60 ESTABLISHMENT OF VARIETIES IN COLEUS 



hybrid forms always segregate out into only these types is somewhat 

 in doubt, for Shull (p. 9) finds that a plant classed as simplex gave 

 unexpectedly a mixed progeny with defective ratio, so that it is clear 

 that the assumed hereditary "gene" became less potent. Hus (1914) 

 distinguished in a culture of Capsella hursa-pastoris 4 forms different 

 from those of Shull and added another factor which he considers 

 determines the narrow character of early leaves in certain forms. 



In none of these studies have the individuals of the F2 exhibited 

 greater variation among themselves than have single plants of Coleus 

 with the laciniate leaf, nor have the individual parent plants been 

 more distinct and uniform as a plant than the plants derived from 

 seed progeny of Coleus, and grown for a period of several months. 

 Furthermore, there has been evidently no attempt to select persistently 

 intermediate types for modified potency of characters. Furthermore, 

 emphasis has not been laid to selection of variations in a seed progeny 

 of a single individual or in a line propagated vegetatively. The 

 more intensive Mendelian studies, such as those by Shull, Hus, and 

 Gregory, indicate that the character of leaf-shape is complex and that 

 selection studies along the lines indicated ma\' reveal further data 

 on variations in the potency of characters. 



The historical evidence and the studies of seed progeny reported 

 above indicate that the strain of Coleus studied is most certainly of 

 mixed parentage. If this strain had been studied solely in its seed 

 progeny the variations obtained would be attributed by many modern 

 geneticists, I venture to imagine, to chance combinations of hereditary 

 units. Yet, as has been fully reported above, these variations are 

 analogous to and even identical in nearly all cases with those arising 

 by bud variation. 



At this point we may note that modern genetics have furnished no 

 evidence as to the real nature of the characters considered. What 

 we may call the genetical or breeding value of characters has alone 

 been emphasized. Characters have been considered solely in regard 

 to their unity of expression in successive generations of plants of 

 selfed or hybrid origin. Special emphasis has been placed on the 

 reappearance of characters and upon their phylogenetic significance. 

 From this standpoint we may further consider specifically the char- 

 acters concerned in these studies of Coleus. 



Considering first the color characters, we note that the pattern of 

 the parent plants was a mosaic of green, yellow, red (or blue), and non- 

 red cells. In the mature leaf the cells are apparently qualitatively 

 different, and furthermore, the color differences between the various 

 cells are identical with the color differences between entire leaves 

 and branches derived by bud variation. The contrast between pure 

 j^ellow and pure green leaves in the bud sports and on plants of seed 

 origin is the same that exists between green and yellow cells that may 



