METHODS OF INVESTIGATION 185 



Each seed is then combed with a small -tooth comb 

 from the tip of the seed towards the butt, and outwards, 

 at first lightly to disentangle the basal portion of the 

 hairs, and then firmly, holding the seed and all the dis- 

 entangled basal portions tightly between the finger and 

 thumb of the left hand to prevent them from being torn 

 apart. Finally a few strokes of the comb carry away 

 any broken or detached fibres, and the seed is left with 

 a halo of lint around it, chiefly at the basal portion. If 

 the regularity of the lint is also under examination, more 

 careful combing is employed to set each hair out along 

 the radius of a circle with its centre in the seed. 



The seed with its flat halo of lint is laid on a dark 

 background, held down by the forefinger of the left hand 

 resting on the seed ; one leg of the dividers is then brought 

 up against the butt of the seed, and the other is swung 

 around and adjusted until it moves along the edge of 

 the halo. Successive measurements made in this way on 

 the same seed vary only 1 millimetre, so that the halo 

 edge is obviously quite definite, even in poor cotton. 



The mean of seven measurements has a probable error 

 of less than 2 per cent., even in samples which consist of 

 cotton damaged by premature opening, boll- worm, etc., 

 and is thus better than twenty-five measurements made 

 on single fibres. 



The time occupied in the complete cycle of operations 

 is ten minutes, or, when two persons are working together, 

 about four minutes. 



The reason for the superiority of the seed-combing 

 method is obvious; it eliminates systematic variations 



