LECTURE VIII. 219 



the north; such are the mussels which find their .vital con- 

 ditions in the arctic sea^ in Greenland and Spitzbergen, where 

 they acquire their full development. Among the arctic mussels 

 on the French coast^ a particular type has been established, 

 which, having become permanent as an independent race, no 

 mussels are now produced with the size or the markings of the 

 arctic type. The smaller the region tenanted by a certain 

 species, the more defined is its type ; the wider the region, 

 the greater the number of races and varieties. That which 

 here interests us most is the inference that the abnormal form, 

 and generally any deviation from any given type, by whatever 

 influences it may have been produced, may, by propagation 

 and transmission, give rise to a variety, and that this variety 

 may, by the constancy of its distinguishing characters, become 

 a race, and be propagated as such. 



With regard to propagation, we observe a difference among 

 races, some becoming obliterated in the progress of inter- 

 mixture with other races ; whilst some impart their peculia- 

 rities to their descendants for many generations. Every dog- 

 fancier knows that the blood of a Newfoundland dog is almost 

 indestructible, — that one crossing is sufficient to perpetuate 

 some of the characters through many generations. But it is 

 also known that the Newfoundland dog belongs to a race very 

 probably the product of the country where it is found, and of 

 the natural conditions of that country, and that this race may 

 justly be considered as a well-marked species. Those who 

 include all dogs, from the Dingo of Australia to the Polar dog, 

 in one species, will only look upon the Newfoundland dogs as 

 a race which is distinguished above many others by the con- 

 stancy of its characters. 



It is looked upon as characteristic of races or varieties that 

 they admit of crossing, and that their descendants are in- 

 finitely prolific. We shall, for the present, consider this axiom 

 as established, though we cannot help observing that the 

 proofs are not complete, and that some results in the breeding 

 of domestic races do not accord with it. This much seems 

 established, that in proportion as races become constant, the 

 difficulty in pairing them increases, and that in the free state 



