on the Surface of the Earth. 107 



give birth to descendants capable of reproducing themselves, and we 

 consider them, probably with reason, as having had a common ori- 

 gin ; while the horse and ass, which give birth to steril mules, are 

 regarded as proceeding from two stocks different in their origin. 

 Now this is exactly what it is necessary to demonstrate ; for what 

 assurance have we that unknown circumstances, operating during an 

 indefinite period, might not have sufficiently modified a primitive 

 type, so as to induce the differences which distinguish the ass from 

 the horse, and that these circumstances have so changed their respec- 

 tive natures, that their copulation is no longer attended with the 

 same results as those which we observe between individuals less mo- 

 dified? We may further add, that everything relating to these 

 crossings presents numerous anomalies. We thus find, particularly 

 in warm countries, examples of fertile mules (of the ass and horse) ; 

 and in the race of dogs, cross breeds are spoken of whose fecundity 

 ceases after a certain number of generations. 



The second character, that is, the tendency to return to the pri- 

 mitive type, presents us with numerous difficulties, for we cannot 

 always distinguish the action of the external agents modifying a 

 normal type, from that of the same agents destroying these anterior 

 modifications. Thus, as we have said, Flemish horses taken to Ara- 

 bia, acquire, after some generations, the forms of a horse of pure 

 blood, while Arabian horses, placed among the moist pastures of 

 Flanders, acquire heavy forms. Which of these two modifications is 

 the alteration, and which is the return to the type ? This is a point 

 not easy to decide. It is true that there are many cases where no 

 uncertainty exists ; this, in particular, takes place in the study of the 

 influence of domestication, and in that of reversion to the type by 

 resuming a wild state. 



With regard to the third character, consisting in this, that the 

 action of exterior agents, operating on beings which proceed from a 

 common stock, can affect only certain superficial characters, authors 

 have been obliged, in order to render it available, to assign more pre- 

 cision to it than the strict observation of fact authorises. In some 

 well-known species, the limits of the variations produced by climate, 

 change of life, nourishment, &c., have been studied ; and it has been 

 thence concludcMi, that the extreme of the modifications obtained by 

 these means, was the limit of the differences which can be prodr.ced 

 among individuals of common origin. In this assertion, hypothesis 

 has again exceeded what has been really demonstrated, for it would 

 have been necessary to prove that there never were circumstances 

 more active, and that we know all the effect which a very long du- 

 ration of time could produce. I believe it is true, as I have said 

 elsewhere (Paleontology, i., 86), that we must take care not to ex- 

 aggerate the effect of these unknown causes, and that the study of 

 actual agents is the true basis of our knowledge ; but I cannot re- 



