LINNE.VN SOCllCTV OF I.ON'JiON. 75 



writes that " iV very frequen.t feature of diver<;eiit a(la])tat ion is 

 the loss of parts .... Tht^se h)st parts are never reacquired .... 

 Nature often resorts to other remedies to repair her looses, namely 

 to substitution of parts, or to chanye of function.'' 



The most reiiiai-k:ib!e instances of iireversibility of evolution 

 cited by palaeontologists are tliose associated with alternations of 

 habit: it must be borne in mind in considering such cases that 

 they necessarily involve a large element of hypothesis. DoUo* 

 himself deals particularly with tlie Marsupials, which he regards 

 as originally arboieal. In the Tree Kangaroo, Dendrola{/r(S, there 

 seems reason to believe that reversion to the ancestral arboreal 

 habit has occurred in the course of pbylogeiiy, after an inti'rveninr/ 

 terrestrial phase ; according to Dollo we ha\e here " nn Kangarou 

 qui est retourne vivre dans les arbres.'" This reversion of habit 

 has been associated with a failure to revert to the ancestral ty])(' 

 of ai-boreal foot-structure with its characteristic opposable thumb 

 or great toe, but, in connexion with the secondary arboreal phase, 

 the claws have become very large and curved, thus serving a 

 similar purpose. 



DoUot has worked out another example from among the Turtles, 

 which is interesting from the botanical standpoint, since it presents 

 an analogy to something that may have occasionally occurred 

 among ])lants. But the whole subject of the phvlogeny of these 

 animals is so obscure that I quote this inslance, on which parti- 

 cular stress has been laid by DeperetJ, merely on account of its 

 botanical suggestiveness, while recognising that it may pi'ove to be 

 itself invalid. In the case of the Chelonida; which live in the sea, 

 the bony case is reduced and its weight lessened by the development 

 of empty spaces or fontanelles. In the Atheca3, to which Spharr/is 

 (Dcrmocheh/s), the Leathery Turtle, belongs, the original f-hield 

 has been entirely lost, but this loss has been compensated by the 

 development of a distinct secondary shield, composed of polygonal 

 dermic plates, superposed upon the rudiment of the first, but not 

 welded to it. According to Dollo, tiie development of this 

 secondary shield is to be associated with a return to littoral life 

 on the part of a hypothetical ancestor resembling Psei>hopliorvs. 

 However this may be, it seems at least probable from the worl< of 

 Case§ that tiie Leathery Turtle traces its descent through Protosfer/a 

 and Protospharrjis — forms in which the carapace has disappeared, 

 but the secondary shield is not yet de\ eloped — back to an ancestor 

 with the normal Chelonian carapace. In otlu^r words, the normal 

 i'arapace once gone could not be reconstituted, but was re[)laced 

 by a new organ of different origin. 



There is no dilHculty in imagining that similar alternations of 

 habitat may have occurred among i)lants in the course of their 

 phyletic history; it is indeed more than probable that future 



* Dollo, L. (1890); see also O.^boni, H. F.(1910) p. 23. 

 ■\ Dollo, L. (11)01). 



i Deperet, C. (1907) p. '221). 



§ Case, E. C. (1897). 



