FUNCTIONALLY-CAUSED MODIFICATIONS. 619 



independent variable, must have had its particular determinant. Now there 

 are about 300 threads on the shaft of a large feather, and each of them 

 bears on the average 1,600 processes, making for the whole feather 480,000 

 of these processes. For one feather alone there must have been 480,000 

 determinants, and for the whole tail many millions. And these, along with the 

 determinants for the detailed parts of all the other feathers, and for the 

 variable components of all organs forming the body at large, must have been 

 contained in the microscopic head of a spermatozoon ! " [And each of them 

 must, throughout all the complex developmental processes, have preserved 

 the ability to find its way to the exact place where it was wanted !] 



If my Cambridge correspondent is able to conceive this process 

 implied by the hypothesis of Weismann, I can only say that he 

 has an enviable power of imagination. 



But now comes the strange fact that an impossibility of thought 

 implied by Weismann's hypothesis does not cause rejection of it, 

 but yet is urged as a reason for rejecting an alternative hypo- 

 thesis which does not imply it. One objector cannot conceive 

 that "a change in the development of a muscle or nerve can 

 effect a corresponding change in that part of the germ which is 

 destined to produce a corresponding part in the descendant " ; 

 and another objector says it is " very hard to believe " that a 

 functionally-changed organ will so affect spermatozoa and ova 

 that " one particular part of them will be so altered that the 

 organisms which grow up from them will be able to present the 

 same modification on the application of a different stimulus." It 

 is tacitly assumed by both that, as in the hypothesis of Weis- 

 mann so in the counter-hypothesis, a particular part of the 

 germ-plasm gives origin to a particular part of the developed 

 organism. But nothing of the kind is implied. The nature of 

 the counter-hypothesis (at any rate as held by me) is entirely 

 misapprehended. Anyone who turns back to the chapters in the 

 first volume where the conception of physiological units (or con- 

 stitutional units) was set forth, or who re-reads the foregoing 

 appendix, will see that there is altogether excluded any idea of 

 correlation between certain parts of the germ and certain parts 

 of the resulting organism. The units are supposed to be all 

 alike, and during the progressive embryological changes local 

 groups of them are supposed to take on different forms and 

 structures under the combined forces, general and local, brought 

 to bear on them. This conception is necessitated by all the 

 evidence. The fact disclosed by the experiments of Driesch, 

 Wilson, and Chabry, that from fractions of an ovum structures 

 may be obtained like that obtained from the whole ovum, only 

 smaller, necessitates it. The fact that any sufficiently large 

 fragment of a polyp or planarian, no matter from what part of 

 the body taken, will develop into a complete polyp or planarian 

 necessitates it. The fact that from an undifferentiated portion 



