PETERS: MOISTURE REQUIREMENTS OF SEEDS. 27 



when mixing was complete. The wet sand was placed in the 

 tumbler, the seeds were spaced more or less evenly about four 

 centimeters below the surface, and the sand was settled by 

 jarring the tumbler against the table. Enough of the melted 

 paraffin-vaseline mixture (20 per cent vaseline in paraffin 

 having a melting point of 45' C.) was poured over the surface 

 to seal it effectively, and the labelled tumbler was set aside at 

 room temperature for two weeks. As sufficient growth did not 

 occur for photosynthesis to become a factor, light was disre- 

 garded. 



In this connection, it should be stated that the first series of 

 tests, some thirty, failed because the seeds were planted about 

 a centimeter only below the surface of the sand. The clue was 

 found when a sample was taken from the top and another 

 from the bottom of the sand at the close of one of these tests, 

 run for moisture content, and compared. That from the bot- 

 tom showed a higher moisture content than the upper one, 

 where the seeds were. A series was then run upon a tumbler 

 machine (the one described by Shull, Bot. Gaz., 62:10-11). 

 The bottles were half filled with the wet sand, the seeds were 

 added, heavily shellacked corks were sealed in place, and the 

 bottles fixed upon the wheel of the machine so that they had 

 fifteen complete rotations a minute. This so mixed the con- 

 tents of the bottles that there could be no question as to the 

 moisture content in the various parts of the soil mass. The 

 results were checked with another series in which the seeds 

 were placed near the center of the sand mass, the tumblers 

 sealed as usual, and set aside for the regular time. As results 

 corresponded closely, the more troublesome machine method 

 was not further used. 



While filling the tumblers a carefully chosen sample of the 

 sand was placed in a tared weighing bottle and this was imme- 

 diately covered. Although this sample was taken when the 

 tumbler was half filled, and although all speed commensurate 

 with careful work was used, yet on dry days considerable loss 

 of water must have occurred from the sand not yet in the 

 tumbler and from the surface of that already in it. This 

 sample was carefully weighed upon a standard balance sensi- 

 tive to .0001 gram and was then placed with cover removed in a 

 drying oven at 100 to 104' C. until a constant weight was ob- 

 tained. Another source of error is to be noted here. The par- 



