The American Mineralogist, Yolume60, pages 840448, 1975

Primary NontroniteFrom the VenezuelanGuayana

W. C. ISpHoRpllrC

Department of Geology and Geography, Uniuersity of South Alabama Mobile, Alabama 36688 Abstract

Nontronite, the -rich end member of SmectiteGroup clays, has previouslybeen reported as a hydrothermalalteration product, as a constituentof soil clays formed by weatheringof a variety of rock types,and as an alterationproduct of volcanicrocks. This paper describesthe chemical,optical, and X-ray propertiesof a new occurrenceof the mineralwhere its origin is apparentlyrelated to directcrystallization from high temperature waters.The mineral is further unique in that it containsonly a very small percentageof aluminaand therebyclosely approaches the previouslyunreported wholly iron end-member. Significantdifferences were found in the optical propertiesof the Venezuelannontronit€ when comparedwith those of an earlier study, especiallywith regard to the variation in refractiveindex with Fe2Oscontent. It now appearsthat this variationhas been over-simpli- fied and may not be as orderlyas was previouslybelieved. The variouslines of evidencethat argue againstan origin by either hydrothermalalteration or chemicalweathering are dis- cussed,as well as thosesupporting an origin by direct crystallizationfrom high temperature watersof probablemagmatic or metamorphicorigin.

lntroduction product in gneissesand slates.At the other sites,non- Nontroniteisa Smectire Group which. ff,T'Lffi'"fff"#:j""1|# ,XXT::""."t1:j'# occurs as the iron-rich end-memberin the d.ioc- rr-;i",;-be an arterationproduct. At the type tahedral sub-group -beidellite- io.uri,v,Nontron, France, Berthier (1g27) descri6ed nontroniteseries' It haspreviously been *p":l:9^:t^1 ,tr" ri".r"r-: as occurringas "onion-like" massesas- constituentof sedimentaryrocks, as the productof -:-,--; -,,, alterationorvolcanic grasses andas a hydrother.mal alterationproduct. This paper describes the chemical, il:Tfl:iil;ilt"1T'ifi::"":'."!'j#fiit"fjo.ll: ;d; ;;ntronit. in alteredbasalts of the Columbia optical, and X-ray propertiesof a specimenfrom iii;; .;gi;" and attributed it to the weatheringof recently discovered new occurrencein southern" L"r"r,i. Ei"rs, palagonite,iddingsite, and augiteun- venezuela'Here the nontronite is apparently.un- orr.""olii"ns of poor drainage.similarly, Sherman related to or alteration of previous ii describednontronite in weatheredbasalts minerals and has formed by crystallizationfrom "i(lgZal from Australia,New Zealand,Fiji, and waters of hydrothermal or possibly metamorphit H";;i;;Jre"na "na"ri,"s it wasreported to haveformed under a origin.The chemistryof the mineralfrom thisloca- varlety.. _,^:-.oI^;.,. cllmatlcconqluons' oy cnemlcalweatner- tion is also unusualin that it contains essentiallvno rng' alumina and approachesthe compositionoi ttr" previouslyunreportedtheoreticalendmember. venezuelanoccurrence The mineral describedin this paper was found 2 Geologicaloccurrence km southof the OrinocoRiver on FederalHighway Rossand Hendricks(1945) in their detailedtreat- 19, l5 kilometersnorth of the turnoffto Guri Dam. ment of montmorilloniteclays listed l3 locations Hereit occurredas an intrusionwithin granitic rocks wherenontronite had previouslybeen described. Of of the PrecambrianImataca Complex (see map, Fig. these,7 werein the United States,3 in the Soviet l). Themain intrusion, where exposed in a newroad- Union,2 in Germany,and I in Hungary.At 8 of the cut, is approximately2 metersthick and 372meters l3 locations, the mineral was associatedwith high, Numeroussmaller veinlets could be seenex- metamorphicrocks and occurred as an alteration tending outward into the surroundinggranites. The

840 PRIMARY NONTRONITE FROM VENEZUELA 841

dioctahedraliron end-memberof the beidellite- nontroniteseries was given as: (Si?.ssAlo.6zXFeo)Orr(OH)a but analysescorresponding to this chemicalformula had not been reported. The type mineral from Nontron,France, was included with analysesfrom an additionall3 samplesthat werejudged to containan "iron-bearing"impurity, based on the lact that the chemicalanalyses showed these to containmore iron thancould be accommodatedin the octahedralsites. 36pure" The 13 samplesare shownin TableI along with one analysisreported by Kerr and Pill (1950); sample115 is the Venezuelannontronite described in this paper.Analysis ll6 is the "type" mineralfrom Nontron, France,and is includedfor comparative purposes. The most striking difference between the Venezuelannontronite and the otherslisted in Table I is that of aluminacontent. The other published analysesshow a rangein AlzO, from 2.94to 17.65 percent,with a mean value of 9.55 percent.The Venezuelannontronite. with an aluminacontent of Ftc. t. Map showing Venezuelan nontronite locality and outcrop only 0.27 percent, contains one-tenththat of the of Precambrian and younger rocks in study area. lowestpreviously reported analysis. The significance of this low aluminacontent has bearing not only on the existenceof a completesolid solution series nontronitewas a uniform,yellow-green (l0Y 6/6), ex- tendingfrom beidellite,(Si? densemass containingoccasional thin (2-5 mm) ,'Alo.6?XAln)Oro(OH)n,to nontronite,but on theorigin of thisunique deposit stringersof specularhematite. Though not chemical- as well. ly analyzed,a similar, but smaller,exposure of the The existenceof a completesolid solutionseries samematerial was also seen 25 km to the southeastin in- cluding montmorilloniteand beidellitewas first the open pit iron mine of the BethlehemSteel Cor- proposedby (1877) porationat El Pao. Collins andhas since been largely confirmedin severallater reports (see Gruner, 1935; Laboratory Analysis Rossand Hendricks,1945). The existenceof a solid Chemistry solution seriesextending from beidelliteto the iron- Ross and Hendricks(1945) listed 13 reported rich end member,nontronite, is a later development analysesof samplesconsidered to representrelatively and wasproposed by Rossand Shannon(1925) and purenontronite (those containing down to two atoms Larsenand Steiger(1928) and, similarly,Ross and of iron per unit cell layer).The compositionof the Hendricksdid much to verify its existence.Thev dis-

Tnsle l. Chemical Analyses (in percent) of Nontronitel

Nwber: #I #2 #3 #lo #t1 *11

49,75 43.51 49.50 46.06 40.72 40,48 +o.c) 19.54 42.74 40.42 41.t8 41.88 40.O2 40.54 49.90 4A.A2 10.06 2,94 ).6.1A 12.22 4.96 6.oo Lt,o) 6.o4 9.76 ro.42 9.84 11.90 rO.55 5.r9 o.27 4.to 22.06 29,61 11.08 18.82 10.24 29.59 r4.39 10.32 26.7t 25.76 27.47 26.2o 27.53 31.61 o.25 0.05 0.61 ).,62 o.74 0. 1 tr tr 1.06 O.It tr 0.10 0.06 1 ?C n i6 2.6e 2.22 2.57 r.66 r.98 2.90 r,21 o.O4 \.94 o.3a tr 0.67 I.92 o.77 o.55 O.32 tr O.5? O.zt+ o.01 o.r5 u,z+ trr v,)z u.I4 a.L2 22.27 17.16 21.80 2r.35 l].45 2t,81 20,75 16.12 9 .66 t pilt Samples 1-13 from Rossand Hendicks (1945); Sample t4 from Kerr and (1950); Sample 15 lromVenezuelanlocality; Sample 16 from Nonton, France. W. C. ISPHORDING cussed,at length,the thenavailable information on probably discounted these, believing that the higher thechemical variation among nontronites, but admit- MgO percentages(1.45 percent averageversus 0.37 ted that wider variationmight exist.Evidence was percent in the uncontaminated samples)could be at- present,they believed,to supporttheir contention tributed to MgO in mixed-layerclays. Basedon the that a solid-solutionseries existed with limited-to- samples available for their analysis, they cannot be extensivesubstitution of aluminumfor siliconin the faulted for this logic. However, the Venezuelannon- tetrahedralsites, and of iron for aluminumin the oc- tronite does not support their conclusion. Though tahedralsites. The chemicalanalyses carried out on containing significantly lower octahedral and the Venezuelannontronite confirm this belief.Thus, tetrahedralalumina than any of the samplesused by unlikethe wholly iron end-membersof both the en- Ross and Hendricks, the MgO content is nearly five statite-ferrosiliteand -anniteseries, the times higher than that found in their analyses.X-ray iron-rich end-memberof the beidellite-nontronitediffraction studies on the Venezuelan material failed seriesis stableand canform duringnatural geologic to disclose any reflections from mixed-layer clays; processes.The extremelylow aluminumsubstitution henceit is concluded that the is presentin in the octahedraland tetrahedralsites results in an octahedral coordination in the nontronite and that oxide chemicalformula for this mineral ol the correlation of low alumina content with low magnesiumwas more apparent than real. 4FerOr.MgO.lgsior.2lHrO (Alros),CaO, NazO or, in termsof octahedral-tetrahedralsite occupancy: Optical Analysis With regard to the optical properties of mont- Si? Fer.urrMgo.nr2Alo.oo6020(OH)1 erAlo.o6 morillonites, Ross and Hendricks (19a5)noted that The very limited substitution of aluminum for these can be lessdefinitely determined than those for, creates only a slightly unbalanced charge in possibly,any other mineral group. This resultsfrom a the tetrahedral layer, Hence the presence of small number of factors, among which are: (l) the lack of amounts of sodium (0.02Vo)and (0.77Vo)in observablecrystals, (2) the tendency of theseminerals the interlayer sites is largely due to excess charges to absorb immersion oils and, hence, undergo a arising in the octahedral layer caused by the undeter- changein refractiveindex, (3) the stacking order of mined amount of iron presentas Fe'+. The cation ex- the individual sheets, (4) their susceptibility to changecapacity was predictably low, therefore, when chemical weathering, and (5) the finely crystalline compared with other SmectiteGroup clays, with a nature of theseminerals in generalcoupled with their value of 48.9 m.e.q./100 grams soil obtained using tendency to occur as micro-crystalline aggregates. the ammonia extraction procedure. Using samplesof the Venezuelannontronite that The information derived from the chemical had been crushed and then dis-aggregatedin an ultra- analysis of the nontronite has also forced a re- sonic separator,optical measurementswere made on examination of some generalizationsmade by Ross finely dispersed dried films, using the method and Hendricks ( 1945)in their exhaustivestudy of the describedby Grim (1934),and on individual crystals SmectiteGroup minerals.For example,a major con- in the 44-62micro-size range. The nontronite proved tradiction was found to exist in the interpretation of to be biaxial (-) with a : 1.554,0 : 1.590,I : optical data (discussedin the next section) and with 1.594,all t0.002. The 2V anglewas measuredas 35o, regard to statements concerning the general using a micrometer ocular. Grains yielding centered chemistryof nontronites.Ross and Hendricks noted biaxial (-) interferencefigures were non-pleochroic that (1945,p. 45) ". . . nontronitescontain relatively but varied in color from pale brown to yellowish small amounts of magnesium. This relationship is brown. particularly true for the specimens having chiefly The optical information obtained from the Fe+3in octahedralcoordination. . . . of the four sam- Venezuelannontronite does not agree with certain ples containing less than 0.12 Al+3 in octahedral observationsmade by Ross and Hendricks, especial- coordination, the highest was 0.04 Mg+2." Though ly with regardto variation of the a and 7 indiceswith this generalization does hold for the l3 nontronites FezOa content. Their chart (see Fig. 2), since consideredby them as "pure" samples,it is not war- reproduced in several standard references dealing rantedif the MgO content of the l3 other samplesin- with the optical properties of smectiteclay minerals cluded in their paper is considered(those containing (seeHeinrich, 1964;Deer, Howie, and Zussman,Vol. an iron-bearing impurity). Ross and Hendricks 3, 1962), predicts for the Venezuelan nontronite a PRIMARY NONTRONITE FROM VENEZUELA 843

all the optical and chemical information listed in Ross and Hendricks' article, along with other, z recent, published nontronites. The o more data on FI U data points shown in Figure 3 include the 7 analy- e ses used by Ross and Hendricks to construct their

4 curves but also included are 15 analysesomitted by (l) f,r them because either the samples contained ex- cessiron and were considered"impure" or (2) the data did not fit the a and'y curvesbecause of low iron content.The writer has includeddata points from the "impure" samples because Ross and Hendricks noted that the amount of excessiron in thesesamples was generally very low and becauseseveral of the samplesyielded refractive indices not greatly different VoFerO" from the Venezuelannontronite. Seven of the 15sam-

Ftc 2. Ross and Hendricks' (1945) chart showing variation in pfes plotted are those excluded by Ross and FerOr content with a and "y indices of refraction. Hendricks becauseof low iron content.They are in- cludedhere because, since Ross and Hendricksused 5 such samples to derive their curves, it would seem value of approximately 15 percent FerO3from its 7 reasonableto include all such samplesavailable to index and of 11.5percent FerO3 using a. Thesediffer determine the actual curve configuration. significantlyfrom the 29.5percent obtained using the Examination of Figure 3 immediatelysuggests that atomic absorption spectrophotometer.Because of the use of refractive index to predict ferric iron con- this unexpecteddifference, an examinationwas made tent in nontronite will. at best.be hazardousand will of the basisand rationalefor the constructionof Ross frequently yield only a crude estimateof the actual and Hendricks' chart. These are summarizedbelow: percentage.Considerably more variation exists in (l) The curvesare basedon optical and chemical iron content versus refractive index than is suggested data from 7 sampleswhose ferric iron content by the smooth curves published by Ross and ranged from 0.06 to 27.47 percent.Only 2 of Hendricks.Their assumptionthat the ferric iron con- the samples were true nontronites; the tent increaseswith refractive index is, roughly, valid remainder were members of the montmoril- up to approximately 28 percentferric iron. Past this lonite-beidelliteseries. value, there is a visible trend toward a decreasein in- (2) The curves represent the "best fit" of the 7 dex of refractionwith increasein iron. In light of the analysesand were simply extended beyond the largeamount of scatterseen between iron contentsof ferric iron content of the last analysis (27.47 0 to l0 percentand 28 to 32 percent,the writer ques- percent)to 40 percent,based on the beliefthat tions whether any significant conclusions are war- the refractiveindex would continue to increase ranted from refractive index data. Even in the central with FerOecontent. portion of the curves (between l0 and 28 percent Ross and Hendricks noted (p. 55) that, from the iron), the curvesare basedon only three data points standpoint of variation in refractiveindex with iron and future analyses could, conceivably, alter the content, those data points plotted show ". . . a very smoothnessof this portion of the a and 7 curves.Un- consistent relation." They further state, however, til such additional data is forthcoming, the published (p. 55) that ". . . a number of indices of low-iron curves of Ross and Hendricks should be used with members given in (their) Table 14 fail to fit this advised caution. curve." They attributed this to several factors but believed that in high and moderately high iron X-ray Diflraction Analysis members that the iron content was the dominant Only a limited amount of data has beenpublished factor controlling refractive index, whereas in low dealing with the X-ray analysis of nontronite. iron members other factors did so (e.g., aluminum: Gruner (1935) determined the average dimensions silicon ratio, water content, etc). To determine the of the nontronite unit cell asao = 5.23A, bo : 9.06A total actual amount of variation that exists in the based on the analysis of six samples.Since the 0 beidellite-nontronite series,it was decided to plot crystallographicangle was unknown, the co dimen- 844 W, C ISPHORDING

to't'

--llgbe-

% ter0,

Frc. 3 Variation in a and ry indices of refraction with FezOecontent. Solid circles are samples used by Ross and Hendricks (1945); crosses are from other samples for which optical and chemical data was available. sion was not calculated; Gruner did list the du di- the aodimension for the Venezuelannontronite was mension as 12.4 - 12.7 A, however, depending on calculatedas 9.1 l7 v3 = 5.264A. tne dozw?svariable the amount of water present in the inter-layer posi- but averaged14.80 A. Treatmentwith glycerine tions. Nagelschmidt (1938) examined nontronite causedthe du peak to shift (expand)to approxi- clays from the type location (Nontron, France) and mately 19.4A (seeFig. 4). Heatingto 300oCfor 7 from Behenjy, Madagascar, and calculated the aoand hours collapsedthe peak to 14.3A; further collapse bo dimensionsfor both as 5.23 and 9,11 A, respec- to 9.6 A occurredwhen the samplewas heatedto tively. The dp dimension for these two samples was 650'C for l0 hours.Total collapseof the latticewas somewhat larger than that determined by Gruner effectedby heatingto 1,000'Cfor 4 hours and, as and was given as 15.6A for the specimenfrom Mad- seenon Figure 4, only cristobalitepeaks are visible agascar and 15.4 A for the sample from France. on the diffractogram.Table 2 lists a comparisonof Since nontronite closely approximates an orthohexa- powder cameradata from Gruner (1935),Nagle- gonal lattice, the do, can be considered as a very schmidt(1938), and for the Venezuelannontronite. close approximation of the codimension. The sample from Venezuelaproduced results comparable to those Origin of Nagleschmidt(1938) in that the bodimension was found to be 9.117A. ttris dimensionwas determined GeneralDiscussion using the equation of Brindley and MacEwan (1953) Montmorilloniteclays, as a group,have been most for dioctahedral clays for which D(A) : 8.92 + frequently attributed to the weatheringof volcanic 0.06x * 0.094 + 0.18r * 0.27s(where p, q, r, and s detritus. They have also been describedas the are the numbers of Al8+, Fe8+,Mg2+, and Fe2+ions products of the alterationof ,, per half cell layer in octahedral coordination and x chlorite, serpentine,and hornblendein a variety of is the number Al3+ ions in tetrahedralcoordination). rocks and, as such,are often found as the principal Since the cell is essentially orthohexagonal, Brown constituentsof somesoil clays(see Weaver, 1958). (1961) states that the a:b ratio is l:1/3. Therefore More recently,Jeans (1968) has shownthat many PRIMARY NONTRONITE FROM VENEZUELA 84s

Ftc. 4 X-ray diffractograms for untreated nontronite, glycerated sample, and samplesheated to 300, 650 and 1,000"C(Copper Ka radiation). may actually be the products of tion products of volcanic glassin the basaltflows of direct crystallization in marine waters of normal the Columbia River Plateau. Ross and Hendricks salinity. Ross and Hendricks (1945) discussedthe (1945)noted that a number of samplesof nontronite origin of vein and gouge clays of the montmoril- have been reported as forming veins of unknown lonite-beidellite series and attributed them to the origin in the schistsof the southern Piedmont. alteration prior of aluminous minerals (feldspar, Becausethe Venezuelannontronite is exposedon tourmaline, spodumene,etc) by late stagehydrother- the surface, an origin by chemical weathering was mal fluids and by crystallization from low considered.Since it is also an obvious intrusion into temperaturewaters. Nontronite clays were reported the surroundinghost-rock granites,an origin involv- by Allen and Scheid (1946) to have formed as altera- ing either hydrothermal alteration or direct crystal- lization from high temperaturewaters was also in- TAsr-s 2. Diffraction Data for Nontronite Samples vestigated.

Venezuelan Behenuy, Nontron, Faratsiho. Origin by Weathering Nontronite Madagascar* France Madagascar** o -; AI 8- r I r 8"- Genesis of the Venezuelan nontronite by surficial 14.80 VS vs t? o VS weatheringcan be quickly 4.ro VS 7.r discquntedfor a number 4.27 1.11 1.11 4.44 s 2.9e of reasons.First, the surrounding granitic rocks into vs 2.79 which the nontronite is intruded show only slight z.or 2.56 2,59 w-s evidenceof chemicalweathering, and this only in the 1 .81 r.72 z.+) 2.5I 1 A7 S r-72 2)a upper half meter of the recentlyexposed highway cut. r.72 Even if selective chemical weathering is postulated r.5a S r.3z 1,.52 a.06 for the primary vein minerals,the normally expected L.37 w s I.32 f.40 w s r-52 weathering product for this warm humid region L.31 *(z ) r.27 r.zo L.29 w(?) t,45 would be , accompaniedby sesquioxidesof L.25 w(?) 1.37. alumina and iron. If the parent rock were presumed an iron-rich serpentine (rather gabbro x,Data fron Nagelschridt (t9l8) than or *x Data fron cnrner (19f5) diabase), other problems are encountered. Not- Venezuelan specimen analyzed using 57.3 millimeter diameter withstanding the fact that no such rocks have De bye-Scherrer camera, nickel-filte red, CuKa radiation. been previously reported in this area, the weathering 846 W. C ISPHORDING chemistry of serpentine would present difficulties. Wildman, Whittig, and Jackson(1971) have shown that within the usualrange of pH and low activitiesof Fe3+, Al3+, AlO2-, and Mg'+ encountered in soil matrix solutions,both Fe(OH)' and AI(OH)' will be stable in the octahedral layers. If silica were present in sufficientconcentration in the matrix solution, the formation of a layer silicatewould provide a lower energy state for the iron and aluminum ions than is provided by their hydroxides.The resultwould be the formation of montmorillonite, rather than non- tronite. This is causedby the fact that in the pH range of tropical soils,the percentageof alumina and iron remains essentiallyunchanged (in the pH range 4.5- 9.0) because of their similar low solubilities (see Loughnan, 1969). If, in order to explain the low alumina content,a peridotiteis assumedas the parent rock, then the end products of weathering,according to Loughnan (1969), would be iron oxides and (the magnesiumanalog of nontronite). If an iron-rich peridotiteor picritic diabasewere the parent rock, then weatheringcould resultin the formation of nontronite.Such an occurrencehas been described by Sherman et al (1962) but only under conditions of low rainfall (15 inches per year). Even under such conditions, however, the mineral was reported to form only coatingssurrounding unweatheredolivine grains in picrite . Again, it should be pointed out that although diabase, meta-gabbros, and Frc. 5 Electronphotomicrographs of Venezuelannontronite sam- amphibolites are present throughout the Imataca ples.a. x 1000;b. X 3000. Complex; no serpentines,, nor peridotites have been reported (see de Ratmiroff, 1964; Chase, of resistantminerals such as anatase, spinel, leucox- 1963);hence, these are unlikely to have served as the ene, ilmenite, apatite, sphene,zircon, garnet,etc, parent rock. shouldhave survived the alterationprocess and be An additional line of evidencethat arguesagainst present.These have beenreported (Chase, 1963) as an origin for the nontronite by the chemicalaltera- common constituentsof the igneousrocks of the tion of any previous mineral is the high purity of the ImatacaComplex. Their completeabsence is difficult material. Optical and X-ray diffraction analysesdis- to explainif the nontroniteoriginated by simple closeno "unaltered" mineralsthat could haveserved chemicalweathering. Finally, if the nontronitewere as a source for the nontronite, and electron the product of the weatheringand alterationof photomicrographs at 1000X and 3000X similarly previousrock-forming minerals of anytype, at leasta confirmed the material to be pure and essentially smallquantity of eitherkaolinite or mixed-layerclays homogeneous(see Fig. 5). Further, crushedsamples (or both) would be expectedto appearon the diffrac- of the nontronite were placed in acetylene tograms.Instead,these were found to be completely tetrabromide (sp. gr. : 2.96) to determine what lacking. In an occurrenceof nontronitereported by heavyminerals, other than visually observedspecular Hosking(1957), where the mineralwas present as a hematite,might be present,with the belief that these weatheringproduct ofred-brown earths derived from might give evidenceas to the identity of some possi- hornblendegranite, the samplewas found to consist ble parent rock type. None were found to be present chieflyof ,with nontronitepresent only in in any of the four crushed samples analyzed. If the minor amounts.Allen and Scheid(1946), in their dis- nontronite were, in fact, the product of the alteration cussion of nontronite veins resulting from the of either acid, intermediate,basic, or ultrabasic ig- weatheringof basalt,presented photomicrographs neousrocks, it would be expectedthat small amounts that showediddingsite, augite, plagioclase, volcanic PRIMARY NONTRONITE FROM VENEZUELA 847

glass, and ilmenite partially altered to nontronite. In mineralsor residualheavy minerals, some of which many casesthe original mineralsremained unaltered should survive any alterationprocess. Regardless and were clearly discernible.Hence, the lack of any whetherthe parent rocks were acid, basic,or ultra- kaolinite, halloysite,or mixed layerclays in the Vene- basicin composition,some remnant grains should be zuelan nontronite, coupled with the completelack of found. Partially alteredgrains were reportedto be remnant minerals or residual heavy minerals, is widespreadin the nontronite clays found within strong evidence against its origin by chemical basaltsin the Columbia River region(see Allen and weathering. Scheid,1946). Heavy minerals such as rutile, spinel, ilmenite,and ziron alsoare resistantto the attack Origin by Hydrothermal Alteratton of hydrothermalsolutions and would be expected to re- Some of the argumentsraised in opposition to an main essentiallyunaffected. origin by chemical weathering apply also to one call- ing upon hydrothermal alteration of primary vein Primary Origin for the VenezuelanNontronite minerals. Again, the difficulty exists in finding a The unusuallyhigh iron, low alumina,low afkali suitableparent rock that could give rise to the highly compositionof the Venezuelannontronite, coupled pure nontronite, even assuming an open ther- with the lack of otherassociated clay minerals, rem- modynamic systemwith addition and,/or removal of nant minerals,and resistantheavy minerals, suggests mobile ions. Grim (1953) states that, in many or that it may be the product of direct crystallization perhapsmost instances,the hydrothermal alteration from siliceousfluids rich in dissolvediron. Chase productsconsist of a mixture of severalclay minerals, (1963)has noted that all the rocks in the Imataca with mixed-layer clays invariably present.As men- Complex,except the diabasedikes, have been sub- tioned under the discussionof chemical weathering, jected to high grade regionalmetamorphism, rang- other clay minerals were conspicuouslyabsent in ing from low amphibolite to the granulite facies. diffractograms of the Venezuelan nontronite. The iron formations, trondhjemites,gneisses, and Another feature commonly exhibited by clays amphibolitesof the ImatacaComplex underwent a formed by hydrothermal alteration, which is also major metamorphicevent approximately 2,000 m.y. lacking in the Venezuelan occurrence, is the typical ago and weresubsequently intruded by largegranite zonal arrangement of the constituent clays. Clays plutonsand laterby diabasedikes. The diabase dikes formed by hydrothermal alteration commonly dis- arepost-tectonic and have not beenmetamorphosed; play a central core of fine-grained white the ageof thesepost metamorphicintrusions is un- (sericite)which passesoutward to kaolinite and/or known;however, igneous events are knownto have montmorillonite, followed by an outer zonemade up takenplace in the areaapproximately 1,500 m.y. ago largely of chlorite. No zonal arrangementwas found and during an interval from 800-1,070m.y. ago, in the Venezuelanintrusion, and samplestaken near based on radiogenic ages from granites. Chase the contact with the host rock were identical, (1963) notes that these dates were obtainedus- mineralogically,to thosein the centerportions of the ing only one methodand are, therefore,not reliable vein. indicatorsof either metamorphicor igneousevents. A further argumentagainst an origin by hydrother- Regardless,the presenceof the ubiquitoussiliceous mal alteration involvesthe extremelylow alkali con- iron formationsthroughout the ImatacaComplex tent of the nontronite. Krauskopf (1967)notes that, testifiesto the availabilityof both a sourceof iron in addition to any metals and sulfur that might be and silica.Though probably magmatic in origin,the present,hydrothermal waters would be expectedto watersneeded to dissolvethese ions may be relatedto contain silica,the cationsNa+, K+, Ca2+,Mg2+, and Precambrianmetamorphic events that occurredin anions such as Cl , SO42-,HCO'-, erc. Becauseof the region. It has long been establishedthat high their mobility in hydrothermal fluids, particularly temperaturewater with its high dielectricconstant those containing chlorides,Na+, K+, and Ca2+ions can react with and dissolvea wide range of ionic commonly are found in the inter-layerexchange posi- crystallineminerals. Following dissolution, a change tions in hydrothermal montmorillonites. Their near in envirdnmentalconditions (reduction in tempera- complete absencein the Venezuelan nontronite is ture, changein pH, etc) may reprecipitatehydrated or conspicuous and argues against its formation by hydroxyl-bearingminerals. Past studieshave shown hydrothermal alteration (seeTable l). that montmorillonitecan form, hydrothermally,and A final argument againsthydrothermal alteration existup to temperaturesof 450oC(Ellis, 1967). Ewell again concerns the lack of any remnant primary and Insley(1935) synthesized nontronite by heating 848 W. C. ISPHORDING

FezOs-2SiOzgels for 6 days in a sealedbomb and Dnnn,W. A., R. A. Howte, lNo J. Zussrrl,lN(1962) Rock-Forming noted that it was stableup to temperaturesof 340o- Minerals,vol. 3, John Wiley and Sons'Inc., New York' A. J. (1967)The chemistryof someexplored geothermal 350"C.Hence, the abilityof variousclay minerals to Elus, systems.In, H. Barnes,Ed.,Geochemistry of HydrothermalOre form by directcrystallization from high temperature Deposits.Holt, Rinehart,and Winston,New York' waters has been clearly established.It is therefore Ewrll, R. H., nro H. INsuv (1935)Hydrothermal synthesis of postulatedthat high temperaturewaters, either of kaolinite., and nontronite.U.S' Nat' Bur. Stand',J' Res' magmaticor metamorphicorigin, locallydissolved 15,173-185. R. E. (1934)The petrographicstudy of clay minerals.,L from the abundantiron formations Gnru, iron and silica SedimentPetrol. 4, 45-46. andlater, on cooling,directly deposited nontronite in - (1953)Clay Mineralogy.McGraw-Hill' 384 p. their presenthost rocks. Such an origin explainsthe GnuNen,J. W. (1935)Structural relationships of nontronitesand unusualchemistry of this mineral and the other montmorillonites.Am. Mineral. 20, 4'75-483. of Minerals' anomalousfeatures associated with its occurrence.HetNnrcu,E. W. (1964)Microscopic ldentifcation New York, 414 presenceof mostclay minerals found in McGraw-Hill, P' Thoughthe HosrrNc,J. S.,M. Ntslsox, nNoA. C,cnrHrw(1957) A studyof veins has long been attributed to hydrothermal clay mineralogyand particlesize. Aust' J. Agric Res.E'45-74' alteration(see Grim, 1953;Deer, Howie, and Zuss- JrlNs, C. (1968)The neoformationof clay mineralsin brackish man, 1962; etc), the fact that kaolinite, dickite, and marineenvironments. Clays Clay Minerals,9,2W-217' (1950) data on reference beidellite,and nontronitehave been directly syn- Ksnn. P. F.. aNo R. J. PIr-l Analytical Preliminaryreport No. 1 Ref. Clay Minerals, high temperaturewaters indicates that, clay materials. thesizedfrom Am. Pet. Insl.,Res. Proj' 49,Columbia Univ.' 38 p. underappropriate circumstances,a primary origin for Knnusroen, K. A. (1967) Introductionto GeochemistryMc- them in somegeologic occurrence would not be en- Graw-Hill,New York, 712p. tirely unexpected. L,qnseN.E S., rr.roG. Srrlcen (1928)Dehydration and optical studiesof alunogen,nontronite, and griffithite.Am. J. Sci.15' Acknowledgments l-t9 LoucHNAN,F. C. (1969) ChemicalWeathering of the Silicate Field work for this projectwas supportedby grantsfrom the Minerals.Elsevier, New York, 154P. Army ResearchOffice and Universityof SouthAlabama Research NncrLscsutor, G. (1938)On the atomic arrangementand Committee.This studycomprises a portion of a largerproject deal- variability of the membersof the montmorillonite group' ing with tropicalweathering financed by the Army ResearchOffice Mineral. Mag. 25, 140-155. under Grant Number DA-ARO-D-31-124-73-G145.Special oe Rrrvrnorr, G. N. (1964)Origin and Metamorphismof the thanks are also extendedto Dr. W D. Keller, Departmentof Major Paragneissof the PrecambrianImataca Complex: Upata Geology, University of Missouri, for supplying the electron Quadrangle,State of Bolioar, Venezuela,Ph.D. Thesis,Rutgers photomicrographsused in this investigation. University,New Brunswick,New Jersey. Ross,C S.,nro HsNontcrs,S. B. (1945)Minerals of the mont- morillonitegroup. U.S. Geol.Suru. Prof Pap. 205-8,79 p. References -, ANDE. V. Sn,lNNoN(1925) Chemical composition and op- Ar-r-rN,V. T.. ero V. E. Scurto(1946) Nontronite in theColum- tical propertiesof beidellite.Wash. Acad. Sci' /. 15' 467-468. bia River region,Am. Mineral. 3l' 294-311. Sgrnl,r,,rN,G., H. Ixnwl, G UeHlne, ,,rNnE. Ox,tznrt (1962) BrnrHtsn, P. (1827)Nontronite noveau mineral. Ann. Chim. Typesof occurrenceof nontroniteand nontronite-likeminerals Phys.36, 22-27. in soils.Pac. Sci.16,57-62. BnrNoLev,G. W., nNoD. M. M,ccEweN(1953) Structural aspects WEAvER,C. E. (1958)The effectsand geologicsignificance of of the mineralogyof clays.Ceramics, A Symposium,p. l5-59. postassiumfixation by expandableclay mineralsderived from Bnowr, G. (1961)The X-Ray ldentffication and Crystal Structures muscovite.biotite. chlorite, and volcanicmaterial. Am. Mineral' of Clay Minerals Mineral. Soc.,London, 544 p. 43,839-861. Cuasr, R. (1963)The Imataca Complex, the Panamo Amphibolite, wrLDr\4AN.w. 8.. L. D. WsIrrtc, eno M. L. JecrsoN (1971) and the Guri Trondjemite,Precambrian Rocks of the Adjuntas- Serpentinestability in relationto formationof iron-richmont- lz. Mineral.56,587-602. PanamoQuadrangle, State of Boliuar; Venezuela.Ph' D. Thesis. morillonitesin someCalifornia soils. PrincetonUniversity, Princeton, New Jersey. Collrrls, J. H. (1877)Remarks on gramenitefrom Smallcombe, Manuscript receiued,August 12, 1974;accepted and the chloropalgroup of minerals.Mineral. Mag' l, 67-82. for publication,MaY 5, 1975.