138 PROBLEMS OF EVOLUTION 



already the system of multiplication by fission is in full swing 

 and further evolution is conditioned by this. If there is to be 

 great increase in size it can only be by one method : the resultant 

 cells when fission takes place may cling together, and thus we 

 shall have one of the metazoa or multicellular organisms. This 

 stage attained, further possibilities lie open. There may be 

 division of labour ; certain cells may specialise and undertake 

 particular duties. As we ascend to higher and higher stages 

 we see the same principle, limitation of possibilities by heredity. 

 To show this I take a genealogical tree giving the supposed 

 lines of descent of the various existent classes and sub-classes 

 of animals. Hackel's tree (Evolution of Man, vol. ii. p. 1 88), 

 is generally acceptable enough for our purpose, though, no 

 doubt, every biologist would prefer to draw out his own. We 

 will begin by taking a point more than half way up the tree, 

 where the Selachii, primitive fishes, are found. From these we 

 find branching off the various sub-classes of fishes and the 

 amphibians. Should the former course be followed there are 

 paths leading towards the ganoids, the osseous fishes and the 

 mud-fishes. There may have been other possibilities which 

 never became actualities, other actualities of which we have no 

 record. But when once one of the other alternatives, whatever 

 their number, had been definitely chosen, development on 

 amphibian lines had become an impossibility. It is true there 

 may be retrogression, but when this takes place it is always 

 partial : some features are retained while others are lost or 

 dwindle. This being so, every advance may be looked upon 

 as a point gained from which a further advance may be made. 

 Every advance cuts ofF certain possibilities and opens up others. 

 As the stages traversed recede further and further into the past, 

 return to them becomes increasingly difficult and at last im- 

 possible. Evolution can only in theory reduce elephants to 

 protozoa. Its method of advance may be compared to the 

 system of case-law, in which the interpretation of the law by 

 a judge, at any rate if upheld on appeal, becomes binding on all 

 judges before whom similar cases come up for trial. As we 

 advance upward from the lowest forms of life, we see this 



