360 Journal of Agricultural Research voi. xvm. No. 7 



to its original moisture content by withdrawing as much of the new 

 solution as was required to accomplish this. To prevent loss of water by 

 evaporation, all the cultures were sealed at the beginning of the experi 

 ment by pouring a thin layer of Briggs and Shantz wax {2) over the sur- 

 face of the sand around the seedlings. 



In all experiments of this kind it is, of course, practically impossible 

 to maintain absolute uniformity in the moisture conditions of the sub- 

 stratum. This ideal condition may not even be very closely approxi- 

 mated because of transpi rational water loss, which tends to decrease 

 the moisture content of the cultures and at the same time to increase 

 the total concentration of the solution. In the present work, however, 

 excessive variation in the moisture content of the cultures and in the 

 total concentrations of the solutions was prevented by the frequent 

 addition of distilled water in quantities sufficient to restore the cultures 

 to their original weights. The cultures were weighed daily, and during 

 the later growth stages whenever the atmospheric conditions were such 

 as to produce high rates of transpiration more frequent weighings were 

 made and the original moisture conditions of the sand cultures were 

 restored by the addition of distilled water. The highest water loss from 

 any culture of the three series during the interval between two successive 

 weighings was not greater than 4 per cent of the original volume of the 

 solution present in the culture; or, on the dry-weight basis, this was not 

 greater than 0.4 per cent for the series with the lowest moisture content, 

 0.6 per cent for the series with medium moisture content, and 0.8 per 

 cent for the series with highest moisture content. This, of course, repre- 

 sents the extremes in the variations of the moisture conditions. Ordi- 

 narily the decrease in the moisture content of the cultures and the 

 consequent increase in the total concentration of the solution during the 

 intervals between two successive weighings were very much less than 

 this. Small variations in the moisture conditions of the substratum, 

 such as were encountered in these sand cultures, could scarcely be expected 

 to have any material influence upon the physiological properties of the 

 solutions as these affect the growth of the plants. 



The three series of cultures here considered were conducted simul- 

 taneously for a period of 28 days after the seedlings had been trans- 

 planted to the sand cultures. The first triple series was conducted from 

 November 30 to December 27, 1917. The second triple series, which was 

 exactly like the first, was carried out between January 1 8 and February 

 15, 1 91 8. At the end of the growth period the wax seals were removed 

 from the cultures, the plants harvested in the usual manner, and the dry 

 weights of the tops and roots obtained separately. 



