Compound Celluloses 127 



The results observed, which were somewhat unexpected, 

 are given below. 



The N in these analyses was determined by the Kjeldahl 

 method, in the previous cases by the soda-lime combustion. 

 It appears, therefore, that the blue cyanide deposited in the 

 earlier stages of the reaction is fairly constant in composition 

 notwithstanding considerable variations in the proportions of 

 the reagents in the solution from which it is abstracted by the 

 fibre ; and the ratio of Fe to CN in the fibre cyanide complex 

 is approximately 1:3. By the last-cited experiment it is also 

 shown that the rearrangement of the Fe and (CN) which takes 

 place within the fibre-substance is to a certain extent inde- 

 pendent of the condition of the Fe in the solution, i.e. 

 whether added as basic Fe or as in the ferricyanogen complex. 



The main points of the reaction have been sufficiently 

 elucidated by the results described, and it is unnecessary to 

 put on record a large number of quantitative results which 

 merely confirm those already given. It should be noted, how- 

 ever, that the limit of the reaction has not been investigated : 

 on occasions we obtained increases of weight of 50 and even 

 80 p.ct., but at this degree of loading the natural lustre of 

 the fibre had given way to the dull and dusty look of a fibre 

 weighted to excess. 



We have now to consider the mechanism of the reaction 

 which has been in some measure elucidated by further experi- 

 ments. 



To explain it as the result of reduction of the ferric iron 



