On Cell-wall Structure as seen in Cotton Hairs. 437 



The actual data were obtained by selecting usually twenty hairs of equal 

 length on each day over a sequence of thirty days, and counting the convolu- 

 tions in successive intervals of 1'5 mm. from base to tip ; they thus comprise 

 some 20,000 measurements, but statistical considerations make it evident that 

 they need to be greatly extended in order to be conclusive, and in practice it 

 should be found easier finally to attempt the proof by means of direct 

 physiological experiment and by observations of the pit spirals. 



A fairly close correspondence was indicated as between the general form of 

 the hair-length growth curve (formerly ascertained)* and the shift of any 

 convolution-form locus from day to day. 



(c) A renewal of a former attempt was also made, in order to see whether 

 any forms or markings could be found along the length of the hair which 

 would correspond for linear extension to the daily marks of the growth-rings 

 in secondary thickening. The system of convolution mapping was extended 

 to measure the length and direction of every separate convolution. These 

 measurements were then plotted as shown (fig. 14), using rectangles of equal 

 area for each one, the bases of which were equal to the convolution length, 

 and in this form of plotting it is very evident that the convolution sequence 

 along any one hair is at least wave-like. Phases of steep-pitched and short 

 convolutions alternate with phases of slow and long convolutions. The 

 discrepancies between convolutions and pit spirals seem to happen chiefly in 

 the latter phase, which also seems, as might be expected, to contain more 

 convolution reversals (fig. 14, b). The number of peaks (short convolution 

 phases), in the curve, seems to tend towards correspondence with the number 

 of days (about 25), during which the hairs used in these observations (daily 

 pickings samples) were growing in length. 



Here, again, no rigid conclusion can be drawn, but the facts are certainly 

 very suggestive of a daily environmental effect, acting by predetermination on 

 the convolutions through the pit spiral patterning of the primary wall. 



(d) We have not forgotten the probability that mutual pressure inside the 

 growing capsule, together with the curling grouping which the hairs thereby 

 take up, may influence both pit spiral and convolution, but we anticipate that 

 this will be found subsidiary to the other causes indicated. 



(e) A source of error in other observations made on material which had 

 been acted upon by softening and swelling re-agents should be noted. When 

 the wall cellulose is softened with the hair held in slight tension, the 

 convolution spiral may become mechanically unstable and instantaneously 

 jump to a new conformation, the hair becoming a cylindrical helix, like an 



* W. L. B., " Raw Cotton," p. 76. 



