26 THE UNIVERSITY SCIENCE BULLETIN. 



the water movement in the soil, especially v^hen the amount 

 present was quite low; in other words, the water was evapo- 

 rated from the leaves more rapidly than it could be absorbed 

 from the soil, and wilting followed as the result of this back 

 pull before the amount of water in the soil was lowered to the 

 point reached in the corresponding tests of Briggs and Shantz. 



METHODS. 



Since the purpose of this investigation is to determine if 

 germination can occur with far less moisture than is com- 

 monly thought necessary, since transpiration is not a factor 

 in the tests (thus making them somewhat similar to those of 

 Briggs and Shantz in that they had always a high humidity 

 present in theirs), and since the Briggs and Shantz' figures 

 are lower than Caldwell's, they are retained as the standard for 

 this test. Nevertheless, it is recognized that this may not be a 

 fixed standard for all conditions but may vary with differing 

 atmospheric conditions whenever transpiration is a factor. 



Because quartz sand and its data were available, it was 

 used. It is designated as No. 2/o by its manufacturers, the 

 Wausau Quartz Company, and passes over a 147-mesh screen 

 but through a 124-mesh one, thus making the average diameter 

 of the particles about .10 mm. It contains by analysis : 



Per cent. 



Silicon dioxide 99 . 07 



Iron oxide . 17 



Aluminum oxide . 52 



Hygroscopic moisture . 06 



Undetermined 0.18 



100.00 



Its wilting coefficient, as determined at the biophysical lab- 

 oratory of the bureau of plant industry, Washington, D. C, of 

 which Mr. Briggs is director, is 1.31 per cent (8). 



Two hundred grams of this sand, roughly weighed, was 

 chosen as the unit, merely because it lacked about three centi- 

 meters of filling the common heavy glass tumblers used. The 

 unit of sand was spread upon a glass plate and water to pro- 

 duce the desired percentage of moisture was added from a 

 burette, and thoroughly mixed in with a spatula. Owing to 

 varying humidity conditions in the air during mixing at differ- 

 ent times, accuracy was approximate only, but as a rule about 

 twenty per cent more water had to be added than was desired 



