No. 642] 



NATURE OF BUD VARIATIONS 



65 



tion or of somatic segregation without basis from critical 

 inheritance or cytological studies is to prejudge the 

 nature of the observed modification. 



Feequency of Somatic Variations 

 Attempts have been made to estimate the relative fre- 

 quency of vegetative and seminal variations in plants, but 

 little definite information has been gained. The problem 

 is beset with grave difficulties inherent in most attempts 

 to determine coefficients of mutability. The possibility 

 of overlooking even prominent variations until they have 

 once been noted, together with the readiness with which 

 they are found after one's attention has been focused on 

 them, will hardly be questioned by anyone who has given 

 attention to the discovery of new variations in almost any 

 organism. One may attempt witli some assurance an 

 estimation of the frequency of recurrence of a particular 

 mutation, for instance, whether it appears in vegetative 

 parts of individuals or in sexually produced progenies, 

 but it is a hazardous undertaking to estimate the fre- 

 quency of variations in general. Until some one can de- 

 vise a scheme for estimating the frequency of hiid vai-ia- 

 tions as MuUer has done for determiniiii;' imitation fre- 

 quencies in Drosophila, little progress can l)e looked for 

 other than tliroiigli imcstigations of the somatic muta- 

 tion or segri'.u'atioii of specific genes. 



The probk'in of the relative frequency of occurrence of 

 somatic and gametic variations meets the further diffi- 

 culty that it is often impossi1)le to .letermiiie tlie ontoge- 

 netic stage at which particular \ ariatioiis have arisni-a 

 fact that has been noted for i)laiits l)y various writers 

 (deVries, 1910; Emerson, 1913; East, 1017). Both 

 Bridges (1919) and Muller (1920) have discussed this 

 problem from the standpoint of studies of particular mu- 

 tations ill Drnsopliila. The i.revaleiit opinion tliat varia- 

 tions arise in the uanirte> or at alxnit the tiiiu- of their 

 fonnalion may have eonir in part from a belief that aber- 

 rant chroinosome l)..haN ior is most likely to oeeiir at the 

 time of the ivdnetion division. It seems likely, however, 

 that the situation 1ms lieeii confused by failure to realize 



