THE TANGANYIKA PROBLEM. 333 



appearance, engendered by some unspecified conditions, 

 what would become of palaeontological determinations of 

 shells in general ? # 



We could understand, and accept it as a fact, that one 

 form in the whole fauna of a lake had acquired the concho- 

 logical characters of a form belonging to a fauna of a 

 remote age, and yet be independent of it, as is probably the 

 case with the Paramelania of Tanganyika and the Pyrgu- 

 lifera of the upper chalk. But the chance of this sort of 

 thing happening among a large number of forms in two 

 faunas which have no connection is improbable in the 

 extreme. 



However, waiving for the present this whole comparison, 

 and waiving as well the early marine characters of the 

 halolimnic gastropods altogether, the advocates of con- 

 vergence are still not by any means out of the wood. 

 There are other marine organisms in Tanganyika besides 

 gastropods ; there are prawns, and sponges, and jelly-fish, 

 and polyzoa, which are not found elsewhere in the lakes of 

 Central Africa. Where did these things come from, and 

 out of what fresh-water form could they possibly converge ? 



In order to account for the presence of the halolimnic 

 fauna in Tanganyika through convergence, we must, in 

 fact, either resuscitate the old doctrine of special creation or 

 adopt Bastien's theory of heterogenesis, whereby a puppy 

 could breed pigs. To the present writer the view that the 



* In the comparison referred to, and with which I shall have to deal subsequently, I 

 may point out that I took great trouble to consult all the experienced palaeontologists 

 who were available, in order that I might not be accused of making a comparison which 

 experienced palaeontologists themselves would not. And, as I pointed out in the memoir 

 dealing with this matter, the palaeontologists thus consulted were unanimous in affirming 

 that, so far as conchological identifications are ever justifiable, the above comparisons 

 were so. In this matter I used, as I shall explain, the methods adopted and accepted 

 by palaeontologists, and, if these are not to be trusted, it is simply so much the worse 

 for palaeontology. And it does not prove convergence either. 



