3 1 4 Cellulose 



levulose. The reaction is therefore a main reaction, and shows 

 that cellulose under these conditions breaks down, at least in large 

 parf, to ketohexose units. By these investigations therefore the 

 polyaldose view of the constitution of cellulose is directly called in 

 question. We have found on other grounds that a ketonic 

 formula is to be preferred (ist ed. p. 77), the fifth O atom having 

 ketonic rather than aldehydic function. This is consistent either 

 with an open chain or closed ring formula for the assumed C 8 

 unit. There are general grounds of preference for the latter. But 

 this is a matter of speculation and hypothesis. 



A point to be noted in connection with Fenton's researches is 

 that the normal celluloses (of the cotton group) give higher yields 

 of the furfural derivative than the cereal celluloses (group C), 

 which on the other hand are characterised by high yields of furfural 

 under the action of aqueous condensing acids. This decomposi- 

 tion is referred by many chemists to the presence in the cereal 

 celluloses of a pentose anhydride. In view of these later facts the 

 explanation, which is on other grounds doubtful, becomes unneces- 

 sary. It is clear that the transition from the normal chain to the 

 C 4 O ring is equally characteristic of hexose as of pentose units, 

 and the assumption that 'furfural yielding ' is equivalent to * pentose ' 

 or * pentosane ' carbohydrate, falls away. 



A second point to be noted arises in connection with the ex- 

 haustive study of the action of ethereal hydrobromic acid on the 

 celluloses. In a succession of treatments with the acid, diminish- 

 ing yields of the brom-methyl furfural are obtained, and the final 

 residue has the composition and character of the humic or ulmic 

 series of complex derivatives described on p. 240. M. Gostling. 

 Proc. Chem. Soc. 18, 250). 



It is probable from later investigations of our own that pyrone 

 groups are formed as an alternative or complementary course of 

 condensation of he carbohydrates, and are represented in these 

 complex products. 



On the broad and general question of the actual constitution of 

 cellulose there is as yet but little positive evidence. It is a ques- 

 tion of proximate arrangement and configuration of ultimate con- 

 stituent groups which we assume to be of C rt dimensions, and to be 

 represented by the ordinary molecular formulas. But we have no 

 conception of a molecule of cellulose, and no data as to its dimen- 

 sions. We have positive evidence as to a reacting unit, but of 

 variable dimensions, and the more definite synthetical reactions of 



