16 BULLETIN 1248, IT. S. DEPARTMENT OF AGRICULTURE. 
be planted whole or cut into halves or quarters. Therefore, in order 
to insure the greatest degree of 'uniformity all tubers used in this 
experiment were carefully weighed. For example, in separating the 
different lots of tubers only those were taken that came within 5 
grams of the desired weight of 1, 2, 3, 4, 5, and 6 ounces, respectively; 
that is, the maximum variation between the tubers in any given lot 
did not exceed 10 grams. 
Three sets of observations were made in most of the tests. The 
first was the relation of the size of the set to the weight of the crop 
obtained. The second factor studied was the correlation of the 
size of the set and the number of tubers produced, while the third 
was the correlation between the size of the set planted and the num- 
ber of stems produced. The last observation also permits a fourth 
comparison relating to the correlation between the average number 
of stems and the average set of tubers to the stem. In the Virginia 
and Maine experiments the tubers from each set or hill were separated 
into primes and culls on the basis of weight. All tubers weighing 85 
grams or over were regarded as primes or merchantable stock, while 
all below this weight were classed as culls, and counts were made of 
the number of each class of tubers and the weight of each lot recorded. 
At Greeley and Jerome the tubers were separated on the same basis, 
but a record was not kept of the performance of each individual set. 
Closer spacing of the sets in the row was practiced in the earlier 
experiments than after 1915. This was particularly true in the Vir- 
ginia and Maine tests. At the Virginia Truck Experiment Station the 
sets were spaced 15 inches in the row in 1914 and 14 inches in 1915. 
In Maine 12-inch spacing of the sets was practiced in both 1914 and 
1915. In 1916 and thereafter the spacing at all points where the 
work was conducted (except at Greeley in 1915 3 ) was at a distance 
of 16 inches. At the Greeley and Jerome stations the crop was grown 
under irrigation. 
In making the count of stems it was found that prior to harvesting 
the plants field counts were not so reliable as those made at the time 
of digging. This unreliability was due to the fact that in many 
instances what appeared to be two distinct stems turned out upon 
closer examination to be only a single stem, the branching having 
occurred well beneath the surface of the soil. Such branching was 
almost invariably found to be the result of injury to the primary shoot 
before its emergence from the ground. This injury was in most 
cases due to an attack of Rhizoctonia or to a mechanical injury 
from cultivator or harrow teeth in tillage before the plants appeared. 
The quickest and most satisfactory method of making the count of 
stems at harvest time, especially when the stems had dried down, 
was to have a workman partially lift the hill with a digging fork, 
leaving the stems in position in the soil but in such condition as to be 
easily removed and recorded. By having each hill in the row num- 
bered those conducting the work were able to study the relation 
between the number of stems and tubers to the size oi sets. 
In studying the data it is at once apparent that seasonal variations 
as regards temperature and rainfall produce marked variations in 
yield and adversely affect the behavior of whole or cut seed. When 
then' is an abundance of available moisture and plant food the sets 
i ii L918 the planting was done with a two-man plant or with the result thai the spacing was somewhat 
closer, w ii b rows spaced 3 feel aparl and t be sets L 6 inches apart in rows, it requires 1 1 .35 bushels ofseed 
to plant an acre with 1-ouik B sets. The 1918 planting at Greeley required 13 bushels. 
