densation, and the burners should be surrounded by asbestos shields to keep off drafts of air. The water in G also is allowed to drip slowly, at the rate of about one drop per second, to insure plenty of liquid on the brass collars.

At the beginning of the distillation it is necessary to watch the still rather carefully and occasionally to relieve the steam pressure through the valve J, until the dephlegmator is evenly heated and the rate of reflow uniform in each section.

It is customary, in ordinary fractional distillation, to change the receiver at regular temperature intervals, which have been previously determined. In this method, however, the receiver is not changed according to prescribed temperature intervals, but according to the nature of the distillation. If the oil contains a large proportion of one substance, the temperature will remain nearly constant while that portion of the oil is being distilled. During this part of the distillation but few fractions need be taken. When, however, the oil is complex and the thermometer is rising rather rapidly, it is better to take more frequent and smaller fractions. Each time the cylinders are changed the temperature is recorded. Usually from ten to fifteen fractions are collected during one distillation. The cylinder with its contents is weighed and the weight of the dry cylinder subtracted. This gives the weight of the total distillate, which consists of both oil and water. Since by this method no greater accuracy than to 10 mg. is essential, the volume of the water in cubic centimeters as read on the cylinder can be assumed to be numerically equal to the weight of the water in grams. Subtracting this from the weight of the total distillate gives the weight of the turpentine in the fraction. By such a procedure results are obtained from which a table can be made giving for each fraction the steam temperature and the weight of oil it contains.

The odor and color of each fraction are observed and recorded at once. The specific gravity is determined, according to accepted methods, with a 10 c. c. pycnometer. The determinations are made at ordinary temperatures and corrected to give the value at 15° C. If the temperature at which the determinations are made is above 15°, then the correction factor 0.00083 is added for every degree of difference. If the temperature at which the determinations are made is below 15°, then 0.00083 is subtracted for every degree of difference. This correction factor was obtained experimentally by using a very pure wood turpentine. The index of refraction is read by means of an Abbe refractometer compensated to the D line of the spectrum, and the values corrected to 15° C. by means of factors. The factor 0.00049 is applied to all observed values below 1.4750, and is added for each additional degree of temperature above 15° C. and subtracted for each below. For oils having an index of refraction above 1.4750 the correction factor 0.00044 is similarly used. These values were

determined on two pure fractions of wood turpentine with indices of refraction above and below 1.4750, respectively.

There is now for each fraction the steam temperature, the weight, color, odor, specific gravity, and index of refraction.

RESULTS.

The results of the analyses of four different turpentines are given in the accompanying tables and curves. In the tables the data at the top show the properties of the original sample; the first column gives the serial number of the fraction; the second, the steam temperature after correction for barometric fluctuations; the third, the weight of the turpentine in each fraction. The fourth column is made by adding the weight of each turpentine fraction to the weight of all the preceding turpentine fractions, which gives the total weight of oil distilled up to each point. Thus in Table 1 up to and including fraction 3, the total weight of turpentine distilled was 267.66 grams. After all the fractions had been collected, the total weight distilled was 493.85 grams, and the undistilled residue amounted to the difference between the original weight taken, 500 grams, and the total weight distilled, 493.85 grams, or 6.15 grams. For purposes of reference and calculation it is more convenient to deal with percentage weight than with total weight, and hence the fifth column. is prepared from the fourth by finding the percentages which each total weight is of the original sample. The sixth and seventh columns give the corrected values for the specific gravity and index of refraction of each fraction. From these data three curves are drawn, showing the variation of steam temperature, specific gravity, and index of refraction, respectively, with the percentage weight of the distillate.

TURPENTINE No. 19-A GUM TURPENTINE.

Table 1 and figures 6, 7, and 8 give the data obtained from the analysis of a gum turpentine according to the methods which have been described.

1.

TABLE 1.-Turpentine No. 19-A gum turpentine.

[Specific gravity, 0.8693 at 15°; index of refraction, 1.4739 at 15°; weight of sample, 500 grams.]

2.

3.

4

5

6

7

8.

9.

10.

1. 4757

1. 4849

1.4740

The ten points indicated on the curve in figure 6 show the specific gravity of the ten fractions and the corresponding percentage weight

7

و

0.90

0.84 5 6
8
SPECIFIC GRAVITY-15°C.

FIG. 6.-Percentage weight-specific gravity curve for a gum turpentine.

of each fraction. The specific gravity of the first 924 per cent distilled is seen to be very uniform, as it varied only from 0.8660 to

FIG. 7.-Percentage weight-index of refraction curve for a gum turpentine.

0.8685, which is .0025, or 0.29 per cent. The 6 per cent between 92 and 99 per cent varied widely in specific gravity, while the last 1 per cent remained undistilled. Whether this 92 per cent was composed entirely of one substance, the curve does not decide, for it is possible that two substances of the same specific gravity were present. The term "uniformity of composition" in this paper refers only to physical properties.

On reference to the index of refraction-percentage weight curve, figure 7, it is seen that this physical property varied in about the same manner as the specific gravity. Over the first 92 per cent distilled, the index of refraction varied only between 1.4718 and 1.4757, 0.0039, or 0.27 per cent, and beyond 924 per cent the refractive index increased very rapidly.

The steam temperature curve as shown in figure 8 indicates about the same proportion of uniform composition, since the abrupt change in the direction of the curve takes place at about 924 per cent. Over this range the steam temperature varies from 94.9° to 95.7° C. About 92 per cent of this turpentine was therefore nearly uniform in composition, and the remainder consisted of a mixture of heavier oils. The original sample was colorless and had a sweet, pine-like odor. Each fraction was colorless and had a pine-like odor, and only the residue was yellow. Practically, Turpentine No. 19 would be considered a high-grade product.

TURPENTINE No. 14-A CRUDE STEAM TURPENTINE.

Table 2 gives the data from the analysis of an unrefined, steam distilled, wood turpentine. The specific gravity, index of refraction, and steam temperature curves are shown in figures 9, 10, and 11, respectively.

TABLE 2.-Turpentine No. 14-A crude steam turpentine.

[Specific gravity, 0.8834 at 15°; index of refraction, 1.4749 at 15°; weight taken, 500 grams.]