screw r, so arranged that by raising or lowering it, the water level in m is made to stand at zero.

Fifty cubic centimeters of water are put in the vessel A. B is nearly filled with the petroleum to be tested, the screw preplaced, and the whole placed in warm water until the temperature has become constant. The water level in m is placed at zero, and the plate C is moved until the opening of B comes over the opening o. The petroleum spreads upon the surface of the water in A, and by the expansion of its vapor causes the water to rise in the tube m, when its height is read. By a comparison of this number with the known expansion of the vapor of normal petroleum at a corresponding temperature, the combustibility of the oil is determined. For this purpose a table accompanies the apparatus, which gives the obtained vapor expansion of normal petroleum in m, for different temperatures sought.

This method depends upon the supposition that the numbers which express the expansion of the petroleum vapor run parallel with the temperature of the inflammability of all kinds of petroleum. It has been found, however, that this supposition is not correct for all cases, inasmuch as the presence of a small quantity of a very volatile hydrocarbon occasions, by increased temperature, a correspondingly greater pressure in the tube m, without it being sufficient to form an explosive mixture with air. Experiments made by Engler, on samples of petroleum, prepared by mixing in varying proportions oils of low and high boiling points, showed that a small percentage of a volatile constituent, notwithstanding the equal inflammability of the oils, occasions an uncorresponding increase of the vapor expansion. A similar conclusion may be drawn from Biel's investigations, recently published, as will be seen from the appended table showing the results obtained with various oils.

While, according to the above, vapor expansion may be considered a valuable criterion for judging in general the properties of an oil, it does not allow of drawing a direct and exact conclusion as to its inflammability. The researches of Arthur H. Elliot' also lead to the conclusion that the vapor expansions obtained in this apparatus are very irregular.

II. Apparatuses in which the inflammability is directly observed by igniting the vapors evolved at a certain temperature. Extraordinary difficulties have been found to attend the construction of exact apparatuses for heating petroleum and determining thereby the degree of temperature at which a small flame brought over the level of the oil may effect the ignition of the accumulated vapors. The execution of such experiments also requires great attention, because rapidity of heating, motion, quantity of oil, form of the oil-holder, size and intensity of the igniting flame, etc., are factors which may considerably impair the results of the determination of the flashing point. The results to be obtained are largely influenced by the construction of the petroleum-holder. In this respect a distinction is made between apparatuses with open or closed petroleum-holder.

1 Report on the Methods and Apparatus for Testing Inflammable Oils.

A. Apparatuses with open petroleum-holders.

Of the different petroleum testers of this kind, those of Tagliabue and Saybolt will here be described.

Tagliabue's open tester. This apparatus, Fig. 128, was employed in the official testing of petroleum in this country until 1879. It consists of a brass water-bath, A, upon the stand, B, heated by the lamp, C. D is the glass petroleum-holder, in

FIG. 128.

which is immersed the thermometer E. The bath is nearly filled with cold water, allowing for the displacement by the oilholder. D is filled to the top with the petroleum to be tested, care being taken not to wet the rim, the thermometer placed in position and the lamp lighted. The heating should be gradual, and if necessary, the lamp be occasionally removed. When the oil has reached the temperature at which the operator wishes to begin testing, a small flame, either from a wooden splinter or a gas-jet, is slowly and carefully passed over the petroleum,

about half an inch from the surface. If no flashing takes place, this is repeated as the temperature rises until the flashing point is reached. During the testing the apparatus should be protected from draughts of air.

To the group of open testers belong also the apparatuses of Ernecke-Hannemann and of Lenoir, as well as the Saybolt tester which was formerly used in England and the United States, it having been adopted in the latter country, in 1879, by the Produce Exchange of New York, in testing refined petroleum.

Saybolt electric tester. This apparatus resembles the open

tester of Tagliabue, differing chiefly in the use of the electric spark in place of the burning splinter. It is shown in Fig. 129, and consists of the copper water-bath F, containing the petro

leum-holder, which, with the other parts of the apparatus, are placed on the tray C, and, for transportation, can be enclosed in the box A. DD are the covers of two battery elements. H is a current-breaker, E an induction coil, and e e the conducting wires for producing the spark over the surface of the petroleum; a is the thermometer of the oil-holder, and a' that of the water-bath. The further details of the water-bath F may be seen from Fig. 130. b is the glass petroleum holder, c a

small ebonite beam secured by the brass pins dd; ff are two strips of brass ending towards the centre in platinum wires which pass through the ebonite beam, and below the latter are placed opposite to one another so that the spark can leap

across.

In using this apparatus the bath is filled with water, and heated to 100° F., after which the lamp is removed, the oil cup, filled to within an eighth of an inch of the top with the petroleum to be tested, is placed in the bath, and the thermometer is immersed in the oil until the bulb is just covered. When the temperature of the oil is raised to 90° F. a spark is produced by the key H, and after replacing the lamp the operation is repeated every two or three degrees until the flashing point is reached. The temperature of the oil when brought into the water bath should not be below 55° F. and not above