UNITED STATES PATENT OFFICE,

Washington, D.C., February 11, 1891. DR. H. CARRINGTON BOLTON, University Club, N.Y. DEAR SIR,In response to your request I take pleasure in giving you the following information regarding the past accomplishments and plans for the future of the Scientific Library in the matter of technological indexing.

The work of indexing periodicals has been carried on in the Library for some years in a somewhat desultory fashion, taking up one journal after another, the object being, apparently, more to supply clerks with work than the pursuance of any well-defined plan. However, one important work has been substantially completed, viz., a general index to the whole set of the Scientific American and Supplement from 1846 to date.

It is unnecessary for me to point out to you the importance of this work, embracing a collection which has held the leading place in the line of general information on invention and progress, the labour of compiling which has been so formidable that no movement in that direction has been attempted by the publishers, except in regard to the supplement only, and that very imperfectly. This index embraces now 184,000 cards, not punched, and at present stored in shallow drawers and fastened by rubber bands, and, of course, they are at present unavailable for use. There is little prospect of printing this index, and I have been endeavouring for some time to throw the index open to the public by punching the cards and fastening them with guard rods, but as yet have made no perceptible impression upon the authorities, although the expense of preparation would be only about 70.00 dollars. There has also been completed an index to the English journal Engineering, comprising 84,000 cards, from the beginning to date.

An index to Dingler's Polyt. Jour. was also commenced as long ago as 1878, carried on for six or seven years and then dropped. I hope, however, at no remote date, to bring this forward to the present time.

On taking charge of the library I was at once impressed with the immense value of the periodical literature on our shelves, and the great importance of making it more readily accessible, and have had in contemplation for some time the beginning of a card index to all our periodicals on the same general plan as that of Rieth's Repertorium. I have, however, been unable to obtain sufficient force to cover the whole ground, but have selected about 150 journals, notably those upon the sub. jects of chemistry, electricity, and engineering, both in English and foreign languages, the indexing of which has been in progress since the 1st of January. This number includes substantially all the valuable material in our possession in the English language, not only journals, but transactions of societies, all the electrical journals, and nearly all the chemical in foreign languages. This index will be kept open to the public as soon as sufficient material has accumulated. In general plan it will be alphabetical, following very nearly the arrangement of the periodical portion of the Surgeon General's catalogue. I shall depart from the strictly alphabetical plan suffi ciently to group under such important subjects as chemistry, electricity, engineering, railroads, &c., all the subdivisions of the art, so that the electrical investigator, for instance, will not be obliged to travel from one end of the alphabet to the other to find the divisions of gene rators, conductors, dynamos, telephones, telegraphs, &c., and in the grouping of the classes of applied science the office classification of inventions will, as a rule, be adhered to, the subdivisions being, of course, arranged in alphabetical order under their general head, and the title of the several articles also arranged alphabetically by authors or principal words.

With many thanks for the kind interest and valuable information afforded me, I remain, Very truly yours,

HOWARD L. PRINCE, Librarian Scientific Library.

IN the following table all the foregoing analyses o Norwegian material have been re-calculated to the percentages found, excluding the insoluble matter, in order that their true relations may appear at a glance, whereby the sum of the rare earths combined is given instead of each earth or group of earths by itself. Only those constituents are tabulated which may be considered unquestionably in whole or in part as belonging to the uranium mineral. Silica, ferric or ferrous oxide as it may be, the small amounts of magnesia, alkalies, &c., may be thrown out as derived from admixed impurities. Lime from its close relationship to the other earths has been retained, though it is undoubtedly, in part at least, derived from foreign silicates.

TABLE V.—Analyses of Norwegian Uraninite

Re-calculated.

An examination of analyses XII. to XVI. as recalculated above hardly allows of any other conclusion than that the specimens from the above four different quarries about Moss belong to one and the same mineral species. While analysis XII. differs in some respects reproduced in its original form and as modified by subfrom Blomstrand's analysis of bröggerite, which is here stitution of Fe2O3 for FeO, it simply serves to show, assuming correct analyses, that the mineral may vary in composition in the same quarry, for, as before mentioned, this is the original material from which Blomstrand made the species.

The estimation of UO2 in the analyses of Table IV. is thoroughly to be depended on, aside from the slight deduction to be made for possible ferrous oxide, while in

From "Bulletin No. 78, U.S. Geological Survey, 1889-90." An abstract of this paper was published in the Am. Jour. Sci., vol. xl.,

p. 384.

NEWS

that

view of the want of evidence as to the manner of filling his tubes with carbonic acid there is no guarantee Blomstrand's figure may not be too low. Leaving out of consideration for the moment the nitrogen, it is clear in any case that the formula evolved by him from his analysis

cannot be derived from No. XII.

In regard to analyses XIII. and XIV., there can exist no doubt whatever as to the identity of the material derived from Profs. Brögger and Nordenskiöld inde. pendently. A comparison of them with Lorenzen's analysis, coupled with the fact that not one of the specimens from the neighbourhood of Moss above analysed contains less than 8 per cent of earths, gives rise to the strongest possible suspicion that that analyst has overlooked thoria altogether, notwithstanding the fact that the present material is from Elvestad and Lorenzen's (Nat Mag. f. Naturv., vol. xxviii., p. 249) was from Huggenäskilen.

the others would differ from it more widely than in the above exhibit.

The correctness of the orthouranate formula for bröggerite itself having been invalidated by the difference between his own analysis and anal. XII. above, it is hardly worth while to discuss its applicability to the Bohemian and Saxon uraninites, of which no complete and reliable analyses seem to have been made, except perhaps the single one by Ebelmen in 1843 on Joachimsthal material, which on re-calculation by Blomstrand was found to conform to his view. No more singular example can be furnished of incorrect conclusions founded on seemingly the best of evidence. No blame can attach to Blomstrand, but it can hardly be doubted that had he been able to analyse material from more than one quarry about Moss he would have seen the impossibility of reconciling the discrepancies in composition so as to admit of the application of one general formula.

and one or two other discrepancies, that analysis XVII. It is apparent, notwithstanding the deficiency of earths

reliable.

was really made upon cleveite, as the label indicated. The density, too, corresponds almost exactly with that found by Nordenskiöld (Geol. För. Förh., 1878, vol. iv., P. 28; Zeit. f. Kryst., vol. iii., p. 201) and Lindström. In the latter's analysis, which is quoted below, "water and a trace of CO2 are given as loss on ignition, a result which under the circumstances must be considered unThe same objection applies to Hidden and Mackintosh's analysis of nivenite (Am. Jour. Sci., [3] 1889, vol. xxxviii., p. 481), where it is stated plainly that the water represents loss on ignition. The analyses of clevite and nivenite are here given together with XVII. from Table IV. for convenient comparison. It may be mentioned that the 23'07 per cent of UO2 in clevite is the mean of two results differing by nearly 1 per cent. Taking the higher as more nearly correct than the mean, almost exact agreement with the figures of analysis XVII. for UO2 and UO3

is obtained.

UO3 UO2 PbO FeO CaO H2O SiO2

38.23 50'42

37'73 50.89 9'72

Fe2O3 0.28

O'21

SiO2
Ign.
Insol.
CO2 and N

Prof. Brögger, as quoted (ante), says, however, that the material sent by him is to the best of his knowledge from the same place as that analysed by Lorenzen, and this appears probable from the fact that the specimens from Huggenäskilen sent by Prof. Nordenskiöld show a totally different ratio between UO2 and UO3 (anal. XVI.). If it should prove that Lorenzen erred in overlooking thoria, another of Blomstrand's supports in favour of the orthouranate formula for all uraninites, including bröggerite and cleveite, is knocked away, the first being the earlier and, as shown, incorrect analysis of Branchville uraninite.

The oxygen ratios calculated for analyses XII. to XVI., counting all earths as thoria, whereby the comparison is very little affected, since the percentages of the other earths are almost alike in all analyses, are as given below, as also the ratios for Blomstrand's bröggerite calculated from the second column (see above) instead of the first in order to compare properly with the others.

It is seen that none of the analyses conform even approximately to the ratio for Blomstrand's, except that of the mineral from Huggenäskilen. A re-calculation of all on the basis of FeO for Fe2O3 and consequent changes in UO2 and UO3, and separation of the earths, would give the normal ratio 1: 1 for Blomstrand's analysis, but

UO2 Th02

Ce2O3, &c... Y2O3, &c.

Род Fe2O3.. CaO MgO

(a) And alk.

The appearance of clevite is at once suggestive of alteration, and, in view of the uncertainty attaching to the water determinations for both nivenite and cleveite, it is probable that they represent nearly the same stage of alteration of the same species. What this species is is pretty clearly indicated by analysis XVIII. of Table IV., where the earths are in about the same proportions and total amounts as in XVII., but the UO, and UO2 stand in a very different ratio. The material for this, as before said, came from Arendal, and presumably from the precise locality of the clevite of analysis XVII., since the pieces

of felspathic rock containing them were in one package without distinguishing labels, but it was unquestionab y fresher. The extreme solubility of this material com. pared with the other Norwegian uraninites is shared by that of analysis XVII., by cleveite, and by nivenite, and is to be explained probably not so much by advanced decomposition as by the preponderance here of a more soluble yttrium-uranium compound. Whether or not this last Arendal material is the source whence cleveite and its American representative have been derived by alteration, as seems most probable, it is in any event a true uraninite of more basic character than any of the Norwegian thorium uraninites, and consequently conforms still less than those to the orthouranate formula.

Bohemian and Saxon Uraninite.-Nitrogen was carefully sought for in uraninite from Przibram, Joachimsthal, and Johanngeorgenstadt, using the apparatus figured (ante). The extremely small bulk of gas finally obtained from all of them, after cleansing with potassium hydrate, measured about 0'2 c.m.3 for Przibram and Joachimsthal, and much less for Johanngeorgenstadt, I grm. of mineral having been used. The above volume represents about 0.02 per cent in weight of what can hardly be anything else than nitrogen.

None of the specimens contained zirconia, thoria, or other rare earths. Owing to the uncertainty of being able to determine with any close approach to truth the proportions of UO2 and UO3 in the presence of sulphides, and compounds of arsenic and vanadium of unknown degree of oxidation, no quantitative analyses have been carried out, but the attempt will yet be made to solve their composition.

Discussion of Analyses.

Hitherto the analyses have been considered in groups and without special reference to the nitrogen. It has been sought to show on grounds which would be valid even without its presence that the orthouranate formula is capable of no general application to uraninite, and that in the one or two cases where it does seem to apply this agreement is probably accidental. Taking into consider. ation the low atomic weight of nitrogen as compared with uranium, thorium, and lead, it is plain that it must play an important part in the constitution of the molecule, and that therefore its discovery alone, without other evidence furnished by the analyses, is sufficient to invalidate entirely the practically identical formulæ of Comstock and Blomstrand.

Throughout the whole list of analyses in which nitrogen has been estimated, the most striking feature is the apparent relation between the UO2 and the nitrogen. This is especially marked in the second table of Norwegian uraninites (Table V.), from which the rule might almost be formulated that, given either nitrogen or UO2, the other can be found by simple calculation. The same ratio is not found in the Connecticut varieties, but if the imperfect determination of nitrogen in the Branchville mineral is to be depended on, the rule still holds that the higher the UO2 the higher likewise is the nitrogen. The Colorado and North Carolina minerals are exceptions, but it should be remembered that the former is amorphous, like the Bohemian, and possesses the further similarity of containing no thoria, though zirconia may take its place, and the North Carolina material is so much altered that its original condition is quite unknown.

In the absence of all positive knowledge whatever as to the rôle which nitrogen plays in the mineral, it would be idle to speculate at present upon the proper position of the latter in mineral classification. Much remains to be done before this question can be elucidated, but the general direction which the work must take can be fairly indicated.

But two explanations seem possible to account for the wide differences in the oxygen ratios for UO3 and total bases, varying as they do from 1 : 4'37 for the Branchville

material of analysis VI. to 1:1 for Blomstrand's brög. gerite, and even to a ratio indicating acidity for nivenite. Either all the others are alteration products of a mineral having the composition of the Branchville occurrence, or even of some unknown body entirely free from UO3; or they are mixtures of two or more substances which need not necessarily be isomorphous, for Baumhauer's investigations have shown that even well developed crystals of smaltite and chloanthite are often mechanical mixtures. Fractional solution might throw light on this point, and the following are the results of experimentation :

I. In one of the nitrogen estimations in Glastonbury uraninite by the original method the operation had to be discontinued before solution was complete. The undissolved residue, amounting after extraction of lead sulphate to o'1234 grm., was found to contain 71°94 per cent UO2, the original percentage having been 59'93.

II. 1095 grms. were boiled for twelve hours with dilute hydrochloric acid in a flask connected with a reversed condenser. The residue was o'9193 grm., showing that but 16 per cent had been dissolved, and it contained 64.66 per cent UO2. The nitrogen was lost.

III. 2012 grms., when boiled with the same hydrochloric acid for forty-two hours, afforded a residue of 1*4975, showing solution of 25.5 per cent. Its composition

Here it is seen that the UO2 has increased, and likewise the nitrogen, both in about the same proportion, but it is not clear why the earths and lead oxide should not show a greater change.

These experiments require repetition and extension to other than the Glastonbury mineral, but so far as they go they indicate that uraninite as we know it is not a single chemical substance, and that, as was to be expected in such case, the portion containing most UO2 is the least soluble. They are not decisive as between the hypothesis of original mixture and that of mixtures resulting from alteration. In this respect the following observations are of interest :-In dilute hydrochloric acid a large piece of North Carolina uraninite falls wholly to finest powder in course of time, which would hardly happen were it one original substance in process of alteration, unless the change affected every smallest particle of the mineral simultaneously and to the same extent. Hydrofluoric acid speedily covers the faces of a bright crystal of Branchville uraninite with a network of fine cracks visible only under the glass, and soon powder can be rubbed off with the fingers.

Whatever may be the eventual conclusion, it will be found that the small amount of water afforded by all analyses must be carefully considered. Small as this amount is, in consequence of its low molecular weight as opposed to uranium, thorium, and lead, it must play an important part in the mineral as a homogeneous whole or in one of its parts if a mixture. In the latter case it will unquestionably be found to belong to the more soluble component.

It is significant that the two uraninites highest in UO2 give the lowest water, and while it is true that the Branchville mineral appears to contain more than that from Glastonbury, this may be only apparent, for the former was simply ignited in a current of air, while the latter was fused with sodium carbonate, and it is known that the vapours from both possess an acid reaction, and

in the latter case give an increased weight to the calcium chloride tube when no alkaline carbonate is used. It is important that in all future analyses stress shall be laid upon the utmost accuracy in the water determinations, and if insoluble matter is present, it must be ascertained if possible how much, if any, of the water is to be ascribed to this source.

Before leaving this portion of the subject it is necessary to call attention to one difficulty presented by the analyses for which no satisfactory explanation yet presents itself. In the majority of cases in which all constituents have been estimated, the summation is considerably in excess of 100, and in some cases where only one or two are wanting it is evident that they would, if determined, produce a like result. Although it did not seem possible that this almost constant excess could be due to impurities from reagents, or to want of sufficient care in washing or igniting precipitates, it was determined to make an analysis on Glastonbury material with the utmost possible care, using specially prepared reagents and only platinum vessels and reducing the number of weighings to be used as the basis for the summation to a minimum. This analysis, executed partially in duplicate, is No. V. of Table I., and the excess still appears. It seems as if some one of the weighings must have been uniformly too high, but how this can be is a mystery. Efforts will still be made to find a solution. It cannot be sought in a replacement of oxygen by nitrogen in combination with uranium, like the replacement of oxygen by fluorine in many minerals, for since the nitrogen is freed as a gas by sulphuric acid, it is immaterial, so far as the summation is concerned, whether the proportions of UO2 and UO3 as found by titration are correct or not. A certain amount of oxygen has been used, and, assuming its correct determination, it does not alter the result whether this has been employed to oxidise a suboxide of uranium to UO3, supposing nitrogen to have replaced a part of the oxygen in UO2, or only in oxidising UO2 to UO3. In the former case the actual percentage of UO, in the mineral would be increased, but the oxygen

consumed would be the same.

In a former notice (Am. Journ. Sci., 1889, [3] vol. xxxviii., p. 329) it was mentioned that a relation appeared to subsist between the nitrogen and the UO2, but this statement was based on two experiments which subsequent work showed to be illusory. There is certainly a relation, but not exactly such a one as was then surmised. It was found, namely, on examining the portions used for water determinations (by ignition in dry air without sodium carbonate) for UO2 and nitrogen, that the residual amounts of these were exactly in proportion to their original percentages, showing that the loss of nitrogen kept step with the oxidation of UO2 to UO3. later found that these results were accidental, for by sufficiently long heating the nitrogen was entirely eliminated, so far as the usual tests showed.

(To be continued).

It was

Volumetric Determination of Hydrocarbon Vapours in Coal-gas.-W. Hempel and L. M. Dennis (Ber. Deutsch. Chem. Gesell.).-The gas is measured in a simple gas burette over water saturated with coal-gas; it is then conveyed through a capillary tube into a gas pipette, in which it is in contact with 1 c. c. of absolute alcohol over mercury, with which it is shaken up for three minutes. The gas is then returned to the burette, taking care that no alcohol penetrates into the connecting capillary. In order to absorb the alcoholic vapour out of the remaining gas, it is again conveyed into a pipette, in which it is shaken up for three minutes with i c.c. of water over mercury, after which the gas is conveyed back into the burette. The difference of volume corresponds to the vapours.

OPENING UP CHROME IRON ORE BY HYDROCHLORIC ACID UNDER PRESSURE.

By PAUL JANNASCH and HANS VOGTHERR.

My recently published method of opening up silicates by means of strong hydrochloric acid under pressure in a special platinum apparatus, has been since applied, in concert with H. Vogtherr, to an extensive series of minerals with great success. We have succeeded in completely opening up silicates, such as garnet (pyrope of Meronitz and Arendal, &c.), also black horn-blende (from Granatilla and from Schima), also cerite, so as to be quite suitable for the purpose of an accurate quantitative analysis.

The more detailed results, as well as the different modifications in the execution of the process, will be published in a future communication. In the present short communication we wish in the first place to direct attention to the energetic action of concentrated hydrochloric acid under pressure upon chrome iron ore, which is well known to be very refractory under the common method of opening up by fusion. It is perfectly opened up by hydrochloric acid at 250°, which will greatly facilitate the accurate and complete analysis of the many important deposits of this mineral. The opening up of the mineral is effected as follows:--

I.-Chrome Iron Ore from Baltimore.

The

The ore, in fine powder, was mixed with 2 grms. salammoniac and 10 c.c. of hydrochloric acid (4 vols. acid at 1119 specific gravity to I vol. water), sealed up in a tube of potash-glass and heated for 8-10 hours to 2752900. The acid employed had been previously saturated with ammonium chloride by shaking up with the latter salt. The quantity of ore taken was 10426 grms. silica, 00194 grm. = 1'86 per cent, was volatilised by heating on the water-bath with hydrofluoric and sulphuric acids; the residual sulphuric acid was evaporated on the air-bath, and the residue was ignited and weighed. It amounted to o'0018 grm. 0'17 per cent. II.-Chrome Iron Ore from Frankenstein in Silesia, and III.-Chrome Iron Ore from Orsova.

=

Both these specimens were treated in a similar

manner.

In practice, if it is merely required to determine the percentage of chrome in a specimen, the author's platinum apparatus may be dispensed with, and the opening up may be effected in a sealed tube.-Ber. d. Deutsch. Chem. Gesell., vol. xxiv., p. 3206.

A Spectro Colorimeter.-D'Arsonval (Séances Soc. Franc. de Phys.).—In this instrument the eye-piece of an ordinary Dubosq colorimeter is replaced by a small direct vision spectroscope, the slit of which is fixed vertically to the line separating the two halves of the field of vision. If the liquid under examination is placed in one of the tubes and the standard of comparison in the other, and if the substance produces sufficiently characteristic absorption-phenomena, a given constituent can be recog nised even in a mixture of bodies of different colours. The same author describes (ibidem) a differential spectrometer without polarisation. The rays coming from the sources of light to be compared are thrown upon two parallel tubes closed by two similar objectives, and placed before a spectrum apparatus. The light of each source is thrown upon one-half of the slit by two pairs each of two total reflective prisms. In order to vary in a measurable manner the luminosity of the two spectra, which appear superimposed, a sliding screen is placed before the two object-glasses, which cuts off the light differently, according to its position.

NEW METHOD FOR THE ANALYSIS OF TIN ORES, AND FOR THE SEPARATION OF COPPER AND CADMIUM.*

By J. S. C. WELLS, Ph.D.

TIN ORE.

THE methods for the analysis of tin ores are so tedious and unsatisfactory that I was led, some time ago, to try and discover some simpler way of decomposing the ore; that being the chief difficulty in all the old methods.

In my first experiments, fusing the ore with borax was tried, and although this was found to effect the desired result, as far as the decomposition of the ore, still it introduced other difficulties that caused me to give it up.

The borax, at the high temperature of the fusion, attacked the platinum of the crucible to such an extent that it became necessary to separate the platinum from the tin. Boracic acid, as was to be expected, acted the same as borax. The idea then occurred to me that the ore might possibly be reduced in the same way as the On artificial oxide, i.e., by means of nascent hydrogen. trying the experiment, I found that in this way the tin contained in the ore could easily be obtained in the metallic state. My first tests were made as follows:About I grm. of the finely pulverised ore was placed in a large test-tube with a few pieces of zinc and some dilute hydrochloric acid. The first trials were made on an ore from Cornwall, and it was found to be very easily reduced. A sample of Durango ore, consisting of nearly pure cassiterite, was then tried, but the reduction did not take place as readily as with the ore from Cornwall; the reduction of the latter was nearly complete at the end of an hour, whilst the former at the end of three hours showed little change. I then added a piece of platinum with the zinc and acid, and found the result to be very satisfactory, the ore being quickly reduced. Heating the test to boiling was also found to aid the reaction. The addition of the platinum also facilitates very much the subsequent solution of the reduced tin in hydrochloric acid, tin alone dissolving but slowly in the acid.

It is advisable to shake the test frequently during the reaction, so as to keep the ore in contact with the zinc and platinum. If this is not done, the ore settles to the bottom and the reduction takes place very slowly, if at all.

As soon as the decomposition of the ore appears to be complete, the remaining zinc and the reduced tin are dissolved in hydrochloric acid and filtered from any undecomposed ore or gangue. This residue should again be tested in the same way, with fresh zinc, platinum, and hydrochloric acid, to see if all the tin has been extracted by the first operation. After the tin has been ob tained as chloride, it can, of course, be determined by any of the usual methods.

Instead of dissolving in hydrochloric acid, the tin might be dissolved in ferric chloride, after removal of the excess of zinc, and then determined volumetrically. The metal, being in such a finely divided state, would be very easily soluble in this reagent.

SEPARATION OF COPPER AND CADMIUM.

I find that copper and cadmium may be easily separated by the following method:

To the neutral solution containing these metals (ammonia salts must not be present) add sodium thiosul. phate (hyposulphite) until the solution becomes colourless, then add sodium carbonate, and the cadmium will be precipitated as carbonate (white); filter, and to the filtrate add HCI, and boil, and the copper will be precipitated as sulphide. To use this method in the ordinary course of analysis, the solution, after removal of the bismuth, would

* From School of Mines Quarterly, xii., No. 4.

have to be evaporated to dryness and ignited so as to remove all ammonia salts.

ON NEW QUANTITATIVE SEPARATIONS OF MANGANESE AND NICKEL, MANGANESE AND COBALT, AND OF MANGANESE, NICKEL AND COBALT.

By PAUL JANNASCH and CARL J. FRANZEK.

SOME time ago one of the present writers, in concert with Mr. McGregory, succeeded in ascertaining and laying down those conditions under which manganese and zinc may be very conveniently and exactly separated by means of hydrogen peroxide in a strongly ammoniacal solution, and in presence of very much ammonium chloride. We are now in a position to state that the separation of manganese and nickel can be effected under the same conditions and with the same accuracy as that of manganese and zinc.

In view of these experimental facts we hoped to separate manganese and cobalt in the same or in an analogous manner. But our analyses soon showed the impossibility of a quantitative separation of the two metals in the manner adopted, for all our experiments failed by reason of the marked obstinacy with which very considerable quantities of cobalt (on the average several per cents) adhered to the manganese hydroperoxide, so that even precipitations of the manganese several times repeated proved quite unsatisfactory.

After a long series of the most different modifications of the experimental conditions, we found the solution of the double (potassic) cyanide of the metals an excellent means for the quantitative separation of manganese and cobalt. From such a solution hydrogen peroxide throws down manganese absolutely free from cobalt. The behaviour of a solution of potassiummanganese nickel cyanide is quite similar, and, as it might be expected, the corresponding solution of all the three metals concerned. But in the execution of this new process for separating manganese, nickel, and cobalt, certain definite proportions of the reagents to be used are absolutely necessary, as is also the attention to certain precautions, the full description of which will form the subject of a future communication. Berichte der Deutsch. Chem. Gesell., vol. xxiv., p. 3204.

PROCEEDINGS OF SOCIETIES.

CHEMICAL SOCIETY. November 19th, 1891.

SIR HENRY ROSCOE, F.R S., in the Chair. CERTIFICATES were read for the first time in favour of Messrs. Hugh Brown Collins, West Balgray, Glasgow ; Albert Henry Luckett, Brighton College, Brighton; James Alexander Schofield, University of Sydney, New South Hill Millar, 13, Waterloo Road, Wolverhampton; James Wales; Morris William Trewers, 2, Phillimore Gardens, Kensington, W.; Hugh Woods, 11, Archway Road, Highgate, N.

The following papers were read :

67. "Iron Carbonyl." By LUDWIG MOND, F.R.S., and Dr. LANGER.

The authors have isolated two compounds of iron with carbon monoxide, represented by the formulæ Fe(CO)5 and Fe2(CO), for which they propose the names ferropentacarbonyl and diferroheptacarbonyl.