260 The Pedigree Families. 



as the young seedlings were identifiable they were re- 

 corded and removed. Those which bore the characters 

 of their parents were simply pulled up and thrown away 

 with the exception of some which did not interfere with 

 the growth of their neighbors and could therefore be 

 allowed to flower. The mutants were always carefully 

 removed and planted singly in pots with rich garden-soil 

 for further experiment. This sorting lasted from the 

 middle of May till the middle of June. 



The lata and nanella plants were recognizable at a 

 very young stage and so could be transplanted in large 

 numbers before they became overshadowed and over- 

 grown by the plants which surrounded them. 



The same was true of alhida which was however al- 

 ways very delicate. Not until a much later stage of 

 growth was oblonga recognizable ; and later still rubri- 

 nervis. At this time the bed was thickly covered with 

 the rosettes, the stronger ones overgrowing the weaker 

 ones. There can be hardly any doubt that many mutants 

 especially of the riihrinervis and the other rare types per- 

 ished in this way before I was able to recognize and trans- 

 plant them. Therefore the figures given underestimate 

 rather than overestimate the actual number of mutants. 



About the middle of June the plants were so crowded 

 that there was no longer any hope that any particularly 

 late mutant could maintain itself. I then simply thinned 

 them out as much as was necessary to give those that re- 

 mained room to flower. In August about 700 plants 

 flowered. Later on two of these attracted attention b}^ 

 the great length of their stems which were much longer 

 than those of the rest and towered above them. Thev 

 turned out to be a new form which has appeared again 

 in other strains and will be termed O. Icptocarpa. Apart 



