82 WATER REQUIREMENT OF PLANTS. 
Of the crops tested, those most efficient in the use of water are 
the millets, the sorghums, and the corns, their water requirement 
ranging from 275 to 370. The legumes, alfalfa, sweet clover, and 
Canada pea, have the highest water requirement, ranging from 710 
to 1,070, while the small grains, wheat, barley, oats, and rye, have 
a water requirement ranging from 510 to 700. Potatoes and sugar 
beets are intermediate between the small grains and the corns. The 
efficiency of the grain sorghums and millet is again apparent in grain 
production. Spring rye and Canada peas are least efficient, while 
oats, wheat, barley, and buckwheat are intermediate. Varieties of 
barley, wheat, oats, and sorghum show great differences in the 
water requirement for grain production. 
In the investigations which have been discussed heretofore the 
determinations of the water requirement have been limited to 
different crops. In this investigation the water requirement of 
different varieties of the same crop was also determined, and measur- 
able differences were found. These varietal differences suggest not 
only the superiority of certain strains but also the desirability of 
applying water-requirement measurements in the selection of varie- 
ties for semiarid regions. 
CONCLUSIONS. 
In comparing the results obtained by different investigators rela- 
tive to the water requirement of different species of plants, two points 
should be kept in mind: (1) A real difference in the results is to 
be expected if the climatic conditions differ at the places where the 
experiments were conducted. From this point of view the observed 
differences in the water requirement may be looked upon as a meas- 
ure of the climatic differences. (2) Differences due to the method 
of experiment, to the use of different varieties, to the amount of plant 
food available, to the period during which the crop was grown, and 
to other similar factors must be classed as experimental errors in a 
comparison of this kind. The summary of the results of different 
investigators as presented in Table LXXII should then be considered 
with the points mentioned above in mind. Often the results are not 
strictly comparable, even when forming part of the same investiga- 
tion, owing to differences in the duration or period of growth, and 
the more detailed statement given elsewhere in the bulletin should 
be referred to in this connection. 
285 
