18 SURVEY OF THE REPrtODUCTIVE PROCESS 



pursuit, which could never be awarded to the laborious but 

 desultory researches of the old alchymists. In the main, 

 chemistry still is, and must long remain, an empirical 

 science : the very fluctuations of its nomenclature show that 

 its so-called theories are no more than hypotheses, still they 

 have the important use of giving system to what would 

 otherwdse be a chaos of unconnected facts ; and there is 

 undoubtedly a certain amount of truth in the relations thus 

 indicated, though in any fully developed system this is as 

 obviously eked out by conjectures. To those who have the 

 acuteness to distinguish between them, the former furnishes 

 important indications for the course of farther investigations; 

 and these, when completed, serve to determine the truth of 

 the conjectural superstructure. 



But, in fact, the natural tendency of the human mind to 

 generalize makes it impossible to prevent the introduction 

 of hypothesis. Few persons can address themselves to the 

 consideration of any extensive series of phenomena without 

 attempting to trace certain relations of this kind among 

 them, and the validity of their conclusions is entirely a 

 question of degree. In proportion as the facts are numerous 

 and accurately known, and as they are carefully and 

 judiciously comj)ared, the results arrived at will be entitled 

 to rank as weU ascertained general law^s ; but far short of 

 this conclusions will be drawn. With the averao^e constitu- 

 tion of minds, it is impossible it should be otherwise ; and 

 there is, therefore, probably more good likely to be done by 

 insisting on the necessity of distinguishing the well-founded 

 from the merely conjectural, than by entirely excluding the 

 latter. 



Though w^e may admit, therefore, as is the opinion of some 

 well qualified to judge, that we have not yet in most 

 branches of biological science a sufficient knowledge of 

 actual facts to admit of our conclusions from them ranking 

 much above the level of hypothesis, we still reasonably ex- 



