182 Report OF THE HORTICULTURIST OF THE 
with the spading fork to the required. depth, after which the sur- 
face was made fine and level with the rake. Sometimes after 
slight rains the surface of all was raked over without using the 
fork. The weeds were pulled from the other plats as they 
required it. 
Once each week, commencing May 15 and ending October 3, a 
sample of soil was taken to the depth of one foot, for drying, from 
each of the ten plats. The sampling was done with a tool resem- 
bling in form a butter tester, but much larger and stronger. It 
was made by splitting a section of one and a half inch steel tube 
through the center. The lower end of one piece was then rounded 
and made somewhat thinner, to form the point, and the edges were 
sharpened on the inside, while to the upper end a very strong 
shank, provided with a socket to receive a handle, was riveted. 
In dry weather it was found impracticable to turn the tool 
about, preparatory to drawing it out, if inserted the whole twelve 
inches at once. The first insertion was therefore made to the 
depth of about six inches, and in the bottom of the hole thus 
formed, a second was made of such a depth that the two repre- 
sented a cylinder of soil twelve inches long. The loose soil, that 
fell into the hole as the first section was drawn out, was rejected 
from the sample. In dry weather, the surface soil on the tilled 
plats was scraped away with the foot to the depth of a fourth or 
half an inch before taking the sample, it being assumed that this 
dry soil serves no purpose to vegetation except through its office 
asamulch. As, in the mulched plats, the mulch itself was not 
included in the sample, it seemed hardly proper to include that 
part of the soil on the tilled plats which could act only as a 
mulch. 
The samples from each pair of duplicate plats were placed in a 
fruit jar as they were drawn. ‘The jars, which were sealed as they 
received the sample, to prevent evaporation, were then carried to 
the laboratory, where the soils, after being weighed, were exposed 
in a drying oven surrounded by water until they ceased to lose 
weight, after which each sample was washed through a sieve of 
ten meshes to the inch and the gravel and other matters that 
would not pass the sieve, dried, and their weights deducted from 
that of the fresh and dried sample. The per cent of moisture 
