I 



TWO OBJECTIONS CONSIDERED 



during which newly introduced sets of circum- 

 stances have tended to enhance and accumulate 

 variations on either side of a hitherto established 

 mean. Such a conclusion is implied by the 

 theory of natural selection, according to which 

 specific variation occurs, not in conformity to 

 some mysterious law of progress uniformly 

 operating, but only in conformity to some more 

 or less conspicuous alteration in the sum total 

 of the conditions of existence. 



It follows, therefore, that in general, when 

 incipient varieties are differentiated into well- 

 marked species, the number of intermediate 

 forms must be immeasurably smaller than the 

 numbers of forms contained in the resulting 

 species to which they serve as the transition. 

 During epochs of rapid divergence, the means 

 may all be extinguished after a few hundred 

 generations, while the generations of the ex- 

 tremes which persist thereafter may be num- 

 bered by tens of thousands. Suppose, for ex- 

 ample, two great islands separated by a shallow 

 sea. During long ages, while the floor of this 

 intervening sea is constantly rising, the specific 

 changes occurring on either island may be quite 

 few and unimportant, and such fossil records 

 as are left will indicate a general persistence 

 of type. But when in course of time the pro- 

 cess of elevation has converted this interven- 

 ing channel into an isthmus connecting the two 



S3 



