Canadion MineralogistVol. 26, pp. 5l?-539 (1988)

SUBMERSIBLE INVESTIGAiNON OF AN EXTINCT HYDROTHERMAL SYSTEMON THE GALAPAGOS RIDGE: SULFTDE MOUNDS, STOCKWORK ZONE,

AND DIFFERENTIATED LAVAS

ROBERT W. EMBLEYNational Oceanographic and Atrnospheric Administration, Hatftetd Maine Science Center, Newport,

Oregon 97365, U.S.A.

IAN R. JONASSONGeological Samey of Canada, 601 Booth St., Ottawa, Ontario KIA OES, Canada

MICHAEL R. PERFITDepartment of Geologlt, University of Florida, Gainesville, Florida 3261t, U.S.A.

JAMES M. FRANKLINGeological Sumey o! Canada, 60j Booth St., Ottawa, Onturto KIA hEg, Canada

MAURICE A. TIVEYSchool oJ Oceanography, University of Washington, Seattle, Washington 98195, U.S.A.

ALEXANDER MALAHOFFDepartment of Oceanography, University of Howaii, Honolulu, Hawaii 96822, U.S.A.

MICHAEL F. SMITHDepartment of Ceology, University of Florida, Gainesville, Florida 32611, U.S.A. ,

TIMOTHY J.G. FRANCISInstitute of Oceanographic Sciences, Wormley, Godalming, Suney, England GVA SUB

ABSTRA T

Fifteen dives along the Galapagos Ridge in the regionbetween 85"49'W and 85o55'W were made to examine thedetailed relationships'among tectonics, hydrothermalactivity and lava compositions. Extensive tectonic activiryand physical weathering have exposed the inner parts oflarge Cu-Zn sulfide mounds and the uppermost part oftheunderlying stockwork zone. The mineralization occurs artlte top and southern base of a horst block, 40 to 80 m high,that separates the present Neovolcanic Zone to the northfrom an older rift valley to the south. The lavas in theNeovolcanic Zone are homogeneous MORB pillows; thoseon the horst block and within the southern valley areevolved MORB to andesite pillow and sheet flows. Thealteration zone exposed beneath the sulfide nounds com-prises a network of fracture-controlled pipe and sheetJikebodies of highly altered material which changes outwardinto relatively fresh but similarly closely fractured rocks.The hydrothermal upflow zone is extensively brecciated ona centimeter scale and encloses a stockwork ofveinlets nowfiIled largely by silica, clays and sulfides. The most highlyaltered rocks are strongly depleted in Ca, Na, K and Mn,and are enriched in S, Fe, Cu and Zn relative to their freshanalogs. Si and Mg are variable, the latter showing localdepletions and enrichrnents according to the proportion and{istribution of chlorite. Depletions in 18O with increasing873r/865r suggest extensive seawater-rock interaction(W/R up to 100:l) at Iup to 350"C. Deep-tow and ALVIN-based magnetic profiles have a relative mapetization lowcentered over the southern valley and the horst block that

could reflect more extensive hydrothermal alteration zonesassociated with the older seafloor. The Galapagos stock-work is most analogous to the alteration zones associatedwith massive sulfide deposits in the ophiolites of Cyprusand Oman.

Keywords: Galapagos Ridge, spreading center, alterationpipe, sulfides, stockwork, copper, chlorite, smectite,andesite, basalt, magnetic profile, isotopes.

SouuatnnNous avons effectu6 quinze plong6es le long de Ia cr€te

des Galapagos, entre 85"49'W et 85o55 'W, afin de d6ter-miner les relations entre l'activitd tectonique et hydrother-male et la composition des laves. Une activit6 tectoniqueet une d6gradation physique importantes ont contribu6 dexposer les parties internes de vastes amonoblements de sul-fur* de Cu et de Zn, ainsi que la partie supdrieure de lazone sous-jacente i <<stockwerk>r. La min6ralisation setrouve au sommet et e la base de l'extr6rnitd Sud d'un horstde 40 i 80 m de hauteur qui s6pare la zone n6ovolcaniqueprdsentement active, au nord, d'une vallde de rift plusancienne au sud. La zone ndovolcanique contient des lavesen coussins, homogbnes, de type MORB; les laves en cous-sins et les couldes en feuilles affleurant sur le horst et dansla vallde du Sud correspondent d une composition variantentre celle d'un MORB 6volu6 i celle d'une andesite. Lazone d'alt6ration en dessous des monticules de sulfures con-tient un r6seau de roches fortement alterees, dispos€es le

517

5 1 8 THE CANADIAN MINERALOCIST

long des fissures verticales et horizontales, en s'6loiguantdesquelles i leur p6riph6rie, ces domaines d'alt€ration fontplace i des roches relativement fraiches mais 6galement tra-vers€es par de nombreuses fissures. La zone de flux hydro-thermal est fortement br6chifi6e d l'6chelle centimetrique,et renferme un stocl(work de fissures qu'occupent mainte-nant de la silice, des argiles et des sulfures. Les roches lesplus fortement alt6rees sont nettement appauvries en Ca,Na, K et Mn, et enrichies en S, Fe, Cu et Zn par rapportaux roches non alt6rdes. Les teneurs en Si et Mg sont varia-bles; les variations en teneurs du Mg reflOtent la propor-tion et la distribution de la chlorite. L'appauvrissement enl8O avec l'augmentation du rapport 8?5r/865r t6moigned'une interaction importante entre roches et eau de mer ides tempdratures atteignant 350oC, avec un rapporteau/roche pouvant atteindre 100:1. Les profils magn6ti-ques obtenus par remorquage de magndtom€tres i grandeprofondeur et i bord du sous-marin ALVIN r6vblent unerdgion i faible champ magn€tique centrde sur la vall€e duSud et sur le horst: elle pourait indiquer I'existence de zonesd'altdration hydrothermale plus 6tendues, impliquant lesocle. Le stockwerk de la cr6te des Galapagos ressemblefortement aru( zones d'alt6ration associ6es aux gisementsde sulfures massifs de Chypre et d'Oman.

Mots-clds: cr€te des Galapagos, centre de s6paration desplaques, altdration d'un conduit, sulfures, stockwerk,cuiwe, chlorite, smectite, anddsite, basalte, profilmagndtique, isotopes.

INTRoDUCTIoN

Since the discovery of active submarine hydrother-mal vent systems in the late 1970s (Corliss et ol. 1979,Rise Project Group 1980), comparisons have beenmade with orebodies exposed in ophiolite sequences(e.g., Oman, Cyprus and Newfoundland) or Precam-brian Shield areas (Franklin et ol. 1981). An impor-tant feature of ancient deposits is the relative acces-sibility to the deeper sections of the hydrothermalsystem (either through outcrops or drilling).However, the "stockwork" or feeder zones thatcomprise the hydrothermal upflow regions beneathsuch sulfide bodies have been overprinted to varia-ble degrees by later metamorphism and deformationthat obliterated original mineral assemblages and tex-tures. This complicates interpretations of chemicaldata, which must necessarily be based on mineralog-ical and textural observations.

In modern oceans the upwelling portion of thehydrothermal system is rarely exposed. Detailedstudies of altered lava samples come from dredgedsamples taken in the Atlantic (Mottl 1983, Bonattiet al. 1976, Humphris & Thompson 1978, Delaneyet ol. 1987); these samples have provided a data baseused widely in discussions of hydrothermal plumb-ing systems. However, the inherent inaccuracy indredging, and the tectonic complexity of the Mid-Atlantic Ridge, make it difficult to place their sam-ples in context within a "typical" hydrothermalsystem.

Delaney et al. (1987) described quartz-cementedsulfide-bearing breccias recovered in dredges and bysubmersible from near the Kane Fracture Zone.Renardet ol. (1985) described veined stockwork-likesamples from a scarp on the East Pacific Rise @PR)at 18o30'S, but their bulk chemical data indicateminimal alteration relative to the Galapagos samples.Similarly, Biicker et al. (1985) reported "mineralizedbasalt breccia and basalt with disseminated pyrite"from dredges on the EPR at 2l.5oS' but no analyseswere given.

Deep Sea Drilling Project Hole 5048 on the flankof the Costa fuca Rift provides a "reference section"through oceanic crust that has undergone axial andoff-axial phases of hydrothermal alteration (Alt elal. L98A, A zone of stockwork-like alterauon occursat 635-653 m. Although the stockwork zonepresumably formed when the site was above the axialmagma chamber (5.9 Ma), the exact timing anddepth below the seafloor, and its relationship to anymassive sulfide deposit is unclear.

In 1980 and 1981, several ALVIN dives along theeastern Galapagos (Cocos-Nazca) Spreading Center(Fig. l) discovered and sampled an extensive mas-sive sulfide deposit that formed from an extincthydrothermal system (Malahoff et ol. 1983). Bothsyn-depositional and post-depositional faulting hadoccurred in the mineralized zone: initial observationssuggested the likelihood of underlying stockworkzones exposed along the normal fault systems. In1985, an ALVIN dive series was condircted in thisarea to search for such stockwork zones. This reportdescribes the geological setting and the physical andchemical nature of a fossil hydrothermal $ystem thatwas discovered directly beneath the Galapagos sul-fide deposits. This is the first discovery ofa seafloorsulfide deposit that has a well-exposed contiguousttstringer zone".

GsolocrcAl SsrrtNc

The study areais centered at 0o45'N, 85o50.5'Wat a water depth of 254A m along the south side ofthe main Galapagos rift valey (Fies. l, 2). The activeGalapagos vents (Corliss et al. 1979) are 20 to 30 kmwest, between 86"01 'W and 86'14'W. Wateremanating from the latter vents at up to l7'C sup-ports diverse suites of live animal communities(Corliss et al. 1979). No massive sulfides have beensampled in this area, although they have been notedin camera surveys (Ballard et al. 1982).

The recent structural evolution of this part of theGalapagos Rift probably has been complex. Forinstance, the presence of curved ridges and sharpalong-strike discontinuities and topography (e.9.,southern valley), particularly south of the axis(Fig.2), may mark the position of old DEYALS(Langmuir et al. 1986) or overlapping spreadingcenters that have been partly bypassed by late mag-

EXTINCT HYDROTHERMAL SYSTEM. GAI.APAGOS RIDGE 5r9

COCOS PLATE

e8qt'*

NAzcA PLA'E

"34'"r l"5,|n'J[=ffi: ( /:N:rd8 63 EG

soo- 8r w

csoo\ 2,

\ \_/

5Ecuador Rm,

Al oIto

\ v

\

j--5 Galapagos b. F.Z\{---': ^

Nlt

*8";J

860 W &00 w 820 W

GSC

Ftc.-l ' T.ocation of study area. Bathymetric contours in &ousands of meters. Solid lines mark axes of spreading centers;dotted lines mark fracture zonc. Location ofFigure 2 shown by rectangular box near intersection ofEastem GalapagosRift and Inca Fracture Zone.

matic events and rafted off-axis (Macdonald et al.1984, 1986). The morphology of the GalapagosSpreading Center in the vicinity of the study area isunusual in that there appears to be a "double rift"strudure.

The northern 1.5 km-wide rift valley is separatedfrom the I km-wide southern depression by a nar-row (30-60 m relief) ridge (Figs. 2 to 5). The north-facing scarp of the central ridge is on line with aprojection of the main southern boundary fault ofthe Galapagos Neovolcanic Zone between i5.5g, Wand 86:12'W, where many warm vents are located(Ballard et al. 1982). A band of young, glassy pil-low flows a few hundred meters north of the ridee

marks the trace of the Neovolcanic Zone through this'part of the Galapagios Spreadine Center. Cores taken

in the southern valley contain up to 40 im ofpelagicsediment, indicating up to 15 thousands of yearsof sedimentation on the lava fields (Lalou ef a/.1983). At least l0 cm of sediment also cover pillowlavas and oxidized sulfide mounds on top of theridge. The sparsely sedimented lavas within the cen-tral part of the northern rift valley (NeovolcanicZone) are apparently younger.

Most of the ALVIN dives were made in a smallarea Q x I km) centered at 0o45'N, 85"50.5'W,where major sulfide mounds were sampled duringreconnaissance dives in 1981. The bathymetry shown

s20 TIIE CANADIAN MINERALOGIST

in Figure 4 was derived from pressure-depth reao-ings from ALVIN. The horst on the map shoals toa narrow, flat top that is only about 100 m wide inthe vicinity of the sulfide mounds. The discrepancybetween the minimum depth on the Sea Beam bathy-metric map (Fig. 3) at about 2590 m, versus t}l.eALVIN-based map (about 2535 m) is largely a resultof the resolution of the Sea Beam system; i.e., tbewidth of the top of the horst is less than the "foot-print" (120 m) of the Sea Beam system at this depth.In addition to a relatively dense grid of mapping andsampling conducted within the area on Figure 4,there were also lava sampling excursions to the northand south within the Neovolcanic zones and in anarea of andesite flows, respectively. In particular,dive 1652 traversed the entire rift valley from northto south (Frg. 3). Another area of sulfides, and anarea of silica-manganese dioxide mounds emanat-ing warm water (- l0'C) are located about 5 km tothe west (85"54'W, Fig. 2); these also were samPled(dives 1660, 1661, and 1663). All mineralizationseems to have formed along an extension of the samegeologic boundary that occurs within the primarystudy area. We interpret the central ridge as a fault-bounded horst on the basis of pervasive talus slopes,vertical walls of truncated pillows, and the presenceof truncated sulfide mounds on the top and also atthe base of tJte southern scarp.

LAVA PETRoLocY, GnocHEMIstnYAND SPATIAL MORPHOLOGY

The 1980/81 dive series recovered a diverse suiteof evolved N-type mid-ocean ridge (MOR) tholeiiticlavas, including FeTi basalt and andesite from anarea along the ridge within 60 km of the GalapagosRift - Inca Transform intersection (Fornari et al.1983, Perfit et al. 1983). Some of the most chemically evolved lavas were sampled near the sulfidedeposits (Malahoff et al. 1983). Within the Neovol-canic Zone, from the study area to about 86o13'W,the rocks are significantly less differentiated; FeTibasalt is rare and andesite absent (Fornari el 4L 1983,Ballard et al. 197 9), On the basis of petrological andgeochemical data, Perfit et ol, (L983') proposed thatextensive crystal-liquid fractionation occurred atshallow levels in isolated magma chambers (approx-imately 2 km depth) in response to increased ther-mal budgets during attempted ridge propagationclose to the Galapagos Rift-Inca Transform inter-

section (Fig. l). They alsoproposed that massive sul-fide generation in this region might have beenenhanced by increased concentration of sulfides inhighly evolved lavas, and their subsequent remobili-zation during hydrothermal qirculation.

Extensive sampling near the sulfide mounds dur-ing the 1985 series of dives, together with subsequentgebchemical investigations, have allowed us to placefurther constraints on the spatial association betweenthese sulfide deposits and highly fractionated MORmagmas. The predominant rock type recovered in9 dives (lU7-1655, Table 1) that traversed thesouthern rift valley and ceutral horst area was FeTibasalt (FeO, >L2wt.Dlo, TiO2 >2wt.9o), andsome of these are highly evolved @eO1 > 16 wt.Vo'TiO2 >3 wt.Vo). In addition, basaltic andesite andand&ite (SiO2 5+CI wt.Vo) are abundant within thesouthern valley and are very common along the cen-tral horst (Fig. a). Chemical analyses of representa-tive rock types are given in Table 1; the FeTi basaltand andesite recovered from this limited area havechemical compositions that closely match thosedetermined in previous studies (Fornari et al. 1983,Perftt et al. 1983). Typical mafic MORB was rarelyrecovered in the southern rift valley and around thehorst near sulfide mounds (Figs. 4' 5), However,MORB was recovered a few hundred meters west ofthe alteration zone (Fig. 4; samples l65Gl, l@9'5).The dichotomy of rock types between the northernrift valley (Ballard et al. 1979), and horst/southernrift valleys is evident if indices of fractionation suchas Mg-number are comPared (Fig. O.

Voluminous lobate flows of andesitic basalt andandesite occur near sulfide mounds in the southernrift valley and along the top and scarps ofthe horst.Some of these flows have a characteristic blockyappearance and lack the surface decoration (i.e, sur-face comrgations or protuberances) common tobasaltic pillow lavas. The andesite is slightly vesicu-lar, extremely glassy, has conchoidal fracture, andranges from dark brown to green to bfuish grey.

FeTi basall was the most common lava type sam-pled and was recovered with andesite at a few local-ities. It occurs as small pillows with thin glass crustsor as slightly vesicular sheet flows (< 4 cm thick).On the horst top (16514) FeTi basalt forms small,irregular-shaped spires (about I m high) similar to"driblet spires" elsewhere found associated withsubaerial "aa" flows.

Frc. 2. Multibeam bathymetric chart of axis of Galapagos Ridge in vicinity of study area. Combination of SASSbathymetry collectedin l976by U.S.N.S. llutton hd Sea Beam collected in 1985by Atlanlirl/. Contour intervalis 2Om and color interval is u10 m. A continuation of the best fit line through the warm-water vents (dashed line)east of 86o00,W corresponds approximately to fault controlling locations of older sulfide mineralization and thesouthern part of the Neovolcanii Zone. Stars (*) indicate warm-water vents (Ballard et al. 1982); triangles locatemassive sulfide mounds (this study and H. B?icker, pers. comm.). D4, D5 refer to dredge hauls; AL1650, 1661, 1663refer to ALVIN dives discussed in text. The box shows location of Figure 3; N-S line traces deep-tow magnetometerline (Fig. ll).

EXTINCT HYDROTI{ERMAL SYSTEM, CALAPAGOS RIDCE 521

522 TIIE CANADIAN MINERALOGIST

Frc. 3. Sea Beam bathymetry of Galapagos Rift in vicinity of study area. The triangles show locations of largest mas-sive sulfide mounds; the arrow indicates the stockwork-zone outcrop. Neovolcanic Zone shown by stippling. Dashedline gives location of cross-section shown in Figure 5. Box is area of Figure 4. Sample designation same as Figure 5.

Lavds. from the Neovolcanic Zone within thenonhern valley @igs. 3, 5), in contrast with thoseto the south, are mainly moderately fractionatedMORB or ferrobasalt (Table l). The lavas typicallyform bulbous piUows with numerous small buds pro-truding fiom comrgated surfaces that are glassy andonly lightly dusted with sediment. FeTi basalt andandesite have not been identified in the NeovolcanicZone in this section of the Galapagos Rift (Fig. 5).However, some moderately evolved FeTi basalt wasrecovered from the medial rift zone during the 1980dives. Notably, lavas recovered around the activevents at 86o06 ' to I I 'W (thermal springs of Corlisset al. 1979) are significantly less evolved and are moremafic than those associated with the inactive ventsites in the study area. Pillow and sheet-flow MORBsamples (Ballard et ql. 1979) contain 1.33 to2.67 vt.tlo TiO2 and have Mg-numbers of about 53(Fig. O. Earlier work indicated that lavas eruptedalong the Galapagos Rift within 60 km of the IncaTransform fault are more fractionated and chemi-cally less homogeneous than those farther from thefault @erfit et ol. 1983). The results of the 1985 diveprogranx support the earlier observation that themost recent period of volcanism along the GalapagosSpreading Center is associated both with less evolvedlavas and, app.rently, with little sulfide formation.Hydrothermal activity associated with the Neovol-caric Zone has generated a series of clay-silica

deposits around warm springs (Fig. 2; dives 1660,1663).

SuLrDs DEPoSITS

In the study area the sulfide deposits occur as sin-gle chimneys and composite mounds scattered overan area of 0.5 km by 0.5 km (Figs. 3, 4), both in thesouthern valley and on top of the horst block.Mounds and chimneys are most extensive at thesouthern base of the horst block at a water depthof 2570-2580 m, where the horst block widens to theeast into a series of fault-sliced blocks @igs. 3, 4).The sulfides in the southern valley (1649 mounds)are in two morphologic settings: (1) individual chim-neys and small mounds along the base of the mainsouth-facing scarp of the horst block, and (2) largecomposite mounds of coalesced and collapsed chim-neys, exceeding 1@ m in diameter and about l0 mhigh, that occur along subsidiary parallel faults southof the primary fault (Fig. 7a). Most,of the moundsare clearly associated with faults, but some largeaccumulations of sulfides may be covering underb-ing faults. Although most of the faults in this areaare east-west trending, at least one sulfide moundwith a north-south trend seems to be fault-controlled, suggesting the possibility of mineralizedcross-structures. In several places, small chimneysproject out of the talus slope and drape against the '

main south scarp. As it is unlikely that such chim-

EXTINCT HYDROTHERMAL SYSTEM. CALAPAGOS RIDGE 5238 6 0 6 0 . 7 5 ' 850 60.50' ,

Ftc. 4. Detailed bathymetry of dive area showing dive tracks, location of samples (e,g.,544 desipates Dive 1654, sta-tion 4), and lava types. Samples 165+6,7 ,8 all are from the same site; 1655-l not located for space reasons. Sam-ples on wall projected to 2537 m (horst top). Box on Figure 3 shows area covered on map. Depths (m) based onALVIN pressure gauge. N-S line indicates position of cross-section in Figure 5.

neys could survive rock falls, this observationstrongly implies some post-tectonic mineralization.

A nest of sulfide mounds and chimneys on top ofthe horst at a depth of 2535 m was also sampled(Fig. 4; 1650 mounds). These deposits.are heavilysedimented, are scattered over an area of abour100 m in diameter, and also appear to be associatedwith east-west faulting. Some mounds are truncatedby the north scarp that exposes underlying alteredlavas (Fig. 7b).

The unweathered sulfides consist principally ofpyrite and chalcopyrite with minor sphalerite, pyr,rhotite and marcasite. From the areal extent andexposed thickness of sulfides observed, the totalamount of sulfide is 1.5 m tonnes in the study areaGig. ). The high ratio of Cu to Zn (about 6:l) dis-tinguishes the Galapagos sulfides from all other sub-marine deposits (Rona 1984, Malahoff et al. 1983,Kappel & Franklin 1988). Unweat[ered sulfides ontop of the horst are compositionally.indistinguisha-

ble from most of those on the adjacent downthrownblock (Table 2), suggesting that all the mounddeposits in the. area were formed under similarhydrothermal conditions.

Local oxidation and mass wasting of sulfides iscommon, both around mounds and within theprimary talus slopes. Disintegrating chimneys arestrongly depleted in all sulfideg but contain silica (asquartz or amorphous silica), goethite, hematite,magnetite, and secondary copper minerals whichinclude covellite, digenite, roxbyite, atacamile, botal-lackite and the Fe sulfate, szomolnokite. Rare bariterosettes and axiolites are enclosed in, or partlyreplaced by amorphous silica and pyrite in the mar-gins of disintegrating chimneys.

Some ochrous mounds are composed mainly ofsiliceous iron oxide crusts and ferruginous mud.Thesb represent the remains of wholly oxidizedexteriors of clqiqneys and sulfide interiors, respec-tively. Areas where unweathered sulfides axe exposed

850 50.26' tY

/ \

mounds / ' \1648{ /

mwnds /

52-54

ffi

Metera0 roo 200

ooa6.26 'N

oo 46 .00 '

o.11.76'

A sz-gY

f enorortr"t"I Seabeam

I'Footerm' a6eo

50-2.I

tl Ir '

'/z

ir

t --549

I\ *tr"+-=-.=

---i-\-

\

M o R B . . . . . . . O

Fe t l baoah . . . 8

A n d o o l t s . . . . . . a

Ane redhYa . . .X

S u m d e . . . . . . . . . A

A[oratlon zone. . . i'j'l

Snlltronto.rt'n.../Z

I) , lvetrad(s. . . -J

SAMPLE SITES

tri i'""'1'"' \

, - J . . \ /\v' a\+--f(-l:\,"r-l]

524 TTM CANADIAN MINBRALOGIST

sourH --foRsr, NoRrH -' vALLEY - -VALLEY-

Ngovolcanlc zone

A N. . n . o . t r

gtoakwork

traa

ntt n

probably represent cores of composite stasks nowexposed by mass wasting or exfoliation (e.9.,Fie. 7a). The massive sulfides that survivedprolonged supergene alteration have interiors thatare richer in chalcopyrite and pyrite than in theirouter walls. The latter consist of assemblages ofpyrite-chalcopyrite-sphaletilg:pyrrhotite.

ALTERATToN elvn Srocxwom MnmnauzATloN

An extensive outcrop of altered lavas from theuppermost section of a zone of primary hydrother-mal upflow occurs on the north wall of the horst(Figs. 3, 4). Small outcrops were also o6r.*e6 highon the south-facing wall. The outcrop on the northwall extends 100 m laterally and 35-45 m verticallydown to the talus pile at the foot of the wall. Vol-canic rocks on the north wall consist of layers ofglassy pillow lavas (predominantly basaltic andesiteand andesite) interbedded with lobate and sheet flows(FeTi to ferrobasalt), and ponds of hyaloclastite

.? Norttrem,t boundaryi tault

o,-- ut* '904:1 vortloal

6ragg6retlon

(andesite). The most accessible exposrues lie withinl0 to 15 m of the top of the horst @igs. 7b to 7h);lower portions are partly obscured by talus and sedi-ment. Alteration is pervasive and brecciation is exten-sive; however, pillow lobes, their selvages, and bedsof glass shards generally are well-preserved as pseu-domorphs and were easily mapped (Figs. 7f'g).Alteration is most pervasive in the hyaloclastite lensesbecause of their original high permeability. Masswastage has differentially removed softer materialfrom the horst and has given rise to a spectacularlocal topography ofspurs, hoodoos, and overhangsthat protrude from the face of the main escarpment(Fig. 7c). The amount of mass wastage depends onthe nature of the parent volcanic rocks, the extentof brecciation, and the intensity of chemical altera-tion.

The innermost part of the alteration zone (Fig. 8)is typically light grey to white. Mud-free areas con-sist of smooth-faced slopes of soft, highly alteredmaterial and narrower patches of less altered, shat-

E 2400 -l

6 - l= 2000] ? \

Oouthem Iboundary ;fauf! i7

'S..\

\

ooo

,:E

Fault . . .

M o r b . . . . . . . . . . . O

A n d o 3 | t o . . . . . . . . 4

;:":i: gul l lde mounds. . . .A

GSc

Frc. 5. Top: generalized cross-section of Galapagos Rift through the study area showing location of features discussedin text. Lava samples have been projected onto cross-section (see Fig. 4 locations). Bottom: expanded cross-sectionthrough horst structure.

525EXTINCT HYDROTHERMAL SYSTEM. GALAPAGOS RIDGE

o d r 6 o o r .@ d - d O 6 t s .

o O d o O o { .6 n F 6 n F .

o 6 0 6 0 0 0 ,< 6 9 S N O i i

$ a i

O o F N O o i .6 6 F 3 6 6 i i

o r r o o o < .6 A < i @ O A i

N N .

O . . t E tO € F o ! g - i 6

r4 i i l f ; i * r s " .qE* , -F "e" 'Q ,? t, 9 8 8 ' 5 9 ' " I " j * S " 8 ; d ! f f g B g g j; _ ' E * I H . H E ! i r E r E I d 4 1 S ! $ l s q EE - e s H 3 H * . E . i u t u E E c ' o . ! i o ! f l 9 ' b6 $ " E d ' " f i , 5

" t . " , 9 " , 9 H o € F g d g ! a g n

p i E c b l ' ? E : 2 E l P u C ' i 8 d ; g 8 . E E E 9 E, f l : i . t : E hf i 6 ! b g s" ' - .8 E r$f sgdi E: s* f f > *H b f l " " 6 r s r $ , { f r : . qsgs I8?^ " ,jFIeje, ; aiEf i * I et gE: = !Y"i-r lB*$o s o N b - r . q - o o o r q e f l F : 5 t E i . A ! " ! 5 3 t g By l q s q E a i ! * o o u ? , 5 5 u " 8 8 8 d o , E . i 5 S i ^ d j; i ; i ; i f$ Ht i *s , lE jc jg [ , E *$,E;€sf - f i ;i g E f l i B a d b i h e E I . o I d i a - l o - 9 ; 5 4 ; S f - ;- - o r d o o 6 o o t

- r i * " - ' t , 6 t t t l o 3 " : ii 9 ? : ? E 1 ' . 9 n 3 ! i €E!6 ' 6 i : s l bsessa j . s t iE : s { , r g€ : i g : iE j, H . 5 , 3 . } i ? o 1 . , ' P 3 j ? 9 t g . f a g S e g i r * € 9 ut E € , ? € i g g g g h g g ' l i ? E i i d E ' : 3 d . q - ? g - 9 o . '[ S i E H g E d 8 i r * " 9 E d c @ o o s o o o E @ ; d i o + h 's r q s s s n i e o > . d s o * q r € 6 8 6 ,

" 3 o " € ? . d 9 5 f , ' . \- U ! i r b E 5 ? " q , - - * ? i o I i u . 9 . 9 - E e g E . g d g o " B R f i$ Ef l? iE3EF E$; { f rEH EeE"rs .8 {ssE:d€6f i :Fsi ; e f i i , i ; a 3 ; i . ? 1 9 f , f . 6 . . i i E " , i @ . ; e j g ; d E g J s t , 8 e J 3

o i o < a 6 N d o <o r _ J 6 0 6 N d H N F 9 0 0 N O O d @ r o o 6 . .r I . . N H O d F 6 F . .

; F 3 ^ ! : o . F o o o H o o o d H d . .

! a € o < H 6 < o - f< t . 6 N 6 H o s € o s g @ i < o r i o o o r . i oI I . . . . . . 6 . F { d s @ ; o o

; a 3 ^ 5 9 - . @ o o o d o ! o R

4 @ o 0 6 0 0 0 6 do I o o o 6 N < d N 6 0 0 N <I o . . , . . . d .

x 3 H ^ = : - . r m o o d o r o

I d 3 @ i € 6 rN d F S 6 0 N 6 A O d N O i i o

r g€ o o N N < O 6 0 N O O O 6 b o6 H O i d d o

0 | o o @ o o o F r{ d c F o 6 N a 0 6 d m 9 i l N 6

s E f i ^ : g - r 6 N o o o o o o

| 6 < a 6 6 0 S i6 d n 9 0 0 N r d o d < @ N o r

6 0 d N d o o < 6 N o o o o o o! q 6 i d o

@ @ F 6 d @ < N d . . . . 6

i d o o o o a o o , . . . on d d o

9 6 6 i 6 F N 6r 6 0 6 0 0 0 6 a 3 d o o r

o a 6 0 0 d o o o o N d < so d 6

6 i o < d d <@ N O @ d O 0 6 H S 6 < 6 0

a $ o r o @ i i o J a o 6 0o N d a d o

d R 6 < d r 6 F r oo I o 6 0 < o < < F a o r 6 0 Nt o

; g 3 * 5 3 o ^ o o o o a F o 6

@I | < O 6 r < O d 6o d < m d o o o F F d 0 o @ H rt g

o o o o F @ o 9

oo | < 6 0 9 6 0 0@ d o o o 6 0 6 0 d d 6 0 N o 4t ,

< o o N i o o N o o o o o - o @n q o d d i 6

$ f . S . . i - N 3 5 S - . . .r g

< d o o t s N o i o o o o N o o o6 d e i N J

E a 6 d < 6 i N oN I d d N < d d 6 0 d < 0 6 < dr !

g ? g " R E . S N d o i o d a o

a E o 6 d o 6 <N | 6 0 6 0 0 N d 6 N F O 9 6 N

o i o 9 @ @ o i d N o o o o N o< 6 I i d a o N o

O I t s @ < O N N F NN p r F d N o d F @ o N N o 6 o

g . 8 s ^ 5 9 0 - N d o o r o 3 0

o | 6 0 r G d < 6 0 0 0N ! F N 6 i O 6 0 3 0 0 A O < O O O O N O O O O o NI d . ! a d N a o 6 a 6 F ,

3 3 S - 5 : o - d d o o o o o o d o o N

R .

4 | @ 3 H d N 6 0N , 6 < o < O 6 6 F o i o N f < O H f d d O F i . S

o A O N i O o N o o o o < N O O 6 d F . ! i< ! o i i a 6 6

H Fd | 6 0 0 N o o < o oN ! t N O O O 6 0 N O N d @ O @ d 6 d d 6 9 6 @ O i

I a . . d o o o i F d N @ .6 @ O N d O O F N i O O < O O O O g O O i< ! 6 i a

o . > o ! oz F a a t

! i @ d 0o b ( { o c o r @ 6 d o !H \ O O i O O O O n O O d ' \ Oa o o r d d o E o d 6 ( n o o 6 ! t d c 6 ! ! !

f i i $ 3 " " " " E o 2 & q q a o H F o o z > N r @ o 3

o q I

@

o o < N o 0 0 0 0a d o d < o d o N

N Or

6

o @ 9 $ 6 0 0 r @o H o N F O i r d

o n < 6 0 0 6 0 9r H o N o o i t s N

F

H @ 0 9 3 0 0 F5 { d 9 N O 6 F d G 6 d 6

. . ' . ' . o .d o a F o 6 6 d o o o o ! o6 i d

o N o c i do € N o , 6 9 N 3 6 . . . O

F d a d t d € o o o t . . 6n d d 6

N F o o a < 6 <m @ 6 F N 9 0 N O r . d . O

9 d o 6 0 H o o o o . o . r

s r N o 6 0 0 0d d N F d O O O i O . d . O

@ N O O o d g d o o . o . 6

N O o @ d r o o3 F N O N 6 O d 6 O . t a . 6

o

- o

0 ar d

@

r d

r €

r €

r.r q

t i

N N N S O ' 0 6d d o d < . d dt o

O A

@N

. . . ' . . . . 6

. . . . . . . . o

tN

o o o o o o i s rd o r o F o 0 d

r

o o o o o * o o f i4 6 6 @ O O O N

o or

No N o N o o d o ni o r 6 6 0 0 N

o a o

N d N N O 6 6 . .@ d d o o N t .

r O N . .

d

o N . 6 0 0 0 6 0i i r N O o N i O

N d o 6 d o F 6 6a d o i o o N d @

I N O O3 F

o o . o o o N i .i N t i N H N . .

. N 9 I !

F N d o O O r . 'N O O F 9 O | .

i o € i . .

0 6 o @ O 0 6 . 'O d O d O < O i .

d O N . .

@

a

oo

Iq

:I

4

o

o

I

o

d

o

x&

nd

cll

i

d € @ o o 6 F i6 F O g 6 d F O A 6 6 i l N

d N N o o 6 6 N o o o o i ao a d

@ N O 6 d N @s @ o o N N € o a N . a . o

H d N A O O O N O O . O . 66 d d d o

. ,] . 3 * - * $ - ! 3 S - 3 .I H r . . . . . d .

6 0 0 i o d o r d N o o o o r oI ! 6 d i a v 6

r ? o r o 6 Ni r . l o d o o . o o o o r . . . r6 4 J i < O . T N N O O . . . 66 g n d i

N I O N F d A O Id g o r d o N o

f r . 8 ; a : s - e d N o o . o . 6

o | 6 a o N o 6 uH ! < O O N N 6 0 N H H . d . O

N p o i 6 N o o o N o o . o . @6 A 6 A a a 6

| / d F o d o oq \ r i 9 r d @ d o @ d H a t d . o

N E O a A N O O O N O O . O . 6n ! 6 d d

N 6 d < r or r o o . a F o N o

N N N O . 0 6 o o O

tered rocks that typically form cliffs and promon-tories. The inner alteration zone is well-defined bysharp transitions from soft, highly altered rocks to

N@ <

o 6 0 0 0 0 d 6 N o6 6 o d < d N d @ .d < d r N o N

F T

N d d O O O b . . id o N F N @ d . . .

A G d o . . i

harder, less fragmented and less altered rocks(Fig. 7d). flssg innsr areas consist of a series of nar-row anastomosing pipes or sheets of highly altered

. q iZ Eu 6 - P

t,{ , d.{3heB?3"^4b" f; ",,. ",,i. E i F 1 " o " " E o z ! 4 t s @ o F r 0 o u z > f r J d o@ e F r d

6

t s N

€ A

- d

' l ' e@o {

g 6

o 6q p

t oo

o

q

o

I

{ F

@ oo q q

q o

o a o

d 6 ao > cd - . qg u o

, E 8g a zd o

d o9 Zo

q

c oF F

NORTHERNRIFI ZONE

526

2

TIIE CANADIAN MINERALOGIST

HORST AND SOUTHERNRIFT ZONE

Ug Nmbot

Ftc. 6. Histograms of Mg-numbers for lavas from north-ern and southern valleys. Bars give range of Mg-numbers of MORB from 86'W, as calculated from Bal-lard et al. (1979); number of samples in each range isindicated.

materials @ig. 7e), probably related to the positionsof minor faults, flow-bedding planes and joints(Fie. 8).

Peripheral to the inner alteration zone, the closelyfractured lavas commonly have red to yellowish coat-ings or stains primarily concentrated along fracturesor parted intraflow surfaces (Fig. 7f). Microscopi-cally, these are thin (<5 mm) palagonitized andochrous zones within the fractured lavas or clay-silica coatings that washed offpalagonitized surfaces.Chemical and petrologic comparisons of these"stained" samples with nearby fresh glass indicatethat major and minor elements have not beenaffected by this type of "alteration", which is sim-ply low-temperature precipitation (compare l65zt-3Awith 1654-6B, Table l).

Meter-scale alteration channels locally breach thetop of the horst; here sulfide mounds have formed.These channels terminate within several meters ofthe seafloor and form patches and layers ofintenselyaltered material within well-preserved pillow lavas(Fig. 7d). Centimeter-scale conduits extend thtoughthe pillows, following local fractures and selvagesup into the sulfide mounds on the top of the horstblock @g. 8). The relatively unaltered nature of thelava flow covering part of the top of the horst is enig-matic (1654-3, Table l). The trapped hydrothermalfluid may have been rapidly cooled through dilutionwith ambient seawater within a couple of meters ofthe seafloor. Less likely, a final extrusive event mayhave occurred after hydrothermal activity ceased.

The hydrothermal upflow zone encloses a com-plex stockwork of anastomosing veinlets whose den-sity is controlled by the presence of fractures andextent of brecciation @igs. 7i to 7l). Glassy andesiticpillows are fractured radially and have exfoliated(Fig. 7g). Basaltic sheet and lobate flows are frac-tured irregularly and along flow planes (Fig.7i).Hyaloclastite beds have undergone pervasive inva-sion and alteration by hydrothermal fluids(Figs. 7d,l,m,n).

The veinlets are typically filled with a matrix ofspheroidal, locally desiccated amorphous and cryp-tocrystalline silica either replaced or overgrown byFe-rich chlorite, euhedral pyrite, chalcopyrite,cristobalite and minor sphalerite (Figs. 7i,k). Sulfidemineralization is largely restricted to the center ofthe upflow zone a:rd occurs as veins and dissemina-tions @ig. 7j). In peripheral areas, fractures aremuch less dense and more typically filled with amor-phous silica, amorphous hydrous iron oxides, andtraces of iron-rich clay minerals.

Chemical analyses (Table l) of rocks that enclosethe alteration zone indicate that they are andesite,FeTi basalt and ferrobasah. In the inner alterationzone, rocks were strongly depleted in Ca, Na, K, Mn,and enriched in S, Fe, Cu and Zn (e.9., Fig. 9). Siis highly variable but also tends to be depleted;however, Al remained relatively immobile in all butthe most altered samples. Mg is highly variable,showing local redistribution or addition leading toprecipitation of chlorite on the margins of conduits,and almost total depletion in the cores of conduits.Ti and Al are unaffected even in highly altered lavas,and can be used as indicators of protolith composi-tion.

Highly altered rocks comprise assemblages ofsmectite, silica (weakly crystalline cristobalite), minorchlorite and uncharacterized clay minerals, rutile ortitanite, and "leucoxene". No primary plagioclase,augite, olivine, pigeonite, titanian magnetite orilmenite are preserved. Most clay minerals are toofine-grained for microprobe analysis, but X-raydiffraction patterns, together with SEM-EDS ana-lyses, are consistent with a chlorite/smectite mixture.Preliminary microprobe analyses of the mixed-layerclay minerals and chlorite indicate that both have abroad range of compositions, even within a singlesample (Fig. l0). For example, sample l@9-2,fromthe center ofthe alteration zone, has a very iron-richclaylchlorite assemblage in the most itensely sulfidicportion (sample 2a), but is quite magnesian awayfrom this area (sample 2e). The Mg content of theclays seems to reflect the Mg content of the wholerock (Table l). Pyrite, the most cornmon sulfide, isfinely disseminated in veinlets and their immediatelyadjacent wallrocks (Figs. 7j,k).

Less altered rocks that surround or flank thehighly altered areas @g. 8) have preserved ilmenitecores within "lsucoxene" (or titanite) envelopes;magnetite has been converted to hematite, and someplagioclase survives within clay envelopes. Intensealteration is restricted to within a few centimeters offractures and is characterized by chlorite in the wall-rocks immediately enclosing pyrite-cristobalite vein-lers (Fis. 7i).

On a microscopic scale, alteration seems to havebeen controlled by episodic fluid movement throughnetworks of fractures, because multiple layers of

EXTINCT ITYDROTI{ERMAL SYSTEM. GALAPAGOS RIDGE 5n

differing mineral assemblages are arranged concen-trically about fractures or fragment surfaces (Figs. 7ito 7l). Highly altered glassy pillows and interbeddedhyaloclastite display intense fracturing and complexalteration zonation that is broadly and distinctivelybanded (Figs. 7m,n). However, mineral assemblagesoverlap considerably between adjoining zones. Fur-ther evidence for multi-phase alteration proccses liesin rhythmic microlayering (mm to <5 pm)(Figs. 7o,p) within broader (cm) layers of kaolinite-smectite-chlorite-cristobalite-sulfide, and multiplecrack-seal structures within the veinlets themselves(Fig. 7k). Rocks from the surrounding outer altera-tion zone display only faint bleaching around thinfractures (2m pm) or around veicles; these may con-tain silica or an amorphous Fe-Ti oxide, accordingto microprobe and SEM-EDS analyses., Strontium and orygen isotopic analyses of a wide

variety of fresh to highly altered rocks provide addi-tional information about the conditions of altera-tion and mineralization in and aroundthe stockworkzone (Smith et al. 1986, Smith 1987). Fifteen hand-picked and acid-cleaned glasses have 875r/863rvalues that range from O.70251L to 0.7W22, anddrsO (relative to SMOW) from 5.5 to 6.2 (Table 1).Most of the samples, however, have Srisotope ratiosless than 0.702665 and dr8O less than 5.9. Thesevalues are typical of fresh, normal (N-type) MORBfrom Pacific spreading centers (lto et ol. 1987, Mac-Dougall & Lugmair 1986, Whrte et al. 1987). In fact,the range of Sr isotope compositions of our sampleset is even more restricted than those previously

reported from the Galapagos Rift (Perfit et al. 1983,White et al. 1987), suggesting a fairly homogeneoussource and little or no crustal contamination duringmagma evolution.

In contrast, 19 altered lava and hyaloclastite sam-ples exhibit extremely variable Sr isotopic composi-tions (0.702678-0.7089M, Table 1). Oxygen isotopicratios in seven of these sample range from 6.6 in theleast altered sample to a low of 3.9 in a hiehly alteredand minslalized rock (Smith 1987). In general, themore altered a sample appears in thin ssctiea 1s.9.,fewer relict minelsls, more chlorite), the higher875r/865r aud lower dlEO ft has. Samples that haveexperienced only mild alteration (e.9., clay replace-ment of primary phases, oxidation of mafic andoxide phases) have slightly elevated uSr/86Sr valuesand heavier l8O relative to fresh samples (e.g., sam-ple 54-4, Table l). Samples such as these are com-monly described as "weathered" rather thanhydrothermally altered. Samples of massive sulfidechimneys and Si-Mn-Fe mounds have the highestSr isotope ratios (0.707080-0.709401).

Strontium concentrations in fresh la'ias range fromaround 60 ppm in basalt to 80 ppm in andesite(Table l). The weathered samples have similar toslightly elevated concentrations (80-100 ppm). Mostof the hydrothermally altered and mineralized sam-ples have Sr abundances less than 55 ppm. The lowerconcentrations correspond to depletions in Ca whichcan be related directly to the extent of feldspar des-truction and accompanying chloritization ofindividual samples. Depletions in these elements can,

Ftc. 7. (See following pages.) Submersible-based photography. (a) Weathered remnants of sulfide chimneys on southmounds (1651-l). Average compositions are given in Table 2. (b) Looking southwest into top edge of north wall(2536 m) ntzr site 1654-6. Ten cm of sediment cover oxidized sulfide mound and underlying alteration zone.(c) Hyaloclastite pinnacles exposed on north cliff face l0 m below horst top Q547 m). Looking south near site 1654-4; slightly altered outer zone of pipe; field width about 4 m. (d) Oxidized sulfide (yellow) at contact of highly alteredhyaloclastite (white) and relatively fresh glassy lobate flows (25216 m) near site 1654-4; field width about 2 m. (e) Closelyfractured sheet flows display oxidized veins of iron-rich sulfides. Fractures are normal to layering. Near site 165r[-8Q5a2 $ north wall; field width about I m. (f) Glassy pillow lava is fractured, bleached and oxide-stained, espe-cially along selvages. North wall looking south (25,0 m); field width about 2 m. (g) Highly altered glassy pillowbasalt (near 165U). Pronounced exfoliation and intensively fractured; selvages consist of soft clay and pyrite. Northwall looking south (2540 m); pillow about 0.5 m. (h) Hiehly altered piilow andesite; sample 16521-8L was taken fromspace marked by a talus of pyritic sanid Q542 m); field width abour 1.5 m.Hand specimens and photomicrographs: (i) Crackle brec{a of partly altered ferrobasalt (1649-2C, Table l) cementedby pyrite-cristobalite veinlets from inner zone of pipe (north wall,2543 m). Alteration forms Iayers ibncentric tofractures, but is also localized on vesicles and ceitain flgiw layers. (i) Slab of highly altered pyritic stringer (1649-2Al) with pronounced alteration from white zones (kaolinite, silica, smectite) to gxeen zones (chlorite, chlorite-smectite). (k) Stockwork veins show multiple crack-seal striping of pyrite-chalcopyrite (black) and cristobalite (white).Glassy, microlitic ferrobasalt (1649-2A) from north wall of horst; field width 6.5 mm, plane light. (l) Chloritizedandesitic hyaloclastite (1654-5) is cemented by cristobalite and sphalerite. Many shards are hollow. North wall at2546 m. (m) As above; large brown sphalerite crystals in silica matrix (white). Patchy and variable alteration in glass.Field width 6.5 mm, plane light, crossed nicols. (n) As above; highly altered hollow fragments with concentric bandsof chlorite, chlorite-smectite, silica. Interiors infilled with precipitates of similar compositions. Field width6.5 mm,plane light, crossed nicols. (o) Highly maenified interior of large fragment in 7(n) displays micrometer-scale tlyth-mic precipitation of chlorite and silica; field width 1.7 mm, plane light. (p) SEM backscattered electron image ofa single chlorite spheroid (upper rigbt in 7o) shows rhythmic 5 pm layering in a continuously grown crystalline matrix.Darker layers and spheres (SEM-EDS sites 3, 4, 6) are more Ti-rich and Fe-, Al-poor than white layers (as in site 2).Bright grains (5) are pyrite, erey band (l) is interpreted to be apatite.

528 THE CANADIAN MINERALOGIST

EXTINCT HYDROTHERMAL SYSTEM, GAL{,PAGOS RIDGE 529

530 TIIE CANADIAN MINERALOGIST

Table 2

Massive Sulfide ComDositions

Dive No. ofsamples

Cu(wt.%)

Zn(wt.%)

Fe(wt.%)

s(wl.oA)

Pb(ppm)

Ao(pp-m)

Au(ppb)

16/;6

1648

1649

1650

1 551

1662

341

5

2

2

2.17

5.52

0.41

5.42

6.90

7.86

v.zJ

0.45

0.07

0.93

2.44

o.47

16 .6

20.s

45.5

32.0

29.4

28.2

31.6

49.8

54.9

41.4

49.3

47.3

104

145

180

126

240

77

t 1

20

1 2

1 4

33

1 8

66

65

91

7 1

143

81

Averagesall samDles 4.7 0.8 28.7 45.7 145 1 7 80

1 650 mounds: top of hont; all otherr: south base of horst and southern valley (Figure 4)

in part, be a result of dilution of the whole-rock Srby the addition of Sr-poor sulfide minerals duringhydrothermal alteration and mineralization, butseawater-rock interaction is likely the moredominant cause.

In general, there is an inverse relationship betweenSr concentration and 875r/865r in the hydrother-mally altered samples. A positive correlation existsin the weathered samples. In a few rocks that havediscrete concentric zones of alteration (cm-scale),subsamples were drilled for isotopic analysis. Theresults showed that the highly chloritized zones (com-monly on the rims or closest to mineralized veins)have the highest 875r/865r and lowest Sr ppm, bothopposite to the least altered cores of the s€une sam-ples. Intermediate zones, typicdly composed ofmixed chlorite-smectite, have intermediate values.The Sr isotopic variations in individual samplesparallel the trend observed in the whole rocks. Itappears then, that macro-scale hydrothermalprocesse$ within the stockwork zone also occurredon a micro-scale within individual rocks.

The full results of our isotopic studies will be dis-cussed elsewhere; however, our initial data providesome constraint on temperatures and water-rockratios during alteration. Many studies havedocumented that increased 6rEO relative to mantlevalues (5.6-6.0) are the result of seawater alterationat low to moderate temperatures of 0 to <200oC(e.9., McCulloch et ol. 1981, Barrett & Friedrichsen1982, Stakes et al. 1983, Bohlke et al. L984, Nt etat. 1986). Other studies have shown that low-temperature alteration or weathering by isotopicallyheavy seawater will effect modest increases in875r/865r with little or no shift in Sr concentrations(Dasch et al. 1973, McCulloch et ql. 1981, Spooneret al. 1977).In contrast, metamorphosed and highlyaltered rocks from the seafloor, particularly thoseassociated with metalliferous deposits, have high6180 depletions and more radiogenic 875r/865r

caused by interaction with hydrothermal fluids attemperatures >300oC (Vidal & Clauer l98l' Alba-rede et al. 1981, McCulloch et al. 1981, Ah et sl.1986).

The temperaturedependent fractionation of ory-gen isotopes make 6180 values extremely useful fordetermining temperatures of alteration, whereas thesmall mass differences between Sr isotopes makesthem insensitive to temperature differences. Shiftsin the Sr isotopic composition of oceanic rocks can,in the simplest sense, be considered a mixing processbetween unaltered igneous material and seawater orhydrothermal fluids (Spooner et ol. 1974, McCul-loch et al. l98l\.

To determine the temperature and water-rockratios operative in the fossil hydrothermal system wemust assume the compositions of the end-members.Our data allow us to place fairly tight constraintson the composition of protoliths in the study areaand the composition of seawater is well-known. Wecannot, however, ignore the fact that as a hydrothei-mal system develops, the composition of the alter-ing fluids must change as do the compositions of thecrust through which they flow. Our results suggestthat multiple pulses of fluids have affected somerocks in the stockwork zone, or that even smallvolumes of rock did not attain isotopic equilibrium.Relatively systematic changes in 875r/865r, Sr ppm,and 6180 in mineralogically zoned samples indicatethat: (l) the composition of hydrothermal fluidscould not have been much different than those meas-ured in vents from the EPR at 21oN (Albarede e/al, L98l), and (2) the isotopic values measured inwhole rocks probably represent the integrated effectsof hydrothermal alteration over the lifespan of thesystem.

Using the methods discussed by McCulloch el a/.(l9Sl) and Stakes & O'Neill (1982), alteration is esti-mated to have occurred from 150 to 320'C. Theupper value is in good agxeement with mineralogi-

EXTINCT HYDROTHERMAL SYSTEM. GALAPAGOS RIDGE

15omteruglnoug mud

531

w

eroded chlmneyo

hyaloclaetlte

OUTER ZONEtraoturod

unelt€red - w-?akly-anotoc(Mg, Fe-chlorlte)

andpglaglc aodlnsnl

Dlltoi ttows

le6s-arsrod Ilocaly I

Ftc. 8. Schematic section of the north wall (central horst) in the vicinity of the alteration zone, depicting stnrcturaland petrological features and their relationships to sulfide mounds. Reconstruction is based on visual observations,video, and still photogaphy from ALVIN dives.

INNER ZONE (PIPE)

clossly ftactursd and brocclated:hlohly alterod banda flank traclurea:kaolln, Itnoctlte, Fe-ohlorlts, sulphlds,slllca atockwork ot volnlets. Patchoaot leoa-altsrod rock3 Inigryong:Fe, M0-chlorlte, slllaa

OUTER ZONEfraclured

weaklY+ unalteredelloredlnclplont alteratlon ontraclureo: gllloa, Fe ory-hydroxldeo, clay

GSC

MAGNETICS EXPERIMENTS

A deep-towed magnetometer profile was obtainedacross the rift valley at an altitude of 100 m abovethe seafloor to determine the presence and extent ofmagnetization anomalies generated by the sulfidedeposits and the associated alteration zone. Magneti-zation lows had previously been detected over somecontinental hydrothermal areas, such as Wairakei(Studt 1959) and the Salton Sea (Koenig 1967), andalso over some ancient massive sulfide deposits, asdescribed by Johnson et al. (1982) for a deposit inthe Troodos area of Cyprus. Similar anomalies havebeen detected over submarine hydrothermal areas(McGregor & Rona 1975, Phillips et ol. 1969,Rora1978). The marine studies, however, had been con-ducted using surface-towed magnetometers withreceivers several km above the ocean floor. Ourexperiment was designed to place the receiver in closeproximity to the seafloor, and centered overdocumented massive sulfide deposits.

The profile obtained from the submersible-mounted magnetometer was upward-continued to alevel plane 2.2km below the sea surface, and then

cal data and measured exit temperatures of fluidsfrom active vent fields (Edmond et al. 1979, Delaneyet al. 1984). A simple two-component mixing calcu-lation using fresh Galapagos basalt and seawater asend-members roughly matches the variations in Srand 875r,/863r in altered samples (Smith 1987). Cal-culated water-rock ratios for whole-rock samplesvary from 0.25:l to 95:1, assuming perfectequilibrium mixing. In one of the concentricallyzoned samples, calculated water-rock ratios varyfrom slightly greater than I in the least altered coreto approximately 13 in the altered rim only 2 mmaway. As the efficiency of exchange decreases, thecalculated water-rock ratios will increase in a recipro-cal manner. Consequently, if the mixing process wasonly l09o efficient, the water-rock ratios would bel0 times greater than those discussed above. In sum-mary, it seems that the most highly altered andmineralized samples in the stockwork zone haveinteracted with large volumes of seawater-likehydrothermal fluids al Z -300'C, and that multi-ple pulses of this hot fluid preferentially passedthrough small volumes of rock along closely spacedfractures that led to the seafloor.

532

ALTERATIONPATHWAYS

SlllcaKaolhgmect l te

o1 2 r o a a o M g O ,

n " l

Ftc. 9. Na2O versas MgO plot (top) compares fresh andaltered lavas from the central horst. Na2O is stronglydepleted in samples that have ferrobasalt and basalticandesite protoliths. CaO versus MgO plot (bottom)shows similarly strong depletions in CaO.

was inverted for the magnetization solution (over aconstant thickness layer of 500 m) using the Parker& Heustis (1974) Fourier inversion technique(Fig. ll). The profile shows a 40OGgamma anomalylocated approy<imately 200 m north of the horstblock that divides the survey area into two zones ofrelative magnetization. The northern valley has arelatively high magnetization compared to the horstblock and southern valley @ig. l1). A detailedsubmersible-based magnetometer survey alsoencountered the same magnetization contrast 200 mto the north of the horst. This corresponds to theposition of the Neovolcanic Zone. The mapped sul-fide deposits displayed no local anomalies, possiblybecause of their dominantly pyritic composition andtheir oxidized natlue. The dramatic contrast in mag-netization2.AO m north of the horst is unlikely to besimply a function of age (i.e.,low-temperature alter-ation to maghemite and ultimately to titanite) sincethe age contrast is only of the order of a few thou-sand years. Petrologic differences between the north-ern and southern valleys are also unlikely to be thesource of this contrast: FeTi basalt and andesite areas magnetic lf not more so in comparison to MORB(e.g., Vogt & De Boer 1976). The presence of per-vasive alteration of Fe-Ti oxides and brecciation inthe horst block area suggests that the magnetizationcontrast more likely reflects the presence of freshunaltered crust in the northern valley and hydrother-mally altered crust in the southern valley.

TI{E CANADIAN MINBRALOGIST

a Froah lava: F9rroba8oll FeTlbaell . andoalt€

^ Htdrothgroally alt6r0d laya, contral hor6i (pyrlle-oor.soted)

DrscussroNPet ro che micql re lat io ns hips

The striking spatial association of highly fractio-nated rocks and the sulfide deposit suggests the pos-sibility of a genetic link. The coincidence of fractio-nated rocks and massive sulfide deposits has beendemonstrated in other massive sulfide areas, such asCyprus (Schminke et ql. 1983). In other hydrother-mal areas, notably along the EPR, little has beenreported about the composition of the associatedlavas. The significance of highly evolved lavas is two-fold: 1) sulfur and volatile contents in FeTi basaltincrease dramatically but sulfur decreases in theandesite; 2) small isolated magma chambers arerequired to create an environment conduciye toextreme fractional crystallization (Perfit et al. 1983).

A decrease in sulfur content of andesite melt mayhave resulted either from immiscible separation ofsmall amounts (<0.5 wt.9o) of monosulfide liquidtogether with titanian magnetite micro-phenocrysts(Fornari & Perfit 1982, Perfit & Fornari 1983), orby degassing of sulfur from the magma. The formerprocess would result in removal of most Ni, Cu andCo from the residual melt @erfit et al. 1983, Clagueet al, l98l). Tlre separated sulfide liquid would, inthis case, result in a portion of the subvolcanic cumu-late enriched in monosulfide solid solution (n.s.s.)relative to most MORB-related subvolcanic intru-sions.

Zinc does not partition as strongly into monosul-fide liquid and would remain in similar amounts inandesite as in basalt (Shimazaki &MacLean 1976).Escape of S as a volatile phase would not be accom-panied by depletion of involatile elements such asCu from andesite. Mor@ver, low vesicle contentsand high volatile concentrations (HrO, Cl, F) donot indicate de-gassing (PerfiI et al. 1983). Prelimi-nary analyses (Table l) indicate that andesite sam-ples have only half the Cu content of most of thebasalt samples. Petrographic observations confirmthe presence oftiny zn.s.s. droplets, closely associatedwith magnetite, in the andesite.

The presence of highly evolved lavas on theEastern Galapagos Rift suggests extensive crystalli-zation of small magma bodies (<30 km3) at depthsless than 2 km beneath the seafloor (Perfit et al.1983). These magma chambers provide heat to local-ized hydrothermal systems, enabling them to attaintemperatures of at least 350'C; such systems prob-ably exist for less than 10,000 years (Cann et al.1985/1980. These ephemeral chambers may developas small cupolas above larger steady-state mag@aticreservoirs that may be related to relatively small off-sets in the ridge crest (DEVALs of Langmuir et al.1986, OSCs of Macdonald et al. 1984). Along theEastern Galapagos Rift, development of these shal-low intrusions seems to be enhanced by the thermaleffects ofrift propagation (Fornariet ol. 1983, Per-

EXTINCT HYDROTHERMAIT SYSTEM. GALAPAGOS RIDCE 533

Frc. 10. Preliminary analyses of chlorite and smectiteminerals from the central area of alteration @ig. 8). (A):composition of chlorite from tow samples, with com-parative data for the Mid-Atlantic Ridge Q{umphris &Thompson 1978); Kane Fracture Zone (Delarrey et al.1987); Bayda massive sulfide deposit, Oman (Collin-son 1986). @): compositions (mole proportions) of chlo-rite (riangles) and (mainly mixed-layer) clays (circles)from the central part ofthe alteration zone. Field forHole 5048 chlorite,/smectite compositions from Alt etal. (1986). All analyses by wavelength dispersionmicroprobe techiques.

sulfur-rich hydrothermal fluid.Fractional precipitation of Zn relative to Cu within

the zone of ascent can be ruled out, as the sulfidesin both the Galapagos alteration pipe and in ancientalteration pipes are all Cu-rich. Furthermore, asdemonstrated by many solubility calculations (seeFranklin et al. l98l for summary), Cu precipitatespreferentially from a high-temperature hydrothermalfluid. Copper minerals are thus most abundant inthe stringer zone, and the remaining fluid is relatively

fit et al, 1983). In contrast, where sub-rift chambersare continually replenished, erupted basalts are moremafic and uniform in composition; this may be thesituation in the Neovolcarric Zone (north valley).Langmuir et al. (1986) have recovered similariyhighly fractionated lavas only at DEVALs along thenorthern EPR. A crucial test of this genetic mbdellinking fractionation processes to massive sulfidegeneration on the seafloor will result either frommore exploration for sulfides in areas of fractionatedlavas, or through a more thorough assessment ofpetrographic trends in lavas associated with majorseafl oor-sulfi de deposits.

Sulfide compositions

The most striking aspect of the sulfides in theGalapagos mounds is their copper-rich composition(Cu:Zn 6:1; Table 2) in comparison with otherlpflogr deposits (Cu:Zn -0.1; Bischoff et ot. 1983,Kappel & Franklin l98S). As with all of the volcanic-associated deposits on the modern seafloor, these arewithin the Cu-Zn group (Franklin et al.lggl). TheGalapagos deposits are similar to sulfides on an off-axis-seamount at l2.N (H6kinian & Fouquet l9g5),to the deposits at Cyprus, and to the Cu-rich por_tiom of many ancient deposits (Franklin et a/. lggl).At least two mechanisms can explain the Cu-rishnature: l) the fluids were relatively Cu-rich comparedwith fluids in active vent sites, either throughiomepeculiarity in the source region, or 2) through frac_tional precipitation of Cu minerals from the ascend-ing hydrothermal fluid, either in the zone of ascenror within the sulfide mound.

Although the source region cannot yet be exa-mined, the dykes and lavas that comprise the lowerpart of the oceanic crust are most probablyrepresented by the rocks at surface. Perfit & For-nari (1983) noted an increase in Cu content fromMORB through ferrobasalt to FeTi basalt, thedominant rock type. Zinc content ofall types is cons-tant. The relatively Cu-rich FeTi basalt may haveprovided more Cu to the hydrothermal fluid, but asthe Cu:Zn ratio increases from only I in MORB toabout 1.5 in the FeTi basalt, using the simplisticexplanation of metal content of the source rocks toexplain the large difference in sulfide compositionsis not adequate. As demonstrated experimentally bySeyfried & Janecky (1985), the metal content of ahydrothermal fluid is obtained only where the pHis sufficiently acid to promote chloride-ion complex-ing; this acidity is reached in basalt at about 385.C,and the metal content of the fluid is more or lessindependent of rock composition at that point. Thus,although it cannot be ruled out, variation in metalcontent of the source rocks is not a likely explana-tion for the high Cu content at Galapagos. The highcontent of sulfur (as n.s.s.) in the FeTi basalt,however, may have resulted in an exceptionally

/ 161e.4

'r""r#"r"',-W;J

b q-"

534 THE CANADIAN MINERALOGIST

GALAPAGOS 8505f W DEEP TOW PROFILE

MAGNETIC FIELD aT -2.2Km

rite, is consistent with metasomatic alteration thatoriginated from both high-temperature fluid andlocally-warmed seawater. Iron-rich minerals musthave precipitated from the hydrothermal fluid, butthe more magnesian minerals may have formed dueto interaction with locally derived seawater as metal-liferous hydrothermal fluids are notoriously deficientin Mg (Edmond et al. 1979). The delicate concen-tric rhythmic layering of alteration minerals mayindicate a repetitive change in the flux ofhydrother-mal fluid. Crack-seal veins indicate that, at least forshort periods, the fluid pressure exceeded thelithostatic load. The large variation in FelMg ratioof the claylchlorite assemblages, even within singlesamples, indicates that alteration involving directinteraction with the hydrothermal fluid may havegone on immediately adjacent to the major fluidchannelways. Within a few cm of the channelways,rapidly heated, locally derived seawater was contem-poraneously reacting with the basalt.

The loss of alkalis, and strontium isotope data,indicate that the water-rock ratio was very high inthe alteration core, consistent with observations inpreserved deposits (Mottl 1983, Roberts & Reardon1978). The general increase in 875r/865r with inten-sity of alteration indicates that much of the altera-tion formed from reaction with something other thanthe ascending metalliferous fluid. Several studies(e.g., Hinkley & Tatsumoto 1987) have shown thatthe Sr isotopic composition of unmodified hydro-thermal fluid is close to that of the host basalt. Thehigh 875r/863r compositions must have been gener-ated from reaction with advecting, locally derivedseawater.

The range of temperatures calculated from theoxygen isotopic data are consistent with highly vari-able mineral assemblages associated with the wall-rocks to the stockwork veins. Both progressivelyheated local seawater and cooled hydrothermal fluidcould produce this range of temperatures. Magne-sian clays must have originated from interaction withadvecting seawater, or by redistribution of Mg fromthe zone of interaction of hydrothermal fluidimmediately adjacent to the stockwork veins.

In summary, the mineralogical and isotopic dataindicate that a large volume of locally derived sea-water had a major role in producing the alterationwithin the basalt. The vein fillings (silica, pyrite) andalteration products in their immediate wallrocks wereobtained, however, primarily from the metalliferoushydrothermal fluid.

CoupanrsoN wrrH OTHER HvonornenualSYSTEMS

The presence of alteration halos around massivesulfide orebodies has long been recognized as animportant exploration tool. Franklin et al. (1981),Franklin (1986), and Morton & Franklin (1987)

M A G N E r l z A r r o N

" ( \ ]

BATHYMETFY NBFa

,aa

ABF

$1,"t,-;1,.,*9;-^- ]

cs

Frc. ll. Deep-tow magnetometer profile over study area.Location in Figure 2. Top profile is the upward con-tinued magnetic field at a level of 2.2 km below sea-level. The magnetization profile (center) was computedusing a Fourier inversion technique.

Zn-rich. Similarly, zone refining takes place withinthe chimney and mound structures, with Cu precipi-tated in the high-temperature chimney linings, andZn precipitated in the lower temperature outer por-tions (Haymon & Kastner l98l), or lost altogether.It may be this last process, taken to completion inmature systems such as Galapagos, that explains theCu-rich composition. As hydrothermal fluid movedthrough the well-established mounds at Galapagos,the cooling high-temperature fluid probably diffusedoutward, causing dissolution of sphalerite andreplacement by copper minerals. As the moundsheated, they became more Cu-rich.

Alterotion assembloge

The petrographic and submersible observationsprovide some new insights into the alteration processand timing of hydrothermal activity. Alterationchannels cut the entire lava sequence, so hydrother-mal activity must have been initiated near the endof volcanic activity, and continued afterward.Although hydrothermal activity probably accompa-nied volcanism (cl, Rise Project Group 1980), themost prolific period of sulfide formation post-datedlocal volcanism. This observation is consistent withthose made at the Endeavor Ridge (Kappel & Frank-lin 1988) and the l2oN seamount (H6kinian & Fou-quet 1985). Evidently the optimum time of forma-tion of large seafloor massive sulfide deposits isimmediately following, and not during, the culmina-tive volcanic event in any specific portion of a ridgecrest.

The mineral assemblage, dominated by smectite,cristobalite, amorphous silica, pyrite and minor chlo-

Anpa/m

o

-2.2

kn-2.4

g 2 r 0 1 2 sKuorutoF

presented compositional classifications of massivesulfide deposits and suggested that genetic relation-ships exist between the type of alteration and theenvironment of formation of these deposits. Underthe above classification schemes most of the ridge-crest deposits are the "Cu-7.n" type. The Galapagosalteration zone is an excellent example of a totallyunmetamorphosed and undeformed alteration zone;the characteristics of such examples are the basis forinterpreting preserved pipes throughout the geolog-ical record, and are of great significance in estab-lishing an ever more precise genetic model for mas-sive sulfide deposits. It is thus important to establishthe amount of similarity between the Galapagos pipeand alteration under major ancient deposits.

The sulfide deposits within the Abitibi greenstonebelt of northern Quebec and Ontario are some ofthe largest and best studied ofthe Cu-Zn class. Theyare in bimodal (basalt and rhyolite) sequences inwhich mafic rocks predominate. Deftriled models ofalteration pipes typical of this area are available forthe Mattagami Lake mine in northern Quebec.Roberts & Reardon (1978) and Costa e/ al. (1983)have identified the following alteration trends: (l)removal of Na and K, (2) addition of Mg, Mn andFe (in chlorites), (3) removal of silica to producechlorite-rich rock, and (4) late potassic alterationcoupled with transformation of chlorite to phlogo-pite.

The basic model of Costa et al. (1983) suegests thatheat flow associated with a magmatic source hasinduced a thermally driven convective flow systemwhereby ambient seawater was drawn into the ris-ing hydrothermal fluid. Furthermore, the highest Mgconcentrations commonly are in the upper and outerparts of the pipe, where a high flux of locally warmedseawater through the country rocks interacted withthe lavas to produce Mg-en'iched chlorite. As thezone of high flux of hydrothermal fluid in the coreof the pipe is approached, the Mg available from coldambient seawater decreased and the relativelyreduced, iron-rich fluids reacted with the volcanicrock to produce Fe-rich chlorite. Therefore, in themost simplistic case, the Fe:Mg ratio would increasetowards the center of the alteration pipe. However,as Cosla et al. (1983) and Franklin (1986) point out,the Fe:Mg ratio in any given system probablydepends on the total flux and the rate of discharge,which directly relate to the amount of mixing withcold seawater and consequent overprinting of a"primary" Fe chlorite-rich core alteration zone bylater Mg-rich chlorite. The Galapagos alteration zonegenerally conforms to this model, with Femetasomatism more dominant in the center of thealteration zone, and Mg metasomatism moredominant outside. The Mattagami deposit, however,has much more Mg alteration than Galapagos.

The Cu-Zn massive sulfide deposits that formed

535

in tectonic settings most analogous to the Galapagossulfides are those in ophiolite complexes. The bestdocumented ophiolitic stockwork zones are in Omanand Cyprus. The Bayda deposit in Oman occurs ina pillow-lava sequence of the Semail ophiolite inter-preted by Alabaster & Pearce (1985) as part of anoff-axis seamount province on the flank of a margi-nal ocean-basin spreading center. As at Galapagos,the stockwork zone consists mostly of quartz, plusFe-rich chlorite and pyrite veinlets within highly frac-tured wallrocks. Qellinssa (1980 outlined four alter-ation stages, viz., qlJartz + hematite; quartz +pyrite (main stage); quartz + chlorite; quafiz +hematite (ate stage). The earliest and latest stagesof alteration have not been recognized at Galapagos,imply'rng that the brecciation-hydrothermal flowevents were triggered and terminated abruptly. AtGalapagos, heated seawater was unable to mix andreact with hydrothermal fluid in the central zone ofthe alteration pipe. However, a peripheral zone ofsilica * hydrated iron oxides on the margins of thepipe indicates that some shallow downwelling andmixing did take place on the flanks of the upwellingzone, but this formed only minimal, superficial alter-ation along fractures (contrast with Bayda whereperipheral alteration is more pervasive: hematite +albite + quafiz + zeolites). The preservation ofcristobalite and cryptocrystalline silica rather thanchalcedonic quartz throughout the inner alterationzon.e at Galapagos supports the concept of very lit-tle post-alteration fluid flow. Silica gel formed bysuper-cooling (quenching) first forms alpha cristoba-lite and would anneal very slowly at low tempera-ture (ambient seawater) to alpha quartz (Plyusnina1984). The similarity in style of the main stages atthe two sites is further emphasized by similarly highFe:Mg chlorites and common mineral assemblages.

The stockwork zones at Mathiati and Kok-kinopezoula in Cyprus (Lydon & Galley 1986,Richards et ol. 1988) are similar to Galapagos inalteration-mineral assemblages and chemical com-position. The Mathiati deposit in particular is anexcellent analog to Galapagos in all respects, includ-ing chlorite compositions (Lydon & Galley 1986).The basic mineralogical and chemical changes inalteration zones at Mathiati, Bayda, and Galapagosare similar; all have undergone loss of Ca and Na,massive precipitation of silica, and the formation ofFe-rich chlorite.

In colrtrast, stockwork alteration zones not over-lain by exhalative massive sulfides (DSDP Site 5048,Mid-Atlantic Ridge and the Pitharkhoma deposit ofCyprus) show somewhat different chemical charac-teristics (Nt et al. 1986, Richards et al. 1988,Fig. 10). As at Galapagos, the stockwork zone inHole 5MB contains sulfide stringers in quartz vein-lets, and displays evidence of multiple mineralizingevents (Honnorez 1981, Honnorez et al. 1985').

EXTINCT HYDROTHERMAL SYSTEM. GAI-APAGOS RIDGE

536 TIIE CANADIAN MINERALOGIST

However, Hole 5048 stockwork is enriched in alkalis,whereas Galapagos is strongly depleted. The altera-tion described by Delaney et al. (1987) for Mid-Atlantic Ridge greenstone is also similar; there, chlo-rite has a high Fe:Mg ratio that distinguishes it fromother seafloor metabasalt. Delaney et al. interpretthe chlorite alteration as a product of a highlyevolved, high-temperature, high-salinity fluid mov-ing through a hydrothermal upflow zone.

It is important to note that massive sulfide depositsand stockwork zones are commonly associated withhighly evolved lavas in ancient greenstone belts andin some ophiolites, similar to the exceptional associ-ation with andesite and FeTi basalt at Galapagos.Most of the lavas in the Troodos exfusive sequence arebasaltic andesite or andesite (Schmincke et al. 1983)and range up to rhyodacite (Robinso\ et al. 1983).Similar andesitic lavas and shallow-level differen-tiated plutonic rocks (plagiogranites) have beendescribed near alteration zones in the Semail ophio-lites (Alabaster et ql. 1979). Campbell et sl. (1982)have demonstrated anomalous fractionation histo-ries for felsic volcanic rocks associated with severalmassive sulfide districts in the Precambrian Shield.They attribute these to very efficient cooling ofhigh-level subvolcanic magma chambers by heat transferinto a hydrothermal system. Thus, the determina-tion of such anomalous petrogenetic trends both infelsic and mafic sequences can be an important guideto new resources.

CoNcr-usroNs(l) The Galapagos sulfide mounds at 85"50.5'W

were formed several thousands of years ago duringa hydrothermal episode associated with faultingalong a horst-type structure. The horst separates asouthern rift valley from a northern rift valley thatcontains the Neovolcanic Zo\e.

(2) The lavas in the Neovolcanic Zone are aver-age MORB, whereas those from the horst and thesouthern valley range from evolved MORB to ande-site. The extreme differentiation nec€ssary to producethis cogenetic suite of volcanic rocks probablyresulted from cooling of a small, shallow magmachamber. Heat from this magma enabled thedevelopment of a large volume of metalliferoushydrothermal fluid at a shallow crustal depth. Evi-dence for such anomalous fractionation of high-levelmagma chambers has been noted in Precambriansequences (Campbell et al. 1982), and may be a keyexploration guide.

(3) The unusual nature of the double rift valleyis also reflected in the magnetic-field data. The highlymagnetic northern rift valley has a magnetic expres-sion that contrasts with the less magnetic horst blockand southern rift valley. This magnetization contrastprobably reflects limited or no hydrothermal alter-

ation in the northern valley, but pervasive hydrother-mal alteration in the mineralized horst and underly-ing the southern valley.

(4) The stockwork-like zone exposed on the scarpdirectly beneath the massive sulfide mounds on thenorth side of the horst is similar in general form andcharacter to those associated with ophiolite-hostedmassive sulfides, in particular those in Cyprus andOman. The amount of alteration is directly propor-tional to permeability, e.9., hyaloclastites are totallyaltered to clays, whereas pillowJava alteration is con-trolled by pre-existing fractures with less alterationoccuring in a zoned pattern inward from silica- andsulfide-filled fractures. Layering represents a hithertounknown pulsation in a hydrothermal system.

Altered rock is extremely depleted in Ca, K andNa, and enriched in Fd* and S. Mg and Si contentsare variable, indicating different amounts of mix-ing of hydrothermal fluid and seawater within thealteration zone. Chemical and isotopic data indicatethat large volumes of locally derived, progressivelyheated seawater, advecting near the upwelling zone,were responsible for much of the alteration. Altera-tion directly effected by the metalliferous hydrother-mal fluid is confined to the immediate zones of fluidupwelling.

The Galapagos sulfide site is the first examplewhere the relationships between alteration and sul-fide accumulation can be studied in the absence ofmodification by post-depositional thermal and defor-mational effects. With exhaustive study it will pro-vide further new insights into the mineralizationprocess.

AcrNIowLmcBlvlsNrs

We are grateful to Susan Hanneman, who workedtirelessly on all aspects of the data collection and onthe laborious and frustrating task of piecing togetherthe dive data. Chris Fox and Joyce Miller devotedconsiderable time and effort to merging the SeaBeam and SASS multibeam bathymetric data setsand we thank them and the URI Sea Beam group fortheir efforts. AJlof IheATLANTIS l/ship and scien-tific complement, in particular Michael Jackson, Tet-suro Urabe, Stan Jones, Lowell Fritz and MakotoYuasa contributed to the success of the dive series.We thank Harald Bdcker and members of the FSSONNE for their valuable on-site collaboration. Thesuccess of our dives reflects the dedication andprofessionalism ofthe pilots, technicians and officestaff responsible for the submersible, and we heart-ily thank pilots Ralph Hollis, Dudley Foster, Wil-liam Sellers and the rest of the ALVIN group fortheir efforts. We gratefully acknowledge the supportof the NOAA Office of Undersea Research for finan-cial support of the dive series. Rachel Haymonprovided valuable discussion and information on theOman stockwork zone. At the Geological Survey of

EXTINCT HYDROTHERMAL SYSTEM, GALAPAGOS RIDGE 537

Rise; Juan de Ftrca Ridge; and Galapagos rift. Bulkchemical composition and economic implications.Econ. Geol. 78, 17 ll-1720.

Boru,rr, J.K., Alr, J.C. & MusHr.rNBAcHs, K. (1984):Orygen isotope - water relationships in altered deepsea basalts: low temperature mineralogical controls.Can. J. Earth Sci.2l,67-77.

Boxarrr, E., Gunnsrnnr-Hoxxonrz, B.M. & HolrNor.nz, J. (1976): Copper-iron sulfide mineralizationsfrom the equatorial Md-Atlantic Ridge. Econ. Geol.71, l5t5-1525.

CAMparLL, I.H., Coat, P., FnANKLTN, J.M., GontoN,M.P., Scorr, S.D., Sowe, J. & Tnunsron, P.C.(1982): Rare earth elements in volcanic rocksassociated with Cu-Zn massive sulphide minerali-zation: a preliminary report. Can. J. Earth Sci. 19,619-03.

CaNN, J.R., Srnnss, M.R. & Rrcr, A. (1985/86): Asimple magma-driven thermal balance model for theformation of volcanogenic massive sulphides. .Earf&Planet. Sci. Lett. 76, 123-134.

CLacuE, D.A., Fnrv, F.A., TnouesoN, G. & Rrmocr,S. (1981): Minor and trace element geochemistry ofvolcanic rocks dredged from the Galapagos spread-ing centre: Role of crystal fractionation and man-tle heterogeneity. J. Geophys. Res. 86,9469-9482.

CorrnsoN, T.B. (1986): Hydrothermal Mineralizationand Basalt Alteration in Stockwork Zones of theBayda and Lasail Massive SulJide Deposits, OmanOphiolite. M.A. thqsis, Univ. California, Santa Bar-bara, California.

Contrss, J.8., Dvtr,totto, J., GoRDoN, L.I., Eolaoxo,J.M., voN HrnznN, R.P. & vaN ANnsr, T.H.(1979): Submarine thermal springs on the GalapagosRift. Science 203, 1073-1083.

Cosra, V.R., BenNrrr, R.L. & Krnmcir, R. (1983):The Mattagami Lake Mine Archean Zn-Cu sulfioedeposit, Quebec: Hydrothermal coprecipitation oftalc and sulfides in a seafloor brine pool - evidencefrom geochemistry, 180/166 and mineral chemistry.Econ. Geol. 78, ll4+1203.

DascH, E.J., Hrocn, C.E. & DruoNo, J. (1973):Effect of seawater interaction on strontium isotopecomposition of deep-sea basalts. Earth Planet. Sci.Lett. 19, 177-183.

DELANEy, J.R., McDurT', R.E. & LurroN, J.E. (1984):Hydrothermal temperatures of 4@"C on theEndeavor segment, northern Juan de Fuca Ridge.EOS, Trans. Amer. Geophys. Union 650973.

Mocr, D.W. & Morrl, M.J. (1987): Quartz-cemented, sulfide-bearing greenstone from Mid-Atlantic Ridge; sample of a high-temperature, high-salinity hydrothermal upflow zone. J. Geophys. Res.92.9r75-9192.

Canada, Gwendy Hall and Gerry Lachance super-vised geochemical analyses, Gina LeCheminantprovided microprobe data on sulfide and claymineral$, Adrienne Larocque carried out XRD deter-minations of clay minerals, and David Walker sup-plied SEM analyses.

At the University of Florida, R. Shuster providedanalytical expertise on samples prepared by M.Lebron. The analytical work has been supported inpart by the U.S. Department of Commerce (NOAA-50ABNR-7-00131) and the National Science Foun-dation (OCE 87 rc\n).

REFERENcESALABAsTER, T. & Pnancn, J.A. (1985): The interrela-

tionship between magmatic and ore-forminghydrothermal processes in the Oman ophiolite.Econ. Geol.80, l-16.

+, Pnancn, J.A., Marr-rcr, D.I.J. & ElnoNsnr,I.M. (1979): The volcanic stratigaphy and locationof massive sulphide deposits in the Oman ophiolite.1z Ophiolites: Proc. Int. Symposium l9?9, Cyprus(A. Parayiotou, ed.). Cyprus Ministry of Agricul-ture, Natural Resources and Geology, Geol. Surv.Dep.,751-757.

ALBAREnn, F., MtcHeno, A., MrNsrrn, J.F. &MrcHato, G. (1981): 87Sr/81Sr ratios in hydrother-mal waters and deposits from the East Pacific Riseat 2loN. Earth Planet. Sci. Lett. 55,229-236,

Arr, J.C., HoNNonuz, J., Lewntr, C. & ErurrmnuaNN,R. (1986): Hydrothermal alteration of a l-km sec-tion thtough the upper oceanic crust, Deep Sea Drill-ing Project Hole SMB: Mineralogy, chemistry, andevolution of seawater-basalt interactions. "LGeophys. JRes. 91, 10309-10336.

BAcrnn, H., Letcr, J. & MancHrc, V. (1985):Hydrothermal activity and sulphide formation inaxial valleys of the East Pacific Rise crest between8o and 22oS. Earth Planet. Sci. Lett.72,9-22.

BaLLARD, R.D., Hor-couo, R.T. & vaN ANonr, T.H.(1979): The Galapagos rift at 86oW, 3. Sheet flows,collapse pits and lava lakes of the rift valley. "I.Geophys, Ras. 84, 5407-54:22.

+, var Axonl, T.H. & Horcoun, R.T. (1982):The Galapagos Rift of 86"W: Variation in volcan-ism, structiue and hydrothermal activity along a 30-kilometer segment of the rift valley. "L Geophys.Res. 87, 1149-116l,

BanRETT, T.J. & FnrEpnrcHsEN, H. (1982): Strontiumand oxygen isotopic composition of some basaltsfrom Hole 5048, Costa Rica Rift, DSDP Legs 69and 70. Earth Planet. Sci. Lett.60,27-38.

BrscHorn, J.L., Rosnvoauen, R.J., Anuscavacr, p.J.,BAEDEcKER, P.A. & Cnocr, J.G. (1983): Sea-floormassive sulfide deposits from 2loN, East pacific

538 THE CANADIAN MINERALOGIST

EouowD, J.M., MsesuRFs, C., Mercuu, B., GRANI,B., Scransn, F.R., Colr,rcn, R., HuDsoN, A., Gon-ooN, L.I. & Coru-ns, J.B. (1979): On the formaiionof metal-rich deposits at ridge crests. Earth Planet.Sci. Lett.6,19-30.

Fomqaru, D.J. & Pnnrn, M.R. (1982): Sulfide fractio-nation: its role in the evolution of massive sulfidedeposits along the mid-ocean ridge crest. .605,Trans. Amer. Geophys. Union 63, 1135.

-, -, Melauorr, A. & ENasrev, R.W.(1983): Geochemical studies of abyssal lavas reco-vered by DSRV ALVIN from eastern Galapagos rift,Inca transform and Ecuador rift, l. Major elementvariations in natural glasses and spatial distributionof lavas. J. Geophys. Res. 88, 519-530.

FnamlrN, J.M., (1986): Volcanic-associated massivesulphide deposits, an update. .Ill Geology and Gen-esis of Mineral deposits in lreland. Irish Assoc.Econ. Geol. Spec. Vol. (C.J. Andrew, R.W.A.Crowe, S. Finlay, W.M. Pennel & J.F. Pyne, eds.),49-69.

-, Lynor, J.W. & SeNcs'rnn, D.F. (t981):Volcanic-associated massive sulfiide deposits. Ecoz.Geol. 75th Anniv. Vol., 485-627.

HayuoN, R.M. & Kesrurn, M. (1981): Hot springdeposits on the Easi Pacific Rise at 2lo: preliminarydescription of mineralogy and genesis. Earth Planet.Sci. Lett.53,363-381.

HExrNw, R. & Foueusr, Y. (1985): Volcanism andmetallogenesis of axial and off-axial structures onthe East Pacific Rise near l3oN. Ecoz. Geol. 84,22t-49.

HrNrLEy, T.K. & Tersutvtoto, M. (1987): Metals andisotopes in Juan de Fuca Ridge hydrothermal fluidsand their associated solid materials . J. Geophys. Res.92, tt4oo-tt{l0.

HomNonrz, J. (1981): The aging of oceanic crust at lowtemperature. .In The Sea 7 (C. Emiliani, ed.). JohnWiley & Sons, New York.

Alr, J.C. HoNNonsz-GuERsrElN, B.-M.,LavEnmn, C., MurrrnmacHs, K., Rurz, J. & Serrz-tr,ten, E. (1985): Stockworklike sulfide mineraliza-tion in young oceanic crust. In lfitid Reports DeepSea Drilling Project 83,263-282. U.S. Gov. Print-ing Office, Washingfon, D.C.

Hururgnn, S.E. & TnoursoN, G. (1978): Hydrother-mal alteration of basalt by seawater. Geochim. Cos-mochin. Acta 42, ln-145.

Iro, E., WHnr, W.M. & Goerl, C. (1987): The O, Sr,Nd, and Pb isotope geochemistry of MORB . Chem.Geol. 62, 157-176.

JottNsolt, H.P., KensmN, J.L., VrNs, F.J., SMtttt,G.C. & ScHonsarurc, G. (1982): A low level mag-netic survey over a massive sulphide orebody in theTroodos ophilite complex, Cyprus. Marine Technol.Soc. J.16,76.

Karrn, E.S. & Fnaltrlttt, J.M. (1988): The relation-ship between geological development ofridge crestsand sulfide deposits in the northeast Pacific Ocean.Econ. Geol. (in press).

KonNrc, J.B. (196?): The Salton-Mexicali geothermalprovince. Catifornia Div. Minq Geol., Mineral Inf.Semice 20.

Lelou, C., BmcHel, E., JeHnNNo, C. & PsREz-Lscr.AlRE, H. (1983): Hydrothermal manganeseoxide deposits from the Galapagos mounds, DSDPLeg70, hole 5MB and "Alvin" dives 729 and72l.Earth Planet, Sci. Lett. 63,63-75.

LANcMUrn, C.H., Brwoen, J. & Barze, R. (1986):Petrological and tectonic segmentation of the EastPacific Rise, 5o30'-14"30'N. Trans, Amer.Geophys. Union ff|, 422429.

LynoN, J.W. & Gellev, A. (1986): The chemical andmineralogical zonation of Mathiati alteration pipe,Cyprus, and genetic significance. .Ire Metallogeny ofBasic and Ultrabasic Rocks (M.J. Gallagher, R.A.Ixer, C.R. Neary & H.M. Prichard, eds.). Inst. Min-ing Metall., London, U.K., 49-68.

MecooNato, K.C., Srverne, J.C. & Fox, P.J. (1984):East Pacific Rise from Siquerros to Orozco Frac-ture Zones: Along-strike continuity of axial neovol-canic zone and structure and evolution of overlap-ping spreading centers. J. Geophys. Res. 89t6M9-6069.

-, & - (1986): Reply: The debateconcerning overlapping spreading centers and mid-ocean ridge processes. J. Geophys. Ras. 91'10501-105 I l.

MacDoucer,l, J.D. & Lucrvran, G.W. (1986): Sr andNd isotopes in basalts from the East Pacific Rise:significance for mantle heterogeteity. Earth Planet.Sci. Lett.77,273-284,

Melanorr', A., Etr,tsl-Bv, R.W., CnoNat, D.S. & Srtn-Row, R. (1983): The geological selling and chemis-try of hydrothermal sulfides and associated depositsfrom the Galapagos Rift at 86'W. Mar. Mining 4'123-137.

McCur-locH, M.T., GnncoRY, R.T., WASSERBURc,G.J. & Tevron, H.P., Jn. (1981): Sm-Nd, Rb-Sr-4 l867t6g isotopic systematics in an oceaniccrustal section: Evidence from the Samail ophiolite.J, Geophys. Res. 86' 2721-2735.

McGnncon, B.A. & RoNe, P.A. (1975): Crest of theMid-Atlantic Ridge at 26'N. "/. Geophys. Res. 8Q3307-3314.

Monror, R.L. & FRANTLIN, J.M. (1987): Two-foldclassification of Archean volcanic-associated mas-sive sulfide deposits. Econ. Geol. 82, 1057 -1063.

Morrl, M.J. (1983): Metabasalt, axial hot springs,and the structure of hydrothermal systems at mid-ocean ridges. Geol. Soc. Amer. Bull.94' 161-180.

EXTINCT HYDROTIIERMAL SYSTEM. CALAPACOS RIDGE 539

PIRKER, R.L. & Hrusrrs, S.P. (1974): The inversion ofmagnetic anomalies in the presence of topography.J. Geophys. Res. 79, 1587-1593.

PeRFrr, M.R. & FonNeRr, D.J. (1983): Geochemicalstudies of abyssal lavas recovered by DSRV ALVINfrom Eastern Galapagos rift, Inca transform, andEcuador rift, 2. Phase chemistry and crystallizationhistory. J. Geophys. i?es. 88, 10530-10550.

-, -, Melenoru, A. & Err,rnr,nv, R.W.(1983): Geochemical studies of abyssal lavas reco-vered by DSRV ALVIN from eastern Calapagos rift,Inca transform and Ecuador rift, 3. Trace elementabundances and petrogenesis. J. Geophys. Ras. 88010551-10572.

PHrr,urs, J.D., Wooosror, J. & BowrN, C.O. (1969):Magnetic and gravity anomalies in the Central RedSea. .trc Hot Brines and Recent Heavy Metal Depositsin the Red Sea (8.T. Degens & D.A. Ross eds.), 98-l13. Springer-Verlag, New York.

Pr-vusNrxa, I.I. (1984): A special state of low-temperature silica in the transition, gel>cristoba-lite >quartz. Dokl. Akad. Naak SSSR n9, M7+1478. (Eng. transl. 189-193).

Rnano, V., HErnneN, R., Fnaxcuerteu, J., Barlem,R.D. & BAcx-en, H. (1985): Submersible observa-tions at the axis of the ultra fast-spreading EastPacific Rise (17 " 30' to 21.30, S). Earth planet. Sci.Lett. 75,339-353.

RrcuaRDs, H.G., CarN, J.R. & JBNseNsrus, J. (1988):Mineralogical and metasomatic zonation of the alter-ation zone of Cyprus sulfide deposits. Econ. Geol.

-(in press).

Rtse Pnorecr Gnoup (1980): East Pacific Rise: hotsprings and geophysical experiments. Science 207,t42t-t433.

Rorrnrs, R.G. & RnanooN, E.J. (1978): Alteration andore-forming processes at Mattagami Lake Mine,Quebec. Can. J. Eqrth Sci.15, l-21.

RonnsoN, P.T., MnnoN, W.G., O'HnanN, T. &Scurvrncrr, H.-U. (1983): Volcanic glass compos!tions of the Troodos ophiolite, Cyprus. Geologltll,4004M.

RoNa,-P. A. (1978): Magnetic signatures of hydrother-mal alteration and volcanogenic mineral deposits inoceanic crirst. "I. Volcanot, Geothermal'Ris. 3.2r9-22s.

(1984): Hydrothermal mineralization at seafloorspreading centers. Earth Sci. Rev. Z0, l-l(/,,

Scryrlrycqr, H.-U., RarrmNscHLErN, M., RoarNsor.r,P.T. & Mnrecnn, J.M. (1983): Troodos extrusivesequence of Cyprus: A comparison with oceaniccrust. Geology ll, 405409.

SEvrruno, W.E. & JaNecrv, D.R. (1985): Heavy metaland sulfur transport during subcritical and super_

critical hydrothermal alteration of basalt; influenceof fluid pressure and basalt composition and crys-tallinity. Geochirn. Cosmochim. Acta 49, 25454l,@.

Sgrvezarr, H. & MacLreN, W.H. (1976): Anexperimental study on the partition of zinc and leadbetween the silicate and sulfide liquids. Mineral.Deposita ll, 125-132.

SrrtrrH, M.F. (1987): Geochemical Constraints onAlteration in the Galapagos Fossil HydrothermalSystem. M.Sc. thesis. Univ. of Florida, Gainesville,Florida.

PERrrr, M.R., JouassoN, I.R. & Eunmv,R.W. (1986): Geochemical constraints on thedevelopment of the Galapagos massive sulfidedeposit. EOS, Trans. Amer. Geophys. Union 91,I 185.

Spooxrn, E.T.C., BrcrrNsors, R.D., Fvre, W.S. &SMewwc, J.D. (1974): r8O-enriched ophiolite meta-basic rocks from East Liguria (Italy), Pendos(Greece) and Troodos (Cyprus). Contr. Mineral.Petrologlt 47,41-62.

-, Cnaruar, H.J. & SunwrNc, J,D. (1977):Strontium isotopic contamination and oxidationduring ocean floor hydrothermal metamorphism ofthe ophiolitic rocks of the Troodos Massif, Cyprus.Geochim. Cosmochin. Acta 4L, 873-890.

Sraxes, D.S. & O'Neu,l, J.R. (1982): Mineralogy andstable isotope geochemistry of hydrothermallyaltered oceanic rocks. Earth Planet. Sci. Lett.57.285-344.

-, Tevron, H.P. & Frsurn, R.L. (1983): Orygen-isotope and geochemical characterization ofhydrothermal alteration in ophiolite complexes andmodern oceanic cru$. .fn Ophiolites and OceanicLithosphere (I. c. Goss, S. J. Leppard & A. W.Shelton, eds.). Geol. Soc. London, Blackwell Sci.Publ., Oxford, 199-214.

Sruor, F.E. (1959): Maenetic survey of the Wairakeihydrothermal field. New Zealand J. Geol. Geophys.2,74G752.

Vrnel, PH. & Cleurn, N. (1981): Pb and Sr isotopicsystematics of some basalts and sulphides from theEast Pacific Rise at 21'N (project RITA). EarthPlanet. Sci. Lett. 55,237-46.

Vocr, P.R. & Dr Born, J. (1976): Morphology, mag-netic anomalies and basalt magnetization at the endsof the Galapagos high-amplitude zone. EorthPlanet. Sci. Lett. 33, 145-163.

WHrm, W.M., Honrr n, A.W. & Pucrrlt, H. (1987):Isotope geochemistry of Pacific mid-ocean ridgebasalt. /. Geophys. Res. V2,4881-4893.

Received August 8, 1987; revised manuscript acceptedMay 27, 1988.