the dry paper filter then introduced in the ordinary manner, moistened, and freed from all adhering air-bubbles by pressure with the finger. A filter so arranged and in perfect contact with the glass, when filled with a liquid will support the pressure of an entire atmosphere without the least danger of breaking; and the interspace between the folds of the platinum-foil is perfectly sufficient to allow of the passage of a continuous stream of water.

In order to be able to produce the additional pressure of an atmosphere, the filtered liquid is received in a strong glass flask instead of in beakers. This flask is closed by means

These flasks must be somewhat thicker than those ordinarily used, in order to prevent the possibility of their giving way under the atmospheric pressure.

of a doubly perforated caoutchouc cork, through one of the holes of which the neck of the glass funnel is passed to a depth of from 5 to 8 centimeters (fig. 3); through the other is fitted a narrow tube open at both ends, the lower end of which is brought exactly to the level of the lower surface of the cork, to the other is adapted the caoutchouc tube connected with the apparatus (fig. 4) destined to produce the requisite difference in pressure: this apparatus will be described immediately. The flasks are placed in a metallic or porcelain vessel (fig. 3), in the conical contraction of which several strips of cloth are fastened. This method of supporting the flask has the advantage that, in one and the same vessel, flasks varying in size from 0.5 to 2.5 liters stand equally well, and that, by simply laying a cloth over the mouth of the vessel, the consequences of an explosion (which through inexperience or carelessness is possible) are rendered harmless.

It is impossible to employ any of the air-pumps at present in use to create the difference in pressure, since the filtrate not unfrequently contains chlorine, sulphurous acid, hydric sulphid, and other substances which would act injuriously upon the metallic portions of these instruments. I therefore employ a water air-pump constructed on the principle of Sprengel's mercury-pump, and which appears to me preferable to all other forms of air-pump for chemical purposes, since it effects a rarefaction to within 6 or 12 millimeters pressure of mercury. Fig. 4 shows the arrangement of this pump. On opening the pinchcock a, water flows from the tube into the enlarged glass vessel b, and thence down the leaden pipe c. This pipe has a diameter of about 8 millims., and extends downward to a depth of 30 or 40 feet, and ends in a sewer or other arrangement serving to convey the water away. The lower end of the tube d possesses a narrow opening; it is hermetically sealed into the wider tube b, and reaches nearly to the bottom of the latter. A manometer is attached to the upper continuation of this tube d by means of a side tube at d; at d2 is attached a strong thick caoutchouc tube possessing an internal diameter of 5 millims, and an external diameter of 12 millims.; this leads to the flask which is to be rendered vacuous, and is connected with it by means of the short narrowed tube k. Between the air-pump and the flask is placed the small thick glass vessel f, in which, when one washes with hot water, the steam which may be carried over is condensed. All the caoutchouc joinings are made with very thick tubing, the internal diameter of which amounts to about 5 millims., the external diameter to about 17 millims. The entire arrangement is screwed down upon a board fastened to the wall, in

such a manner that each separate piece of the apparatus is held by a single fastening only, in order to prevent the tubes being strained and broken by the possible warping of the board. On releasing the pinchcock a, water flows from the conduit 7 down the tube c to a depth of more than 30 feet, carrying with it the air which it sucks through the small opening of the tube d in the form of a continuous stream of bubbles. No advantage is gained by increasing the rapidity of the flow, since the friction exerted by the water upon the sides of the leaden pipe acts directly as a counter pressure, and a comparatively small increase in the rapidity of the flow is accompanied by a great increase in the amount of this friction. Accordingly at g is a second pinchcock, by which the stream can be once for all so regulated that, on completely opening the cock a, the friction, on account of the diminished rate of flow, is rendered sufficiently small to allow of the maximum degree of rarefaction. Such an apparatus, when properly regulated once for all by means of the cock g, exhausts in a comparatively short time the largest vessels to within a pressure of mercury equal to the tension of aqueous vapor at the temperature possessed by the stream. The tension exerted by the water-stream in my laboratory, in which six of these pumps are used, amounts to about 7 millims. in winter and 10 millims. in summer. The filtration is made in the following manner: The flask standing in the metallic or porcelain vessel (fig. 3) is connected by means of the slightly drawn-out tube with the caoutchouc tube h attached to the pump, the cock a having been previously opened and the properly fitted moistened filter filled with the liquid to be filtered. As usual, the clear supernatant fluid is first poured upon the filter; in a moment or two the filtrate runs through in a continuous stream, often so rapidly that one must hasten to keep up the supply of liquid, since it is advisable to maintain the filter as full as possible. After the precipitate has been entirely transferred, the filtrate passes through drop by drop, and the manometer not unfrequently now shows a pressure of an extra atmosphere. The filter may be filled (in fact this is to be recommended) with the precipitate to within a millimeter of its edge, since the precipitate, in consequence of the high pressure to which it is subjected, becomes squeezed into a thin layer broken up by innumerable fissures. As soon as the liquid has passed through and the first traces of this breaking up become evident, the precipitate will be found to have been so firmly pressed upon the

The time required to obtain the above degree of exhaustion in a flask of from 1 to 3 liters capacity ranges from six to ten minutes; the quantity of water necessary amounts to about 40 or 50 liters.

paper, that on cautiously pouring water over it it remains completely undisturbed. The washing is effected by carefully pouring water down the side of the funnel to within a centimeter above the rim of the filter: the washing flask for this purpose is not applicable; the water must be poured from an open vessel. After the filter has in this manner been replenished four times with water and allowed to drain for a few minutes, it will be found to be already so far dried, in consequence of the high pressure to which it has been subjected, that without any further desiccation it may be withdrawn, together with the precipitate, from the funnel, and immediately ignited, with the precautions to be presently given, in the crucible.

If the porosity of a paper filter containing a precipitate were as unalterable as that of a pumice-stone filter, the experiments above described would show that the times required for filtration, according to the old method on the one hand, and the new one on the other, would be inversely proportional to the difference in pressure in each case; that is, by using the pump under the full pressure of about 740 millims., the time needed to wash a precipitate, occupying by the old process an hour, would at the utmost not amount to more than 30 seconds. In using these pumice filters (about which I will speak presently) to drain crystals from adhering mother li quors, or, say, to wash crystals of chromic acid by means of concentrated sulphuric acid and fuming nitric acid, the time occupied in the filtration is scarcely longer than that needed to pour a liquid slowly from one vessel to another. In filtering by means of paper, the precipitate gradually closes up the pores of the filter, and accordingly such an extraordinary acceleration as this can no longer be expected. But the following examples will show the saving of time and labor the method effects, even under all unfavorable conditions. For these experiments I have purposely chosen the hydrated chromium sesquioxyd, since it is one of the most difficult of precipitates to wash thoroughly. A solution of chromium chlorid was prepared by acting with fuming hydrochloric acid upon potassium dichromate; and by means of a measuring-vessel, which allowed the amount of chromium to be estimated to within 0.0001 grm., successive portions of the liquid were withdrawn, and the chromium oxyd contained in them precipitated with the usual precautions by ammonia. The volume of liquid, the quantity of ammonia employed, the time occupied in boiling and in permitting the precipitate to settle, the angle of inclination possessed by the funnel, and the size of the filter were the same in all the experiments. All the precipitates

AM. JOUR. SCI.-SECOND SERIES, VOL. XLVII, No. 141.—MAY, 1869.

were washed with hot water, and, after burning the filter, ignited over the blowpipe for a few minutes; in weighing, the platinum crucible was tared by one of about equal weight, and the position of equilibrium of the beam determined by

vibrations.

v

I first attempted to filter one of the precipitates in the ordinary way. amounted to 2; and consequently, from the table, 8.4 fresh additions of water were required in order to wash the precipitate to the T part. The times required were as follows:

In transferring the precipitate from the beaker
and allowing it to drain

For the first addition of water to run through,

}

40'

48

70

80

238

Total length of time

At this point the experiment was discontinued, as the filtrate became turbid. A second experiment failed from the

same cause.

Accordingly I attempted to wash the precipitate by decantation. The volume of the precipitate amounted to about 30 cub. centims.; the quantity of water required to fill the beaker was seven times the volume of the precipitate; hence was 7,

V

v

and the requisite number of decantations to reduce the amount of impurity to the 3 part was 52. The times observed were as follows:

II.

For the first decantation to run through the filter... 15'

Experiment repeated. Number of decantations 7. Other cir

cumstances the same as in the foregoing determination.

Time required in washing

Weight of the precipitate

Volume of wash-water