74 ESTABLISHMENT OF VARIETIES IN COLEUS 



SUMMARY. 



(1) A single variety of Coleus propagated vegetatively by cuttings 

 in two main clones has shown (a) gradual fluctuations and (6) sudden 

 mutations, giving a total of 16 distinct and characteristically different 

 color patterns. 



(2) (a) A total of 15 patterns (see diagram 2) arose by sudden muta- 

 tion affecting a part of a leaf or a branch, or a series of associated leaves 

 or branches; (b) 6 of these 15 patterns (figs. 2, 4, 5, 10, 13, and 13a) 

 also appeared among the fluctuating variations; (c) one type of color 

 pattern (fig. 15) has thus far appeared only as a fluctuating variation. 



(3) (a) Six (diagram 2, and figs. 1, 4, 5, 6, 8, and 12) of the 15 color 

 patterns arose directly from the parent type bj^ sudden bud variation. 

 One of these 6 (fig. 4) also appeared as a fluctuating variation. (6) Five 

 of these 6 tj^pes (figs. 4, 5, 6, 8, and 12) propagated by cuttings showed 

 further fluctuations and bud variations, giving (a) the parent type 

 (fig. 2), (6) 4 of the 6 types already directly derived (figs. 1, 2, 5, and 12), 

 and (c) 8 new types (figs. 9, 10, 11, 13, 14, and 16, also 8a, 13a, not 

 illustrated). 



(4) The variations in the development of color patterns mentioned 

 above involve (a) increase and decrease of green and yellow, (b) increase 

 and decrease of red pigmentation, (c) reversals of the relative positions 

 of the green and yellow by which a type with green center and yellow 

 border (fig. 2) gave one with yellow center and green border (fig, 6), 

 and (d) changes in the distribution of the red pigmentation especially 

 giving concentration in the epidermis of the upper surface of the leaves. 



(5) Progeny of 11 types of color pattern have been grown through 

 from 2 to 6 generations, as follows: 



Type of figure 1 2 4 5 6 8 9 10 12 13 14 



Number of generations 2 6 6 6 6 5 2 2 2 2 2 



Total number of plants 11 337 198 90 41 54 8 7 7 7 4 



Some of these types have shown themselves more constant than the 

 parent type (fig. 2), others were less constant. All varied about a new 

 mode and all would be considered good horticultural races. 



(6) The relative constancy of color-pattern types derived by the 

 accumulation of fluctuating variations was tested in two cases: Type 

 of figure 2 (see clone 14 of table 2), 3 generations, total, 45 plants; 

 type of figure 4 (see clone 13 of table 3) 3 generations, total, 79 plants. 

 In both cases the constancy of the progeny showed no essential differ- 

 ence from that of the same types obtained by sudden bud variations. 



(7) Variations in leaf form, even more striking than changes in color 

 patterns, from entire to deeply laciniate-leaved forms, appeared in 13 

 instances as fluctuations affecting an entire plant and in one case 

 (during the winter of 1914-14) as a bud variation. The striking 



