Vol. I] VAN DEN BURGH—SNAKES OF THE GALAPAGOS ISLANDS 333 



to believe that Charles and Hood islands were connected, and 

 formed a single island for a long time after their separation 

 from the more northern or central islands. 



The snakes of the two Gardner islands agree in every detail 

 with those of the larger islands to which these are adjacent, 

 so that the separation of the one Gardner from Charles, and 

 of the other Gardner from Hood, must have occurred still 

 more recently. 



The second and third subgroups are much more closely 

 related to each other than to the first. This may be considered 

 to indicate that all of the central islands from Narborough 

 to Barrington and from James to Brattle — with the possible 

 exception of Duncan — remained connected for a considerable 

 period after the separation of the northern and the southern 

 islands. 



The distribution of the second and third subgroups, and 

 of D: slevini and D. steindachneri, indicates that there occurred 

 at a still later date the separation of this central land into two 

 large islands; an eastern, including the present James, Jervis, 

 Indefatigable and Barrington Islands ; and a western, of which 

 Narborough, Albemarle and Brattle formed parts. 



The more recent changes are much less clearly indicated by 

 the ophidian fauna, but certain color-differences render it 

 probable that Narborough became separated from Albemarle 

 before breaks in the eastern island occurred, first between 

 Barrington and Indefatigable, then between James and Inde- 

 fatigable, and lastly between James and Jervis. 



The snakes of Albemarle are at present known only from 

 two specimens — one Dromictis slevini from Cowley Moun- 

 tain, and one Dromicus occidentalis helleri from Cape Berkeley. 

 Under such conditions little can be deduced as to the past 

 history of this island without the use of evidence furnished 

 by other groups of its inhabitants. This evidence I do not 

 now wish to use; for I believe more accurate results can be 

 attained by attempting to read the story of each group sep- 

 arately, and then comparing results. The mixing of evidence 

 here, it seems to me, would be only less confusing than the 

 jumbling together of data derived from distribution, geology, 

 paleontology, and ocean-soundings. Each should be worked 



