fix a lower limit. Besides it is advisable, in order to obtain a complete picture of the nature of a petroleum, to divide the distillation into sub-fractions of from 25° to 25° or from 50° to 50° and determine these according to volume or weight. If, for instance, an oil contains 60 per cent. of constituents boiling between 3020 and 572° F., it is by no means unimportant whether this quantity consists of 30 per cent. from 302° to 392o, 20 per cent. from 392° to 482° and 10 per cent. from 482° to 572° F., or vice versa of 10 per cent. from 302° to 392° F., 20 per cent. 392° to 482°, and 30 per cent. 482° to 5720 F., an oil of the former constitution burning much better in a lamp than one of the latter. In short, a sufficient decision regarding the quality of an oil can only be obtained by dividing it into separate parts.

For the execution of such distillation the arrangement and method prescribed by Engler, which has been given in Chapter viii., p. 248, is recommended.

That the values obtained by this method do not entirely agree with the results of Beilstein's process is natural, but in executing the distillation according to Engler's method the results obtained by different experimenters are likely to agree better than with the use of special boiling tubes, particularly if the dimensions of the latter are not accurately determined and controlled.

The results given in the appended table were obtained by fractional distillation of different commercial varieties of Caucasian (Nobel's) and Pennsylvania oils according to Engler's method. The fractions were caught from 20° to 20°. It is still better, according to Engler's later proposition, to divide the fractions from 50° to 50° or from 25° to 25°, because the principal fraction from 150° to 300° C., (302°-572° F.), which may be designated standard illuminating oil, is more readily reduced to the latter sub-divisions.

For distillation 100 cubic centimeters of each of the petroleum varieties were used.

On an average there were obtained from the separate portions of the four Caucasian and the ten American kinds of petroleum :

However, as previously mentioned, it must not be supposed that Pennsylvania oils burn worse in proportion to their content of heavy oils than Caucasian oils, because experiments have shown that the heavy portions in American oil injure the illuminating power somewhat less than corresponding quantities of Caucasian heavy oils of the same boiling point, which very likely is connected with the considerably higher specific gravity of the latter.

Below are given the results of distillation made by Engler and C. Schestopal, of Alsace petroleum known in commerce as "brilliant petroleum," as compared with Kaiserœl (Imperial oil) and ordinary Pennsylvania petroleum.

The specific gravities of the oils are:

For distillation 100 cubic centimeters of each were used; the

experiments were made with Engler's apparatus.

From the table it will be seen that brilliant oil as well as Kaisercel are within the limits of a normal petroleum, while ordinary Pennsylvania petroleum is far outside of them and consequently is inferior.

The content of "normal burning oil" in thoroughly purified varieties of petroleum, for instance, in American astral oil, Kaiserol, etc., exceeds 80 and even 90 per cent., this being the principal reason for the safety as regards fire and the excellent illuminating value of these varieties of oil.

As regards the examination of lubricating oils and their fractions, the determination of their content of readily volatile oils is, as a rule, sufficient, a distilling test, similar to that for illuminating oil, being made according to Engler's method. Engler in his laboratory uses a small glass retort of the form and size shown in Fig. 125 and distils in it 100 cubic centimeters of oil so that about 21⁄2 cubic centimeters per minute pass over. With properly prepared lubricating oil by far the greater portion should boil above 590° F. Hence distillation is interrupted at this temperature and the distillate measured in the measuring apparatus, when distillation is continued.

If the asphaltum-like residuum, which cannot be distilled, is also to be determined, distillation is continued, after removing the thermometer, as long as anything passes over, and the resi

duum determined by weighing the retort. Properly distilled lubricating oils do not leave behind over three per cent. of coke, and lubricating oils, not distilled but simply chemically purified, not over 7 per cent., according to their derivation.

However, by direct distillation, the lubricating oils are decomposed, thinly-fluid (light) oils being formed, and conse

quently show more light constituents than they actually contain. Hence a true picture of the content of more or less volatile, and especially of heavy oils, is not obtained by this mode of distillation, and it is therefore recommended to execute distillation either in vacuum or, still better, with superheated steam. The latter method, which also serves for determining the value of petroleum residuum (the portions boiling above 572° F.) for working into lubricating oils, is carried on by means of