This portion of the investigation is an extension of that of Suringar and Tollens, these latter confining themselves to celluloses of the 'normal' groups, i.e. textile and paper-making celluloses. The present communication is a study of the tissue and cell-wall constituents of the following types:—
1. Green plants of false oat grass (Arrhenatherium, E.).
2. Green plants of lucerne (Medicago sativa).
3. Leaves of the ash (Fraxinus).
4. Leaves of the walnut (Juglans).
5. Roots of the purple melic grass (Molinia cærulea).
6. Roots of dandelion (Taraxacum officinale).
7. Roots of comfrey.
8. Coffee berries.
9. Wheat bran.
These raw materials were treated for the quantitative estimation of cellulose by the method of Lange (b), Hoffmeister (c), and Schulze (d), and the numbers obtained are referred for comparison to the corresponding yields of 'crude fibre' (Rohfaser) by the standard method (a).
As a first result the author dismisses Lange's method as hopeless: the results in successive determinations on the same materials showing variations up to 60 p.ct. The results by c and d are satisfactorily concordant: the yields of cellulose are higher than of 'crude fibre.' This is obviously due to the conservation of 'hemicellulose' products, which are hydrolysed and dissolved in the treatments for 'crude fibre' estimation. A modified method was next investigated, in which the process of digestion with acid chloroxy- compounds (c and d) was preceded by a treatment with boiling dilute acid. The yields of cellulose by this method (e) are more uniform, and show less divergence from the numbers for 'crude fibre.'
The author's numerical results are given in a series of tables which include determinations of proteids and ash constituents, and the corresponding deductions from the crude weight in calculating to 'pure cellulose.' The subjoined extract will illustrate these main lines of investigation.
| Raw Material | Crude Fibre | Pure Cellulose | |
| Weende Method. (a) | Hoffmeister Method. (c) | Hoffmeister, modified by Author. (e) | |
| Oat grass | 30.35 | 34.9 | 31.5 |
| Lucerne | 25.25 | 28.7 | 20.5 |
| Leaves of ash | 13.05 | 15.4 | 13.8 |
| Roots of melic | 21.60 | 29.1 | 21.4 |
| Coffee beans | 18.30 | 35.1 | 23.3 |
| Bran | 8.2 | 19.3 | 9.3 |
The final conclusion drawn from these results is that the method of Hoffmeister yields a product containing variable proportions of hemicelluloses. These are eliminated by boiling with a dilute acid (1.25 p.ct. H2SO4), which treatment may be carried out on the raw material—i.e. before exposure to the acid chlorate, or on the crude cellulose as ordinarily isolated.
Determination of Tissue-constituents.—By the regulated action of certain solvents applied in succession, it appears that such constituents of the plant-complex can be removed as have no organic connection with the cellular skeleton: the residue from such treatments, conversely, fairly represents the true tissue-constituents. The author employs the method of digestion with cold dilute alkaline solutions (0.15 to 0.5 p.ct. NaOH), followed by exhaustive washing with cold and hot water, afterwards with cold and hot alcohol, and finally with ether.
The residue is dried and weighed as crude product. When necessary, the proportions of ash and proteid constituents are determined and deducted from the 'crude product' which, thus corrected, may be taken as representing the 'carbohydrate' tissue constituents.
Determination of Hemicelluloses.—By the process of boiling with dilute acids (1.25 p.ct. H2SO4) the hemicelluloses are attacked—i.e. hydrolysed and dissolved. The action of the acid though selective is, of course, not exclusively confined to these colloidal carbohydrates. The proteid and mineral constituents are attacked more or less, and the celluloses themselves are not entirely resistant to the action. The loss due to the latter may be neglected, but in calculating the hemicellulose constants from the gross loss the proteids and mineral constituents require to be taken into account in the usual way.