Oct. IS. I9I9 Relation of Nutrient Medimn to Plant Absorption 77 



sand, at least for limited periods of time. The physical analysis of the 

 sand was as follows : 



Per cent. 



Passing 60-mesh sieve 60. o 



Passing 80-mesh sieve 26. 7 



Passing loo-mesh sieve 9. 7 



Passing 150-mesh sieve 2.1 



Passing 200-mesh sieve 8 



Passing finer than 200-niesh sieve 6 



The most valuable method of carrying out sand-culture experiments is 

 that described by McCall {26, 28, 2g). By this procedure it is possible to 

 change the nutrient solution as desired. In our experiments large 

 glazed earthenware jars with a capacity of 5 gallons were used with only 

 four or five plants in each jar. It is possible in this way to grow plants 

 without restriction of root development, which might not be possible 

 in small jars. It was not found convenient to withdraw the solution 

 through tubes placed in the bottom of the jars, so a siphon arrangement 

 was finally adopted. A wide-bore glass tube was placed so that it 

 extended to the bottom of the jar. The upper part was bent at right 

 angles and connected through a receiving bottle to a Nelson suction 

 pump. To prevent the sand from being carried over with the liquid a 

 filter of two thicknesses of muslin cloth was fixed on the lower end of 

 each glass tube and held in place by a tightly fitting rubber tube. The 

 use of a little air pressure served to free the filter from any finely dis- 

 solved material which might tend to stop the flow of solution. When 

 such arrangements are properly made, six jars simultaneously can be 

 sucked free of all excess solution in a short period. 



Water transpired or evaporated was replaced by the addition of dis- 

 tilled water, making up to original weight. The jars were moved about 

 by means of a traveling pulley. One difficulty connected with the addi- 

 tion of water at the top is its tendency to wash down nutrients to the 

 bottom of the jar so that the concentration of the solution is not uni- 

 form. In the final sand-culture experiments this has been largely over- 

 come by the use of a glass percolator of i -liter capacity inverted in the 

 middle of the jar and not filled with sand. Most of the water was added 

 through the projecting tube of the percolator and became equally dis- 

 tributed throughout the jar. In order to test the concentration of 

 the solution, portions of sand were removed at various intervals and 

 freezing-point lowerings determined. 



METHODS OF CONTROL 



Various methods of control were used during the investigation. These 

 included determinations of osmotic pressure by the freezing-point method 

 both in solutions and when necessary in the soil or sand directly, de- 

 scribed by Bouyoucos and McCool (j). For estimating hydrogen-ion con- 



