It sometimes occurs that ether fails to take up all of the alkaloid present in the fluid treated with bicarbonate of soda. Under these circumstances the fluid should be mixed with caustic potassa, the mixture shaken, and the ether decanted; this operation being repeated several times, until the entire amount of the alkaloid is removed; the ethereal fluids are then united in a capsule, and allowed to spontaneously evaporate. The result of the evaporation may be solid; more frequently, however, a milky liquid remains which restores the blue color to reddened litmus paper; if so, the presence of a vegetable alkaloid is certain. In order to purify the residue, a few drops of water, slightly acidulated with sulphuric acid, are added to the capsule, and the latter turned, so as to bring the fluid in contact with the substance at all points; in this manner a colorless and limpid fluid is obtained, the fatty substances adhering to the dish. The liquid is decanted into a second capsule, the remaining residue washed with a little acidulated water, and the washings likewise added to the principal solution. The fluid is now evaporated either _in vacuo_, or over sulphuric acid, to about three-fourths of its original volume a concentrated solution of neutral carbonate of potassa added, and the mixture treated with absolute alcohol, which dissolves the liberated alkaloid, and separates it from the sulphate of potassa formed and the excess of carbonate of potassa. The alcoholic solution is decanted and allowed to evaporate _in vacuo_ or in the air: the alkaloid now crystallizes out in a state suitable for further examination. MODIFICATIONS TO STAS'S METHOD, PROPOSED BY OTTO. In Stas's method, the loss of morphine is possible, for, if ether is not added immediately after the addition of carbonate of soda, this alkaloid crystallizes and is then no longer soluble in that menstruum; and, if the ethereal solution is not quickly decanted, the portion dissolved will likewise separate out in small crystals. In both cases, morphine will remain in the aqueous solution from which the other alkaloids have been extracted by the ether. _M. Otto_ recommends the addition of chloride of ammonium and a little soda-lye, in order to dissolve the alkaloid. Upon allowing the solution so obtained to stand for some time exposed to the air, crystals of morphine are deposited. According to the same authority, it is advisable to omit the distinction drawn by Stas between volatile and fixed alkaloids, and submit both to the treatment recommended for those that are volatile.