332 Journal of Agricultural Research voi. xix, no. 7 



It will be noted in Table VII that in 191 7 certain factors seemed to be 

 more favorable for the spread of mosaic than in 191 8 — namely, higher 

 percentage of diseased hills (rogued) in the plots, greater proximity of 

 unrogued mosaic stock, and higher percentage of mosaic in the nearest 

 unrogued diseased plot. However, there was less spread in 191 7 than 

 in 1 91 8, as shown by the lower percentage of mosaic in 191 8 than in 

 191 9, in correlation with the greater abundance of aphids in 191 8. Fur- 

 thermore, these observations indicate how difficult the problem is of pro- 

 ducing perfectly mosaic-free stocks from susceptible varieties wherever 

 these agents of transmission exist. 



EFFECT OF VARIATION IN THE TIME OF HARVESTING IN I918 



It was expected that if aphids were a deciding factor in mosaic trans- 

 mission the lots of tubers harvested at progressively later dates during 

 their increase in numbers would show an increasing percentage of mosaic. 

 Seventy-eight healthy hills (66 Green Mountain and the rest Bliss Tri- 

 umph or Irish Cobbler) were selected in 191 8 in a plot containing many 

 small lots all with more or less mosaic. Aphids became noticeable on 

 potatoes the last part of July and increased in numbers so that they 

 were very numerous about the middle of August and more excessively 

 abundant as the end of the month was approached. Tubers about an 

 inch in diameter were harvested on August 8 but did not keep with the 

 methods used. Another set of tubers was harvested on August 15 and 

 a third on August 26, one tuber being removed from every hill on each 

 date. On September 12 the remaining tubers — 321 in all — were har- 

 vested. The tubers were planted uncut in 191 9 and transmitted 6, 

 14, and 50 per cent of mosaic, respectively, for the three lots. Ap- 

 parently some of the infection occurred before August 15, but most of 

 it was too late to affect many of the tubers harvested by August 26. 

 This difference can be explained best by the great increase of aphids 

 during August, together with the results obtained in the experiments on 

 aphid transmission. 



TEST OF THE SEED-CUTTING KNIFE 



In 1 91 9 stock was available from 191 8 all-mosaic plots and rogued 

 plots. One hundred tubers from the former were divided by three paral- 

 lel transverse cuts so that no two cut surfaces joined in a seed piece, 

 while 100 tubers from the latter were quartered by a transverse and a 

 longitudinal cut so that each seed piece had two cut surfaces joining at 

 a right angle. The same knife w^as used, cutting alternately tubers from 

 the two lots. The 800 sets were left mixed in the same sack for over a 

 day and planted by hand at 15-inch intervals in two rows. Another 

 mixture was prepared in the same way with 200 tubers from the same 

 two barrels, but in this case the pieces from the all-mosaic lot were sorted 



