I^OTATO Investigations 31 
was used this year as in 1916 and 1917. The seed was limited to 
one-, two- and three-ounce sizes which included whole seed and seed 
cut in halves and quarters so that three sets of comparisons were 
obtained with each of the three kinds and sizes of seed. It has been 
found by previous experience that the planter would not, without 
modification, successfully plant sizes larger than three ounces, in 
fact, three-ounce seed, especially that which was cut in halves, gave 
some trouble. One row, 246 feet long, of each size and kind of seed 
was planted and the series was duplicated. Soil and moisture con- 
, ditions were ideal at the time of planting which are considered 
largely responsible for the nearly perfect stand that was secured. 
The crop was practically free from disease. The results are shown 
; in the 6th, 9th, 12th and 16th columns of Table 12. 
In each of the three groups shown in Table 12 the number and 
weight of tubers per hill increased directly with the increase in the 
j size of the seed piece. The percentage of the weight of tubers 
which were marketable varied inversely with the size of the seed 
piece. In comparing the calculated yields per acre the two-ounce 
seed appears to have produced slightly better results than either 
of the other two sizes though in the two groups where cut seed 
Fig. 5— Residence and laboratory at experiment station. 
