Sorting and blending of unit characters. 233 



respondino; particle of the "germ plasm". The implication is, rather, 

 that portions of the hereditary substance bear materials which tend to 

 repeat processes in the cells of the progeny which have led, in the 

 preceding generations at similar stages of development of the indi- 

 vidual and under like epigenetic conditions, to the expression of those 

 forms, etc., which we call "characters". 



It is well known that language handicaps us in the use of terms 

 which may be objectionable from some points of view, or which are 

 inadequate to express our conceptions of life processes. But if we are 

 to give expression to our ideas "words" must be used. With this 

 apology I proceed to outline briefly a hypothesis which may offer an 

 explanation of the production of the "twin" segregate hybrids, and the 

 blend "triplet" in the Fi generation of reciprocal crosses between Oe. 

 pycnocarpa and Oe. nutans. 



We have seen that the parent species differ by a large number 

 of homologous but contrasting characters. When the bearers or factors 

 for these qualities meet in the zygote, by the entrance of the sperm 

 of one species into the egg of another, we may conceive that the 

 situation is very different from that w='here the sperm of one species 

 enters the egg of the same species. In the latter case homologous but 

 like factors, or tendencies, meet in the zygote. Because of this and 

 also because of the long history of similar processes in the past, they 

 form a reciprocal working organization with comparative ease. But the 

 situation is very different in interspecific crosses where the homologous 

 characters are so strikingly different. The homologous factors or ten- 

 dencies of each pair are more or less foreign to each other. The or- 

 ganization of a reciprocal action is more difficult. This is shown by 

 the great difficulty of securing progeny in many interspecific crosses, 

 and the failures in many more. The different positions which the 

 bearers of the factors, or tendencies, occupy as they approach the 

 nuclear plate of the first division in the zygote may well be due in 

 a large measure to chance. In this way different combinations might 

 establish a working relation in different eggs. 



If we now apply this hypothesis to the Fi hybrids from the reci- 

 procal crosses of Oe. pycnocarpa and Oe. nutans, an explanation of the 

 origin of the different hybrid t^q^es may be formulated as follows. In 

 the blend, lujbrida nutella, the factors representing pairs of homologous 

 contrasting characters approach the nuclear plate in such a way as to 

 form a mutually reciprocal working combination of all the vegetative 



