SPONTANEOUS VARIATIONS 51 



do not occur in a haphazard manner all round the specific 

 mean, but trend in a perfectly definite direction. If we know 

 the parents we are able to predict with tolerable accuracy the 

 kind of variations to which the union of their dissimilar germ- 

 plasms will give rise. It follows, contrary to the doctrine so 

 widely held by Weismann and his followers, that bi-parental 

 reproduction cannot provide materials for the work of Natural 

 Selection. Nevertheless we shall see that it plays a great 

 part in adapting complex species to their environments. 



