LHELM HOFFMEISTER (Landw. Versuchs-Stat., 1897, 48, 401-411).

To separate the hemicelluloses, celluloses, and the constituents of
lignin without essential change, the substance, after being freed from
fat, is extracted with dilute hydrochloric acid and ammonia, and the
residue frequently agitated for a day or two with 5-6 p.ct. caustic soda
solution. It is then diluted, the extract poured off, neutralised with
hydrochloric acid, treated with sufficient alcohol, and the
hemicellulose filtered, dried, and weighed. The residue from the soda
extract is washed on a filter with hot water, and extracted with
Schweizer's reagent.

When the final residue (lignin) is subjected to prolonged extraction
with boiling dilute ammonia (a suitable apparatus is described, with
sketch) until the ammonia is no longer coloured, a residue is obtained
which mostly dissolves in Schweizer's reagent, and on repeating the
process the residue is found to consist largely of mineral matter. The
dissolved cellulose-like substances often contain considerable amounts
of pentosanes.

According to the nature of the substance, the extraction with ammonia
may take weeks, or months, or even longer; the ammonia extracts of hard
woods (as lignum vitae) and of cork are dark brown, and give an odour of
vanilla when evaporated down. The residues, which are insoluble in
water, but redissolve in ammonia, have the properties of humic acids.
Other vegetable substances, when extracted, yielded, besides humic
acids, a compound, C_{6}H_{7}O_{2}, soluble in alcohol and chloroform,
but insoluble in water, ether, and benzene; preparations from different
sources melted between 200 deg. and 210 deg..

FOOTNOTES:

[4] The original paper is reproduced with slight alterations.

[5] This purple colour would appear to be due to a highly dissociable
compound of omega-brommethylfurfural with hydrogen bromide. The aldehyde
gives yellow or colourless solutions in various solvents, which are
turned purple by a sufficient excess of hydrogen bromide. Dilution, or
addition of water, at once discharges the colour.

[6] Other forms of cellulose were also examined--for example, pinewood
cellulose--and the substances separated from solution as thiocarbonate
(powder and film). All of these gave good yields of
omega-brommethylfurfural.

[7] The change is empirically represented as

C_{6}H_{12}O_{6} + HBr - 4H_{2}O = C_{6}H_{5}O_{2}Br.

SECTION IV. CELLULOSE GROUP, INCLUDING