ioo#.] IMPROVEMENT OF THE POTATO. 417 



7. DISCONTINUOUS VARIATIONS 

 MUTATIONS 



It has long been recognized that in potato varieties there some- 

 times appear marked bud variations which when propagated are 

 true to type. If we accept DeVries' idea of mutations, and if we 

 believe that bud mutations are of practically the same nature as seed 

 mutation ; then they may be either progressive, degressive or retro- 

 gressive. Such variations which have been noticed have always 

 been relatively wide ones, but it is not inconsistent with the theory 

 to have mutations which are within the limit of fluctuating varia- 

 tions. As the writer understands the question, these mutations may 

 be either bud-mutations, which may not be propagated by seed ; or 

 mutations affecting the gametic structure, which will always be in- 

 herited unless new mutations intervene. 



Darwin (21) mentions three cases, one in which a single white 

 eye in the purple variety Forty fold, became the "parent" of a white 

 variety. In another case this same variety produced a whole white 

 tuber which bred true. The third case was that of -the white Kemp 

 potato which produced a red spot which was propagated and 

 yielded a variety of much prominence, which was called Taylor's 

 Forty-fold. In the United States there has been a number of vari- 

 eties on the market for several years which have originated in this 

 way. Among them are Thorburn's Late Rose, the White Victor and 

 White Early Ohio. I should roughly estimate that less than 0.5 

 percent of our present varieties are from bud-mutation. The gen- 

 eral belief is that these variations are confined to tuber color or pos- 

 sibly to tuber color and shape. Wohltmann (43) has offered an ex- 

 ception; a variation in flower color in the Leo variety although it 

 was unknown whether the variety came from one or several seed- 

 lings. It certainly seems that the variations are almost always con- 

 fined to the tubers, but this should be expected as the tubers are the 

 modified part. The probability is that color is the only character 

 that is easily gauged, and that if accurate metnuus of estimating 

 other characters were used, they too would be found to vary. The 

 supposed rarity of these occurrences has made them of little com- 

 mercial importance, but it is very possible that with strict search, 

 they might be shown to occur much oftener than is expected. 



As stated before, data which I have collected appear to show 

 that bud-mutations are usually and possibly always the loss of 

 the dominant character of an allelomorphic pair, with the conse- 

 quent appearance of the recessive character. This data will be pub- 

 lished in a separate paper. 



