188 THE DEVELOPMENT OF RAW COTTON 



balance, and the weight calculated to that of 1 centimetre 

 of a single fibre. 



The micro-balance used by the author was home-made, 

 the torsion spring being fine capillary glass rod, mounted 

 with sealing-wax at either end into a frame made of glass 

 tube. The transverse lever was made of fine capillary 

 glass tube, with a hook at one end to hold the fibres, and 

 a counterpoising tail, the motion of which was observed in 

 a mirror; a drop of sealing-wax united the lever to the 

 torsion rod where they crossed, and the particular instru- 

 ment employed was thus easily made to give 50 milli- 

 metres deflection for 1 milligramme, which was sufficiently 

 sensitive for preliminary purposes. 



The probable error of fibre-weight determinations 

 would seem to be high, but the greater part of this is due 

 to difficulties in sampling. 



Strength. The original work upon the breaking strain 

 of single fibres is that of O'Neill, who rightly observes 

 that : " Experimenters appear to have been deterred from 

 manipulating with the individual hairs, on account of 

 their ^mallness and lightness." O'Neill's paper has 

 suffered from endless citation, but it was a very neat and 

 accurate piece of work. The number of fibres he exam- 

 ined was not sufficient to give the general certainty to his 

 figures which have since been attributed to- them, only 363 

 fibres in all having been tested from seventeen different 

 samples, representing about two weeks' steady work with 

 the method he employed. But he published all his figures. 

 From these we can work out the statistical significance, 

 which is the same as Mr. Hughes and the author have 



