lOOfi.] PUBLIC DOCUMENT — No. 33. 139 



ceptU^le to disease. In other words, much of the selection 

 would be done more cheaply and easily by sifting the seed 

 than as it is done at the present time, by selection in the 

 seed bed. A sieve of the size mentioned can easily be made 

 by purchasing from almost any tin shop a piece of colander 

 tin of 1 millimeter mesh, which can very readily be soldered 

 to a suitable rim, or even fastened to a tomato can which 

 has had the bottom removed. 



Our experiments in sifting seed have been confined, how- 

 ever, for the most part to lettuce seed; and, while these 

 seeds can be separated in this way very easily, much more 

 difficulty would be experienced in separating some other 

 types of seed, such as turnip or tobacco ; and another ob- 

 jection to this method is that the size and weight of seeds do 

 not necessarily correspond, that is to say, a large seed may 

 not necessarily 1)e a heavy one. When seeds are separated 

 by sifting, while the largest size contains practically all of 

 the heaviest seeds and the lighter seed is practically all in the 

 small sizes, there will be a few light seeds in the larger size 

 and a few heavy seeds in the smaller sizes, so that this method 

 by no means gives an absolute division of the seeds by 

 weight, which is the ideal method of seed selection. 



Mr. A. D. Shamel of the Connecticut (New Haven) Ex- 

 periment Station uses a very satisfactory method for the 

 separation of tobacco seed, which we can do no better than 

 describe in his own words : — 



This seed separator consists of a glass tube 1 inch in diameter 

 and 5 feet long, and a glass receptacle for holding the seeds, 

 having the diameter of the long glass tube, and so arranged 

 with a finely woven wire screen in the bottom as to hold the 

 seeds in the receptacle, and at the same time freely admit a 

 current of air directly into the seed. The top of this recep- 

 tacle is fitted with a coupling into which the long glass tube 

 can be set and held in place. The current of air is developed 

 by a common foot bellows and regulated with a valve. The 

 seed to be separated is poured into the receptacle, usually about 

 1 to 2 ounces at a time, the glass tube set in place and a cur- 

 rent of air pumped into the seed. The lightest seed and chalf 

 are first blown out of the tube, and next the small seed. Small 



