peters: moisture requirements of seeds. 31 



and working at first in July in a room where it at times be- 

 came exceedingly warm, a number of the early tests failed 

 because the vapor caused the seal to buckle and loss of mois- 

 ture resulted. The unnoticed loss of sand particles in remov- 

 ing covers when placing bottles in the oven, caused on one 

 series alone some seventy useless weighings in the endeavor 

 to secure constant weights. But when the difficulties had been 

 overcome, results were secured as shown in table 3, the first 

 ones naturally being too high. 



Only those tests are quoted which may be of assistance in 

 reaching conclusions. By "weight of bottle" is meant the tare 

 of the weighing bottle in which the particular sand sample 

 was placed for drying. "Weight with wet sample" is the 

 weight of this bottle and the wet sand sample before going 

 into the oven. "Weight with dry sample" means the weight 

 of this bottle and the sand when a constant weight had been 

 secured by drying. "Loss of water" is the difference be- 

 tween the two just given. "Weight of dry sample" is the 

 net weight of the sand sample after drying. "Per cent of 



Loss of water 



water" = ■. The upper line of figures in 



Weight of dry sample 



each test is the record of the sample taken at the beginning of 

 the test ; the lower one, that at the close of it. 



