sign in sign up
<<
Home , Caroline Island, Horizon Guyot, Line Islands

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 89, NO. B13, PAGES 11,261-11,272,DECEMBER 10, 1984

Geology and Geochronologyof the

S. O. SCHLANGER,1 M. O. GARCIA,B. H. KEATING,J. J. NAUGHTON, W. W. SAGER,2 J. g. HAGGERTY,3 AND J. g. PHILPOTTS4

Hawaii Institute of ,University of , Honolulu

R. A. DUNCAN

Schoolof Oceano•Iraphy,Ore,ion State University,Corvallis

Geologicaland geophysicalstudies along the entire length of the Line Islands were undertakenin order to test the hot spot model for the origin of this major linear island chain. Volcanic rocks were recoveredin 21 dredge hauls and fossiliferoussedimentary rocks were recoveredin 19 dredge hauls. Volcanic rocks from the Line Islands are alkalic and hawaiites. In addition, a tholeiitic and a phonolite have been recoveredfrom the central part of the Line chain. Microprobe analysesof groundmassaugite in the alkalic basaltsindicate that they contain high TiO2 (1.0-4.0 wt %) and A1203 (3.4-9.1 wt %) and are of alkaline to peralkaline affinities.Major element compositionsof the Line Islandsvolcanic rocks are very comparableto Hawaiian volcanicrocks. Trace elementand rare earth elementanalyses also indicatethat the rocksare typical of oceanicisland alkalic ' the Line Islands lavas are very much unlike typical mid-oceanridge or fracturezone basalts.Dating of theserocks by '•øAr-39Ar,K-Ar, and paleontologicalmethods, combined with DeepSea Drilling Projectdata from sites 165, 315, and 316 and previouslydated dredgedrocks, provide ages of volcanic eventsat 20 localities along the chain from 18øN to 9øS,a distanceof almost4000 km. All of thesedates define mid- to late Eoceneedifice or ridge-buildingvolcanic events. volcanic events took place from 15øN to 9øS, and Late Cretaceousevents took place from 18øN to 9øS. In the southern Line Islands both Cretaceousand Eoceneevents took place on the sameedifice or ridge, indicatingrecurrent volcanism at a singlelocality. The irregular distribution of in the chain, the fact that Late Cretaceousreefs flourishedalong a distanceof approximately2500 km in the central and southernLine Islands,and the observationthat spatiallyclosely related seamountsexhibit different subsidencehistories are interpreted as indicatingthat largesegments of the chainhave not followeda t •/2 relatedsubsidence path. Magnetic surveysof 11 seamountsshow that four of the ,from the central Line chain, give virtual geomagneticpoles which fall well to the north of virtual geomagneticpoles of Cretaceousseamounts. Thesefour polesagree with other paleomagneticdata of middle-lateEocene-early age from the Pacific.One of thesefour seamountsyielded a '•øAr-39Artotal fusionage of 39 Ma. Becausethe poles of all four seamountsfall into a tight group we infer that they are probably of middle-lateEocene age to early Oligoceneage. The other three seamountsas well as one seamontfrom the Line Islandsanalyzed previouslyall give virtual geomagneticpoles which agreewith Late Cretaceouspaleomagnetic data from the Pacific.Of thesefour seamounts,three give '•øAr-39Ar ages ranging from 71 to 85 Ma; another, in the southern Line Islands, is interpreted,on paleontologicalevidence, to be of Late Cretaceousage. The origin of the Line Islandshas beenascribed by previousworkers to the effectsof a singlehot spot and to the action of four hot spots.The singlehot spot model cannot account for all of the volcanicedifices in the Line Islands, although it does explain a general age progressionof 9.6 _+0.4 cm/yr from north to south along the chain derived from a number of dated edifices.The four hot spot model accountsfor more of the dated volcanicedifices but still doesnot explain all of the availabledata. The petrologic data argue against a mid-ocean ridge or transform origin proposed by earlier workers.The complexvolcanic province represented by the Line Islandsremains a challengeto existing models for the origin of midplate volcanism in the Pacific. An drilling program which could determine edifice building histories,paleolatitudes of formation, and subsidencerates and patternsin the Line Islandsis needed.

INTRODUCTION AND BACKGROUND and JohnstonIsland in the north to the northeastend The Line Islands geologicalprovince, capped by the atolls of the Tuamotu Islands in the south (Figure 1). The Line of Johnston, Palmyra, Washington, Fanning, Christmas, and Islands are a major bathymetric feature of the central Pacific , is made up of dozens of simple and com- comparable in size to the Hawaiian-Emperor chain to the north and the Marshall-Gilbert-Ellice chain to the west. posite seamountsand linear ridges that extend from Horizon Morgan [1972], arguing from the fact that the Line-Tuamotu geometric trends appear to parallel the Emperor-Hawaiian • Now at Departmentof GeologicalSciences, Northwestern Uni- trends,postulated that the Line and Tuamotu chainswere the versity,Evanston, Illinois. 2 Now at Departmentof Oceanography,Texas A & M University, temporal and genetic equivalentsof the Emperor and Ha- College Station. waiian chains, all four of these major featureshaving devel- 3 Now at Departmentof Geosciences,University of Tulsa, Tulsa, oped through the actions of two hot spots fixed relative to Oklahoma. each other for the past 70 Ma. Jarrard and Clague [1977] '• Now at U.S. GeologicalSurvey, National Center, Reston,Vir- pointed out that geometric and paleomagnetic evidence ginia. argued against the proposition that the Line and Emperor Copyright1984 by the AmericanGeophysical Union. chainswere entirely coeval. Paper number4B0781. At the time the above argumentswere put forward, reliable 0148-0227/84/004B-0781 $05.00 dates for volcanic activity over any significantportions of the 11,261 11,262 SCHLANGERET AL.' GEOLOGYAND GEOCHRONOLOGYOF THELINE ISLANDS

of midplate volcanismalong such crosstrends were influenced by older Line Islands structures. He implied that several chainsof volcanoson crosstrends were formed by individual hot spots that crossedthe main Line chain. Both DSDP drill- / ing in the central Line Islandsand the dating of rocksdredged from seamounts in the cross trend showed that both Late Cretaceous and Eocene volcanism had occurred there [Lan- phere and Dalryrnple, 1976; Saito and Ozirna, 1977]. Further, paleontological studies of rocks dredged from seamountsin the southern Line Islands [Haggetty et al., 1982] showedthat both Late Cretaceousand Eocenevolcanism had taken place in the vicinity of (Figure 1, RD-44 and RD- ß 45). The wide geographicdistribution of both Cretaceousand O Eocene volcanism over a distance of several thousands of ki-

/.,• ' ! lometersalong the Line chain is regardedby us as establishing Kingman I the fact that the Line Islands were not formed by the action of Palmyra / a single hot spot. That the southern Line Islands as well as the central Line Islandswere formedby a complexoverprinting of •0o H•a*shingtonIs. • ZOo volcanic events was suggestedby Crough and darrard [1981]. They showed that a depth and geoid anomaly extends from IS, o o RD-35 the Marquesas Islands into the southern Line Islands and RD-39 o ascribedthis anomaly to the trace of a hot spot which transi- PCOD-6 -- Christmas Is. ted the Line Islands in Eocene time. Vivid evidence that the Line chain has a complexhistory of volcanismas comparedto JarvisIs. eRD_4 o that of the Emperor chain is containedin W. Haxby's geotec- C'/•ø•e•ton.. •o tonic imagery map of the Pacific Basin [see Francheteau, e 1983], a sketch map of which is shown here as Figure 2. In addition to the marked Line cross trend that cuts across the central Line Islands at 10øN a number of other linear ridges intersect the northern and central Line Islands; the southern- most of these extendstoward the Marquesas Islands and lies along the depth and geoid anomaly of Crough and darrard [1981]. The en echelon appearance of much of the central Line Islands could be due to the presence of these short NW-SE trending features which intersect the main Line Is- lands trend. Bathymetrically then, the Line chain bears little resemblanceto the relativelysimple Emperor chain. Hendersonand Gordon [1982] and Duncan [1983] have ad- dressedthe complexity of the Line Islands and argue, based on plate motion reconstructionsover the past 150 Ma, that several hot spots are needed to model the evolution of the Line Islands. If all linear volcanic features or chains of vol-

Fig. 1. Chart of the Line Islands showingthe track of R/V Kana Keoki (dashedline) through the area. Dredge stations are indicated by RD numbers; piston core stations by PCOD numbers; seamount magnetizationsurveys by L numbers.

LineIslands •-• /•"' Cross-Trend...,• G•ppe•tø• Line Islands were not available. Driling along the Line Islands on Deep Sea Drilling Project (DSDP) legs 17 and 33 (see Winterer et al. [1973] and Schlangeret al. [1976] for details) ' Main Line x showedthat volcanic episodesin the northern Line Islands at "•, o•Island Chain • site 171 on Horizon Guyot and at sites 165, 315, and 316 in the central Line Islands were more temporally complex than could be accountedfor by using the Emperor chain as a model for the development of the Line Islands. Winterer [1976] in an extensive review of the data relevant to the for- Fig. 2. Sketchmap of major featuresin the Line Islandsprovince mation of the Line Islands emphasizedthe importance of the (unlabeledlinear trends are after the geotectonicimagery map of W. Line crosstrend (Figure 2) and argued that repeatedepisodes Haxby from Francheteau[1983]). SCHLANGERET AL.' GEOLOGYAND GEOCHRONOLOGYOF THE LINE ISLANDS 11,263

canic edificesoriginate solely by hot spot activity of the crosstrend is the relic of a central Pacific rift systemthat Hawaiian-Emperorstyle, then a number of hot spotsmust underwent limited crustal extension. have contributed to the formation of the Line Islands. Other Strontium isotopic ratios have been determinedon some of mechanismsfor the linear localization of volcanic activity these rocks using both untreated and HCl-leached material. have been put forward.Farrat and Dixon [1981] postulated Unfortunately, most of these samplesare at least somewhat the presenceof a long transformfracture zone that extended altered,and the valuesreported have very large uncertainties from the Emperor Trough through the Gardner Seamounts (e.g.,0.7030 _+0.0009). Thereforeit is difficult to interpret the and the Line Islands.Orwig and Kroenke[1980] also pro- valueswith any confidence.Nevertheless, the reportedinitial posed a origin for the Line Islands. Natland 87Sr-SaSrratios for basaltsvary from 0.7028 to 0.7039(all but [1976] consideredthat the petrologyof rocksfrom the central one between0.7035 and 0.7039),virtually overlappingthe re- Line Islands indicatedtheir similarity to rift associatedvol- ported range for Hawaiian volcanic rocks (0.7029-0.7041 canismof the African type. Winterer [1976] also invokedfis- [Lanphereet al., 1980]) and generallyhigher than typical mid- sures and fractures in the Line Islands to account for some of oceanridge basalts(MORB) ratios (0.7026-0.7035[Schilling et the volcanicactivity there. at., 1983]). It is the purposeof this paper to presentthe geological, Previouspetrologic studies have all been on samplesrecov- geochemical,and palcomagneticdata that will outline the eredfrom the northernhalf of the Line chain; no sampleshad constraintson any modelsproposed for the evolution of the been recovered from the chain south of DSDP site 315 near Line Islands. Fanning Island at about 4øN. During the R/V Kana Keoki 1979 cruises,samples were dredged from along the entire CRUISE OPERATIONS lengthof the chain (Figure 1). The rocks recoveredfrom dredge stations and those pre- The geologicand geophysicalstudies of the Line Islands viouslyreported from the Line Islands [Basset al., 1973;Jack- discussedhere were undertakenduring three cruises(of R/V sonet al., 1976; Natland, 1976] are almost entirely moderately Kana Keoki) to the area in the fall of 1979 (Figure 1). The to severelyaltered alkalic basalts and hawaiites. In addition, objectivesof these cruiseswere to dredge edifice volcanic tholeiitic basalts and a phonolite have been recoveredfrom rocks and cappinglimestone, take piston coresin submarine the central part of the chain [Natland, 1976]. The alkalic ba- valleysincised into the sedimentaryapron of the central Line salts contain phenocrystsof , (replacedby Islands province, conduct seamount magnetization studies, iddingsite),and, rarely, augitc. The hawaiites are generally and to collect magnetic,gravity, and bathymetric data rele- aphyric, although some contain rare phenocrystsof plagio- vant to the problem of the evolution of the Line Islands.Our claseand/or brown hornblende.The of both rock types bathymetricmapping showedthat some existingbathymetry consistpredominantly of plagioclase,titaniferous augitc, titan- was incorrectand misleading.Among our discoverieswere an iferousmagnetite, and, locally, clays replacingglass and oli- unmappedseamount at 1lø40'N, 160ø40'W,a ridge whichrises vine. Mineralogically,these rocks are typical of Hawaiian to 2700 m in depth connectinga seamountat 6ø20'N, 158øW lavas. with a ridge at 6ø20'N, 159ø30'W;and a ridge nearly 300 km Microprobe analysesof groundmassaugitc in the Line Is- long consistingof eight peaks ranging from 3000 to 2150 m lands lavas indicate that they contain high TiO2 (1.0-4.0 wt depth connectingCaroline Island at 10øS, 150ø20'W with a %) and A1203(2.0-9.1 wt %) [Jacksonet al., 1976]. A plot of seamountat 7ø35'S,151ø35'W. In addition, Carol Guyot (9øN, SiO: versusAIeO3 showsthe Line Islands lavas to be of alka- 158øW)and Vostok Island (10øS,152øYW) were found to be line to peralkaline affinitiesbased upon the classificationof incorrectlycontoured. LeBas [1962] (Figure 3), although two samples,RD-33-4 and A total of 21 dredgesrecovered volcanic rocks. Sedimentary RD-43-1 from the central Line Islands are of the nonalkaline rock sampleswere recoveredin 19 dredges.In addition, sedi- type. ments(and in one casebasalt pebbles) were recoveredfrom six The freshestsamples from eachdredge haul were chemically piston cores.Underway gravity and magneticswere recorded analyzed(Table 1). All of the samplesare moderatelyaltered. over the entire lengthof the Line chain and over the margin of Some sampleshave high Fe:O3/FeO ratios and P:O5 and the Tuamotu Ridge. Eleven seamountmagnetization surveys H20 contents.Nevertheless, the least alteredsamples are simi- were conducted. lar chemically to Hawaiian alkalic lavas. They plot in the Hawaiian alkalic field and span the range from alkalic PETROLOGY olivine basalt to mugearite using the classificationof Coombs Previous petrologicwork on volcanicrocks from the Line and Wilkinson [1969]. Islands is almost entirely reported on in the DSDP initial Rare earth elements(REE) were determinedby isotopedilu- reports of legs 17 and 33 [Winterer et al., 1973; Schlangeret tion massspectrometry on selectedsamples. One of the sam- al., 1976]. At DSDP site 315, six flow units of transitional ples (RD-60-6) has the REE pattern of a typical Hawaiian basalt (intermediate in composition between Hawaiian tho- tholeiite (Figure 4). The other samplesshown have REE con- leiitic and alkalic lavas) were recovered[Jackson et al., 1976]. centrations similar to those of Hawaiian alkalic lavas. Hawaiites,some containing brown amphibole,were recovered Jacksonet al. [1976], in addressingthe suggestionthat the at DSDP site 165 [Bass et al., 1973]. A wide variety of rocks Line Islands are an abandoned ridge crest, argue that the dredged from the northern and central portions of the Line chemistryof the basaltsdrilled at DSDP sites165 and 315 are Islandsinclude tholeiites, alkalic basalts,hawaiites, phonolites, inconsistentwith a ridge crestorigin. In their review of analy- and potassicnephelinites. Natland [1976] consideredthe po- sesin hand following leg 33 [Jacksonet al., 1976] they came tassic nephelinitesdredged at four locations in the Line Is- to the conclusionthat the Line Islandsare made up of basaltic lands area as similar in compositionand origin to potassic rocksthat are similar to basaltsof the Hawaiian chain. Analy- mafic lavas from African rift zones.He argued that the pres- ses reported here support the argument that the main Line ence of such rocks in the Line cross trend indicated that the Islandschain was built of effusionsof typical alkalic and tho- 11,264 SCHLANGERET AL.: GEOLOGYAND GEOCHRONOLOGYOF THE LINE ISLANDS

x 52

51

NONAL 50- • :"•iiii!•',:,•-;ii•ii!!i-'.iii!i":'"'•••-:ii•i•!ii•!iiii•ii•:."i::'7:•.:-;:•il;•:.':•i•'•iiiiii!i:i!iii!i!i•:.. :.. '" :::?•i!i!iii::::. 59-1361-3 • • .... ,.::::..:...... :.:.-?•!,;•,;:.:?-, A .o ] .0_ 47 ALKALINE lO ,-,,. 46- E, PERALKALINE K La Ce Nd Sm Eu Gd Dy Er Yb Lu 45' Fig. 4. Chondrite-normalized rare earth element patterns for some representativelavas from the Line Islands. Fields for Hawaiian 44 ' ] [ I ] I ] I ' I • I ' I ' I ' I ' 0 I 2 3 4 5 6 7 8 9 tholeiitesand alkalics and mid-ocean ridge basalt (MORB) from Ba- saltic VolcanismStudy Project [1981]. ' AI203 Wt.% Fig. 3. SiO2 and AI203 wt % in groundmassclinopyroxene [Winterer et al., 1973; Schlangeret al., 1976]. At DSDP site grains from lavas from the Line chain. Fields for nonalkaline, alka- line, and peralkaline rocks are from LeBas [1962]. A, RD-41-1; B, 165 a complex and long history of Cretaceousvolcanism was RD-41-2; C, RD-44-2; D, RD-54-2; E, RD-59-1, H, RD-59-9; I, RD- revealed [Lanphere and Dalrymple, 1976; Winterer et al., 59-10; J, RD-59-15; K, RD-63-7; X, RD-52-7. 1973]. Three basalt units interbedded with fossiliferoussedi- ments were cored at the site. Between the two deepestbasalt leiitic Hawaiian type lavas and that there is no compelling flows of the Eiffelithus eximus zone were recovered. evidencethat a ridge crest origin is warranted for the Line Taking the base of this zone to be at the - Islands. boundary dated at 83 Ma by Harland et al. [1982], we can deduce that volcanism was active at least 83 GEOCHRONOLOGY Ma. The age of the deepestbasalt, based on sedimentation The basic data set constrainingdiscussions of the origin of rate arguments could be as old as 93-101 Ma [Lanphere and linear island chains consistsof determinations of the ages of Dalrymple, 1976]. Higher in the section at site 165 are thick volcanic edifice formation along the chain. The data used in volcanic breccia and beds which indicate that late- this paper consistof datesderived from interpretationof vol- stagevolcanism persisted into the Tetralithus trifiduszone of canicand sedimentarysequences drilled on DSDP legs 17 and late Campanian age dated at 74 Ma by Harland et al. [1982]. 33, edifice ages determined by dating of assemblages A rock dredgedfrom the large seamountjust to the east of site dredgedfrom seamounts,and radiometricages determined by 165 was dated at 72 Ma usingthe 4øAr-39Armethod [Saito 4øAr-39Arand K-Ar methodson dredgedrocks collectedby and Ozima, 1977]. Thus major volcanism in the immediate the Hawaii Institute of Geophysics(RD and PC prefix) and vicinity of site 165 occurred over a time span of at least l l the ScrippsInstitution of Oceanography(D suffix)(Figures 5 Ma, from 72 to 83 Ma and possibly longer. At site 315 and 6 and Tables 2 and 3). Becauseage determinationslie at [Schlangeret al., 1976] the east flank of the Fanning Island the heart of any discussionon the origin of a chain suchas the edifice was drilled. The basalt cored at the bottom of the hole Line Islands, the data used are discussedbelow in the geologi- 180 øW 160 140 cal context of the various locationsshown on Figures 1 and 5. Three DSDP siteswere drilled along the central Line chain in an attempt to reconstructthe history of edifice building •id_-Pacificfox 40Ar-39Arß This paper Ages 20 øN TABLE 1. Major Element Analysesof Volcanic Rocks From the :L/-•Mrl/••••t 8(143D) 0$aito&Ozima(1977) Line Islands Chain 2 "••_•4 •)•6(RD63)• •81 83(RD61) Fossilf(>78)Ages 56 0 , Sample (RDS9) 083 (133D) •• x 79•128D) 63-7 61-1 60-6 59-13 41-2 43-1 44-3 45-1 52-2 _& • •39•B•33)

SiO2 41.70 45.20 46.45 42.90 45.20 42.75 41.75 38.65 44.55 TiO 2 3.48 3.88 2.02 2.78 2.88 2.60 3.22 2.34 2.93 AI203 13.53 17.02 16.20 17.64 13.74 17.97 15.80 15.72 15.87 Fe20 3 9.19 9.48 8.44 5.81 11.12 13.27 13.68 11.36 9.70 (pc2•o • (RD4') FeO 3.62 2.60 2.56 5.56 1.96 0.97 0.92 0.48 3.04 DS•P31• •9(RD43) . MnO 0.14 0.09 0.09 0.18 0.18 0.13 0.12 0.21 0.15 f(>78) LINE ISLANDS 2.20 4.57 4.57 6.85 0.97 1.92 1.13 4.05 MgO 5.48 o CaO 11.85 8.66 10.30 8.11 12.55 8.58 10.19 14.76 11.53 Na20 1.97 3.22 2.95 2.86 2.58 3.34 3.02 2.83 2.60 ø(RO44)172 Marque 2.46 1.03 1.94 0.91 1.75 1.29 1.71 0.92 o /•._•_1 -10 K20 1.21 f(44-50) (RD45) 1.98 0.58 1.88 0.37 2.12 2.81 5.87 1.48 P205 0.93 0 1000 km 0 DSDP 318 f(~52) H 20 + 3.29 1.95 2.43 3.60 0.62 3.39 2.54 2.51 1.50 H20- 2.48 1.26 1.92 1.93 0.74 1.90 2.16 1.37 1.27 CO 2 1.05 0.02 0.08 0.17 0.16 0.29 0.37 0.83 0.24 Total 99.92 100.02 99.62 99.93 99.78 99.85 99.79 99.77 99.83 I I

Analyst' K. Ramlal, Universityof Manitoba. Fig. 5. Stationlocations and agedeterminations in the Line chain. $CHLANGERET AL.' GEOLOGYAND GEOCHRONOLOGYOF THE LINE ISLANDS 11,265

Mid-Pacific Mtns 40Ar-39Ar Ages ß This paper 85 0 8aito&Ozima( 1976,1977)

45 motus al

4500

85 K-Ar Ages

,TH

ß o o H•'... 45 o ß o '.... ß bl ß ß 4500 a•o 27•o ,•oo 9•)0 b Distanc'e from Line-Tuamotu Bend km Fig. 6. Radiometricand paleontologicages in the Line chain.(a) '•øAr-39Arages shown as solidcircles were deter- mined in this study; open circlesare datesfrom Saito and Ozima [1976, 1977]. The solid line showsa rate of migration of volcanismof 9.6 + 0.4 cm/yr which would accountfor many of the agesdetermined. (b) K-Ar agesand biostratigraphic age determinations(DSDP sites165, 315, and 316 and seamounts(H) from Haggetty et al. [1982]).

yielded a feldspar fraction K-Ar age of 91.2 + 2.7 Ma [Lan- [Schlangeret al., 1976]. In summary the DSDP resultsfrom phere and Dalrymple, 1976]. Their age convertsto 93.3 Ma the main Line Islandssystem show the following' upon usingnew constants[Harland et al., 1982], the age used 1. Volcanism at site 165 and the nearby seamount from in this paper. Nannofossilsof the Marthasteritesfurcatus zone which dredgehaul 130-D was taken occurredfrom 83 Ma (or were recovered 85 m above the basalt. These sediments are of possiblyealier) to 72 Ma. age, the M. furcatus zone being dated at 88 Ma 2. At site 315, major edifice constructionceased by ,--93 [Hadand et al., 1982]. The correlationof the fossilage and the Ma; initiation of volcanism is undated but could be prior to K-Ar age at site 315 arguesfor a cessationof major edifice ,,- 95 Ma. building activity at ,,-93 Ma. Further, Premoli-Silvaand Brusa 3. At DSDP site 316 the cessation of volcanic edifice [1981] record the presence of a mid-Cretaceous Cuneolina building is estimatedat 81-83 Ma; initiation of volcanismis assemblagein turbiditesat this site,implying the existenceof a undated. seamount in the photic zone in - time, In our study of rocks dredgedfrom the Line Islands, both >90 Ma. K-Ar and '•øAr-39Ardating techniqueswere employedto At site 316, basalt basement was not reached. The oldest examinethe distributionof ageswithin this volcaniclineament sedimentspenetrated yielded foraminiferaand nannofossilsof (Figures 5 and 6 and Tables 2 and 3). Previous attempts at early Campanian age (77-80 Ma) associatedwith basaltic measuring reliable crystallization ages by the conventional breccias and volcanic . On the basis of the corre- K-Ar method [Davis et al., 1980] have been frustratedby the lation of stratigraphicsections between sites 316, 315, and 165, effects of seawater alteration. Both K addition and Ar loss basalt basementat 316 probably lies ,,-75 m below the deep- during low-temperature and formation, in the est sedimentsrecovered; the data of cessationof major edifice- presenceof seawater,will cause the measuredage to be sys- building volcanism at 316 can be estimated at ,,- 81-83 Ma tematicallylower than the rock crystallizationage. Reconnais- 11,266 SCHLANGER ET AL.' GEOLOGY AND GEOCHRONOLOGY OF THE LINE ISLANDS

TABLE 2. The 4øAr-a9ArTotal FusionAge Data on DredgedVolcanic Rocks From the Line Islands Chain

Percent Radiogenic Age + 1• Sample Location '•øAr-a6Ar '•øAr-a9Ar 37Ar-'•øAr* ½øAr x 106years

143D-102 19ø30'N, 169ø0YW 3011. 32.03 0.0538 90.0 88.1 + 0.4 142D-11 18ø00'N, 169ø05'W 57028. 38.16 0.0092 99.4 93.4 + 1.3 RD63-7 16ø27'N, 168ø1YW 1456. 35.36 0.1525 79.4 86.0 + 0.9 RD59-12 12ø31'N, 167ø0YW 2463. 71.24 0.0106 88.0 85.0 + 1.1 RD61-1 14ø59'N, 166ø27'W 965.9 86.32 0.0149 69.4 81.4 + 1.1 RD61-5 14ø59'N, 166ø27'W 835.3 79.74 0.0221 65.1 82.6 +_0.7 128D-11 9ø15'N, 160ø45'W 642.1 59.24 0.0089 53.8 78.7 + 1.3 RD33-1 8ø11'N, 161ø55'W 379.8 128.84 0.0371 22.2 39.3 + 1.5 123D-15 5ø50'N, 160ø45'W 2251. 28.74 0.1058 76.4 76.4 + 0.5 PCOD-6-2 2ø35'N, 158ø30'W 2524. 58.05 0.0243 88.2 69.8 + 1.0 RD41-1 2ø06'N, 157ø21'W 705.8 23.57 0.1929 57.7 35.5 + 0.9 RD43-1 0ø42'S, 155ø17'W 702.4 33.15 0.0956 57.6 59.0 + 0.8 RD44-3 7ø35'S, 151ø3YW 461.8 65.24 0.0445 35.9 71.9 + 1.4 RD45-26D 9ø04'S, 150ø42'W 633.7 53.57 0.1238 53.1 70.5 + 1.1 RD52-1 15ø01'S, 149ø02'W 544.1 75.62 0.0869 45.6 47.4 + 0.9 RD52-2 15ø01'S, 149ø02'W 472.7 81.12 0.0914 37.4 41.8 + 0.9

*Ar 37 correctedfor decay sinceirradiation.

sanceK-Ar agesdetermined in this study(Table 3) were simi- erally fresh,particularly in the more trachytic-texturedrocks larly scatteredand significantlylower than indirect age esti- (143D-102, 142D-11, RD-59-12, and PROD6-2). Olivine mates basedon paleontologicdata [Haggerry et al., 1982] or phenocrystsare commonly altered to iddingsite.The matrix seamountmagnetization studies [Sager, 1983a]. phasessuch as feldsparand clinopyroxeneare generallyfresh, Recently,the applicationof the '•øAr-39Artotal fusion but where devitrified glass occurred, it has been altered to method to dating moderately altered submarine volcanic clays. and are common in vesiclesand along rocks has proven successfulin determiningconsistent ages microfractures.A further indication of the degreeof alteration [Dalrympleand Clague,1976; Duncan,1978, 1982; Dalrymple of these samplesis the high-H20 content in most samples et al., 1981]. According to these studiesit is probable that (between0.6 and 3.6 wt %) and in rare cases,high-CO2 con- 39Ar, derivedfrom K-bearing alterationminerals or from K tent (up to 1 wt %). residingon grain boundaries,escapes from the sampleduring Argon isotopic compositionswere measuredusing an AEI neutronirradiation or extractionline bake-out.Radiogenic MS-10Smass spectrometer equipped with 3BArspike and air '•øAr was lost from preciselythese sites during alteration,so calibrationpipette systems. For 4øAr-39Arages, split aliquots the age signalfrom the low-temperaturealteration phases is of crushed sampleswere irridiated at the U.S. Geological essentiallyerased prior to samplefusion. Thus the 4øAr-39Ar SurveyTRIGA Reactor in Denver, Colorado, for 4 hours at 1 age commonlyapproaches the crystallizationage becausethe MW power. Operating conditions and isotope interference predominantsource of nonatmosphericargon is the primary correctionshave beendescribed by Dalrympleet al. [1981]. crystallinephases. That conventionalK-Ar analysesunderestimate the agesof Sampleswere selectedfrom the Line Islandscollections for Line Islands volcanic rocks is particularly apparent in the K-Ar and '•øAr-39Argeochronology on the basisof thin sec- scatteredages of samplesfrom the same seamount(RD-45-1, tion examination.Petrographically, all samplesshow slight to RD-45-26D) or neighboringseamounts (Table 3 and Figure 6). moderatealteration. Augite and feldsparphenocrysts are gen- Where indirect age estimatesare available,K-Ar age determi-

TABLE 3. K-Ar Age determinationsFrom Dredged VolcanicRocks From the Line IslandsChain

Percent Percent Radiogenic'•øAr Radiogenic Age* + 1• Sample Location K x 10-s cma/g Ar x 106 years

143D-102 19ø30'N, 169ø0YW 2.097 0.6128 88.8 73.7 + 0.8 RD63-7 16ø27'N, 168ø1YW 1.194 0.2589 78.6 55.0 + 0.6 RD61-1 14ø59'N, 166ø27'W 2.098 0.4840 67.3 59.8 + 0.6 133D-9 12ø04'N, 165ø50'W 1.666 0.4808 65.8 72.8 + 1.3 128D-11 9ø15'N, 160ø45'W 0.725 0.1370 56.6 48.0 + 0.6 RD33-1 8ø11'N, 161ø55'W 0.695 0.0625 44.8 23.5 + 0.3 123D-15 5ø50'N, 160ø45'W 1.996 0.3478 86.6 44.3 + 0.5 PCOD-6-2 2ø35'N, 158ø30'W 1.601 0.3783 90.2 61.2 + 0.6

RD41-1 2ø06'N, 157ø21'W 1.016 0.1005,, 75.6 25.3 + 0.3 RD43-1 0ø42'S, 155ø17'W 1.411 0.2047 53.8 37.8 + 0.4 RD44-3 7ø35'S, 151ø3YW 1.120 0.1780 32.8 41.4 + 0.5 RD45-1 9ø04'S, 150ø42'W 1.301 0.2259 34.4 45.2 + 0.6 RD45-26D 9ø04'S, 150ø42'W 0.915 0.2133 58.4 59.0 + 0.7 RD52-2 15ø01'S,149ø02'W 0.742 0.0712 35.9 25.1 + 0.4

*Agescalculated from the following decay and abundance constants' 2• = 0.581x 10-•ø yr-" 2• = 4.962 x 10-'ø yr-" '•øK/K = 1.167x 10-'• mol/mol. SCHLANGERET AL.' GEOLOGYAND GEOCHRONOLOGYOF THE LINE ISLANDS 11,267

nations are always younger,by 9-30 Ma, and are not reliable studied,but a usefulresult was obtainedfrom only one, L3 in indicatorsof samplecrystallization age, despite generally high Figure 1 [Harrison et al., 1975]. The magnetizationparame- proportionsof radiogenicargon. ters and virtual geomagneticpoles (VGP's) derived from the The '•øAr-39Artotal fusionages are muchmore consistent, eight seamountmagnetic models that yieldedapparently reli- however(Table 2 and Figure 6). Where multiple samplesfrom ableresults are givenalong with '•øAr-S9Arages determined the sameseamount were analyzed (RD-61-1, RD-61-5, RD-52- for them in Table 4. 1, and RD-52-2), age agreement was good. Samples from The goodnessof fit ratio (GFR) [Richards et al., 1967], neighboring seamounts(143D-102, 142D-11, and RD-63-7; measuring the match between the observed and calculated RD-44-3 and RD45-26D) gave similar ages.Further, the con- anomalies,is high for all of the seamountsanalyzed in this cordance of our radiometric dates with age determinations study,ranging from 3.1 to 5.2. The larger the GFR the better based on paleontologicaland seamountmagnetization data the agreement between the observed and calculated anoma- indicatethat the '•øAr-39Arages are reliable.Haggetty et al. lies; usually a GFR of 1.8-2.0 is consideredto be the mini- [1982] determineda minimum age of 70-75 Ma, basedon the mum acceptablevalue for a reliable magneticinversion [Har- occurrence of Late Cretaceous shallow water fossils, for the risonet al., 1975; Sager,1983a]. Not only do the GFR's of the seamountfrom which RD-45 was taken; as shown on Table 2, Line seamountsindicate excellent results, but the paleo- RD-45-26D yieldeda '•øAr-39Arage of 70.5+ 1.1 Ma. Sea- magnetic poles derived from these seamountsare consistent mount L8 has a virtual geomagneticpole which lies closeto a among themselvesand with other Pacific seamount VGP's. Maestrichtian pole of both Gordon [1982a] and Sager Moreover, the VGP's of the reliably dated seamountsfrom the [1982a];RD-44-3 from this seamount yielded a '•øAr-S9Arage chain that have been studiedpaleomagnetically are in close of 71.9 + 1.4 Ma, within the 65-73 Ma span of the Maestrich- agreementwith other Pacific paleomagneticdata of the same tian stageaccording to Harland et al. [1982]. The majority of age,as is discussedbelow. sampleages falls along a linear array, from older seamountsin Four of the Line seamountsgave VGP's closerto the geo- the north (-,•93 Ma) to youngerseamounts in the south(-,•44 graphicpole than any VGP derivedfrom a CretaceousPacific Ma at the Line-Tuamotu bend). The overall consistencyof seamount(Figure 7) implying that thesefour seamountsare this data set encouragesconfidence in theseestimates of vol- younger than Cretaceousage. These four VGP's are located cano ages. It is difficult to imagine, for instance,a process closeto the VGP's of three Pacific seamountsthat probably which would produce progressivelygreater alteration to the formed during the late Eocene or early Oligocene.Of the south to yield the observedpattern. A rate of migration of three,one is inferredto be approximately41-42 Ma of age by volcanismof 9.6 + 0.4 cm/yr bestfits the linear array of sea- its positionin the Hawaiianseamount chain and by its mag- mount ages. A number of ages,however, do not fall on the netic polarity [Sager, 1984]. The other tWO seamountsare main linear trend, these are also considered reliable and are located in the eastern Pacific on seafloor 45 and 37 Ma old as discussed below. indicated by magnetic lineations [Sager, 1983b]. In addition Earlierstudies have reported '•øAr-39Ar age determinations to this data, a DSDP sedimentpa!eolatitude (site 166) calcu- from a number of seamounts in the northern and central Line lated from samplesspanning the late Eoceneand early Oligo- Islands[Saito and Ozirna,1976, 1977]. Where both '•øAr-39Ar cene records a paleopole locus in agreement.with the four total fusion and incrementalheating ageswere reported,we Line IslandsVGP's as doesa paleoequatortransit determined have plotted the total fusionages considered to be reliableby from sediment facies examined at DSDP Site 163 [Sager, Saito and Ozima (shownas open circles)for comparisonwith 1983b]. the agesdetermined in the presentstudy (shown as solidci r- Of the four seamountsthat give high-latitude VGP's, only des) in Figure 6. Two seamounts,130D and 133D, yield Cre- one has been dredgedand dated. SeamountL4 (RD-33) has taceousages which fall along the main linear trend. Another is an 4øAr-39Artotal fusion age of 39.3-+_ 1.5 Ma. SeamountL5 Eocenein age. Saito and Ozirna [1977] reported an isochron is directly adjacent (and possiblyconnected •to) to seamount age of 127.5 + 5.0 Ma and a total fusion age of 114.8 + 1.3 L4 [Sager and Keating, this issue] and is thus likely to be the Ma for 142D dredged from the northern end of the Line sameage, particularly consideringthe nearnessof their VGP's Islands.We determinedan age of 93.4 + 1.3 Ma for this site. (Figure 7). Neither L6 nor L7 have been dated, but the prox- The reasonfor this discrepancyis unknown. Other agesfrom imity of their VGP's to the other Eocene-Oligocenepaleo- neighboring seamountssupport our younger age determi- magneticpoles coupledwith the fact that a significantnumber nations, however. of Eocene ages have been determinedfor seamountsin the Line Islands(Table 3) [Saito and Ozima, 1977; Haggerry et al., SEAMOUNT PALEOMAGNETISM 1982] suggeststhat these two seamountsare also likely to be Measurement and analysisof the magnetic field of a sea- the sameage. mount can yield information on the paleoazimuth and lati- The other four Line Islands seamounts studied paleo- tudinal transport of a seamountsubsequent to its formation magnetically produced VGP's that are located among the and on the location of its virtual geomagneticpole (VGP); the VGP's of Pacific seamountsof Cretaceousage. Both L1 and location of its VGP allows inferences to be drawn as to the L8 have VGP's locatednear the Maastrichtianpaleomagnetic age of the seamount. In this paper the implications of the polescalculated by Gordon[1982a] and Sager [1983a] (Figure paleomagneticresults for periodsof volcanismalong the Line 8). L1 is undated,but a rock from L8 yieldedan '•øAr-39Ar chain are stressed.The details of the surveys and data re- total fusion age of 71.9 _+1.4 Ma (RD-44) which agreeswell duction methods are describedby Sager and Keating [this with the age inferredfor the volcanofrom fossils[Haggerry et issue]and Sager [1983a]. al., 1982] and from its reversedpolarity [Sager and Keating, To these ends, 11 seamounts in and around the Line chain this issue]. were surveyedfor paleomagneticanalysis; of these,seven gave Seamounts L2 and L3 are located close to one another in apparently reliable results.Previous to our work, eight sea- the Line chain(Figure 7) and have4øAr-39Ar total fusionages mounts in and around the chain had been paleomagnetically that are virtually identical, 85.0 _+1.1 Ma (L2, RD-59) and 11,268 SCHLANGERlET AL.' GEOLOGYAND GEOCHRONOLOGYOF THE LINE ISLANDS

TABLE 4. Magnetization Parameters for Line Islands Seamounts

Location VGP

Latitude Longitude Latitude Longitude Inclination Declination Intensity, Age, Name ID N E N E ( + Down) ( + East) A/m GFR m.y. Reference

Watkins L1 17.5 190.8 68.8 33.5 -4.3 352.0 5.8 5.2 1 Nagata L2 12.5 193.0 61.6 4.0 - 29.0 4.4 3.8 3.7 85.0 + 1.1 2 Kapsitotwa L3 12.0 194.2 47.5 333.5 -36.6 28.0 5.1 3.7 82.7+ 3.6 3 Stanley L4 8.2 198.1 75.6 356.5 - 10.5 5.3 3.2 4.7 39.3 + 1.5 4 Willoughby L5 7.9 198.1 78.2 14.6 - 7.8 0.7 3.5 4.4 4 Chapman L6 3.4 199.9 75.7 37.8 - 19.7 355.6 4.7/8.0 4.3 4 Clarke L7 - 3.3 206.0 80.0 20.4 - 25.3 1.0 3.3 4.0 4 Uyeda L8 - 7.5 208.5 68.5 345.6* 40.1 195.7 6.9 3.1 71.9 + 1.4 4

GFR, goodnessof fit ratio. References'1, Keating and Sager [1980]' 2, Sager et al. [1982]' 3, Harrison et al. [1975]' 4, Sager and Keating [this issue]. *Reverselypolarized, south pole given.

82.7 + 3.6 Ma (L3, 133D) [Saito and Ozirna, 1977], yet their age of two samplescould result in a larger separationof their VGP's are separatedby over 22ø. This phenomenonwas puz- paleopoles.Consequently, we hypothesizethat L2 is slightly zling until it was noted that apparent polar wanderjust prior youngerthan L3, which is allowedwithin the analyticaluncer- to 80 Ma was very rapid (Figure 8) [Gordon, 1983; Sager, taintyof their 'tøAr-S9Artotal fusionages. 1983a], so that a differenceof only a few million years in the VOLCANIC EVENTS The above describedDSDP results, radiometric ages,bio- stratigraphic ages, and paleomagneticresults determined for + NP seamounts and linear ridges in the Line Islands reveal the complexityof volcanicevents in this province.Figures/5 and 6 + summarize the relevent radiometric and fossil age determi- •o nationsfrom the site of Horizon Guyot south to the Tuamotu

20 Islands. Table 2 lists the data referred to in this section. For the purposesof this discussionthe well-definedLine chain is

3 taken to extend from Horizon Guyot and the northern end of the linear ridge from which dredge 142-D was taken to the southern end of the line of seamounts from which RD-45 was

+ NP •'

L5

L4

Fig. 7. Comparison of high-latitude seamount paleopoles and Pacific Eocenepaleomagnetic data. The VGP's of seamountsL4-L7 are shown by open stars. L4 has a total fusion age of 39 + 1.5 Ma. SeamountsEl, E2, and HR1 have VGP's shown by the solid stars. HR1 has been assignedan age of 41-42 Ma by its position in the Hawaiian chain [Sager, 1984] and E1 and E2 have maximum agesof 45 and 37 Ma, respectively,determined by the age of the seafloor upon whichthey rest [Sager,1984]. The heavyline is a segmentof the small circle that is the locus of the paleomagneticpoles measured from sedimentsamples from DSDP site 166. These samplesrange in age from Oligoceneto Eocene [Jarrard, 1973]. The thin line is a Fig. 8. Comparison of Line Islands seamount paleomagnetic similar segmentof the polar circle estimatedfrom the identificationof poles with Pacific apparent polar wander paths. The Line Islands equatorialsediments in the coresfrom DSDP site 163.It wasdeter- VGP's are shownby solid stars.Solid circlesshow the locationsof the mined that this site crossed the 43 + 6 Ma [Suarez and meanpaleomagnetic poles calculated by Sager[1983a, 1984],whereas Molnar, 1980]. The large dot is the mean position of the Eocene the solid squaresrepresent the mean pole positionsof Gordon[1982a, seamountVGP's and the ellipseis the 95% confidenceregion of the 1983]. The solid line is the apparent polar wander path of Sager mean paleomagneticpole [Sager, 1984]. The small dots connectedby [1983a]' the dashedline, the path of Gordon[1983]. The 95% confi- the dashedlines are positionsof the paleopolepredicted by the Paci- denceregions surrounding the mean paleomagneticpoles are shown fic plate/hot spot rotation model of Jarrard and Clague [1977] (with as ellipses,and the numberswithin the ellipsesare the mean agesof rotation rate corrected by Dalryrnpleet al. [1977]) labeled at 10-Ma the poles in millions of years. SeamountsL2, L3, L4, and L8 have intervals. The diamonds are the VGP's of Cretaceous seamounts from total fusion agesof 85.0 + 1.1, 82.7 + 3.6, 39.3 + 1.1, and 71.9 + 1.4 Harrisonet al. [ 1975].Map projectionis polar equalarea. Ma, respectively.Map projectionis polar equal area. SCHLANGERET AL.' GEOLOGYAND GEOCHRONOLOGYOF THE LINE ISLANDS 11,269 taken that extendnorth from Caroline Island. The sequenceof At the southern end of the central Line Islands the cluster of volcanic events at both the northern and southern boundaries dates represented by PCOD-6 (69.8 _+ 1.0 Ma), RD-41 of the Line chain are also complex. Drilling on Horizon (35.5 _+0.9 Ma), RD-43 (59.0 _+0.8 Ma), and DSDP site 316 Guyot [Winterer et al., 1973] showed this ridge to have a (minimum age of cessationof volcanismat 81-83 Ma) presents history of two Cretaceous volcanic events separated by a further complicationsfor a single hot spot linear progression period of Cenomanian growth. Further, dates obtained in model. Sample PCOD-6 consistedof a relatively fresh cobble our study from the northeast end of the Tuamotu Islands of of plagioclase-richbasalt in a piston core taken in a turbidite- 41.8 +_0.9 and 47.4 _+0.9 Ma (RD-52) complicate the gener- filled basin which lies downslopefrom the site of RD-41. This ally accepted argument that the Line-Tuamotu elbow is basalt cobble must representdebris from a seamount.Thus we coevalwith the Emperor-Hawaiian elbow dated at 43 Ma. On see in a very close geological conjunction edifice ages of the basisof fossilevidence derived from drilling at DSDP site 69.8 + 1.0 and 35.5 + 0.9 Ma. The PCOD-6 and RD-43 dates 318 in the Tuamotu chain [Schlan•teret al., 1976], where late fit on the linear progressionline of Figure 6. The cessationof early Eocene shallow water fossils in turbidites overlie the volcanism at DSDP site 316 greatly predates both PCOD-6 presumedbasalt basement,it was argued that the volcanic and RD-43; three distinct periods of volcanismare seenin this edificesof the Tuamotu ridge could have formed more than 51 region. Seamount L7 of this paper, of Eoceneage, could also Ma. Also, Ha•l•lerty et al. [1982] determined, paleon- be ascribedto the main age-progressivetrend linear ridge. tologically,that Eocenevolcanism, -,•44-50 Ma, took place in The final set of dates to be discussed are those from the the southern Line chain at the site of RD-45. Therefore the southernLine Islands,dredge hauls RD-44 and RD-45. Prior data available at this time show that volcanism in the area of to the determination of radiometric age dates from these the Line-Tuamotu elbow took place from more than 51 to 42 dredgehauls, Ha•t•terty et al. [1982] had determinedon fossi. Ma. evidence that the seamount from which RD-45 was obtained Within the Line chain the following chronologicdata needs had a minimum age of 70-75 Ma. This seamountis capped by to be taken into account. As shown on Figures 5 and 6, the coarse-grainedbioclastic of reefal origin which con- majority of dated volcanic edificesdefine an overall age pro- tains rudists, calcareousalgae, and fragments. gression along the Line chain from the location of dredge Further, volcanic peperite was dredged from the RD-45 sea- 142-D (93.4 _+ 1.3 Ma) to the Line-Tuamotu elbow. Watkins mount. Tests of planktonic incorporated into this Seamount(L1 of this paper; Figure 1) is also consideredto be volcanicrock have been dated as middle Eocenein age, 44-50 of Late Cretaceousage. A rate of migration of volcanism of Ma [Ha•t•terty et al., 1982]. The ages of 71.9 q-1.4 and 9.6 +_0.4 cm/yr best fits this trend. However, deviations from 70.5 q- 1.1 Ma reported here are consistent with the Mae- the progressivetrend exist. The cluster of agesrepresented by strichtian age of shield-buildingvolcanism at these seamounts RD-63 (86.0_+ 0.9 Ma), RD-61 (81.4 _+ 1.1 Ma, 82.6 _+0.7 which must have been above sea level in Late Cretaceous time Ma), RD-59 (85.0 _+ 1.1 Ma), and 133-D (84.4 _+0.9 Ma) are as determinedby the fossilevidence of Haggerry et al. [1982]. associatedwith 137D dated at -,•57 Ma by Saito and Ozima These Cretaceous dates fall far above the postulated line of [1977]. The elongate ridge from which 137-D was dredged progressionshown on Figure 6 and argue persuasivelyagainst exactly parallels the ridges from which RD-63 and RD-61 the formation of the Line Islands by a single hot spot. In fact, were taken; RD-59 is from an isolated seamountlying off the when consideredtogether with the DSDP results from 165, southernend of the ridge from which 137-D was taken. It is 315, and 316, one could argue that an older progressivetrend difficult to reconcilethis parallelism of ridgeswith the path of existedwhich paralleled the trend line defined in Figure 6. The a single hot spot to account for both the Cretaceous and phase of Eocene volcanic activity at RD-45 can be explained Paleocene ages in this short sector of the Line Islands. Fur- by hot spot volcanism along the main trend. We wish to ther, 137-D lies to the north of the well-defined Line cross emphasizethe following points concerningvolcanic events' trend (Figure 1). SeamountsL2 and L3 of this paper (Table 4 1. The majority of dated can be interpreted as and Figure 1) yielded Cretaceous ages based on paleo- defining an age-progressivevolcanic trend from 93 (142-D) to magneticswhich are in accordwith the radiometricage deter- 44-50 Ma (RD-45 southward along the Line chain, which minations of RD-59 and 133-D. In the Line cross trend itself supports an origin by hot spot volcanism now active in the (Figure 2), which has a geometrictrend distinct in the prov- vicinity of Easter Island. During this period the Pacific plate ince, the only radiometric age obtained at 128D (78.7 +_ 1.3 moved at 9.6 + 0.4 cm/yr acrossthis hot spot. Ma) plots in the main trend line shown in Figure 6 even 2. Three volcanos exhibit Eocene and Paleocene vol- though the seamount from which the rock was taken lies far canism (between 38 and 60 Ma) which does not fit this main to the east of the main Line chain. trend' 137-D at 15øN, RD-33 in the Line Islands cross trend, Further, the long time span of volcanismat site 165 and the and RD-41 at 2øN. This small group of agespossibly shows an age of 39.3 +_ 1.5 Ma obtained for RD-33 argue against a age progression which could be related to volcanic over- straightforward age progressionalong the Line cross trend, printing by hot spot volcanism along the Line-Marquesas although both of these locations,it could be argued, lie to the Swell [Crou•th and Jarrard, 1981]. Alternatively, these vol- south of the well-definedcross trend and may representvol- canos representa period of rejuvenescencenot related to hot canic activity of Cretaceous age along the 9.6 _+0.4 cm/yr spot volcanism but part of a Pacific-wide Eocene volcanic progressionshown in Figure 6 complicated by Eocene vol- event [Ha•t•terty et al., 1982]. In either case, the Eocene and canism along one of the SE trending ridges which intersects Paleoceneedifices and ridges sampled are in close proximity the Line Islands chain. SeamountsL4 and L5 of this paper are to older Cretaceous edifices and ridges, as in the case of of age based on paleomagneticdata. In the central RD-45 and the pair of ages shown by PCOD-6 and RD-41. Line Islands the date of 76.4 _+0.5 Ma for dredge 123-D falls The same edificesmay be compositesof both Cretaceousand on the line of age progressionsshown on Figure 6. However, Eocene volcanic episodesand clearly require a volcanic his- at site 315 major edifice building ceasedby 93.3 Ma and a tory requiring more than a singlehot spot. nearby seamount,L6 of this study, is probably of Eoceneage. 3. The Cretaceous ages determined at DSDP sites 165, 11,270 SCHLANGERET AL.' GEOLOGYAND GEOCHRONOLOGYOF THE LINE ISLANDS

315, and 316 and from RD-44 and RD-45 do not fit 9.6 + 0.4 north of Caroline Island (dredge hauls 44 and 45). The eXis- cm/yr trend and could be interpreted as representinga pro- tenceof reefsof late Campanianto early-middleMaestrichtian gressiveage trend that is systematically10-20 Ma older than age along a 2500-km length of the Line chain from DSDP site the linear trend shown in Figure 6 which connectsthe Tua- 165 to RD-45 shows reef development over this distance be- motu elbow with the northernmostLine Islands.It is particu- tween 70 and 75 Ma. larly difficult to fit the 93.3 Ma age of volcanismat DSDP site Evidencefor post-Cretaceousreef growth is only recorded 315, which is concordantwith the age of the Cuneolinaassem- from DSDP site 165, where late Eocene reef-derived fossil blagethere, to the 9.6 + 0.4 cm/yr trend. debris was cored in Oligocene turbidites [Premoli-Silva and Brusa, 1981] and at Caroline Island (RD-46 and 47) where REEF GROWTH AND dredge hauls recovered reef of Eocene through One of the major implicationsof the hot spot model for the Plio-Pleistoceneage; Caroline Island is an atoll today [Ha•t- developmentof a linear island chain is that the history of reef •terty et al., 1982]. At the site of dredge haul RD-44, late growth along the chain should reflect the progressivechrono- Cretaceousreefs were succeededby a late Paleoceneneritic- logy of volcanismalong the chain, provided that the volcanic shelfcarbonate unit, but by middle Eocenetime, pelagicsedi- edificesbuilt from the seafloorinto the photic zone in latitudes mentation prevailed at this site. We can assumethat there are amenable to vigorous reef growth. The thicknessof the reef sectionsof post-Cretaceousreefs below the atolls of the cen- cap on an edifice would depend on the subsidencepath of the tral Line chain. It is important to note that dredge hauls edifice. Along a hot spot path the edifices and the seafloor RD-57 through 64 recovered no reefal limestone debris be- would be expectedto sink along an age-depthcurve such as tweenJohnston Island and .The morphologyof one of those proposed by Parsons and Sclater [1977], the chain and the pattern of reef growth along it contrasts Heestandand Crou•th [1981], and Crou•th [1983] where depth with the Emperor-Hawaiian chain. Long sectionsof the Line is equalto a functionof t •/2 chain lack atolls, i.e., from Johnston Island to Kingman Reef The Emperor-Hawaiian chain is the best documented hot and from Christmas Island to Caroline Island. This lack of spot trace in the Pacific Basin; along this chain the history of atolls, consideringthat the Line chain was in latitudes amen- reef growth and development follows a classical Darwinian able to reef growth throughout its history from Late Cre- pattern [see Jackson et al., 1980]. Around the island of taceousto the present day, argues against a hot spot model Hawaii, now over the hot spot, only Recent reefs are known. that would produce a chain capped by a more regularly dis- At Midway Island, drilling revealedthat reef growth has been posed chain of atolls. Further, the irregular distribution of continuous since time [Ladd et al., 1970]. Drilling atolls arguesfor a subsidencehistory along the chain at vari- along the Emperor chain showed that Paleocenereefal car- ance with one predicted by a regular age-depthrelationship bonates cap seamountsin that part of the chain. Failure of relatedto t •/2. Where reefinghistory is known from closely atolls to maintain themselvesat sealevel in the Emperor chain spacededifices such as the sites of RD-44, RD-45, and Caro- can be ascribedto the transport of the Emperor edificesinto line Island (RD-46 and 47), the subsidencehistory appearsto colder, northern waters as these edificesmoved NW atop the have beencomplex. At RD-44 the apparentsuccession of Late Pacificplate. Moreover, an almostcontinuous chain of closely Cretaceousreefs followed by late Paleoceneneritic shelf car- spacedatolls extendsfrom Midway Island SE along the chain bonatesand then by middle Eocenepelagic sediments suggests to Gardner Pinnacle where the volcanic edifice breaks sea that the edifice subsidedtoo quickly for the maintenanceof level. reef growth; subsidenceat this site amounted to 1500 m since Therefore,if the Line chain formed through the action of a Late Cretaceoustime. At RD-45, reef growth ceasedby Mae- single hot spot, we would expect to see a history of reef strichtian time, and the edifice has subsided 1200 m since that growth and subsidencealong the chain somewhatparallel but time. At Caroline Island, reef growth has persistedat least older than that seen in the Emperor-Hawaiian chain. How- from Eocenetime to the presentday. We note that Epp [this ever, in the case of the Line chain, two kinds of data needed issue] estimated subsidencerates at DSDP sites 165 and 316 for a detailed reconstructionof a subsidencehistory are large- and at the sites of the seamountsfrom which dredge hauls ly lacking. First, the age of the seafloor in the vicinity of the RD-44 and RD-45 were taken (see also Haggetty [1982] for Line chain is poorly known, and second,the history of reef original data) and came to the conclusionthat each of these growth can only be inferred from data from DSDP sites 165, sites underwent different subsidence histories no one of which 315, and 316 in the central part of the chain and dredge haul fit the subsidencepaths predicted by the age-depthcurves of sitesRD-44, 45, 46, and 47 in the southernpart of the chain. Parsonsand Sclater [1977], Heestand and Crough [1981] and The oldest known reefs in the main Line chain were in the Crough [1983]. vicinity of Fanning Island as shown by data from DSDP site 315. There Prernoli-Silva and Brusa [1981] report a mid- DISCUSSION CretaceousCuneolina assemblage of Cenomanian-Albian age The purpose of the researchcarried out by the present au- in turbidites of both late Campanian-middle Maestrichtian thors was the testing of the various models put forward to and Oligocene age. The presenceof this mid-Cretaceousfauna account for the Line chain. Beginningwith Morgan's [1972] is in concordancewith the K-Ar age of 93.3 Ma reported for proposition that the Line chain was generatedby a single hot the basal basalt at DSDP site 315. If we consider Horizon spot other models followed. Natland [1976] argued that the Guyot to be the northern extensionof the Line chain, then we petrologiccharacter of several basaltic rocks recoveredfrom must consider the Cuneolina-bearingreef limestone drilled the Line chain and the crosstrend indicated that at least parts there at DSDP site 171 [Winterer et al., 1973] to be coeval of the chain were part of a central Pacific rift systemof limited with the mid-Cretaceous reef reported at DSDP site 315. crustal extension. Winterer [1976] postulated a ridge crest During late Campanian through early-middle Maestrichtian origin for the main Line chain at ~ 110 Ma followedby ridge time (70-75 Ma) reefsexisted in the vicinity of Kingman Reef jumps; he recognizedthe importanceof the NW-SE trending (DSDP site 165), Fanning Island (DSDP site 315), south of crossridges and proposedthat they were youngerthan the ChristmasIsland (DSDP site 316), and on seamountsjust main Line trend and that the complexityof the chainwas due SCHLANOERET AL.: GEOLOOYAND GEOCHRONOLOGYOF THE Ll•u• ISLANDS 11,271

to the passageof severalhot spotsacross the chain. Crough transformfaults, or activity at ridge crestsand we do not, for and Jarrard [1981] proposedthat one suchcross trend contin- the Line chain,accept ridge erestor transformfault associated ued into the Line chain from the Marquesas Islands, Farrar volcanismas a major factor, then we are drawn, inductably, and Dixon [1981] proposedfracture zone volcanismas a to the hot spot mechanismto accountfor the chain. Our data major factor in the formation of the Lines. Hendersonand can be interpretedto show that two hot spots traversedthe Gordon [1982] have attempted to explain the complex vol- present chain. However, both of these would have had to canic history of the Line chain through the action of four hot follow parallel pathswith a time interval of from 10 to 20 m.y. spotsthat traversedthe area of the presentchain at different betweenthem. One would expect that a secondhot spotpass- times.Finally, Epp [this issue]proposed two models,each of ing so closely along an earlier hot spot trace would have which employsa singlehot spot,for the evolutionof the Line produceda more morphologicallycoherent chain with many chain. He arguesthat perturbationsof the signal,i.e., the sur- more atolls than we now see in the Line chain. The geotec- face expressionof volcanism,of the hot spot accountfor the tonicimagery map of W. Haxby [seeFrancheteau, 1983] cer- complexchronology of the volcanicevents along the chain. tainly indicatesthat many possiblehot spot traces pass into The purposeof this sectionthen is to presentarguments the Line chain, at least one of which coincides with the hot against certain modelsthat tend to invalidate them and that spottrack between the Line chain and ihe Marquesas aspro- place constraintson other models.This sectionmight be posedby Croughand Jarrard [1981]. Hendersonand Gordon viewed as constructed after the Popperian argument that [1982] propose that four hot spots formed the Line chain, whileit maynot be logically possible to validatean hypothesis althoughtheir model doesnot accountfor many of the dated by discoveringnew confirmingdata, it is possibleto invalidate edifices in the chain. an hypothesisby discoveringeven a single new piece of evi- Beforecompletely accepting the efficacyof multiple hot spot dencethat contradictsthat hypothesis. modelsin explainingindividual island chains we wish to em - The ridge crest model set is difficult to justify on palco- phasizethat complexvolcanic thronelogics are not restricted magneticgrounds inasmuch as the magneticanomaly patterns to the Line chair• but characterize the entire Pacific Basin implicit in suchmodels are not k,nownin and around the Line from the Line Islands to the western subduetion zone bound- chain. A strongerargument against the ridge crest model lies ariesof the present Padfie plate [Schlanoer and Premoli-$ilv a, in the petrologiccharacter of the basalts from the chain as 1981; Schlanoeret al., 1981; Haooerty, 1982; Hagoerty et al., discussedabove. Our data and that of Jacksonet al. [1976] 1982]. Both Cretaceousand Eocenevolcanism blanketed this show that the basalts in the Line chain are similar to those westernpart of the Pacific Basin; Cretaceous.volcanism took found in the Empe•ror-Hawaiianchain, the classichot spot placeover a vastarea between ~ 110and ~ 70Ma [Wattset trace; the basaltsare largely normal oceanicisland basalts. el., 1980; Menard, 1964; Larson and Schlanoer,1981; Rea and The fracture zone model is attractive in that it would ac- Vallier, 1983]. Indeed, mid-Cretaceousvolcanism is a global phenomenon[Schlanoer et al., 1981]. We propose that a count for the simultaneity of much of the Cretaceousvol- , canismalong the Line chain. However, a fracturezone origin mechanismexists that would allow the surfaceexpression of would imply large amounts of strike-slipdisplacement along coevalvolcanism over wider areas than that allowed for in the axis of the chain. Farrat and Dixon [1981] proposed1700 necessarilylaterally restrictedhot spot models. km of strike-slipmotion along a fracturezone that extended from the Emperor Trough through the Gardner Seamounts Acknowledgments.This researchwas SUl•po•rtedby the Officeof and through the entire Line chain. Gordon[1982b], basedon Naval Research.We especiallythank the captainand crew of the R/V Kana Keokifor theirefforts during three cruises of the shipin atoll- a review of available pale0magneticdata has effectivelyrefu- Studdedwaters. The governmentof the Republicof kindly ted the Farrar-Dixon model. allowed us to conduct research in their territorial waters. Also we Models that ascribethe Line chain to the action of a single appreciatethe help and hospitalityof the peopleof Fanning Island hot spot run into difficultiesin the face of the distribution while R/V Kana Keoki lay disabledoff the island. Discussionson the geologyof linear island chainswith R. Gordon and L. Henderson along the entirechain of radiometricand bigstratigraphicdata wereenlightening and helpful.R. Gordonkindly reviewed the manu- relevent to the timing of volcanic edifice formation and script.Over the past decade,as a shipmate'on cruisesto the Line periods of reef growth. If we take the Emperor-Hawaiian Islands and in the Courseof many fruitful discussions,E. L. Winterer chain [see Jacksonet al., 1980] as the exampleof a singlehot hascontributed greatly to our efforts.Analyses of reefalfaunas in the spot trace,we seethat individual edificesin that chain built up LineIslands province by I. Premoli-Silvashowed the extentof Cre- taceousreef growth. We thank J. Morgan and R. Jarrardfor critical from the seafloorto sealevel in 1-2 m.y. and the reefsimmedi- reviewsand helpfulsuggestions for the improvementof the original ately formedon and around theseedifices. This example manuscript.Hawaii Instituteof Geophysicscontribution 1486. cannot be applied to the Line chain. A singlehot spot moving through the Line chain cannotaccount for (1) the formationof REFERENCES Cenomanianor older reefsat both Horizon Guyot and in the vicinity of Fanning Island (DSDP site 315), (2) the Late Cre- Basaltic Volcanism Study Project, Oceanic intraplate volcanism,in Basaltic Volcanismon the Terrestrial Planets, pp. 161-192, Perga- taceousand Eocenevolcanism recorded at siteRD-45, (3) the mon,New York, 1981. simultaneousappearance of reefsof late Campanian to early- Bass,M. N., R. Moberly, J. M. Rhodes,C.-Y. Shih, and S. E. Church, middle Maestrichtianage from near DSDP site 165 to the site Volcanicrocks cored in the centralPacific, leg 17, DSDP, Initial of RD-45, and (4) the irregular subsidencehistory of at least Rep. Deep Sea Drill. Proj., 17, 429-466, 1973. • Coombs,D. S., and J. F. G. Wilkinson, Lineagesand fractionation parts of the Line chain. trendsin undersaturatedvolcanic rocks from the East Otago vol- Further, as shownon Figure 6, the trace of a singlehot spot canicprovince (New Zealand)ant! relatedrocks, J. Petrol.,10, which may have causedthe progressivevolcanism along the 440-501, 1969. chain at a plate motion rate of 9.6 + 0,4 cm/yr does not ac- Crough, S. T., Hot spot swells,in Annu. Rev. Earth Planet. Sci., 11, count for the agesof volcanismseen at DSDP sites 165, 315, 165-193, 1983. Crough,S. T., and R. D. Jarrard,The Marquesas-LineSwell, J. Gee- and 316 and at sites RD-44 and RD-45. phys.Res., 86, 11763-11771, 1981. If we adhere to the generallyaccepted idea that significant Dalrymple, G. B., and D. A. Clag.ue,Age of the Hawaiian-Emperor volcanismin the oceanic basins is due to hot spots, leaky Bend, Earth Planet. $ci. Lett., 31, 313-329, 1976. 11,272 SCHLANGERET AL..' GEOLOGYAND GEOCHRONOLOGYOF THE LINE ISLANDS

Dalrymple,G. B., D. A. Clague,and M. A. Lanphere,Revised age for Natland, J. H., Petrology of volcanicrocks dredged from seamounts Midway , Hawaiian volcanicchain, Earth Planet. Sci. Lett., in the Line Islands, Initial Rep. Deep Sea Drill. Proj., 33, 749-777, 37, 107-116, 1977. 1976. Dalrymple, G. B., E. C. Alexander,Jr., M. A. Lanphere, and G. P. Orwig, T. L., and L. W. Kroenke, Tectonicsof the easterncentral Kraker,Irradiation of samplesfor '•øAr/39Ardating using the Geo- Pacific basin, Mar. Geol., 34, 29-43, 1980. logical Survey TRIGA Reactor (GSTR), U.S. Geol. Surv.Prof. Pap., Parsons,B., and J. G. Sclater,An analysisof the variation of ocean 1176, 1-54, 1981. floor bathymetry and heat flow with age, J. Geophys.Res., 82, Davis, G. T., J. J. Naughton, and J. A. Philpotts,Petrology, geochem- 803-827, 1977. istry, and agesof Line Islandsbasaltic rocks, central Pacific Ocean Premoli-Silva, I., and C. Brusa, Shallow-water skeletal debris and (abstract),Eos Trans. AGU, 61, 1144, 1980. larger foraminiferasfrom Deep Sea Drilling Project site 462, Nauru Duncan, R. A., Geochronologyof basaltsfrom the NinetyeastRidge Basin,western equatorial Pacific,Initial Rep. Deep Sea Drill. Proj., and continental disperisonin the eastern Indian Ocean, J. Vol- 61, 439-473, 1981. canol. Geotherm. Res., 4, 283-305, 1978. Rea, D. K., and T. Vallier, Two Cretaceousvolcanic episodesin the Duncan, R A., A capturedisland chain in the Coast Range of Oregon western Pacific Ocean, Geol. Soc. Am. Bull., 94, 1430-1437, 1983. and Washington,J. Geophys.Res., 87, 10827-10837, 1982. Richards,M. L., V. Vacquier,and G. D. Van Voorhis,Calculation of Duncan, R. A., Overprinted volcanismin the Line Islands and Pacific the magnetizationof upliftsfrom combiningtopographic and mag- plate motion since 100 m.y. (abstract),Eos Trans. AGU, 64, 843, netic surveys,Geophysics, 32, 678-707, 1967. 1983. Sager,W. W., Seamountpaleomagnetism and Pacific plate tectonics, Epp, D., Possibleperturbations to hot spot traces and implications Ph.D. dissertation,472 pp., Univ. of Hawaii, Honolulu, 1983a. for the origin and structureof the Line Islands J. Geophys.Res., Sager,W. W., A late Eocenepaleomagnetic pole for the Pacificplate, this issue. Earth Planet. Sci. Lett., 63, 408-422, 1983b. Farrar, E., and J. M. Dixon, Early Tertiary rupture of the Pacific Sager,W. W., Paleomagnetismof Abbott Seamountand implications plate: 1700 km of dextral offset along the Emperor Trough-Line for the latitudinal drift of the Hawaiian hot spot, J Geophys.Res., Islands lineament, Earth Planet. Sci. Lett., 53, 307-322, 1981. 89, 6271-6284, 1984. Francheteau, J., The oceanic crust, Sci. Am., 249, 114-129, 1983. Sager,W. W., and B. H. Keating, Paleomagnetismof Line Islands Gordon, R. G., The Late paleomagneticpole of the seamounts:Evidence for Late Cretaceousand early Tertiary vol- Pacificplate, Geophys.J. R. Astron.Soc., 70, 129-140, 1982a. canism,J. Geophys.Res., this issue. Gordon, R. G., Paleomagnetictest of the Emperor fracture zone hy- Sager,W. W., G. T. Davis, B. H. Keating, and J. A. Philpotts,A pothesis,Geophys. Res. Lett., 9, 1283-1286, 1982b. geophysicaland geologicstudy of Nagata Seamount,northern Line Gordon, R. G., Late Cretaceousapparent polar wander of the Pacific Islands,J.Geomagn. Geoelectr., 34, 283-305, 1982. plate: Evidencefor a rapid shift of the Pacifichot spotswith respect Saito,K., and M. Ozima,'•øAr-39Ar ages of submarinerocks from the to the spin axis, Geophys.Res. Lett., 10, 709-712, 1983. Line Islands:Implications on the origin of the Line Islands,in The Haggerty, J. A., The geologic history of the southern Line Islands, Geophysicsof the Pacific OceanBasin and Its Margin, Geophys. Ph.D. dissertation,202 pp., Univ. of Hawaii, Honolulu, 1982. Monogr. Ser., vol. 19, edited by G. H. Sutton, M. H. Manghnani, Haggerty, J. A., S. O. Schlanger, and I. Premoli-Silva, Late Cre- and R. Moberly, pp. 369-374, AGU, Washington,D.C., 1976. taceous and Eocene volcanism in the southern Line Islands and Saito,K., and M. Ozima,4øAr-39Ar geochronological studies on sub- implicationsfor hot spot theory, Geology,10, 433-427, 1982. marine rocks from the western Pacific area, Earth Planet. Sci. Lett., Harland, W. B., A. V. Cox, P. G. Llewellyn, C. A. G. Pickton, A. G. 33, 353-369, 1977. Smith, and R. Walters, A GeologicTime Scale, 131 pp., Cambridge Schilling,J.-G., M. Zajac, R. Evans, T. Johnston,W. White, J. D. University Press,New York, 1982. Devine, and R. Kingsley, petrologic and geochemicalvariations Harrison, C. G. A., R. D. Jarrard, V. Vacquier, and R. L. Larson, along the Mid-Atlantic Ridgefrom 29øN to 73øN, Am. J. Sci.,283, Paleomagnetismof CretaceousPacific seamounts,Geophys. J. R. 510-586, 1983. Astron. Soc., 42, 859-882, 1975. Schlanger,S. O., and I. Premoli-Silva,Tectonic, volcanic, and pa- Heestand, R L., and S. T. Crough, The effect of hot spots on the leogeographicimplications of redepositedreef faunasof Late Cre- oceanicage-depth relation, J. Geophys.Res., 86, 6107-6114, 1981. taceousand Tertiary age from the Nauru Basin and the Line Is- Henderson,L. J., and R. G. Gordon, Fixed hot spots and recurrent lands, Initial Rep. Deep Sea Drill. Proj., 61, 817-827, 1981. volcanismalong the Line Islands chain, Geol. Soc. Am. Abstr. Pro- Schlanger,S. O., et al., Initial Reportsof the Deep Sea Drilling Project, grams,14, 513, 1982. 33, 1976. Jackson,E. D., K. E. Barbar, B. P. Fabbi, and C. Heropoulos, Pet- Schlanger,S. O., H. C. Jenkyns,and I. Premoli-Silva,Volcanism and rology of basalticrocks drilled on leg 33 of the Deep Sea Drilling vertical tectonicsin the Pacific basin related to global Cretaceous Project, Initial Rep. Deep Sea Drill. Proj., 33, 571-613, 1976. transgressions,Earth Planet. Sci. Lett., 52, 435-449, 1981. Jackson, E. D., I. Koisumi, G. B. Dalrymple, D. H. Clague, R. J. Suarez,G., and P. Molnar, Paleomagneticdata and Kirkpatrick, and H. G. Greene, Introduction and summary of re- faciesand the motion of the Pacific plate relative to the spin axis sults from DSDP leg 55, the Hawaiian-Emperor hot spot experi- sincethe Late Cretaceous,J. Geophys.Res., 85, 5257-5280, 1980. ment, Initial Rep. DeepSea Drill. Proj., 55, 5-31, 1980. Watts, A. B., J. H. Bodine, and N.M. Ribe, Observations of flexure Jarrard, R. D., Paleomagnetismof leg 17 sedimentcores, Initial Rep. and the geologicalevolution of the Pacific Ocean basin, Nature, Deep Sea Drill. Proj., 17, 365-275, 1973. 283, 532-537, 1980. Jarrard, R. D., and D. A. Clague, Implications of Pacific Island and Winterer, E. L., Bathymetryand regionaltectonic setting of the Line seamount ages for the origin of volcanic chains, Rev. Geophys. Islandschain, Initial Rep. DeepSea Drill. Proj., 33, 731-748, 1976. SpacePhys., 15, 57-76, 1977. Winterer, E. L., et al., Initial Reportsof the Deep Sea Drilling Project, Keating, B., and W. Sager, Watkins Seamount: Preliminary paleo- 17, 1973. magneticresults, J. Geophys.Res., 85, 3567-3571, 1980. Ladd, H. S., J. I. Tracey, Jr., and M. G. Grass, Deep drilling on R. A. Duncan, School of Oceanography,Oregon State University, Midway Atoll, U.S. Geol.Surv. Prof. Pap., 680-A, A1-A21, 1970. Corvallis, OR 97331. Lanphere, M. A., and G. B. Dalrymple, K-Ar ages of basalts from M. O. Garcia, B. H. Keating, and J. J. Naughton,Hawaii Institute DSDP leg 33: Sites 315 (Line Islands)and 317 (Manihiki Plateau), of Geophysics,University of Hawaii, Honolulu, HI 96822. Initial Rep. Deep Sea Drill. Proj., 33, 649-653, 1976. J. A. Haggerty,Department of Geosciences,University of Tulsa, Lanphere, M. A., G. B. Dalrymple, and D. Clague, Rb-Sr systematics Tulsa, OK 74104. of basalts from the Hawaiian-Emperor volcanic chain, Initial Rep. J. A. Philpotts,U.S. GeologicalSurvey, National Center, Reston,VA Deep Sea Drill. Proj., 55, 695-706, 1980. 22092. Larson, R. L., and S. O. Schlanger,Geological evolution of the Nauru W. W. Sager,Department of Oceanography,Texas A & M Univer- Basin, and regional implications,Initial Rep. Deep Sea Drill. Proj., sity,College Station, TX 77843. 61, 841-862, 1981. S. O. Schlanger,Department of GeologicalSciences, Northwestern LeBas, M. J., The role of aluminum in igneousclinopyroxenes with University,Evanston, IL 60201. relation to their parentage,Am. J. Sci.,260, 267-288, 1962. Menard, H. W., Marine Geologyof the Pacific,271 pp., McGraw-Hill, New York, 1964. (ReceivedJ, anuary 5, 1984; Morgan, W. J., Deep mantle convectionplumes and plate motions, revisedMay 2, 1984; Am. Assoc.Petrol. Geol. Bull., 56, 203-213, 1972. acceptedMay 25, 1984.)