Here again much has been written of late[194] concerning the two former methods, their liabilities to error and the avoidance of these, yet without bringing to light anything which calls for a specially detailed treatment in this book. 11. ~Determination of Albuminates.~ The determination of albumin is frequently required in the analysis of cacao powder and is necessary to the ascertainment of its nutritive value. The determination of nitrogen is determined by mixing 0·5 grammes of finely powdered bean with soda lime and burning the mixture in a tube. (This determination of nitrogen is a necessary part of the process.) Thus ammonia is formed, which is passed through a known quantity of sulphuric acid. When the combustion is finished, the acid solution is titrated with a standard solution of barium hydroxide, and from the quantity consumed the percentage of nitrogen is calculated. But as the diureides also contain nitrogen (31·1 % of the theobromine and caffeine present) the nitrogen corresponding to this amount must be deducted from the total quantity of nitrogen yielded by combustion and the remainder multiplied by 6·25 will indicate the amount of albumen present as a constituent. Another and better method of determining the nitrogen is by Kjeldahl’s[195] process. It has been frequently subjected to modifications, but was originally carried out as follows. 0·25 grammes of the nitrogenous substance (cacao preparation) is heated on the sand bath together with 20 ccm of concentrated sulphuric acid and a little quicksilver, till the solution becomes colourless or only of a very pale yellow. After diluting with about 200 ccm of water, it is made alkaline by the addition of soda lye (which must of course be entirely free from nitrogen, the same remark applying to the sulphuric acid used) and, potassium or sodium sulphide being added, it is then distilled, and the ammonia given off collected and determined as above described. As this method also determines the total amount of nitrogen, an allowance must be made for the nitrogen in the theobromine and caffeine before multiplying the result by 6·25. This modification is still to be recommended as the best and most reliable. In rare cases an excessive amount of albumen may be due to the admixture of earth-nut cake or gelatine. As to the detection of the latter adulteration, see page 254. Bileryst[196] says that earth-nut cake can be recognised by its high percentage of albumen content, amounting to between 45 and 47 percent.