POT EXPERIMENTS WITH FERTILIZERS. 165 
If the greater part of the water used was poured on the 
surface of the soil in the pots it might dissolve out the nitrog- 
enous fertilizers as it percolated through the soil, and carry 
them to the bottom of the pots and perhaps out of the immedi- 
ate vicinity of the roots of the plants. ‘To avoid this as much 
as possible, the plants were watered mainly from the bottom 
of the pots, by means of the pipe on the outside. ‘The water 
used was taken from a driven well at the College, as analysis 
had shown this water to be practically free from nitrogen. 
The water was supplied to the plants as it seemed to be needed 
to keep up a vigorous growth. The need of water was seldom 
shown by the wilting of the plants, which would seem to indi- 
cate that the quantity supplied was sufficient for normal growth. 
Sampling and analyzing the crops.—When the crops were 
ready to harvest the total growth on each pot was cut about 
one inch above the soil and in each case (except the straw of 
the soy beans) the total crop was taken fora sample. In the 
oat experiments the seeds were separated from the straw after 
the sample was thoroughly air-dried, and the loose hulls were 
added to the straw, the seeds and the straw being analyzed sep- 
arately. In the experiments with Hungarian grass and those 
with orchard grass the crops were harvested in the early seed 
stage and the samples included the total amount of fodder from 
each pot. In the soy bean experiments only the seeds were 
sampled, no effort being made to save the stems and leaves, as 
it had been found in field experiments that many of the leaves 
dropped off before the seeds were fully matured. All of the 
crops were thoroughly air-dried and kept in sealed glass jars 
until prepared for analysis. 
The only analyses of the crops from the pot experiments 
were determinations of nitrogen, as these were sufficient for 
the purpose of the experiments. The total nitrogen was de- 
termined by the Kjeldahl method. Nitric nitrogen was also 
determined in a number of samples, as described in a preceding ~ 
article on the Analyses of Fodders and Feeding Stuffs (page 
Sa. 
The data in the tables below for the experiments with differ- 
ent crops show for each pot the amount of nitrogen in the fertil- 
izer, the total weight of crop produced, the percentage and total 
weight of nitrogen and the estimated percentage of protein. 
