THE UISTRII!UTR)N ANI/ CHARACTERS OF REI'TFLES. 239 



of animals, but for the present I use it only in a geographical or 

 topographical sense. 



,H. Seebohm has stated, " So far as is known, no species ot 

 birds has ever been differentiated without the aid of geographical 

 isolation." Dr. Jordan, dealing with birds and fishes in North 

 America, attached the greatest importance to barriers which 

 effect isolation : in regions broken by barriers which isolate 

 groups of individuals, we find a great number of related species. 

 The Rev. Mr. Gulick makes a good case for isolation from a 

 study of the land shells which live under apparently unifonn 

 conditions in the forests of the Hawaiian islands, where an 

 original variable stock seems' to have separated into a number 

 of distinct species unaided by any process of selection, and 

 simply as a result of the geographical separation of the colonies 

 which gave rise to the various species. For a similar investiga- 

 tion in South Africa the land shells of the genus Achatina 

 would probably prove eminently suitable as the genus abounds 

 in local forms : in this genus no case of divergence which is 

 clearly a result of unassisted isolation has yet been brought to 

 my notice, but a more thorough investigation than I have been 

 able to make is much to be desired. In the case of most 



reptiles in South Africa we are not likely to find anything 

 strictly parallel to the phenomena recorded by Gulick, seeing 

 that their active life makes geographical isolation almost im- 

 possible within uniform environments, which are unbroken by 

 large natural barriers. On the other hand, this subcontinent 

 presents such a number of dift'erent climates and environments 

 even in limited areas that in attempting to account for the 

 differences between two forms, which are geographically 

 separated, we can hardly ever be certain that the factor of 

 environment has been eliminated. Amongst the lizards, per- 

 haps the most sluggish or most homeloving are the species of 

 Zonurus. There are ten species which are distributed in dis- 

 tinct areas, those areas in most cases coinciding with regions of 

 very distinct climatic or physical conditions. These specific 

 areas are contiguous, but for the most part do not overlap 

 except to a slight extent in the case of the group of related 

 species, including Zonurus cordylus. The species of the 

 cordylus group are distri1:)Uted as follows: ^irdyhis occupies the 

 coastal belt of Cai^e Colony : a very closr allied form, vittifer 

 lives in Natal and parts of the Transvai,:, whilst another near 

 ally occupies the central plateau from the Transvaal to Angola, 

 its distribution somewhat overlapping that of vittifer in the 

 Transvaal : two other species, which are not so closely related 

 to cordylus as vittifer or joncsii seem to occu])y small areas 

 within the larger area of cordylus. Z. capcnsis being known 

 only from the Hottentot Holland mountain, and Z. coeruleo- 

 punctatus occurring on a hill between Knysna and Avontuur. 

 If we assume that these forms all originated from the same 

 stock we cannot satisfactorily explain the occurrence of the two 



