3 grammes of the defatted and atmospherically dried substance are boiled for ½ hour with 200 ccm of a 1·24 percent solution of sulphuric acid. It is allowed to settle, then decanted, and the residue boiled twice with the same volume of water. The decanted liquids are allowed to settle in cylinders and the sediment added to the rest of the substance, which is then boiled half an hour with 200 ccm of a 1·25 percent solution of caustic potash, filtered through a weighed filter and the residue twice boiled with 200 ccm of water. The cellulose-like substance collected on the filter is washed first with hot water, then with cold, afterwards with alcohol, and finally with ether. After being dried and weighed, it is incinerated and the necessary corrections made for ash. The process worked out by Henneberg is the one usually adopted for the determination of crude fibre in vegetable matter. Recently H. Suringer and B. Tollens[186] and more particularly König[187] have pointed out that in Weender’s process the so-called pentosan (sugar derivative) of the composition C_{5} H_{10} O_{5}, which comprises a not inconsiderable portion of crude fibre, would undergo a disproportionate alteration, so that the analytical results thus obtained would not represent the amount of cellulose correctly. The crude fibre must therefore be treated in such a manner as to eliminate the pentosan. König attains that result by treating 3 grammes of the defatted substance with 200 ccm of glycerine (1·23 sp. gr.) containing per litre 20 grammes of concentrated sulphuric acid, under a pressure of three atmospheres, for one hour. It is then filtered through an asbestos filter whilst hot, and after being successively washed with hot water, alcohol and ether, it is weighed, incinerated and the ash weighed. The difference between the two weighings expresses the amount of ash-free crude fibre. Filsinger has determined by König’s method the amount of crude fibre in a series of different varieties of bean, the results of which have already been given on page 72. Which process is the better has yet to be established, and in issuing results as data the method employed has always to be indicated owing to the many variations which arise.