248 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4th Ser. 



(21.6) — therefore, considering its size,^ not from Hood, Ab- 

 ingdon, Bank's Bay, Narborough, Tagus Cove, James, Chat- 

 ham, or Duncan ; vertical diameter nine inches ; lateral diam- 

 eter fourteen inches ; marginals reflected upward anteriorly and 

 over limbs; height to marginals apparently great — therefore 

 not from Chatham, Tagus Cove, Bank's Bay, nor Narborough ; 

 front of carapace elevated — therefore not from Indefatigable 

 Island nor Cowley Mountain, Albemarle. 



In the foregoing list we have excluded most of the races 

 of the Galapagos Archipelago. Of all the localities where tor- 

 toises ever had been found in the archipelago up to the date 

 of the visit of the present expedition, there remain to be con- 

 sidered only two — Charles Island and southern Albemarle.'^ 

 I doubt if it be possible to decide with certainty from either 

 the description or the plate whether Harlan's type came from 

 Charles Island rather than from Albemarle ; but there is other 

 evidence which throws some light upon the question. First, 

 there is the circumstance that most of the early voyagers se- 

 cured their tortoise from Charles, James, and Hood islands. 

 Second, Porter stated in his journal that the tortoises of Hood 

 Island were similar in appearance to those of Charles Island, 

 the form of the shell being elongate and turned up forward 

 in the manner of a Spanish saddle. Third, Harlan's plate 

 strikingly resembles my specimens from Hood Island, although 

 his measurements show that he had a different and much 

 broader species. Fourth, the few specimens in collections which 

 can be pretty definitely traced to Charles Island agree with 

 Harlan's specimen in having the length over the curve less 

 than the breadth over the curve. 



I hold, therefore, that Harlan's specimen came from Charles 

 Island, although we cannot positively prove this to have been 

 the case. This being true, the name Testudo elephantopiis 

 cannot be used for the distinct race to which Giinther applied 

 it.^ Some might think it best not to use the term at all, sub- 



6The young are narrower than the adults, so that this statement is true, although 

 in some of the races here enumerated adult individuals may have the curved width 

 greater than the curved length. 



'i'The shape of the carapace of the tortoise found on Jervis Island is quite dif- 

 ferent from Harlan's plate of T. elephantopus, and although the carapace of the Bar- 

 rington Island tortoise is unknown, I think that these islands may safely be ignored 

 as possible places of origin of Harlan's specimen, for the reason that tortoises had 

 never been found upon them by any of the earlier explorers. 



8See Testudo giintheri infra. 



