The La evifolia- Family. 269 



character was highly variable. A whole series of varia- 

 tions in the flowers not hitherto described for this spe- 

 cies, was found in the first few days and many more of 

 them were noticed when I visited them later. 



In 1886, 1887 and 1888 I spent the whole summer 

 at a spot a few minutes walk from the field and so had 

 the opportunity of studying this single species for hours 

 at a time every week and often, even, every day. Since 

 then I have visited the place nearly every year, often 

 indeed twice or oftener, every year ; or I have had ac- 

 counts of it from others ; so that I have been able to fol- 

 low the progress of events step by step by collecting ac- 

 counts of the behavior of old mutations, and of the 

 origin of new ones on the field. 



But it soon became evident that to follow the se- 

 quence of events more closely the investigation must be 

 continued on the lines of experimental culture. Mr. 

 Six kindly agreed to my taking some first year rosettes 

 and seed from the field. My object in doing this was 

 twofold. In the first place (we will suppose for the mo- 

 ment that we are justified in regarding the deviations 

 exhibited by an individual as already present in the seed 

 which produces it) we learn by sowing in the garden 

 the seed gathered in the field, in a far more certain and 

 precise manner, what new forms would be produced in 

 the field. In the second place I have saved seed from 

 the plants thus raised in the garden and have sown them 

 and have in this way raised the different "Families" each 

 of which has its origin on the field at Hilversum, and 

 has, in great part, repeated the process of mutation under 

 daily control in my experimental garden. It is evident 

 that in this wav a much more accurate investis^ation of 

 heredity was made possible than could ever ha\'e been 



