556 
AMERICAN JOURNAL OF BOTANY 
[Vol. io. 
culture where the roots of the plants are mainly developed. Tests of the 
moisture content of the sand in the culture vessels here used, when one 
liter of solution percolated through this substratum in 24 hours, showed 
that in the bottom of the vessel the moisture content of the sand approached 
saturation, while in the median regions it approximated 60 percent of the 
maximum water-retaining capacity and decreased progressively toward the 
surface. 
Aeration in both solution and sand cultures was accomplished by means 
of the improvised air pump previously described and illustrated (1). By 
this means a continuous stream of air bubbles was passed through both the 
solution and the sand cultures of series II and IV, the point of entrance of 
the air in both cases being near the bottom of the culture vessel. 
A brief discussion of the question of maintaining a continuous flow of 
the nutrient solution through the culture vessel has been presented recently 
by Trelease and Livingston (8), who have described a satisfactory apparatus 
for obtaining- such a flow. In their argument for the desirability of such a 
modification in the customary methods of studying the salt requirement 
and concentration relations in culture solutions, they have considered the 
comparatively prompt change in the concentration and chemical make-up 
of such solutions when brought into contact with the roots of rapidly grow¬ 
ing plants. Further argument in favor of the continuous renewal of the 
mineral nutrients in solution cultures might well be based upon a considera¬ 
tion of reaction change brought about in the solutions by contact with the 
plant roots, and of the aeration which such renewal affords. Both of these 
factors will be emphasized. 
The apparatus here used, by means of which a continuous renewal of the 
nutrient solutions was accomplished, is diagrammatically represented in 
figure 1. This apparatus consists of a solution reservoir, a constant-level 
reservoir, and a main conducting tube C with its branches leading to the 
cultures. The constant level as indicated in figure 1 is found to fluctuate 
within a range of ^ to I cm., depending upon the quantity of solution in the 
upper reservoir at any given time. It is readily seen that, as the level drops 
away from the opening of the air inlet A, the equilibrium through the tube 
is broken and air enters the solution reservoir and the solution drops through 
B until the level in the lower vessel rises over the opening of the air inlet A, 
and equilibrium is again restored. In siphoning from the constant-level 
reservoir through the main tube C to the cultures, the rate of flow must 
necessarily be controlled, since, as above stated, it was desired to deliver 
approximately one liter of solution per day per culture. It has been found 
repeatedly that the rate of solution flow can not be controlled satisfactorily 
by the use of metal clamps on rubber connections. In order to obtain the 
required rate of solution flow through the culture, a capillary tube of one- 
half-millimeter bore was used to carry the solution from the conducting tube 
C to the culture, the rate of flow being regulated by the length of this capil- 
