SELECTION AND SAMPLING OF SEED. 16T 



to ten bags, from ten when twenty bags, and twenty 

 samples when there are fifty bags, etc., for over 500 

 bags, one in every five. The seeds thus obtained are 

 placed in a flask, well corked, and remain there until 

 needed for analysis in a germinator. If several sam- 

 ples are to be taken from the large sample, the seeds 

 are spread out on a table and divided into as many 

 parts as there are tests to be made. 



The unique sampling is more accurate, for the 

 more the seeds are manipulated, the greater will be 

 their loss of impurities. The germinating power of 

 the seed varies from year to year; owing to existing 

 frauds, it is most difficult to get an average sample. 

 There is great need of some uniformity of method of 

 purchasing and testing seed. However, when making 

 the first sampling from the sacks of seed, an average 

 should be obtained from the start; it would be a mis- 

 take to select only from the upper surface, as the seed 

 there is the lightest, but samples should be taken from 

 the bottom, middle and top, so that the total seed 

 obtained should weigh at least io| ounces, or 300 

 grams. The ultimate selection may be made from this 

 preliminary sample. 



By the Nobbe method 300 grams are thoroughly 

 mixed and then emptied into a funnel-shaped hopper, 

 the bottom opening of which is sufficiently small to 

 permit very few seeds to pass through at a time. At 

 regular intervals, timed by a watch, samples of seed are 

 taken, which are received in a special spoon. After a 

 given number of spoonfuls are obtained, they are 

 spread over a black surface, from various parts of 

 which are taken twenty or thirty seeds, this operation 

 being repeated about twenty times, until 600 seeds are 

 obtained, which are divided into three lots of 200 each, 

 and are respectively used for the determinations of 

 moisture, impurities and germination. 



Another method for sampling differs from the 



