JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 96, NO. B8, PAGES 13,485-13,507,JULY 30, 1991

Large-ScaleCrust Formation and LithosphereModification Beneath Middle to Late Cenozoic Calderas and Volcanic Fields, Western North America

CLARK M. JOHNSON

Department of Geology and Geophysics, University of Wisconsin, Madison

Over 500,000 km3 of intermediate-to silicic-compositionash flow tuffs and lavaswere eruptedfrom large-volume volcanic fields in western North America during the middle to late Cenozoic. Of the commonlyused isotope systems, Nd isotopedata provide the best constrainton the proportionsof crust and mantle componentsin the tuffs; all tuffs that have been studied contain a major, often dominant, mantle component. The proportion of mantle to crustal componentsis best constrainedfor ash flow tuffs that were erupted on Precambrian crust. There is no simple correlation between inception of extensionalfaulting with development of calderas, although all calderas formed in or adjacent to regionsthat ultimately underwentcrustal extension to somedegree. Similarly, there are no first-order correlationsbetween the proportionof mantle-derivedcomponent in the silicic magmasand the tectonic setting. The common thread to all caldera complexes is that they are generatedby large fluxes of mantle-derivedbasaltic magmas. Detailed isotopicand petrologic studies of severalcaldera complexes indicate that the silicic magmas were fundamentally derived by fractional crystallization of mantle- derived magmas, accompanied by assimilation of continental crust. Cenozoic caldera-related magmatismin westernNorth America representsa major episodeof crustalgrowth and hybridization; crustalgrowth rates in the Cenozoicmay rival thosethat occurredin the Proterozoicduring initial crust formation. New results of petrologic and isotopic studiesof multicyclic caldera complexesin the northern Rio Grande rift region, in addition to several other areas of the western United States,confirm previous models that predict large changesin O, Sr, and Nd isotope compositionsof the crust during injection of mantle-derived basaltic magmas. Continued injection of basaltic magmas results in an increasein the mean density of the crust, which may be reflected in upward movement of the locus of silicic magmaevolution in the crust;such changes may be monitoredin areasunderlain by ancientcrust using Pb isotope ratios if the crust is stratified in U/Pb ratios. If accompaniedby assimilation, crystallization of mantle-derivedmagmas that stall near the crust-mantleboundary will return crustal componentsto the mantle in the form of mafic and ultramafic cumulates,in addition to producing a petrologicallycomplex crust-mantle boundary. This model can explain many geophysicalcharacteris- tics of the crust and upper mantle in western North America and supportsthe conclusionsof many recent studiesof lower crustal xenoliths, which proposerecent magmatic underplatingin tectonically active regions.

INTRODUCTION caldera complexesalso developedat 40-23 Ma along the easternmargins of what is now the ColoradoPlateau [e.g., Over 500,000 km3 of ash flow tuffs have been erupted Stucklessand Sheridan, 1971; Steven and Lipman, 1976; onto the surface of western North America during the Rattd et al., 1984; Elston, 1984; Pallister and du Bray, middle to late Cenozoic,approximately equally distributed 1989; Lipman et al., 1986]. Caldera-relatedvolcanism in between the western United States and Mexico [Mackin, the Trans-PecosTexas region occurredat 38-28 Ma [Henry 1960; Smith, 1960; McDowell and Clabaugh, 1979; and Price, 1984], and the large volume of ash flow tuffs Swanson and McDowell, 1984; Lipman, 1984]. Taking that belongto the Upper Volcanic Supergroupof the Sierra into accountthe voluminousplutonic rocks that must un- Madre Occidental were erupted at 34-23 Ma [McDowell derlie volcanic centers[Crisp, 1984; Lipman, 1984; Shaw, and Clabaugh, 1979; Swansonand McDowell, 1984]. Vol- 1985], as much as 5,000,000 km 3of intermediate-to silicic- canismin the Mojave and northern SonoranDeserts began compositionmagmas may have been associatedwith this at 30 Ma and migratednorthward over a 20-m.y. period but Cenozoic volcanism. Voluminous, caldera-related Ceno- producedrelatively few calderas [Glazner and Supplee, zoic volcanismin westernNorth America began at 49 Ma 1982]. Well-known young (< 2 Ma) calderasin the western in the Challis volcanic field of central Idaho [Cater et al., United States include those in the Yellowstone and Jemez 1973] and swept southwestalong an arcuate front [e.g., volcanic fields, as well as the Long Valley caldera Armstrong et al., 1969; Best et al., 1989]. Caldera com- [Christiansenand Blank, 1972; Smith et al., 1970; Bailey et plexesin the northeasternGreat Basinformed at 35-30 Ma, al., 1976], and these lie along the outer margins of the whereas calderas in central and western Nevada formed at caldera-related volcanism in the western United States. 30-20 Ma. Calderasat the westernand southernmargins of Cenozoic caldera complexesin westernNorth America the Great Basin formed less than 20 m.y. ago. Large formed in a diverse range of tectonic settings. Although Gans et al. [1989] proposethat major volcanicepisodes are Copyright 1991 by the AmericanGeophysical Union. related to periodsof extensionin the Great Basin, detailed Paper number 91JB00304. studiesof a large number of stratigraphicsections in the 0148-0227/91/91JB-00304505.00 eastern Great Basin indicate that caldera-related volcanism

13,485 13,486 JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHEREMODIFICATION in the region is not associatedwith extension[Taylor et al., 1989; Best and Christiansen, this issue]. Several large- volume caldera complexes formed significantly prior to extension or in areas that underwent little extension, in- cluding the Sawatch Range calderas in central Colorado, the central Sierra Madre Occidental calderas, and the Marysvale, San Juan, and parts of the Mogollon-Datil vol- canic fields [e.g., Lipman et al., 1970; McDowell and Clabaugh, 1979; Steven et al., 1984; Cather and Johnson, 1986; Shannon, 1988]. Early caldera-relatederuptions in the southwestNevada volcanic field began approximately contemporaneouswith initiation of extension[œkren et al., 1968], and ash flow tuffs were eruptedfor another8 m.y. [Byers et al., 1989]. The McDermitt, Jemez, and Long Valley calderas formed after initiation of extension (see Rytuba and McKee [1984] and referencesabove). Regard- less of the tectonic setting in which calderaswere formed, all developed within or adjacent to areas that underwent mid-Miocene to Quaternaryextensional faulting. Geologistshave long debatedif silicic magmasare gen- eratedprimarily by crystal fractionationof basalticmagmas or partial melting of the crust [e.g., Daly, 1914; Bowen, 1915; Holmes, 1932]. Early Sr isotope analysesof ash- flow tuffs in the Great Basin have been usedto supportboth crystalfractionation and crustalmelting models[Noble and Hedge, 1969; Scott et al., 1971; McKee et al., 1972; Noble et al., 1973; Stuckless and O'Neil, 1973]. In the first Pb isotope study of a large-volume caldera complex, Lipman et al. [ 1978] highlight the stronginfluence of the crust on caldera-relatedmagmatism. This report summarizesNd isotopedata for silicic rocks from middle to late Cenozoiccaldera complexes and volca- Fig. 1. Locationsof Cenozoicvolcanic fields and calderasdiscussed nic fields from westernNorth America (Figure 1). The data in text for which Nd isotopedata are available. Other localitiesnoted highlight the large mantle componentin the ash flow tuffs, in italics. B, Batopilas;BCS, Baja California Sur; GP, Grizzly Peak which, based on detailed study of several caldera com- caldera; J, Jemez volcanic field; K, Kalamazoo volcanic sequence; KSW, Kane SpringsWash volcanic field; L, Latir volcanicfield; LO, plexes,is interpretedto reflect fractionalcrystallization of La Olivina; LH, Los Humeros volcanic field; LV, Long Valley large volumesof basalticmagmas, accompanied by interac- caldera region; MCD, McDermitt volcanic field; M-D, Mogollon~ tion with continentalcrust. Of the three commonlyused Datil volcanic field; MO, Molango; N, Novillo; O, Oaxaca; P-W, radiogenic isotope systems,Rb-Sr, Sm-Nd, U-Th-Pb, de- Source area for Peach SpringsTuff and Woods Mountains volcanic termination of the relative proportionsof crust and mantle center; S, Sonoma volcanic field; SA, San Antonio; SJ, San Juan volcanic field; SLC, Sierra los Cajones; SLP, Sierra La Primavera; componentsis probably least ambiguoususing Nd isotope SV, Sierra Virulento; SWNV, southwest Nevada volcanic field; Z, ratios. Neodymium contentsof mantle-derivedbasalts and Zacatecas. 87Sr/86Sr=0.706line for Mesozoicand Tertiary granitic continental crust are similar, and where the basement con- rocks and is interpretedto reflect the westernmargin of the Precam- sistsof Precambriancrust, the Nd isotopecompositions of brian basement in the western United States [Kistler and Peterman, 1973, 1978]. Farmer and DePaolo [1983] interpret granitic rocks the crust and mantle are often distinct. Present-day that haveend=-7 as indicating the Precambrianbasement margin; this 87Sr/S6Srratios of low-Rb lower crustmay be similar to "line" lies just eastof the 87Sr/a6Sr=0.706line. SP in box refersto those of the modern mantle, which would disguisea crustal Sierran Province (stippled pattern), which containsTertiary basalts component. Conversely, the low-Sr contentsof highly that haveexceptionally low endvalues (see references in text). evolved rhyolitesmake their Sr isotoperatios susceptibleto modificationby even minor amountsof late stageassimila- THE LOWER CRUST tion of radiogenicupper crust;this would over emphasizea crustalcomponent. Lead isotoperatios of intermediate-to Contrastsbetween Nd isotope ratios of the crust and silicic-composition rocks are probably dominated by mantleare greatestwhere the crustis Precambrianin age. crustalcompositions, given the relatively high Pb contents The largest body of Nd isotope and rare earth element of the crustas comparedto mantle-derivedbasalts. Tempo- (REE) data for Proterozoic rocks in western North America ral variations at several multicyclic caldera complexesare existsfor thoseexposed in Coloradoand New Mexico, and discussed,with particular emphasis on the northern Rio the majorityof theserocks have low Sm/Nd ratios. Many Granderift region. Data from multicyclic systemsprovide Early Proterozoictholeiite lavas have low Sm/Nd ratios, in strong evidence for extensive hybridization of the lithos- additionto all mafic calc-alkalinerocks yet analyzed,and phere and formation of new crust, and broad consideration thesewould producelow present-day8Nd values (Figures of available Nd isotope data for silicic volcanic rocks 2a and2b). SomeEarly Proterozoictholeiite lavas that are throughoutwestern North America suggeststhat thesepro- associatedwith bimodal greenstonesequences have high cesseswere common at many volcanic centers. Sm/Nd ratios, however, which have producedpositive JOHNSON: LARGE-SCALECRUST FORMATIONAND LITHOSPHEREMODIFICATION 13,487

20 F7]S AND CNEgA/ MEXICO i A comprise a relatively small part of rocks exposed in the [7'] PECOSTHOLElITE m lower Proterozoic sections,but it is likely that they were ,[• PECOSCALC-ALKALINE BASALTm more voluminous in the lower crust in the Early Protero- zoic, when the crust consisted of accreted arcs [e.g., Condie, 1982]. Widespread plutonic activity occurred in 0 ß ß ß ß ß ß ß ß ß ß ß ß ß the southwestern United States between 1000 and 1700 Ma, after formation of the greenstonesequences, and all of the OTHER GREENSTONES plutonic rocks have low Sm/Nd ratios that would produce present-day/•Nd values <-10 (Figures 2d-2f). The Sm/Nd ratio of the Proterozoic crust is determined by the integrated effects of magmatism that has occurred since the time of initial crust formation. Intrusion into the 20 1650-1760 Mo VOLCANICROCKS crustof tholeiite magmasthat had high Sm/Nd ratios (Fig- ures 2a and 2b) was probably restrictedto the early stages [] INTERMEDIATE- of arc formationin the Early Proterozoic,or duringback arc COMPOSITION ROCKS II•--'q SlLIClCROCKS / spreading [e.g., Boardßan and Condie, 1986; Knoper and Condie, 1988; Robertson and Condie, 1989], inasmuch as light REE (LREE) depletedmagmas are relatively rare in • 20 1600-1700MoPLUTONIC ROCKSD thick continental arcs as comparedto oceanic arcs [e.g., Pearce, 1982, 1983]. It seemslikely that basaltic magmas which intruded the lower crust during major plutonism in COMPOSITIONROCKS• the western United States were LREE enriched, similar to continentalbasalts erupted during the Cenozoicin the west- 0 m m m m m•mmm m m m m m m- m m' m 20 ern United States. If the Sm/Nd ratios of mafic magmas 1500 Mo PLUTONS intruded into the lower crust during the Proterozoic were similar to those measured for Cenozoic continental basalts, COMPOSITION ROCKS I!?"•1 INTERMEDIATE- / the mafic lower crust would have low present-day/•Nd val- ues that are similar to the average of exposedProterozoic crust (Figures 2g and 3). o ßmr??'J, •m • m • • • ß • • • m m • 20 It is likely that the lower crust is more mafic than the 1000-14.00 Iv(o PLUTONS middle and upper crust [e.g., Rudnick and Taylor, 1987], and the possibility that the Nd isotopecomposition of the 1•SILICIC ROCKS) exposed crust doesnot reflect that of the lower crust must be assessed. Amphibolite, granulite, and eclogite facies xenoliths from central Arizona, southern New Mexico, and northern and easternMexico have averagepresent-day /•Nd 6o G values of-4 to 0, which are substantiallyhigher than those 4O

2O

0 -18 -14 -10 -6 -2 2 6 10 I PREDICTEDVALUES FOR MAFIC LOWER CRUST

USING BASIN AND ENd(0) RANGE LAVAS GRANDE RIFT LAVAS Fig. 2. Predictedpresent-day ENd values for Proterozoicrocks in 40 L• NORTHERN LATETHOLElITE LAVAS Colorado and New Mexico, based on measured Sm and Nd concen- [] SOUTHERN USINGEARLYNORTHERNAND FLANK RIOLAVAS trations. Sourcefor oldest rocks assumedto be depletedmantle at 1800 Ma [DePaolo, 1981a]. Younger rocks assumedto have been TRANSITIONCOLORADO PLATEAUZONE AND derivedfrom oldestsource as it evolvedto a present-dayENd value of -14. The peakat ENDS-13in the predictedhistogram (Figure 2g) is the sameas that measuredfor intermediate-to silicic-compositionrocks 0 ß ß ß [DePaolo, 1981a; Nelson and DePaolo, 1985]; similar rocks in -18 -14 -10 -6 -2 2 6 10 southeasternArizona are predicted to have an average present-day ENdvalue of-15 ICondieet al., 1985]. Additional datafrom Barker et (• Nd(0) al. [1976b], Condie and Budding [1979], Condie and Nuter [1981], Condie and McCrink [ 1982],Nelson and DePaolo [ 1984],Boardßan and Condie [1986], Knoper and Condie [1988], Robertson and Fig. 3. Predictedpresent-day ENd values for Proterozoicmafic lower Condie [1989], and Stein and Crock [1990]. crust, assumingcrust had REE contentssimilar to those of Cenozoic continental basalts from the western United States. Calculated as- suming crust was derived from depleted mantle at 1800 Ma [DePaolo, 1981a] andusing Sm/Nd ratios measuredfor mafic Ceno- present-day/•Nd values (Figures 2a and 2b) [Nelsonand zoiclavas erupted through Proterozoic basement inthe southwestern DePaolo, 1984]. Silicic volcanic rocks associatedwith the UnitedStates. Data from Phelpset al. [1983],Alibert et al. [1986], EarlyProterozoic greenstone sequences have low Sm/Nd Croweet al. [1986],Dungan et al. [1986],Kempton etal. [1987], Johnson and Lipßan [1988], Dungan et al. [1989], Wittke et al. ratios, which would produce low present-day/•Nd values [1989], Thompsonet al. [this issue]and R. Thompson(unpublished (Figure2c). Thetholeiite lavas that have high Sm/Nd ratios data,1990). 13,488 JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHEREMODIFICATION of exposedProterozoic crust [Esperancaet al., 1988; Ruiz On Phanerozoic Basement and Precambrian et al., 1988a; Kempton et al., 1990; K.L. Cameron et al., Craton Margin Granulite-facies xenoliths from north central Mexico: Evi- West-east traverses across northern Mexico and the dence for a major pulse of mid-Cenozoic crustal growth, submittedto Journal of GeophysicalResearch, 1991]. If Mexican NeovolcanicBelt provide a comparisonof mag- the xenolithsrepresent Proterozoic lower crust, the major- mas erupted on Phanerozoic basement in the west with ity of Nd isotope data for Cenozoic ash flow tuffs from those erupted on Proterozoic basementin the east. The voluminous tuffs of the Sierra Madre Occidental in central westernNorth America can be explainedlargely by melting of the crust,as noted by Ruiz et al. [ 1988a]. Mafic xenoliths Mexico were probably erupted largely on Phanerozoic at the Geronimo volcanic field, New Mexico, and at La basement,although the natureof the basementis uncertain Olivina, Chihuahua,however, are interpretedto reflect re- [e.g., Campaand Coney, 1983]. The McDermitt and Long cent intrusion of mantle-derivedmagmas and associated Valley calderasprovide a comparisonof large-volumeash cumulatesinto the deep crust [Kemptonet al., 1990; K.L. flow tuffsthat were erupted just westand east, respectively, Cameron et al., submittedmanuscript, 1991]. In contrast, of the Precambriancraton margin in the western United States. xenolithsfrom Arizona and easternMexico that have high present-day ENdvalues have been interpreted to reflect Northern Mexico. Oligocene dacite lavas in northwest- lower crust that is Proterozoic in age [Esperanca et al., ern Mexico in the Batopilas region of the Sierra Madre 1988; Ruiz et al., 1988a]. This interpretationis basedon Nd Occidentaland mid-Miocene to Quaternarylavas in Baja and Pb-Pb model ages. Interpretationof thesemodel ages California Sur have high ENdvalues (+0.2 to +4.8) that as indicating a Proterozoic age for the xenoliths has been overlapthose of Oligocenemafic lavas in the region (Fig- recently debated, however, highlighting the difficulty of ure 4, northwesternfield), suggestingthat they evolved using xenoliths to constrainthe isotopic compositionsof largely by crystal fractionationof mantle-derivedbasaltic the Proterozic lower crust [Cameron and Robinson, 1990; magmas [Cameron and Cameron, 1985; Cameron et al., Esperanca et al., 1990; Johnson, 1990; Ruiz et al., 1990]. 1989]. Theselavas, however, are not underlainby Precam- In summary,it appearslikely that the Proterozoiclower brian crustbut probablyby Mesozoicand younger plutonic, crustprior to initiation of Cenozoicmagmatism had low ENd volcanic, and sedimentarycomplexes [Campa and Coney, values that were similar to those of exposedbasement 1983], and there may be little contrastin Nd isotopecom- rocks, consistentwith the observationthat most granulite- positionsbetween basement and mantle-derivedmagmas. facies rocks have low and rather consistent Sm/Nd ratios Lavas in northeast Mexico in the Chinati Mountains, San [e.g., Weaver and Tarney, 1980; Ben Othman et al., 1984]. Antonio, and Sierra los Cajonesregions, in addition to an Although the crust is probably vertically zoned in major ash flow tuff near Sierra Virulento, are similar in age and element and Sr and Pb isotopecompositions, it is probably major element compositions to lavas in northwestern relatively homogeneousin Nd isotopecompositions. This Mexico but were erupted on Precambrianbasement [e.g., is consistentwith Nd isotopedata for peraluminousgran- Campa and Coney, 1983]. The northeasternsilicic rocks ites from the southwestern United States, which have an have ENdvalues that are similar to those of the western average present-dayENd value of-14, and are thought to lavas, varying from +0.1 to -2.0 (Figure 4, northeastern dominantlyrecord lower and middle crustNd isotopecom- fields and symbols[Cameron and Cameron, 1985]). Tem- positions [Farmer and DePaolo, 1984; Bennett and porally relatedmafic lavas have ENdvalues of-2.2 to +3.5 DePaolo, 1987; J. Wooden and C. Johnson,unpublished [Cameronet al., 1989]. Theserocks were probablyerupted data, 1989]. Moreover, studiesof crustally contaminated throughProterozoic basement rocks that have low present- mafic- and intermediate-compositionlavas in the northern day ENdvalues of-14 to -4, as indicated by pelitic and Rio Grande rift suggestthat the Proterozoic lower crust in granuliticxenoliths from La Olivina and exposedbasement the region has low present-dayENd values [Johnsonand at Novillo [Cameron and Cameron, 1985; Ruiz et al., Thompson,this issue]. 1988a,b; K.L. Cameron et al., submitted manuscript, 1991). As Cameronand Cameron [ 1985] note, the high ENd values of the silicic volcanic rocks are consistent with MIDDLE TO LATE CENOZOIC CALDERAS generationby crystal fractionationof mafic magmas,ac- AND VOLCANIC FIELDS companiedby little crustalassimilation. Mexican Neovolcanic Belt. The Sierra La Primavera Neodymium isotope data are now available for caldera center lies near the western margin of [he Mexican complexesand volcanic fields that span nearly the entire Neovolcanic Belt and Quaternaryvolcanism at this center age range and tectonic setting of the middle to late Ceno- included the 40 km 3 Tala Tuff [Mahood, 1981]. The Sierra zoic volcanic episodethat coveredmuch of westernNorth La Primaveracenter is probablyunderlain by Mesozoicand America. Discussion of calderas and volcanic fields is Cenozoic basement [Campa and Coney, 1983]. Mahood divided into three groups,based on basementage and tec- and Halliday [ 1988] report a restrictedrange of ENdvalues tonic setting: (1) Phanerozoicbasement and at the margin for all Sierra La Primaverarocks of +4.8 to +5.8 (Figure 4), of the Precambrian craton, which includes extended and including an older basalt. Mahood and Halliday [1988] non-extendedregions, (2) the highly extendedGreat Basin reject crystal fractionation of basaltic magmas as the and Mojave Desert regions, which are underlain by Prot- mechanismfor producingthe silicic magmas,based prima- erozoic basement, and (3) the modestly extended Rio rily on argumentsthat the volume of parentalbasalt and Grande rift region, which is also underlainby Proterozoic associated cumulates would be excessive, and instead favor basement(Figure 1). partial melting of Mesozoic or Cenozoicplutonic rocks. JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHERE MODIFICATION 13,489

I MEXIC0/ l EASTERNMEXICO BASEMENT]

ß ,-)• SIERRALA PR.[_NI_AVERA ----.•-'• J A !-7-] XENOLITHS . ß -' NORTHWESTERN ,,/ • I EXPOSEDPRECAMBRIAN

-2 QUATERNARY ZACATECAS NORTHEASTERN -6 BASANITES LAVAS

,• • LOS HUMEROS -10 ß • ZACATECAS ß /x NORTHE.ASTERN REGION , , < -14 -14 ß ß ß ,

wf % SiO2 FREQUENCY

Fig. 4. (a) 8Nd-SiO2variations for Cenozoicvolcanic rocks in Mexico.Dashed lines enclose data for rockson Phanerozoicbasement, and solid lines indicate data for rocks on Proterozoicbasement. Los Humeros and Northeast- ern Regionrocks on Proterozoicbasement; Zacatecas probably on Phanerozoicbasement. For Figures4-8, solid symbolsindicate ash flow tuffs,open symbols indicate lavas, and all •mdvalues are initial values. Vertical scales for Figures4-8 span20 •mdunits, although maximum and minimumlimits vary. Data from Verma [1983, 1984], Cameronand Cameron [1985], Mahood and Halliday [1988], Cameronet al. [1989], and Pier et al. [1989]. (b) Histogramfor present-day•md values of exposedbasement rocks and lower crustal xenoliths in easternMexico. For Figures4b, 5b, 6b, and 8b, verticalscales match those in Figures4a, 5a, 6a, and 8a. Data from Cameronand Cameron [1985] and Ruiz et al. [1988a, b].

The Los Humerosvolcanic center erupted ~220 km3 of [1984], the proportion of mantle in the silicic rocks would basalt to high-SiO2 rhyolite magma at 0.47-0.02 Ma and be considerablyhigher than previously estimated. was associated with three caldera-forming eruptions McDermitt. Seven major ash flow tuffs, totalling over [Ferriz and Mahood, 1984]. This center lies near the 1700 km3, were eruptedfrom four overlappingand nested easternmargin of the Mexican NeovolcanicBelt and occu- calderasof the McDermitt calderacomplex and three satel- pies the borderbetween Proterozoic rocks that are part of lite calderasbetween 16.1 and 15 Ma [Rytubaand McKee, the easternmargin of the Sierra Madre terrane and Paleo- 1984]. Ail of the tuffs are peralkalinerhyolite or comendite zoic rocks of the Coahuila terrane [Campa and Coney, and are part of a sequenceof mid-Miocene alkaline caldera 1983]. Proterozoic basement rocks are exposed to the complexesthat developedalong the westernand southern northwestat Molango and southat Oaxaca,and theserocks marginsof the Great Basin [Noble and Parker, 1974]. The have low present-dayENd values of-16 to -6 (Figure 4) McDermitt volcanic field was developed on Phanerozoic [Ruiz et al., 1988b]. ENdvalues of mafic- to silicic-compo- basementand lies west of the 87Sr/86Sr=0.706line for plu- sition rocks at Los Humeros vary from +4.1 to -1.4 and tonic rocksthat is interpretedto reflect the westernlimit of decreasewith increasingSiO 2 contents(Figure 4 [Verma, Precambrianbasement (Figure 1) [Kistler and Peterman, 1983]). These variationsare interpretedby Verma [1983] 1973, 1978]. as reflecting generationof the silicic rocks primarily by The Tuffs of Double H and Members 2 and 3 of Long crystal fractionationof mafic parentalmagmas, accompa- Ridge, the first two tuffs of the McDermitt caldera com- nied by little assimilationof crust. plex, have ENdvalues of +2.0 to +6.4 [Tegtmeyerand Central Mexico. Althoughonly a very small fractionof Farmer, 1990]. The lowest ENdvalues are interpreted to the voluminous ash flow tuffs and lavas of the Upper Vol- reflect late stage assimilationin the pre-eruptive magma canicSupergroup of theSierra Madre Occidental have been chamber [Tegtmeyer and Farmer, 1990], indicating that studied in detail, rocks near Zacatecas have a relatively magmasparental to the ash flow magmashad ENdvalues restrictedrange of ENdvalues, from -2.3 to + 1.4 (Figure 4) >+6, and that the local crust had ENdvalues <+2. The [Verma, 1984]. Theserocks crop out in the westernSierra highest ENdvalues (least contaminatedat late stages)over- Madre terrain, which may be underlain by Phanerozoic lap thoseof the underlyingSteens Basalt (lgNd='!'4.1 to +6.9 rocks [Campa and Coney, 1983]. Verma [1984] suggests [Carlson and Hart, 1987]). Probable basement rocks in- that the silicic rocksare dominatedby a crustalcomponent, clude Mesozoic and Tertiary plutonic rocks of the north- althoughuncertainties in constrainingbasement composi- western Great Basin which lie immediately south of the tions make calculation of the relative contributions of crust McDermitt field and have present-dayENd values of-6.4 to and mantle difficult. If a basaltlisted by Verma [ 1984] that +1.4 [Farmer and DePaolo, 1983, 1984], or Paleozoic hasan ENdvalue of + 1.4 is morerepresentative of the mantle sedimentaryrocks that have present-dayENd values <-11 compositionthan the value of +7 assumedby Verma [Farmer, 1985]. The high ENdvalues of the McDermitt 13,490 JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHEREMODIFICATION magmasthat were least contaminatedby late stageassimi- zoic granitic rocks of the Sierra Nevada batholith, and the lation suggesta dominantmantle componentin the silicic granitic rocks that lie east of the 0.706 line have present- magmas. day/•Ndvalues of-8.6 to -3.6 [DePaolo, 1981b]. Paleozoic Long Valley. Greaterthan 600 km3 of high-SiO2rhyolite sedimentaryrocks intruded by the Sierra Nevada batholith was erupted as the Bishop Tuff from the Long Valley have present-day /•Ndvalues of-11.8 to -3.4 [DePaolo, calderaat 0.7 Ma [Bailey et al., 1976]. Premonitoryvolca- 1981b]. Thesedata suggestthat the silicic volcanicrocks in nism included eruption of small volume mafic- and inter- the Long Valley region, including the Bishop Tuff, contain mediate-compositionlavas between3.2 and 2.6 Ma [Bailey a large mantle component and that this component in- et al., 1976], followedby eruptionof N15km 3 of high-SiO2 creasedwith continuedmagmatism in the area. rhyolite at Glass Mountain between 2.1 and 0.8 Ma [Metz and Mahood, 1985]. Neodymium isotope data form a striking trend of increasingend values with decreasingage On Highly ExtendedProterozoic Basement for silicic rocks at Glass Mountain, the Bishop Tuff, and Mono Craters (Figure 5). Older lavas at Glass Mountain This region includes several multicyclic caldera com- (1.35 Ma) have endvalues of-3.9 to -2.6, whereasyounger plexes that were associatedwith basaltic volcanism and lavas (1.06 Ma) have higher end values of-1.2 to -0.8 Proterozic basementthat had unusual isotopic composi- (Figure 5) [Halliday et al., 1989]. Although the Bishop tions. The basementof the Mojave Desert is 200-400 m.y. Tuff magma chamber was thermally and chemically zoned older than that of the central Great Basin, and this is re- [Hildreth, 1979], endvalues of the tuff are constantat -1.2 flected in markedly lower present-day/•Nd values for Pre- to -0.8 (Figure 5) [Halliday et al., 1984], and identical to cambrianrocks. The basementunderlying the northeastern the younger Glass Mountain lavas. The Inyo-Mono craters Great Basin probably also has present-day/•Nd values that chain containsthe youngest silicic lavas and domes in the are lower than those of the central Great Basin, due to the region, and these rocks have end values of-1.4 to +0.9 proximity of the Archean craton. /•Ndvalues for basaltic [Sampson and Cameron, 1987; Kelleher and Cameron, lavasin the GreatBasin and Mojave Desertrange from very 1990]. Plioceneand Quaternarybasalts and andesitesfrom low values (down to - 11) for the Sierran Province rocks to the Long Valley area have endvalues of-2.5 to +2.9, similar relatively high values (up to +9) for young basaltsof the to those of the silicic rocks [Ormerod et al., 1988; Kelleher central Great Basin and easternMojave Desert. and Cameron, 1990]. Mojave Desert. Ash flow magmatism was rare in the Becausethe Long Valley caldera is locatedjust east the Mojave Desert region comparedto the adjacentBasin and 87Sr/SaSr=0.706line [Kistler and Peterman, 1973, 1978], it Range. The 18.5 Ma PeachSprings Tuff is the only region- is inferred to be underlain by Proterozoicand Phanerozoic ally extensiveCenozoic ash flow tuff in the Mojave Desert basementrocks. The caldera partly collapsed into Meso- area, and it representsat least several hundred cubic kilo-

EASTERN CALIFORNIA

I VOLCANICANDPLUTONIC ROCKS I BASEMENTJ

• PALEOZOIC SED RXS U[0[] MONOBISHOP CRATERSTUFF 0 .l•)..... A m PROTEROZOIC o o YGM -•-'m.....m• •

-4 ,,//' SNP'"", ,,. -8 . $PB ',, - --%<._._Z. LZ.::..•,,• .. - \ ) / • MMP 18.5 -12 ß. \ •100DS MOUNTAINS / • WILD HORSE MESA TUFF -16 0 LAVAS ß', MPP\ < -17 : ß ß ß ß 45 55 65 75 0 4 8 wf % SiO2 FREQUENCY

Fig. 5. (a) ?_.Nd-SiO2variations for Cenozoicvolcanic rocks and Mesozoicplutons in easternCalifornia that lie east of 87Sr/g6Sr=0.706line, on Proterozoicbasement. OVB, OwensValley basalts;SPB, Sierran Province basalts; SNP, SierraNevada plutons that crop out east of 87Sr/g6Sr=0.706line; MMP, Mojavemetaluminous plutons; MPP, Mojave peraluminousplutons; PST, PeachSprings Tuff; OGM, older GlassMountain lavas; YGM, youngerGlass Mountain lavas. (b) Histogramfor present-dayENd values of exposedbasement rocks in easternCalifornia. Data fromDePaolo [1981b], Menzies et al. [1983], Farmer and De?aolo [1983, 1984], Nelson and De?aolo [1985], Bennett and De?aolo [1987], Ormerod et al. [1988],Musselwhiteet al. [1989],Farmer et al. [1989],Halliday et al. [1984, 1989], Kelleher and Cameron, [1990], J. Wooden and C. Johnson(unpublished data, 1989) and C. Johnson,(unpublished data, 1990). JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHEREMODIFICATION 13,491 meters of silicic magma, based on estimatedoutflow vol- tions of •Ndvalues for silicic rocks from the Timber Moun- umes alone [Glazner et al., 1986; Nielson et al., 1990]. The tain-OasisValley complex. Early extracaldera(14 Ma) PeachSprings Tuff probablyerupted from a sourcein the intermediate-and silicic-composition lavas have the lowest easternMojave Desert, near the intersectionof the Califor- E:Ndvalues of the complex(-13 to -12; Figure 6), whereas nia-Nevada-Arizona borders [Hillhouse and Wells, 1991], the large-volumeash flow tuffs have E:Ndvalues that in- severalmillion years after a period of short-lived,though creaseto -9 with decreasingage (Figure 6). Caldera-related voluminous,intermediate-composition volcanism in the re- mafic- and silicic-compositionlavas have E:Ndvalues that gion [e.g., Armstrongand Higgins, 1973; Glazner, 1990]. are similarto thoseof the tuffs, in contrastto the early Two glasssamples from pumicein the lower PeachSprings lavas. Tuff haveend values of- 11.7 and-11.0 (C. Johnson,unpub- Basaltic volcanismin the region largely occurredafter lished data, 1990), similar to the averagepresent-day end the peakin ashflow tuff eruptionsat ~ 11 Ma [Croweet al., values of metaluminous Mesozoic granites from the 1986;Byers et al., 1989],and Nd isotopedata appear to be Mojave Desert region (Figure 5) [Farmer and DePaolo, bimodally distributedbetween-10.4 to -7.8 and- 1.3 to +3.7 1983, 1984; J. Wooden and C. Johnson,unpublished data, (Figure 6) [Farmer et al., 1989, 1991]. Other mid-Miocene 1989]. and youngermafic lavas in the region includethose of the The Woods Mountains volcanic center contains one of "SierranProvince basalts" (Figure 6); seealso Figure 5), the few calderasrecognized in the Mojave Desert region. and thoseof northwestArizona (Figure 6) [Alibert et al., The 15.8 Ma caldera-formingWild Horse Mesa Tuff (~80 1986]. Proterozoic rocks in southern Nevada, northern km3 magma equivalent)largely consistsof peralkaline Arizona,and eastern California generally have present-day rhyolite, but mingled rhyolite and trachytepumices in the E:Ndvalues _<-16 (Figure 6), significantlylower than those of upperparts of the tuff are thoughtto reflect tappingof more ash flow tuffs from the Timber Mountain-OasisValley mafic magma at depth [McCurry, 1988]. Trachyte and complex,implying a largemantle component in the tuffs. rhyolite have similar endvalues of-7.5 to -6.2 (Figure 5), Kane Springs Wash. The Kane SpringsWash volcanic which have beeninterpreted as reflectingextensive crystal center in eastern Nevada was active from 14 to 13 Ma, and fractionationof mantle-derivedmagmas and 20 to 40 wt % mafic lavas both underlie and overlie lavas and tuffs of the crustal assimilation [Musselwhite et al., 1989]. Postcaldera center, at 16-15 Ma and 13-11 Ma, respectively[Novak, mafic lavas (11.2-10.2 Ma) that unconformablyoverlie the 1984]. Extension-relatedfaulting largely postdatesthe tuff have endvalues of-4.2 to +2.4 [Musselwhite et al., youngermafic lavas. The Kane Wash Tuff eruptedat 14 1989]. The PeachSprings Tuff and Wild Horse Mesa Tuff Ma and represents>130 km3 of metaluminousand mildy eruptedfrom the samegeneral vicinity, and althoughcur- peralkaline rhyolite magma [Novak and Mahood, 1986]. rent isotopicdata are sparse,their Nd isotopecompositions Two Nd isotopeanalyses of Kane Wash Tuff indicatethat it are suggestiveof a trend of increasingeNa with decreasing has relatively high E:Ndvalues of-6.9 and -6.7 (Figure 6) age (Figure 5). [Novak, 1985; Leeman and Hawkesworth, 1986]. Precambrianbasement rocks in the Mojave Desertregion Precalderatrachyte and postcalderalavas have similar val- have present-dayend values that are significantly lower ues, between-6.4 and -5.3. Extracalderabasalts have E:Nd than thoseof Tertiary lavas and tuffs, includingthe Peach valuesof-8.3 to -2.0, and a precalderatholeiite lava has an SpringsTuff (Figure 5). Proterozoicrocks have Nd TDM E:Ndvalue of-10.4. Trachytic lavas, common in both the agesof 2.3-2.0 Ga, and low present-dayend values of-24.1 precalderaand postcaldera volcanic sequence, are probably to -9.9 (Figure 5) [Nelsonand DePaolo, 1985; Bennett and similar to the magmasthat were parentalto the Kane Wash DePaolo, 1987; Musselwhite et al., 1989]. These low Tuff magma. Novak and Mahood [ 1986] suggestthat com- valuessuggest that the ashflow tuffs containa largemantle positionaland isotopicvariations in the intermediate-com- component, particularly the younger Wild Horse Mesa positionrocks are bestexplained by assimilation/fractional Tuff. crystallization(AFC) of alkali basalt magmas,accompa- SouthwestNevada VolcanicField. Five major calderas nied by relatively small amountsof crustal assimilation; developedat 14-6.5Ma in southwestNevada, accompanied continuedcrystal fractionation is thoughtto be the primary by eruptionof >5000 km3 of ashflow tuffs [e.g., Byers et mechanismfor producingthe Kane Wash Tuff magma. al., 1989]. The 14.3-11.4 Ma Timber Mountain-Oasis Val- Kalamazoo. Forty to 35 Ma andesiteto rhyolite lavas ley calderacomplex produced >4000 km3 of calc-alkaline underlie the regional Kalamazoo Tuff in northeasternNe- and alkali-calcicash flow tuffs [Byerset al., 1976a, 1976b; vada and northwestern Utah [Gans et al., 1989]. The Christiansen et al., 1977; Broxton et al., 1989] which are pretuff lavasgenerally predate initiation of regionalexten- intercalatedwith peralkalinetuffs of the slightly older Si- sion in the area, which occurred at ~36 Ma. The dacite to lent Canyon caldera [Orkild et al., 1969; Sargent and rhyolite KalamazooTuff erupted>500 km3 of magmaat 35 Orkild, 1973; Sawyerand Sargent,1989] and overlainby Ma from an inferred sourcein or adjacentto the Schell peralkaline tuffs of the 7.5 Ma Black Mountain caldera Creek Range and is overlain by an equally voluminous [Christiansen and Noble, 1968; Noble and Christiansen, sequenceof dacite lavas [Gans et al., 1989]. E:Ndvalues of 1968; Vogel et al., 1989]. The last calderato form in the pretuff andesiteand dacite lavas decreasewith increasing field is the ~6.5 Ma Stonewall Mountain caldera [Ekren et SiO2 contents,from -7.4 to -16.0 (Figure 7) [Gans et al. al., 1971; Weissand Noble, 1989]. Early volcanismin the 1989]. Two samplesof the KalamazooTuff have slightly volcanicfield correlateswith the onsetof regional exten- lower E:Ndvalues of - 17.7 and- 16.7 (Figure7); a postcaldera sion [Ekren et al., 1968]. dacite has a similar E:Ndvalue of-16.5. These relations are In one of the first detailed Nd isotope studies of a interpretedby Gans et al. [1989] as reflecting AFC of multicycliccaldera complex, Farmer et al. [1991] present mantle-derivedbasaltic magmas, involving assimilationof importantdata which document significant temporal varia- >50 wt % crust. 13,492 JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHERE MODIFICATION

SOUTHERN GREAT BASIN

I VOLCANICROCKS BASEMENT

ß ß KANE WASH TUFF PALEOZOIC SED RXS NW'•"i;• '",':"•'! SWNVB-LCB ',1 , • KANESPRINGS LAVAS k J ARIZONA,E CALIF, -2 NEVADA PROTEROZOIC SPBIi

... -6

z • /F-"6,,.' AT . • //I ',,,,, -10 • I..•li Ma TS-TV-RM "---"• - - ' ', ; /.' -,...=,,...=,-•,,-,,.•,• ,'...... ;'" PL ,,/ :, PL"; -14 " ...... ' 14 l•a SWNVB TIMBER MTN-OASIS VALLEY < -18

-18 ß ß z5 55 65 75 0 5 lO wf % Si02 FREQUENCY Fig. 6. (a) eNd-SiO2variations for Cenozoicvolcanic rocks in thesouthern Great Basin. SWNVB, southwest Nevada volcanicfield basalts;LCB, LunarCraters basalts; SPB, Sierran Province basalts; NWA northwestArizona basalts. Fieldsat high-SiO 2side of diagramdenote lavas and ash-flow tuffs from the Timber Mountain-Oasis Valley caldera complexof thesouthwest Nevada volcanic field. TSand TV, TopopahSprings Member and Tiva Canyon Member, respectively,ofthe Paintbrush Tuff; RM and AT, Rainier Mesa Member and Ammonia Tanks Member, respectively, ofthe Timber Mountain Tuff. PL,precaldera lavas from the Timber Mountain-Oasis Valley Caldera complex. Kane WashTuff from Kane Springs Wash caldera. (b) Histogram for present-day ENd values of exposedbasement rocks in thesouthern Great Basin. Data from DePaolo [1981b], Farmer and DePaolo [1983], Menzies et al. [1983],Novak [ 1985],Nelson and DePaolo [ 1985],Alibert et al. [ 1986],Leeman and Hawkesworth [ 1986], Bennettand DePaolo [1987],Ormerod et al. [1988],Farmer et al. [1989, 1991],Lum et al. [1989],Musselwhite et al. [1989],and Tegtmeyer and Farmer [1990].

This interpretationis consistentwith variations in Nd >25,000km 3 of ashflow tuffs from 36 Ma to 1 Ma. A large isotopecompositions of probablemantle-derived magmas body of Nd isotope data exists for the calderas of the and crust. The/•INdvalue of the most mafic Kalamazoolava northernrift regionin Coloradoand northernNew Mexico. may be similar to that of the Oligocene mantle in the A large data set of O, Sr, and Pb isotoperatios is also region,and lies within the rangeof the 1OW-•mdgroup of availablefor theserocks and formsthe primarybasis for youngermafic lavas in the TimberMountain-Oasis Valley discussionlater in thispaper on the originof ashflow tuff caldera complex region (Figure 7; see also Figure 6) magmas and temporal and compositionalvariations in [Farmer et al., 1989, 1991]. Few Proterozoicrocks in the large,multicyclic caldera complexes. regionhave beenanalyzed for Nd isotoperatios, but two NorthernArea. An exceptionallylarge middle Tertiary rocksfrom the RubyMountains to the westhave present- volcanicfield is exposedin the northernRio Granderift day /•INdvalues of-16 and -23.8 [Farmer and DePaolo, region in southern Colorado and northern New Mexico 1983]. LatePrecambrian and early Paleozoic miogeoclinal [Steven,1975]. TheGrizzly Peak Tuff erupted>600 km 3 of sedimentaryrocks in the Deep Creek and Schell Creek dominantlyrhyolite magma at 34 Ma fromthe Grizzly Peak Rangeshave present-day•md values of- 18 to -11, andthose caldera, which is the northernmost of three calderas that in the Pilot Rangeand Ruby Range have present-day /•INd formed during the Oligocenein the SawatchRange valuesof-26 to -18 (Figure7) [Farmer,1985] and probably [Fridrichet al., 1991]. Althoughthe SawatchRange indicatea 1OW-/•INdbasement at depth. Low /•INdvalues for calderascurrently occupy the westernside of the northern the basementare predicted,considering the proximityto Archean crust which is inferred to extend south to within 50 Rio Granderift, theyformed 8-10 m.y. priorto develop- mentof regionalextension in the area[Epis and Chapin, km of the Kalamazoosequence [e.g., Bryant, 1988]. These 1975; Scott, 1975; Taylor, 1975; Shannonet al., 1987]. factorsindicate that it is likely that basementrocks under Eruptionof theGrizzly Peak Tuff tappedan exceptionally the Kalamazoorocks have present-day•md values that are large compositionalrange, from 57 to 77 wt % SiO2 significantly lower than those of Precambrianrocks in Colorado, New Mexico, and the central Great Basin. [Fridrichand Mahood, 1987]. /•Imdvalues of thetuff range from-13.0 to -11.3 (Figure8) andare interpreted to reflect AFC of mantle-derivedbasaltic magmas that assimilated Rio Grande Rift Region 20-40 wt % crustwhich had/•INd values <-20; suchcrust is typical of-1400 Ma Silver Plume-typeintrusions in the Severallarge-volume caldera complexes formed adja- region[Johnson and Fridrich, 1990]. Althoughthe /•Imd centto theRio Granderift, from the SawatchRange, Colo- valuesof theGrizzly Peak Tuff arerelatively low, the mafic rado, to the U.S.-Mexicoborder; these calderas erupted partsof the tuff have SiO2 contentsthat are too low to have JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHEREMODIFICATION 13,493

NORTHEASTGREAT BASIN mediate-composition lavas contemporaneouswith forma- tion of the nearby SawatchRange calderasand continuedto 30 Ma [Lipman et al., 1970, 1978; Lipman, 1989]. end VOLCANICANDPLUTONIC ROCKS values of these lavas range from -9 to -6 and generally

z decreasewith increasingSiO 2 contents(Figure 8) [Riciputi

v and Johnson, 1990; Colucci et al., this issue]. Ash flow

-4 ._1 eruptionsbegan at 30 Ma and continueduntil 23 Ma, result- ing in emplacement of approximately 15 major ash flow

-8 tuffs and as many calderas;the total volume of ash flow tuff z magmaerupted exceeded 10,000 km3 [Stevenand Lipman,

-12 1976;Lipman, 1989]. More than6000 km3 of largely dacite

z and rhyolite ash flow tuffs were erupted from at least six calderas in the central caldera complex at 28-26 Ma, in- -16 o % • cludingthe >3000 km3 Fish CanyonTuff. The centraltuffs and related lavas have endvalues between -8 and -6 and are -2O interpreted to contain a large mantle component [Riciputi 45 •5 Iß CMP•5 l TO and Johnson, 1990]. wt • SiO2 -2s Precaldera volcanism at the Latir volcanic field was con- Fig. 7. •:Nd-SiO2variations for Cenozoicvolcanic rocks and Meso- temporaneouswith the peak in ashflow tuff magmatismin zoic and Cenozoic plutonic rocks in the northeast Great Basin. the nearbySan Juanfield, at 28 Ma [Lipmanet al., 1986]. SWNVB, southwest Nevada volcanic field basalts; LCB, Lunars Early magmatismin the Latir field consistedof approxi- Craters basalts; SRPB, Snake River Plain basalts (east of 87Sr/86Sr=0.706line); MMP, Mesozoicmetaluminous plutons; PP, mately ~1000 km3 of intermediate-to silicic-composition Mesozoic and Cenozoic peraluminousplutons; CMP, Cenozoic lavas and related shallow intrusions,followed by eruption metaluminousplutons. Few Nd isotope data are available for ex- of 500-1000 km3 of high-SiO2peralkaline rhyolite as the posedPrecambrian basement in the Kalamazooregion, althoughlow Amalia Tuff [Lipman, 1983, 1988; Johnson and Lipman, ENdvalues for upperPrecambrian and lower Paleozoicsedimentary 1988]. The Questa caldera formed during initiation of rocksin the region (shownas bars on right side of figure) suggestthat the basementfrom which the sedimentaryrocks were derived has regional extension at 26 Ma [Lipman et al., 1986; very low ENdvalues. Data from Menzieset al. [1983], Farmer and Hagstrum and Lipman, 1986]. Postcalderamagmatism is DePaolo [1983, 1984], Farmer [1985], Hart [1985], Farmer et al. preservedas lavason a horstin the Rio Granderift [Thomp- [1989], Gans et al. [1989], and Lum et al. [1989]. son et al., 1986], and as intermediate- to silicic-composi- tion intrusive rocks as young as 19 Ma on the east side of been producedby melting of most crustal rocks [Johnson the rift [Lipman et al., 1986]. endvalues of the precaldera and Fridrich, 1990]. lavas vary from -1.8 to -7.3 and generally decreasewith Early precalderavolcanism at the San Juanvolcanic field increasingSiO 2 contents(Figure 8) [Johnsonet al., 1990]. consistedof eruptionof voluminous(>25,000 km3) inter- The Amalia Tuff and related lavas and resurgentintrusions

NORTHERN RI0 GRANDE RIFT

VOLCANICROCKS I BASEMENTk 2 i B >2 $LH21 I• Aß UR AMALIA •O•C TUFFFm DI A

$LH1 ß ',,, LVF ...... (œ6MoJ // .-.• -.... ß

z l ...... , ß .. -10-

ß

ß -'14.- SAN JUAN VOLC FIELD ß

ß

ß TUFF < -18

ß 45 •5 wt %SiO 2 FREQUENCY Fig. 8. (a) •:Nd-SiO2variations for Cenozoicvolcanic rocks of the compositevolcanic field that is exposedin the northernRio Granderift region [Steven,1975]. SLH1 and SLH2, mantle sourcesfor early rift (26 Ma) basaltsat San Luis Hills; LVF, precalderalavas of the Latir volcanic field; ESJL, early San Juan lavas; NWCOB, Miocene to Quaternarybasalts from northwesternColorado. (b) Histogramfor present-dayENd values of exposedbasement rocks in the northern Rio Granderift region. Data from DePaolo [1981a], Nelson and DePaolo [1984, 1985], Leat et al. [1988, 1989, 1990],Johnsonet al. [1990],Johnson and Fridrich [1990],Riciputi and Johnson[1990], Colucci et al., [this issue];Johnson and Thompson,[this issue],and L. Riciputi and C. Johnson,(unpublished data, 1990). 13,494 JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHEREMODIFICATION have ENdvalues of-5.9 to -6.9 (Figure 8) [Johnsonet al., SILICIC MAGMATISM AND RELATIONS TO 1990]. Postcalderaintrusions generally have ENdvalues of EVOLUTION OF THE LITHOSPHERE -4 to -7 and are interpretedto reflect the waningstages of the magmaticsystem. Detailed petrologic,chemical, and ENdvalues of Cenozoic ash flow tuffs from westernNorth isotopic studiesof the Latir field indicate that most of the America are closelyrelated to thoseof the local mantle, as evolvedmagmas were generatedby AFC of mantle-derived estimated by caldera-relatedmafic lavas, or those that basalticmagmas [Johnson and Lipman, 1988; Johnsonet erupted shortly after caldera magmatism(Figure 9). ENd al., 1990]. values of the tuffs also correlate with those of the local basement,indicating both crustaland mantle influenceson Primitive basaltsin the northernRio Granderiff region the isotopiccompositions of the silicic rocks (Figure 9). were erupted after initiation of regional extensionwithin, Althoughthe Nd isotopecomposition of the crustis likely and alongthe marginsof, the broadmiddle-Tertiary volca- to be variable and the compositionof the lower crust re- nic field that includesthe SawatchRange calderas and the mainsuncertain, it is relativelyhomogeneous as compared San Juan and Latir volcanic fields [Steven, 1975]. ENd to Sr and Pb isotopevariations, and thereforeNd isotope valuesof primitive, early-extensionlavas at San Luis Hills ratiosprovide the bestconstraint possible on the proportion form two clustersat -4 and 0 (Figure 8) [Johnsonand of mantle and crustalcomponents in the tuffs. Where low Thompson,this issue];this may be the range of valuesof ENdvalues of mafic lavas are due primarily to lOW-ENd magmasthat were parental to the ash-flow tuff magmas. lithosphericmantle [e.g., Fitton et al., 1988; Kemptonet Miocene and youngermafic lavaserupted northwest of the al., this issue],and not crustalcontamination, the propor- Rio Granderift have ENdvalues of-4 to -8 and may repre- tion of crustin the ash flow tuffs basedon inspectionof sentmelts derivedfrom enrichedlithospheric mantle [Leat Figure 9b will be overestimated. et al., 1988, 1989, 1990]. Present-dayENd values of most Detailedstudies of severalcaldera complexes support an Proterozoicrocks in Coloradoand New Mexico rangefrom origin for the ash flow tuffs that involve assimilation/frac- -16 to -10, although values as low as -24 are found in tional crystallization(AFC [DePaolo, 1981c]) of basaltic LREE-rich, 1400 Ma anorogenic granites (Figure 8) magmas,and it is proposedbelow that this may be gener- [DePaolo, 1981a; Nelson and DePaolo, 1984, 1985]. ally applicableto most calderas. This conclusionrequires The relatively high ENdvalues of the San Juan and Latir large volumesof basalticmagmas, and evidencefor this is ash-flow tuffs suggesta large mantle componentin the discussedbelow. Finally, the significanceof temporal magmas. Although the Grizzly Peak Tuff has ENdvalues variationsin isotopecompositions of multicyclic caldera that overlap those of average Proterozoiccrust in the re- complexesis discussedin termsof modelsfor masstransfer gion, Johnsonand Fridrich [ 1990] concludethat the mag- in the lithosphereand recentcrustal growth. mas interactedwith crust that had ENd--<-20. Johnsonand Given the evidence that the ash flow tuffs discussed here Thompson[this issue]conclude that the lower crustbeneath are not exclusivelymelts of ancientcrust, three models can the northern Rio Grande rift region must have ENd-< -12, explain the Nd isotopedata for the silicic rocks: (1) evolu- based on studies of contaminated lavas at San Luis Hills. tion primarily by assimilationof continentalcrust and frac- ENdvalues of the southernRocky Mountain ash flow tuffs tional crystallization of mantle-derived basaltic magmas, increasewith decreasingage, both in a regionalcontext and (2) melting of precursor mafic magmas that had stalled within the large multicyclic central caldera cluster of the deep in the crust but had assimilated crust so that their San Juan field. These important temporal variations are isotopic compositionsare midway between the ancient discussed below. crust and mantle, and (3) mixing between mantle-derived Central Area. Volcanic rocks of the Jemez volcanic field magmasand melts of ancient crust. range from mid-Miocene to Quaternaryin age. Early vol- Detailed petrologicand isotopicstudies of calderacom- canismconsisted of-•13 to -•7 Ma basaltthrough rhyolite plexes that contain mafic- and intermediate-composition composition lavas [Smith et al., 1970; Gardner et al., precalderalavas suggestthat AFC involving basalticparen- 1986]. The precalderabasalts and andesiteshave ENdvalues tal magmasmay be the dominantprocess in severallarge- of-1.5 to +0.5 [Loeffier and Futa, 1985], slightly lower volume caldera complexes. Examples include volcanic than thoseof temporally related tholeiite and alkali basalts rocks at Los Humeros [Verma, 1983], Zacatecas [Verma, in the region [Perry et al., 1987]. Younger andesiteand 1984], Kane Springs Wash [Novak, 1985; Novak and dacite lavas of the Tschicoma Formation (-•7 to -•4 Ma) Mahood, 1986; Leeman and Hawkesworth, 1986], have ENdvalues of-3.0 to -0.8, and the 2 Ma E1 Rechuelos Kalamazoo [Gans et al., 1989], Timber Mountain-Oasis rhyolite has ENdvalues between-3.5 and-1.2 [Loeffier and Valley [Farmer et al., 1991], Latir [Johnsonand Lipman, Futa, 1985]. The Otowi Member and TshiregeMember of 1988; Johnsonet al., 1990], San Juan [Lipman et al., 1978; the BandelierTuff eruptedat 1.4 and 1.1 Ma, respectively, Riciputi and Johnson, 1990], and Grizzly Peak [Johnson and represent>600 km3 of mildly peralkalinerhyolite and Fridrich, 1990]. In most of these cases, the calcula- [Smith, 1979]; these magmas were erupted after block tions indicate that the tuffs contain >50 wt % mantle- faulting had establishedthe Rio Grande depression. One derived material. A particularly strong case for AFC of pumice samplefrom the Guaje Pumice Bed of the Otowi parental basaltscan be made for the Grizzly Peak Tuff, Member has an ENdvalue of-0.3, and one pumice sample which is zonedto cogeneticcompositions that are too mafic from the lower part of the Tshirege Member has an ENd to be partial melts of most crustalrocks (57-77 wt % SiO2 value of-2.3 [C. Johnson,unpublished data, 1990]. The [Fridrich and Mahood, 1987; Johnson and Fridrich, 1990]. relatively high ENdvalues of the tuffs, as comparedto Prot- The magma chamber of the Carpenter Ridge Tuff also erozoic basementin New Mexico (Figure 8), suggestsa apparentlycontained silicic and mafic compositions(down large mantle component. to 55 wt % SiO2,Figure 8) that had similar Nd isotoperatios JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHEREMODIFICATION 13,495

8 A INCREASING B MANTLE MCDm , r 1:1 LINE CONTRIBUTION • MEX AT KSW AT /- ' WM ß -..c :..'-1-• /•. e• - •''''J SJ •M'OV.. I.....• b'J/ ',b' . • -12

GPT INCREASING z -16 ANCIENT KAI, GPT/ - CRUSTAL I I . KAI., CONTRIBUTION -20 ß , /, ß , , , , , , -12 -8 -4 0 4 • Nd(BASALT) • Nd(CRUST) Fig.9. (a) Summaryof rangesin Nd isotopecompositions of silicic volcanic rocks (mostly ash flow tuffs)and those of basalticmagmas that erupted contemporaneous with silicicvolcanism or arepostulated to havebeen parental magmasto thesilicic rocks. AT, AmaliaTuff; BT, BishopTuff; GPT, GrizzlyPeak Tuff; J, Jemez;KAL, Kalamazoo; KSW, Kane SpringsWash; MCD, McDermitt;E MEX, easternMexico; W MEX, westernMexico; PST, Peach SpringsTuff; SJ,San Juan; TMOV, TimberMountain-Oasis Valley; WM, WoodsMountains. Dashed lines indicate ashflow tuffs from the northern Rio Granderift. Silicicrocks that have Nd isotopecompositions that plot farthest fromthe 1:1 line probably have the largest crustal components. (b) Summaryof rangesin Nd isotopecompositions of silicicvolcanic rocks (mostly ash flow tuffs) and those of theregional crust. Silicic rocks that have Nd isotope compositionswhich plot farthest from the 1:1line probably have the largest mantle components. Labels and lines as in Figure9a. Datafor the BandelierTuff of the Jemezvolcanic field andPeach Springs Tuff fromC. Johnson (unpublisheddata, 1990).

[Riciputi and Johnson, 1990; L. Riciputi and C. Johnson, values and markedly higher 2ø6pb/2ø4pbratios [Lipman et unpublished data, 1990]. Interpretation of Nd isotope al., 1978; Johnsonet al., 1990; Riciputi and Johnson, 1990; variationsin volcanic centersthat only erupted silicic mag- Colucci et al., this issue]. Although a possibleexception to mas are more ambiguous because the nature of mafic- this argumentcould involve melting of lower crust that is magmadifferentiation is unknown. For example,the silicic similar to that sampled by many high-grade xenoliths, partsof the Grizzly Peak Tuff have very low endvalues (- 11 which have radiogenic Pb isotoperatios [Esperanca et al., to -12; Figure 8) that might have been interpretedas reflect- 1988; Rudnick and Goldstein, 1990; Kempton et al. 1990; ing crustalmelting, had thesevalues not been found in the K.L. Cameron et al., submittedmanuscript, 1991 ], this is an mafic parts of the tuff as well [Johnson and Fridrich, unlikely scenario for ash flow tuffs in the northern Rio 1990]. It remains difficult, however, to estimate the rela- Granderift region. A lower crustthat has high 2ø6pb/2ø4pb tive proportionsof crust and mantle componentsin ash ratios is probably not a major reservoir in this area as flow tuffs that were eruptedthrough basement that may not virtually all mafic- and intermediate-compositionlavas that have a large isotopiccontrast with the mantle, suchas much contain a crustal componenthave low 2ø6pb/2ø4pbratios; of the Sierra Madre Occidental and at least the westernpart this generalization applies to 34 Ma precaldera lavas to 3 of the Mexican Neovolcanic belt. Ma lavas of the Taos Plateau Volcanic field (references Several caldera complexeserupted mafic lavas that have above and Dungan et al. [1986] and Johnson and Thomp- endvalues that are lower than those of the ash flow tuffs, son [this issue]). which is best explained by an AFC model and not crustal Hildreth and Moorbath [ 1988] convincingly demonstrate melting. Examples include the Timber Mountain-Oasis the stronginfluence of the crust on the chemical and isoto- Valley complex [Farmer et al., 1991], the Kane Springs pic compositionsof Andean arc rocks and suggestthat most Wash caldera [Novak, 1985; Leeman and Hawkesworth, of the evolved magmas were generatedby mixing crustal 1986], and the San Juan volcanic field [Riciputi and John- and mantle melts in the lower crust. It seemslikely that the son, 1990; Colucci et al., this issue]. Special pleading is "blending" of crustal melts and more primitive magmas required to explain the relatively high end values of the that Hildreth and Moorbath [1988] envision in their silicic rocks as reflecting melting of contaminatedprecur- "MASH" model occurs in most volcanic fields. Because sor basalts,while simultaneouslyexplaining the lower end the end-membersthat may be involved in mixing are prob- values of the mafic precaldera lavas as the result of ably not seen at the surface, it is difficult to evaluate the crustally contaminatedmantle-derived magmas. significance of this process in the caldera complexes of Calling upon a relatively high-eNdlower crust as an ex- western North America. If magma mixing is the primary planationfor deriving the ash flow tuffs by crustalmelting mechanismfor determiningthe Nd isotopecompositions of predicts that Nd and Pb isotope ratios of volcanic rocks the ash flow tuffs, the tuffs contain mantle componentsthat shouldbe anticorrelated. Magmas that were generatedby are equal to or exceedthat of related intermediate-composi- partial melting of mafic lower crust that was not as LREE tion lavas at the Timber Mountain-Oasis Valley complex, enriched as the exposedcrust should have relatively high the Kane SpringsWash caldera, and the San Juan and Latir •43Nd/•44Ndratios but low 2ø6pb/2ø4pbratios. This contrasts volcanic fields. with the fact that mafic precalderalavas in the Latir and San The largest volumes of basaltic magma are required to Juan volcanic fields have low endvalues and 2ø6pb/2ø4pbgenerate silicic magmasin models 1 and 2 (above). Produc- ratios, whereas ash-flow tuffs have similar or higher end tion of silicic magmasby AFC will require of the order of 13,496 JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHERE MODIFICATION

10 km3 of basalt magma to produce 1 km3 of rhyolite dergone large-scale tectonism, but accompaniedby little magma. Although the contaminatedprecursor magmas magmatism or extension, such as in the Himalaya-Tibet envisionedin model 2 (above) probably involved less ex- orogenic belt (see review by Jarchow and Thompson tensivecrystallization than that requiredin the first model, [1989]). additional basaltic magma is required to melt the contami- Primarywave velocities (Vp) obtained in seismicrefrac- natedprecursor magmas; a mafic:silicicratio of the orderof tion studiesin the Basin and Range province and the Rio 10:1 also seemsappropriate for model 2. Granderift regionindicate that the middle crustcommonly Although smaller volumes of basaltic magma are re- hasVp that is generallylower (~6.2-6.4 km/s) than those of quired to generatesilicic magmasin model 3 (above), the the middle crustunder regionsthat have not been extended, volumes are still substantial. Becausethe enthalpy of fu- suchas the Great Plains [McCarthy and Thompson,1988; sion for common silicate minerals in mafic and silicic rocks Prodehl and Lipman, 1989; Mooney and Braile, 1989]. The are similar within a factor of two [e.g., Robie et al., 1978], lower crustin the Great Basin commonlyappears strongly approximately1 km3 of basalticmagma will be requiredto layeredin reflectionstudies, and has relatively high Vp of generate1 km3 of rhyolitic magma by crustalanatexis; it ~7.3-7.5 km/s in some areas [e.g., Mooney and Braile, seemslikely that larger amountsof basalticmagma often 1989; Thompsonet al., 1989; Smithson,1989]. A high- will be required,because some mafic magmasmay undergo velocity lower crust is missing, however, beneath the only limited crystallization,and substantialheat lossesmay Mojave Desert and Arizona, along the PACE transect occurthrough convection and conduction.Additional ba- [McCarthy and Thompson, 1988]. It is notable that this saltic magma is, of course,required for the mafic compo- region contains few documented Cenozoic calderas, in nent involved in mixing. All three models, therefore, re- comparisonto the Great Basin, althoughnumerous Meso- quire substantialvolumes of mafic magmato generateand zoic calderas have been identified in southern Arizona sustain silicic volcanism at the caldera complexes dis- [e.g., Lipman and Sawyer, 1985]. Klemperer et al. [1986] cussed here. Distinction between these models require andSmithson [1989] notethat a prominentreflector (X) lies detailed petrologic studiesof caldera complexesthat con- ~4 km abovethe Moho in the Great Basin and suggestthat tain premonitory mafic- and intermediate-composition this marks the boundary of a transition zone between the lavas, so that possiblegenetic links betweenthe silicic and crustand mantle. Primary wave velocitiesin the upper 100 mafic parts of the systemcan be tested. km of the mantlebeneath the Basinand Range province are ~7.8 km/s, significantlylower thanthose under the adjacent ColoradoPlateau (-8.0 to 8.2 km/s) [Thompsonand Burke, Evidencefor Basalt UnderplatingFrom Geophysicaland 1974; Hauser and Lundy, 1989; Braile et al., 1989]. These Petrologic Studiesof the Lower Crust characteristicssuggest that the crust beneathlarge Ceno- There is growing consensusthat heat flow and seismic zoic ash flow tuff fields is composedof large volumesof data obtained in the Basin and Range province are best intermediate- to silicic-compositionintrusive rocks in the explained by recent intrusion or underplatingof basaltic middle crust,which are underlainby layered mafic sills in magmas at or near the base of the continental crust. the lower crust,which in turn is underlainby uppermantle Reduced heat flow in the Basin and Range province is thathas anomalously low Vp. generally50-100 % greaterthan that in stablecratons and The relatively flat Moho, high lower crustal seismicve- can be well explainedby underplatingthe crustby mantle- locities, and stronglayered reflections in the lower crustof derived basaltic magmas [Lachenbruchand Sass, 1978]. the Great Basin have been interpreted as indicating em- Lachenbruch and Sass [1978] note that if basaltic magma placementof large volumesof basalticmagmas and related underplatingis the primary mechanismfor producinghigh cumulatesinto the lower crust, which would create a petro- heat flow in the western United States, then intrusion rates logically mixed boundary between the crust and mantle in the Basin and Range province must be of the order of [e.g., Furlong and Fountain, 1986; Thompsonet al., 1989]. flood basalt eruptionrates in the Columbia Plateau. Indeed, Klemperer [ 1989] arguesthat the seismicreflection The reflection Moho is relatively flat at depthsof 30-35 and refraction data obtainedin the Basin and Range prov- km in the Basin and Range province, and this has been ince are best explainedby underplating~3x10 6 km3 of interpretedas indicating that the Moho is relatively young basalticmagmas near the crust-mantleboundary during the [Allmendingeret al., 1983, 1987; Klemperer et al., 1986]. Cenozoic and that significantcrustal growth may occurby The reflection Moho in the Basin and Range cuts across this mechanism. This is equivalent to addition of ~6 km of crustal sectionsthat have had markedly different tectonic mantle-derived magma to the crust, averaged over the and accretionaryhistories, which included large-scalever- ~5x105km 2 area of the preextensionBasin and Range prov- tical movementsin the crust, as well as large variations in ince. Furlong and Fountain [1986] calculate larger addi- the amount of extension recorded in upper crustal rocks tions of basalticmagma at the base of the crust 10-15 km, [Allmendinger et al., 1983, 1987; Klemperer et al., 1986; based on thermal models of magma extraction from the McCarthy and Thompson,1988; Thompsonet al., 1989]. mantle. Unfortunately,geophysical evidence for magmatic Where high-quality reflection and re•l,,•,,,li data are avail- underplatingcannot constrain if underplatingoccurred dur- able for coincident traverses, the reflection and refraction ing caldera-relatedvolcanism in the Great Basin or after Moho occur at the samedepths in the westernUnited States most calderas formed. within error of the determinations [e.g., Mooney and Sufficiently deep crustal exposuresare not available in Brocher, 1987; W. Mooney, personal communication, the westernUnited Statesto test models that involve large- 1989]. The relatively flat Basin and Range Moho contrasts scale intrusioninto the lower crust of mafic magmas.Evi- stronglywith that observedin other regionsthat have un- dence for large mafic intrusionsin the lower crust can be JOHNSON: LARGE-SCALECRUST FORMATIONAND LITHOSPHEREMODIFICATION 13,497 found in other regions [e.g., Fountain and Salisbury, 1981]. tions of the crust, and these changesreflect profound hy- For example, the large Border Ranges mafic-ultramafic bridization of ancient crust and mantle componentsin the complex, Alaska, demonstratesthat some regions of the crustal column. Johnson et al. [1990] calculate that Sr and lower crust are dominatedby cumulatesthat underlie more Nd isotope ratios of the hybridized crust may shift N80% silicic plutonic and volcanic rocks [Burns, 1985]. Ultra- and -•50% closer to those of the mantle, respectively, dur- mafic xenolithsin island arc lavas have been interpretedas ing evolution of the Latir volcanic field, which had a single cumulatesthat were generatedby crystal fractionation of caldera cycle. In contrast,Pb isotoperatios of the hybrid- islandarc magmas[e.g., Debari et al., 1987]. Mass balance ized crust were estimatedto shift only -•20% toward those calculationsbased on cumulate xenolith abundancessug- of the mantle, due to the low Pb contents of mantle-derived gestthat it is reasonableto expect the lower parts of island basaltscompared to those of the crust. arcs to have a large cumulate component [Kay and Kay, Temporal trendsin Nd isotopecompositions for ash flow 1985]. tuffs of the multicyclic central caldera cluster of the San Early studiesof crustal and mantle xenoliths in the west- Juan volcanic field are remarkably similar to those pre- ern United Statesnote the complexity of the lower crust and dicted to occur in the crust based on models for the Latir upper mantle [e.g., McGetchin and Silver, 1972]. volcanic field. As noted by Riciputi and Johnson [1990], McGetchin and Silver suggestthat the crust-mantlebound- •mdvalues of ash flow tuffs from the central caldera cluster ary is a complex and diffuse zone that includes mafic and increaseby-•2 units with decreasingage; this trend can be ultramafic regions in the lower crust, as well as unusually modeledby a shift of-•4 •mdunits of the isotopiccomposi- silicic regionsin the uppermostmantle. Subsequentstudies tion of the crust toward mantle compositionsas a result of of xenolithsfrom other regionsin the westernUnited States continuedinjection of basaltic magmas. Additional data have confirmed these observations[e.g., Wilshire et al., have identified two trends in the central caldera cluster. 1988; Kempton et al., 1990; K.L. Cameron et al., submitted Relatively rapid shifts in Nd isotope compositionsof ash manuscript,1991]. Xenolith studiesfrom other Cenozoic flow tuffs from the San Luis caldera complex may reflect magmatic provinces support models for large-scale rapid injection of mafic magmasinto the lower parts of this underplatingof mantle-derivedbasaltic magmas and asso- multicyclic system (Figure 10). Ash flow tuffs that define ciated cumulate formation near the crust-mantle boundary the younger end of the main trend for tuffs of the central [e.g., Wass and Hollis, 1983; Griffin and O'Reilly, 1987; cluster,as well as thosefor the San Luis caldera complex, Rudnick et al., 1986; Rudnick and Taylor, 1987; Rudnick have •md values that are higher than those of most and Williams, 1987]. precaldera intermediate-composition lavas (Figure 10), Relations of Magma Sourcesand Tectonic Setting stronglysupporting an increasingmantle componentin the crust during evolution of the volcanic field. Regardlessof the tectonicsetting in which large-volume Temporal variations for ash flow tuffs in the northern Rio ashflow tuff eruptionsoccurred in westernNorth America, Grande rift region follow those that occur within the San all of the silicic magmascontained a major mantle compo- Juan volcanic field (Figure 11). Eruption of the Grizzly nent, that in somecases appears to be >50%. Although the PeakTuff occurredcontemporaneous with early precaldera discussionabove suggeststhat large volumes of mantle- volcanism in the nearby San Juan field, and ash flow derivedbasaltic magmas were emplacedinto the crustin all magmatismin the central San Juan calderacluster tempo- regions of caldera formation in western North America, rally overlaps precalderaand caldera-relatedvolcanism at regardlessof the amountof extension,injection of volumi- the nearby Latir volcanic field [Lipman et al., 1970, 1986; nous basaltic magmasis not uniformly correlatedwith the Fridrich et al., 1991]. The youngestcaldera in the SanJuan onsetof extensionand cannot, therefore, be called upon as field erupted 3 m.y. after formation of the Questacaldera in the uniform driving mechanismfor extension [e.g., Gans, the Latir field [Lipman et al., 1970, 1986]. These factors 1987; Fountain, 1989]. The common thread to all caldera indicatethat caldera-relatedmagmatism in the northernRio complexesis that when they are formed, they are funda- Granderift region may be broadly consideredpart of the mentally associatedwith large volumesof basalticmagma same coeval volcanic field [e.g., Steven, 1975]. Similar that were injected into the crust. temporalvariations in isotopiccompositions are well docu- mentedat the Timber Mountain-OasisValley calderacom- plex, where •mdvalues of ash-flow tuffs that did not un- TemporalEvolution and Modification of the Crust dergolate stageroof assimilationincrease 1.2 •Ndunits over Sustained(10-20 m.y. duration) eruption of magma at a 2-m.y. period (Figure 6) [Farmer et al., 1991]. •mdvalues many magmatic centers,including multiple caldera-form- for silicic rocks in the Long Valley area generallyincrease ing eruptions, has led several workers to suggestthat a 4 •mdunits over a 2-m.y. period(Figure 5) [Halliday et al., generalrequirement for developingcaldera complexes is a 1984, 1989; Sampson and Cameron, 1987; Kelleher and large flux of mantle-derivedbasaltic magma into the lower Cameron, 1990]. Although the Peach Springs Tuff and parts of the system[e.g., Christiansenand Lipman, 1972; Wild Horse Mesa Tuff in the easternMojave Desert are not Lachenbruch et al., 1976; Smith, 1979; Hildreth, 1981; part of the samemagmatic system,they probably erupted Spera and Crisp, 1981]. Based on detailed study of the from the sameregion, and the markedly higher •mdvalues Latir volcanic field, which was active for 9 m.y., Johnsonet of the youngerWild Horse Mesa Tuff may reflect an in- al. [1990] calculatethat N29,000 km3 of basaltic magma creasingmantle component in the crustof the regionduring were requiredto generateand sustainthe magmaticsystem. late Cenozoic magmatism(Figure 5) [Musselwhiteet al., Injection of large volumesof basalticmagma into the crust 1989; C. Johnson,unpublished data, 1990]. Further work will producelarge changesin Sr and Nd isotopecomposi- on othermulticyclic calderacomplexes are neededto test if 13,498 JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHERE MODIFICATION

I SANJUAN VOLCANIC FIELD -4 PRECALDERA ß ASH FLOW TUFFS CONEJOS MANTLE /x TUFF-RELATED LAVAS LAVA S -5 34-29

-6

z

-7

-8 ß FCT ORIGINAL PROTEROZOIC CRUST

-9 i ß ß ß I ß ß ß ß ß ß ß ß ß ß ß ß 28.0 27.0 26.0 FREQUENCY AGE (Ma)

Fig. 10. Temporalvariations of ENdvalues for ashflow tuffsand related lavas of thecentral caldera cluster of theSan Juanvolcanic field (symbols) and prechldera lavas of theConejos Formation in thecentral and southeastern parts of thefield (histogram on left). FCT andCRT denote the Fish Canyon Tuff andCarpenter Ridge Tuff, respectively, whichhave variable Nd isotopecompositions that are probably due to latestage crustal interaction. Ash flow tuffs fromthe San Luis caldera define a temporaltrend that is distinctfrom that of themain trend and may reflect the multicyclicnature of theSan Luis caldera. Data from Riciputi and Johnson [1990], Colucci et al. [thisissue], and L. Riciputi and C. Johnson(unpublished data, 1990).

the temporal variationsnoted here are common,particu- of the crustduring continuedinjection of mantle-derived larly in the caldera complexes of the Great Basin, Sierra basalt,large variations in Pb isotoperatios for precaldera Madre Occidental,and Mogollon-Datil volcanicfield. andcaldera-related volcanism and postcaldera ore deposi- Although models presentedby Johnson et al. [1990] tionhave been noted for theSan Juan volcanic field (Figure predict only small changesin the Pb isotopecompositions 12) [Lipmanet al., 1978; Doe et al., 1979; Riciputi and Johnson, 1990; Colucci et al., this issue]. The low 2ø6pb/2ø4pbratios of precalderalavas have been interpreted NORTHERNASHRIO FLOWGRANDE TUFFS RIFT REGION by the above workers to reflect interaction with the lower 2 crust,and the markedlyhigher 2ø6pb/2ø4pb ratios of the ash 0 \ \ \ \ \ \ \ \ \ \ \ \ \ \ flow tuffs are interpretedto reflecta largerupper crustal -2 ORIGINAL LITHOSPHERIC MANTLE component.Riciputi and Johnson[1990] suggestthat the SO AT decreasein 2ø6pb/2ø4pbratios of theyounger ash flow tuffs of the centralcaldera cluster can be interpretedas due to ß'0 -6 """--•': .... transportof nonradiogeniclower crustalPb into the middle z .':*' and upperparts of the magmaticsystem during continued magmatism. This decreasein 2ø6pb/2ø4pbratios of the ash flow tuffscontrasts with thevery high2ø6pb/2ø4pb ratios that IJJ---8 12 GP HY•CRUST characterizepostcaldera ore leads(Figure 12) [Doe et al., -10• EVOLUTIONI 1979] whichpresumably reflect the Pb isotopecomposi- - 16 , ORIGI'NAZ,PROTEROZOIC CRU'$T tions of the upper crust in the vicinity of the ash-flow •5 • • 2• 27 25 magmachambers that had not beenhybridized. ^•E (•o) 2ø6pb/2ø4pbratios for ash flow tuffs in the northern Rio Fig. 11. Variationsin •Ndvalues for ashflow tuffsof thenorthern Rio Granderift regiongenerally increase with decreasingage Granderift regionvolcanic fields with time. GP, Grizzly PeakTuff; (Figure13), suggestingthat the evolvedmagmas interacted $J, San Juan ash flow tuffs of the central caldera cluster; AT, Amalia with progressivelymore evolved, and presumablymore Tuff(Latir volcanic field-Questa caldera). Hybridized crust evolu- shallow, Proterozoic crust. That the younger ash flow tuffs tioncurve reflects changes inisotopic compositions ofa hybridized in the central caldera cluster of theSan Juan field do not mixture of original Proterozoic crust and new mantle-derivedmate- rialbeneath thevolcanic fields; curve from Johnson etal. [1990] and followthis trend may indicate that transfer of lowercrustal variesfrom ENd = -12to ENd = -6 for9 m.y.period (only part of curve Pb to the uppercrust may only occurin multicycliccaldera plottedfor simplicity). Data from Johnson etal. [1990], Riciputi and complexes,where crust hybridization and magma fluxes JohnsonandC. Johnson[1990], (unpublished and Johnson data, and 1990).Fridrich Field[1990] for Originaland L. Protero- Riciputi are concentrated inareally restricted regions. Incontrast to zoicCrust from DePaolo [1981a] and Nelson and DePaolo [1985]. the ash flow tuffs, 2ø6pb/2ø4pb ratiosof mafic- and interme- Fieldfor Original Lithospheric Mantle from Johnson and Thompson diate-composition lavas in the regionremain low (Figure [thisissue]. 13), andthis may reflectrestriction of large-scaleinterac- JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHERE MODIFICATION 13,499

FREQUENCY 6 19.2

19.0 I SANJUAN VOLCANIC FIELD / 18.8 PRE- 18.7 - CALDERA CONEJOS LAVAS 18.6 34-Z9 Ma 18.5'

- 18.4

18.3 ' POST-

. MANTLE COMPOSITION CALDERA 18.2 ORE lEADS 18.1- 25-24 Mc• &• ASHTUFF-RELATED FLOWTUFFS LAVAS 18.0 ß '2.0 .... 2.0 .... 2.0 17.9 - (uo)

Fig. 12. Temporal variationsof 2ø6pb/2ø4pbratios for ash flow tuffs 17.7 and related lavas of the central caldera cluster of the San Juan volcanicfield (symbols)and precaldera lavas of the ConejosForma- tion in all regionsof the San Juanfield (histogramon left) and ore leadsin the centralcaldera cluster (histogram on right). The high 17.5 2ø6pb/2ø4pbratios of early ash flow tuffs are interpretedto reflect ascentof the locus of silicic magmatisminto the upper crust; de- •I•I I I creasesin 2ø6pb/2ø4pbratios for youngertuffs are interpreted to reflect transportof lower crustal Pb into the upper crust during continued FREQUENCY magmatism in the region [Riciputi and Johnson, 1990]. Data from Lipman et al. [1978],Doe et al. [1979],Riciputi and Johnson[1990], and Colucciet al. [ 1991]. Mantle Compositionfor regionfrom Doe et al. [1969] and Johnsonand Thompson[this issue]. tion betweenmantle-derived magmas and crust to the lower becausetheir Sr isotoperatios are very susceptibleto modi- crust. fication by late stageassimilation [e.g., Noble and Hedge, Temporalchanges in O and Sr isotoperatios for ashflow 1969; Johnson, 1989]. tuffs in the northernRio Granderift regionare alsoconsis- It is unlikely that systematictemporal changesin four tent with an increasingmantle component in the hybridized isotope systems(O, Sr, Nd, and Pb) for the tuffs in the crust. g•80 valuesfor ashflow tuff samplesthat did not northernRio Grande rift region is due to significantgeo- interact with meteoric water decrease 2.5 %o with decreas- graphicheterogeneity of the crust. If the variation in O, Sr, ing age for the Grizzly Peak, San Juan, and Latir ash flow and Nd isotoperatios is due to preferentialinteraction with tuffs [Larsonand Taylor, 1986; Johnsonet al., 1990; John- "isotopically primitive" Proterozoic crust to the south, a son and Fridrich, 1990]. Similar, but larger, decreasesin decreasein 2ø6pb/2ø4pbratios is expected,if mafic lower g•80 valuesof 4 %oin a 3-m.y. periodoccurred for ashflow crustthat has low g•aO valuesand 87Sr/a6Sr ratios and high tuffs from the Sonoma volcanic field in western California present-dayENd values is envisionedas the contaminant,as [Johnson and O'Neil, 1984]; these variations are inter- the lowercrust in theregion has low 2ø6pb/2ø4pbratios [e.g., preted as reflecting replacementof the original Mesozoic Johnsonand Thompson,this issue]. Barker et al. [1976a] sedimentarycrust in western California with mafic intru- note that Rb/Sr ratios of 1700 Ma plutonic rocks increase sions that were associated with the late Cenozoic tuffs. north to south from central Colorado to northern New Initial 87Sr/86Srratios of the mostmafic partsof the ash Mexico, suggestingthat the Early Proterozoiccrust beneath flow tuffs in the northernRio Granderift region decrease the Latir field should have higher present-day87Sr/86Sr with decreasingage from 0.7099 for the Grizzly Peak Tuff ratiosthan under the Grizzly Peak caldera. This is opposite (34 Ma [Johnsonand Fridrich, 1990]), to 0.7055-0.7075 to the trend observed for the ash flow tuffs. g•80 and for the San Juantuffs (29-25 Ma [Lipman et al., 1978]), to present-dayENd values are relatively high and low, respec- 0.7057 for the Amalia Tuff (26 Ma) that was eruptedfrom tively, for 1000-1400 Ma granitic rocks in Colorado the Questa caldera of the Latir volcanic field [Johnson et [Barker et al., 1976b; DePaolo, 1981a] and are thoughtto al., 1990]. Cautionmust be usedin interpretingthe initial be the primary contaminantof the magmasthat were paren- 87Sr/86Srratios of the silicic,low-Sr parts of ashflow tuffs, tal to the Grizzly Peak Tuff magma [Johnsonand Fridrich, 13,500 JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHEREMODIFICATION

Johnsonet al. [ 1990] independentlydeveloped a quanti- INORTHERN ASHRIO FLOWGRANDE TUFFSRIFT REGION tative model for lithospheremodification that may occur during evolution of large-volumecaldera complexes, along 19.8 I ORICINAL HYBRIDIZED similar lines as the qualitative model discussedby Lipman 19.4 AVERAGE CRUST AT [1988]. Johnsonet al. concludethat significant transfer of CRUST EVOLUTION i crustal componentsto the upper mantle occurs through

19.0 SJ -- incorporationof crystalcumulates that are generatedduring 18,6 AFC of mantle-derived magmas at the base of the crust.

18,2 This model is based on detailed petrologic and isotopic

\ studiesof the Latir volcanic field, which documenta major 17.8 •_...--...... GP...... ESJ '"'/ mantle-derivedbasalt input into the magmatic system,in addition to recognition that Sr and Pb isotope ratios of 17,40P,J'C/N,,I./., ...... , precalderavolcanism in many volcanic fields provide evi- 17.0 dence for extensive interaction with the lower crust. The 35 33 31 29 27 25 AGE (Mo) uplift problem in the modestly extendedRio Grande rift region is also recognizedin this model and was one reason Fig. 13. Variations in 2ø6pb/2ø4pbratios for volcanicrocks of the for proposingreturn of materialto the mantlein the form of northernRio Grande rift region with time. GP, Grizzly Peak Tuff; cumulates. Expected uplifts associatedwith large-scale SJ, San Juan ash flow tuffs of the central caldera cluster; AT, Amalia magmaticinjection in highly extendedregions suchas the Tuff (Latir volcanic field-Questa caldera); ESJ, early intermediate- Basin and Range provinceshould be significantlysmaller. compositionSan Juanlavas (ConejosFormation); EL, early interme- If basalticmagmas stagnate at the crust-mantleboundary diate-compositionLatir lavas; SLH, early rift mafic lavas at San Luis Hills. Hybridized crust evolution curve from Johnsonet al. [ 1990]. and form crystalcumulates during assimilation,then the Sr, Data from Lipman et al. [1978], Johnsonet al. [1990], Riciputi and Nd, and especially Pb isotope compositionsof the local Johnson[ 1990], Johnsonand Fridrich [ 1990], Johnsonand Thomp- uppermantle will be largely overprintedby crustalcompo- son [this issue], and Colucci et al. [this issue]. sitions. In their model for the Latir field, Johnson et al. [1990] calculate that the mass ratio of magma originally extracted from the mantle relative to the mass of crust 1990]. Although the middle Proterozoicintrusive rocks are returnedto the mantle is approximately5:1, indicatingthat locally more abundantin the Grizzly Peak and San Juan despitecrustal recycling, significantnet crustalgrowth oc- regions,as comparedto the Latir field, theserocks crop out curred on at least a local scale, in addition to significant within 50 km of the Latir field, and U-Pb zircon data crust formation. If the source regions for late Cenozoic indicate that 1400 Ma granites underlie the Latir field at mafic magmas in western North America intersectedthe shallow depths [Johnson, 1989; I. Williams and C. John- contaminated upper mantle, the isotopic compositions son,unpublished data, 1984]. It is alsopossible that g•80 would be similar to thosecommonly interpreted as reflect- values of the crust beneath the Latir field may have been ing derivation from ancient enriched mantle. Localized unusuallyhigh, sinceLaramide age thrustfaults that bound crustalrecycling into the mantleoffers an alternativeexpla- the easternmargin of the Latir field may have thrustPaleo- nation for the low ENdvalues of some basaltic lavas that zoic sedimentary rocks beneath the field [Reed, 1984; erupted after caldera formation, which have been inter- Lipman and Reed, 1989]. It may be significant in this preted by some workers to reflect ancient lithospheric respectthat the only unit in the Latir Field that has a high mantle [e.g., Leeman, 1982; Menzies et al., 1983; Fitton et magmaticg•80 value is the Lucero Peak pluton, which al., 1988; Farmer et al., 1989; Kempton et al., this issue]. intruded in the southernmostpart of the field, possibly outside the main locus of magmatism [Johnson et al., SUMMARY AND CONCLUSIONS 1990]. Middle to late Cenozoic ash flow tuffs and silicic lavas at large-volume volcanic fields in western North America Mass Transfer in the Lithosphere contain a large, commonly dominant, mantle component. Following the ideas of Cox [1980] and Ewart et al. Detailed studiesof severalcaldera complexes indicate that [ 1980] regardingcrustal growth by magmaticunderplating, the silicic magmasfundamentally owe their originsto pro- Arndt and Goldstein [1989] proposethat the crust-mantle tractedassimilation/fractional crystallization of basalticpa- boundary representsnot only an interface for transfer of rental magmas. To a first order, the proportionof mantle- mantle componentsto the crust but also return of crustal derivedcomponents in the ashflow tuffs discussedhere is material to the mantle, if mantle-derivedmagmas stagnate approximatelythe same,regardless of the tectonicsetting near the crust-mantleboundary. The Amdt and Goldstein in which the calderas formed or the extent to which volumi- model envisionscrustal recycling to largely occurby foun- nous high-SiO2magmas were developed. The widespread dering of the crust, and the motivation for this model episodeof Cenozoiccaldera-related magmatism in western largely lies as an alternativemechanism to mantle metaso- North America representsa major period of rapid crust matism for generatingenriched reservoirs in the mantle, as formationthat may rival crustalgrowth ratesin the Protero- well asrecognition of the fact that the large basalticmagma zoic [e.g., Nelson and De?aolo, 1985]. fluxes which Cox [1980] and Ewart et al. [1980] propose Addition of mantle-derivedbasaltic magmasto the crust shouldresult in large-scaleuplift that is not generally ob- resultsin a shift toward more mafic bulk compositionand served. an increasein crustaldensity (Figure 14), particularlyin the JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHERE MODIFICATION 13,501

CALDERA-RELATED MAGMATISM

DENSITY (gm/cm3) 2.6 2.8 3.0 3.2 3.4 O m I •K•, ':'""":'* MAGMAS HIGHU/Pbll I ø ZONEOFOLDER •S,•?•EG• /]'-ZsO!L!•!C• T•lii•GAER ½• •'"';::':';:; !:'::::;

-"187, /86- •::};?'.:.JREGIONOFOLDER I-- • / :•r, ::ir ,'•,'•:i: i: -I'"" STALLED / 40- '"'"'/ / .,," ø ' --;'1-'[-4ø'

60 GROW,• I 5.8I I 6.6I I 7.4 8.2 ( LITHOSPHERIC BASALTICMAGMAS Vp(kin/s)

IOO

Fig. 14. Crosssection through the crust and upper mantle and accompanying depth-Vp-density variations during evolutionof caldera-relatedmagmatism (generally 35-10 Ma) in westernNorth America. Depth to Mohofixed at 40 km for simplicity,although depth to Mohowill varywith the amount of magmaticunderplating and extension. The crust-mantleboundary is envisionedto havebeen relatively sharp prior to caldera-relatedmagmatism. Continued injectionof mantle-derivedmagmas into the crust during evolution at largevolcanic fields is envisionedto increase thedensity of thecrust, resulting in ascentof thelocus of bothmafic and silicic magma chambers with time; this may be reflected in an overall increasein 2ø6pb/2ø4pbratios of ash flow tuffs from the northernRio Grande rift region (see Figure13). Theincrease in averagedensity of thecrust will beaccompanied bya decreasein the average 87Sr/86Sr ratioand increase in theaverage •md value of thecrust as the mantle component in thecrust increases. Return of cumulatematerial to themantle is envisioned(unpatterned arrow) to partiallycompensate for thelarge flux of mantle- derivedbasalt into the crust(solid arrow), although the differencein thesefluxes will producenet crustalgrowth (stippledarrow). The overlapping nature of thesefluxes will producea petrologicallyand seismically transitional Moho(Figure 15). Basis for depth-Vp-density diagram is fromGlazner and Usslet [1988]. lower crust [e.g., Glazner and Usslet, 1989]. These com- magma evolution to higher crustal levels. Assuming a positionalchanges can explain the anomalouslyhigh Vpof generalupward increasein U/Pb ratios of the crust, this the lower crust in some areas of the Basin and Range processmay explain the generalincrease in 2ø6pb/2ø4pb province, as comparedto those of stable cratons [e.g., ratios that is observed for ash flow tuffs in the northern Rio Mooneyand Braile, 1989]. Thesecharacteristics may be Grande rift volcanic fields (Figure 13), which produced accompaniedby significantincreases in SNdvalues and approximately20 calderasover a 12-m.y. period. decreasesin 87Sr/86Srratios of the crust,particularly for the Pondingof parentalbasaltic magmas in the lower crust regionsof the crustthat arehybridized mixtures of original may explainthe commonobservation of horizontalreflec- crust and new mantle-derived rocks. Precaldera mafic- and tors in the lower crust [e.g., Klemperer et al., 1986; intermediate-compositionmagmas are envisioned to as- McCarthyand Thompson,1988; Klemperer, 1989; Thomp- cend into the middle crust and continuehybridization pro- son et al., 1989]. The baseof thesestrong reflectors prob- cessesthat began in the lower crust. Lead isotoperatios for ablybelongs to a petrologicallycomplex transition zone of the crust,including the hybridizedcrust, will be little af- original lower crust and young basalticrocks and their fected, however, due to the low Pb content of mantle- associatedcumulates, as indicatedby combinedgeophysi- derived basalts; Pb isotope ratios, therefore, provide a cal and xenolith studies(Figure 15). Young cumulatesare meansto "see through"the effects of crust hybridization consideredto lie below the Moho, becausethey should and retain somerecord of the original vertical zonationsof haverelatively high seismic velocities. Although the low Vp Pb isotoperatios in the crust. Only at multicycliccaldera in the uppermantle beneath the Basinand Range has been complexeswill magmatismbe sufficientlyfocussed so that explainedas a resultof high heatflow [Blackand Braile, transportof Pb from the lower crust to the upper crust 1982], silicic (crustal)material alsohas low velocities[e.g., occurs[e.g., Riciputi and Johnson,1990]. Continuedin- Rybachand Buntebarth,1982; Fountainand Christensen, creasesin the mean crustal density during evolution of 1989], and local recyclingof crust into the upper mantle large volcanicfields will tendto focusthe zonesof silicic couldexplain the low V,(Figure 15). The low Vpand heat 13,502 JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHEREMODIFICATION

POSTCALDERA MAGMATISM

YOUNGBASALT DENSITY(gm/crn 3)

•L VOLCANICFIELDS 2.6 2.8 :5.0 :5.2 :5.4 0- .B.,D,ZEDC.UST g• •O•;G;NiLC.UST •.•'- ."•'."5

CRUST

x -20 ...... ] uW[1 "W'sr ;:'..•.'. •.• x •E:FLECTO•

40- ! ".'.s.'.'-'-'c:c:ø""'='ø"'='"-'- DoCIo00 0 rl MOHO ' - •k• -•o•-- -40 [3 OCi 0 O0 CU;•ULATE o COMPONENT•,• o rl 0 rl CI CI o o • I• , Ci C]

I I • • • 60 6.6 7.4 8 I I Vp(krn/s) LI•,•,,••..ERE ,eel ASTHENOSPHERE J•ASTHENOSPHERICBASALTIC MAGMAS

Fig. 15. Crosssection through the crust and upper mantle and accompanying depth-Vp-density variations during postcalderamagmatism (generally <10 Ma) in westernNorth America. Large fluxes of mantle-derivedmagmas into the crust during earlier, caldera-relatedmagmatism (Figure 14) is thought to have produceda petrologically and seismicallytransitional Moho that has anomalously high lower crust Vp and low upper mantle Vp. Although the largest volumes of mafic lavas in western North America were generally erupted after caldera-relatedmagmatism, [e.g., Christiansenand Lipman, 1972; Lipman et al., 1972], the largestfluxes of basalticmagma into the crustare thought to have occurredduring earlier, caldera-relatedmagmatism, due to the large mantlecomponent in ashflow tuffs from westernNorth America. Basis for depth-Vp-densitydiagram is fromGlazner and Ussler [1988]. flow beneaththe Sierra Nevada cannot be explainedby Acknowledgmentis also made to the Donors of the Petroleum Re- expectedVp - T relations[Black and Braile, 1982], and searchFund, administeredby the American Chemical Society, for interaction between the voluminous magmas associated additionalsupport of this work. with the batholithand the uppermantle is a possibleexpla- nation. Neodymium isotope data for ash-flow tuffs discussed REFERENCES here provide strong, independentsupport for large-scale Alibert, C., A. Michard, and F. Albar6de, Isotope and trace element basaltinjection that has been postulatedbased on heat flow geochemistry of Colorado Plateau volcanics, Geochim. data [e.g.,Lachenbruch and Sass,1978], seismicdata [e.g., Cosmochim. Acta, 50, 2735-2750, 1986. Furlong and Fountain, 1986; McCarthy and Thompson, Allmendinger, R.W., J.W. Sharp, D.V. Tish, L. Serpa, L. Brown, S. 1988], studieson the origin of lower crustalxenoliths [e.g., Kaufman, J. Oliver, and R.B. Smith, Cenozoic and Mesozoic structureof the eastern Basin and Range province, Utah, from Griffin and O'Reilly, 1987] and granulite-faciesmetamor- COCORP seismic-reflectiondata, Geology, 11,532-536, 1983. phic rocks [e.g., Bohlen and Mezger, 1989], the high 3He Allmendinger, R.W., K.D. Nelson, C.J. Potter, M. Barazangi, L.D. contents that are observed in extensional zones [e.g., Brown, and J.E. Oliver, Deep seismicreflection characteristicsof O'Nions and Oxburgh, 1988], and modelsfor crustalexten- the continentalcrust, Geology, 15, 304-310, 1987. sion [Okaya and Thompson, 1986; Gans, 1987; Fountain, Armstrong,R.L., and R.E. Higgins, K-Ar dating of the beginningof Tertiary volcanism in the Mojave Desert, California, Geol. Soc. 1989]. Am. Bull., 84, 1095-1100, 1973. Armstrong, R.L., E.B. Ekren, E.H. McKee, and D.C. Noble, Space- Acknowledgments.The author is indebtedto Peter Lipman and time relations of Cenozoic silicic volcanism in the Great Basin of Gerry Czamanske for many rousing discussionsin the field, the the western United States, Am. J. Sci., 267, 478-490, 1969. office, and throughthe mail on the origin and evolutionof caldera- Arndt, N.T., and S.L. Goldstein, An open boundary between lower related magmatism; Lipman is thanked for a detailed review of an continental crust and mantle: Its role in crust formation and early versionof the manuscript.Allen Glaznerand Walter Mooney crustalrecycling, Tectonophysics,161,201-212, 1989. also reviewed an early version of the manuscript. Journalreviews Bailey, R.A., G.B. Dalrymple, and M.A. Lanphere, Volcanism, were provided by Ken Cameron and an anonymousreviewer; in structure, and geochronology of Long Valley caldera, Mono addition,Eric Christiansenis thankedfor an exceptionallythorough County, California, J. Geophys.Res., 81,725-744, 1976. review. JaneNielson providedsamples of the PeachSprings Tuff. Barker, F., J.G. Arth, Z.E. Peterman, and I. Friedman, The 1.7- to Michelyn Hass cheerfully enduredmultiple revisionsof the manu- 1.8-b.y.-old trondhjemitesof southwesternColorado and northern script, and Paul Dombrowski drafted Figures 1, 14, and 15. This New Mexico: Geochemistryand depthsof genesis,Geol. Soc.Am. work was supportedby NSF grantsEAR-861836 and EAR-8803892. Bull., 87, 189-198, 1976a. JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHERE MODIFICATION 13,503

Barker,F., C.E. Hedge,H.T. Millard, Jr., andJ.R. O'Neil, PikesPeak Cater, F.W., D.M. Pinckney, W.B. Hamilton, R.L. Parker, R.D. batholith: Geochemistryof some minor elementsand isotopes, Weldin, T.J. Close, and N.T. Zilka, Mineral resources of the Idaho and implicationsfor magmagenesis, Prof. Contrib. Colo. Sch. Primitive area and vicinity, Idaho, U.S. Geol. Surv. Bull., 1304, Mines, 8, 44-53, 1976b. 431 pp., 1973. Bennett, V.C., and D.J. DePaolo, Proterozoic crustal history of the Cather, S.M., and B.D. Johnson,Eocene depositional systemsand westernUnited Statesas determinedby neodymiumisotopic map- tectonics in west-central New Mexico and eastern Arizona, AAPG ping. Geol. Soc.Am. Bull., 99, 674, 1987. Mem., 41,623-652, 1986. Ben Othman, D., M. Polv•, and C.J. All•gre, Nd-Sr isotopiccompo- Christiansen,R.L., and H.R. Blank, Jr., Volcanic stratigraphyof the sitionof granulitesand constraintson the evolutionof the lower Quaternaryrhyolite plateau in Yellowstone National Park, U.S. continental crust, Nature, 307, 510-515, 1984. Geol. Surv. Prof. Pap., 729B, 18 pp., 1972. Best, M.G., and E.H. Christiansen,Limited extension during peak Christiansen,R.L., and P.W. Lipman, Cenozoicvolcanism and plate- Tertiaryvolcanism, Great Basin of Nevadaand Utah, J. Geophys. tectonic evolution of the western United States, II, Late Cenozoic, Res., this issue. Philos. Trans. R. Soc. London, Set. A, 271,249-284, 1972. Best, M.G., E.H. Christiansen, A.L. Deino, C.S. Gromme, E.H. Christiansen,R.L., and D.C. Noble, Geologic map of the Trail Ridge McKee, and D.C. Noble, Eocenethrough Miocene volcanismin quadrangle,Nye County,Nevada, U.S. Geol. Surv.Map, GQ-774, the Great Basin of the western United States, Mem. N.M. Bur. 1968. Mines Miner. Resour., 47, 91-133, 1989. Christiansen,R.L., P.W. Lipman, W.J. Carr, F.M. Byers, Jr., P.P. Black,P.R., and L.W. Braile,Pn velocity and cooling of thecontinen- Orkild, and K.A. Sargent,Timber Mountain-OasisValley caldera tal lithosphere,J. Geophys.Res., 87, 10,557-10,568, 1982. complex of southernNevada, Geol. Soc. Am. Bull., 88, 943-959, Boardman,S.J., and K.C. Condie,Early Proterozoicbimodal volca- 1977. nic rocks in central Colorado, U.S.A., part II, Geochemistry, Colucci, M.T., M.A. Dungan, K.M. Ferguson,P.W. Lipman, and S. petrogenesisand tectonicsetting, Precambrian Res., 34, 37-68, Moorbath, Precaldera lavas of the southeast San Juan volcanic 1986. field: Parent magmasand crustal interactions,J. Geophys.Res., Bohlen, S.R., and K. Mezger, Origin of granulite terranesand the this issue. formation of the lowermost continental crust, Science, 244, 326- Condie, K.C., Plate-tectonics model for continental accretion in the 329, 1989. southwesternUnited States,Geology, 10, 37-42, 1982. Bowen, N.L., The later stagesof the evolutionof the igneousrocks, Condie, K.C., and A.J. Budding, Geology and geochemistryof Pre- J. Geol., suppl., 23, 1-91, 1915. cambrian rocks, central and south-central New Mexico, Mem. N. Braile, L.W., W.J. Hinze, R.R.B. von Frese, and G.R. Keller, Seismic M. Bur. Mines Miner. Resour., 35, 1-58, 1979. propertiesof the crustand uppermost mantle of the conterminous Condie, K.C., and T.P. McCrink, Geochemistryof Proterozoicvolca- United Statesand adjacentCanada, Mem. Geol. Soc. Am., 172, nic and granitic rocks from the Gold Hill-Wheeler Peak area, 655-680, 1989. northern New Mexico, Precambrian Res., 19, 141-166, 1982. Broxton,D.E., R.G. Warren, F.M. Byers, and R.B. Scott, Chemical Condie, K.C., and J.A. Nuter, Geochemistry of the Dubois green- andmineralogic trends within the Timber Mountain-OasisValley stonesuccession: An Early Proterozoicbimodal volcanic associa- calderacomplex, Nevada: Evidencefor multiple cyclesof chemi- tion in west-central Colorado, Precambrian Res., 15, 131-155, cal evolution in a long-lived silicic magma system,J. Geophys. 1981. Res., 94, 5961-5985, 1989. Condie, K.C., G.P. Bowling, and R.K. Vance, Geochemistryand Bryant,B., Evolutionand Early Proterozoichistory of the marginof originof earlyProterozoic supracrustal rocks, Dos CabezasMoun- the Archean continent in Utah, in Metamorphism and Crustal tains, southeasternArizona, Geol. Soc. Am. Bull., 96, 655-662, Evolutionof the WesternUnited States, Rubey Volome 7, editedby 1985. W.G. Ernst,pp. 432-445, Prentice-Hall,Englewood Cliffs, N.J., Cox, K.J., A model for flood basalt vulcanism.,J. Petrol., 21,629- 1988. 650, 1980. Bums, L.E., The Border Ranges ultramafic and mafic complex, Crisp, J.A., Rates of magmaemplacement and volcanic output,J. south-central Alaska: Cumulate fractionates of island-arc Volcanol. Geotherm. Res., 20, 177-211, 1984. volcanics, Can. J. Earth Sci., 22, 1020-1038, 1985. Crowe, B.M., K.H. Wohletz, D.T. Vaniman, E.Gladney, and N. Byers, F.M., Jr., W.J. Carr, R.L. Christiansen,P.W. Lipman, P.P. Bower, Status of volcanic hazard studies for the Nevada waste Orkild, and W.D. Quinlivan, Geologic map of the Timber Moun- storageinvestigations, vol. 2, Los AlamosNatl. Lab. Rep., LA- tain caldera area, Nye County, Nevada, U.S. Geol. Surv. Misc. 9325-MS, 101 pp., 1986. Geol. Invest. Map, 1-891, 1976a. Daly, R.A., IgneousRocks and Their Origin,563 pp., McGraw-Hill, Byers, F.M., Jr., W.J. Carr, P.P. Orkild, W.D. Quinlivan, and K.A. New York, 1914. Sargent,Volcanic suitesand related cauldronsof the Timber DeBari, S., S.M. Kay, and R.W. Kay, Ultramafic xenolithsfrom Mountain-OasisValley caldera complex, southernNevada, U.S. AdagdakVolcano, Adak, Aleutian Islands,Alaska: Deformed Geol. Surv. Prof. Pap., 919, 70 pp., 1976b. igneouscumulates from the Moho of an islandarc, J. Geol., 95, Byers, F.M., Jr., W.J. Carr, and P.P. Orkild, Volcanic centersof 329-341, 1987. southwesternNevada: Evolution of understanding,1960-1988, J. DePaolo, D.J., Neodymium isotopesin the Colorado Front Range Geophys.Res., 94, 5908-5924, 1989. and crust-mantle evolution in the Proterozoic, Nature, 291, 193- Cameron,K.L., and M. Cameron,Rare earth element, S7Srff6Sr,and 196, 1981a. 143Nd/144Ndcompositions of Cenozoicorogenic dacites from Baja DePaolo, D.J., A neodymiumand strontiumisotopic study of the California, northwesternMexico, and adjacent west Texas: Evi- Mesozoic calc-alkalinegranitic batholithsof the Sierra Nevada dencefor the predominanceof a subcrustalcomponent, Contrib. and PeninsularRanges, California, J. Geophys.Res., 86, 10,470- Mineral. Petrol., 91, 1-11, 1985. 10,488, 1981b. Cameron, K.L., and J.V. Robinson, Comments on "Nd-Sr isotopic DePaolo, D.J., Trace element and isotopic effects of combined compositionsof lowercrustal xenoliths -- Evidencefor the origin wallrock assimilationand fractionalcrystallization, Earth Planet. of mid-Tertiary felsic volcanics in Mexico", by J. Ruiz, P.J. Sci. Lett., 53, 189-202, 1981c. Patchett, and R.J. Arculus, Contrib. Mineral. Petrol., 104, 609- Doe, B.R., P.W. Lipman, C.E. Hedge, and H. Kurasawa,Primitive 614, 1990. and contaminatedbasalts from the southernRocky Mountains, Cameron, K.L., G.J. Nimz, D. Kuentz, S. Niemeyer, and S. Gunn, U.S.A., Contrib. Mineral. Petrol., 21,142-156, 1969. Southern Cordilleran basaltic andesite suite, southern Chihuahua, Doe, B.R., T.A. Steven, M.H. Delevaux, J.S. Stacey, P.W. Lipman, Mexico: A link betweenTertiary continentalarc and flood basalt and F.S. Fisher, Genesisof ore depositsin the San Juanvolcanic magmatismin North America,J. Geophys.Res., 94, 7817-7840, field, southwesternColorado: Lead isotope evidence, Econ. 1989. Geol., 74, 1-26, 1979. Campa, M.F., and P.J. Coney, Tectono-stratigraphicterranes and Dungan, M.A., M.M. Lindstrom,N.J. McMillan, S. Moorbath, J. mineral resource distributions in Mexico, Can. J. Earth Sci., 20, Hoefs,and L.A. Haskin,Open system magmatic evolution of the 1040-1051, 1983. Taos Plateau volcanic field, northernNew Mexico, 1, The petrol- Carlson,R.W., and W.K. Hart, Crustal genesison the Oregon Pla- ogy and geochemistryof the ServilletaBasalts, J. Geophys.Res., teau, J. Geophys.Res., 92, 6191-6206, 1987. 91, 5999-6028, 1986. 13,504 JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHEREMODIFICATION

Dungan, M.A., R.A. Thompson,J.S. Stormer, and J.M. O'Neill, Rio Furlong,K.P., andD.M. Fountain,Continental crustal underplating: Grande rift volcanism: Northeastern Jemez zone, New Mexico, Thermal considerationsand seismic-petrologicconsequences, J. Mem. N.M. Bur. Mines Miner. Resour., 46, 435-486, 1989. Geophys.Res., 91, 8285-8294, 1986. Ekren, E.B., C.L. Rogers, R.E. Anderson, and P.P. Orkild, Age of Gans,P.B., An open-system,two-layer crustal stretching model for Basin and Range normal faults in Nevada Test Site and Nellis Air the eastern Great Basin, Tectonics, 6, 1-12, 1987. Force Range, Nevada, Mem. Geol. Soc.Am., 110, 247-250, 1968. Gans, P.B., G.A. Mahood, and E. Schermer, Synextensional Ekren, E.B., R.E. Anderson,C.L. Rogers,and D.C. Noble, Geology magmatismin the Basin and RangeProvince: A casestudy from of northernNellis Air Force Base Bombing and Gunnery Range, the easternGreat Basin,Spec. Pap. Geol. Soc.Am., 233, 53 pp., Nye County, Nevada, U.S. Geol. Surv. Prof. Pap., 651, 91 pp., 1989. 1971. Gardner,J.N., F. Goff, S. Garcia, and R.C. Hagan, Stratigraphic Elston, W.E., Mid-Tertiary ash flow tuff cauldrons,southwestern relationsand lithologic variations in the Jemezvolcanic field, New New Mexico, J. Geophys.Res., 89, 8733-8750, 1984. Mexico, J. Geophys.Res., 91, 1763-1778, 1986. Epis, R.C., and C.E. Chapin, Geomorphicand tectonicimplications Glazner, A.F., Recycling of continentalcrust in Miocene volcanic of the post-Laramide,late Eoceneerosion surface in the southern rocks from the Mojave block, southernCalifornia, Mem. Geol. Rocky Mountains, Mem. Geol. Soc. Am., 144, 45-74, 1975. Soc. Am., 174, 147-168, 1990. Esperanca,S., R.W. Carlson, and S.B. Shirey, Lower crustalevolu- Glazner,A.F., and J.A. Supplee,Migration of Tertiary volcanismin tion undercentral Arizona: Sr, Nd and Pb isotopicand geochemi- the southwestern United States and subduction of the Mendocino cal evidence from the mafic xenoliths of Camp Creek, Earth fracture zone, Earth Planet. Sci. Lett., 60, 429-436, 1982. Planet. Sci. Lett., 90, 26-40, 1988. Glazner, A.F., and W. Ussler III, Trappingof magmaat midcrustal Esperanca,S., R.W. Carlson, and S.B. Shirey, Reply to commentof densitydiscontinuities, Geophys. Res. Lett., 15, 673-675, 1988. C.M. Johnson on "Lower crustal evolution under central Arizona: Glazner,A.F., andW. UsslerIII, Crustalextension, crustal density, Sr, Nd and Pb isotopic and geochemicalevidence from the mafic and the evolution of Cenozoic magmatismand the Basin and xenoliths of Camp Creek", Earth Planet Sci. Lett., 99, 406-409, Range of the westernUnited States,J. Geophys.Res., 94, 7952- 1990. 7960, 1989. Ewart, A., K. Baxter, and J.A. Ross,The petrologyand petrogenesis Glazner. A.F., J.E. Nielson, K.A. Howard, and D.M. Miller, Correla- of the Tertiary anorogenicmafic lavas of southernand central tionof thePeach Springs Tuff, a large-volumeMiocene ignimbrite Queensland,Australia m Possibleimplications for crustal thick- sheetin California and Arizona, Geology,14, 840-843, 1986. ening, Contrib. Mineral. Petrol., 75, 129-152, 1980. Griffin, W.L., and S.Y. O'Reilly, Is the continentalMoho the crust- Farmer, G.L., A Sm-Nd isotopic study of Precambrian/Cambrian mantleboundary?, Geology, 15, 241-244, 1987. sedimentaryprovenance in the Great Basin and implications for Hagstrum,J.T., and P.W. Lipman, Paleomagnetismof the structur- the tectonic evolution of the western U.S. (abstract), Geol. Soc. ally deformedLatir volcanic field, northern New Mexico: Rela- Am. Abstr. Programs, 17, 578, 1985. tionsto formationof the Questacaldera and development of the Farmer, G.L., and D.J. DePaolo, Origin of Mesozoicand Tertiary Rio Granderift, J. Geophys.Res., 91, 7383-7402, 1986. granite in the westernUnited States and implicationsfor pre- Halliday, A.N., A.E. Fallick, J. Hutchinson,and W. Hildreth, ANd, Mesozoic crustal structure,1, Nd and Sr isotopic studiesin the Sr, and O isotopicinvestigation into the causesof chemicaland geoclineof the northernGreat Basin,J. Geophys.Res., 88, 3379- isotopiczonation in the BishopTuff, California,Earth Planet. Sci. 3401, 1983. Lett., 68, 379-391, 1984. Farmer, G.L., and D.J. DePaolo, Origin of Mesozoicand Tertiary Halliday,A.N., G.A. Mahood,P. Holden,J.M. Metz, T.J. Dempster, granite in the westernUnited Statesand implicationsfor pre- andJ.P. Davidson, Evidence for longresidence times of rhyolitic Mesozoic crustal structure, 2, Nd and Sr isotopic studies of magmain theLong Valley magmaticsystem: The isotopicrecord unmineralizedand Cu- andMo-mineralized granite in the Precam- in precalderalavas of GlassMountain, Earth Planet. Sci. Lett., 94, brian craton,J. Geophys.Res., 89, 10,141-10,160, 1984. 274-290, 1989. Farmer, G.L., F.V. Perry, S. Semken,B. Crowe, D. Curtis, and D.J. Hart, W.K., Chemicaland isotopicevidence for mixing between DePaolo,Isotopic evidence on the structureand origin of subcon- depleted and enriched mantle, northwestern U.S.A., Geochim. tinentallithospheric mantle in southernNevada, J. Geophys.Res., Cosmochim.Acta, 49, 131-144, 1985. 94, 7885-7898, 1989. Hauser,E.C., and J. Lundy, COCORP deepreflections: Moho at 50 Farmer, G.L., D.E. Broxton, R.G. Warren, and W. Pickthom, Nd, Sr, km (16 s) beneaththe ColoradoPlateau, J. Geophys.Res., 94, and O isotopic variations in metaluminous ash-flow tuffs and 7071-7081, 1989. related volcanic rocks at the Timber Mountains/OasisValley Henry,C.D., andJ.G. Price,Variations in calderadevelopment in the caldera complex, SW Nevada: Implicationsfor the origin and Tertiaryvolcanic field of Trans-PecosTexas, J. Geophys.Res., 89, evolutionof large volume silicic magmabodies, Contrib. Mineral. 8765-8786, 1984. Petrol., in press, 1991. Hildreth,W., The BishopTuff: Evidencefor theorigin of composi- Ferriz, H., and G.A. Mahood, Eruption rates and compositional tional zonationin silicicmagma chambers, Spec. Pap. Geol. Soc. trends at Los Humeros volcanic center, Puebla, Mexico, J. Am., 180, 43-75, 1979. Geophys.Res., 89, 8511-8524, 1984. Hildreth,W., Gradientsin silicicmagma chambers: Implications for Fitton, J.G., D. James, P.D. Kempton, D.S. Ormerod, and W.P. lithosphericmagmatism, J. Geophys.Res., 86, 10,153-10,192, Leeman, The role of lithosphericmantle in the generationof late 1981. Cenozoic basic magmas in the western United States,J. Petrol., Hildreth, W., and S. Moorbath, Crustal contributions to arc Spec.Lithosphere Issue, 331-349, 1988. magmatism in the Andes of central Chile, Contrib. Mineral. Fountain, D.M., Growth and modification of lower continental crust Petrol., 98, 455-489, 1988. in extended terrains: The role of extension and magmatic Hillhouse, J.W., and R.E. Wells, Magnetic fabric, flow directions, underplating,in Properties and Processesof Earth's Lower Crust, and sourcearea of the lower Miocene Peach SpringsTuff in Geophys.Monogr. Ser., vol. 51, editedby R.F. Mereu, S. Mueller, Arizona,California, and Nevada, J. Geophys.Res., in press,1991. and D.M. Fountain,pp. 287-299, AGU, Washington,D.C., 1989. Holmes,A., The origin of igneousrocks, Geol. Mag., 69, 543-558, Fountain,D.M., and N.I. Christensen,Composition of the continen- 1932. tal crust and upper mantle: A review, Mem. Geol. Soc. Am., 172, Jarchow,C.M., and G.A. Thompson,The natureof the Mohorovi•i6 711-742, 1989. discontinuity,Annu. Rev. Earth Planet. Sci., 17, 475-506, 1989. Fountain,D.M., andM.H. Salisbury,Exposed cross-sections through Johnson,C.M., Isotopiczonations in silicicmagma chambers, Geol- the continentalcrust: Implicationsfor crustalstructure, petrology, ogy, 17, 1136-1139, 1989. and evolution, Earth Planet. Sci. Lett., 56, 263-277, 1981. Johnson, C.M., Comment on "Lower crustal evolution under central Fridrich, C.J., and G.A. Mahood, Compositionallayers in the zoned Arizona: Sr, Nd, andPb isotopicand geochemical evidence from magmachamber of the Grizzly Peak Tuff, Geology,15, 299-303, the mafic xenoliths of Camp Creek", by S. Esperanca,R.W. 1987. Carlson, and S.B. Shirey, Earth Planet. Sci. Lett., 99, 400-405, Fridrich, C.J., R.P. Smith, E. DeWitt, and E.H. McKee, Structural, 1990. eruptive, and intrusive evolution of the Grizzly Peak caldera, Johnson,C.M., and C.J. Fridrich, Non-monotonic chemical and O, SawatchRange, Geol. Soc.Am. Bull., in press,1991. Sr, Nd, andPb isotopezonations and heterogeneity in themafic- to JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHEREMODIFICATION 13,505

silicic-compositionmagma chamber of the Grizzly Peak Tuff, Leeman, W.P., Tectonic and magmatic significance of strontium Colorado, Contrib. Mineral. Petrol., 105, 677-690, 1990. isotopic variations in Cenozoic volcanic rocks from the western Johnson,C.M., andP.W. Lipman, Origin of metaluminousand alka- United States, Geol. Soc. Am. Bull., 93,487-503, 1982. line volcanic rocks of the Latir volcanic field, northern Rio Grande Leeman, W.P., and Hawkesworth, C.J., Open magma systems:Trace rift, New Mexico, Contrib. Mineral. Petrol., 100, 107-128, 1988. element and isotopic constraints,J. Geophys.Res., 9!, 5901-5912, Johnson,C.M., and J.R. O'Neil, Triple junction magmatism: A 1986. geochemicalstudy of Neogenevolcanic rocks in westernCalifor- Lipman, P.W., The Miocene Questa caldera, northern New Mexico: nia, Earth Planet. Sci. Lett., 71,241-262, 1984. Relation to batholith emplacementand associatedmolybdenum Johnson,C.M., and R.A. Thompson,Isotopic composition of Oligo- mineralization, in The Genesis of Rocky Mountain Ore Deposits: cene mafic volcanic rocks in the northern Rio Grande rift: Evi- Changes with Time and Tectonics, pp. 133-148, Denver Region dence for contributions of ancient intraplate and subduction Exploration GeologistsSociety, Denver, Colo., 1983. magmatismto evolutionof the lithosphere,J. Geophys.Res., this Lipman, P.W., The roots of ash flow calderasin westernNorth issue. America: Windows into the tops of granitic batholiths, J. Johnson,C.M., P.W. Lipman, and G.K. Czamanske,H, O, Sr, Nd, Geophys.Res., 89, 8801-8841, 1984. and Pb isotope geochemistryof the Latir volcanic field and Lipman, P.W., Evolution of silicic magma in the upper crust: The cogeneticintrusions, New Mexico, and relationsbetween evolu- mid-Tertiary Latir volcanic field and its cogeneticgranitic batho- tion of a continental magmatic center and modifications of the lith, northern New Mexico, U.S.A., Trans. R. Soc. Edinburgh lithosphere,Contrib. Mineral. Petrol., 104, 99-124, 1990. Earth Sci., 79, 265-288, 1988. Kay, S.M., and R.W. Kay, Role of crystalcumulates and the oceanic Lipman, P.W., Oligocene-Miocene San Juan volcanic field, Colo- crust in the formation of the lower crust of the Aleutian arc, rado, Mem. N.M. Bur. Mines Miner. Resour., 46, 303-380, 1989. Geology,13,461-464, 1985. Lipman, P.W., and J.C. Reed, Jr., Geologic map of the Latir volcanic Kelleher, P.C., and K.L. Cameron,The geochemistryof the Mono field and adjacent areas, northern New Mexico, U.S. Geol. Surv. Craters-Mono Lake Islands volcanic complex, easternCalifornia, Misc. Invest. Map, I-! 907, ! 989. J. Geophys.Res., 95, 17,643-17,659, 1990. Lipman, P.W., and D.A. Sawyer, Mesozoic ash-flow caldera frag- Kempton, P.D., M.A. Dungan, and D.P. Blanchard, Petrology and ments in southeasternArizona and their relation to porphyry cop- geochemistry of xenolith-bearing alkalic basalts from the per deposits,Geology, !3, 652-656, 1985. Geronimo volcanic field, southeast Arizona: Evidence for Lipman, P.W., T.A. Steven, and H.H. Mehnert, Volcanic history of polybaricfractionation and implications for mantleheterogeneity, the San Juan Mountains, Colorado, as indicated by potassium- Spec.Pap. Geol. Soc. Am., 215, 347-370, 1987. argon dating, Geol. Soc. Am. Bull., 8!, 2329-2352, 1970. Kempton, P.D., R.S. Harmon, C.J. Hawkesworth, and S. Moorbath, Petrology and geochemistryof lower crustal granulites from the Lipman, P.W., H.J. Prostka, and R.L. Christiansen,Cenozoic volca- Geronimo volcanic field, southeastern Arizona, Geochim. nism and plate-tectonic evolution of the western United States, I, Cosmochim. Acta., 54, 3401-3426. 1990. Early and middle Cenozoic, Philos. Trans. R. Soc. London, Ser. A, Kempton, P.D., J.G. Fitton, C.J. Hawkesworth, and D.S. Ormerod, 27!, 217-248, 1972. Isotopic and trace element constraintson the composition and Lipman, P.W., B.R. Doe, C.E. Hedge, and T.A. Steven, Petrologic evolution of the lithosphere beneath the southwesternUnited evolution of the San Juan volcanic field, southwestern Colorado: States,J. Geophys.Res., this issue. Pb and Sr isotopeevidence, Geol. Soc.Am. Bull., 89, 59-82, 1978. Kistler, R.W., and Z.E. Peterman, Variations in Sr, Rb, K, Na, and Lipman, P.W., H.H. Mehnert, and C.W. Naeser, Evolution of the initial Sr87/Sr86in Mesozoicgranitic rocks and intruded wall rocks Latir volcanic field, northern New Mexico, and its relation to the in central California, Geol. Soc. Am. Bull., 84, 3489-3512, 1973. Rio Grande rift, as indicatedby potassium-argonand fission track Kistler, R.W., and Z.E. Peterman, Reconstruction of crustal blocks of dating,J. Geophys.Res., 9!, 6329-6345, 1986. California on the basisof initial strontiumisotopic compositions Loeffler, B.M., and K. Futa, Sr and Nd isotope systematicsof the of Mesozoic granitic rocks, U.S. Geol. Surv. Prof. Pap., !07!, 17 Jemez volcanic field (abstract), Eos Trans. AGU, 66, 1110, 1985. pp., 1978. Lum, C.C., W.P. Leeman, K.A. Foland, J.A. Kargel, and J.G. Fitton, Klemperer,S.L., Deep seismicreflection profiling and the growth of Isotopic variations in continentalbasaltic lavas as indicators of the continentalcrust, Tectonophysics,!6!, 233-244, 1989. mantle heterogeneity:Examples from the westernU.S. Cordillera, Klemperer, S.L., T.A. Hauge, E.C. Hauser, J.E. Oliver, and C.J. J. Geophys.Res., 94, 7871-7884, 1989. Potter, The Moho in the northern Basin and Range province, Mackin, J.H., Structuralsignificance of Tertiary volcanic rock in Nevada, along the COCORP 40øN seismic-reflection transect, southwestern Utah, Am. J. Sci., 258, 81-131, 1960. Geol. Soc. Am. Bull., 97, 603-618, 1986. Mahood, G.A., Chemical evolution of a Pleistocenerhyolitic center: Knoper, M.W., and K.C. Condie, Geochemistryand petrogenesisof Sierra La Primavera, Jalisco, Mexico, Contrib. Mineral. Petrol., Early Proterozoic amphibolites,west-central Colorado, U.S.A., 77, 129-149, 1981. Chem. Geol., 67, 209-225, 1988. Mahood,G.A., andA.N. Halliday, Generationof high-silicarhyolite: Lachenbruch,A.H., and J.H. Sass,Models of an extendinglithos- ANd, Sr, and O isotopicstudy of Sierra La Primavera,Mexican phereand heat flow in the Basin and Rangeprovince, Mem. Geol. Neovolcanic Belt, Contrib. Mineral. Petrol., !00, 183-191, 1988. Soc. Am., !52,209-250, 1978. McCarthy, J., and G.A. Thompson, Seismic imaging of extended Lachenbruch,A.H., M.L. Sorey,R.E. Lewis, and J.H. Sass,Geother- crustwith emphasison the westernUnited States,Geol. Soc. Am. mal setting and simple heat conductionmodels for the Long Bull., !00, 1361-1374, 1988. Valley caldera,J. Geophys.Res., 8!, 769-785, 1976. McCurry, M., Geology and petrologyof the Woods Mountainsvol- Larson,P.B., andH.P. Taylor, Jr., •80/•60 ratiosin ash-flowtuffs and canic center,southeastern California: Implicationsfor the genesis lavasfrom the centralNevada caldera complex and the centralSan of peralkaline rhyolite ash flow tuffs, J. Geophys. Res., 93, Juan caldera complex, Colorado, Contrib. Mineral. Petrol., 92, 14,835-14,855, 1988. 146-156, 1986. Leat, P.T., R.N. Thompson,M.A. Morrison, G.L. Hendry, and A.P. McDowell, F.W., and S.E. Clabaugh, Ignimbrites of the Sierra Dickin, Compositionally-diverse Miocene-Recent rift-related Madre Occidental and their relation to the tectonic history of magmatismin northwestColorado: Partial melting, andmixing of westernMexico, Spec. Pap. Geol. Soc. Am., !80, 113-124, 1979. mafic magmasfrom 3 different asthenosphericand lithospheric McGetchin, T.R., and L.T. Silver, A crustal-uppermantle model for mantle sources,J. Petrol., Spec.Lithosphere Issue, 351-377, 1988. the Colorado Plateau based on observationsof crystalline rock Leat, P.T., R.N. Thompson, A.P. Dickin, M.A. Morrison, and G.L. fragments in the Moses rock dike, J. Geophys. Res., 77, 7022- Hendry, Quaternaryvolcanism in northwesternColorado: Impli- 7037, 1972. cationsfor the roles of asthenosphereand lithospherein the gen- McKee, E.H., D.C. Noble, C.E. Hedge, and H.F. Bonham, Strontium esis of continental basalts, J. Volcanol. Geotherm. Res., 37, isotopic compositionof some early Miocene rhyolitic tuffs and 291-310, 1989. lavas from northwesternpart of the Great Basin, U.S. Geol. Surv. Leat, P.T., R.N. Thompson,M.A. Morrison, G.L. Hendry, and A.P. Prof. Pap., 800D, D99-D 102, 1972. Dickin, Geochemistryof mafic lavasin the early Rio Granderift, Menzies, M.A., W.P. Leeman, and C.J. Hawkesworth,Isotope geo- Yarmony Mountain, Colorado, U.S.A., Chem. Geol., 8!, 23-43, chemistryof Cenozoicvolcanic rocks reveals mantle heterogene- 1990. ity below westernUSA, Nature, 303,205-209, 1983. 13,506 JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHERE MODIFICATION

Metz, J.M., and G.A. Mahood, Precursorsto the Bishop Tuff: Glass Mexico, 2, Sr and Nd isotopicsystematics, J. Geophys.Res., 94, Mountain, Long Valley, California, J. Geophys.Res., 90, 11,121- 7941-7951, 1989. 11,126, 1985. Prodehl,C., andP.W. Lipman, Crustalstructure of the Rocky Moun- Mooney, W.D., and L.W. Braile, The seismicstructure of the conti- tain region, Mem. Geol. Soc.Am., 172,249-284, 1989. nentalcrust and upper mantle of North America, in The Geologyof Ratt6, J.C., R.F. Marvin, C.W. Naeser, and M. Bikerman, Calderas North America- An Overview, edited by A.W. Bally, and A.R. and ash flow tuffs of the Mogollon Mountains, southwesternNew Palmer,pp. 39-51, GeologicalSociety of America,Boulder, Colo., Mexico, J. Geophys.Res., 89, 8713-8732, 1984. 1989. Reed, J.C., Jr., Proterozoicrocks of the Taos Range,Sangre de Cristo Mooney, W.D., and T.M. Brocher, Coincident seismicreflection/ Mountains, New Mexico, Field Conf. Guideb. N.M. Geol. Soc., refraction studiesof the continentallithosphere: A global review, 35, 179-185, 1984. Rev. Geophys.,25, 723-742, 1987. Riciputi, L.R., and C.M. Johnson,Nd- and Pb-isotopevariations in Musselwhite,D.S., D.J. DePaolo, and M. McCurry, The evolutionof the multicyclic central caldera cluster of the San Juan volcanic a silicicmagma system: Isotopic and chemical evidence from the field, Coloradoand implicationsfor crustalhybridization, Geol- Woods Mountains volcanic cen,ter,eastern California, Contrib. ogy, 18, 975-978, 1990. Mineral. Petrol., 101, 19-29, 1989. Robie, R.A., B.S. Hemingway, and J.R. Fisher, Thermodynamic Nelson, B.K., and D.J. DePaolo, 1,700-Myr greenstonevolcanic propertiesof mineralsand related substancesat 298.15ø K and 1 successionsin southwesternNorth America and isotopicevolution bar (105pascals) pressure and at highertemperatures, U.S. Geol. of Proterozoic mantle, Nature, 312, 143-146, 1984. Surv. Bull., 1452, 456 pp., 1978. Nelson, B.K., and D.J. DePaolo, Rapid productionof continental Robertson,J.M., and K.C. Condie, Geology and geochemistryof crust 1.7-1.9 b.y. ago: Nd and Sr isotopicevidence from the Early Proterozoic volcanic and subvolcanicrocks of the Pecos basement of the North American mid continent, Geol. Soc. Am. greenstonebelt, Sangrede Cristo Mountains,New Mexico, Spec. Bull., 96, 746-754, 1985. Pap. Geol. Soc. Am.,235, 119-146, 1989. Nielson, J.E., D.R. Lux, G.B. Dalrymple, and A.F. Glazner, Age of Rudnick, R.L., and S.L. Goldstein,The Pb isotopiccompositions of the PeachSprings Tuff, southeasternCalifornia and western Ari- lower crustal xenoliths and the evolution of lower crustal Pb, zona,J. Geophys.Res., 95, 571-580, 1990. Earth Planet. Sci. Lett., 98, 192-207, 1990. Noble, D.C., and R.L. Christiansen,Geologic map of the southwest Rudnick, R.L., and S. R. Taylor, The compositionand petrogenesis quarterof the Black Mountainquadrangle, Nye County,Nevada, of the lower crust: A xenolith study,J. Geophys.Res., 92, 13,981- U.S. Geol. Surv. Misc. Geol. Invest. Map, 1-562, 1968. 14,005, 1987. Noble, D.C., and C.E. Hedge, Sr87/Sr86variations within individual Rudnick, R.L., and I.S. Williams, Dating the lower crust by ion ash flow sheets,U.S. Geol. Surv. Prof. Pap., 650C, C133-C139, microprobe,Earth Planet. Sci. Lett., 85, 145-161, 1987. 1969. Rudnick, R.L., W.F. McDonough, M.T. McCulloch, and S.R. Taylor, Noble, D.C., and D.F. Parker, Peralkaline silicic volcanic rocks of Lower crustal xenoliths from Queensland,Australia: Evidence for the western United States, Bull. Volcanol., 38, 803-827, 1974. deep crustal assimilationand fractionationof continentalbasalts, Noble, D.C., C.E. Hedge, E.H. McKee, and M.K. Korringa, Recon- Geochim. Cosmochim. Acta., 50, 1099-1115, 1986. naissancestudy of the strontiumisotopic composition of Cenozoic Ruiz, J., P.J. Patcherr,and R.J. Arculus, Nd-Sr isotopecomposition volcanic rocks in the northwestern Great Basin, Geol. Soc. Am. of lower crustal xenoliths-evidencefor the origin of mid-Tertiary Bull., 84, 1393-1406, 1973. felsic volcanics in Mexico, Contrib. Mineral. Petrol., 99, 36-43, Novak, S.W., Eruptivehistory of the rhyolitic Kane SpringsWash 1988a. volcanic center,Nevada, J. Geophys.Res., 89, 8603-8615, 1984. Ruiz, J., P.J. Patchett, and F. Ortega-Gutierrez, Proterozoic and Novak, S.W., Geology and geochemicalevolution of the Kane Phanerozoicbasement terranes of Mexico from Nd isotopicstud- SpringsWash volcanic center, Lincoln Co, Nevada,Ph.D. thesis, ies, Geol. Soc. Am. Bull., 100, 274-281, 1988b. 173 pp., StanfordUniv., Stanford,Calif., 1985. Ruiz, J., P.J. Patcherr,and R.J. Arculus,Reply to "Commentson Nd- Novak, S.W., and G.A. Mahood, Rise and fall of a basalt-trachyte- Sr isotopiccompositions of lower crustal xenoliths-evidencefor rhyolite magma systemat the Kane SpringsWash caldera,Ne- the origin of mid-Tertiary felsic volcanics in Mexico" by K.L. vada, Contrib. Mineral. Petrol., 94, 352-373, 1986. Cameron and J.V. Robinson, Contrib. Mineral. Petrol., 104, 615- Okaya, D.A., and G.A. Thompson,Involvement of deep crust in 618, 1990. extensionof the Basinand Range province, Spec. Pap. Geol. Soc. Rybach, L., and G. Buntebarth, Relationships between the Am., 208, 15-22, 1986. petrophysicalproperties density, seismicvelocity, heat genera- O'Nions, R.K., and E.R. Oxburgh, Helium, volatile fluxes and the tion, and mineralogicalconstitution, Earth Planet. Sci. Lett., 57, developmentof continentalcrust, Earth Planet.Sci. Lett., 90, 331- 367-376, 1982. 347, 1988. Rytuba,J.J., and E.H. McKee, Peralkalineash flow tuffs andcalderas Orkild, P.P., K.A. Sargent,and R.P. Snyder,Geologic map of Pahute of the McDermitt volcanic field, southeastOregon and north cen- Mesa, Nevada Test Site and vicinity, Nye County,Nevada, U.S. tral Nevada, J. Geophys.Res., 89, 8616-8628, 1984. Geol. Surv. Misc. Geol. Invest. Map, 1-567, 1969. Sampson,D.E., and K.L. Cameron,The geochemistryof the Inyo Ormerod,D.S., C.J. Hawkesworth,N.W. Rogers,W.P. Leeman,and volcanic chain: Multiple magma systemsin the Long Valley M.A. Menzies, Tectonicand magmatictransitions in the western region, easternCalifornia, J. Geophys.Res., 92, 10,403-10,421, Great Basin, USA, Nature, 333, 349-353, 1988. 1987. Pallister,J.S., andE.A. du Bray, Field guideto volcanicand plutonic Sargent,K.A., and P.P. Orkild, Geologic map of the Wheelbarrow features of the Turkey Creek caldera, Chiricahua Mountains, southeastArizona, Mem. N.M. Bur. Mines Miner. Resour., 46, Peak-Rainier Mesa area, Nye County, Nevada, U.S. Geol. Surv. Misc. Geol. Invest. Map, 1-754, 1973. 138-152, 1989. Pearce,J.A., Trace element characteristicsof lavas from destructive Sawyer, D.A., and K.A. Sargent,Petrologic evolution of divergent plate boundaries,in Andesites,edited by R.S. Thorpe,pp. 525- peralkaline magmas from the Silent Canyon caldera complex, 548, John Wiley, New York, 1982. southwesternNevada volcanic field, J. Geophys.Res., 94, 6021- 6040, 1989. Pearce,J.A., Role of sub-continentallithosphere in magmagenesis at active continentalmargins, in ContinentalBasalts and Mantle Scott, G.R., Cenozoic surfacesand depositsin the southernRocky Xenoliths,edited by C.J. Hawkesworthand M.J. Norry, pp. 230- Mountains, Mem. Geol. Soc. Am., 144, 227-248, 1975. 249, Shiva, Nantwich, Mass., 1983. Scott, R.B., R.W. Nesbitt, E.J. Dasch, and R.L. Armstrong,A stron- Perry, F.V., W.S. Baldridge,and D.J. DePaolo,Role of astheno- tium isotopeevolution model for Cenozoicmagma genesis,east- sphereand lithospherein the genesisof late Cenozoicbasaltic ern Great Basin, U.S.A., Bull. Volcanol., 35, 1-26, 1971. rocks from the Rio Granderift and adjacentregions of the south- Shannon, J.R., Geology of the Mount Aetna cauldron complex, westernUnited States,J. Geophys.Res., 92, 9193-9213, 1987. SawarchRange, Colorado, Ph.D. thesis,434 pp., Colo. Sch. of Phelps,D.W., D.A. Gust,and J.L. Wooden,Petrogenesis of themafic Mines, Golden, 1988. fieldspathoidallavas of the Raton-Claytonvolcanic field, New Shannon,J.R., C.W. Naeser,E. DeWitt,and A.R. Wallace,Timing of Mexico, Contrib. Mineral. Petrol., 84, 182-190, 1983. Cenozoicmagmatism and tectonismin the Sawatchuplift and the Pier, J.G., F.A. Podosek, J.F. Luhr, and J.C. Brannon, Spinel- northern Rio Grande rift, Colorado (abstract), Geol. Soc. Am. lherzolite-bearingQuaternary volcanic centers in SanLuis Potosi, Abstr.Progams, 19, 839, 1987. JOHNSON:LARGE-SCALE CRUST FORMATION AND LITHOSPHERE MODIFICATION 13,507

Shaw, H.R., Links between magma-tectonic rate balances, Thompson,G.A., and D.B. Burke, Regionalgeophysics of the Basin plutonism, and volcanism, J. Geophys.Res., 90, 11,275-11,288, and Range Province, Annu. Rev. Earth Planet. $ci., 2, 213-238, 1985. 1974. Smith, R.L., Ash flows, Geol. $oc. Am. Bull., 71,795-842, 1960. Thompson, G.A., R. Catchings, E. Goodwin, S. Holbrook, C. Smith, R.L., Ash-flow magmatism,Spec. Pap. Geol. $oc. Am., 180, Jarchow,C. Mann, J. McCarthy, and D. Okaya, Geophysicsof the 5-28, 1979. western Basin and Range province, Mem. Geol. Soc. Am., 172, Smith, R.L., R.A. Bailey, and C.S. Ross,Geologic map of the Jemez 177-203, 1989. Mountains, New Mexico, U.S. Geol. Surv. Misc. Geol. Invest. Thompson,R.A., M.A. Dungan, and P.W. Lipman, Multiple differ- Map, 1-571, 1970. entiation processesin early-rift calc-alkaline volcanics, northern Smithson,S.B., Contrastingtypes of lower crust, in Properties and Rio Grande rift, New Mexico, J. Geophys.Res., 91, 6046-6058, Processesof Earth's Lower Crust, Geophys. Monogr. Ser., vol. 1986. 51, editedby R.F. Mereu, S. Mueller, and D.M. Fountain,pp. 53- Thompson,R.A., C.M. Johnson,and H.H. Mehnert, Oligocenebasal- 63, AGU, Washington,D.C., 1989. tic volcanism of the northern Rio Grande rift: San Luis Hills, Spera,F.J., and J.A. Crisp, Eruption volume, periodicity, and caldera Colorado, J. Geophys.Res., this issue. area: Relationshipsand inferenceson developmentof composi- Verma, S.P., Magma genesisand chamberprocesses at Los Humeros tional zonationin silicic magmachambers, J. Volcanol. Geotherrn. caldera, Mexico -- Nd and Sr isotopedata, Nature, 303, 52-55, Res., 11, 169-187, 1981. 1983. Stein, H.J., and J.G. Crock, Late Cretaceous-Tertiarymagmatism in Verma, S.P., Sr and Nd isotopic evidence for petrogenesisof mid- the Colorado Mineral Belt: Rare earth element and samarium- Tertiary felsic volcanism in the mineral district of Zacatecas,Zac. neodymiumisotopic studies,Mem. Geol. $oc. Am., 174, 195-223, (Sierra Madre Occidental), Mexico, Isot. Geosci., 2, 37-53, 1984. 1990. Vogel, T.A., D.C. Noble, and L.W. Younker, Evolution of a chemi- Steven,T.A., Middle Tertiary volcanic field in the southernRocky cally zoned magma body: Black Mountain volcanic center, south- Mountains, Mern. Geol. Soc. Am., 144, 75-94, 1975. western Nevada, J. Geophys.Res., 94, 6041-6058, 1989. Steven, T.A., and P.W. Lipman, Calderasof the San Juan volcanic Wass, S.Y., and J.D. Hollis, Crustal growth in south-easternAustra- field, southwesternColorado, U.S. Geol. $urv. Prof. Pap., 958, 35 lia- Evidence from lower crustal eclogitic and granulitic xeno- pp., 1976. liths, J. Metamorph. Geol., 1, 25-45, 1983. Steven, T.A., P.D. Rowley, and C.G. Cunningham, Calderas of the Weaver, B.L., and J. Tarney, Continental crust composition and Marysvale volcanic field, west centralUtah, J. Geophys.Res., 89, nature of the lower crust: Constraintsfrom mantle Nd-Sr isotope 8751-8764, 1984. evolution, Nature, 286, 342-346, 1980. Stuckless,J.S., and J.R. O'Neil, Petrogenesisof the Superstition- Weiss, S.I., and D.C. Noble, Stonewall Mountain volcanic center, Superiorvolcanic area as inferred from strontium- and oxygen- southernNevada: Stratigraphic,structural, and facies relationsof isotopestudies, Geol. $oc. Am. Bull., 84, 1987-1998, 1973. outflow sheets,near-vent tuffs, and intracalderaunits, J. Geophys. Stuckless,J.S., and M.F. Sheridan,Tertiary volcanic stratigraphyin Res., 94, 6059-6074, 1989. the Goldfield and SuperstitionMountains, Arizona, Geol. Soc. Wilshire, H.G., C.E. Meyer, J.K. Nakata, L.C. Calk, J.W. Shervais, Am. Bull, 82, 3235-3240, 1971. J.E. Nielson, and E.C. Schwarzman, Mafic and ultramafic xeno- Swanson, E.R., and F.W. McDowell, Calderas of the Sierra Madre liths from volcanic rocks of the western United States, U.S. Geol. Occidental volcanic field western Mexico, J. Geophys.Res., 89, $urv. Prof. Pap., 1443, 179 pp., 1988. 8787-8799, 1984. Wittke, J.H., D. Smith, J.L. Wooden, Origin of Sr, Nd and Pb Taylor, R.B., Neogene tectonism in south-central Colorado, Mem. isotopic systematics in high-Sr basalts from central Arizona, Geol. $oc. Am., 144, 211-226, 1975. Contrib. Mineral. Petrol., 101, 57-68, 1989. Taylor, W.J., J.M. Bartley, D.R. Lux, and G.J. Axen, Timing of Tertiary extension in the Railroad Valley-Pioche Transect, Ne- C.M. Johnson,Department of Geologyand Geophysics,Univer- vada: Constraintsfrom 4øAr/39Arages of volcanic rocks, J. sity of Wisconsin-Madison,Madison, WI 53706. Geophys.Res., 94, 7757-7774, 1989. Tegtmeyer, K.J., and G.L. Farmer, Nd isotopic gradients in upper (Received April 2, 1990; crustalmagma chambers: Evidence for in situ magma-wall-rock revised December 26, 1990; interaction,Geology, 18, 5-9, 1990. acceptedJanuary 22, 1991.)