In each of two flasks marked at 100 and 200 cubic centimeters, respectively, are placed 65.52 grams of milk, four cubic centimeters of mercuric nitrate added, the volume completed to the mark and the contents of the flask well shaken. After filtering, the polarization is made in a 400 millimeter tube by means of the triple shadow polariscope described in =75=. From the reading thus obtained the volume of the precipitate and the degree of correction to be applied are calculated as in the subjoined example. The flasks should be filled at near the temperature at which the polarizations are made and the observation room must be kept at practically a constant temperature of 20° to avoid the complications which would be produced by changes in the gyrodynat of lactose and the value of the quartz plates and wedges of the apparatus by marked variations in temperature. _Example._—Weight of milk used in each case 65.52 grams. Polarimetric reading from the 100 cubic centimeter flask, 20°.84 ” ” ” ” 200 ” ” ” 10°.15 Then 10.15 × 2 = 20.30 20.84 - 20.30 = 0.54 0.54 × 2 = 1.08 20.84 - 1.08 = 19.76 19.76 ÷ 4 = 4.94, which is the corrected reading showing the percentage of lactose in the sample used. The volume of the precipitate is calculated as follows: 20.84 ÷ 4 = 5.21, the apparent percentage of lactose present. Then 5.21: 4.94 = 100: _x_. Whence _x_ = 94.82. From this number it is seen that the true volume of the milk solution polarized is 94.82 instead of 100 cubic centimeters, whence the volume occupied by the precipitate is 100 - 94.82 = 5.18 cubic centimeters. So little time is required to conduct the analysis by the double dilution method as to render it preferable in all cases where incontestable data are desired. Where arbitrary corrections are made the volume allowed for the precipitate may vary from two and a half cubic centimeters in milks poor in fat, to six for those with a high cream content. For milks of average composition sufficient accuracy is secured by making an arbitrary correction of five cubic centimeters for the volume of the precipitate. SEPARATION OF SUGARS BY CHEMICAL AND CHEMICAL-OPTICAL METHODS. =245. Conditions of Separation.=—In the foregoing paragraphs the optical methods for determining certain sugars have been described. Many cases arise, however, in which these processes are inapplicable or insufficient. In these instances, the analyst, as a rule, will be able to solve the problem presented by the purely chemical methods which have been previously described, or by a combination of the chemical and optical processes. Not only have the different sugars distinctive relations to polarized light, but also they are oxidized by varying quantities of metallic salts and these differences are sufficiently pronounced to secure in nearly every instance, no matter how complex, data of a high degree of accuracy.