34 BULLETIN 195 ; U. S. DEPARTMENT OF AGRICULTURE. 
It is evident from the data presented that other causes than that of 
inherent unproductiveness must have operated to lower the yield in 
selections 1 to 5. This fact is made still clearer by the behavior of 
selection No. 4, in which the first unit gave a reasonably good yield 
of primes or merchantable tubers, while the remaining ones did not 
produce any. It is believed that the seed tubers from the 1911 selec- 
tions were either infected with some obscure disease or happened to 
have been planted in previously infected soil. All the selections were 
grown in the same or in contiguous rows, so there is little likelihood 
that the moisture or plant-food content of the soil was deficient in 
the one case and not in the other. The behavior of the plants during 
the growing season strongly corroborated the disease theory and 
sustains a previous statement in this bulletin in connection with 
varietal tuber-unit studies, namely, that the chief value of such 
studies consisted in the elimination of diseased and weakened plants. 
It would seem probable from the behavior of No. 25 that a strong 
vigorous, and productive strain had been isolated. The term " prob- 
able " is used advisedly, because, as previously stated, the best selec- 
tions, including, of course, No. 25, were unfortunately cut to and 
below the ground by a very severe freeze late in April, 1913. The 
injury sustained was so severe that very few of the plants survived, 
and those that did survive made a very unsatisfactory growth and 
crop. These selections have therefore been lost, and a new start has 
become necessary. 
Thus far, all selections that have been made for disease resistance 
have proved undesirable, as they either did not retain this quality or 
else they were unproductive commercially or otherwise. The highest 
degree of success can only be attained from either the tuber-unit or 
hill-selection method by working with rather large numbers. It is 
hardly conceivable that there exist many strains within a variety that 
are especially productive or commercially desirable, or that are mark- 
edly resistant to disease. Occasionally one may be fortunate enough 
to isolate such a strain with a minimum amount of effort, but the 
chances are strongly in the opposite direction. This statement is not 
made for the purpose of discouraging anyone from attempting to 
improve his seed stock by up-to-date seed-selection methods. The 
intention is rather to encourage the selectionist to make a larger num- 
ber of selections and thereby increase his chances of securing a supe- 
rior strain. 
The selection of a large number of high-yielding hills which are then 
thrown together for mass planting the ensuing year is not likely to 
result in any marked improvement except by the elimination of the 
diseased or the unproductive plants. The only certain method of 
securing a superior strain is to plant each selection separately, asrec- 
