the amount of recycled crust in sources of www ...the amount of recycled crust in sources of...
TRANSCRIPT
DOI: 10.1126/science. 1138113 , 412 (2007); 316Science
et al.Alexander V. Sobolev,Mantle-Derived MeltsThe Amount of Recycled Crust in Sources of
www.sciencemag.org (this information is current as of April 19, 2007 ):The following resources related to this article are available online at
http://www.sciencemag.org/cgi/content/full/316/5823/412version of this article at:
including high-resolution figures, can be found in the onlineUpdated information and services,
http://www.sciencemag.org/cgi/content/full/1138113/DC1 can be found at: Supporting Online Material
http://www.sciencemag.org/cgi/content/full/316/5823/412#otherarticles, 6 of which can be accessed for free: cites 38 articlesThis article
http://www.sciencemag.org/cgi/collection/geochem_physGeochemistry, Geophysics
: subject collectionsThis article appears in the following
http://www.sciencemag.org/about/permissions.dtl in whole or in part can be found at: this article
permission to reproduce of this article or about obtaining reprintsInformation about obtaining
registered trademark of AAAS. c 2007 by the American Association for the Advancement of Science; all rights reserved. The title SCIENCE is a
CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
on
Apr
il 19
, 200
7 w
ww
.sci
ence
mag
.org
Dow
nloa
ded
from
The Amount of Recycled Crust inSources of Mantle-Derived MeltsAlexander V. Sobolev,1,2* Albrecht W. Hofmann,1 Dmitry V. Kuzmin,1,3 Gregory M. Yaxley,4Nicholas T. Arndt,5 Sun-Lin Chung,6 Leonid V. Danyushevsky,7 Tim Elliott,8 Frederick A. Frey,9Michael O. Garcia,10 Andrey A. Gurenko,1 Vadim S. Kamenetsky,7 Andrew C. Kerr,11Nadezhda A. Krivolutskaya,2 Vladimir V. Matvienkov,12 Igor K. Nikogosian,13,14Alexander Rocholl,15 Ingvar A. Sigurdsson,16 Nadezhda M. Sushchevskaya,2 Mengist Teklay17
Plate tectonic processes introduce basaltic crust (as eclogite) into the peridotitic mantle. Theproportions of these two sources in mantle melts are poorly understood. Silica-rich melts formedfrom eclogite react with peridotite, converting it to olivine-free pyroxenite. Partial melts of thishybrid pyroxenite are higher in nickel and silicon but poorer in manganese, calcium, andmagnesium than melts of peridotite. Olivine phenocrysts’ compositions record these differencesand were used to quantify the contributions of pyroxenite-derived melts in mid-ocean ridge basalts(10 to 30%), ocean island and continental basalts (many >60%), and komatiites (20 to 30%).These results imply involvement of 2 to 20% (up to 28%) of recycled crust in mantle melting.
It is widely accepted that the heterogeneity ofthe convecting mantle observed in thecomposition of mantle-derived magmas is
largely due to subduction and recycling ofoceanic crust into the deep mantle (1, 2). Tounderstand the role of crustal material in creatingcompositional heterogeneities in the mantle andto evaluate the geodynamical consequences ofthis contribution, one must quantify the crustal
input to the mantle sources of common, mantle-derived magmas in mid-oceanic ridges basalts(MORBs), ocean islands (OIBs), and largeigneous provinces (LIPs). It is not possible to useincompatible element abundances in basalts toconstrain the proportion of recycled component inthe magma source because concentrations of theseelements are also sensitive to the extent of melt-ing. Similarly, the use of isotope ratios for mak-ing such quantitative estimates is compromisedby the isotopic variability of subducted materialsinvolved in the recycling process (2). We used analternative approach based on a combination ofmajor elements and compatible trace elements inparental melts, because these are more uniform inthe mantle and are strongly controlled by theresidual phases in equilibrium with partial melts(3–5).
Our method has its basis in the experimentaland theoretical prediction that high-pressure (P >3.0 GPa) melting of typical recycled oceaniccrust (in the form of eclogite with a separate SiO2
phase) and reaction of this melt with peridotiteproduces olivine-free pyroxenite (5). We showthat further melting of this hybrid lithology in theabsence of residual olivine is more voluminousthan the melting of peridotite (at a given pressureand temperature) and that pyroxenite-derivedmelts are characteristically enriched in Si andNi but depleted in Mg, Ca, and Mn comparedwith their peridotite-derived counterparts. Thisdifference arises because olivine principallycontrols the composition of melt produced inperidotite, whereas pyroxene mainly controls thecomposition of melt from olivine-free hybridpyroxenite (5–8). Experimental data predict (9)that, as such pyroxenite-derivedmelts rise towardthe surface, the decrease in pressure causes theirsaturation in olivine. This olivine is unusually Nirich and Mn and Ca poor. With use of a new,large data set of high-precision analyses of
olivine phenocrysts from OIBs, LIPs, MORBs,and komatiites, we show that hybrid pyroxeniteis a common lithology in upwelling mantle and amajor contributor to tholeiitic (silica-saturated)and transitional (moderately silica-undersaturated)magmas of OIBs and LIPs emplaced on thickoceanic or continental lithosphere.
Olivine data set. We use olivine phenocrystsas probes of parental melt composition, becauseolivine is the first phase to precipitate at lowpressures in almost all mantle-derived magmasand because its forsterite content is an excellentmeasure of the degree of fractional crystallizationallowing reconstruction of the parental meltcomposition.
Olivine phenocrysts were analyzed by elec-tron microprobe using high probe currents andlong counting times (10). This procedure rou-tinely yields detection limits of around 6 to 15parts per million (ppm) and errors (2 standarderrors) of 15 to 30 ppm for trace elements (Ni,Ca, Mn, Cr, Co, and Al) and 0.01 mole percent(mol %) for forsterite content [defined as Fo =Mg/(Mg + Fe)], checked by repeated analysis ofSan Carlos olivine standard (11). In the followingdiagrams we use only high-precision data.
We have analyzed nearly 17,000 grains ofolivine phenocrysts from 229 samples of tholei-itic to transitional compositions coveringMORBs(40 samples) from Mid-Atlantic Ridge, EastPacific Rise, South-East Indian Ridge, andKnipovich Ridge; OIBs (138 samples) fromHawaiian Islands and Emperor Seamounts,Canary Islands, Reunion, Azores, and Iceland;LIPs (36 samples) from Ontong Java Plateau,Siberia, Emeishan, Karoo, Afar, and NorthAtlantic Province; komatiites and associatedpicrites (15 samples) from the Archean Abitibigreenstone belt in Canada and the Belingwe beltin Zimbabwe and South Africa; Proterozoickomatiitic basalts from Gilmour Island, Canada;and komatiites and picrites from Gorgona Island,Colombia. Most samples are picrites or olivinebasalts containing large amounts of fresh, high-magnesium olivine phenocrysts. The samples aresubdivided into four groups: (i) MORB; (ii)within plate magmas (WPM, magmas eruptedfar from plate boundaries) forming OIBemplaced over thin lithosphere (<70 km thick),WPM-THIN; (iii) WPM (OIB and LIP)emplaced over thick lithosphere (>70 km thick),WPM-THICK; and (iv) komatiites and asso-ciated magmas, KOMATIITES. Details of sam-ple locations, references for sample descriptions,and their group correspondence are presented intable S2a.
The most-magnesian olivine compositions(defined by olivines phenocrysts with Fo within1 mol % from amaximum Fo) for each specimenwere averaged (table S2a) for the plots shown inFig. 1. Individual olivine analyses are presentedon fig. S4 and tables S2, c to f.
In addition to Mn and Ni concentrations,which strongly correlate with Fo (Fig. 1, A and
RESEARCHARTICLES
1Max Planck Institute (MPI) for Chemistry, Post Office Box3060, 55020 Mainz, Germany. 2Vernadsky Institute ofGeochemistry and Analytical Chemistry, Russian Academyof Sciences, Kosygin Street 19, 119991 Moscow, Russia.3Institute of Geology and Mineralogy, Siberian Branch ofRussian Academy of Sciences, Koptuyga prospekt 3,630090 Novosibirsk, Russia. 4Research School of EarthSciences, Australian National University, Canberra, ACT0200 Australia. 5Laboratoire de Géodynamique desChaînes Alpines, Université de Grenoble, 38401 Grenoblecedex, France. 6Department of Geosciences, National TaiwanUniversity, Post Office Box 13-318, Taipei 106, Taiwan.7Australian Research Council, Centre of Excellence in OreDeposits and School of Earth Sciences, University ofTasmania, Private Bag 79, Hobart, Tasmania, 7001, Aus-tralia. 8Department of Earth Sciences, Queen’s Road, WillsMemorial Building, University of Bristol, Bristol BS8 1RJ,UK. 9Department of Earth, Atmospheric, and PlanetaryScience, Massachusetts Institute of Technology (MIT), 77Massachusetts Avenue, Cambridge, MA 02139, USA.10Department of Geology and Geophysics, University ofHawaii, 1680 East-West Road, Honolulu, HI 96822, USA.11School of Earth, Ocean and Planetary Sciences, CardiffUniversity, Main Building, Park Place, Cardiff CF10 3YE,UK. 12P. P. Shirshov Institute of Oceanology of RussianAcademy of Sciences, Nakhimovsky prospekt 36, 117997Moscow, Russia. 13Department of Petrology, Faculty ofGeosciences Utrecht University, Budapestlaan 4, Utrecht,Netherlands. 14Department of Petrology, Faculty of Earthand Life Sciences, Vrije Universiteit, De Boelelaan 1085,1081 HV Amsterdam, Netherlands. 15Department of Earthand Environmental Sciences, University Munich, 80333Munich, Germany. 16South Iceland Nature Centre, Strandvegur50, Vestmannaeyjar, IS 900, Iceland. 17Department of EarthScience, University of Asmara, Asmara, Eritrea.
*To whom correspondence should be addressed. E-mail:[email protected]
20 APRIL 2007 VOL 316 SCIENCE www.sciencemag.org412
on
Apr
il 19
, 200
7 w
ww
.sci
ence
mag
.org
Dow
nloa
ded
from
C), we also plot Mn/Fe and Ni versus Mg/Feratios (Fig. 1, B and D). These ratios do not varysignificantly with olivine fractionation (seemodel curves Frac 1 and Frac 2) but neverthelessrange considerably (Fig. 1, B and D). Mostolivine phenocrysts from MORBs and manyfrom komatiites have Mn and Ni contents similarto those of peridotite-derived melts. In contrast,most olivines from the WPM-THICK group aresignificantly depleted in Mn and enriched in Ni.Their concentrations are not compatible with themelting products of common peridotites. The ol-ivines from the WPM-THIN group have inter-mediate Mn and Ni contents.
Concentrations of Ca also provide somediscrimination in spite of the greater overlap.Most olivines from the WPM-THICK group aretoo low in Ca to have precipitated from peridotite-derived melts (shown as experimental-basedmodel compositions and fractionation trajecto-ries, Fig. 1E).
Chromium is strongly controlled by garnetand spinel in peridotites and thus might be usefulto decipher products of high-degree melting of
peridotite, which leave residuals (restites) free ofCr-rich phases (12). Olivines from Archeankomatiites have the highest Cr values and matchcompositions of olivines from a spinel- andgarnet-free refractory restite (Fig. 1F). Theycould, therefore, be derived directly from high-degree melting of peridotite. In the other groupsof olivines, Cr is markedly lower than expectedin equilibrium with peridotite at high pressures(see experimental data on lherzolite melting, Fig.1F). The lowest Cr contents are found in MORBolivines, indicative of residual Cr spinel.
Cobalt (Fig. 2A) shows nearly uniform cor-relation with Fo for all rock groups, with possiblyonly minor (around 5%) relative enrichment inWPM-THICK and WPM-THIN over MORB(estimated from group average Co/Fe of tableS2a). Decoupling of Co andNi yieldsNi/Co ratiosof many WPM-THICK olivines that are un-usually high for the Earth mantle (Fig. 2B).
Mn/Fe is the parameter least dependent onolivine fractionation (as shown by the modelfractionation curves in Fig. 3). Thus, it is diag-nostic of parental magma compositional differ-
ences. There is a significant negative correlationof Ni/Mg versus Mn/Fe (linear correlation coef-ficient r is 0.66 for 238 samples) in spite of strongdependence of Ni/Mg on the degree of olivinefractionation (see fractionation trajectories inFig. 3A). This correlation improves (r = 0.88 for103 samples) for the subset of olivines with anarrower range of Fo contents (Fo89 to Fo91).MORB olivines are the lowest in Ni and highestinMn, whereas olivines from theWPM-THICKgroup are the highest in Ni and lowest in Mn,with olivines from theWPM-THIN group beingintermediate.
To minimize the effects of olivine fractiona-tion, we show parameters Ni/(Fe/Mg) and Ca/Fein Fig. 3, B and C. This procedure also reducesthe scatter in the ordinate significantly, thushighlighting the differences between geodynamicsettings.
Fate of recycled oceanic crust. In subductionat P > 2.5 GPa, the basaltic and gabbroic portionsof the oceanic crust are transformed completelyto eclogite (clinopyroxene and garnet) with a freeSiO2 phase (13–15). Unless silica has been
Fig. 1. (A to F) Average compositionsof the most highly magnesian olivinephenocrysts in each sample. Concen-trations and their ratios are given inppm versus forsterite content of olivinein mol %. Olivine group names are asdefined in text. PM, Herzberg indicatescompositions of olivine in equilibriumat 0.1 MPa with melt originally gen-erated at 3.0 to 5.0 GPa from fertileperidotite (12), calculated by Petrologsoftware (41) for oxygen fugacity corre-sponding to quartz-fayalite-magnetite(QFM) buffer using the Herzberg model(4), PM, Kinzler, olivine compositionssimilar to PM, Herzberg but with Nicalculated by using Ni partitioning be-tween olivine and melt from Kinzleret al. (28). Frac 1 is the trend of olivinecomposition during fractional crystalli-zation from a melt derived from fertileperidotite at 3 GPa and 1515°C (12).Fractionation of olivine modeled up to20% for oxygen fugacity correspond-ing to QFM buffer using the Herzbergmodel (4). Frac 2 is similar to Frac 1 butcalculated for 35% crystallization ofmelt derived at 4.0 GPa and 1630°C(12). Green ellipse indicates field ofolivine compositions compatible withperidotitic source. In (F), HRZ, Herzbergstands for calculated compositions ofolivine from spinel- and garnet-freeharzburgite restite using (4); LHRZ,Walter and HRZ, Walter indicate exper-imental olivines from lherzolite- andgarnet-free harzburgite residual assem-blages, respectively, produced by high-pressuremelting of fertile peridotite (12).Black ellipse marked “ol + low Ca-Px”indicates field of olivine compositions from refractory garnet- and spinel-free assemblage of olivine and low-Ca pyroxene.
800
1000
1200
1400
1600
86 87 88 89 90 91 92 93 94
Mn
MORB
WPM-THIN
WPM-THICK
KOMATIITES
Frac-1
Frac-2
PM, Herzberg
A
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
86 87 88 89 90 91 92 93 94
100M
n/F
e
B
1000
1500
2000
2500
3000
3500
4000
4500
5000
86 87 88 89 90 91 92 93 94
Ni
PM, Kinzler
C
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
86 87 88 89 90 91 92 93 94
Ni/(
Mg
/Fe)
/100
0
D
0
500
1000
1500
2000
86 87 88 89 90 91 92 93 94
Fo mol%
Cr
Walter HRZ
Walter LRZ
Herzberg HRZ
OL+Low-Ca Px
F
1000
1500
2000
2500
3000
86 87 88 89 90 91 92 93 94
Fo mol%
Ca
E
www.sciencemag.org SCIENCE VOL 316 20 APRIL 2007 413
RESEARCH ARTICLES
on
Apr
il 19
, 200
7 w
ww
.sci
ence
mag
.org
Dow
nloa
ded
from
removed during subduction (16), this combina-tion will also be the relevant assemblage duringrecycling to the upper mantle (17).
In the ascending mantle (e.g., a mantle plumeor upstream flow of convecting mantle), thesilica-oversaturated eclogite starts melting athigher pressures than the peridotite and produceshigh silica melt (18, 19). This melt reacts witholivine from peridotite, producing pyroxenes andgarnet (5, 8, 19). Previous studies have envi-sioned that this reaction creates a refertilizedperidotite enriched in pyroxene (19, 20). Thisconclusion would predict variable mixingproportions of individual ingredients (eclogite-derived high-Si melt and peridotite) that are
drastically different in composition. Melting suchvariable source compositionswould create highlynonlinear correlations of 187Os/188Os and 87Sr/86Srisotope ratios in the melts, and this is contradictedby the strongly linear correlations observed inHawaiian basalts (21, 22), which are thought tohave a significant eclogite component (3, 5).
However, it has been shown experimentally(23) and proposed on the basis of Korzinskii’stheory (24) that, under conditions of local equilib-rium, the reaction between high-Si eclogite-derivedmelt and peridotite produces an olivine-free lithol-ogy enriched in pyroxene (5). This fundamentallydiffers from a partial reaction (19, 20) because itleads to a stable pyroxenite lithology (hybrid py-
roxenite) generated by roughly fixed proportionsof high-Si melt and peridotite [constrained byreaction stoichiometry between 40 and 60weight % (wt %) of melt (5)] irrespective of theinitial proportions of the reaction ingredients.Consequently, the hybrid pyroxenite has nearlyuniform chemical and isotopic composition,thus constituting a single mixing endmember.Binary mixing of melts derived from peridotiteand this pyroxenite leads to near-linear 87Sr/86Srversus 87Os/188Os trends (5).
Other predicted geochemical consequencesof replacement of olivine by pyroxene are a sig-nificant decrease of the bulk distribution co-efficient between crystals and melt (Kd) for Ni
Fig. 2. (A and B) Cobaltand nickel to cobalt ratioversus Fo of average Mg-rich olivine phenocrysts.Pink band at Ni/Co 20 ±1 represents estimatedvalues for bulk silicateEarth (BSE), core, andchondrites (39). Arrowsindicate trend of olivinecompositions due to themantle melting (melting)and magma crystalliza-tion (cryst). All other sym-bols are as on Fig. 1.
100
110
120
130
140
150
160
170
180
190
200
86 87 88 89 90 91 92 93
Co
pp
m
A
10
15
20
25
30
35
86 87 88 89 90 91 92 93
Fo mol%Fo mol%
Ni/C
o
BSE, Core, Chondrites
B
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
2.6
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
Ni/(
Mg
/Fe)
/100
0
PYR, HERZBERG
PYR, KINZLER
MIX, HERZBERG
MIX, KINZLER
PYROXENITE
PERIDOTITE
B
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
2.6
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
100N
i/Mg PYROXENITE
PERIDOTITE
A
0
1
2
3
4
5
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
100Mn/Fe100Mn/Fe
100Mn/Fe
100C
a/F
e PYROXENITE
PERIDOTITE
C
C
Fig. 3. (A to C) Composition of average Mg-rich olivine phenocrysts andequilibrium olivines from peridotite- and pyroxenite-derived melts.Elements and their ratios are shown in ppm. PYR, Herzberg and PYR,Kinzler stand for compositions of olivine equilibrium at 0.1 MPa withmelt generated at 3.5 GPa by melting of pyroxenite (table S3) calculatedfor oxygen fugacity corresponding to QFM buffer using model (4) for Fe-Mgpartitions and models (4, 28) for Ni partition, respectively. MIX Herzberg andMIX Kinzler correspond to equilibrium olivine compositions calculated for themixture of pyroxenite-derived and peridotite-derived melts using model (4)for Fe-Mg partitions and models (4, 28) for Ni partition, respectively. Blackellipse indicates olivine compositions corresponding to pyroxenitic source. Allother symbols are as on Fig. 1.
20 APRIL 2007 VOL 316 SCIENCE www.sciencemag.org414
RESEARCH ARTICLES
on
Apr
il 19
, 200
7 w
ww
.sci
ence
mag
.org
Dow
nloa
ded
from
(5, 6) and a decrease in the ratio of the bulkcoefficients of Mn and Fe (7). These featuresoccur because olivine is the major silicate phasein peridotite concentrating Ni and the only sili-cate phase in peridotite having Kd for Fe greaterthan Kd for Mn (7). The bulk Kd changes willincreaseNi and lower theMn/Fe ratio of pyroxenite-derived melt compared with peridotite-derivedmelts. In addition, melting of pyroxenite yieldslower Ca compared with peridotite (8). Addi-tional predicted differences are higher meltfractions for hybrid pyroxenite than peridotiteand higher Si and lower Mg in pyroxenite-derived melt (5, 25, 26).
These predictions were tested by experimen-tal melting of a model hybrid pyroxenite (5).Experiments were run at P = 3.5 GPa andtemperatures between 1400° and 1570°C in aconventional 1.27-cm piston-cylinder apparatusat the Australian National University (10, 19).These results, together with published experi-mental data for melting of peridotite (12),confirm the Ni and Mn relationships as well asmelting rates predicted above (table S3 and fig.S5). From these data, we calculate equilibriumolivine compositions at low pressure (4, 27, 28)in order to compare them with the naturalphenocryst data (Fig. 3). We used these resultsto estimate mixing proportions of melts derivedfrom the two end-member sources for the olivinedata sets representing different geodynamicsettings. The end-member melt compositionswere calculated from averaging experimentaldata on melting of pyroxenite and peridotite (10).
Quantitative estimates. We assumed mix-tures (in 10% intervals) of the end-member melts(10) and calculated the composition of equilibri-um olivines. The calculated mixing trajectoriesfor the two different models for Ni partitioning
are consistent with natural olivine data (Fig. 3B).The relation between Mn/Fe of modeled olivinesand mixing proportions (10) was used to com-pute the amount of pyroxenite-derived compo-nent for individual samples (Fig. 4). Olivinesfrom the WPM-THICK group of basalts yield anaverage of 61 ± 16% (standard deviation)pyroxenite-derived component, similar to resultsderived from Ni contents in Hawaiian meltinclusions and olivines only (5). The olivinesfrom some continental LIPs (specific suits fromSiberia and Karoo) indicate almost pure pyrox-enitic sources. Corresponding results for theother groups are for WPM-THIN, 30 ± 13%;for Archean komatiites, 21 ± 10%; and forMORB, excluding one unusual sample from theSouthern Atlantic (see below), 17 ± 12% [similarto predictions of (25)]. Because of the uncertain-ties involved in estimating the end-membercompositions, the differences between groupsare better constrained than the absolute numbers.Although MORBs contain the lowest proportionof pyroxenite-derived melt, the spread of MORBdata is significant, and many samples do containsubstantial amounts of pyroxenite-derived com-ponent [the extreme is the enriched in silicaMORB sample from the Southern Atlantic (29)with 100% pyroxenite-derived component]. Thecalculations show that the Archean komatiitescontain a significant amount of pyroxeniticcomponent (maximum of 30% for samples fromCanada and Belingwe), although the largestamount is in Proterozoic komatiitic basalts fromGilmour Island, Canada (up to almost 40%).From these calculations, an estimate of theamount of recycled oceanic crust (10) yields4% for MORB, 11% for WPM-THIN group,16% for WPM-THICK group, and around 13%for Archean komatiites. The highest estimate of
the amount of recycled oceanic crust (10) yieldOntong Java high-Mg lavas: 13% to 28%.
Silica-undersaturated basalts. Most of themagmas analyzed in this study are silica saturated(tholeiites or transitional). Only a few samples aremoderately silica-undersaturated alkali magmas(e.g., Azores and Afar). Our database is repre-sentative of the normal oceanic crust, several ofthe world’s major suites of flood basalts (LIPs),several of the major modern mantle plumes (30),and some komatiites. The strongly silica-undersaturated associations not covered hereinclude continental rift basalts, many smallerocean islands consisting mostly of alkali lavas,and also some larger-flux plumes (30) such asPitcairn, Tahiti, and Cape Verde islands. Why aresuch basalts that are highly enriched in in-compatible elements, and therefore presumablygenerated by very low degrees of melting, nearlyalways undersaturated in silica? This observationappears to contradict our model; one wouldexpect that silica-saturated melts generated fromhybrid pyroxenites should be prevalent, especial-ly at very low melt fractions. There are severalpossible explanations. (i) Avolatile (mostly CO2-)triggered melting of peridotite may be thedominant mechanism forming strongly silica-undersaturated alkaline magmas at temperatureslower than hybrid pyroxenite melts (31). (ii)Low-degree melts of silica-saturated eclogitemay be retained in the source because of theirhigh viscosity, thus preventing production of thehybrid pyroxenite (5). (iii) Melting of hybridpyroxenite at the contact with peridotite mayproduce low-degree, silica-undersaturated meltsat lower temperatures than melting of hybridpyroxenite itself (8). (iv) Melting of bimineraliceclogites (no free silica phase) formed fromsilica-undersaturated recycled crust producesundersaturated alkaline magmas (16).
What controls the amount of pyroxenite-derived melt? By following the method outlinedabove, we estimated the proportions of meltderived from pyroxenite and peridotite for eachparental magma. These proportions depend onseveral interrelated parameters, namely the thick-ness of lithosphere, the potential mantle temper-ature (Tp) (32), and the amount of recycled crustin the upwelling mantle (Fig. 5). Because at thesame Tp pyroxenite melts at higher pressure thanperidotite (26), a thick lithosphere (which im-poses a high lower limit on the depth of melting)will suppress low-depth peridotite melting andtherefore favor a high proportion of pyroxenite-derived melts (33, 34). The extreme case is foundin some continental flood basalts (specific suitesof Siberia and Karoo at table S2a) where theamount of pyroxenite-derived melt is nearly100%. In such a case, the amount of recycledmaterial cannot be estimated because the perido-titic component contributes no melt. In contrast, athin lithosphere (MORB, Iceland, Azores, andDetroit seamount) favors a higher proportion ofperidotite-derived melt because of the increasingdegree ofmelting of peridotite at shallower depths.
KOMATIITES
0
1
2
3
4
5
Fraction of pyroxenite derived meltFraction of pyroxenite derived melt
Fraction of pyroxenite derived melt Fraction of pyroxenite derived melt
Nu
mb
er o
f sa
mp
les
WPM-THICK
0
5
10
15
20
25
30
35
Nu
mb
er o
f sa
mp
les
WPM-THIN
0
1
2
3
4
5
6
7
8
Nu
mb
er o
f sa
mp
les
MORB
0
2
4
6
8
10
12
0 0.1 0.2 0.3 0.6 0.7 0.8 0.9 10.2 0.3 0.4 0.5 0.4 0.5
0 0.1 0.2 0.3 0.6 0.7 0.8 0.9 10.4 0.50 0.1 0.2 0.3 0.6 0.7 0.8 0.9 10.4 0.5
0.6 0.7 0.8 0.9 1
Nu
mb
er o
f sa
mp
les
0.10
Fig. 4. Estimated amounts of pyroxenite-derived component in the parental melt for 229 samplesof four different groups.
www.sciencemag.org SCIENCE VOL 316 20 APRIL 2007 415
RESEARCH ARTICLES
on
Apr
il 19
, 200
7 w
ww
.sci
ence
mag
.org
Dow
nloa
ded
from
A high Tp is an important condition to main-tain sufficient buoyancy of mantle plumes or anyother upstream mantle flow, and this buoyancylimits the amount of dense eclogite that they cancarry (35). Also, a high Tp affects mostly theproportion of peridotite-derived melt becausefractional melting imposes a rather stringent up-per limit to further melting at high melt fractions(36). High melt fractions are restricted to theeclogite and hybrid pyroxenite assemblages (26)(fig. S5). The peridotite assemblage produceslower melt fractions than pyroxenite (fig. S5) oreclogite (19, 25) at any given temperature andpressure, and its actual extent ofmelting thereforedepends strongly on the specific Tp.
Lastly, why are the proportions of recycledcomponent lower beneath mid-ocean ridges thanin thick-lithosphere settings? We suggest severalexplanations: (i) Relatively low amount of denserecycled component inMORBs is limited by theirTp, which is too low to carry more. (ii) For sta-tistical reasons, plumes are more likely to en-counter more-common thick lithosphere thanless-common thin lithosphere and few plumesimpinge directly on ridges, so we are forced todeal with very-small-number statistics. (iii) De-troit seamount represents one case where a(Hawaiian) plume has encountered thin litho-sphere and where our results do indicate a highfraction of recycled crust, similar to those foundon the island of Hawaii on the thick lithosphere.
This amount is significantly higher than for Ice-land, probably reflecting the effect of a higher Tpof the Hawaiian plume (4). (iv) The surfaceexpression of a plume emplaced under thicklithosphere requires high Tp, which is necessaryfor carrying a significant amount of recycledcrust (35), allowing melts to form at higher pres-sures than for ordinary peridotite (5, 25, 26),and melting a peridotite at higher pressures [e.g.,komatiites (37)].
Heterogeneous versus homogeneousmantle.The model presented here assumes that therecycled crustal component was not fully mixedwith peridotite during subduction and mantleconvection and thus that the formation of theolivine-free hybrid lithology may take place. Onthe other hand, homogenization of crustal ma-terial within the peridotite mantle should create arange of ultramafic lithologies with variableamounts of olivine, similar to a model byKelemen et al. (20). Under these circumstances,the major-element contents of partial melts willcorrespond to the eutectic-like composition, buf-fered by the peridotite assemblage, whereas thecompatible trace elements (Ni and Mn) will becontrolled by the bulk partition coefficients ofthis assemblage and thus by the amount of ol-ivine and pyroxene present in the system. There-fore, the amount of recycled crust can still beestimated on the basis of these trace elements, buttheir abundances will no longer correlate with the
buffered major elements (Si, Ca, and Al). ForHawaiian basalts, such correlations with Ni arepresent (5), which requires a strongly heteroge-neous source.
Input from Earth’s core? The Ni excess inmantle olivines from Siberian LIP (38) and theelevated Fe/Mn ratios in Hawaiian lavas (7) havebeen explained by input from Earth’s core to thesources of mantle plumes. This suggestion isconsistent with 186Os/188Os ratios for someHawaiian and Gorgona lavas (39) but is contra-dicted by the fact that concentrations of highlysiderofile (Pt) and chalcophile (Cu) elementsreported for Hawaiian basalts are not affected bythis process (5). Our olivine data provide strongarguments against any notable core contribution toNi or Fe excess in the sources of mantle-derivedmagmas. Cobalt does not show significant excessin olivines (Fig. 2 and fig. S3) and is effectivelydecoupled fromNi. As a result, the Ni/Co ratio inmost Ni-rich mantle plume olivines exceeds 30 atthe typical mantle Fo range of 89 to 91 (Fig. 2).This is not expected from a core contribution,because Ni/Co ratios for both mantle and core arealmost equal and close to the chondritic value ofabout 20 (40). In addition, Ca is significantlydepleted in many high-Ni and low-Mn olivinesfrom the WPM-THICK group, which cannot beexplained by core contribution. Lastly, theolivines from Gorgona komatiites, which doshow significant excess in 186Os (39), do notindicate large anomalies in Ni and Mn/Fe,whereas Koolau lavas with the highest Niexcess and lowest Mn/Fe ratio in olivines donot show significant elevations in 186Os/188Osratios (39). This suggests complete decouplingof these potentially strong indicators of core-mantle exchange.
References and Notes1. A. W. Hofmann, W. M. White, Earth Planet. Sci. Lett. 57,
421 (1982).2. A. W. Hofmann, in The Mantle and Core, vol. 2 of Treatise
on Geochemistry, H. D. Holland, K. K. Turekian, Eds.(Elsevier, Amsterdam, 2003), p. 61.
3. E. H. Hauri, Nature 382, 415 (1996).4. C. Herzberg, M. J. O'Hara, J. Petrol. 43, 1857 (2002).5. A. V. Sobolev, A. W. Hofmann, S. V. Sobolev,
I. K. Nikogosian, Nature 434, 590 (2005).6. Ni was first proposed as a monitor for replacement of
olivine by pyroxene in (20).7. M. Humayun, L. P. Qin, M. D. Norman, Science 306, 91
(2004).8. C. Herzberg, Nature 444, 605 (2006).9. S. M. Eggins, Contrib. Mineral. Petrol. 110, 387 (1992).
10. Materials and methods are available on Science Online.11. E. J. Jarosevich, J. A. Nelen, J. A. Norberg, Geostand
Newsl. 4, 43 (1980).12. M. J. Walter, J. Petrol. 39, 29 (1998).13. V. S. Sobolev, A. V. Sobolev, Dokl. Akad. Nauk SSSR 237,
437 (1977).14. T. Kogiso, M. M. Hirschmann, D. J. Frost, Earth Planet.
Sci. Lett. 216, 603 (2003).15. T. Kogiso, M. M. Hirschmann, M. Pertermann, J. Petrol.
45, 2407 (2004).16. T. Kogiso, M. Hirschmann, Earth Planet. Sci. Lett. 249,
188 (2006).17. A. Yasuda, T. Fujii, K. Kurita, J. Geophys. Res. 99, 9401
(1994).18. A. E. Ringwood, D. H. Green, Tectonophysics 3, 383
(1966).
Xpx ~ 10-15%Xrc ~ 5 %
Tp~1350oC
Siberian LIP Hawaii Detroit Smt MORB
per
ido
tite
+ e
clo
git
ep
yro
xen
ite
mel
ts
per
ido
tite
mel
ts
Xpx ~ 100%Xrc ~ 20 %
Tp~1550oC
Xpx ~ 60%Xrc ~ 20 %
Tp~1550oC
Xpx ~ 40%X rc~ 20 %Tp~1550oC
Iceland
LITHOSPHERE
Xpx ~ 20%X rc~ 10 %
Tp~1450oC
LITHOSPHERELITHOSPHERE
LITHOSPHERE
eclo
git
e m
elts
Fig. 5. Schematic model illustrating interplay between amount of recycled crust, thickness oflithosphere, and Tp. Blue, upwelling peridotitic mantle; red, recycled oceanic crust (eclogite withfree SiO2 phase); black dots, melting; yellow, reaction zones forming hybrid pyroxenite; pink,refractory restite after eclogite melting; and green, lithosphere. Xpx, amount of pyroxenite derivedmelt in the mixture with peridotite-derived melt, and Xrc, amount of recycled oceanic crust in theperidotitic mantle (42).
20 APRIL 2007 VOL 316 SCIENCE www.sciencemag.org416
RESEARCH ARTICLES
on
Apr
il 19
, 200
7 w
ww
.sci
ence
mag
.org
Dow
nloa
ded
from
19. G. M. Yaxley, D. H. Green, Schweiz. Mineral. Petrogr.Mitt. 78, 243 (1998).
20. P. B. Kelemen, S. R. Hart, S. Bernstein, Earth Planet.Sci. Lett. 164, 387 (1998).
21. E. H. Hauri, M. D. Kurz, Earth Planet. Sci. Lett. 153, 21(1997).
22. J. C. Lassiter, E. H. Hauri, Earth Planet. Sci. Lett. 164,483 (1998).
23. R. P. Rapp, N. Shimizu, M. D. Norman, G. S. Applegate,Chem. Geol. 160, 335 (1999).
24. D. S. Korzhinskii, Theory of Metasomatic Zoning(Clarendon, Oxford, 1970).
25. M. M. Hirschmann, E. M. Stolper, Contrib. Mineral. Petrol.124, 185 (1996).
26. G. M. Yaxley, Contrib. Mineral. Petrol. 139, 326 (2000).27. P. Beattie, Contrib. Mineral. Petrol. 115, 103 (1993).28. R. J. Kinzler, T. L. Grove, S. I. Recca, Geochim.
Cosmochim. Acta 54, 1255 (1990).29. V. S. Kamenetsky et al., Geology 29, 243 (2001).30. N. H. Sleep, Annu. Rev. Earth Planet. Sci. 20, 19 (1992).31. G. H. Gudfinnsson, D. C. Presnall, J. Petrol. 46, 1645 (2005).32. D. McKenzie, M. J. Bickle, J. Phys. Earth 38, 511 (1990).33. J. P. Morgan, Geochem. Geophys. Geosystems 2,
2000GC000049 (2001).34. G. Ito, J. J. Mahoney, Earth Planet. Sci. Lett. 230, 29 (2005).
35. M. Pertermann, M. M. Hirschmann, J. Petrol. 44, 2173(2003).
36. I. Kushiro, Annu. Rev. Earth Planet. Sci. 29, 71(2001).
37. N. Arndt, J. Geophys. Res. 108, XX (2003).38. I. D. Ryabchikov, Dokl. Earth Sci. 389, 437 (2003).39. A. D. Brandon, R. J. Walker, Earth Planet. Sci. Lett. 232,
211 (2005).40. W. F. McDonough, S. S. Sun, Chem. Geol. 120, 223
(1995).41. L. V. Danyushevsky, J. Volcanol. Geotherm. Res. 110, 265
(2001).42. D. McKenzie, M. J. Bickle, J. Petrol. 29, 625 (1988).43. We thank B. Schulz-Dobrick for supervising purchase and
establishing electron microprobe facility in MPI; HawaiianScientific Drilling Project team, Koolau Scientific DrillingProject team, Ocean Drilling Program team, A. T. Anderson,E. A. Mathez, and N. Gitahi for providing samples;E. J. Jarosevich for supplying microprobe standards;A. Yasevich for sample preparations and N. Groschopffor maintaining the electron microprobe. The paperbenefited from discussions with M. Hirschmann,P. Kelemen, C. Herzberg, B. McDonough, I. Ryabchikov,L. Kogarko, A. Kadik, E. Galimov, V. Batanova. Constructivereviews of C. Herzberg, and P. Kelemen improved the
clarity of the manuscript. The study was supported byWolfgang Paul Award of the Alexander von HumboldtFoundation, Germany, to A.V.S. Partial support from thefollowing sources is also acknowledged: RussianFoundation for Basic Research (RFBR) grants 06-05-65234and НШ-4264.2006.5 and Russian Academy of Sciencesand Deutsche Forschungsgemeinschaft grant HO1026/16-1 to A.V.S., RFBR grant 06-05-64651 to N.M.S.,NSF grants EAR03-36874 to M.O.G. and EAR-0105557to F.A.F., and Netherlands Research Center for IntegratedSolid Earth Science grant 6.2.12 to I.K.N. This is Schoolof Ocean and Earth Science and Technology, Universityof Hawaii, contribution no. 7104.
Supporting Online Materialwww.sciencemag.org/cgi/content/full/1138113/DC1SOM TextFigs. S1 to S5Tables S1 to S4Data set
29 November 2006; accepted 16 March 2007Published online 29 March 2007;10.1126/science.1138113Include this information when citing this paper.
Genes Required for Mitotic SpindleAssembly in Drosophila S2 CellsGohta Goshima,1,3* Roy Wollman,2,3 Sarah S. Goodwin,1 Nan Zhang,1Jonathan M. Scholey,2 Ronald D. Vale,1,3† Nico Stuurman1,3
The formation of a metaphase spindle, a bipolar microtubule array with centrally alignedchromosomes, is a prerequisite for the faithful segregation of a cell’s genetic material. Using afull-genome RNA interference screen of Drosophila S2 cells, we identified about 200 genes thatcontribute to spindle assembly, more than half of which were unexpected. The screen, incombination with a variety of secondary assays, led to new insights into how spindle microtubulesare generated; how centrosomes are positioned; and how centrioles, centrosomes, andkinetochores are assembled.
The diamond-shaped mitotic spindle hasbecome one of the most widely recog-nized images in biology, emblematic of
life’s propagation through cell division. In highereukaryotes, the process of spindle formationbegins after nuclear envelope breakdown (NEB)when microtubules (MTs), generated both fromcentrosomes and from the vicinity of chromatin,are organized into a bipolar array (1–5). Sisterchromatids bind toMTs emanating from oppositepoles, are aligned in the middle of the bipolarMTnetwork, and then ultimately separate and moveapart during anaphase. Failures early in mitosisresult in the formation of an abnormal metaphase
spindle, which can lead to mitotic delay and,potentially, chromosome missegregation duringthe ensuing anaphase.
To understand the mechanism of metaphasespindle assembly, it is critical to identify the pro-teins required for this process and then dissecthow they function. Many mitotic proteins havebeen identified through genetic and RNAi screens(6–10), but the inventory is likely incomplete.Here, we present a genome-wide screen for mi-totic spindle morphology in Drosophila S2 cellsand the functional analysis of unexpected genesdiscovered through the screen.
Identification of genes involved in meta-phase spindle formation by high-throughputmicroscopy. Because the percentage of S2 cellsin mitosis is low (~1%), we conducted ourRNAi screen in the presence of dsRNA (double-stranded RNA) to Cdc27 (a subunit of theanaphase-promoting complex) to delay ana-phase and thereby increase the percentage ofmetaphase cells (~10% of the population). Thus,our screen was designed to investigate the as-sembly of the metaphase spindle, but not ana-phase or cytokinesis. We also rescreened the
final hits without Cdc27 RNAi–induced mitot-ic arrest. The majority (88%) showed identicalphenotypes, although a few genes only manifestclear phenotypes under conditions of mitoticarrest (table S1).
Using our custom, full-genome (14,425 genes)Drosophila RNAi library (11), we treated S2cells with dsRNA for 4 days, conditions thatgenerally reduce protein levels by >80% (12, 13).After dsRNA treatment, cells were fixed andstained for DNA, g-tubulin, MT, and phospho-histone H3 (pH3) in 96 well plates, and about 40sites per well were imaged by automated micros-copy with a high numerical aperture air ob-jective to obtain relatively high-resolution imagesof mitotic spindles (Fig. 1A). To reduce the com-plexity of this large amount of image data, a cus-tom computer code was used to identify, crop,and arrange mitotic spindles into galleries, whichwere then blindly scored by an observer (Fig. 1Band fig. S1). In addition, computer algorithmsmeasured eight parameters of spindle shape, aswell as the intensity of g-tubulin, cell number, andmitotic index (Fig. 1C) (11). More than 4,000,000spindles were analyzed in this screen.
Before beginning this screen, we annotated49 genes that produce mitotic defects in S2 cells(table S2). Of these 49 genes, 45 were identifiedas hits in the primary screen, indicating a highsuccess rate of identifying mitotic phenotypes.However, our final list of genes should not beconsidered as a complete or universal inventory,because genes can be missed (particularly thosewith subtle phenotypes), and some phenotypes(or lack thereof) may be specific to S2 cells.False positives by off-target effects of dsRNAcan occur in RNAi screens (14, 15), so pre-cautions were taken to minimize gene overlap inthe dsRNA design, and all unexpected hits wereconfirmed with another dsRNA that had nooverlap with the first dsRNA (11). To learnmore about the functions of interesting genes,
1Howard Hughes Medical Institute and the Department ofCellular and Molecular Pharmacology, University of Cal-ifornia, San Francisco, 600 16th Street, San Francisco, CA94158, USA. 2Department of Molecular and Cellular Biol-ogy, University of California, Davis, One Shields Avenue,3203 Life Sciences, Davis, CA 95616, USA. 3PhysiologyCourse, Marine Biological Laboratory, 7 MBL Street, WoodsHole, MA 02543, USA.
*Present address: Institute for Advanced Research, NagoyaUniversity, Furo-cho, Chikusa-ku, Nagoya, 464-8601, Japan.†To whom correspondence should be addressed. E-mail:[email protected]
www.sciencemag.org SCIENCE VOL 316 20 APRIL 2007 417
RESEARCH ARTICLES
on
Apr
il 19
, 200
7 w
ww
.sci
ence
mag
.org
Dow
nloa
ded
from
Corrected 19 April 2007: Minor typographic erros have been corrected.
www.sciencemag.org/cgi/content/full/1138113/DC1
Supporting Online Material for
The Amount of Recycled Crust in Sources of Mantle-Derived Melts
Alexander V. Sobolev, Albrecht W. Hofmann, Dmitry V. Kuzmin, Gregory M. Yaxley, Nicholas T. Arndt, Sun-Lin Chung, Leonid V. Danyushevsky, Tim Elliott, Frederick A. Frey, Michael O. Garcia, Andrey A.Gurenko, Vadim S. Kamenetsky, Andrew C. Kerr, Nadezhda A. Krivolutskaya, Vladimir V. Matvienkov, Igor K. Nikogosian, Alexander
Rocholl, Ingvar A. Sigurdsson, Nadezhda M. Sushchevskaya, Mengist Teklay
Published 29 March 2006 on Science Express
DOI: 10.1126/science.1138113
This PDF file includes:
Materials and Methods Figs. S1 to S5 Tables S1, S2a, S2b, and S3 References
Other Supporting Online Material for this manuscript includes the following: available at www.sciencemag.org/cgi/content/full/1138113/DC1
Table S2 (complete) as zipped file
Sobolev et al, Supporting Online Material www.sciencemag.org/cgi/content/full/1138113/DC1
20 APRIL 2007 VOL 316 SCIENCE www.sciencemag.org
S1
Electron probe microanalysis. Olivine phenocrysts were analyzed for Si, Fe, Mg, Ca, Ni, Mn, Cr, Al and Co with a Jeol Jxa 8200 electron probe at Max-Planck Institute of Chemistry. Each olivine grain was analyzed in the geometrical center at 20 kV accelerating voltage and probe current of 300 nA. A small olivine subset was analyzed at 30 kV and 200nA. Details of analytical conditions are presented in Table S1. These conditions have been found optimal to obtain best detection limits (see Fig S1). The intensity of the Co Kα line was additionally corrected for overlap with the shoulder of FeKβ second order line using the following linear equation established by analyzing Fe bearing standards free of Co: CoOc=CoOm-0.0011×FeOm-0.013. [S1] Where CoOc and CoOm, FeOm-are corrected and measured values correspondingly.
The constant in equation [S1] was obtained from analyses of San Carlos olivine standard with known Co content (Table S1). Note that the linear approximation of equation [S1] may not be valid for highly Mg (Fo>93) and low Mg (Fo<70) olivines.
In addition San Carlos olivine standard USNM 111312/444 (S1) was run (3 points per each 30-50 measurement points) as an
unknown to monitor drifts in calibration and estimate accuracy of analysis. All measurements of Si, Fe, Mn, Ni, Ca and Al were corrected for deviation of this standard from the reference values (Table S1).
For trace elements, the above measurement conditions routinely yield detection limits of around 6-15 ppm based on 3 sigma criteria by Jeol standard software, and errors of 15-30 ppm for trace elements and 0.01 mol% for Fo content (2 standard errors) established by counting statistics and reproducibility of olivine standard (see Fig S1-S2). Precision and
El St Cryst Line Peak BG(+) BG(-) S.C.O.
Si 1 TAP Kα 90 90 - 190740 *
Al 2 TAP Kα 240 120 120 170 #
Fe 1 LIF Kα 90 90 - 74230 *
Mn 3 LIFH Kα 120 60 60 1084 *
Mg 1 TAP Kα 90 45 45 298040 *
Ca 4 PETJ Kα 120 60 60 665 #
Ni 5 LIF Kα 150 70 70 2907 *
Co 6 LIFH Kα 120 120 - 140 **
Cr 7 PETJ Kα 120 60 60 95 #
Table S1. Typical analytical conditions of electron probe microanalysis of olivine. Probe current 300 nA, acceleration voltage 20kV.
The Amount of Recycled Crust in Sources of Mantle-Derived Melts
A. V. Sobolev, A. W. Hofmann, D. V. Kuzmin, G. M. Yaxley, N. T. Arndt, et al.
Supporting Online Material Methods and Samples
1.41
1.43
1.45
1.47
1.49
1.51
1.53
1.55
90 90.1 90.2 90.3 90.4 90.5
100M
n/Fe
100
110
120
130
140
150
160
170
90 90.1 90.2 90.3 90.4 90.5
Fo mol%
Co
ppm
600
620
640
660
680
700
720
740
760
90 90.1 90.2 90.3 90.4 90.5
Ca
ppm
2800
2820
2840
2860
2880
2900
2920
2940
2960
2980
3000
90 90.1 90.2 90.3 90.4 90.5
Ni p
pm
measuredcorrected
40
50
60
70
80
90
100
110
120
130
90 90.1 90.2 90.3 90.4 90.5
Fo mol%
Cr p
pm
1040
1050
1060
1070
1080
1090
1100
1110
1120
1130
1140
90 90.1 90.2 90.3 90.4 90.5
Mn
ppm
Fig. S2. Reproducibility of San Carlos olivine standard analyzed three times every 30 to 50-th measurement points. Measured- represent uncorrected measurements, corrected – stands for measurements corrected for the calibration drift.
Notes for Table S1: Standards (St) - 1- San-Carlos olivine USNM 111312/444 (S.C.O.); 2-pure Al2O3; 3-Rhodonite; 4-Wollastonite; 5-pure NiO; 6- pure Co-metal; 7-pure Cr2O3. Peak and background (BG) counting time in seconds. *-Accepted values for (S.C.O.) from ref (S1) in ppm, **-value measured by LA-ICP MS.
Fig. S1. Detection limit for Mn as function of probe current and peak counting time (in seconds, sec). Detection limit was determined for San Carlos olivine standard with Jeol software using background statistics and 3 sigma criterion. Typical measurement conditions used in literature are marked as “routine analysis”. Analysis at higher beam current and counting times published by (S3) is marked as “precision analysis”. The analytical conditions used in this study are indicated as “high precision analysis”.
0
10
20
30
40
50
60
70
0 50 100 150 200 250 300 350
Probe current, nA
Det
ectio
n lim
it, p
pm
30 sec60 sec120 sec180 sec
high precision analysis
precision analysis
routine analysis
Mn
800
900
1000
1100
1200
1300
1400
1500
800 900 1000 1100 1200 1300 1400 1500
Mn by EPMA, ppm
Mn
by L
A-I
CPM
S, p
pm
100
150
200
250
300
100 150 200 250 300
Co EPMA, ppm
Co
LA
-IC
PMS,
ppm
Fig. S3. Comparison between Co and Mn concentrations in olivine measured by EPMA and LA-ICPMS. Error bar corresponds to ± 2 standard errors for both EPMA and LA-ICP MS
Sobolev et al, Supporting Online Material www.sciencemag.org/cgi/content/full/1138113/DC1
20 APRIL 2007 VOL 316 SCIENCE www.sciencemag.org
S2
accuracy of Co and Mn measurements were independently checked by LA-ICP MS analysis using Thermo Finnigan Element 2 mass spectrometer and New Wave Up213 YAG laser (213 nm wavelength) setup. KL2-G (S4) and NIST 612 glasses were used as a external standards and Ca in olivine as reference value. Figure S3 shows that both Co
and Mn concentrations measured by EPMA correspond to those analyzed by LA-ICP MS within 20 ppm (2 standard errors).
Olivine database. The averaged most highly magnesian olivines for each sample are presented in Table S2a. Individual olivine analyses are presented on Fig S4 and in Tables S2,c,d,e,f separately for each group: MORB in Table S2c, WPM-THIN in Table S2d, WPM-THICK in Table S2e and KOMATIITES in
Table S2f. These tables are included as separate spreadsheets together with Table S2a in the Excel file called Table S2. Table S2a is also placed in the end of this file. Tables S2c,d,e,f include concentrations of oxides and their relative standard deviations in % (signal counting statistics) for individual olivine grains. Each analytical point in addition to
sample name has a unique number. Table S2a includes group title, sample name, reference for sample description (if published), information on locality, number of averaged high-Mg grains, forsterite content, element concentrations in ppm, standard errors of the mean in ppm, characteristic ratios of elements and calculated amount of pyroxenitic component (X px, see below for explanation).
Melting of pyroxenites. The model hybrid pyroxenite composition from Sobolev et al. (S3) (their Table S2, column 50%) was chosen for the high pressure experimental investigation. A synthetic starting material with this composition (Px-1) was prepared for the experiments by blending high purity oxides and carbonates under analytical grade
acetone for several hours, until the material was homogenous and very fine-grained. The resultant dried powder was then pelletised and fired for 12 hours at 1100°C to decarbonate and partially fuse the components. FeO was added after firing in the form of synthetic fayalite (Fe2SiO4), and again blended under acetone. The final mixture was then dried overnight at 200°C and subsequently stored at 110°C. The actual composition of the Px-1
A
500
1000
1500
2000
2500
3000
3500
70 75 80 85 90 95
Fo
Mn
ppm
WPM-THICKKOMATIITESWPM-THINMORB
C
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
70 75 80 85 90 95
Fo
Ni p
pm
E
0
500
1000
1500
2000
2500
3000
3500
70 75 80 85 90 95
Fo
Ca
ppm
B
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
70 75 80 85 90 95
Fo
100M
n/Fe
D
0
0.5
1
1.5
2
2.5
70 75 80 85 90 95
Fo
Ni/(
Mg/
Fe)/1
000
F
0
500
1000
1500
2000
2500
70 75 80 85 90 95
Fo
Cr p
pm
Fig. S4. Composition of olivine phenocrysts: all data from database. Group of MORB olivines shown on top of all groups.
Sobolev et al, Supporting Online Material www.sciencemag.org/cgi/content/full/1138113/DC1
20 APRIL 2007 VOL 316 SCIENCE www.sciencemag.org
S3
starting material was checked by electron probe microanalysis of quenched glass prepared by fusing the starting mix at 1300°C in an Ar atmosphere in a box furnace and quenching in water (Table S3).
High pressure experiments were run at a pressure of 3.5 GPa and temperatures between 1400-1570°C in a conventional 1.27cm piston-cylinder apparatus at the Australian National University. The Px-1 starting material (≈1 mg) was loaded into graphite capsules, which were sealed by arc-welding in Pt outer capsules. The
capsules were placed centrally into NaCl-pyrex sleeves with MgO inserts, an assembly which requires no friction correction. Type B thermocouples (Pt6Rh94/Pt30Rh70) were employed, with the thermocouple join placed within a fraction of a millimeter of the Pt capsule. Runs were brought to run temperature and pressure simultaneously. Temperature was controlled throughout by Eurotherm controllers attached to the thermocouple and is accurate to ±10°C. Pressure is accurate to ±0.1 GPa. Experiments were maintained at the
desired PT condition long enough to allow a close approach to equilibrium, and then quenched by terminating power to the furnace.
After recovery from the post-run assembly, the Pt capsules were mounted in 1 inch diameter epoxy buttons, sectioned longitudinally and polished, in preparation for analysis by scanning electron microscopy at the Australian National University and electron probe microanalysis at the Max Planck Institute for Chemistry in Mainz.
Pyroxenite-derived component in melts. In order to estimate endmember compositions we have averaged compositions of melts of pyroxenite and peridotite (see Fig.S5). For pyroxenite we averaged all melt compositions except numbers C-2298 and C-2327 (Table S3) produced at low temperature and low percentage of melting. For peridotites we averaged 7 compositions of melts from runs 30.12, 30.07, 30.14, 30.1 (3.GPa), 40.06, 40.07, 40.05 (4.GPa) of Walter, 1998 (S2). The endmember compositions are shown in Table S3. We assumed that the mixtures consist of endmembers with similar temperatures, and we therefore excluded (from the averaging procedure) experimental melts with very low or very high temperatures (see Fig.S5).
The calculated endmember melts were then mixed (in 10% intervals) producing mixed melt compositions. For these melts the composition of equilibrium olivines were
calculated using Herzberg’s model (S5). The calculated Mn/Fe ratios of the olivines and the amounts of pyroxenitic endmember (Xpx) yield a straight line: Xpx = 3.48–2.071×(100Mn/Fe) [S2]
We used this equation to calculate the amount of pyroxenitic endmember for each average olivine composition in Table S2a.
The calculated endmember compositions are derived from experiments at high pressures and temperatures. Therefore, they are directly relevant primarily to thick-lithosphere settings, but we have applied them for all settings. Equivalent estimates for lower pressures and temperatures (1.0 GPa and 1300-1350oC) have been calculated using pMELTS (S6). This yielded similar degrees of melting and residual assemblages (50 to 60% of melting and low-Ca pyroxene dominated residue) as those derived from high pressure and temperature runs (Table S3). This gives us some confidence that the same Mn/Fe ratios can be applied to lower pressures, but experimental confirmation will be needed. Additional confidence in the results is derived from the observation that the extreme compositions observed in both MORB and WPM-thick settings match the calculated endmember compositions reasonably well.
Run DT T Phases proportions, wt% Melt compositions in oxide wt%, Ni in ppm
N h o C melt Opx Cpx Ga SiO2 TiO2 Al2O3 FeO MnO MgO CaO Na2O K2O Cr2O3 Ni
C-2298 45 1400 8 0. 79 14 60.70 2.76 14.40 6.80 0.068 4.44 7.71 2.34 0.70 0.07 453
C-2317 48 1450 16 0 71 12 55.76 2.33 14.20 7.63 0.081 7.45 7.95 2.16 0.44 0.06 626
C-2301 5 1500 33 0 60 6 53.30 1.58 14.21 8.25 0.106 10.67 8.61 1.75 0.13 0.09 810
C-2330 5 1520 42 0 55 3 52.54 1.29 14.22 8.65 0.112 11.66 8.65 1.98 0.14 0.11 850
C-2332 5 1540 72 24 4 0 52.00 0.86 13.55 8.19 0.125 14.08 9.00 1.80 0.08 0.20 796
C-2318 6 1550 76 18 7 0 52.38 0.83 13.39 7.95 0.117 14.67 8.77 1.74 0.09 0.20 843
C-2319 5 1575 79 20 0 0 52.20 0.79 13.07 8.11 0.126 15.44 8.52 1.67 0.07 0.22 841
Px-1, Bulk composition 52.67 0.64 11.26 7.55 0.119 18.48 7.05 1.52 0.06 0.25 1000 Peridotite-derived endmember 46.79 0.85 11.50 9.68 0.185 19.07 10.00 0.84 0.47 0.39 642
Pyroxenite -derived endmember 52.56 1.07 13.71 8.24 0.117 13.32 8.72 1.79 0.10 0.16 830
Table S3. Proportions of phases produced and melt compositions from melting of pyroxenite at 3.5 GPa and 1400-1575oC.
Note for Table S3: Melt, Opx, Cpx, Ga, - proportions of melt, orthopyroxene, clinopyroxene and garnet in experimental runs in mass fractions, calculated using list square method from bulk composition and compositions of phases. DT- run duration in haurs, T- temperature in oC. The amount of Ni has been calculated using the method described in Sobolev et al, 2005 (S3). This method uses known phase proportions and the distribution of Ni between phases, to calculate Ni in the phases on the basis of mass balance and known bulk Ni content. The assumed bulk Ni concentrations are: 1000 ppm for pyroxenite (Px-1) and 1900 ppm for peridotite (S3).
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1350 1400 1450 1500 1550 1600 1650 1700
Mel
t fra
ctio
n
Pyroxenite 3.5 GPaPeridotite 3.0 GPaPeridotite 4.0 GPa
0.800
1.000
1.200
1.400
1.600
1.800
2.000
2.200
1350 1400 1450 1500 1550 1600 1650 1700
100M
n/Fe
mel
t
0.000
0.200
0.400
0.600
0.800
1.000
1.200
1.400
1.600
1.800
1350 1400 1450 1500 1550 1600 1650 1700
Temperature C
100N
i/Mg
mel
t
Fig. S5. Melt fraction and composition of experimental melts of pyroxenite and fertile peridotite. Temperatures for experimental runs of fertile peridotite (S2) are extrapolated to 3.5 GPa by using slope 100C/GPa. Data for pyroxenite from Table S3. Crystalline phase coexisting with melt are: Ol - olivine; Opx - orthopyroxene; Cpx -clinopyroxene; Ga - garnet. Blue and red ellipses indicate analyses averaged to calculate peridotite-derived and pyroxenite-derived endmembers respectively.
Sobolev et al, Supporting Online Material www.sciencemag.org/cgi/content/full/1138113/DC1
20 APRIL 2007 VOL 316 SCIENCE www.sciencemag.org
S4
Oceanic crust in magma sources. A final step is to estimate the actual amount of recycled oceanic crust (Xcrc) from the determined proportion of pyroxenite-derived melt. This quantity is linked to the proportion of hybrid pyroxenite-derived melt (Xpx) by the degree of melting of eclogite (Fe), the amount of eclogite-derived melt needed to produce hybrid pyroxenite from peridotite (Xe), and the degrees of melting of peridotite (Fpe) and pyroxenite (Fpx) (S3):
[S3]
Following assumption of Sobolev et al,
(2005) (S3), we propose that the extent of melting of eclogites reaches a maximum of 50%. This limit is unlikely to be exceeded significantly during fractional melting, because this process removes Na (S7) and thereby renders the residue highly refractory. Therefore it is also not strongly dependent on the potential temperature of the rising mantle material. The amount of pyroxenite produced by reaction of the primary (eclogite-derived) melt with peridotite is prescribed by the stoichiometry of the reaction, which also yields proportions of the reactants of close to 50:50 (S3). For these, proportions, the amount of reaction pyroxenite equals the amount of initial eclogite. The melt fraction produced by the pyroxenite can be estimated at low pressure (1-2 GPa by modeling using pMELTS software (S6)), yielding a maximum of about 60 % for batch melting at 1 GPa and 1320 to 1350°C. At higher pressures (3.5 GPa) we must rely on the experimental data (Table S3, Fig. S5), which also yield a maximum melt fraction of about 50-60% for batch melting for 1540-1550°C.
These values will be significantly lower for fractional melting (S8), for the same reason as discussed for eclogite melting (i.e. early Na removal). Adopting an amount of 50 % pyroxenite melting will therefore result in a minimum estimate in the amount of recycled crust present. Finally, we estimate the melt fraction of peridotite using published models (S5) for MORB (10%), Iceland, as a proxy for the WPM-THIN group (20%), Hawaii, as proxy for the WPM-THICK group (10%) and Archaean komatiites -40% (S9). This yields the following average estimates for the amounts of recycled oceanic crust in the mantle sources: 4% for MORB, 11% for WPM-THIN group, 16% for WPM-THICK group, and around 13% for Archean komatiites. According to (S10) the melt fraction of peridotite for Ontong Java high-Mg magmas corresponds to 30%, which yield the amount of recycled oceanic crust of 13-28% in the mantle source of these magmas.
Supporting Online References
S1. E. J. Jarosevich, J. A. Nelen, J. A. Norberg, Geostand. Newsl. 4, 43 (1980).
S2. M. J. Walter, Journal of Petrology 39, 29 (J1998).
S3. A. V. Sobolev, A. W. Hofmann, S. V. Sobolev, I. K. Nikogosian, Nature
434, 590 (2005).
S4. K. P. Jochum et al., Geostandards Newsletter-the Journal of Geostandards and
Geoanalysis 24, 87 (2000).
S5. C. Herzberg, M. J. O'Hara, Journal of Petrology 43, 1857 (2002).
S6. M. S. Ghiorso, M. M. Hirschmann, P. W. Reiners, V. C. Kress,
Geochemistry Geophysics Geosystems 3 (2002).
S7. T. Kogiso, M. Hirschmann, Earth and Planetary Science Letters 249, 188
(2006).
S8. I. Kushiro, Annual Review of Earth and Planetary Sciences 29, 71 (2001).
S9. N. Arndt, Journal of Geophysical Research-Solid Earth 108 (2003).
S10. J. G. Fitton, M. Godard, in Origin and Evolution of the Ontong Java PlateauJ. G.
Fitton, J. J. Mahoney, P. J. Wallace, A. D. Saunders, Eds. (Geological
Society, London, Special Publication, 229 London, 2004) pp. 151-178.
S11. B. Sylvander, D. M. Christie, L. V. Danyushevsky, in Eos. (1996), vol. 77,
pp. F691.
S12. M. R. Perfit et al., Earth and Planetary Science Letters 141, 91 (1996).
S13. A. V. Sobolev, Petrology 4, 209 (1996).
S14. V. S. Kamenetsky et al., Earth and Planetary Science Letters 160, 115 (1998).
S15. V. Kamenetsky, Earth and Planetary Science Letters 142, 479 (1996).
S16. A. P. Leroex, H. J. B. Dick, L. Gulen, A. M. Reid, A. J. Erlank, Geochimica Et
Cosmochimica Acta 51, 541 (1987).
S17. V. S. Kamenetsky et al., Geology 29, 243 (2001).
S18. N. M. Sushchevskaya et al., Geochemistry International 43, 222 (2005).
S19. in Table S2b.
S20. T. Elliott, J. Blichert-Toft, A. Heumann, G. Koetsier, V. Forjaz, Geochimica
Et Cosmochimica Acta 71, 219 ( 2007).
S21. S. Huang, M. Regelous, T. Thordarson, F. A. Frey, Geochemistry Geophysics
Geosystems 6 (2005).
S22. A. C. Kerr, R. W. Kent, B. A. Thomson, J. K. Seedhouse, C. H. Donaldson,
Journal of Petrology 40, 873 (1999).
S23. G. M. Yaxley, V. S. Kamenetsky, M. Kamenetsky, M. D. Norman, D.
Francis, Contributions to Mineralogy and Petrology 148, 426 (2004).
S24. L. M. Larsen, A. K. Pedersen, Journal of Petrology 41, 1071 (2000).
S25. K. G. Cox, B. G. Jamieson, Journal of Petrology 15, 269 (1974).
S26. Y. G. Xu, S. L. Chung, B. M. Jahn, G. Y. Wu, Lithos 58, 145 (2001).
S27. S. L. Chung, B. M. Jahn, Geology 23, 889 (1995).
S28. M. Teklay, A. W. Hofmann, G. E. Brugmann, J. C. Lassiter, Geochimica Et
Cosmochimica Acta 66, A767 (2002).
S29. A. V. Sobolev, I. K. Nikogosian, Petrology 2, 111 (1994).
S30. J. M. Rhodes, M. J. Vollinger, Geochemistry Geophysics Geosystems 5, Q03G13
(2004).
S31. F. A. Frey et al., Journal of Geophysical Research-Solid Earth and Planets 96,
14347 (1991).
S32. F. A. Frey, M. O. Garcia, M. F. Roden, Geochimica Et Cosmochimica Acta 58,
1441 (1994).
S33. E. H. Haskins, M. O. Garcia, Contributions to Mineralogy and Petrology 147,
162 (2004).
S34. M. D. Norman, M. O. Garcia, Earth and Planetary Science Letters 168, 27
(1999).
S35. M. O. Garcia, D. J. P. Foss, H. B. West, J. J. Mahoney, Journal of Petrology
36, 1647 (1995).
S36. V. V. Matvienkov, O. G. Sorokhtin, Okeanologiya 38, 742 (1998).
S37. A. A. Gurenko et al., Chemical Geology 233, 75 (2006).
S38. A. A. Gurenko, T. H. Hansteen, H.-U. Schmincke, in Proceedings of the
Ocean Drilling Program, Scientific Results P. P. E. Weaver, H.-U. Schmincke, J.
V. Firth, W. Duffield, Eds. (1998), vol. 157, pp. 375-401.
S39. N. T. Arndt, A. J. Naldrett, D. R. Pyke, Journal of Petrology 18, 319 (1977).
S40. N. T. Arndt, Journal of Petrology 27, 279 (1986).
S41. M. J. Bickle et al., in The geology of the Belingwe Greenstone Belt, Zimbabwe M. J.
Bickle, E. G. Nisbet, Eds. (Balkema, Rotterdam, 1993) pp. 175-214.
S42. N. T. Arndt, G. E. Brügmann, K. Lehnert, C. Chauvel, B. W. Chappell, in
Geochemistry and Mineralization of Proterozoic Volcanic Suites T. C. Pharaoh, R.
D. Beckinsale, D. Rickard, Eds. (Geol. Soc. Spec. Publ., 1987) pp. 133-145.
S43. A. C. Kerr, Lithos 84, 77 (2005).
Table S2b. Locations (GPS coordinates) and description for Icelandic samples
Sample Locality Description latitude longitude
01-7 Reykjanes, Haleyjabunga Glassy olivine phyric picrite N63o 48.954' W22o 38.646' 03-102 Reykjanes, Lagafell Glassy olivine phyric picrite N63o 52.906' W22o 32.611'
03-106 Reykjanes, Haleyjabunga Glassy olivine phyric picrite N63o 48.873' W22o 38.724'
01-15 Reykjanes, Sulur Glassy olivine –phyric basalt N63o 54.094' W22o 32.558'
01-12 Reykjanes, Stapafell Glassy olivine-plagioclase phyric picrite N63o 54.601' W22o 31.370'
01-8 Hengill, Midfell Glassy olivine phyric picrite with rare plagioclase and clinopyroxene xenocrysts N64o 10.140' W21o 04.257'
01-10 Hengill, Midfell Glassy olivine phyric picrite with rare plagioclase and clinopyroxene xenocrysts N64o 10.103' W21o 04.118'
01-19 Hengill, Maelifell Glassy olivine phyric basalt with rare plagioclase and clinopyroxene xenocrysts N64o 06.252' W21o 10.744'
03-131 Kistufell Glassy olivine phyric basalt N64o 47.806' W17o 13.745'
03-164 Kistufell Glassy olivine phyric basalt N64o 46.918' W17o 12.713'
03-140 Kistufell Glassy olivine phyric basalt N64o 47.861' W17o 12.202'
03-161 Kistufell Glassy olivine phyric basalt N64o 47.808' W17o 13.957'
01-55 Theistareykir, Laufrandarhraun Glassy olivine phyric picrite N65o 56.285' W17o 05.047'
01-57-4 Theistareykir, Laufrandarhraun Glassy olivine phyric picrite N65o 55.791' W17o 04.463'
01-56-1 Theistareykir, Laufrandarhraun Glassy olivine phyric picrite N65o 56.277' W17o 05.374'
01-41 Theistareykir, Theistareykjahraun Glassy olivine phyric basalt N65o 57.547' W17o 04.120'
01-44 Theistareykir, Laufrandarhraun Glassy olivine phyric picrite N65o 56.068' W17o 05.246'
01-56-2 Theistareykir, Laufrandarhraun Glassy olivine phyric picrite N65o 56.277' W17o 05.374'
01-54 Theistareykir, Laufrandarhraun Glassy olivine phyric picrite N65o 56.281' W17o 04.624'
01-51 Theistareykir, Langavitishraun Glassy olivine phyric basalt with rare clinopyroxene phenocrysts N65o 56.058' W16o 52.282'
03-224 Snaefellsness, Enni Glassy olivine phyric basalt N64o 54.146' W23o 45.776'
03-226 Snaefellsness, Sydri-Raudamelur Olivine phyric basalt N64o 52.296' W22o 17.368'
03-220 Snaefellsness, Ytri-Raudamelur Olivine phyric basalt N64o 52.717' W22o 20.860'
)111
( +−
+−
=
ee
e
pe
px
px
pxe
ecrc
XF
FFF
XX
F
XX
Sobo
lev
et a
l, Su
ppor
ting
Onl
ine
Mat
eria
l ww
w.sc
ienc
emag
.org
/cgi
/con
tent
/full/
1138
113/
DC
1
20 A
PRIL
200
7 VO
L 31
6 SC
IEN
CE
ww
w.s
cien
cem
ag.o
rg
S5
Tab
le S
2a. A
vera
ge c
ompo
sitio
ns o
f mos
t Mg-
rich
oliv
ine
phen
ocry
sts (
see
text
of p
aper
for d
efin
ition
). GR
OUP
Geod
ynam
ic se
tting
Refer
ence
s Sa
mple
Geog
raph
ic
Loca
lity
N Fo
Al
ppm
Fe pp
m Mn
ppm
Mg pp
m Ca
ppm
Ni pp
m Co
ppm
Cr pp
m 10
0*Mn
/Fe
100*
Ni/M
g Ni
/(Mg/F
e)/10
00
100C
a/Fe
X px
Mn/F
e ST
E Fo
wt%
Al p
pm
Fe pp
m Mn
ppm
Mg pp
m Ca
ppm
Ni pp
m Co
ppm
Cr pp
m
MORB
MO
RB
(S11
) 11
3-16
Ind
ian O
cean
10
5.22 o
E SE
IR
24
89.49
27
0 79
979
1293
29
6327
19
95
2109
389
1.617
0.7
12
0.569
2.4
94
0.135
0.07
7 50
2 10
26
8 13
38
5 MO
RB
MORB
(S
11)
145-
3 Ind
ian O
cean
11
6.72 o
E SE
IR
9 89
.90
298
7703
2 12
24
2982
38
1761
25
00
36
9 1.5
90
0.838
0.6
46
2.285
0.1
91
0.0
5 4
422
8 23
4 9
27
9
MORB
MO
RB
unpb
30
-29
Indian
Oce
an
124.2
8 oE
SEIR
10
89
.60
236
7918
5 12
80
2968
74
1843
21
15
39
0 1.6
17
0.712
0.5
64
2.328
0.1
35
0.1
0 10
72
4 14
49
9 20
71
8 MO
RB
MORB
(S
11)
144-
13
Indian
Oce
an
115.2
1 oE
SEIR
29
89
.16
269
8231
5 13
16
2944
71
1921
19
93
36
3 1.5
99
0.677
0.5
57
2.334
0.1
71
0.0
5 6
377
8 26
5 20
39
5 MO
RB
MORB
un
pb
18-1
0 Ind
ian O
cean
12
6.74 o
E SE
IR
24
89.60
27
2 78
878
1239
29
5854
19
21
2303
372
1.571
0.7
78
0.614
2.4
35
0.229
0.06
4 43
8 9
280
12
38
7
MORB
MO
RB
(S11
) 69
-1
Indian
Oce
an
88.92
oE S
EIR
11
89.89
31
9 76
966
1205
29
7733
20
36
1972
299
1.566
0.6
62
0.510
2.6
46
0.239
0.05
10
417
8 28
7 10
29
4 MO
RB
MORB
(S
11)
88-2
2 Ind
ian O
cean
96
.83 oE
SEI
R 10
88
.18
213
8862
7 14
53
2878
37
2153
15
72
38
0 1.6
39
0.546
0.4
84
2.429
0.0
88
0.1
1 8
755
14
526
10
18
7
MORB
MO
RB
unpb
5-
2 Ind
ian O
cean
12
6.52 o
E SE
IR
33
89.63
25
7 78
288
1238
29
4600
18
75
2383
364
1.582
0.8
09
0.633
2.3
95
0.207
0.04
3 30
7 6
235
3 17
2 MO
RB
MORB
un
pb
16-1
9 Ind
ian O
cean
12
7.58 o
E SE
IR
18
88.40
23
5 87
179
1357
28
9113
19
20
1953
348
1.557
0.6
76
0.589
2.2
02
0.258
0.15
5 11
12
18
675
21
48
6
MORB
MO
RB
unpb
J-P
E-1
Pacif
ic Oc
ean
Garre
tt F.Z
. 32
90
.82
395
6962
7 10
84
2997
51
2071
28
17
48
0 1.5
57
0.940
0.6
54
2.975
0.2
59
0.0
2 10
17
1 3
169
3 38
11
MORB
MO
RB
(S12
) SQ
-1
Pacif
ic Oc
ean
Siqu
eiros
F.Z
. 89
90
.38
377
7245
4 11
31
2962
61
1985
26
61
137
413
1.561
0.8
98
0.651
2.7
39
0.250
0.03
4 19
1 3
201
2 15
1
5 MO
RB
MORB
(S
13)
FZ 64
9-5(
i) At
lantic
Oce
an
9 N M
AR
36
89.75
30
1 76
919
1213
29
3090
19
53
2444
374
1.577
0.8
34
0.642
2.5
39
0.216
0.04
5 30
9 4
200
5 15
4 MO
RB
MORB
(S
13)
FZ 64
9-5(
ii) At
lantic
Oce
an
9 N M
AR
28
88.76
30
2 84
161
1235
28
9205
18
32
2410
371
1.467
0.8
33
0.701
2.1
77
0.444
0.04
7 27
6 6
173
14
30
6
MORB
MO
RB
(S13
) 64
9(i)
Atlan
tic O
cean
9N
MAR
12
89
.63
314
7804
4 12
22
2933
79
1929
25
01
144
360
1.566
0.8
53
0.665
2.4
72
0.239
0.05
7 41
0 6
334
18
20
2 5
MORB
MO
RB
(S13
) 64
9(ii)
Atlan
tic O
cean
9N
MAR
10
88
.87
310
8345
2 12
40
2898
34
1875
24
34
155
390
1.486
0.8
40
0.701
2.2
47
0.406
0.08
9 63
9 10
58
5 20
55
5
8 MO
RB
MORB
un
pb
AII-6
0, St
n9, D
9, 9-
20
N. A
tlanti
c 43
N MA
R 88
89
.81
332
7641
6 11
84
2931
34
1896
25
43
43
0 1.5
49
0.868
0.6
63
2.481
0.2
74
0.0
2 2
152
2 10
3 2
11
3
MORB
MO
RB
unpb
AI
I-32 1
1-92
N.
Atla
ntic
43N
MAR
92
90.70
24
2 70
116
1151
29
7618
20
32
2461
502
1.642
0.8
27
0.580
2.8
98
0.082
0.01
5 10
9 2
97
9 7
5
MORB
MO
RB
unpb
AI
I-32 1
2-6(
i) N.
Atla
ntic
43N
MAR
74
90.83
24
0 69
316
1124
29
8835
23
40
2221
428
1.622
0.7
43
0.515
3.3
76
0.124
0.03
3 20
0 3
134
14
19
10
MO
RB
MORB
un
pb
AII-3
2 12-
6(ii)
N. A
tlanti
c 43
N MA
R 13
90
.88
252
6893
4 11
75
2989
50
2174
22
91
51
0 1.7
05
0.766
0.5
28
3.153
-0
.048
0.0
1 3
57
2 16
6 33
10
3 MO
RB
MORB
un
pb
AII-3
2 12-
2 N.
Atla
ntic
43N
MAR
85
91.13
25
4 66
863
1089
29
8972
22
21
2339
531
1.629
0.7
82
0.523
3.3
21
0.110
0.03
4 20
1 4
132
27
18
10
MO
RB
MORB
(S
14)
AII-3
2 12-
7 N.
Atla
ntic
43N
MAR
55
91.03
21
3 67
956
1115
30
0251
23
93
2252
453
1.641
0.7
50
0.510
3.5
22
0.084
0.03
4 21
1 4
152
22
11
10
MO
RB
MORB
(S
15)
AII-3
2 11-
178
N. A
tlanti
c 43
N MA
R 68
90
.88
260
6915
5 11
40
2998
97
2072
24
79
52
1 1.6
48
0.827
0.5
72
2.996
0.0
70
0.0
3 5
202
3 12
6 12
15
6 MO
RB
MORB
(S
15)
CH-3
1 Dr0
8 695
N.
Atla
ntic
Famo
us
17
91.21
30
6 66
586
1039
30
0597
21
00
2721
592
1.561
0.9
05
0.603
3.1
54
0.251
0.09
9 67
0 12
37
1 8
81
8
MORB
MO
RB
(S15
) CH
-31 D
r08 1
36-0
2 N.
Atla
ntic
Famo
us
19
91.38
29
8 64
581
1012
29
7930
20
19
2620
595
1.567
0.8
79
0.568
3.1
27
0.237
0.07
10
523
9 26
5 7
89
8
MORB
MO
RB
(S15
) AR
P-73
10.03
N.
Atla
ntic
Famo
us
52
91.30
29
5 65
425
1039
29
8895
20
46
2535
582
1.589
0.8
48
0.555
3.1
26
0.193
0.05
9 34
5 5
180
5 31
8 MO
RB
MORB
(S
16)
AG32
4-68
S.
Atla
ntic
Bouv
et 44
90
.72
319
6999
6 10
88
2978
91
2024
26
53
45
0 1.5
54
0.891
0.6
23
2.891
0.2
64
0.0
4 7
270
5 18
5 7
41
10
MO
RB
MORB
un
pb
AG-3
2 3-3
6 S.
Atla
ntic
Bouv
et 98
89
.95
282
7547
4 12
17
2940
27
1869
22
39
42
7 1.6
12
0.761
0.5
75
2.476
0.1
44
0.0
4 4
264
4 15
6 4
40
6
MORB
MO
RB
unpb
AG
-32 3
-41
S. A
tlant
ic Bo
uvet
73
90.22
30
1 74
107
1199
29
7481
19
02
2432
474
1.618
0.8
18
0.606
2.5
66
0.132
0.02
5 14
9 3
122
10
35
7
MORB
MO
RB
(S17
) 51
8-60
/1 S.
Atla
ntic
Bouv
et 5
86.70
22
5 99
549
1199
28
2320
11
15
3478
536
1.204
1.2
32
1.226
1.1
20
0.99
0.0
5 5
308
6 34
5 11
24
22
MORB
MO
RB
unpb
13
-12/4
9a
Atlan
tic O
cean
Ro
manc
he F
.Z.
60
89.35
19
8 80
862
1332
29
5314
23
92
1999
13
9 27
0 1.6
48
0.677
0.5
47
2.958
0.0
70
0.0
3 3
230
4 20
6 6
13
1 6
MORB
MO
RB
unpb
13
-12/4
9c
Atlan
tic O
cean
Ro
manc
he F
.Z.
58
88.98
19
8 82
768
1357
29
0684
24
55
1953
14
5 26
4 1.6
40
0.672
0.5
56
2.966
0.0
87
0.0
4 5
268
5 17
0 13
13
1
5 MO
RB
MORB
un
pb
13-1
2/49d
At
lantic
Oce
an
Roma
nche
F.Z
. 19
88
.90
185
8379
5 13
88
2919
45
2415
18
72
143
250
1.656
0.6
41
0.537
2.8
81
0.053
0.06
7 41
5 7
229
31
18
2 10
MO
RB
MORB
un
pb
13-1
2/49e
At
lantic
Oce
an
Roma
nche
F.Z
. 36
87
.69
185
9253
9 15
14
2867
53
1715
20
76
154
322
1.636
0.7
24
0.670
1.8
54
0.095
0.02
4 10
2 3
160
4 10
2
5 MO
RB
MORB
un
pb
13-1
2/49f
Atlan
tic O
cean
Ro
manc
he F
.Z.
17
89.39
19
5 80
132
1331
29
3645
24
43
1985
14
1 26
0 1.6
61
0.676
0.5
42
3.048
0.0
44
0.0
6 6
478
9 33
9 11
12
2
8 MO
RB
MORB
un
pb
13-1
2/49g
At
lantic
Oce
an
Roma
nche
F.Z
. 21
89
.23
201
8129
9 13
38
2929
63
2402
19
96
145
262
1.646
0.6
81
0.554
2.9
54
0.074
0.06
3 43
8 8
358
11
15
2 6
MORB
MO
RB
unpb
13
-11/1
At
lantic
Oce
an
Roma
nche
F.Z
. 21
88
.11
240
8921
9 13
58
2876
67
1705
19
96
154
307
1.522
0.6
94
0.619
1.9
11
0.330
0.01
3 75
2
130
2 9
2 4
MORB
MO
RB
unpb
13
-11/2
At
lantic
Oce
an
Roma
nche
F.Z
. 36
87
.44
205
9409
0 13
36
2851
63
1854
15
68
164
341
1.420
0.5
50
0.517
1.9
70
0.542
0.05
3 34
1 6
219
6 11
1
4 MO
RB
MORB
un
pb
13-1
2/66
Atlan
tic O
cean
Ro
manc
he F
.Z.
2 90
.56
283
7090
2 11
93
2960
41
2433
22
79
130
356
1.682
0.7
70
0.546
3.4
32
-0.00
1
0.01
13
253
0 15
71
18
8 4
7 MO
RB
MORB
un
pb
13-1
2/78
Atlan
tic O
cean
Ro
manc
he F
.Z.
13
89.34
21
1 79
948
1314
29
1469
23
95
2059
14
7 28
3 1.6
43
0.707
0.5
65
2.996
0.0
80
0.0
8 5
562
10
433
18
23
3 9
MORB
MO
RB
unpb
13
-13/5
At
lantic
Oce
an
Roma
nche
F.Z
. 23
89
.16
230
8145
1 13
09
2914
85
1722
17
67
161
357
1.607
0.6
06
0.494
2.1
14
0.154
0.03
2 23
8 3
195
3 3
2 2
MORB
MO
RB
unpb
13
-13/1
4 At
lantic
Oce
an
Roma
nche
F.Z
. 2
88.27
16
4 87
742
1448
28
7388
18
83
1501
14
2 28
4 1.6
51
0.522
0.4
58
2.146
0.0
64
0.0
8 0
653
0 39
11
8
0 17
MO
RB
MORB
(S
18)
22-4
0 Ar
ctic o
cean
Kn
ipovic
h ridg
e 31
87
.56
213
9354
7 15
12
2865
42
1813
17
75
155
288
1.616
0.6
19
0.579
1.9
38
0.136
0.04
6 21
6 4
350
4 35
8 MO
RB
MORB
(S
18)
22-2
3 Ar
ctic o
cean
Kn
ipovic
h ridg
e 17
87
.42
219
9421
1 15
24
2849
00
1824
17
47
157
291
1.618
0.6
13
0.578
1.9
36
0.132
0.08
4 40
6 7
602
6 34
7 W
PM-T
HIN
OIB
(S19
) 01
-7
Icelan
d Re
ykjan
es, H
aleyja
bung
a 52
90
.22
344
7415
6 11
76
2976
83
2335
25
18
48
7 1.5
86
0.846
0.6
27
3.148
0.1
98
0.0
3 10
25
2 4
217
11
12
9
WPM
-THI
N OI
B (S
19)
03-1
02
Icelan
d Re
ykjan
es, L
agafe
ll 31
90
.66
355
7078
0 11
42
2989
21
2440
25
20
54
9 1.6
14
0.843
0.5
97
3.447
0.1
41
0.0
5 14
36
2 7
264
17
23
16
W
PM-T
HIN
OIB
(S19
) 03
-106
Ice
land
Reyk
janes
, Hale
yjabu
nga
65
90.17
31
4 74
317
1175
29
6602
23
08
2498
462
1.581
0.8
42
0.626
3.1
05
0.208
0.03
7 19
7 4
210
5 10
7 W
PM-T
HIN
OIB
(S19
) 01
-15
Icelan
d Re
ykjan
es, S
ulur
8 88
.39
237
8768
7 13
72
2905
73
1989
23
07
154
434
1.564
0.7
94
0.696
2.2
68
0.243
0.12
8 87
3 15
68
1 31
43
2
16
WPM
-THI
N OI
B (S
19)
01-1
2 Ice
land
Reyk
janes
, Stap
afell
7 87
.61
249
9289
2 14
57
2857
89
1987
21
23
36
8 1.5
69
0.743
0.6
90
2.139
0.2
34
0.1
3 5
952
15
695
35
73
18
W
PM-T
HIN
OIB
(S19
) 01
-8
Icelan
d He
ngill,
Midf
ell
11
90.25
47
4 73
667
1150
29
6824
24
60
2464
671
1.561
0.8
30
0.612
3.3
39
0.251
0.10
18
722
9 39
7 34
39
26
WPM
-THI
N OI
B (S
19)
01-1
0 Ice
land
Heng
ill, M
idfell
25
89
.80
373
7695
3 12
24
2947
65
2672
23
03
54
3 1.5
91
0.781
0.6
01
3.472
0.1
88
0.0
5 12
39
1 5
347
19
30
14
W
PM-T
HIN
OIB
(S19
) 01
-19
Icelan
d He
ngill,
Mae
lifell
5 90
.24
426
7353
3 11
55
2958
22
2444
27
31
149
643
1.571
0.9
23
0.679
3.3
24
0.228
0.19
9 13
81
23
719
17
80
6 23
W
PM-T
HIN
OIB
(S19
) 03
-131
Ice
land
Kistu
fell
67
88.43
29
5 86
728
1335
28
8535
21
70
2305
16
9 30
7 1.5
40
0.799
0.6
93
2.502
0.2
94
0.0
2 9
178
4 19
6 7
13
1 6
WPM
-THI
N OI
B (S
19)
03-1
64
Icelan
d Ki
stufel
l 25
88
.84
331
8404
6 12
87
2911
32
2174
24
11
167
356
1.531
0.8
28
0.696
2.5
87
0.311
0.05
10
320
8 42
9 18
18
2
8
Sobo
lev
et a
l, Su
ppor
ting
Onl
ine
Mat
eria
l ww
w.sc
ienc
emag
.org
/cgi
/con
tent
/full/
1138
113/
DC
1
20 A
PRIL
200
7 VO
L 31
6 SC
IEN
CE
ww
w.s
cien
cem
ag.o
rg
S6
GROU
P Ge
odyn
amic
settin
g Re
feren
ces
Samp
le Ge
ogra
phic
Lo
cality
N
Fo
Al pp
m Fe
ppm
Mn pp
m Mg
ppm
Ca pp
m Ni
ppm
Co pp
m Cr
ppm
100*
Mn/F
e 10
0*Ni
/Mg
Ni/(M
g/Fe)
/1000
10
0Ca/F
e X
px M
n/Fe
STE
Fo w
t% A
l ppm
Fe
ppm
Mn pp
m Mg
ppm
Ca pp
m Ni
ppm
Co pp
m Cr
ppm
WPM
-THI
N OI
B (S
19)
03-1
40
Icelan
d Ki
stufel
l 45
88
.56
305
8584
3 13
11
2892
25
2161
23
59
167
321
1.527
0.8
16
0.700
2.5
17
0.320
0.03
9 22
3 4
186
13
17
1 6
WPM
-THI
N OI
B (S
19)
03-1
61
Icelan
d Ki
stufel
l 50
88
.48
286
8640
8 13
26
2886
57
2157
23
41
169
317
1.535
0.8
11
0.701
2.4
97
0.303
0.04
10
263
6 21
5 8
17
1 6
WPM
-THI
N OI
B (S
19)
01-5
5 Ice
land
Theis
tarey
kir,P
icrite
s 24
90
.30
356
7331
3 11
68
2969
01
2439
24
89
144
584
1.594
0.8
38
0.615
3.3
27
0.182
0.06
19
430
9 35
6 31
33
2
21
WPM
-THI
N OI
B (S
19)
01-5
7-4
Icelan
d Th
eistar
eykir
,Picr
ites
32
90.06
37
9 74
874
1187
29
5369
23
71
2467
589
1.586
0.8
35
0.625
3.1
67
0.199
0.04
15
268
6 34
3 22
26
15
WPM
-THI
N OI
B (S
19)
01-5
6-1
Icelan
d Th
eistar
eykir
,Picr
ites
48
90.21
29
9 73
806
1189
29
5885
24
35
2423
552
1.611
0.8
19
0.604
3.2
99
0.147
0.04
10
284
6 22
7 15
18
13
WPM
-THI
N OI
B (S
19)
01-4
1 Ice
land
Theis
tarey
kir
15
89.76
43
7 77
047
1178
29
4013
21
77
2800
14
7 63
1 1.5
29
0.952
0.7
34
2.826
0.3
16
0.0
8 17
58
4 9
406
8 20
2
15
WPM
-THI
N OI
B (S
19)
01-4
4 Ice
land
Theis
tarey
kir
8 89
.16
415
8159
2 12
67
2920
04
2169
25
70
155
580
1.553
0.8
80
0.718
2.6
58
0.266
0.11
48
803
17
608
76
73
4 53
W
PM-T
HIN
OIB
(S19
) 01
-56-
2 Ice
land
Theis
tarey
kir,P
icrite
s 9
90.49
35
0 71
973
1149
29
7838
24
25
2491
15
3 58
6 1.5
96
0.836
0.6
02
3.370
0.1
77
0.1
0 30
71
6 18
41
4 43
67
3
44
WPM
-THI
N OI
B (S
19)
01-5
4 Ice
land
Theis
tarey
kir,P
icrite
s 45
87
.95
293
9033
5 14
07
2870
24
2183
22
97
172
399
1.557
0.8
00
0.723
2.4
17
0.258
0.04
9 29
8 5
226
13
14
1 8
WPM
-THI
N OI
B (S
19)
01-5
1 Ice
land
Theis
tarey
kir
6 91
.39
367
6565
0 10
35
3032
39
2433
22
87
138
885
1.577
0.7
54
0.495
3.7
07
0.217
0.15
38
1176
25
35
9 10
8 10
7 4
71
WPM
-THI
N OI
B (S
19)
03-2
24
Icelan
d Sn
aefel
lsnes
s 4
89.14
29
2 82
132
1272
29
3273
19
94
2363
15
0 51
3 1.5
49
0.806
0.6
62
2.428
0.2
75
0.2
5 12
18
23
29
1054
17
86
3
27
WPM
-THI
N OI
B (S
19)
03-2
26
Icelan
d Sn
aefel
lsnes
s 15
88
.28
232
8838
3 14
09
2896
92
2079
19
84
161
364
1.594
0.6
85
0.605
2.3
52
0.181
0.08
9 58
0 11
73
6 69
33
3
16
WPM
-THI
N OI
B (S
19)
03-2
20
Icelan
d Sn
aefel
lsnes
s 5
88.93
24
5 83
462
1344
29
1820
20
73
2067
15
6 45
8 1.6
11
0.708
0.5
91
2.483
0.1
47
0.2
1 23
14
84
17
1033
93
55
6
32
WPM
-THI
N OI
B (S
20)
SM-6
Az
ores
Sa
o Mige
l 26
87
.43
166
9461
0 14
69
2862
86
2062
17
06
161
308
1.552
0.5
96
0.564
2.1
80
0.268
0.05
3 34
1 5
243
16
19
2 5
WPM
-THI
N OI
B (S
20)
SM-1
0 Az
ores
Sa
o Mige
l 11
7 88
.58
184
8525
2 11
81
2877
85
1919
23
87
151
553
1.385
0.8
29
0.707
2.2
51
0.615
0.01
2 78
2
79
7 6
1 3
WPM
-THI
N OI
B (S
20)
SM-1
6 Az
ores
Sa
o Mige
l 12
2 89
.18
233
8108
2 10
94
2907
82
1812
23
12
149
672
1.349
0.7
95
0.645
2.2
34
0.689
0.02
2 14
8 3
100
10
26
1 3
WPM
-THI
N OI
B (S
20)
SM-3
6a
Azor
es
Sao M
igel
13
89.88
22
2 75
771
1140
29
2770
19
22
2605
13
9 52
5 1.5
04
0.890
0.6
74
2.537
0.3
68
0.1
0 9
761
12
481
26
40
3 13
W
PM-T
HIN
OIB
(S20
) SM
-36
Azor
es
Sao M
igel
75
87.91
17
0 89
944
1397
28
4715
23
57
1774
15
4 35
5 1.5
53
0.623
0.5
60
2.620
0.2
67
0.0
3 2
205
4 12
6 13
8
1 3
WPM
-THI
N OI
B (S
20)
SM-4
Az
ores
Sa
o Mige
l 4
88.20
17
7 87
865
1297
28
5844
18
87
1925
14
5 41
1 1.4
76
0.674
0.5
92
2.147
0.4
25
0.1
1 3
796
5 46
5 5
23
5 16
W
PM-T
HIN
OIB
(S20
) SM
-33
Azor
es
Sao M
igel
59
88.68
16
3 84
172
1269
28
7065
23
94
2159
14
9 41
4 1.5
08
0.752
0.6
33
2.844
0.3
59
0.0
2 2
175
3 12
1 17
14
1
3 W
PM-T
HIN
OIB
unpb
T-
18
Azor
es
Terce
ria
3 88
.45
199
8605
0 12
03
2867
61
1894
24
49
152
447
1.398
0.8
54
0.735
2.2
01
0.587
0.12
9 83
8 32
65
3 98
21
9 6
44
WPM
-THI
N OI
B un
pb
T-6
Azor
es
Terce
ria
43
88.73
24
3 84
248
1275
28
8597
18
97
2537
14
4 45
0 1.5
13
0.879
0.7
40
2.251
0.3
49
0.0
4 5
277
4 16
3 14
16
1
7 W
PM-T
HIN
OIB
(S21
) 12
03a -
038R
01W
Pa
cific
Ocea
n De
troit s
mt
57
89.53
34
4 78
587
1160
29
2482
18
37
2987
15
3 60
9 1.4
76
1.021
0.8
03
2.338
0.4
26
0.0
2 7
173
3 18
2 8
16
1 7
WPM
-THI
N OI
B (S
21)
1203
a -03
2R
Pacif
ic Oc
ean
Detro
it smt
13
89
.73
369
7692
4 11
20
2924
12
1813
30
25
151
628
1.456
1.0
34
0.796
2.3
57
0.467
0.07
14
520
9 49
5 10
32
2
16
WPM
-THI
CK
LIP
(S22
) BR
-2
Atlan
tic O
cean
Mu
ll isla
nd
17
87.25
34
4 96
138
1384
28
6280
20
00
2239
445
1.440
0.7
82
0.752
2.0
80
0.500
0.07
19
524
9 32
8 5
35
17
W
PM-T
HICK
LIP
(S
22)
BR-5
At
lantic
Oce
an
Mull i
sland
17
87
.99
421
9066
5 12
66
2890
61
1921
29
30
43
1 1.3
96
1.014
0.9
19
2.119
0.5
92
0.0
8 11
55
8 7
423
15
17
12
W
PM-T
HICK
LIP
(S
22)
BR-6
At
lantic
Oce
an
Mull i
sland
11
87
.83
435
9178
2 13
22
2881
14
1982
24
59
55
2 1.4
41
0.854
0.7
83
2.159
0.4
99
0.1
2 17
90
4 14
37
6 8
28
16
W
PM-T
HICK
LIP
(S
22)
BHL-
15
Atlan
tic O
cean
Mu
ll isla
nd
33
88.24
40
4 89
264
1333
29
1557
21
19
2386
510
1.493
0.8
18
0.730
2.3
74
0.391
0.04
7 32
6 5
230
3 17
7 W
PM-T
HICK
LIP
(S
22)
BHL-
19
Atlan
tic O
cean
Mu
ll isla
nd
16
87.83
46
3 92
074
1353
28
9015
21
51
2562
567
1.469
0.8
87
0.816
2.3
36
0.440
0.07
21
557
11
305
31
72
22
W
PM-T
HICK
LIP
(S
22)
BHL-
34
Atlan
tic O
cean
Mu
ll isla
nd
24
87.77
41
0 92
639
1335
28
9280
20
43
2655
476
1.441
0.9
18
0.850
2.2
05
0.499
0.05
10
396
6 25
9 5
22
13
W
PM-T
HICK
LIP
(S
22)
BCH-
14
Atlan
tic O
cean
Mu
ll isla
nd
10
86.50
46
1 10
1517
13
75
2829
42
1904
27
01
41
1 1.3
55
0.955
0.9
69
1.875
0.6
77
0.1
2 19
87
5 12
59
4 9
32
16
W
PM-T
HICK
LIP
(S
22)
BCH-
24
Atlan
tic O
cean
Mu
ll isla
nd
14
89.35
43
7 80
685
1195
29
4511
20
43
2912
654
1.482
0.9
89
0.798
2.5
33
0.414
0.09
6 68
8 11
49
5 16
28
14
WPM
-THI
CK
LIP
(S22
) BC
H-27
At
lantic
Oce
an
Mull i
sland
24
88
.90
426
8408
3 12
33
2929
30
2042
28
49
62
6 1.4
66
0.973
0.8
18
2.428
0.4
46
0.0
6 8
423
6 26
7 8
8
7 W
PM-T
HICK
LIP
(S
22)
BCH-
33
Atlan
tic O
cean
Mu
ll isla
nd
20
88.38
41
9 87
922
1262
29
0959
19
58
2782
550
1.435
0.9
56
0.841
2.2
27
0.510
0.06
10
434
7 29
5 7
16
9
WPM
-THI
CK
LIP
(S22
) AM
-7a
Atlan
tic O
cean
Mu
ll isla
nd
18
86.21
40
7 10
2948
14
38
2801
59
1956
23
96
32
8 1.3
97
0.855
0.8
80
1.900
0.5
90
0.0
6 13
43
0 7
303
20
18
11
W
PM-T
HICK
LIP
(S
22)
BB-2
2 At
lantic
Oce
an
Mull i
sland
12
86
.68
319
9973
1 13
45
2823
19
1621
25
49
38
5 1.3
49
0.903
0.9
00
1.626
0.6
90
0.1
2 53
78
5 19
65
4 12
1 63
32
WPM
-THI
CK
LIP
(S22
) B-
26
Atlan
tic O
cean
Mu
ll isla
nd
3 86
.66
450
9999
0 14
17
2826
39
2101
26
06
48
1 1.4
17
0.922
0.9
22
2.101
0.5
47
0.3
1 52
21
77
39
1420
33
97
75
WPM
-THI
CK
LIP
(S22
) B-
29
Atlan
tic O
cean
Mu
ll isla
nd
12
87.29
39
3 95
617
1355
28
5840
19
55
2696
448
1.417
0.9
43
0.902
2.0
44
0.547
0.09
21
613
10
532
13
23
15
W
PM-T
HICK
LIP
(S
23)
BB
Atlan
tic O
cean
Ba
ffin B
ay
1 92
.65
386
5570
9 89
8 30
5505
20
51
3450
11
8 82
8 1.6
13
1.129
0.6
29
3.682
0.1
43
0.0
0
W
PM-T
HICK
LIP
(S
24)
9597
6 At
lantic
Oce
an
Gree
nland
16
92
.05
487
6047
3 94
4 30
4500
18
94
3749
12
5 88
6 1.5
62
1.231
0.7
45
3.132
0.2
49
0.0
8 14
56
4 6
391
12
9 2
14
WPM
-THI
CK
LIP
(S24
) 27
141
Atlan
tic O
cean
Gr
eenla
nd
32
90.93
46
9 68
080
1032
29
6916
19
44
3499
13
6 84
8 1.5
15
1.178
0.8
02
2.856
0.3
44
0.0
5 14
39
0 4
286
11
14
1 16
W
PM-T
HICK
LIP
(S
24)
G-27
142
Atlan
tic O
cean
Gr
eenla
nd
21
91.06
37
9 67
782
1062
30
0460
18
72
3128
14
3 72
5 1.5
67
1.041
0.7
06
2.762
0.2
36
0.0
7 8
500
7 27
7 16
24
2
12
WPM
-THI
CK
LIP
unpb
Di
s-1
Atlan
tic O
cean
Di
sko
75
88.50
31
0 86
270
1329
28
8974
20
28
3006
468
1.541
1.0
40
0.897
2.3
50
0.292
0.03
6 21
7 3
142
8 6
5
WPM
-THI
CK*
LIP
(S10
) 11
85A
9R 2
92-9
6 Pa
cific
Ocea
n On
tong J
ava P
latea
u 80
87
.36
250
9424
2 14
63
2835
14
2121
21
56
167
377
1.553
0.7
60
0.717
2.2
50
0.267
0.02
3 15
7 3
124
6 7
1 3
WPM
-THI
CK*
LIP
(S10
) 11
87A
10R
7 46-
49
Pacif
ic Oc
ean
Onton
g Jav
a Plat
eau
15
88.76
26
9 83
907
1115
28
8325
15
84
3313
15
2 74
3 1.3
28
1.149
0.9
64
1.887
0.7
31
0.0
8 6
604
8 36
2 15
48
2
30
WPM
-THI
CK*
LIP
(S10
) 11
87A
6R 6
116-
119
Pacif
ic Oc
ean
Onton
g Jav
a Plat
eau
4 89
.33
292
7992
8 12
02
2913
33
1910
24
48
157
506
1.504
0.8
40
0.672
2.3
90
0.367
0.07
25
377
7 70
4 32
70
3
41
WPM
-THI
CK
LIP
(S25
) P2
3-9
Afric
a Ka
roo
21
88.28
33
3 87
816
1245
28
7906
16
23
2681
15
5 83
6 1.4
17
0.931
0.8
18
1.848
0.5
27
0.0
4 8
491
7 33
9 11
44
2
31
WPM
-THI
CK
LIP
(S25
) P2
3-32
Af
rica
Karo
o 25
89
.49
361
7908
0 11
38
2930
22
1632
28
22
149
1106
1.4
39
0.963
0.7
61
2.063
0.4
84
0.0
5 7
303
4 24
0 9
11
2 24
W
PM-T
HICK
LIP
(S
25)
N356
Af
rica
Karo
o 29
90
.42
124
7208
2 92
2 29
6170
10
84
4392
486
1.279
1.4
83
1.069
1.5
04
0.835
0.05
4 39
0 5
276
9 9
11
W
PM-T
HICK
LIP
(S
26)
EM-4
3 Ch
ina
Emeis
han
26
90.47
25
6 72
181
1286
29
8304
26
12
2566
15
5 47
0 1.7
81
0.860
0.6
21
3.619
-0
.207
0.0
4 14
28
4 5
301
39
18
1 19
W
PM-T
HICK
LIP
(S
27)
EM-5
5 Ch
ina
Emeis
han
5 90
.59
213
7131
0 10
36
2988
30
2084
34
18
55
1 1.4
53
1.144
0.8
16
2.922
0.4
73
0.2
6 5
1869
29
12
61
163
80
24
W
PM-T
HICK
LIP
(S
27)
EM-5
7 Ch
ina
Emeis
han
4 90
.69
198
7101
3 10
20
3010
66
1728
33
99
48
7 1.4
37
1.129
0.8
02
2.433
0.5
07
0.4
0 27
30
14
41
1492
38
19
9
51
WPM
-THI
CK
LIP
(S27
) EM
-58
China
Em
eisha
n 4
90.76
20
8 69
952
1001
29
8979
17
96
3548
523
1.431
1.1
87
0.830
2.5
67
0.519
0.35
22
2640
36
11
04
186
143
60
W
PM-T
HICK
LIP
un
pb
4270
Si
beria
, Nor
ilsk
Gudc
hikhin
skay
a unit
15
79
.90
170
1466
01
1770
25
3513
19
14
2942
299
1.207
1.1
61
1.702
1.3
06
0.982
0.09
3 60
1 9
371
8 28
7 W
PM-T
HICK
LIP
un
pb
XC51
-130
Si
beria
, Nor
ilsk
Gudc
hikhin
skay
a unit
7
82.41
16
0 12
8631
16
19
2622
06
1698
30
16
37
9 1.2
58
1.150
1.4
80
1.320
0.8
77
0.1
6 6
1071
18
66
0 25
24
12
WPM
-THI
CK
LIP
unpb
XC
51-1
30(n
) Si
beria
, Nor
ilsk
Gudc
hikhin
skay
a unit
9
82.74
13
0 12
6790
15
82
2645
03
1805
31
02
200
383
1.247
1.1
73
1.487
1.4
24
0.899
0.10
5 74
7 11
46
6 12
18
2
19
Sobo
lev
et a
l, Su
ppor
ting
Onl
ine
Mat
eria
l ww
w.sc
ienc
emag
.org
/cgi
/con
tent
/full/
1138
113/
DC
1
20 A
PRIL
200
7 VO
L 31
6 SC
IEN
CE
ww
w.s
cien
cem
ag.o
rg
S7
GROU
P Ge
odyn
amic
settin
g Re
feren
ces
Samp
le Ge
ogra
phic
Lo
cality
N
Fo
Al pp
m Fe
ppm
Mn pp
m Mg
ppm
Ca pp
m Ni
ppm
Co pp
m Cr
ppm
100*
Mn/F
e 10
0*Ni
/Mg
Ni/(M
g/Fe)
/1000
10
0Ca/F
e X
px M
n/Fe
STE
Fo w
t% A
l ppm
Fe
ppm
Mn pp
m Mg
ppm
Ca pp
m Ni
ppm
Co pp
m Cr
ppm
WPM
-THI
CK
LIP
unpb
SU
-50
Sibe
ria, N
orils
k Gu
dchik
hinsk
aya u
nit
22
80.80
16
3 14
0093
16
91
2565
05
1930
30
17
32
7 1.2
07
1.176
1.6
48
1.378
0.9
83
0.0
7 3
501
6 34
1 7
19
6
WPM
-THI
CK
LIP
(S28
) AK
-31
Afric
a, Af
ar
Zawl
-Kelk
elti, E
mba T
eker
a, Er
itrea
5
87.10
37
4 96
459
1382
28
3503
19
33
2112
16
9 37
1 1.4
32
0.745
0.7
19
2.003
0.5
16
0.1
6 31
11
49
17
711
16
58
2 22
W
PM-T
HICK
LIP
(S
28)
AK-3
3 Af
rica,
Afar
Za
wl-K
elkelt
i, Emb
a Tek
era,
Eritr
ea
9 89
.34
475
7997
2 11
86
2917
58
1891
27
12
158
480
1.483
0.9
29
0.743
2.3
64
0.412
0.10
17
711
9 42
1 8
26
3 12
W
PM-T
HICK
LIP
(S
28)
AG-1
0 Af
rica,
Afar
Ad
i Ghe
bray
, Emb
a Tek
era,
Eritre
a 1
88.30
51
3 87
011
1286
28
5738
19
08
2562
15
4 49
3 1.4
77
0.897
0.7
80
2.193
0.4
23
0.0
0
W
PM-T
HICK
LIP
(S
28)
AG-1
5 Af
rica,
Afar
Ad
i Ghe
bray
, Emb
a Tek
era,
Eritre
a 17
84
.66
358
1127
74
1530
27
0733
17
43
2154
18
2 29
5 1.3
56
0.795
0.8
97
1.546
0.6
73
0.0
7 8
483
8 25
9 8
35
3 15
W
PM-T
HICK
OI
B (S
29)
RY-7
Ind
ian O
cean
Re
union
, Pito
n de
Neige
73
85
.86
224
1054
45
1567
27
8606
19
97
2377
329
1.486
0.8
53
0.900
1.8
94
0.405
0.02
3 14
9 3
104
11
5
3 W
PM-T
HICK
OI
B (S
29)
RY-1
1 Ind
ian O
cean
Re
union
, Pito
n de
Neige
13
89
.82
262
7702
8 11
25
2957
86
1675
29
70
53
4 1.4
60
1.004
0.7
74
2.174
0.4
59
0.0
8 9
545
11
444
28
38
12
W
PM-T
HICK
OI
B (S
29)
OPF
Indian
Oce
an
Reun
ion, P
iton d
e la F
ourn
aise
82
83.98
18
5 11
8585
16
96
2704
79
1988
20
16
24
3 1.4
30
0.745
0.8
84
1.677
0.5
21
0.0
3 2
186
3 10
4 5
3
2 W
PM-T
HICK
OI
B un
pb
IKI-2
2 Ha
waii
Kilau
ea Ik
i 8
88.28
25
5 87
957
1205
28
8325
17
80
3059
16
1 60
0 1.3
70
1.061
0.9
33
2.023
0.6
45
0.1
3 16
95
7 12
44
1 10
49
3
17
WPM
-THI
CK
OIB
unpb
IK
I-44
Hawa
ii Ki
lauea
Iki
20
88.54
28
8 85
808
1176
28
8577
17
42
3210
15
8 61
4 1.3
71
1.112
0.9
54
2.030
0.6
43
0.0
5 4
368
5 31
9 12
33
2
6 W
PM-T
HICK
OI
B un
pb
K97-
12
Hawa
ii Ki
lauea
Iki
34
87.81
23
6 91
085
1259
28
5538
18
68
2915
15
9 54
4 1.3
82
1.021
0.9
30
2.050
0.6
21
0.0
5 6
365
6 20
7 20
17
1
7 W
PM-T
HICK
OI
B un
pb
K97-
14b
Hawa
ii Mo
una L
oa, P
uu W
ahi
67
88.41
25
5 86
630
1195
28
7459
13
76
3023
16
1 60
7 1.3
79
1.051
0.9
11
1.589
0.6
27
0.0
3 15
4 4
220
7 14
1
7 24
7 W
PM-T
HICK
OI
B (S
30)
SR-1
7-9.0
Ha
waii
Maun
a Loa
, HSD
P-2
31
90.22
26
4 73
484
998
2949
88
1352
33
00
165
663
1.357
1.1
19
0.822
1.8
40
0.671
0.02
2 13
4 2
131
2 26
1
15
WPM
-THI
CK
OIB
(S30
) SR
-18-
0 Ha
waii
Maun
a Loa
, HSD
P-2
19
90.53
25
7 71
919
994
2993
11
1359
39
18
69
2 1.3
82
1.309
0.9
41
1.889
0.6
20
0.0
8 11
57
8 8
279
6 37
27
WPM
-THI
CK
OIB
(S30
) SR
-31-
1.0-2
.0 Ha
waii
Maun
a Loa
, HSD
P-2
7 89
.32
196
8135
6 11
33
2959
86
1791
28
83
58
4 1.3
93
0.974
0.7
92
2.201
0.5
99
0.1
3 18
93
2 20
67
5 11
8 14
3
14
WPM
-THI
CK
OIB
(S30
) SR
-98-
2.5-3
.0 Ha
waii
Maun
a Loa
, HSD
P-2
31
89.94
27
8 76
622
1031
29
7952
14
28
3928
710
1.346
1.3
18
1.010
1.8
63
0.695
0.05
7 39
4 5
275
6 49
18
WPM
-THI
CK
OIB
(S30
) SR
-98-
2.5-3
.0(n)
Ha
waii
Muan
a Loa
, HSD
P2
60
89.75
29
9 76
967
1045
29
3323
13
98
3830
15
1 73
6 1.3
58
1.306
1.0
05
1.817
0.6
70
0.0
2 4
172
3 16
5 7
26
1 11
W
PM-T
HICK
OI
B (S
30)
SR-1
01-7
.0 Ha
waii
Maun
a Loa
, HSD
P-2
50
90.00
31
8 74
911
1035
29
3507
13
85
3727
15
0 78
4 1.3
82
1.270
0.9
51
1.849
0.6
21
0.0
3 4
188
3 12
5 3
24
1 6
WPM
-THI
CK
OIB
(S30
) SR
-104
-5.5
Hawa
ii Ma
una L
oa, H
SDP-
2 33
89
.89
276
7700
6 10
54
2979
65
1405
37
00
74
5 1.3
68
1.242
0.9
56
1.824
0.6
49
0.0
5 4
387
6 21
7 3
21
9
WPM
-THI
CK
OIB
(S30
) SR
-104
-5.5(
n)
Hawa
ii Mu
ana L
oa, H
SDP2
70
89
.77
287
7719
7 10
57
2947
44
1413
37
29
144
735
1.370
1.2
65
0.977
1.8
30
0.646
0.03
2 20
3 3
149
3 13
1
8 W
PM-T
HICK
OI
B (S
30)
SR-1
13-6
.5 Ha
waii
Maun
a Loa
, HSD
P-2
63
89.58
27
3 77
943
1057
29
1517
14
09
3702
568
1.357
1.2
70
0.990
1.8
08
0.673
0.03
3 21
0 2
153
2 10
5 W
PM-T
HICK
OI
B (S
30)
SR11
7-4.0
Ha
waii
Maun
a Loa
, HSD
P-2
13
89.67
30
5 76
422
1046
28
8664
13
97
3680
836
1.368
1.2
75
0.974
1.8
28
0.649
0.11
10
782
8 37
3 6
39
15
W
PM-T
HICK
OI
B (S
30)
SR-1
20-2
.0-2.5
Ha
waii
Maun
a Loa
, HSD
P-2
35
89.65
25
8 78
973
1075
29
7548
14
29
3632
732
1.361
1.2
21
0.964
1.8
09
0.664
0.04
6 29
6 5
203
4 22
8 W
PM-T
HICK
OI
B (S
30)
SR-1
20-2
.0-2.5
(n)
Hawa
ii Mu
ana L
oa, H
SDP2
40
89
.30
275
8046
2 10
85
2923
28
1453
35
65
150
1.3
49
1.220
0.9
81
1.806
0.6
89
0.0
4 5
305
5 19
7 4
19
1 0
WPM
-THI
CK
OIB
(S30
) SR
-133
-10.0
Ha
waii
Maun
a Kea
, HSD
P2
7 87
.33
211
9519
3 13
10
2854
97
2011
28
15
42
8 1.3
76
0.986
0.9
39
2.113
0.6
33
0.1
3 9
942
11
649
23
70
12
W
PM-T
HICK
OI
B (S
30)
SR-1
34-2
.6 Ha
waii
Maun
a Kea
, HSD
P2
25
86.71
21
2 98
879
1354
28
0742
20
66
2725
17
2 40
6 1.3
70
0.971
0.9
60
2.089
0.6
46
0.0
6 6
397
6 27
0 19
13
2
9 W
PM-T
HICK
OI
B (S
30)
SR-1
36-6
.1 Ha
waii
Maun
a Kea
, HSD
P2
16
89.20
21
0 82
042
1101
29
4758
15
37
3140
630
1.342
1.0
65
0.874
1.8
73
0.703
0.08
6 55
0 8
426
21
63
12
W
PM-T
HICK
OI
B (S
30)
SR-1
38-6
.5 Ha
waii
Maun
a Kea
, HSD
P2
31
89.28
26
3 80
686
1096
29
2513
15
31
3197
15
7 66
7 1.3
59
1.093
0.8
82
1.898
0.6
68
0.0
5 5
347
6 20
6 9
33
1 8
WPM
-THI
CK
OIB
(S30
) SR
-139
-9.5
Hawa
ii Ma
una K
ea, H
SDP2
39
89
.05
243
8233
8 11
25
2912
74
1666
30
77
69
8 1.3
66
1.056
0.8
70
2.024
0.6
53
0.0
5 5
349
5 25
3 10
22
6 W
PM-T
HICK
OI
B (S
30)
SR-1
44-2
.0 Ha
waii
Maun
a Kea
, HSD
P2
23
88.58
16
8 86
738
1181
29
2769
17
19
2961
464
1.361
1.0
12
0.877
1.9
82
0.664
0.07
5 47
0 8
323
20
20
14
W
PM-T
HICK
OI
B (S
30)
SR-1
45-2
.5-2.6
Ha
waii
Maun
a Kea
, HSD
P2
39
88.55
25
3 86
149
1170
29
0005
17
28
2962
16
4 65
6 1.3
59
1.021
0.8
80
2.005
0.6
69
0.0
5 4
340
5 18
9 9
11
1 7
WPM
-THI
CK
OIB
(S30
) SR
-148
-8.5
Hawa
ii Ma
una K
ea, H
SDP2
19
88
.00
201
9007
4 12
41
2874
69
1877
26
79
55
2 1.3
77
0.932
0.8
39
2.083
0.6
30
0.0
8 7
561
9 38
5 29
21
13
WPM
-THI
CK
OIB
(S30
) SR
-149
-1.0
Hawa
ii Ma
una K
ea, H
SDP2
31
87
.92
190
9145
4 12
58
2895
28
1908
26
55
54
8 1.3
75
0.917
0.8
39
2.086
0.6
34
0.0
6 8
414
7 26
7 14
19
7 W
PM-T
HICK
OI
B (S
30)
SR-1
57-6
.1 Ha
waii
Maun
a Kea
, HSD
P2
24
89.11
19
6 82
064
1126
29
2116
16
40
3079
596
1.372
1.0
54
0.865
1.9
99
0.641
0.06
4 42
3 7
272
21
23
13
W
PM-T
HICK
OI
B (S
30)
SR-1
59-2
.6 Ha
waii
Maun
a Kea
, HSD
P2
42
88.75
24
4 84
807
1154
29
1029
16
48
3036
15
9 64
3 1.3
60
1.043
0.8
85
1.943
0.6
65
0.0
4 4
288
3 21
1 9
14
1 8
WPM
-THI
CK
OIB
(S30
) SR
-170
-1.2
Hawa
ii Ma
una K
ea, H
SDP2
19
88
.30
213
8827
1 11
93
2897
85
1648
29
40
62
7 1.3
51
1.014
0.8
95
1.867
0.6
85
0.0
7 8
530
8 34
7 20
40
16
WPM
-THI
CK
OIB
(S30
) SR
-170
-3.1
Hawa
ii Ma
una K
ea, H
SDP2
46
87
.98
218
8982
2 12
17
2861
90
1668
29
33
62
9 1.3
55
1.025
0.9
21
1.857
0.6
77
0.0
4 5
282
4 18
5 11
16
7 W
PM-T
HICK
OI
B (S
30)
SR-1
75-5
.2 Ha
waii
Maun
a Kea
, HSD
P2
36
88.95
24
1 83
748
1123
29
3387
15
46
3210
696
1.340
1.0
94
0.916
1.8
45
0.707
0.05
4 35
2 6
214
16
26
11
W
PM-T
HICK
OI
B (S
30)
SR-1
84-3
.0 Ha
waii
Maun
a Kea
, HSD
P2
4 89
.35
201
8022
2 11
15
2930
05
1746
30
96
165
513
1.390
1.0
57
0.848
2.1
76
0.604
0.29
10
2119
9
1187
81
8
6 9
WPM
-THI
CK
OIB
(S30
) SR
-188
-6.5
Hawa
ii Ma
una K
ea, H
SDP2
6
89.55
24
0 78
725
1053
29
3534
15
41
3340
15
2 55
6 1.3
38
1.138
0.8
96
1.958
0.7
12
0.1
5 10
11
09
25
654
48
213
6 30
W
PM-T
HICK
OI
B (S
30)
SR-2
60-8
.5 Ha
waii
Maun
a Kea
, HSD
P2
5 89
.73
267
7739
6 10
04
2941
14
1434
38
47
134
590
1.297
1.3
08
1.012
1.8
52
0.797
0.13
9 98
0 11
56
7 14
53
6
14
WPM
-THI
CK
OIB
(S30
) SR
-277
-7.5-
8.0
Hawa
ii Ma
una K
ea, H
SDP2
2
90.86
19
8 69
347
906
2998
34
1361
36
93
146
380
1.307
1.2
32
0.854
1.9
63
0.777
0.14
13
999
0 70
9 11
55
1
38
WPM
-THI
CK
OIB
(S30
) SR
-291
-3.0-
3.8
Hawa
ii Ma
una K
ea, H
SDP2
38
89
.24
201
8111
4 11
15
2926
28
1645
33
27
149
603
1.374
1.1
37
0.922
2.0
28
0.637
0.04
7 31
6 4
210
12
18
1 12
W
PM-T
HICK
OI
B (S
30)
SR-2
98-6
.0 Ha
waii
Maun
a Kea
, HSD
P2
27
89.03
24
7 82
446
1102
29
1201
15
75
3328
14
9 66
4 1.3
37
1.143
0.9
42
1.910
0.7
15
0.0
6 6
424
8 23
6 10
49
2
11
WPM
-THI
CK
OIB
(S30
) SR
-308
-0-0
.6 Ha
waii
Maun
a Kea
, HSD
P2
12
89.10
27
7 81
993
1099
29
1738
15
57
3478
15
1 63
9 1.3
40
1.192
0.9
77
1.899
0.7
07
0.1
3 6
952
17
611
26
108
3 8
WPM
-THI
CK
OIB
(S30
) SR
-310
-2.4-
2.8
Hawa
ii Ma
una K
ea, H
SDP2
70
89
.51
265
7943
7 10
82
2947
97
1548
33
05
150
742
1.362
1.1
21
0.891
1.9
49
0.662
0.03
3 19
2 3
130
2 14
1
5 W
PM-T
HICK
OI
B (S
30)
SR-3
19-7
.5 Ha
waii
Maun
a Kea
, HSD
P2
42
89.02
25
7 82
543
1097
29
1302
15
51
3384
15
3 68
0 1.3
29
1.162
0.9
59
1.879
0.7
30
0.0
4 5
316
5 20
2 5
26
1 10
W
PM-T
HICK
OI
B (S
30)
SR-3
32-9
.8-10
.0 Ha
waii
Maun
a Kea
, HSD
P2
7 90
.31
222
7333
6 98
1 29
7371
14
76
4121
13
7 48
7 1.3
38
1.386
1.0
16
2.013
0.7
11
0.1
4 9
1015
15
62
3 15
11
2 2
21
WPM
-THI
CK
OIB
(S30
) SR
-335
-3.2-
3.4
Hawa
ii Ma
una K
ea, H
SDP2
55
89
.89
286
7596
6 10
37
2937
64
1500
33
45
149
785
1.365
1.1
39
0.865
1.9
75
0.656
0.04
4 26
8 4
201
4 25
1
9 W
PM-T
HICK
OI
B (S
30)
SR-3
37-3
.4-3.7
Ha
waii
Maun
a Kea
, HSD
P2
58
89.78
26
5 76
896
1041
29
4088
15
04
3344
15
0 74
0 1.3
54
1.137
0.8
74
1.956
0.6
78
0.0
4 5
318
4 18
7 6
25
1 10
W
PM-T
HICK
OI
B (S
30)
SR-3
53-7
.2-7.4
Ha
waii
Maun
a Kea
, HSD
P2
62
88.01
22
9 89
908
1210
28
7179
16
60
3108
16
5 55
3 1.3
45
1.082
0.9
73
1.846
0.6
97
0.0
3 3
236
3 17
7 9
22
1 6
WPM
-THI
CK
OIB
(S30
) SR
-361
-2.5-
2.8
Hawa
ii Ma
una K
ea, H
SDP2
26
89
.66
231
7798
5 10
50
2942
11
1498
35
12
148
602
1.346
1.1
94
0.931
1.9
21
0.695
0.04
5 28
6 4
173
5 12
1
11
WPM
-THI
CK
OIB
(S30
) SR
-363
-7.0-
8.0
Hawa
ii Ma
una K
ea, H
SDP2
57
89
.03
228
8275
1 11
07
2923
65
1584
33
46
151
603
1.337
1.1
44
0.947
1.9
14
0.713
0.03
6 22
6 4
155
4 14
1
13
WPM
-THI
CK
OIB
(S30
) SR
-375
-0.5-
1.0
Hawa
ii Ma
una K
ea, H
SDP2
20
89
.65
243
7802
7 10
48
2942
22
1534
34
68
146
738
1.343
1.1
79
0.920
1.9
67
0.702
0.06
8 44
4 6
244
6 32
1
11
WPM
-THI
CK
OIB
(S30
) SR
-377
-6.7-
6.9
Hawa
ii Ma
una K
ea, H
SDP2
38
89
.36
276
7998
9 10
60
2923
96
1534
36
17
146
731
1.325
1.2
37
0.990
1.9
18
0.739
0.04
5 25
7 3
210
11
8 1
11
WPM
-THI
CK
OIB
(S30
) SR
-382
-4.0-
4.4
Hawa
ii Ma
una K
ea, H
SDP2
19
88
.26
238
8764
8 11
66
2868
14
1619
32
56
156
660
1.330
1.1
35
0.995
1.8
47
0.728
0.08
11
561
8 32
0 9
28
2 24
Sobo
lev
et a
l, Su
ppor
ting
Onl
ine
Mat
eria
l ww
w.sc
ienc
emag
.org
/cgi
/con
tent
/full/
1138
113/
DC
1
20 A
PRIL
200
7 VO
L 31
6 SC
IEN
CE
ww
w.s
cien
cem
ag.o
rg
S8
GROU
P Ge
odyn
amic
settin
g Re
feren
ces
Samp
le Ge
ogra
phic
Lo
cality
N
Fo
Al pp
m Fe
ppm
Mn pp
m Mg
ppm
Ca pp
m Ni
ppm
Co pp
m Cr
ppm
100*
Mn/F
e 10
0*Ni
/Mg
Ni/(M
g/Fe)
/1000
10
0Ca/F
e X
px M
n/Fe
STE
Fo w
t% A
l ppm
Fe
ppm
Mn pp
m Mg
ppm
Ca pp
m Ni
ppm
Co pp
m Cr
ppm
WPM
-THI
CK
OIB
(S30
) SR
-389
-1.8-
2.0
Hawa
ii Ma
una K
ea, H
SDP2
35
89
.76
303
7720
7 10
41
2944
10
1505
34
25
147
795
1.349
1.1
63
0.898
1.9
49
0.690
0.04
2 29
7 4
232
3 42
2
11
WPM
-THI
CK
OIB
(S30
) SR
-565
-0.7
Hawa
ii Ma
una K
ea, H
SDP2
4
90.28
28
7 73
350
962
2963
80
1261
41
17
149
547
1.312
1.3
89
1.019
1.7
20
0.766
0.01
8 20
5 5
610
12
12
6 3
WPM
-THI
CK
OIB
(S30
) SR
-682
-9.8-
10
Hawa
ii Ma
una K
ea, H
SDP2
50
88
.70
289
8469
0 11
38
2893
80
1576
32
12
158
723
1.343
1.1
10
0.940
1.8
61
0.701
0.03
4 24
2 4
163
6 17
1
5 W
PM-T
HICK
OI
B (S
30)
SR-6
84-5
.0-5.5
Ha
waii
Maun
a Kea
, HSD
P2
28
88.91
28
4 83
330
1111
29
0795
15
50
3209
16
1 73
1 1.3
33
1.103
0.9
20
1.860
0.7
23
0.0
4 4
274
4 18
4 6
31
2 8
WPM
-THI
CK
OIB
(S30
) SR
-690
-4.0
Hawa
ii Ma
una K
ea, H
SDP2
16
89
.67
307
7786
9 10
32
2941
48
1587
36
51
150
723
1.325
1.2
41
0.967
2.0
38
0.738
0.08
4 62
2 5
293
10
18
3 10
W
PM-T
HICK
OI
B (S
30)
SR-6
94-3
.0-3.2
Ha
waii
Maun
a Kea
, HSD
P2
2 89
.01
278
8282
2 11
04
2920
28
1601
35
44
157
770
1.332
1.2
14
1.005
1.9
33
0.723
0.07
13
428
4 52
5 7
63
6 65
W
PM-T
HICK
OI
B (S
30)
SR-9
30-9
.1 Ha
waii
Maun
a Kea
, HSD
P2
10
89.28
27
6 80
768
1122
29
2698
15
45
3395
15
5 69
5 1.3
89
1.160
0.9
37
1.913
0.6
05
0.1
0 9
746
10
575
27
85
3 27
W
PM-T
HICK
OI
B (S
30)
SR-9
42-5
.7 Ha
waii
Maun
a Kea
, HSD
P2
40
88.14
22
9 88
920
1206
28
7592
14
68
3102
16
4 60
1 1.3
56
1.079
0.9
59
1.651
0.6
74
0.0
4 5
267
4 19
0 7
10
2 8
WPM
-THI
CK
OIB
(S30
) SR
-954
-10.7
Ha
waii
Maun
a Kea
, HSD
P2
48
89.52
28
9 78
227
1059
29
0821
14
87
3631
14
7 77
9 1.3
53
1.249
0.9
77
1.900
0.6
80
0.0
3 4
212
3 15
3 2
13
1 5
WPM
-THI
CK
OIB
(S30
) SR
-961
-14.0
Ha
waii
Maun
a Kea
, HSD
P2
44
89.44
26
1 79
445
1082
29
2711
15
06
3537
765
1.362
1.2
08
0.960
1.8
95
0.662
0.04
4 31
1 4
195
3 29
7 W
PM-T
HICK
OI
B (S
30)
SR-9
62-1
8.1
Hawa
ii Ma
una K
ea, H
SDP2
66
89
.51
284
7902
9 10
75
2933
75
1488
35
80
152
773
1.361
1.2
20
0.964
1.8
82
0.665
0.03
4 20
8 3
147
3 15
1
5 W
PM-T
HICK
OI
B (S
31)
KI-1
1 Ha
waii
Maun
a Kea
post
shiel
d 18
86
.98
132
9816
6 14
23
2854
90
2424
22
33
36
6 1.4
50
0.782
0.7
68
2.470
0.4
80
0.0
7 6
472
8 30
9 20
14
9 W
PM-T
HICK
OI
B (S
31)
KI-8
Ha
waii
Maun
a Kea
post
shiel
d 9
88.26
18
1 89
202
1229
29
1934
18
32
2448
536
1.378
0.8
39
0.748
2.0
54
0.630
0.12
8 81
8 14
60
4 30
60
10
WPM
-THI
CK
OIB
(S31
) KI
-4
Hawa
ii Ma
una K
ea po
st sh
ield
11
87.74
13
9 92
943
1295
28
9526
20
43
2588
446
1.393
0.8
94
0.831
2.1
98
0.598
0.11
9 79
1 11
49
9 68
30
18
WPM
-THI
CK
OIB
(S31
) KI
-3
Hawa
ii Ma
una K
ea po
st sh
ield
8 88
.79
251
8497
9 11
50
2927
85
1590
29
16
65
4 1.3
53
0.996
0.8
46
1.871
0.6
80
0.1
2 11
90
1 17
59
5 29
10
0
13
WPM
-THI
CK
OIB
(S31
) KI
-3(n
) Ha
waii
Maun
a Kea
, pos
t shie
ld 5
88.82
23
4 84
025
1145
29
0374
15
62
2951
15
9 64
2 1.3
62
1.016
0.8
54
1.859
0.6
61
0.2
1 11
15
10
28
1002
68
13
4 3
20
WPM
-THI
CK
OIB
(S32
) Ko
o-10
Ha
waii
Koola
u, Ma
kapu
u 7
87.13
17
9 96
799
1263
28
5148
12
74
2967
479
1.305
1.0
40
1.007
1.3
16
0.780
0.14
7 10
30
10
495
8 24
8 W
PM-T
HICK
OI
B (S
32)
Koo-
17a
Hawa
ii Ko
olau,
Maka
puu
19
89.11
24
1 82
175
1085
29
2709
11
99
3910
613
1.321
1.3
36
1.098
1.4
59
0.747
0.08
6 54
1 11
41
7 13
98
12
WPM
-THI
CK
OIB
(S32
) Ko
o-19
a Ha
waii
Koola
u, Ma
kapu
u 43
88
.82
248
8427
6 11
11
2912
60
1257
37
07
61
0 1.3
18
1.273
1.0
73
1.492
0.7
53
0.0
4 4
275
5 17
0 8
36
6
WPM
-THI
CK
OIB
(S32
) Ko
o-30
a Ha
waii
Koola
u, Ma
kapu
u 2
86.74
18
0 99
604
1348
28
3591
13
54
3505
404
1.353
1.2
36
1.231
1.3
60
0.681
0.04
11
459
8 36
2 32
47
21
WPM
-THI
CK
OIB
(S32
) Ko
o-49
Ha
waii
Koola
u, Ma
kapu
u 34
89
.00
235
8300
5 10
96
2921
94
1272
38
03
57
9 1.3
20
1.302
1.0
80
1.532
0.7
48
0.0
3 5
227
6 21
2 7
37
10
W
PM-T
HICK
OI
B (S
32)
Koo-
55
Hawa
ii Ko
olau,
Maka
puu
41
88.52
21
5 87
012
1156
29
2015
12
94
3584
557
1.329
1.2
27
1.068
1.4
88
0.731
0.03
5 21
3 4
165
6 18
7 W
PM-T
HICK
OI
B un
pb
S10(
i) Ha
waii
Koola
u, Ma
kapu
u 23
86
.78
194
9874
5 12
89
2819
59
1239
31
29
171
475
1.306
1.1
10
1.096
1.2
54
0.751
0.04
3 30
7 5
213
6 35
1
6 W
PM-T
HICK
OI
B un
pb
S10(
ii) Ha
waii
Koola
u, Ma
kapu
u 2
87.86
23
0 90
851
1142
28
6085
11
94
4503
16
9 48
2 1.2
57
1.574
1.4
30
1.314
0.8
79
0.3
5 3
2433
12
16
67
21
149
5 24
W
PM-T
HICK
OI
B (S
33)
R8-1
.1-11
97.1
Hawa
ii Ko
olau,
KSDP
23
88
.67
248
8515
7 11
36
2899
45
1366
35
16
152
620
1.334
1.2
13
1.033
1.6
04
0.720
0.04
6 29
9 5
323
14
49
2 7
WPM
-THI
CK
OIB
(S34
) LO
-02-
04
Hawa
ii Lo
ihi sm
t 28
87
.70
236
9199
3 12
83
2853
14
1984
27
53
52
8 1.3
94
0.965
0.8
88
2.156
0.5
95
0.0
5 7
395
7 24
8 22
29
10
WPM
-THI
CK
OIB
(S34
) LO
-02-
02
Hawa
ii Lo
ihi sm
t 19
88
.29
219
8815
2 12
33
2890
83
2036
28
11
54
2 1.3
99
0.972
0.8
57
2.309
0.5
85
0.0
7 5
515
8 28
2 66
35
17
WPM
-THI
CK
OIB
(S35
) 15
8-9
Hawa
ii Lo
ihi sm
t 44
89
.15
239
8197
7 11
66
2929
71
1781
31
00
65
9 1.4
23
1.058
0.8
67
2.172
0.5
36
0.0
3 5
193
3 13
1 15
12
6 W
PM-T
HICK
OI
B (S
35)
186-
5 Ha
waii
Loihi
smt
3 89
.43
309
7963
7 11
31
2933
26
2015
28
79
71
2 1.4
20
0.981
0.7
82
2.531
0.5
42
0.3
4 14
24
15
45
1594
84
12
3
51
WPM
-THI
CK
OIB
(S35
) 18
7-1
Hawa
ii Lo
ihi sm
t 7
89.67
30
3 78
190
1095
29
5195
17
74
3242
679
1.401
1.0
98
0.859
2.2
69
0.582
0.12
16
898
14
488
36
121
29
W
PM-T
HICK
OI
B (S
36)
M234
3-7
Hawa
ii Lo
ihi sm
t 9
88.14
21
5 89
237
1319
28
8692
21
94
2742
483
1.478
0.9
50
0.848
2.4
59
0.421
0.13
17
973
14
586
26
11
12
W
PM-T
HICK
OI
B (S
36)
M233
5-12
Ha
waii
Loihi
smt
13
89.44
25
8 79
807
1143
29
4008
18
87
3091
683
1.432
1.0
51
0.839
2.3
65
0.516
0.08
11
619
11
378
29
28
23
W
PM-T
HICK
OI
B (S
36)
M234
0-3
Hawa
ii Lo
ihi sm
t 3
87.56
22
2 93
624
1291
28
6763
19
11
3023
511
1.379
1.0
54
0.987
2.0
41
0.628
0.32
26
2393
37
12
03
88
156
39
W
PM-T
HICK
OI
B (S
36)
M234
3-10
Ha
waii
Loihi
smt
37
87.16
21
0 95
978
1382
28
3538
22
72
2781
350
1.440
0.9
81
0.941
2.3
67
0.500
0.03
5 18
6 4
157
11
8
8 W
PM-T
HICK
OI
B (S
36)
M234
3-8
Hawa
ii Lo
ihi sm
t 5
88.16
21
1 88
961
1295
28
8238
21
80
2804
434
1.456
0.9
73
0.865
2.4
50
0.468
0.13
37
914
11
569
51
9
36
WPM
-THI
CK
OIB
(S36
) M2
340-
1 Ha
waii
Loihi
smt
3 88
.21
279
8834
6 12
21
2876
78
1899
29
62
62
0 1.3
82
1.030
0.9
10
2.149
0.6
20
0.1
7 40
13
40
14
269
43
99
28
W
PM-T
HICK
OI
B (S
36)
MK23
37-4
Ha
waii
Loihi
smt
19
87.11
22
4 96
131
1351
28
2756
18
26
2823
468
1.405
0.9
98
0.960
1.9
00
0.573
0.07
10
524
10
364
43
18
11
W
PM-T
HICK
OI
B (S
36)
M234
3-9
Hawa
ii Lo
ihi sm
t 17
87
.09
228
9672
6 13
58
2839
61
2223
25
77
44
5 1.4
04
0.908
0.8
78
2.298
0.5
76
0.0
8 6
585
12
312
64
40
10
W
PM-T
HICK
OI
B (S
36)
M234
3-5
Hawa
ii Lo
ihi sm
t 3
88.87
21
7 83
863
1242
29
1338
22
11
2792
504
1.481
0.9
58
0.804
2.6
36
0.416
0.30
19
2098
23
15
55
74
36
30
W
PM-T
HICK
OI
B (S
36)
M233
6-5
Hawa
ii Lo
ihi sm
t 10
86
.86
220
9850
9 13
85
2832
69
2150
26
95
38
5 1.4
06
0.951
0.9
37
2.183
0.5
70
0.1
0 9
734
13
499
44
56
14
W
PM-T
HICK
OI
B (S
36)
M233
6-4
Hawa
ii Lo
ihi sm
t 21
87
.57
235
9349
6 12
98
2865
00
1996
27
14
53
4 1.3
88
0.947
0.8
86
2.135
0.6
08
0.0
6 5
425
9 30
4 27
20
14
WPM
-THI
CK
OIB
(S36
) M2
338-
5 Ha
waii
Loihi
smt
5 88
.54
231
8629
0 12
21
2902
33
2311
31
49
42
3 1.4
14
1.085
0.9
36
2.679
0.5
53
0.1
9 12
13
56
27
891
22
67
4
WPM
-THI
CK
OIB
(S36
) M2
336-
7 Ha
waii
Loihi
smt
6 87
.20
200
9550
0 13
68
2830
74
2372
25
00
35
4 1.4
33
0.883
0.8
43
2.483
0.5
16
0.1
8 20
13
41
16
790
66
96
29
W
PM-T
HICK
OI
B (S
36)
M233
5-R
Hawa
ii Lo
ihi sm
t 22
89
.42
232
7964
9 11
43
2929
60
1962
30
15
64
1 1.4
35
1.029
0.8
20
2.463
0.5
11
0.0
6 7
391
9 37
9 35
32
15
WPM
-THI
CK
OIB
(S36
) M2
335-
2 Ha
waii
Loihi
smt
5 87
.41
196
9404
8 13
34
2840
70
2341
28
70
32
6 1.4
18
1.010
0.9
50
2.489
0.5
46
0.1
9 4
1438
27
65
9 49
72
19
WPM
-THI
CK
OIB
(S36
) M2
335-
8 Ha
waii
Loihi
smt
23
86.21
19
2 10
3052
14
62
2803
46
2559
24
38
29
8 1.4
19
0.870
0.8
96
2.483
0.5
45
0.0
6 6
420
8 28
0 39
61
11
WPM
-THI
CK
OIB
(S36
) M2
343-
4 Ha
waii
Loihi
smt
13
87.93
20
7 90
549
1334
28
6982
20
45
2715
428
1.473
0.9
46
0.857
2.2
58
0.432
0.09
17
674
12
429
123
42
25
W
PM-T
HICK
OI
B un
pb
5328
Ha
waii
Kaua
i-Hae
na
1 88
.92
259
8303
9 10
46
2899
96
1401
38
74
55
4 1.2
59
1.336
1.1
09
1.687
0.8
75
0.0
0 0
0 0
0 0
0
0 W
PM-T
HICK
OI
B un
pb
0433
c-033
R Pa
cific
Ocea
n Su
iko sm
t 11
88
.46
326
8637
1 11
56
2879
78
1559
33
70
162
655
1.338
1.1
70
1.011
1.8
05
0.711
0.08
9 61
0 9
530
21
19
2 13
W
PM-T
HICK
OI
B un
pb
0433
c-024
R Pa
cific
Ocea
n Su
iko sm
t 21
86
.77
280
9805
4 12
94
2798
60
1627
29
43
174
552
1.319
1.0
51
1.031
1.6
59
0.750
0.06
7 46
0 7
317
12
37
2 12
W
PM-T
HICK
OI
B (S
37)
LG-9
9-58
At
lantic
Oce
an
La G
omer
a 5
89.13
25
2 81
918
1209
29
2195
19
05
2839
15
5 47
7 1.4
76
0.972
0.7
96
2.326
0.4
25
0.1
0 11
71
2 6
557
46
160
3 12
W
PM-T
HICK
OI
B (S
37)
GC-9
8-59
At
lantic
Oce
an
Gran
Can
aria
5 89
.46
221
7978
0 11
82
2946
70
1711
30
46
153
464
1.481
1.0
34
0.825
2.1
44
0.415
0.14
5 87
1 7
1514
37
96
3
6 W
PM-T
HICK
OI
B (S
37)
GC-9
8-35
At
lantic
Oce
an
Gran
Can
aria
5 89
.21
180
8115
2 11
96
2918
39
1753
29
97
153
434
1.474
1.0
27
0.833
2.1
61
0.430
0.16
8 10
22
23
1094
60
20
9 4
15
WPM
-THI
CK
OIB
(S37
) GC
-126
5 At
lantic
Oce
an
Gran
Can
aria
3 89
.56
171
7862
9 12
19
2936
71
2085
27
05
152
367
1.551
0.9
21
0.724
2.6
52
0.271
0.23
19
1481
21
17
05
81
45
1 37
W
PM-T
HICK
OI
B (S
38)
953C
-101
R-5 1
16-1
23
Atlan
tic O
cean
Gr
an C
anar
ia 5
88.23
19
7 87
674
1202
28
6031
13
52
3547
15
4 48
6 1.3
71
1.240
1.0
87
1.542
0.6
44
0.1
7 8
1214
18
77
5 44
71
4
13
WPM
-THI
CK
OIB
(S38
) 95
3C-9
3R-0
5 13-
21
Atlan
tic O
cean
Gr
an C
anar
ia 14
88
.67
222
8479
5 11
61
2886
63
1447
34
71
152
518
1.369
1.2
03
1.020
1.7
06
0.648
0.08
3 55
5 8
335
6 39
3
10
Sobo
lev
et a
l, Su
ppor
ting
Onl
ine
Mat
eria
l ww
w.sc
ienc
emag
.org
/cgi
/con
tent
/full/
1138
113/
DC
1
20 A
PRIL
200
7 VO
L 31
6 SC
IEN
CE
ww
w.s
cien
cem
ag.o
rg
S9
GROU
P Ge
odyn
amic
settin
g Re
feren
ces
Samp
le Ge
ogra
phic
Lo
cality
N
Fo
Al pp
m Fe
ppm
Mn pp
m Mg
ppm
Ca pp
m Ni
ppm
Co pp
m Cr
ppm
100*
Mn/F
e 10
0*Ni
/Mg
Ni/(M
g/Fe)
/1000
10
0Ca/F
e X
px M
n/Fe
STE
Fo w
t% A
l ppm
Fe
ppm
Mn pp
m Mg
ppm
Ca pp
m Ni
ppm
Co pp
m Cr
ppm
WPM
-THI
CK
OIB
(S38
) 95
3C-9
3R-0
6 45-
55
Atlan
tic O
cean
Gr
an C
anar
ia 4
82.31
16
7 12
8922
17
48
2610
52
1930
24
85
187
219
1.356
0.9
52
1.227
1.4
97
0.674
0.20
12
1377
18
90
8 61
74
5
6 W
PM-T
HICK
OI
B (S
38)
953C
-93R
-4 18
-24
Atlan
tic O
cean
Gr
an C
anar
ia 5
87.75
18
1 91
591
1300
28
5495
16
02
3239
16
8 41
6 1.4
19
1.135
1.0
39
1.749
0.5
44
0.1
8 13
12
84
16
702
53
180
3 26
KO
MATI
ITES
KO
MATI
ITES
(S
39)
MUN-
24
Cana
da
Munr
o 2.7
Ga
65
92.34
39
2 58
523
925
3070
10
1562
33
16
14
41
1.580
1.0
80
0.632
2.6
69
0.210
0.03
4 22
3 4
182
4 4
6
KOMA
TIIT
ES
KOMA
TIIT
ES
(S39
) MU
N-41
4 Ca
nada
Mu
nro 2
.7 Ga
63
89
.42
254
8021
8 12
38
2948
87
1625
29
42
73
2 1.5
43
0.998
0.8
00
2.026
0.2
88
0.0
3 3
205
4 16
6 5
5
8 KO
MATI
ITES
KO
MATI
ITES
(S
39)
Ch-1
1a
Cana
da
Munr
o 2.7
Ga
32
88.82
29
9 84
803
1339
29
3033
16
45
3034
791
1.579
1.0
35
0.878
1.9
40
0.212
0.04
5 32
7 7
214
8 6
16
KO
MATI
ITES
KO
MATI
ITES
(S
40)
M-62
6 Ca
nada
Al
exo 2
.7 Ga
11
92
.81
334
5513
2 86
2 30
9510
14
90
3461
1299
1.5
64
1.118
0.6
17
2.703
0.2
43
0.1
0 7
772
13
474
7 10
20
KOMA
TIIT
ES
KOMA
TIIT
ES
(S40
) M-
767
Cana
da
Alex
o 2.7
Ga
8 93
.53
390
4982
9 80
1 31
3247
14
84
3481
1308
1.6
07
1.111
0.5
54
2.978
0.1
55
0.1
4 14
10
82
13
575
10
41
30
KO
MATI
ITES
KO
MATI
ITES
(S
40)
M-66
6 Ca
nada
Al
exo 2
.7 Ga
33
94
.30
421
4388
1 71
9 31
5979
14
84
3353
1263
1.6
38
1.061
0.4
66
3.381
0.0
91
0.0
5 9
404
5 27
4 5
6
20
KOMA
TIIT
ES
KOMA
TIIT
ES
(S41
) MZ
-4
Zimba
bwe
Belin
gwe 2
.7 Ga
9
92.60
36
1 56
373
885
3071
89
1494
36
30
128
1327
1.5
71
1.182
0.6
66
2.651
0.2
30
0.1
1 10
79
6 13
47
7 6
8 2
32
KOMA
TIIT
ES
KOMA
TIIT
ES
(S42
) G-
21
Cana
da
Gilm
our 1
.9 Ga
15
92
.03
359
6084
1 94
8 30
5902
17
62
3351
1058
1.5
58
1.095
0.6
67
2.897
0.2
56
0.0
8 8
626
9 32
7 11
18
25
KOMA
TIIT
ES
KOMA
TIIT
ES
(S42
) G-
18
Cana
da
Gilm
our 1
.9 Ga
2
91.83
35
2 62
561
968
3058
04
1787
33
04
10
71
1.547
1.0
80
0.676
2.8
56
0.278
0.36
3 26
62
46
1468
14
35
38
KOMA
TIIT
ES
KOMA
TIIT
ES
(S42
) G-
15
Cana
da
Gilm
our 1
.9 Ga
63
88
.74
222
8527
7 12
81
2924
37
1796
21
90
65
2 1.5
03
0.749
0.6
39
2.106
0.3
71
0.0
4 3
263
4 15
6 2
3
5 KO
MATI
ITES
KO
MATI
ITES
(S
42)
G-8a
Ca
nada
Gi
lmou
r 1.9
Ga
70
88.67
24
4 85
637
1284
29
1770
17
93
2617
675
1.500
0.8
97
0.768
2.0
94
0.377
0.03
6 20
9 3
124
3 5
9
KOMA
TIIT
ES
KOMA
TIIT
ES
(S43
) GO
R-94
-19
Gorg
ona
Gorg
ona 0
.9 Ga
71
91
.44
549
6460
7 10
51
3004
52
2140
35
93
92
0 1.6
26
1.196
0.7
73
3.313
0.1
15
0.0
2 4
126
2 11
7 2
5
5 KO
MATI
ITES
KO
MATI
ITES
(S
43)
GOR9
4-35
Go
rgon
a Go
rgon
a 0.9
Ga
13
92.58
53
3 56
770
930
3083
31
2120
34
47
11
67
1.638
1.1
18
0.635
3.7
34
0.090
0.08
10
613
11
396
10
16
41
KO
MATI
ITES
KO
MATI
ITES
(S
43)
GOR9
4-34
Go
rgon
a Go
rgon
a 0.9
Ga
27
91.50
47
3 64
764
1042
30
3202
22
74
3247
1003
1.6
09
1.071
0.6
94
3.511
0.1
50
0.0
5 12
40
2 7
256
12
12
15
KO
MATI
ITES
KO
MATI
ITES
(S
43)
GOR9
4-32
Go
rgon
a Go
rgon
a 0.9
Ga
18
92.79
54
4 54
632
907
3059
20
2121
34
82
11
95
1.661
1.1
38
0.622
3.8
82
0.043
0.07
18
527
9 26
6 12
10
29
Note
s for
Tab
le S2
a. Co
ncen
tratio
ns an
d stan
dard
erro
rs of
mean
(STE
) are
show
n for
elem
ents
in pp
m an
d for
Fo i
n mol%
. N-
numb
er o
f ave
rage
d hig
h-Mg
oliv
ines (
see
text fo
r defi
nition
). Sa
mples
mar
ked
(i) a
nd (i
i) re
pres
ent d
iffere
nt co
mpos
itiona
l gro
ups o
f oliv
ine w
ithin
the sa
me sa
mple.
Sam
ples m
arke
d (n
) rep
rese
nt an
add
itiona
l inde
pend
ent p
opula
tion
of oli
vines
analy
zed f
or th
e sam
e sam
ple. R
efere
nces
inclu
de sa
mples
desc
riptio
n, un
pb- u
npub
lishe
d. *-
Onton
g Jav
a Plat
eau s
ample
s wer
e arb
itrary
class
ified a
s WPM
-THI
CK, b
ut ac
tually
may
belon
g to W
PM-T
HIN
at the
time o
f form
ation
.
7 2:21 PM
We hope that you enjoy this section of Earth Pages. It gives one geologist's personal views about some ofthe major stories that appeared in leading journals, such as Geology, Science and Nature, in the last month. The items have been written and compiled by Steve Drury author of a number of Earth science books, including Stepping Stones: The Making of our Home World (Oxford University Press, 1999) and Image Interpretation in Geology (3rd edition) (Taylor & Francis, and Blackwell Science (USA), 2001).
If you have comments on this month's items you can contact the author at [email protected]
• News Archive
News for May 2007Anthropology and geoarchaeologyClimate change and palaeoclimatologyGeobiology, palaeontology, and evolutionGeochemistry, mineralogy, petrology and volcanologyPlanetary, extraterrestrial geology, and meteoriticsTectonics
Anthropology and geoarchaeology
Do Neanderthals sit next to us on the train?
May 2007
Many might answer, ‘Yes, and they speak loudly about their love lives into mobile phones’, but that is a tiredold joke. A much better one is that related by Steve Jones of University College, London. Were an unwashedbut shell-suited, Late Palaeolithic, fully human hunter-gatherer to sit next to us, we would probably changeseats. Jones believes that if our companion turned out to be a freshly showered, shaved and eau decologned Neanderthal in a business suit, we would change trains. Neanderthals were impressive, in themanner of all-in wrestlers with extremely large noses and eyebrows.
A boy’s skeleton turned up in Portuguese 24 ka cave deposits in the late 1990s. The lad had the hallmarkchin of a modern human but the stocky body and short legs of a Neanderthal. He may be the only tangibleevidence of a human inter-species hybrid. There again, he may have been a perfectly normal, stocky boywith short legs. Yet the find re-opened the possibility that Neanderthal genes may have made their way intomodern humans. It certainly does not look like it from the available DNA fragments extracted fromNeanderthal bones, but ongoing attempts to sequence the Neanderthal genome (see Neanderthal genome on the cards in January 2007 EPN) could resolve the issue. But the ambitious genetic plans have sent a thrill through scientific journalists with palaeoanthropological leanings (Jones, D. 2007. The Neanderthal within. New Scientist, v. 193 (3 March 2007), p. 2832).
Another skull claimed to show hybrid features has turned up in Romania, but the most tantalising hints come from existing knowledge of human genetics. The Out of Africa model for all modern humans is based on studies of mtDNA and that from Y chromosomes, which now enable human migrations to be tracked and put into a time frame. But such genetic material forms a tiny proportion of the human genome. It is nuclear DNAthat dominates and is also responsible for how we function and how we look. There is so much of it that work has only just begun in the context of human origins. One haplotype, in the PDHA1 gene, has shown up something odd in a small sample of men from different continents. Two lineages seem to be represented, one that a molecular clock dates to a last common ancestor 1.8 Ma ago [Homo habilis?], the other having split at about 200 ka. That duality should not be present if all living people descended from a small group who lived around 160 ka. Either it resurrects the almost-buried multiregional model, or points to occasional interbreeding with other human species that our forebears encountered: only time and a lot of work will tell. Yet fertile offspring must have emerged from such liasons. In a Linnaean sense that implies that however the partners might have looked and whatever their habits were, they had to have been the same species: one that had lasted almost 2 million years in different guises or polymorphs, as Jonathon Kingdon once suggested.
s
2007
Earth-Pages http://www.earth-pages.com/news.asp
4 of 6 5/8/2007 2:21 PM
See also: Porcelli, D. 2007. When crust is bred. Nature, v. 446, p. 863-864.
Magmas from the mantle and recycled crust
May 2007
Two processes result in the Earth’s mantle consuming rocks that formed oceanic and continental crust:subduction of oceanic lithosphere, and delamination from the base of thickened continental crust. In bothcases the rocks involved are likely to be dominantly basaltic in composition. Those from the continental crustprobably include mafic layered complexes, representing magma chambers in which intermediate magmashad resulted from fractional crystallization of basalt magma, and undifferentiated mafic igneous rocksunderplated to the crust. Highly fractionated materials from the upper continental crust may also make theirway into the mantle in the sedimentary cover to subducted ocean-floor basaltic crust.
New magmas originate in various ways as partial melts of ultramafic mantle rocks. Yet 4 billion years worth or more of consumed masses of crustal rocks must increasingly become involved in the chemistry of mantle melting by adding to its heterogeneity. Mantle heterogeneity is well-established from several lines of evidence provided by isotopic and trace-element analyses of modern basaltic lavas erupted in different tectonic settings. Yet, judging the influence and role of recycled crust have so far been plagued by data that are ambiguous. One outcome of previous research is that some ocean island basalt magmas formed by partial melting of peridotite whose chemistry had been transformed previously by other melts that had flowed through it (see Herzberg, C. 2007. Food for a volcanic diet. Science, v. 316, p. 378-379).
A large team of geochemists, combining forces (and their data) from Russia, Germany, Australia, France, Taiwan, Eritrea, Britain, the USA, the Netherlands and Iceland, have sought to reduce the ambiguities by focussing on the chemistry of olivine phenocrysts found in basaltic lavas, rather than that of whole-rock samples (Sobolev, A.V. and 19 others 2007. The amount of recycled crust in sources of mantle-derived melts. Science, v. 316, p. 412-417). Dominantly basaltic crustal masses in the mantle would melt to form silica-rich magmas. Passing through mantle peridotite, such melts would transform parts of the mantle to olivine-free pyroxenite. Magmas derived by partial melting of pyroxenite in the upper mantle would be basaltic, but enriched in silicon and nickel, and depleted in magnesium, calcium and manganese partly retained in the residual pyroxenes. Olivines crystallising first from basalt magmas carry a chemical signature of the parental melt composition, and thus the source.
The olivine approach by Sobolev and colleagues provides evidence for recycled crust in products of all kindsof basaltic magmatism, ranging from a contribution of 5% in ocean-floor basalts formed at ridges to about20% in within-plate basalts. The contribution of mantle transformed to pyroxenite by chemical interaction withmelt from foundered crustal masses ranges from 10% in mid-ocean ridge basalts to 100% in within-platebasalts formed below thick continental lithosphere. These include the largest volcanic outpourings in Earth’shistory, in the form of continental flood basalts. The largest of these, the Siberian Traps, accompanied thelargest mass extinction of the Phanerozoic at the end of the Permian Period.
378
mis of most land plants is uniseriate (compris-
ing one layer), and if SHR and SCR orthologs
control development in these species, this
mechanism may have evolved early in the his-
tory of land plants. But there are some rare
exceptions. For example, extant members of the
Equisetales (horsetails), whose forebears were
important species in Carboniferous forests 300
million years ago, have two layers of endoder-
mis in some organs (8). These horsetails have
apparently tinkered with their SCR and SHR
genes to allow SHR to escape the clutches of
SCR in the first endodermal layer, thus extend-
ing the endodermis to further layers.
Why is the structure of a single-layered
endodermis so conserved across most land
plants? Autotrophic plants rely on the uptake
of nutrients and water from the soil, so tight
spatial regulation of endodermal development
reflects its significance. The endodermis is
highly specialized in that intercellular spaces
between the cells are sealed, prohibiting the
free diffusion of water and ions through
the interstices. If water and ions are to pass
through this roadblock, they must do so in a
regulated manner, via the cytoplasm of endo-
dermal cells. This affords strict control over
transport and may well have provided a sur-
vival advantage—the highest level of control
over a specialized tissue of very limited size. If
so, this discovery may explain the evolution of
a tightly regulated developmental program in
angiosperms (flowering plants) and also points
to changes in developmental programs that
may have occurred during the past 470 million
years of land plant evolution.
References
1. H. Cui et al. Science 316, 421 (2007).
2. Y. Helariutta et al., Cell 101, 555 (2000).
3. K. Nakajima, G. Sena, T. Nawy, P. N. Benfey, Nature 413,
307 (2001).
4. L. Di Laurenzio et al., Cell 86, 423 (1996).
5. T. Wada et al., Developmemt 129, 5409 (2002).
6. W. J. Lucas et al., Science 270, 1980 (1995).
7. A. Maizel, O. Bensaude, A. Prochiantz, A. Joliot,
Development 126, 3183 (1999).
8. F. O. Bower, Primitive Land Plants (Macmillan, London,
1935).
10.1126/science.1142343
Volcanic eruptions have the power to
reshape Earth’s landscape, alter climate,
and affect life. To understand how this
works requires that we go deep into the Earth to
learn exactly what kind of rock melts to pro-
duce magmas and the chemistry of this source
rock. These are fundamental problems in geol-
ogy, and they are also among the most difficult
to understand. On page 412 in this issue,
Sobolev et al. (1) describe a method for identi-
fying some of these source rocks. We can think
of them as food for volcanoes in the sense that
they melt to provide the magmas that can erupt
to the surface. To understand what Sobolev et
al. have done and the ramifications that go
beyond Earth science, we need to start with a
refresher in geology.
Earth’s mantle consists mostly of peridotite,
a rock rich in the mineral olivine (Mg,Fe)2SiO
4.
When peridotite partially melts, the liquids
collect to magmas that rise to the crust, give off
gases like SO2, CO
2, and H
2O, and solidify to
basalt, a rock rich in the minerals clino-
pyroxene [Ca(Mg,Fe)Si2O
6] and plagioclase
[(Ca,Na)(Al1-2
Si2-3
) O8]. Portions of these outer
layers can be recycled back into the mantle at
subduction zones and below thickened conti-
nents (see the figure). The recycled basaltic
crust is transformed to a new rock called pyrox-
enite, so-called because it is rich in clinopyrox-
ene. It may pile up on Earth’s core, or be mixed
back into the mantle with structures that have
been described as marble cake (2), plum pud-
ding (3), spaghetti (4), and gumbo (5) (see the
figure). Volcanoes like those of Hawaii can melt
from source rocks consisting of peridotite
and/or pyroxenite from recycled crust. Sobolev
et al. describe a method for identifying this rock
based on the chemistry of lavas on volcanoes.
Sobolev et al. determined that many volca-
noes melted from recycled crust, a conclusion
that is not new (6). However, there has always
been some ambiguity with past methods of
identifying recycled crust based on the isotope
and trace-element geochemistry of lavas at the
surface. New interpretations suggest that
many oceanic islands melted from mantle
peridotite that had been modified by melts
that flowed through it (7, 8), a process called
metasomatism (9). Because it makes no dif-
ference to an atom of lanthanum, for example,
whether it ends up concentrated in the crust or
as metasomatized peridotite, using it as a
tracer can be ambiguous and nonunique (7, 8).
A breakthrough came when Sobolev et al.
(10) showed that the nickel contents of many
Chemical analyses of lava can now reveal the
nature of the rocks deep in the Earth that
melted and rose to generate specific volcanoes.Food for a Volcanic DietClaude Herzberg
GEOCHEMISTRY
The author is in the Department of Geological Sciences,Rutgers University, Piscataway, NJ 08854, USA. E-mail:[email protected]
Oceanic ridge
Oceanicridge
Oceanic islands
Oceanic island
Continental crust Oceanic crust
CORE MANTLE
Subduction zone
Subduction zone
Models of Earth’s crust and mantle. Oceanic crust (brown) is solidified liquid that forms by partial meltingof mantle peridotite (green) at oceanic ridges; together with sediment, oceanic crust can be recycled back intothe mantle at subduction zones (2, 3, 6). Continental crust (brown) forms at subduction zones and can berecycled when it thickens by delamination (5, 15). All crust (brown) is transformed to pyroxenite (brown)when recycled. Green arrow denotes melting peridotite. Red arrow denotes melting pyroxenite. Recycled crustmay be distributed uniformly throughout the mantle, or it may be concentrated in certain hemispheres ordepths. Crustal thickness is exaggerated for clarity, but ranges from ~6 to 40 km at the present time.Recycling is expected to reduce crust to dimensions ranging from micrometers to kilometers.
20 APRIL 2007 VOL 316 SCIENCE www.sciencemag.org
PERSPECTIVES
Published by AAAS
olivine crystals in Hawaiian lavas were higher
than those expected from melts of peridotite,
and they preferred to explain this with a recy-
cled crust source instead. But a lingering ambi-
guity is that a high nickel content in olivine can
also arise when peridotite is enriched in pyrox-
ene by melt-rock reaction (11). Supporting
evidence for the recycled-crust interpretation
(10) comes from the calcium contents of the
Hawaiian lavas, which are too low to be easily
explained by melting peridotite (12). Nickel
and calcium are therefore telling the same
story. That is, the main shield-building lavas at
Hawaii were melted from a pyroxenite source
rock that required the involvement of recycled
crust as proposed by Sobolev et al. (10). The
authors go further in that they examine the
problem that arises when nickel, calcium, and
manganese are used, and they extend the
analysis to a larger population of volcanoes.
Their results, together with other recent studies
(7, 8, 11), show that it is unlikely that a single
rock type will be an appropriate source for
all oceanic volcanoes. For example, recycled
crust is an important source rock for the
Hawaiian islands (1, 10, 12), whereas metaso-
matized peridotite is the source rock for the
Cook islands (7). An outstanding question is
whether peridotite sources become metaso-
matized by melted recycled crust (13) or in
some other way (7, 8).
Future studies might allow us to transform
our picture from hypothetical models to
actual three-dimensional views showing the
size and distribution of recycled crust in the
mantle. The implications go far beyond geol-
ogy. For example, it may be no surprise that
Sobolev et al. (1) identify pyroxenite as the
rock that melted to produce the Siberian
Traps. This was a magmatic flood on land so
massive in scale that it triggered the largest
mass extinction of life on Earth, some 250
million years ago (14). Although the exact
causal links remain poorly understood in detail,
one can reasonably imagine a different out-
come if the mantle diet had less pyroxenite
and more peridotite. Under these circum-
stances, less magma would have been pro-
duced and made available for eruptive flood-
ing, and Earth’s biosphere could have evolved
along different pathways.
References1. A. V. Sobolev et al., Science, 316, 412 (2007); published
online 29 March 2007 (10.1126/science.1138113).2. C. J. Allègre, D. L. Turcotte, Nature 323, 123 (1986).3. J. P. Morgan, W. J. Morgan, Earth Planet. Sci. Lett. 170,
215 (1999).4. W. Abouchami et al., Nature 434, 851 (2005).5. D. L. Anderson, New Theory of the Earth (Cambridge
Univ. Press, Cambridge, ed. 2, 2007).6. A. W. Hofmann, W. M. White, Earth Planet. Sci. Lett. 57,
421 (1982).7. S. Pilet et al., Earth Planet. Sci. Lett. 236, 148 (2005).8. Y. Niu, M. J. O’Hara, J. Geophys. Res. 108, 2209 (2003).9. J.-L. Bodinier et al., J. Petrol. 45, 1 (2004).
10. A. V. Sobolev et al., Nature 434, 590 (2005).11. P. B. Kelemen et al., Earth Planet. Sci. Lett. 164, 387
(1998).12. C. Herzberg, Nature 444, 605 (2006).13. R. Dasgupta et al., J. Petrol. 47, 647 (2006).14. I. H. Campbell et al., Science 258, 1760 (1992).15. C. Herzberg et al., Contrib. Mineral. Petrol. 84, 1 (1983).
Published online 29 March 2007; 10.1126/science.1141051
Include this information when citing this paper.
379
Iron-dioxygen relationships. Mononuclear nonheme iron enzymes can accommodate a diverse range ofgeometries with substrates and dioxygen, as illustrated here by the trapped catalytic intermediates in anextradiol ring-cleaving dioxygenase (left and center) reported by Kovaleva and Lipscomb and in superoxidereductase (right) reported by Katona et al.
www.sciencemag.org SCIENCE VOL 316 20 APRIL 2007
PERSPECTIVES
Published by AAAS