NOTES ON WOLFRAMITE, BERAUNITE, AND AXINITE.By Edgar T. Wherry,Assistant Curator, Division of Mineralogy andPetrology, United States National Museum.OBSERVATIONS ON WOLFRAMITE.The studies herein presented were undertaken at the suggestionof Mr. Frank L. Hess, of the United States Geological Survey, whodesired information as to the composition of these two occurrencesof wolframite.No. 1 is a large specimen in the exhibition collection of the Museum(Cat. No. 80179), labeled "Cornwall, England," although unfortu-nately with no statement as to the exact mine or district from whichit came. It shows, however, the characteristic features of the min-eral as found in that region, occurring as long bladed crystals, withgranular chalcopyrite, in white vein quartz. Along cracks in thesolid wolframite a yellowish powder is developed, which apparentlyrepresents an alteration product, although the amount is too smallto determine its character.No. 2 is a mass about 6 by 9 by 12 cm. in size, showing a veryblack wolframite intimately associated with bright green chrysocolla,with here and there between the two, as well as throughout the wol-framite, streaks of an olive green, waxy, copper tungstate mineral.The associated gangue minerals are microcline and gray quartz, theoccurrence being evidently in a pegmatite vein. It came from CaveCreek, north of Phoenix, Arizona, and was presented to the Museumby Mr. S. H. Brockunier, through Mr. Hess (Cat. No. 87283)
.
Material for analysis was selected from portions free from visibleimpurities, finely powdered and dried in a dessicator over sulphuricacid. Preliminary trials of the method of distillation in a streamof sulphur monochloride mixed with chlorine, recently recommendedby Bourion,^ gave rather unsatisfactory results in that it was ex-tremely difficult to recover all the tungsten, some of it sticking tothe glass apparatus, and further, although most of the manganeseand all the silica, calcium, copper, and magnesium remained behindin the boat, some iron was also left, so that it was necessary to run
1 Ann. chim. phys., ser. 8, vol. 21, 1910, p. 87.Proceedings U. S. National Museum, Vol. 47?No. 2060. 501
502 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.47.both the residue and the distillate through the whole course ofanalysis.The method of fusion with sodium carbonate as outlined by Tread-well ^ was therefore adopted, although modifications had to beintroduced to obtain the copper in No. 2. The rare metals werealso separated by the course recommended by Treadwell, that ofdecomposition by nitro-hydrochloric acid. It was found that onwashing the residue insoluble in ammonia with pure water, theoxides of columbium and tantalum tended to become colloidal andrun through the filter paper, a turbid filtrate under these con<litionsconstituting a very delicate test for their presence; adding a littleamLmoiiium nitrate to the wash water coagulated them again, how-ever, and permitted their retention on the filter.In sample 1 the unusually high value of 0.80 was obtained forcalcium. As it seemed of interest to know whether this was due toincluded calcite gangue, or to scheelitc, half a gram was ignited for30 minutes in a platinum crucible at the highest temperature of amodern complete combustion burner, and tested with neutral phe-nolphthalein solution. Not the slightest alkalinity was shown,although calcite should have been readily decomposed under theseconditions. It was then thought that treatment with tartaric acidmight throw some light on the question. Half a gram of the samplewas heated on a steam bath with a solution of 10 grams of tartaricacid in 50 cc. of water for three hours, the undecomposed mineralfiltered out, and ammonimn hydroxide and ammonium oxalateadded to the filtrate. On standing overnight a precipitate of cal-cium oxalate came down, which on being filtered out, ignited, andweighed, gave 0.82 per cent, showing that all the calcium had goneinto solution. In the filtrate from the calcium oxalate ammoniumsulphide gave a black precipitate of ferrous sulphide, which wasfiltered out, dissolved in hydrochloric acid, and its iron precipitatedwith ammonium hydroxide and weighed as ferric oxide. A smallamount of manganese oxide, determined as 0.07 per cent, was alsorecovered here. The filtrate, from which iron, manganese, andcalcium had now been completely removed, was evaporated to dry-ness three times with concentrated nitric acid to destroy the tartaricacid, and the separated tungsten trioxide dissolved off the dish withammonium hydroxide, and poured through a filter into a cruciblein which it was ignited and weighed as tungsten trioxide. Theamount of tungsten trioxide obtained was not sufficient to satisfyall of the bases, but it seemed most probable that it had been unitedwith all of the calcium and part of the iron.To determine whether scheelite is soluble in tartaric acid, some clearcrystals were broken from a specimen from Zinnwald, powdered, and
> Quantitative Analysis, translated by W. T. HaU, 3d edition, 1911, p. 296.
NO. 2060. WOLFRAMITE, BERAVNITE, AND AXINITE?WHERRY. 5030.2250 gram heated on the steam bath with 10 grams tartaric aciddissolved in water, for two days. The amount of powder was seen togradually decrease, and that remaining at the end of thisperiod weighedonly 0.0302 gram, representing thus only about 15 per cent of the origi-nal material. This would no doubt have dissolved completely had itbeen finely enough powdered, or had the heating been continuedlonger, so it is evident that scheeUte is essentially soluble in tartaricacid. This reagent therefore can not be used to differentiate scheelitefrom calcite as an impurity in woKramite, but in the present case thefailure to obtain an alkaline reaction on ignition seems clear evidenceof the absence of calcite, so the calcium has all been regarded as in thescheelite form. As noted in the description of the mineral, a yellowishpowder is visible in some cracks on the specimen; although every-thing of this sort was carefully removed from the material used foranalysis, the presence of invisible cracks along which incipient altera-tion by calcium-bearing solutions has occurred is quite within therange of possibility.Tlie tartaric acid dissolved from the woKramite 4.16 per cent oftungsten trioxide. Tlie 0.82 per cent of calcium oxide found would beunited with only 3.40 per cent, if in the form of scheelite, so that 0.76per cent tungsten trioxide remains to be accounted for. But 0.90 percent ferric oxide and 0.07 manganous oxide were also found to bedissolved. The manganous oxide would take 0.23 tungsten triox-ide, leaving 0.53 which was united with ferrous oxide. The amount offerrous oxide corresponding would be, however, only 0.17 per cent(0.20 as ferric oxide), which, subtracted from the total ferric oxidefound, leaves 0.70 free ferric oxide wliich was dissolved by the tartaricacid. This, which is perhaps present as limonite, has been listedseparately in the analysis, the equivalent amount of feri'ous oxide,0.63 per cent, having been subtracted from the total ferrous oxidefound. But on calculating the mineral composition of the originalmaterial, following Mr. Hess's plan of assigning the tungsten first tothe manganese, calcium, magnesium, etc.,* a total excess of ferricoxide of 2.2 per cent was found. Since the form of the remauiing 1.5per cent of this is indeterminate, it has been necessary to include itwith the ferrous oxide, although it may be present as a ferric oxideinsoluble in tartaric acid, perhaps hematite, or possibly in solidsolution (see below). Mr. Hess ^ has found such an excess of iron tobe very frequent in wolframite.No. 2 was fused with sodium carbonate at as low a temperature aspossible, to prevent alloying of any of the copper with the platinumcrucible, and then, when the oxides insoluble in water were dissolvedin acid, the copper was first precipitated by hydrogen" sulphide beforethe determination of the iron and manganese, and weighed as oxide,
1 U. S. Geol. Surv., Bull. 583, 1914, p. 21. * Idem, p. 38.
504 PROCEEDINGS OF THE NATIONAL MUSEUM.after repeated ignition with dry ammonium carbonate. Here the cal-cium was only moderate in amount, but the cohimbium and tantalumoxides seemed sufficiently high to attempt to determine the relativeamounts of the two.For this purpose the oxides were fused with potassium acid fluoride,reprecipitated, and ignited as recommended by Foote and Langley ^and the specific gravity then determined. Although the impossi-bility of duplicating the conditions exactly renders the result uncer-tain, it is probably quite as dependable as that obtained by any of thedirect methods of determination. The 2.20 per cent of mixed oxidesobtained had a specific gravity of 7.02, corresponding to about 2/3tantalum, so that tlie percentages of the two oxides are stated as tan-talic oxide 1.50, columbic oxide 0.70. No tin or titanium could bedetected in either sample of wolframite.
Analyses.
NO. 2060. WOLFRAMITE, BERAUNITE, AND AXINITE?WHERRY. 505ferred to the use of distinct names for the end members because,while both are ambiguous in being ordinarily used for quite distinctthings, it seems simpler to restrict the chemical terms to one mean-ing. Thus, both ferrowolframite and ferberite may be used eitherfor pure ferrous tungstate, or for a ferrous tungstate containing a lit-tle manganese isomorphously replacing the iron ; manganowolframiteand hiibnerite for either pure manganese tungstate or manganesetungstate containing a few per cent of iron in isomorphous replace-ment; cuproscheelite and cuprotungstite for various mixtures of cop-per and calcium tungstates, etc.If mineralogy is ever to have anything like a quantitative systemof nomenclature, such ambiguity will have to be avoided, and itseems to the wi'iter that the best way to avoid it will be throughrestricting the use of words with chemical prefixes to the end mem-bers, whether found in nature or not, and doing away with distinctnames for them. The method used in the well-known quantitativeclassification of igneous rocks could be adopted here, using the pre-fixes: Permangano when Mn: Fe is greater than 7: 1; domanganobetween 7 : 1 and 5:3; ferromangano between 5 : 3 and 3:5; doferrobetween 3: 5 and 1:7; and perferro less than 1 : 7. Named accordingto this plan, the wolframite from Cornwall would be ferromangano-woKramite and that from Cave Creek doferrowolframite. Such namesare of course too cumbersome for everyday use, although it is possi-ble that in certain cases they might be employed with advantage forpurposes of classification or comparison. It is especially urged,however, that they, like those of end members, be omitted from listsof mineral names, for arbitrarily partitioned-off portions of isomor-phous series are not to be regarded as definite minerals.It is of course impossible to refer in the names to aU constituentsof the minerals, but that does not mean that some of the minor onesare not of considerable importance; and the significance of the colum-bium and tantalum found in these samples is certainly worth dis-cussing. In the table of calculated mineral compositions these havebeen regarded as united with iron and manganese, to form the colum-bite and tantalite molecules. No columbite and tantalite are presentas visible inclusions, for the brilliant cleavage surfaces of the wolfram-ites look perfectly uniform under the microscope. But they mightexist either 1, in chemical combination; 2, as submicroscopic inclu-sions; in one of the types of soUd solution: 3, isomorphous replace-ment; or 4, mix crystals; or, finally, 5, as an adsorption compound.The tendency of the " metallic acids"?columbium, tantalum, tita-nium, tungsten, vanadium, etc.?to enter minerals in these ways isvery evident when the variability in composition of many colum-bates, of the titaniferous magnetites, etc., is considered. But vari-ous interpretations have been put on it by differeiit observers.
506 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.47.Thus Dr. J. T. Singewald, jr./ concluded that the presence oftitanium in magnetites showing no visible ilmenite intergrowthsproved the existence of a " titanomagnetite " ; but he had found thelameUas of ilmenite to vary continuously from 4 mm. down to 0.001mm. in length, and there is no reason at all why they should ceaseto exist at just the latter size, for the limit of microscopic visibiHtyis determined by the wave length of hght, and has no significanceas far as the molecules are concerned, a particle of this size containingthousands of molecules. So submicroscopic inclusions may wellaccount for much of the titanium. Stopford Brunton,^ from similarstudies, assumed the titanium to be present in "sohd solution,"but evidently used this term to cover aU cases where no inclusionsare visible, making no attempt to differentiate the various possibiU-ties Usted above. It seems to the writer that we should be morespecific in stating just what mode of combination the evidencefavors in any particular case, so the data in regard to wolframitewill be further discussed here.The absence of the elements columbiimi and tantalum from somespecimens of the latter mineral, and their variable amount when foimd,clearly inchoate that they are not chemically combined with the tungs-tates. That then* presence might be due to inclusions of columbiteand tantalite has been suggested,'' without defhiite proof; and onlythe existence of visible inclusions would justify the assumption ofsubmicroscopic ones, as in the titaniferous magnetites above men-tioned. There remain, therefore, only the possibihties of sohd solu-tion and adsorption, 3, 4, and 5.A sohd solution, according to van't Hoff, the first to employ theterm, is a sohd homogeneous complex of two or more substances, therelative proportions of which may vary, but the homogeneous charac-ter be retained."* Two principal types can be distinguished, isomor-phous replacement, where the substances are closely related chemi-cally and may be regarded as taking one another's places in the pointsystem constituting the crystal structure; and mix-crystal forma-tion, where the substances are so different in character that mutualreplacement is out of the question. The formation of mix crystals,in which, as has been shown more especially by O. Lehmami,^ two ormore crystalline substances grow together so intimately that theyappear homogeneous under the microscope, yet mutually affect oneanother as to crystallization, producing changes in habit, crystalangle, optical properties, etc., is probably a much more frequent andimportant phenomenon than is ordinarily supposed. It takes the
1 Econ. Geol., vol. 8, 1913, p. 207; U. S. Bureau of Mines, Bull. 64,. 1913.2 Idem, p. 670.? Damour, Bull. soc. g^ol. France, ser. 1, vol. 2, 1848, p. 108; Hess, loc. cit.* Zeits. phys. Chemie, vol. 5, 1890, p. 323.
' Zeits. Kryst. Min., vol. 1, 1877, p. 453; vol. 6, 1882, p. 48, p. 580; vol. 12, 1887, p. 399, etc.
NO. 2060. WOLFRAMITE, BERAUNITE, AND AXINITE?WHERRY. 507place in crystalline substances that adsorption does in colloids, and,indeed, is probably the same effect. At least it is among the elementsthe compounds of which show the greatest tendency to take on thecolloidal form that mix crystal formation is most frequent.Tungsten and columbium are ahke only in that both can act asanions toward the more strongly electro-positive elements, and dif-ferent in practically every other chemical property?in valence,behavior with reagents, etc.?so that they can not be expected toreplace each other isomorphously. The most reasonable explanationof the condition of the columbium and tantalum oxides in wolframiteis, then, that they are present as mix crystals, or, in other words, thatthe several point systems have interpenetrated to such an extent thatthey have become at least pseudohomogeneous.^ Wolframite is ofcourse not susceptible to optical examination, nor can any crystalangle measurements be made on the present specimens, but if theycould be so studied effects such as those shown by other mix crystalswould no doubt be observed. Such intergrowth produces markedeffects on the crystal angles of the columbite group of minerals,which, as will be more fully explained elsewhere, are probably to beregarded as composed of mix crystals of a very few fundamentalcompounds; the principal ones, the compositions of which areR^CbjOg and R'TaaOg, are trimorphous, having isometric, tetragonal,and orthbrhombic forms; so that, even admitting that the analysesof members of this group are to some degree correct, the number of
''species" into which it should be separated is greatly overestimated.In summary, two wolframites have been analyzed and their compo-sitions discussed, a standard nomenclature for such isomorphous mix-tures recommended, and the columbium and tantalum oxides foundshown to most probably exist in mix crystal form.A NEW OCCimilENCE OF BERAUNITE.The rare iron phosphate beraunite was discovered near Hellertown,Northampton County, Pennsylvania, by the department of geologyof Lehigh University in 1911. The exact locality is the northeastcorner of an abandoned iron-ore pit, 1 mile southeast of the centerof the town. It was analyzed by Mr. J. S. Long, assistant in thedepartment of chemistry, and later more thoroughly studied by Mr.Louis H. Koch, assistant m mineralogy, as part of the work for hisdegree of master of science at Lehigh. Specimens were brought bythe writer to the United States National Museum (Cat. No. 87284),and further investigated and the combined results of all the workare here presented.iThe apparent excess of iron oxides, If not due to analytical errors, may be explained in the same way,which would be favored by their similar crystallization: Prior, Mineralogical Magazine, vol. 13, 1903, p. 217.
608 PROCEEDINGS OF THE NATIONAL MUSEUM. VOL. 47.The material shows a deep brown nodular crust up to 5 mm.thick, and fiat radiations up to 1 cm. in diameter, on the surface ofan iron-stained quartzite. In some specimens it bears implantedglobules of psilomelane, yellow needles of cacoxenite, and powderyclay. Internally the crusts are coarsely fibrous, -with, the fibersperpendicular to the surface, and the material was evidently origi-nally a gelatinous precipitate, which has become hard and crystallinein place, thus representing a "meta-coUoid," as defined by thewriter in a recent paper.*After a preliminary analysis (1) had shown the general nature ofthe mineral, material for further study was obtained by crushingselected fragments, and the powder looked completely crystallineand homogeneous under the microscope, except for the presence ofa trace of clay or fine sand. Standard methods of analysis wereused, the iron being determined with permanganate, the manganeseby the bismuthate and the Ford methods, and the phosphorus weighedas magnesium pyrophosphate after separation by molybdate. Asmall amount of water?less than 0.5 per cent?was given off below100?, and the alialyses were made on coarsely powdered materialdried at this temperature, for the fineness of the grinding and thehumidity of the air were found to have a distinct effect on the watercontent. Table of analyses.
NO. 2060. WOLFRAMITE, BERAUNITE, AND AXINITE?WHERRY. 509ever, with beraunite. The specific gravity, determined by a pycno-meter, varied from 2.850 to 2.920; Dana gives 2.95. The indicesof refraction were found to be approximately a and /?=1.78, ^=1.81,extinction straight (elongated parallel to b.), and sign of elonga-tion? . No optical data have been published for beraunite, but Dr.E. S. Larsen, of the United States Geological Survey, kindly exam-ined for comparison a specimen of the variety "eleonorite" (U.S.N.M.Cat. No. 80622) from the type-locality at Giessen, Germany, andobtained the values: ?' = 1.775, /?= 1.786, ;-= 1.815, which are essen-tially the same as those above given.Two possible explanations of the variation in composition shownsuggest themselves. The first, that the law of definite proportionsdoes not hold in this, and other iron phosphates, would be capableof introducing a rather chaotic condition into mineral chemistry,and seems entii-ely improbable. If the mineral were a colloid, bow-ever, the results could at once be interpreted in a second way?thatthe mineral is an adsorption compound of ferric and manganichydroxides with phosphoric oxide and water. It is, however, ameta-coUoid, that is, a colloid which has become crystalline withoutdissolving or losing its soHdity. If, when this crystallization tookplace, the adsorbed constituents united as well as they could intodefinite compounds which formed mix crystals (or, as it is oftencalled, solid solution), the results obtained could easily be accountedfor.The RjOg-.PjOg ratio shown here varies from 1.72:1 in No. 2 to1.93:1 in No. 1, whUe that indicated by the best previous berauniteanalyses is 1.50:1, and the best dufrenite analyses, 2.00:1. Thesimplest explanation of the variation and indefinite ratios shownby the specimens under investigation is that they represent mixcrystals (solid solutions) of these two fundamental compounds. Thedufi-enite molecule is in excess over the beraunite, although theproperties are those of the latter mineral; it might therefore be calleddufreniberaunite.AXINITE FROM DELAWARE COUNTY, PENNSYLVANIA.In the pegmatite cutting the granite gneiss of the Leiper Quarry atAvondale, Delaware County, Pennsylvania, an occurrence of the min-eral axinite has been discovered, certain features of which are sounusual that it has seemed worth while to make it the subject ofspecial study and description. It was in fact at first supposed to be anew mineral, and its true character was only recognized toward theend of the investigation.Specimen 1 (U.S.N.M. Cat. No. 87232) consists of three small frag-ments, showing a columnar mineral with a resinous luster, of a paleyellow to salmon pink color, associated with pink microchne, granular
510 PROCEEDINGS OF TEE NATIONAL MUSEUM. tol. 47.quartz, and muscovite. It was found at the locality in 1904, andanalyzed by the writer when beginning the study of quantitativeanalysis. The results, given in column la in the table below, showedit to have a composition distinctly different from that of any pre-viously known mineral, although, because of lack of experience, it isprobable that they are not altogether accurate, some silica having nodoubt been weighed with the sesquioxides, and some manganese withthe calcium or magnesium ; but the boric acid, having been determinedby the writer's volumetric method,^ is probably exact. The matterwas then laid aside because of the pressure of other work, but whenthe writer became connected with the Museum, it was taken up again.To obtain a check on the previous analysis, all the mineral that couldstill be broken from the specimens, amounting to less than half agram, was sent to Mr. J. E. Whitfield, of the firm of Booth, Garrett& Blair, of Philadelphia, who has been doing considerable analy-tical work for the Museum, and his results are given in column 1&.Meanwhile a mass of yellow plates, about 4 by 2 by 1 cm. in size,was found at the quarry by Mr. J. Watts Mercur, jr., of WalUngford,Pennsylvania. (U.S.N.M. Cat. No. 87233). Although showing noplanes defuiite enough for crystallographic measurement, this ma-terial had the aspect of axinite, and the same specific gravity,3.250. A clear fragment was sent to Mr. Whitfield for analyis, andhe obtained the results in column 2a; but as the boric acid seemedrather low, probably owing to incomplete decomposition, two deter-nmiations were made by the writer, and 5.98 and 6.09, average 6.04,obtained (column 26). Using this value, the agreement with thetheory for axinite is so close as to leave no doubt that it is the mineralrepresented.The question as to whether No. 1 was a new mineral, or only an im-pure axinite, remained unsolved, so powder from both specimens wassubmitted to microscopic examination. No. 2 showed a mean index ofrefraction of 1.680, birefringence 0.008, and sign ? , thus agreeingwith typical axinite. The greater part of No. 1 gave essentially thesame values, but scattered through this material could be seenpinkish, pleochroic grains with a much higher index, 1.700, but stilllower birefringence, 0.005, showing, in fact, ultrablue interferencecolors, also extinguishing straight and + in sign. These propertiesidentified it as zoisite, a mineral which had previously been reportedfrom the locaUty.^ Here, then, was the explanation of the differencebetween analyses la and Ih, as well as their deviation from the theoryfor axinite: the material is not homogeneous, but contains a variable
1 Joum. Amer. Chem. Soc, vol 30, 1908, p. 1687.2 Cardeza, Proc. Acad. Nat. Sci. Phila., 1892, p. 194; discovered by Miss Mary S. Holmes. A small speci-men of this mineral, not associated with axinite, collected by the writer about 1906, has also been addedto the Museum collection (Cat. No. 87234).
NO. 2060. WOLFRAMITE, BERAUNITE, AND AXINITE?WHERRY. 511amount of intergrowii zoisite. And the theoretical composition ofa mixture of 80 per cent axinite with 20 per cent zoisite, given in thelast colunm, corresponds, in a general way, with the analytical results.Table of analyses.