354 THE AMERICAN NATURALIST [Vol. XL VII 



verse even in several respects, a relatively simple and 

 unequivocal complex of causes can, as a rule, be discov- 

 ered, so that the problem for investigation becomes 

 clearly limited. But to search for all the causes of any- 

 thing taken by itself is (in biology at least) a hopelessly 

 indefinite and unlimited task. 



3. Search for a single definite and unequivocal cause 

 or determiner of a given result or characteristic has 

 meaning only when there is at least implicitly a compar- 

 ison with something else, for nothing is in itself com- 

 pletely and exclusively determined by any single pre- 

 ceding condition. What cause or determiner will be 

 found depends upon what comparison is made. When 

 the comparison is not specified, it may be made with di- 

 verse things by different minds; thence arise apparent 

 disagreements. The cause or determiner of brownness of 

 skin in man is some peculiarity of the germ cell, when 

 we compare a given brown individual with a white one 

 that has lived under the same conditions ; it is exposure 

 to sun when we compare a brown individual that has 

 lived in the open with his in-door brother ; if some other 

 comparison is made, the cause is still different. It is 

 really the difference between the two cases that we ac- 

 count for, and both members of the comparison must be 

 considered before the cause can be given. 



4. When seeking the cause of a given result, it may be 

 unnecessary to state what we are comparing it with, be- 

 cause that is evident. But much obscurity and disagree- 

 ment would be avoided if that were always made clear; 

 the investigator himself should at least have thought 

 through the comparison carefully. 



5. While it is helpful if in experimentation the two 

 cases compared can both be concretely present, for clear- 

 ness of thought this is not indispensable. One of them 

 may be supplied mentally. 



6. By successively comparing our given case with 

 others, taking first those which differ from it but little., 

 and passing then to cases which differ from it in other re- 



