Schulze.—This process gave a yield of 47.6 p.ct. cellulose from pine wood. The celluloses themselves, treated by the process, showed losses of 1-3 p.ct. in weight, much less therefore than in the preceding case.

Hönig's method of heating with glycerin to 210° was found to yield products very far removed from 'cellulose.' The process may have a certain value in estimations of 'crude fibre,' but is dismissed from further consideration in relation to cellulose.

Lange.—The purpose of the investigation was to test the validity of the statement that the celluloses are not attacked by alkaline hydrates at 180°. Experiments with pine wood yielded a series of percentages for cellulose varying from 36 to 41; the 'purified wood' gave also variable numbers, 44 to 49 per cent. It was found possible to limit these variations by altering the conditions in the later stages of isolating the product; but further experiments on the celluloses themselves previously isolated by other processes showed that they were profoundly and variably attacked by the 'Lange' treatment, wood cellulose losing 50 per cent. of its weight, and filter paper (cellulose) losing 15 per cent. Further, a specimen of jute yielded 58 per cent. of cellulose by this method instead of the normal 78 per cent. It was also found that the celluloses isolated by the process, when subjected to a second treatment, underwent a further large conversion into soluble derivatives, and in a third treatment further losses of 5-10 per cent were obtained. The authors attach value, notwithstanding, to the process which they state to yield an 'approximately pure cellulose,' and they describe a modified method embodying the improvements in detail resulting from their investigation.

Gabriel's method of heating with a glycerin solution of alkaline hydrate is a combination of 'Hönig' and 'Lange.' An extended investigation showed as in the case of the latter that the celluloses themselves are more or less profoundly attacked by the treatment—further that the celluloses isolated from lignocelluloses and other complex raw materials are much 'less pure' than those obtained by the Lange process. Thus, notably in regard to furfural yielding constituents, the latter yield 1-2 p.ct. furfural, whereas specimens of 'jute cellulose' obtained by the Gabriel process were found to yield 9 to 13 p.ct. furfural.

Cross and Bevan.—Chlorination process yielded in the hands of the authors results confirming the figures given in 'Cellulose' for yield of cellulose. Investigation of the products for yield of furfural, gave 9 p.ct. of this aldehyde showing the presence of celluloses, other than the normal type.

Conclusions.—The subjoined table gives the mean numerical results for yield of end-product or 'cellulose' by the various methods. In the case of the 'celluloses' the results are those of the further action of the several processes on the end-product of a previous process.

Methods
F. Schulze Weende Lange Gabriel Cross and Bevan
Wood cellulose 98.51 91.52 48.22 55.93 —
Filter paper cellulose 99.62 95.63 78.17 79.77 —
Swedish filter paper 96.58 — 84.76 — —
Ordinary filter paper 98.17 93.39 86.58 — —
Cotton ('wool') 98.38 89.98 63.96 67.88 —
Jute — — 57.93 71.64 75.27
Purified wood — —{49.27
{46.56		 — —
Raw wood 47.60 —{40.82
{38.87		 — —

The final conclusion drawn from the results is that none of the processes fulfil the requirements of an ideal method. Those which may be carried out in a reasonably short time are deficient in two directions: (1) they yield a 'cellulose' containing more or less oxycellulose; (2) the celluloses themselves are attacked under the conditions of treatment, and the end product or cellulose merely represents a particular and at the same time variable equilibrium, as between the resistance of the cellulose and the attack of the reagents employed; this attack being by no means confined to the non-cellulose constituents. Schulze's method appears to give the nearest approximation to the 'actual cellulose' of the raw material.

(p. 8) SOLUTIONS OF CELLULOSE—(1) ZINC CHLORIDE.—To prepare a homogeneous solution of cellulose by means of the neutral chloride, a prolonged digestion at or about 100° with the concentrated reagent is required. The dissolution of the cellulose is not a simple phenomenon, but is attended with hydrolysis and a certain degree of condensation. The latter result is evidenced by the formation of furfural, the former by the presence of soluble carbohydrates in the solution obtained by diluting the original solution and filtering from the reprecipitated cellulose. The authors have observed that in carefully conducted experiments cotton cellulose may be dissolved in the reagent, and reprecipitated with a loss of only 1 p.ct. in weight. This, however, is a 'net' result, and leaves undetermined the degree of hydration of the recovered cellulose as of hydrolysis of the original to groups of lower molecular weights. Bronnert finds that a previous hydration of the cellulose—e.g. by the process of alkaline mercerisation and removal of the alkali by washing—enables the zinc chloride to effect its dissolution by digestion in the cold. (U.S. patent, 646,799/1900. See also p. 59.)