12 ESTABLISHMENT OF VARIETIES IN COLEUS 



has led into purely speculative fields in the attempts to represent and ex- 

 plain imperfect segregations and variations in the hereditary qualities. 



Furthermore, it appears in final analysis that the extended studies 

 of seed progenies have not contributed anything fundamentally new to 

 the knowledge of the nature of plant characters. At least, the analysis 

 of characters in terms of hereditary units has failed. 



It would seem that, in considering these problems, the studies of 

 variation among numbers of a seed progeny is no more important 

 than studies of variation in progenies derived by vegetative propaga- 

 tion. The latter should give much more conclusive data regarding 

 such questions as the constancy of characters (or of assumed factors) , 

 the purity of apparent segregations, and the frequency, constancy, and 

 nature of spontaneous changes in the expression of characters. 



It is evident that the facts regarding bud variation involve the funda- 

 mental questions of heredity. When such variations occur in a plant 

 that can be propagated vegetatively, there is opportunity to apply the 

 pedigree method of experimental study to successive generations pro- 

 duced by vegetation propagation. The nature, frequency, and perma- 

 nence of such changes as appear can be studied without the complica- 

 tions that are associated with alternation of generations and fertilization 

 as they normally occur even in selfed seed progeny. Special evidence 

 regarding the "expression" of characters, which also bears on the ques- 

 tion of their inheritance, may thus be obtained. 



THE PROBLEM IN COLEUS. 



For the study of variation along the lines indicated above, I have 

 grown a series of 833 plants, all descended by vegetative propagation 

 from two plants of a variety of Coleus. 



Coleus is particularly favorable for such study in that bud variations 

 are frequent and the plant is readily propagated vegetatively. The 

 leaves are in pairs which alternate on a square stem, making but four 

 rows of leaves. Bud variations that appear sectorially can thus be 

 traced with ease. In a young plant lateral branches usually start to 

 develop from the axils of all the leaves on the main stem. In a large, 

 bushy plant many lateral buds remain dormant, but by proper pruning 

 any bud can be forced to develop or it can be propagated as a cutting 

 which will give it full chance for development. Large, bushy plants 

 3 or 4 feet tall grown out-of-doors often have a total of as many as 

 300 branches. In the greenhouse, also, plants can be grown from cut- 

 tings, with the production of many leaves and branches. The ease with 

 which the plant is propagated vegetatively makes it possible to grow a 

 large series of pedigreed plants from any cutting, and to thus test the 

 frequency with which variations appear, the constancy of the different 

 types, and the purity of any vegetative segregations, all bearing on an 

 analysis of the nature and inheritance of the characters concerned. 



