matter of interest to know that from whatsoever locality it comes this minute quantity of rubidium and cæsium occurs with it. On some future occasion I hope to be able to bring together certain generalities in this connection of more or less interest to mineralogists.

I have also detected rubidium in half a gramme of margarodite and Warwick mica, and have failed to detect it in apophyllite, thomsonite, pectolite, elæolite, chesterlite, cancrinite, and other silicates.

WARWICKITE.

This interesting compound has been known for some time to American mineralogists, having been first described by Prof. C. U. Shepard under the name of warwickite, and considered as a hydrated silico-titanate of magnesia, iron, and alumina. It was afterward described by Mr. T. S. Hunt under the name of enceladite, and in his analyses (Amer. Jour. of Science and Arts, 2d series, xi, 352) considered a trititanate of magnesia.

In the re-examination of American minerals, in which Mr. Brush and myself were engaged, this mineral came up in turn for examination, and to our amazement it is found to contain a large amount of boracic acid, doubtless upward of twenty per cent. Approximative analyses are already made, but owing to the difficulty of obtaining it of sufficient quantity in a perfect state of purity, its final examination may be delayed for some time; and it is for that reason thought advisable to publish the present note on the subject. It is essentially a borotitanate of magnesia and iron; the metallic acid, however, has some anomalies about it not yet cleared up. This is the first borotitanate known, and as such highly interesting; the smallest portion of it when acted on with sulphuric acid will give the strongest indication of the presence of boracic acid.

DETERMINATION OF ALKALIES IN MINERALS.* IINERA

1. In the examination for alkalies in the class of minerals alluded to in this article it is usual to devote a separate portion of the mineral to their special determination, without having reference to any of the other ingredients contained in the mineral. This method of proceeding naturally recommends itself, because a fusion with carbonate of soda is so greatly superior for the determination of all other ingredients that even the attempt to control the result of the soda fusion by making use of the one for the alkalies, to arrive at the other substances as well as the alkalies, will in many instances embarrass the analyst as to his results.

2. It is only in cases of absolute necessity that one portion of the mineral should be used to estimate all its constituents, and this condition of things will be alluded to in another part of this paper, as reference is now had to the quantitative determination of the alkalies, discarding whatever else the mineral may contain.

3. In the determination of the alkalies in silicates not soluble in acids three important points present themselves:

I. The means necessary to render the silicate soluble. II. The separation of the other ingredients from the alkalies, more especially magnesia.

III. The removal of the sal ammoniac unavoidably accumulated in the process of analysis.

In all three of these the processes adopted will be found to differ essentially from those now in use, and they are made known only after much experience by the author, in which their advantages have been most fairly tested, comparatively

*This memoir embraces many important points connected with mineral analysis. The minute practical details for laboratory use are given in another article in this collection of papers, and one written after twenty years' experience with the method.

with methods already employed. In order that these processes may serve equally well in the hands of others, they will be given with some detail.

I. METHOD OF RENDERING THE SILICATE SOLUBLE.

4. To render the silicate soluble various plans have been proposed, all of which have their objections. Among the agents used for the purpose are baryta and several of its compounds; viz., the nitrate, carbonate, and chloride.

5. The first of these is undoubtedly the best decomposing agent of the four, could we use a platinum crucible to heat the mixture of it and the mineral; as it is, a silver crucible is necessary, and this is not always capable of standing the requisite heat. According to Rose, "the silver crucible must be very strong, for if thin the action of a red heat might crack it, and a portion of the fused mass would ooze out through the crevices." It also may happen that a heat higher than the point of fusion of silver is necessary to a complete decomposition of the mineral.

6. All that is here said of caustic baryta is equally applicable to nitrate of baryta.

7. The chlorides of barium and calcium have been lately proposed by Prof. Henry Wurtz, but its decomposing properties are very feeble, as the chlorine in combination with the barium is not liberated at a white heat, and few silicates are able to produce the decomposition. It may succeed with some of the feldspars, but decomposes very imperfectly even the micas. So it is rather a risk to employ it with an unknown substance.

8. The carbonate of baryta is the compound of baryta most generally employed for silicate decompositions; still this is attended with much difficulty, owing to the infusibility of this salt and the impossibility of driving off the carbonic acid by heat alone; and even if this latter were possible, the objection pertaining to caustic baryta would then arise.

9. The following extract from Rose's Analytical Chemistry (translation by Normandy, in a note by the translator) presents fairly the difficulties attending this method of decomposing the silicates:

"The heat applied is so intense that some precautions must be taken. The platinum crucible containing the mixture should be exposed first to the

heat of an argand lamp, and when the mass begins to agglutinate the crucible should be closed and its cover tied down with platinum wire, then placed in a Hessian crucible closed up also; the whole is placed upon an inverted crucible and submitted to the action of the blast of a wind-furnace, beginning first gradually with a red heat, piling on more coke, so as to fill up the furnace, and increasing the heat to the highest possible pitch, until the Hessian crucible begins to soften. It is absolutely necessary to the success of the operation that the Hessian crucible should be closed as well as possible, which is best done by luting the cover with fire-clay; the Hessian crucible and its cover, having fused together, can not be separated except by breaking, etc."

It will be seen in reading this extract that the heat required is not ordinarily at the command of most chemists; in fact, no other variety of furnace than a Sefstroem can be depended on for a complete decomposition.

10. Caustic lime and its salts have also been recommended and long used for the more imperfect decomposition of silicates, as for obtaining lithia from spodumene and lepidolite. Lime or its carbonate, well mixed with many silicates finely pulverized, will decompose them completely at a white heat, but no one salt of lime is capable of meeting the demand of the entire range of alkaline silicates.

11. In consideration of these difficulties Berzelius proposed the use of hydrofluoric acid, and this method, when applied with the numerous precautions required, will serve to decompose all silicates; still, according to Rose, there are siliceous compounds that can not be completely decomposed by hydrofluoric acid. Besides, this acid is a most disagreeable one to manipulate with, whether we employ Brunner's apparatus or Laurent's method, or, what is always the best, the concentrated acid previously prepared. I may also add that the necessity of using sulphuric acid after the decomposition is made is another objectionable feature in this process.

12. The above furnishes a hasty review of the methods we are now possessed of for decomposing the silicates in order to determine their alkalies; their merits can be contrasted with those of the method about to be described.

13. The decomposing agent which I present as a substitute for all others, and as capable of meeting the demands proposed in the commencement of this article, is a mixture of carbonate of lime and fluor-spar.

14. Carbonate of lime I have used for more than six years

for decomposing certain of the alkaline silicates, and more successfully than carbonate of baryta; still in numerous instances the decomposition was far from complete, and the method unsatisfactory. Notwithstanding these failures, I felt convinced that lime was the most powerful decomposing agent that could be conveniently employed for this purpose, as it could be used in its caustic state in a platinum crucible without injuring the latter, although exposed to the highest temperature. When its carbonate is used a red heat sufficed to drive off the carbonic acid and bring the mineral under the action of caustic lime-a circumstance that does not take place with carbonate of baryta; and it is well that it does not, for otherwise the platinum crucible would be seriously injured.

15. It was evident that the only obstacle in the way of lime decomposing the silicates as thoroughly as caustic potash was the impossibility of fusing the mixture, and thereby bringing the pulverized mineral and lime intimately in contact. This difficulty overcome, I felt confident of success. Without detailing the various methods resorted to, it will suffice to state that the object in view was to use some flux along with the mixture of the silicate and lime, which would render the mixture fluid at a bright red heat. The two substances which recommend themselves after many experiments are the fluoride and chloride of calcium, neither of which have any marked decomposing action on the silicates; in fact, their action is simply that of fluxes, which enable the lime and silicate to come in contact in a liquid state, effecting nothing beyond that. It is with the fluoride of calcium that we have to do in this part of the paper, leaving the details on the use of the chloride of calcium until further experiments are made to test fairly its value.

16. The manner in which I proceed is as follows: Pulverize the silicate to a sufficient degree of fineness; it is not required that the levigation be carried to any great extent; mix intimately in a glazed porcelain mortar a weighed portion of the mineral with one part of pure fluor-spar and four or five parts of precipitated carbonate of lime; * introduce it into a platinum

*The fluor-spar used is the transparent variety, free from all impurities. It is easily and abundantly procured in this as well as in all other countries. The carbonate of lime is made by dissolving calc-spar or pure marble in