This analysis and ratio does not seem to agree with any of the hydrous silicates in Dana's Mineralogy, and the species is therefore probably a new one. It occurs very sparingly implanted upon Lesleyite, and like it, is most probably the product of the alteration of the corundum.

I am indebted to Dr. Lea for specimens of all of these minerals and for the use of his large specimens for comparison with

each other.

Cambridge, Mass., Jan. 8, 1869.

ART. XXVIII.-On the Washing of Precipitates; by
R. BUNSEN.

A PRECIPITATE is washed either by filtration or by decantation in the former case the portion of liquid not mechanically retained is allowed to drain from the precipitate; in the latter it is separated by simply pouring it away, the foreign substances contained in the precipitate being then removed by the repeated addition of some washing-fluid, in each successive portion of which the precipitate is, as far as possible, uniformly suspended, this process being continued until the amount of impurity becomes so minute that its presence may be entirely disregarded.

Supposing to represent the volume of the moist precipitate remaining at the bottom of the vessel after decantation, or upon the filtrate after filtration, V the volume of wash-water employed at each successive decantation, n the number of decantations, and the fraction expressing the proportion of the original amount of impurity still remaining in the precipitate after n decantations, then

a

Calling W the total volume of wash-water resulting from n decantations, then

therefore

nV=W; .

If we differentiate W with respect to n and make the differ

* From Lond., Edin. and Dublin Phil. Mag., Jan. 1869. Translated from the Ann, der Chem. und Pharm, vol. cxlviii, p. 269.

ential quotient equal to 0, then the minimum value of W becomes, when n= ∞,

W=v nat. log. a.

(4)

Precipitates obtained in the course of chemical analysis may in all cases be assumed to be sufficiently washed when the impurity retained by them amounts to no more than the To part. Making therefore a = 100000 and v = 1, it results from equation (4) that the least quantity of fluid required in order to remove the impurity contained in a precipitate to the part, amounts to eleven and a half times the volume occupied by the precipitate itself in the liquid in which it exists. It is evident, therefore, that the amount of water actually necessary to wash a precipitate the more nearly approaches this minimum the oftener we decant, and the smaller the quantity of washingwater we employ at each decantation.

Since some of the principal sources of error in analytical work consist in the incomplete or in the too protracted washing of precipitates, it becomes important to know how to ascertain the progress of the washing throughout the several stages of the process. By employing the same volume of water at each successive addition, and est mating its relation to that of the precipitate remaining at th bottom of the vessel or upon the filter, we can find from the following Table, calculated by

means of the formula above given, the number of times it is necessary to decant in order to diminish the amount of impurity in the precipitate to the T', 300, 308 or 1 part.

Column I. shows the relation between the volume of the precipitate and that of the washing-water employed for each successive decantation, column II. the number of decantations required to diminish the amount of impurity to the necessary extent, and column III. the total volume of water obtained from the several decantations.

When the washing-process is performed in a beaker, the relation between the volume of the precipitate and that of the liquid may be easily determined by holding a strip of paper along the side of the vessel and marking upon it the respective heights of the precipitate and supernatant liquid; then on folding the portion of paper lying between the two marks in such a manner that each fold corresponds to the height occupied by the precipitate, the number of folds will give the argument in column I. to find in column II. the number of decantations needed to wash to the required extent. If the washing be conducted as in the ordinary method of filtration, funnels possessing an angle of 60° must be invariably employed, and the capacities of the various-sized filters once for all determined by means of a burette. After the precipitate has been brought upon the filter and allowed to drain, it is mixed as thoroughly as possible with water from a graduated washing-flask. Call the amount of water thus necessary to fill the filter, and the v V capacity of the empty filter U, then in column I; FID v that is, the argument needed to find in column II the number of times it is necessary to refill the filter in order to wash the precipitate to the desired extent.

I by far prefer using this Table to employing the method generally followed of ascertaining the completion of the washing-process by evaporating a quantity of the filtrate on platinum-foil, since in the latter case it is only possible to obtain an infallible proof when we have to deal with a precipitate possessing an extremely high degree of insolubility; if the precipitate be soluble to any marked extent, the result is completely illu

sory.

In the process of filtration as hitherto conducted, the time required is so long and the quantity of wash-water needed so great that some simplification of this continually recurring operation is in the highest degree desirable. The following method, which depends, not upon the removal of the impurity by simple attenuation, but upon its displacement by forcing the wash-water through the precipitate, appears to me to combine all the requisite conditions and therefore to satisfy the need.

The rapidity with which a liquid filters, depends, cæteris paribus, upon the difference which exists between the pressure upon its upper and lower surfaces. Supposing the filter to consist of a solid substance, the pores of which suffer no alteration by pressure or by any other influence, then the volume of liquid filtered in the unit of time is nearly proportional to the difference in pressure: this is clearly shown by the following experiments, made with pure water and a filter consisting of a thin plate of artificial pumice-stone. The thin plate of pumice was hermetically fastened into a funnel consisting of a graduated cylindrical glass vessel, the lower end of which was connected with a large thick flask by means of a tightly fitting caoutchouc cork. The pressure in the flask was then reduced by rarefying the air by means of a method to be described upon another occasion; and for each difference of pressure p, measured by a mercury column, the number of seconds t was observed which a given quantity of water occupied in passing through the filter. The following are the results :

In the ordinary process of filtration, p on the average amounts to no more than 0.004 to 0.008 meter. The advantage gained, therefore, is easily perceived when we can succeed by some simple, practicable and easily attainable method in multiplying this difference in pressure one or two hundred times, or, say, to an entire atmosphere, without running any risk of breaking the filter. The solution of this problem is very easy an ordinary glass funnel has only to be so arranged that the filter can be completely adjusted to its sides even to the very apex of the cone. For this purpose a glass funnel is chosen possessing an angle of 60°, or as nearly 60° as possible, the walls of which must be completely free from inequalities of every description; and into it is placed a second funnel made of exceedingly thin platinum-foil, and the sides of which possess exactly the same inclination as those of the glass funnel. An ordinary paper filter is then introduced into this compound funnel in the usual manner; when carefully moistened and so adjusted that no air-bubbles are visible between it and the glass, this filter, when filled with a liquid, will support the pressure of an extra atmosphere without ever breaking.

The platinum funnel is easily made from thin platinum-foil in the following manner:-In the carefully chosen glass funnel

is placed a perfectly accurately fitting filter made of writingpaper; this is kept in position by dropping a little melted sealing-wax between its upper edge and the glass; the paper is next saturated with oil and filled with liquid plaster of Paris, and before the mixture solidifies a small wooden handle is placed in the center. After an hour or so the plaster cone with the adhering paper filter can be withdrawn by means of the handle from the funnel, to which it accurately corresponds. The paper on the outside of the cone is again covered with oil, and the whole carefully inserted into liquid plaster of Paris contained in a small crucible 4 or 5 centims. in height. After the mixture has solidified, the cone may be easily withdrawn ; the adhering paper filter is then detached, and any small pieces of paper still remaining removed by gently rubbing with the finger. In this manner a solid cone is obtained accurately fitting into a hollow cone, and of which the angle of inclination perfectly corresponds with that of the glass funnel.

Fig. 1 represents the cones. By their help the small platinum funnel is made. A piece of platinum (fig. 2 shows the natural size) is cut from foil of such a thickness that one square centimeter weighs about 0.154 grm., and from the center a a vertical incision is made by the scissors to the edge cbd. The small piece of foil is next rendered pliable by being heated to redness, and is placed upon the solid cone in such a manner that its center a touches the apex of the latter; the sides a b d are then closely pressed upon the plaster, and the remaining portion of the platinum wrapped as equally and as closely as possible around the cone. On again heating the foil to redness, pressing it once more upon the cone, and inserting the whole into the hollow cone and turning it round once or twice under a gentle pressure, the proper shape is completed. The platinum funnel, which should not allow of the transmission of light through its extreme point, even now possesses such stability that it may be immediately employed for any purpose. If desired, it may be made still stronger by soldering down the overlapping portion in one spot only to the upper edge of the foil by means of a grain or two of gold and borax; in general, however, this precaution is unnecessary. If the shape has in any degree altered during this latter process, it is simply necessary to drop the platinum funnel into the hollow cone and then to insert the solid cone, when by one or two turns of the latter the proper form may be immediately restored. The platinum funnel is placed in the bottom of the glass funnel,

* The size of this cut has been reduced about one third. The diameter of fig. 2, in the original drawing, is 2.5 millimeters.-Eds.