580 Journal of Agricultural Research voi. xni, no. h 



content of the soil used and its wilting coefficient. The method of sealing 

 the cans and watering the plants was the same as that used in the deter- 

 mination of the water requirement of corn and the sorghums.^ During 

 the period of the experiment water was added to the cans from three to 

 four times each day, in order to keep the moisture content of the soil as 

 nearly constant as possible. 



The number of plants was reduced to one to each can for BlackhuU 

 kafir, Red Amber sorgo, and the different varieties of corn, while for 

 Dwarf milo, feterita, Dwarf BlackhuU kafir, and Freed 's sorgo the number 

 of plants varied from one to three to each can. The number of plants in 

 each is shown in the tables that record the data of the different experi- 

 ments. The plants thus grown were as large and vigorous as those 

 growing in the field under favorable conditions (PI. 62-63). All the 

 leaves remained vigorous long after the plants had reached their full 

 vegetative growth, while all the plants that were allowed to mature 

 produced a normal yield of grain. 



EVAPORATION 



The evaporation v/as determined by means of the Livingston porous-cup 

 atmometers, and the cups used both in 191 6 and 1917 had a coefficient of 

 74. These atmometers were connected with burettes that were graduated 

 to o. I c. c, and readings were made every two hours at the time of weigh- 

 ing the cans. In 191 6 the atmometers Vv^ere placed at a height of 2 feet 

 in the open, while in 1 91 7 they were placed at the same height in the center 

 of a plot that was planted to corn. The cups were thus shaded after the 

 corn had reached a height of 2 feet, so that the rate of evaporation as 

 shown in the results was not so high in 191 7 as in 191 6. 



DETERMINATION OF TRANSPIRATION 



The cans were mounted on small wooden platforms provided with 

 castors and were moved-about on tracks made of unmatched boards. 

 The cans were left in the open on the surface of the ground for the experi- 

 ments in 1 91 6, but in 1 91 7 they were placed in a pit in the center of a plot 

 that was planted to corn. The pit was of such a depth that the surface 

 of the cans were on a level with the top of the ground (PI. 62, C). This 

 arrangement placed the plants more nearly under field conditions and 

 reduced to a minimum the injury to the leaves on account of the pre- 

 vailing high winds. 



The plants were weighed on scales of the platform type that had a 

 carrying capacity of 1 80 kilos and were sensitive to about 5 gm. The 

 cans were weighed in a scale house in order to avoid any error in weight 

 caused by the wind. The time required to bring the three or four plants 

 used in the experiment to the scale house, weight them, and return them 



' Miller. E. C. op. cit. 



