Since the variation in angular rotation in the mixture at different temperatures is due almost wholly to the change in this property of the levulose it follows that the variation for each degree of temperature and each per cent of levulose can be calculated. Careful experiments have shown that the variation in the rotatory power of levulose between 0° and 88° is represented by a straight line. For 13.71 grams per 100 cubic centimeters the variation for each degree of temperature is equal to 43.37 ÷ 88 = 0.49 divisions on the cane sugar scale, or 15.15 ÷ 88 = 0°.1722 angular measure. If 13.71 grams of levulose in 100 cubic centimeters produce the deviations mentioned for each degree of temperature, one gram would give the deviation obtained by the following calculations:
For the cane sugar scale 0.49 ÷ 13.71 = 0°.0357 and for angular rotation 0.1722 ÷ 13.71 = 0.01256.
The above data afford a simple formula for calculating the percentage of levulose present from the variation observed in polarizing a solution containing levulose, provided that the quantity of levulose present is approximately fourteen grams per 100 cubic centimeters.
Example.—Suppose in a given case the difference of reading between a solution containing an unknown quantity of levulose at 0° and 88° is equal to thirty divisions of the cane sugar scale. What weight of levulose is present? We have already seen that one gram in 100 cubic centimeters produces a variation of 0.0357 division for 1°. For 88° this would amount to 3.1416 divisions. The total weight of levulose present is therefore 30 ÷ 3.1416 = 9.549 grams. In the case given 26.048 grams of honey were taken for the examination. The percentage of levulose was therefore 9.549 × 100 ÷ 26.048 = 36.66 per cent.
If it be inconvenient to determine the polarimetric observations at temperatures so widely separated as 0° and 88° the interval may be made less. In the above case if the readings had been made at 20° and 70° the total variation would have been only ⁵⁰/⁸⁸ of the one given, viz., 17.05 divisions of the cane sugar scale. The calculation would then have proceeded as follows:
0.0357 × 50 = 1.785.
Then, 17.05 ÷ 1.785 = 9.552 grams of levulose, from which the actual percentage of levulose can be calculated as above.
With honeys the operation is to be conducted as follows:
Since honeys contain approximately twenty per cent of water and in the dry substance have approximately forty-five per cent of levulose, about 38.50 grams of the honey should be taken to get approximately 13.8 grams of levulose.
In the actual determination the calculations may be based on the factors above noted, but without respect to the degree of concentration. If half the quantity of dextrose noted be present its specific rotatory power is only reduced to about 52°.75, and this will make but little difference in the results. In the case of honey 13.024 grams of the sample are conveniently used in the examination, half the normal weight for the ventzke sugar scale. The error, however, due to difference in concentration is quite compensated for by the ease of clarification and manipulation. Alumina cream alone is used in the clarification, thus avoiding the danger of heating the solution to a high temperature in the presence of an excess of lead acetate.