 districts may yet be forthcoming. 



h. Testudo oculifera, Kuhl, 1820. 



This species possesses the most distinctive features of all the 

 members of the geomeirica-gron^. Taken together, the assemblage 

 of characters calls for specific recognition perhaps more strongly 

 than any other combination ; indeed, its specific rank has hitherto 

 never been questioned. Only when the characters are taken individu- 

 ally, and traced through the various groups, can its close relationship 

 with the others be established, and afford hope that it may be possible 

 to secure a gradational series of forms presenting the same 

 combination. 



Compared with the remaining sub-groups, there are many 

 extreme stages in the characters of oculifera. This is particularly the 

 case as regards the elaborate coloration of the carapace and plastron, 

 the single axillary shield, the large nuchal shield, and the serrations 

 of the marginal plates ; moreover, these features are all fairly con- 

 stant and fixed. 



Oculifera and geometrica are the only two forms in which there 

 is the combination of a single axilliary shield and a practical absence 

 of any margino-costal angulation. As already described, both of 

 these conditions represent extreme variations, and might easily 

 be interpreted as proving a close relationship between the two species. 

 On account of the first character, Siebenrock, in his synopsis, places 

 them together in a separate section, but Boulenger arranges the two 

 at the extremes of the entire group. The other characters of oculifera 

 and geometrica indicate that the two species are widely divergent, 

 as such relationships go within the geometric a-gxoM'^. Thus the 

 condition of the antebrachial shields in the two sub-groups repre- 

 sents modifications along very different lines ; the femoral tubercle 

 is strongly developed in oculifera, but absent in many geometrica ; 

 the margin is strongly serrated and upturned in oculifera, sniooth 

 and dependent in geometrica ; and the coloration is very different in 

 the two. 



Oculifera and geometrica afford remarkably good examples of 

 variation along similar lines as regards certain characters, and along 

 divergent lines as regards others. Such a condition probably points 

 to a common origin far back before the characters in question were 

 elaborated ; from this the two have been transformed alike in certain 

 directions, while they have diverged along others. On the theory 

 of germinal selection we should assume that the determinants of 



