THE ASSOCIATION OF CHARACTERS 313 



it is considered as a unit, the tint of the stems and foliage, 

 and that of the corolla and all other parts must no longer be 

 considered as so many separate marks, but as the results of a 

 single intimate character. As soon as this is lost, or reduced 

 to a state of inactivity by the production of a pale variety, 

 it becomes a matter of course that the change at once affects 

 all the colored organs, as we have seen is the case of the thorn- 

 apple, the belladonna, and numerous other instances. The 

 same explanation holds good for the correlation of fissures, 

 as seen in the petals and leaves of celandines and brambles. 

 It is evidently one and the same internal unit which affects 

 both organs. 



If this principle of units is true, it must have an over- 

 whelming significance in the study of hybridism. In the 

 first place, all the modes of expression of one unit must stead- 

 ily keep together, whenever the entire groups of characters 

 are thrown into one another in crossing. This rule must 

 hold good in the more simple cases of crossing varieties of 

 the same species as well as in the hybrids of more widely 

 distant parents. In the first case the rule prevails that the 

 hybrid is not intermediate between its parents, but bears 

 the characters of one of them. Hybrids between blue or 

 red flowered species and their white varieties, between hairy 

 and smooth forms, between spiny and unarmed parents, and 

 many other instances could be presented. They show the 

 marks of the colored, hairy, or spiny parents and are often 

 not at all distinguishable from these. Here the unit is repro- 

 duced without being weakened or rather without its divergent 

 expressions being separated from one another. The second 

 generation of the hybrids completely supports this concep- 

 tion. Some plants remain true to the type of the first gen- 

 eration, but others return to the alternative grand-parent. 

 Among red-flowered hybrids whites occur, among a spiny 

 and hairy progeny some smooth ones are seen. But inter- 



