94 University of California Puhlications in Botany [Vol. 5 



probably because of the inequality of the surfaces which the 

 seed presented to the air-blast. The seed that had fallen into 

 the first three boxes — nearest the opening of the tube — was then 

 put into the grader a second and a third time, at the end of 

 which the second box was found to contain the majority of the 

 seed. The seed from this second box was then run through 

 the apparatus three or four times more and, at the end, fifty 

 to sixty seeds remained. By marking certain of the seeds it 

 was found that they continually fell during each of the six or 

 seven gradings into the same receiving box or, rarely, struck 

 the edge to fall into the box just before. The same method was 

 employed in the combined boxes four and five and the majority 

 of the seed was at the end found in box five. Similarly for 

 the heaviest seed in boxes six, seven and eight the seed at the 

 end was taken from box seven. After twenty-one gradings, 

 approximately' two hundred seeds of three grades had altogether 

 been obtained. The remaining seed in the boxes was then mixed 

 and, after two or three hundred seeds from the original packet 

 had been added, the whole process Avas repeated. Over two hun- 

 dred and seventy-five gradings were made and, in all, twenty- 

 seven hundred seeds were separated out into grades. 



All the seed was then examined under a low-power magni- 

 fying glass — i.e., a reading-glass lens four inches in diameter, 

 with a focus of nine inches. The extremely small size of tobacco 

 seed makes any such examination a relatively slow and tiresome 

 process. The seed was moved about under the lens on a glass 

 plate with a fine-pointed camels-hair brush. The removal of 

 the ill-formed, chaffy seed that was found in all the grades was 

 the first object of this examination and some 350 individuals of 

 this type were picked out from the whole 2700 seeds examined. 

 In the heavy seed a remarkable uniformity in size and shape 

 was noted, though there were among the 1100 heavy seeds ex- 

 amined some thirty markedly smaller seeds which were arbi- 

 trarily discarded. In all such cases the discarded seed was 

 thrown away and not used in making up other groups of seed. 

 The examination of the medium seed, which had fallen 

 into the intermediate boxes four and five, showed a considerable 

 variation in size. All seed that approximated the size noted 



