On the Inheritance of Acquired Characters. 137 



we arrive at the conclusion that selection consists in the 

 choice of the most highly nourished.^ Nutrition can 

 obviously not effect the full amount of change to which 

 it may lead, in a single generation. Seeds ripen on the 

 mother-plant; and in the ripening seed the young plant 

 passes a very important and, what is still more important, 

 a very sensitive portion of its life-history. At this period 

 it is obviously dependent on the nutritional conditions 

 of the mother. If the mother-plant is not strong — and 

 not grown itself from strong seed — it cannot give rise 

 to the strongest offspring. It takes, therefore, a few 

 generations for the surrounding conditions to exert their 

 full effect. And as variations are caused by nutrition, so 

 it must be possible to increase them by selection of the 

 best nourished individuals within some few generations. - 



It has been my object in this discussion to offer a so- 

 lution of the question of the inheritance of acquired 

 characters b}^ a critical analysis of the theory of selec- 

 tion. A solution on some such lines as those suggested 

 may moreover lead us to the only path that will help us 

 — that of experimental investigation. Once on this path 

 the first thing we have to do is to find out whether varia- 

 tions which are to a great extent dependent on environ- 

 ment can be intensified or diminished by selection in the 

 ordinary way. 



In conclusion, seeing that the material proof is 

 meagre, we may illustrate this by the case of polycephaly 

 in Papaver somniferum.^ 



^ It seems not unreasonable to regard the effects of use as due to 

 the mcreased nourishment of an organ. 



L'Unife dans la variation, pp. 21-22. (Revue de TUniversite de 

 Bruxelles, Tome III, 1898, Avril. 



^ AUmentatinn et selection. Volume jubilaire de la Societe de 

 Biologic. Paris 27 Dec. 1899, p. 17. Ref. in Biolog. Centraiblatt, 

 Bd. XX, No. 6, 1900. 



