326 THE EVOLUTION THEORY 



If the horns of the Peridinete grow to twelve times the usual 

 length in adaptation to life in sea-water with a salinity increased to 

 the extent of -002 per cent., then undoubtedly not only the proto- 

 plasmic particles of the body which form the horns, but all the rest 

 as well, may be capable of adaptation ; and if the Peridinium proto- 

 plasm has this power of adapting itself to the external conditions, 

 then the capacity for adaptation must be a general character of 

 all unicellular organisms, or rather of all living substance. As will 

 be seen later on, we shall be brought to the same conclusion by 

 different lines of evidence. But a recognition of this must greatly 

 restrict the sphere of operation which we can attribute to saltatory 

 mutations in the sense in which the term is used by De Vries, for 

 adaptations from their very nature cannot arise suddenly, but must 

 originate gradually and step by step, from ' variations ' which com- 

 bine with one another in a definite direction under the influence 

 of the indirect, that is, selective influence of the conditions. 



According to the theory of De Vries it seems as if ' variations,' 

 augmented by selection, could never become constant, and that even 

 the degree to which they can be augmented is very limited. As far 

 as this last point is concerned, De Vries seems to me to overlook the 

 fact that every increase in a character must have limits set by the 

 harmony of the parts, which cannot be exceeded unless other parts 

 are being varied at the same time. Artificial selection, in fact, 

 in many cases reaches a limit which it cannot pass, because it has 

 no control over the unknown other parts which ought to be varied, 

 in order that the character desired may be increased still further. 

 Natural selection would in many cases be able to accomplish this, 

 provided that the variation is useful. But of what use is it to the 

 beetroot when its sugar-content is doubled, or to the Anderbeck 

 oats to be highly prized by man ■? And yet many individual 

 characters have been very considerably increased in domesticated 

 animals by selection : of these we need only call to mind the Japanese 

 cock with tail-feathers twelve feet long. 



But undoubtedly these artificial variations do not usually ' breed 

 true ' in the sense that De Vries's mutations of OEnothera lamarckkma 

 did, that is to say, they only transmit their characters in purity with 

 the continual co-operation of artificial selection. This at least §,ppears 

 to be the case, according to De Vries, in the ennobled cereal races, 

 which, if cultivated in quantities, rapidly degenerate. In many 

 animal breeds, however, this is not the case to the same degree ; many, 

 indeed the majority, of the most distinct races of pigeon breed true, 

 and only degenerate when they are crossed with others. 



