20 Mutual Independence of Hereditary Characters 



yet essentially independent entities, from the union of 

 which the specific characters originate. Now let us see 

 whether or not this conclusion is supported by experi- 

 ment. 



For this purpose let us turn to experiments on the 

 formation of varieties, especially to those which have been 

 made on a large scale by plant breeders. They teach us 

 that almost every character may vary independently from 

 the others. Numerous varieties differ from their ancestral 

 form, in only one attribute, as, for example, the white 

 sports of red-flowered species. The red color changes in 

 the corolla through all gradations, into white ; it may be 

 lacking or it may be present not only in the blossoms, but 

 also in the stems and leaves, and can be developed to every 

 conceivable degree, without any other hereditary quality 

 being necessarily involved in the variation. In the same 

 way the hairiness, the arming with thorns and spines, the 

 green color of the leaves, may each vary by itself, and 

 even disappear completely while all other hereditary char- 

 acters remain quite unchanged. Frequently some charac- 

 ters that belong together vary in groups without exercising 

 any influence on the other groups. Thus an increase in 

 the number of petals is not rarely accompanied by a petal- 

 like development of the calyx or the bract-leaves, while 

 otherwise the plant remains normal. I have cultivated 

 a Dipsacus sylvestris, which offers all conceivable diver- 

 sities in the arrangement of the leaves, and which is other- 

 wise constant in thousands of specimens. The Papaver 

 somnifenim polycephalum deviates only in the transfor- 

 mation of numerous stamens into carpels. It is the same 

 for the cultivated Semperviruni tectorum. Such instances 

 are so numerous, in the plant kingdom as well as in the 

 animal kingdom, that the independent varying of single 



