31 8 Journal of Agricultural Research voi. xix, No. ? 



Another way in which to interpret the results is to consider all tubers 

 of a hill lot as occupying equal parts of a line and to determine the 

 "center of disease," which is the point on the two sides of which there 

 are equal numbers of diseased and, if also possible, of healthy tubers. 

 This center of disease was found, for the 69 hill lots described above, to 

 be on the average closer to the large-tuber end of the hill-lot line, 44 per 

 cent of the line being between the two. That is, there was a greater 

 tendency to show mosaic as the relative size of the tuber was greater. 

 However, this tendency is not marked enough to make it seem desirable 

 to experiment further by selecting tubers according to absolute weight 

 or size. 



Of 357 hill lots planted in another plot, only the 2 to 6 largest tubers 

 of each were planted, in order of decreasing apparent size. On July 

 22 to 26, 98 of the hill lots were mixed — that is, partly affected with 

 mosaic. The results are similar to those given in Table III, the per- 

 centages being 57, 44, 48, and 35, respectively, for groups i, 2, 3, and 4. 

 The average center of disease is 46 per cent of the distance from the 

 large-tuber end of the hill-lot line. Before this, on July 2 to 14, only 

 42 hill lots were mixed; and later, on August 22 to 25, a number of hill 

 lots were either dead or too mature to show mosaic distinctly. 



RElvATlON TO POSITION OF SEED PIECE IN THE TUBER 



On July 29, 1 91 8, 18 tuber units were observed which had been 

 planted with quartered tubers and were mixed. Of the hills from stem- 

 end quarters, 45 per cent were mosaic, while 62 per cent of those from 

 bud-end quarters were diseased. Likewise there were 24 mixed tuber 

 units of six plants each. Of the hills from stem-end sixths, middle-part 

 sixths, and bud-end sixths, mosaic hills constituted, respectively, 43, 

 54, and 61 per cent. No attempt was made to sterilize the knife used 

 to cut the tubers. 



In 1 91 9 each tuber was cut by means of one of several knives used 

 in rotation and kept, when unused, with blades immersed in 4 per cent 

 formaldehyde solution. Observations made June 28 to July 14 dis- 

 closed 44 tuber units, out of 1,109 observ^ed, to be mixed. In these, 

 48 per cent of the plants from stem-end quarters and 51 per cent of 

 those from bud-end quarters were mosaic. This slight difference had 

 become more marked at the time of the next observation on July 22 to 

 26, when 84 tuber units, out of 1,348 observed, were mixed. At that 

 time 28 per cent of the plants from stem-end quarters were mosaic, 

 while 61 per cent of those from bud-end quarters were diseased. This 

 difference was reduced slightly when it was found on August 22 to 25 

 that 20 more of the tuber units were mixed. The preponderance of 

 mosaic in bud-end hills is of no value in the problem of control because 

 of the small percentage of tuber units that are rriixed. Its cause is not 

 understood. 



