C. H. Bailey 437 



to have the higher specific gravity, and further, that small, compact 

 kernels have a higher specific gravity than the large (aud presumably 

 less compact) ones. 



The methods previously employed for the determination of specific 

 gravity did not appear in most instances to be wholly satisfactory. 

 A new method was accordingly developed by the writer in collaboration 

 with L. M. Thomas^. About 10 grams of wheat kernels, freed from dirt, 

 weed-seeds, other grains, and broken kernels, are weighed on an analytical 

 balance, and the exact weight recorded. The wheat is then placed in 

 a 50 c.c. pycnometer, which is provided with a ground-in thermometer, 

 side-tube, and over-flow cap. The grain is just covered with cool 

 toluol, the side-tube plugged, and the neck of the pycnometer connected 

 with pressure tubing to an aspirator. The air is then exhausted to 

 free the mass of grain from air mechanically held in the brush and 

 crease of the kernels. Unless this is done the air so held will materially 

 reduce the apparent specific gravity. Moreover the cpiantity of air 

 present and removed by aspiration varies, as it depends upon the shape 

 and size of the kernels. In nine trials it ranged from 0-lOIc.c. to 

 0-335 c.c. After the bubbles cease to rise through the toluol, air is slowly 

 admitted, and the pycnometer is disconnected from the aspirator and 

 completely filled with toluol at a temperature of about 18° C. The 

 temperature is allowed to rise slowly to 20°, as shown by the thermometer 

 in the pycnometer, the last drop on the side-tube is wiped off, and the 

 over-flow cap set firmly in place. Pycnometer and contents are then 

 weighed on the analytical balance. The exact capacity of the pyc- 

 nometer and the specific gravity of the toluol must of course be known. 

 The latter averages about 0-8665. The specific gravity of the wheat 

 is calculated according to the following formula : 



.„ ., ,. , , specific gravity of toluol X weight of wheat 



specific gravitv of wheat = — ^^ — ^-~, — p^p — ; t , ? — , . 



weight 01 displaced toluol 



From the weight of toluol displaced by the wheat its volume can be 

 calculated and this figure divided by the number of kernels in the 

 pycnometer gives the average volume per kernel. From the weight of 

 wheat in the pycnometer the weight per 1000 kernels can also be calcu- 

 lated when the number of kernels is known. 



These investigations have shown the kernel density to be dependent 

 first upon the proportion of pericarp and germ to endosperm, and second 

 upon the density of the endosperm. As a general rule the small kernels, 



1 Bailey and Thomas. U.S. Bur. Plant Ind. Circular No. 99, 1912 



