22 
AMERICAN JOURNAL OF BOTANY 
[Vol. io. 
loosely plugged with glass wool, was vertically placed against the wall of 
the jar. Its lower end rested on the bottom of the jar and the upper end 
extended just above the rim. The tube was held in place while the sand 
was added. Each jar was filled with 1250 grams of carefully washed sand. 
For the supply orifice, 100-cc. wide-mouthed bottles with the bottoms 
removed were used. Each bottle rested on an inverted glazed porcelain 
dish according to the method of McCall and Richards (11). After five 
seedlings were transferred to the s.and medium, enough solution was added 
through the mouth of the bottle to bring the liquid to about 1 cm. above 
the sand, and the crock was slightly jarred to settle the sand about the 
roots of the seedlings and to give a level surface for the wax seal (80 parts 
paraffin, 20 parts vaseline). The outlet tube was then connected with a 
vacuum pump, and the excess solution was drawn off. The sand was 
flooded again, and suction was applied until the moisture content was 
optimum. 
When the seedlings were in place the total weights of cultures were 
recorded. All solutions were renewed twice a week throughout the con¬ 
tinuance of the experiment. At the time, the cultures were weighed and 
the amount of water lost by transpiration was recorded. 
The growth period extended from January 15, 1921, to March 24, 1921, 
which brought the plants to the stage at which the advanced cultures were 
just beginning to head. 
Determination of Hydrogen-Ion Concentration 
After the plants were harvested from the sand and water cultures, 
samples were taken from each culture for the purpose of determining the 
hydrogen-ion concentration of the medium after the plants had grown in it 
for one 3^-day period. 
The apparatus used in this work included the following pieces obtained 
from the Leeds and Northrup Company: potentiometer, 3-dial resistance 
box, Wall d’Arsonval galvanometer (type P), Eppley standard cell, single¬ 
contact key, one double-throw switch, and two dry-cell batteries. An 
Ostwajd normal calomel electrode and a Hildebrand hydrogen-gas electrode 
were used. The hydrogen, obtained from a tank of compressed hydrogen, 
was purified by bubbling through a 5-percent solution of potassium per¬ 
manganate, then through a 5-percent solution of pyrogallic acid, and finally 
through distilled water. The hydrogen-electrode vessel used was similar to 
the one described by Sharpe and Hoagland (18), consisting of a small wide¬ 
mouthed bottle of 75 cc. capacity. This was fitted with a rubber stopper 
into which were inserted the hydrogen electrode, a small exit tube, and a 
small bent glass tube filled with agar jelly saturated with KC 1 . Connection 
was made with the calomel cell by means of a beaker containing a saturated 
KC 1 solution and the agar tube. 
The solution or suspension to be tested was placed in the hydrogen- 
