SELECTION AND SAMPLING OF SEED. 



157 



determining, if observations are being made, on \vha^ 

 might be considered a single seed with about five 

 germs, or whether it is composed of two seeds held 

 together, which, collectively, have five germs. The 

 Nobbe experiments would tend to show that the seed, 

 or germ proper, represents 31 per cent, of the seed, 

 while the ballast, or pericarp, is 68.8 per cent. 



The data furnished by other agronomists upon 

 this subject differ somewhat from these figures. None 

 are more reliable than those of Knauer, and he declares 

 that, notwithstanding all the precautionary measures 

 taken to determine the weight, the data obtained are 

 certainly not mathematical. In the experiments 200 

 seeds weighed 6.099 grams, of which the ballast 

 weighed 4.487 grams, or 73.6 per cent., and the seed 

 proper 26.4 per cent. There can be no doubt but that 

 the weight of the germ increases with the weight of the 

 seed. The experiments at Grobers were upon 50 

 seeds, but to make these results more readily under- 

 stood, we have based our calculations upon 100. 



Size of Seed. 



Large 



Small 



Smaller... 

 Smallest.. 



One Germ 

 Weighs. 

 3.3 M. G. 

 3.0 " 

 2.7 " 

 2.7 " 



Knauer declares that it is a mistake to suppose 

 for one instant that the large seed is simply a com- 

 bination of two of the smaller seeds; a close examina- 

 tion reveals that such is not the case. For in these 

 large seeds may be found, side by side, much smaller, 

 yet too large to fall through the holes of a 7 m. m. 

 mesh. The same argument applies to the smallest seed 

 of the table. 



Upon general principles, it may be admitted that 

 one hectoliter of beet seed weighs 27 kilos (about 

 twenty pounds per bushel). There is a great variation 

 in the weight of seeds, considerd as a whole. Dr. Bret- 

 feld has declared that there may be 14 to 103 seeds per 



