ON VARIETIES, RACES, SUB-SPECIES, AND SPECIES. 99 



preserve on a stock belonging to tlie same tj-pe. But if new 

 qualities are desired, we should proceed in a different way, and 

 recollect an observation made by Cabanis and fully appreciated 

 by Sageret : we mean that the seeds of a pear grafted on a wild 

 quince have a greater tendency to -produce varieties thati the seeds 

 of the same pear not grafted. It ought also to be remembered 

 that tlie influence of a graft is more especially visible in its seeds 

 and plants derived from tiiem. 



The influence of tlie graft on the stock whilst admitted by 

 some is denied by others, and De Candolle has given a clever, 

 criticism on observations cited in favour of it ; our own opinion is, 

 that some such influence is possible, but that it has not yet been 

 shown to exist. 



It has been stated that propagation by layers and cuttings 

 tends to diminish, and even to destroy, the power which the plant 

 operated upon has of multiplying itself by seed, because it is 

 said that such propagation often favours the development of 

 certain parts, especially of the succulent parts of many fruits, as 

 the Pine-apple, Banana, &c., at the expense of the seed ; this 

 opinion has been combated by BF. Duchesne ; and M. Sageret, 

 far from admitting it in principle, acknowledges that many 

 plants raised from layers or cuttings fructify earlier than those 

 which are not so raised, or than those which proceed from the 

 seeds of the parent of the cuttings. 



Akticle 2. 



Of the Degeneration and Extinction of Plants obtained by the 

 Division of Individuals. 



Towards the end of the last century Knight published an 

 opinion, which had been already formed by sevei-al other people, 

 and more especially by Marshall, to the effect that plants ob- 

 tained from grafts, cuttings, and tubers could not be long-lived ; 

 so that if these grafts, cuttings, and tubers represent varieties, 

 these varieties would tend to disappear. Knight even went so 

 far as to say that the life of the individuals thus obtained could 

 not be longer than that of the parent of the graft, cutting, or 

 tuber respectively. M. Puvis, admitting that all the varieties of 

 known plants will die out, and that, consequently, recourse must 

 be had to seeds to obtain new ones, nevertheless denies Knight's 

 exaggerated statement, and remarks that the variety of the 

 Chaumontel besi, although its type, has been apparently extinct 

 for many years ; and that the St. Germain is still in vigour, 

 although the tree which produced it has long ceased to exist in 

 the forest from which it derives its name. 



According to M. Puvis, the death of a variety of fruit tree is 



