of active substance. b. Filtration, lasting for three minutes, of 50 cubic centimetres of an aqueous solution of nitrate of silver raised the proportion of the solid substance from 9-708 to 9.713 per cent. Temperature, 24.5°. In this case the addition amounts to 0.005 per cent., thus agreeing with the result of the preceding experiment. Alcoholic Solutions.-a. 31.007 grammes of alcohol, of 94 per cent. by weight, took four minutes to filter, and lost 0.067 gramme by evaporation. Temperature, 18° Cent. 71-494 grammes of the same alcohol filtered in ten minutes, and lost 0·114 gramme. Temperature, 19° Cent. Had these solutions contained, to begin with, 10 per cent. of active substance, the process of filtration would have raised it to 10.022 per cent. in the first experiment, and to 10.016 per cent. in the second. b. 50 cubic centimetres of a solution of nitrate of silver in alcohol of 78 per cent. by weight took ten minutes to filter (temperature, 23° Cent.), and the proportion of active substance rose in one experiment from 9.686 to 9.714, and in another to 9.736, or by amounts ranging from 0.028 to 0.050 per cent. Assuming, therefore, that evaporation is independent of the concentration of the solutions, and, further, is proportional to the amount of filtrate (both of which postulates are only approximately true), the result of the foregoing experiments may be taken as proving that for every 10 per cent. of active substance originally present, the proportion is raised by filtration-in aqueous solutions by 0005, and in alcoholic solutions by from 0.02 to 0.05 per cent. Thus it will be seen that in the case of concentrated solutions this increase of percentage may become very considerable, and where alcohol is employed as the inactive solvent, may alter the first decimal by several units. Filtration is therefore to be avoided as far as practicable. The degree to which errors of this kind affect the calculation of specific rotation is stated in § 75. $ 69. Reduction of Weighings to Weight in Vacuo.-If it be desired to determine with great exactness the percentage composition of the solutions, the results of the several weighings should be reduced to their values in vacuo. When pains have been taken to weigh accurately to milligrammes, which should, as a rule, be done, it costs but little extra trouble to apply the trifling correction requisite, which is the more desirable, as neglect of it may affect the value of the percentage to the second place of decimals. As a rule, the error arising from non-reduction of the weights will be greater in proportion to the difference in density between the active substance itself and the solution of it employed. The effect on the specific rotation value will be greater the more concentrated the solution is, and the smaller the angle of rotation. : The following simple method will suffice for the reduction:- d its specific gravity, then the weight y, by which the substance, weighed with brass weights, appears too light, owing to the pressure of the atmosphere, is given by the formula1 The number 0.0012 is the mean density of atmospheric air, and 0.12 is obtained by dividing unity by the specific gravity of brass, which latter may be taken as 84. The weight in vacuo P of the substance will then be 2 P = p + y These coefficients are amply sufficient for the reduction of all weighings that occur in determining the specific rotation of active substances, and it is unnecessary to allow for changes of density in the air, so that we need not take observations of temperature and pressure at the time of weighing.3 To facilitate calculation the following table has been prepared, giving the values of the factor 0·0012 (−0·12) for solutions with d specific gravities ranging from 074 to 30. Putting R for this value, as in the table, the reduced weight becomes Pp+p R. For the rationale of this formula see Kohlrausch, Leitfaden der praktischen Physik, 3 Aufl. S. 29. [Translated into English from the second German edition, under the title of An Introduction to Physical Measurements, by Messrs. Waller and Proctor. Churchill, London, 1873. 8vo, 12s.-D.C.R.] 2 If the density of the substance exceeded 8-4, which, however, is never the case with the substances with which we have to deal, P = p - y. 3 Moreover, the circumstance that the smaller weights are of platinum instead of brass has no appreciable effect. In the calculation it will be sufficient to take the weights p to decigrammes and the specific gravity to two places of decimals. The following example of the preparation of a solution of tartrate of ethyl in wood-spirit, will show the mode of the reduction : p = 10.898 grammes. I. Tartrate of ethyl weighed in air 10.898 or (substantially) 10.9 x 0.00086 = 0.009 Weight in vacuo 10-907 grammes. II. Weight in air of the solution in wood-spirit, p=27.269 grammes. Specific gravity of solution, d = 0.94. Value of R for 0.94 from table = 0.00113. 27.269 or (substantially) 27.3 × 0·00113 = 0.031 29 Weight in vacuo = 27.300 grammes. The percentage composition of the solution therefore stands as In the process of preparing solutions there are thus two separate weighings requiring to be reduced to weight in vacuo :-(1) That of the active substance, for which a knowledge of its specific weight is necessary; and (2) that of the prepared solution, the density of which must be known at any rate for calculating the specific rotation. In case the density of an active substance is entirely unknown, which may occur with solids, the difficulty may be got over by weighing first the solvent, whose specific gravity is known, in the flask, and then adding the active substance. The table below gives the specific gravities of a number of optically active solids and liquids, as also of several substances suitable as solvents. The specific gravities of solutions employed in polariscopic experiments seldom appear outside the limits 0.8 to 1.4. § 70. The only method which affords the requisite degree of exactness is that of weighing a determinate volume. For this purpose we may use narrow-necked pycnometers of 10 to 20 cubic centimetres capacity, which can be filled or emptied by means of a pipette with capillary stem and india-rubber ball (Fig. 50.) The washing and drying of the apparatus is quickest done by rinsing with alcohol, followed by a little anhydrous ether. To bring the level of the liquid to the mark a roll of filter-paper or cigarette paper may be used. If the neck of the pycnometer have a width of about 1 millimetre, differences of from 0.5 to 2 milligrammes may be observed in the weight of the flask in successive adjustments of level, A greater degree of exactness is attainable with Sprengel's pycnometer1 (Fig. 51). This consists of a U-shaped tube of thin glass, 1 Sprengel: Pogg. Ann. 150, 459. |