640 ZOOLOGY SEC*. 



by undergoing some more or less persistent change. Mutilations, 

 the rapid mechanical removal or destruction of parts, are here, 

 by the terms of the above definition, excluded from the class of 

 variations altogether, since, though the change involved is frequently 

 permanent, it is effected by an influence which is temporary in its 

 character. This, as will be seen, is of importance in connection 

 with the next question we have to deal with the inheritance of 

 acquired characters. 



Can acquired characters or extrinsic variations be transmitted 

 by inheritance ? That they can be is of the essence of Lamarck's 

 doctrine of development, which, in fact, may be described as a 

 theory of development by means of the inheritance of extrinsic 

 variations, or, as it is sometimes called, use-inheritance. But 

 the maintenance of the view that extrinsic variations may be 

 transmitted is not inconsistent with the acceptance of natural 

 selection as a true cause of evolution. Evolution might be sup- 

 posed to be due to the selection and inheritance of both intrinsic 

 and extrinsic variations. From the nature of the case, evidence 

 in favour of the inheritance of extrinsic variations on the one hand, 

 and the occurrence of intrinsic on the other, is extremely difficult 

 to obtain. One or the other must occur, or there would be no 

 evolution. But to prove in any given case that a change is due 

 to the one factor rather than to the other is extremely difficult. 

 When a character not present in the parents appears in the offspring, 

 there is, to begin with, great difficulty in proving that it is really 

 new : characters not present in the parents are known to be 

 frequently inherited from a more or less remote ancestor. But, 

 if we suppose it to be established that the character is a new one 

 (and absolutely new characters must appear, or we should have no 

 evolution), then it would require a very accurate knowledge of all 

 the circumstances to enable us to be certain whether the appearance 

 of the character is not due to the action of some external influence 

 on the parent, either during development or in the adult state, 

 rather than to a change arising within the reproductive cells. 

 Instances are frequently brought forward which have been supposed 

 to afford evidence of the transmission of mutilations from parent 

 to offspring ; but such a transmission must, from the nature of the 

 case, always be extremely difficult to prove, and the majority, at 

 least, of such cases are found, on a careful analysis, to be capable 

 of other interpretations. On the other hand, though well-estab- 

 lished cases of the inheritance of mutilations would greatly support 

 the doctrine that acquired characters are transmissible, the negative 

 results that have attended certain experiments on mutilations 

 are of little value in the direction of proving that extrinsic variations 

 cannot be transmitted, since, as has already been pointed out, such 

 experiments in mutilation cannot be said to reproduce the con- 

 ditions under which an extrinsic variation might be supposed to 



