28 ESTABLISHMENT OF VARIETIES IN COLEUS 



new leaves on very many of the plants of this series were strongly 

 laciniate. Cuttings were made from these laciniate-leaved branches. 

 The new leaves that developed on these young plants during November 

 were entire; hence it would seem that the laciniate character in these 

 particular clones of Coleus is in some degree associated with decreased 

 vigor. When most favorable conditions for growth prevail, or when 

 rapid growth is brought about in cuttings, the leaves become entire. 



It is, however, clear that the first appearance of the laciniate char- 

 acter was confined to a few plants and that once it originated it 

 reappeared with marked constancy in the vegetative progeny. 



During the winter of 1914-15 the laciniate character appeared in 

 the manner of a bud variation. A large plant that had been grown out 

 of doors during the summer was in September severely pruned and 

 placed in a pot for further development in the greenhouse. It was 

 intended to use the plant for stock in general border planting and the 

 plant label was not preserved. From the records of the pattern and 

 generation it is clear that the plant itself and all the plants in its line 

 of descent possessed only entire leaves and the plant belonged to the 

 main clone 1. This plant was given the number 9. 



In the course of 3 months numerous new 

 branches developed on the 10 pairs of main 

 lateral branches to which the plant had 

 been pruned. It was noticed that of the 



10 10' 



9 9' 



8 8' 



6' 



4' 



r 



5' 



2' 

 1 1' 



20 main branches, 3 bore branches with 

 laciniate leaves. The positions of these are 

 indicated by numbers 1, 2, and 3 of 

 diagram 2. 



All the branches on all other of the main 

 branches bore only entire leaves. All the 

 branches arising from 2 bore only laciniate 

 leaves, but the branches with entire and 

 with laciniate leaves were sectorially dis- 

 tributed on the branches 1 and 3. The DIAGRAM 2. showing position 

 contrast was most marked, especially when NolT branches on plant 



opposite branches were different, one hav- 

 ing laciniate and the other entire leaves. The sectorial differences 

 appeared in some of the secondary branches and carried the two types 

 into parts of individual leaves. 



The most striking behavior of this series of plants summarized in 

 table 4 is the wide fluctuations in the leaf -shape and in the amount of 

 yellow and green, the marked correlation of decrease of yellow with 

 decrease of leaf area, and the rather pronounced periodicity of these 

 fluctuations. These fluctuations are so general and rhythmic that 

 they can almost be considered constant. Of bud variations there were 

 but 3 cases, giving a ratio of about 1 to 2,530, which, however, shows 



