On Cell-wall Structure as seen in Cotton Hairs. 433 



spinning properties, or on the physical properties of the hair, or of their 

 origin. The senior author has regarded them* as a necessary consequence of 

 the simple pits in the wall, with modifications caused by wall-thickness 

 variations, and we can now extend this to include the pit spiral structure, 

 which should completely explain the convoluted form of the collapsed dead 

 hair. 



Actually, however, the explanation is not yet complete. Mapping pit 

 spiral against convolutions, in any one piece of hair (fig. 10), there is a 

 general similarity, but by no means exact identity. It is probable, however, 

 that if we could construct scale models of the wall structure (growth rings 

 and pit spiral fibrils) in the form of semi-rigid tubes, and then cause them to 

 collapse, we should find that local variations of texture, packing, inter- 

 fibrillar friction, wall-thickness and hardness, etc., would produce similar 

 discrepancies to those we have observed, and we venture to think that the 

 existence of convolutions can be explained in this way, if we include the 

 reinforced spirals already mentioned, whose presence makes the '• arch " 

 structure unsymmetrical. 



This, however, leaves a major problem to solve, to wit, the reason why the 

 pit spirals, and hence the convolutions, vary their pitch and direction. On 

 grounds of convenience, we have studied the convolutions rather than their 

 causative spirals, and while no definite conclusions have yet been reached, a 

 number of suggestive observations have been made on the daily pickings 

 material,f which consists of fruit-capsules (bolls) opening on ninety succes- 

 sive days : — 



(a) On counting the number of convolutions in a unit length at the middle 

 of the hair (fig. 11, heavy line), we found the average number changing from 

 day to day, as in the case of other hair properties, indicating that environ- 

 mental changes affected the convolutions. Similar results were given by 

 counting at other loci (fig. 11), and by taking the percentage distribution as 

 between these various loci (fig. 12). 



(b) Extending these measurements by mapping the convolutions along the 

 whole length of the average hair (fig. 13), there were indications that the 

 locus of any feature (e.g., a low number of convolutions) shifted its position 

 along the hair on successive days. The earlier the day of boll opening the 

 further removed was the particular feature from the base of the hair, thus 

 indicating that the environmental determinant of convolution form must have 

 operated while the hair was still growing in length. Our evidence is based on 

 too slight data to be conclusive in itself, but it will be noticed that it confirms 



* W. L. B., " Eaw Cotton," pp. 79 and 147, loc. cit. 

 t Vide supra. 



2 H 2 



