DOI: 10.1126/science.1131722
, 285 (2006);314Science
et al.Glen M. MacDonald,
Variations2and CO4and Atmospheric CH
Rapid Early Development of Circumarctic Peatlands
www.sciencemag.org (this information is current as of January 30, 2007 ):
The following resources related to this article are available online at
http://www.sciencemag.org/cgi/content/full/314/5797/285
version of this article at:
including high-resolution figures, can be found in the onlineUpdated information and services,
http://www.sciencemag.org/cgi/content/full/314/5797/285/DC1
can be found at:Supporting Online Material
http://www.sciencemag.org/cgi/content/full/314/5797/285#otherarticles
, 5 of which can be accessed for free:cites 14 articlesThis article
http://www.sciencemag.org/cgi/collection/atmos
Atmospheric Science
:subject collectionsThis article appears in the following
http://www.sciencemag.org/help/about/permissions.dtl
in whole or in part can be found at:this article
permission to reproduceof this article or about obtainingreprintsInformation about obtaining
registered trademark of AAAS.
c 2006 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
onJanuary30,2007www.sciencemag.orgDownloadedfrom
26. L. G. Rowan, E. L. Hahn, W. B. Mims, Phys. Rev. 137, A61
(1965).
27. Materials and methods are available on Science Online.
28. M. O. Scully, M. S. Zubairy, Quantum Optics (Cambridge
Univ. Press, Cambridge, 1997).
29. E. L. Hahn, Phys. Rev. 80, 580 (1950).
30. E. van Oort, M. Glasbeek, Chem. Phys. 143, 131
(1990).
31. C. P. Slichter, Principles of Magnetic Resonance (SpringerVerlag,
New York, 1990).
32. We are grateful to J. Doyle and Doyle lab members for
making these experiments possible. We thank A. Mukherjee,
N. Khaneja, E. Demler, C. Marcus, and S. Sachdev for many
stimulating discussions and experimental help and S. Prawer
for providing high-purity diamond samples. This work was
supported in part by the NSF Career Award; Army Research
Office Multidisciplinary University Research Initiative; and the
Packard, Sloan, and Hertz Foundations. F.J. and J.W.
acknowledge support from Deutschen Forschungsgemeinschaft
(SFB/TR21) and the European Commission (Q.A.P.).
Supporting Online Material
www.sciencemag.org/cgi/content/full/1131871/DC1
Materials and Methods
References
28 June 2006; accepted 23 August 2006
Published online 14 September 2006;
10.1126/science.1131871
Include this information when citing this paper.
Rapid Early Development of
Circumarctic Peatlands and
Atmospheric CH4 and CO2 Variations
Glen M. MacDonald,1,2
* David W. Beilman,1
Konstantine V. Kremenetski,1,3
Yongwei Sheng,1
Laurence C. Smith,1
Andrei A. Velichko3
An analysis of 1516 radiocarbon dates demonstrates that the development of the current
circumarctic peatlands began È16.5 thousand years ago (ka) and expanded explosively between
12 and 8 ka in concert with high summer insolation and increasing temperatures. Their rapid
development contributed to the sustained peak in CH4 and modest decline of CO2 during the early
Holocene and likely contributed to CH4 and CO2 fluctuations during earlier interglacial and
interstadial transitions. Given the decreased tempo of peatland initiation in the late Holocene and
the transition of many from fens (which generated high levels of CH4) to ombrotrophic bogs, a
neoglacial expansion of northern peatlands cannot explain the increase in atmospheric CH4 that
occurred after 6 ka.
M
odern northern peatlands cover about
4 million km2 across Eurasia and
North America and represent the biggest
wetland complex in the world (Fig. 1).
Today, these peatlands are thought to store 180
to 455 Pg of sequestered carbon while also
releasing 20 to 45 Tg per year of CH4 into the
atmosphere (1, 2). The potential contribution of
northern peatlands to fluctuations in atmospheric
CH4 and CO2 over the late glacial and
Holocene, and during earlier interglacials, has
been a matter of much speculation and debate
(3–8).
Ice-core records show that CH4 concentrations
rose from È350 to 650 parts per billion by
volume (ppbv) between the last glacial maximum
(LGM), which occurred 20 ka (20,000
calendar years before C.E. 1950), and the
BLlling-AllerLd (BA) warm period (È15 to
13 ka). They then declined by È200 ppbv
during the Younger Dryas (YD) stadial (È13 to
11.5 ka), rose rapidly to levels over 700 ppbv in
the early Holocene (11 to 8 ka), and then declined
again between 8 and 6 ka (3). It has been
maintained that because conditions were not
favorable for widespread circumarctic peatland
formation until after 8 ka, tropical wetlands or
marine clathrates were the likely sources for the
CH4 peak that occurred 11 to 8 ka (4, 9). On
the basis of the assumed late-Holocene development,
it has been suggested that northern
peatlands played little role in the declining atmospheric
CO2, which has also been observed
during the period from 11 to 8 ka (5). Others
argue that the development of the northern
peatland complex contributed substantially to
the early-Holocene CH4 increase and simultaneously
decreased atmospheric CO2 through
carbon sequestration in northern soils (6–8).
Resolving the debate on the potential role of
the northern peatlands in early postglacial CH4
variations has become critical since the recent
analysis of the deuterium and carbon isotopic
composition of CH4 (dDCH4
and d13
CCH4
) from
Greenland ice samples, which suggested that
the destabilization of marine clathrates is an
unlikely explanation for the BA or earlyHolocene
CH4 increases (10, 11). In view of
this evidence, it has been argued that the
sustained high levels of CH4 that developed at
the close of the YD in part require a persistent
terrestrial source linked to the warming climate
at that time (11).
Holocene concentrations of atmospheric
CH4 reached a minimum of G600 ppbv at 6 ka
and then increased again over the late Holocene
to values of about 695 ppbv just before the
industrial revolution (3). This late-Holocene increase
has been variously attributed either to
expansion of northern wetlands due to neoglacial
climatic cooling after the Holocene
thermal maximum (4) or to the product of expanding
anthropogenic activity (particularly the
expansion of rice- and cattle-based agrarian
societies) in the mid- to late Holocene (12).
However, recently collected CH4 data from
Antarctic ice cores reveal that the mid- to lateHolocene
increase is not unique. A similar lateinterglacial
increase in Pleistocene atmospheric
CH4 occurred È400 ka during Marine Isotope
Stage 11 (MIS11), which clearly cannot reflect
anthropogenic sources and has been ascribed
instead to natural factors, including expansion of
northern wetlands (13).
To address the hypothesis that northern
peatland development could have contributed
to the late-Pleistocene and Holocene variations
in atmospheric CH4 and CO2 outlined above,
we collated 1516 basal radiocarbon dates for
peat initiation from wetlands throughout highlatitude
Europe, Asia, and North America from
a wide variety of sources (14). Some areas, such
as Fennoscandia, have numerous basal dates for a
small geographic area, whereas other very large
areas such as central and eastern Siberia have a
limited number of dates (Fig. 1). Therefore, we
analyzed the compiled data set by raw number
of initiation dates, and we also divided the
Northern Hemisphere into grids of 2- latitude by
2- longitude and assigned a value for peatland
initiation based on the oldest basal radiocarbon
date in each cell (Fig. 1).
The lack of basal dates older than about
16.5 ka suggests that there was no extensive
peatland complex in the northern circumpolar
region during the LGM (Fig. 2). This finding is
corroborated by palynological data that indicate
a paucity of Sphagnum (peat moss) spores from
deposits of this age (15). Before 16.5 ka, much
of the North American and European arctic and
subarctic were still covered in ice, and it is
likely that the large ice-free areas of Siberia and
Beringia were too cold and dry (16) to promote
extensive peatland development. This absence
of any significant northern peatland complex
during the LGM is consistent with the depressed
CH4 levels and the relatively low proportion
of northern CH4 sources observed in
ice-core records (Fig. 3).
In concert with increasing summer insolation
and northern high-latitude temperatures, the
current northern peatland complex began developing
in ice-free portions of North America and
Asia between 16.5 and 14 ka and initiating
widely on all three northern continents after
14 ka (Figs. 1 and 2). These results dispel the
earlier assertion that peatland development in
1
Department of Geography, 2
Department of Ecology and
Evolutionary Biology, University of California, Los Angeles,
CA 90095–1524, USA. 3
Russian Academy of Sciences,
Moscow 109017, Russia.
*To whom correspondence should be addressed. E-mail:
macdonal@geog.ucla.edu
REPORTS
www.sciencemag.org SCIENCE VOL 314 13 OCTOBER 2006 285
onJanuary30,2007www.sciencemag.orgDownloadedfrom
western Siberia was in advance of development
in North America or elsewhere (7). The initial
expansion coincides with warming during the
BA period (Fig. 3), and it is likely that increasing
warmth and moisture and decreasing
glacial ice cover promoted peatland growth.
Resultant development of the northern peatland
complex corresponds with increasing atmospheric
CH4 concentrations observed during
the BA (Fig. 3). However, the still-limited extent
of northern peatlands at this early interval
is consistent with evidence for a relatively restricted
role of northern wetlands in producing
the BA CH4 increase (Fig. 3). There is a small
decline in the rate of new peatland formation
during the YD, which may be attributed to the
development of cold conditions over much of
the Northern Hemisphere and readvances of
ice. At the same time, there is a precipitous
drop in CH4 recorded in the ice cores (Fig. 3)
that seems to reflect additional factors. During
the BA and YD periods, it is unlikely that the
extent and growth of the northern peatlands was
great enough to have significantly affected atmospheric
CO2 concentrations.
The initiation of the early Holocene warming
at 11.5 ka after the YD is marked by rapid expansion
of peatlands throughout the north. The
North American peatlands expanded rapidly
during this time (Figs. 1 and 2), even though still
constrained by the ice extent of the Laurentide
Ice Sheet (17). The tempo of subsequent
northern peatland development in North America
was influenced by the rate of continued ice
retreat and the exposure of land surfaces (Fig. 2).
Large areas of northern Asia never supported ice
during the LGM (18), and by 11 ka, high-latitude
climate had warmed (Fig. 3). What remained of
the Scandinavian Ice Sheet was restricted to cenFig.
1. Map of the current distribution of the northern circumarctic peatland
complex and the boreal forest biome developed from a number of sources
(SOM text) and the initiation dates of the peatlands based on radiocarbon
dates from the base of peat deposits (14). The red dots indicate new peatlands
that initiated during each time slice and the gray dots indicate preexisting
peatlands initiated during earlier time slices. The timing of peatland
initiation based on 2- by 2- grids is shown in the right panel in the bottom
row. The grids map the oldest basal peat dates within each grid cell.
Fig. 2. Timing of circumarctic
peatland establishment
in North America
(black), Europe (red), and
Asia (blue) based on the
total number of initiation
dates from each region.
The occurrence frequency
of basal peat radiocarbon
ages is plotted as the
number of calibrated age
ranges that fall in any year (14), smoothed with a 100-year running mean. The decreasing area of the
Laurentide Ice Sheet (dashed line) as it retreated over the late glacial and Holocene (17) is also plotted.
REPORTS
13 OCTOBER 2006 VOL 314 SCIENCE www.sciencemag.org286
onJanuary30,2007www.sciencemag.orgDownloadedfrom
tral Fennoscandia (19), allowing widespread expansion
of peatlands in northern Eurasia (Figs. 1
and 2). A sustained period of maximum rates of
peatland establishment followed and persisted
until about 8 ka.
The rapid expansion of peatlands at the close
of the YD coincides with a È15-Tg increase in
atmospheric CH4 derived from the Northern
Hemisphere (Fig. 3) and a 200- to 250-ppbv
increase in total CH4 concentrations. According
to our data, the northern peatland complex was
likely at G20% of its current aerial extent at the
end of the YD and expanded to about 50% by
8 ka. On the basis of current estimates of overall
CH4 production from northern peatlands (1, 2),
they may have contributed 4 to 9 Tg of Northern
Hemisphere CH4 after the end of the YD
and up to 12 to 27 Tg by 8 ka. However, based
on typical peatland succession stages, higher
summer insolation in the early Holocene, and
evidence of peatland vegetation and type
from our peat cores taken in Siberia (7), it is
likely that many of these newly developed
peatlands were warm and wet minerotrophic
fens, often dominated by sedges. Such fens typically
emit CH4 at rates many times greater than
the ombrotrophic Sphagnum bogs common in
much of the north today (6, 20–22). Therefore,
we suspect that the rates of CH4 production in
northern peatlands may have been considerably
higher in the early Holocene than they
are today.
The d13
CCH4
values of ice from the Pakitsoq
outcrop in western Greenland have been used
to infer the likely origins of early-Holocene
atmospheric CH4 (11). It has been suggested
that major terrestrial sources contributing methane
during the period from 11 to 8 ka likely
produced CH4 with d13C values of between –50
and –60 (11). Although northern peatlands
dominated by ombrogenous bogs may typically
emit CH4 with d13C G –60, pore waters
and emissions of boreal fens are relatively
enriched in 13C (average d13
CCH4
values of –50
to –60), particularly in high-productivity sites
(23). These values are consistent with the
early northern peatland complex, dominated
by minerotrophic fens, which was a major
contributor to the peak in CH4 that occurred
11 to 8 ka and must be considered in addition
to the potential tropical sources suggested previously
(11).
The rapid growth of the circumarctic peatland
complex and its associated sequestration of
phytomass carbon in the early Holocene may
have also contributed to the decline by È7 parts
per million by volume (ppmv) of atmospheric
CO2 observed in the ice-core records between
11 ka and the mid-Holocene (Fig. 3). This
decline has been interpreted as reflecting a total
biosphere uptake of 110 Pg C in the first half of
the Holocene (8). More than half of the peatland
basal dates in our data set are older than
8 ka, indicating rapid initiation and development
of this carbon sink in the early Holocene.
Conservatively, if by 8 ka peat deposits were
0.5 to 1 m thick, covered just one-quarter of
today_s peatland area, and were similar in carbon
characteristics to today_s northern peatlands
(1), they would have been capable of
sequestering 29 to 58 Pg C. Further detailed
reconstruction of net peat accumulation rates
during this period are required to better estimate
the magnitude of the contribution of the northern
peatlands to early-Holocene atmospheric
CO2 declines.
The widespread development of peatlands in
response to increasing summer insolation, BA
warming, and particularly rapid Holocene
warming after the YD supports the hypothesis
that the northern peatlands were a major terrestrial
factor contributing to the early fluctuations in
atmospheric CH4 and to CO2 sequestration
during the current interglacial. It is likely that
they played a similar role in earlier interglacials.
The rapidity and large spatial extent of the
response of the northern peatland complex,
particularly to the onset of post-YD warming,
suggests that they may even have played a role
in CH4 and CO2 variations at the shorter time
scale of earlier interstadials such as MIS3.
The observed decline in peatland initiation,
particularly in Europe and Asia after 8 ka,
corresponds to a decline in atmospheric CH4
concentrations between 8 and 6 ka (Fig. 3).
This decline in initiation rates should not be
confused with a decline in total peatland area;
most peatlands that were extant at 8 ka were
still extant at 6 ka. However, many had transFig.
3. Timing of circumarctic peatland establishment compared with June insolation at 60-N (26),
Greenland Ice Sheet Project 2 (GISP2) temperature reconstruction (27), atmospheric CO2 and CH4
concentrations, and estimates of Northern Hemisphere CH4 emissions derived from the InterPolar
CH4 Gradient (IPG) (3, 4, 6, 28). Atmospheric CO2 and CH4 concentrations show ice-core data from
European Project for Ice Coring in Antarctica (EPICA) Dome C (red for CO2 and blue for CH4) and
GISP2 [green triangles show CH4 (3, 29–31)]. Dome C data are shown on the original EPICA Dome C 1
time scale from 0 to 10 ka before present (circles) and on the GISP2 time scale with CH4
synchronization from 10 to 18 ka before present (diamonds; SOM text). Dome C error bars indicate 1s
uncertainty. (A) The occurrence frequency of 1516 radiocarbon dates of basal peat deposits (14)
shows the number of calibrated age ranges that occur in any year (black curve). (B) Cumulative curve
of 1516 dates (red curve). (C) The oldest basal peat dates within each 2- by 2- grid (gray bars).
REPORTS
www.sciencemag.org SCIENCE VOL 314 13 OCTOBER 2006 287
onJanuary30,2007www.sciencemag.orgDownloadedfrom
formed from early minerotrophic fens to ombrotrophic
Sphagnum bogs, which are typically
weaker sources of CH4 than are fens (20–23).
Therefore, the transformation from high CH4efflux
fens to Sphagnum bogs, coupled with a
declining rate of new peatland formation, would
have contributed to the decline in atmospheric
CH4. What remains puzzling is the role that
northern peatlands played in the subsequent
increase in CH4 between 6 ka and just before
the Industrial Revolution. By C.E. 1700, levels
of atmospheric CH4 had increased once again to
almost 700 ppbv. However, contrary to earlier
speculation (9, 15), new peatland initiation was
relatively modest in the late Holocene, and conversion
of fens producing high levels of CH4 to
Sphagnum bogs with lower production was
ongoing. If the mid- to late-Holocene CH4
increase does not have an anthropogenic
explanation, then its source must lie in factors
other than large-scale resurgent expansion of
the northern peatland complex.
References and Notes
1. E. Gorham, Ecol. Appl. 1, 182 (1991).
2. S. E. Mikaloff Fletcher, P. P. Tans, L. M. Bruhwiler, J. B. Miller,
M. Heimann, Global Biogeochem. Cycles 18, GB4004,
10.1029/2004GB002223 (2004).
3. E. J. Brook, S. Harder, J. Severinghaus, E. J. Steig,
C. M. Sucher, Global Biogeochem. Cycles 14, 559 (2000).
4. T. Blunier, J. Chappellaz, J. Schwander, B. Stauffer, D. Raynaud,
Nature 374, 46 (1995).
5. F. Joos, S. Gerber, I. C. Prentice, B. L. Otto-Bliesner,
P. J. Valdes, Global Biogeochem. Cycles 18,
10.1029/3003GB002156 (2004).
6. A. Da¨llenbach et al., Geophys. Res. Lett. 27, 1005
(2000).
7. L. C. Smith et al., Science 303, 353 (2004).
8. A. Indermu¨hle et al., Nature 398, 121 (1999).
9. J. P. Kennett, K. G. Cannariato, I. L. Hendy, R. J. Behl,
Methane Hydrates in Quaternary Climate Change (American
Geophysical Union, Washington, DC, 2003).
10. T. Sowers, Science 311, 838 (2006).
11. H. Schaefer et al., Science 313, 1109 (2006).
12. W. F. Ruddiman, Clim. Change 61, 261 (2003).
13. R. Spahni et al., Science 310, 1317 (2005).
14. Radiocarbon dates of basal peat from 1516 selected
records from the boreal biome and closely adjacent areas
were collated from published articles and reports (table
S1). Only basal dates with complete laboratory and
geospatial information were included. Radiocarbon ages
were calibrated using CALIB 5.0 (24) and the IntCal04
data set (25). In our calculations, the age of peatland
initiation is the range of the calibrated age probability
distribution at 1s precision rounded to decades. The
median span of the calibrated age ranges (1s) is 300
years. Further details on maps and radiocarbon ages are
provided in supporting online material (SOM) text.
15. K. Gajewski, A. Viau, M. Sawada, D. Atkinson, S. Wilson,
Global Biogeochem. Cycles 15, 10.1029/2000GB001286
(2001).
16. S.-I. Shin et al., Clim. Dyn. 20, 127 (2003).
17. A. S. Dyke, A. Moore, L. Robertson, Deglaciation of North
America (Geological Survey of Canada Open File 1574,
Ottawa, Canada, 2003).
18. L. Gualtieri, S. L. Vartanyan, J. Brigham-Grette,
P. M. Anderson, Boreas 34, 264 (2005).
19. J. Donner, The Quaternary History of Scandinavia
(Cambridge Univ. Press, Cambridge, UK, 1995).
20. H. Almquist-Jacobson, D. R. Foster, Ecology 76, 2503
(1995).
21. K. B. Bartlett, R. C. Harriss, Chemosphere 26, 261
(1993).
22. T. Christensen et al., Geophys. Res. Lett. 30,
10.1029/2002GL016848 (2003).
23. L. M. Bellisario, J. L. Bubier, T. L. Moore, J. P. Chanton,
Global Biogeochem. Cycles 13, 81 (1999).
24. M. Stuiver, P. J. Reimer, Radiocarbon 35, 215 (1993).
25. P. J. Reimer et al., Radiocarbon 46, 1029 (2004).
26. J. Chappellaz et al., J. Geophys. Res. 102, 15987
(1997).
27. A. Berger, M. F. Loutre, Quat. Sci. Rev. 10, 297 (1991).
28. R. B. Alley, Quat. Sci. Rev. 19, 213 (2000).
29. J. Flu¨ckiger et al., Global Biogeochem. Cycles 16, 1010,
10.1029/2001GB001417 (2002).
30. E. Monnin et al., Science 291, 112 (2001).
31. E. Monnin et al., Earth Planet. Sci. Lett. 224, 45 (2004).
32. Research funding was provided by NSF through the Arctic
System Science Program (ARCSS) through the RussianAmerican
Initiative on Shelf-Land Environments of the
Arctic (grant OPP-9818496) and a Doctoral Dissertation
Research Improvement Grant from the Geography
Program (grant SRB-0425625). We thank our many
colleagues from the United States, Russia, Canada, and
Finland for helpful discussions on our results and
assistance in compiling radiocarbon dates. We thank
J. Flu¨ckiger for providing helpful comments on an earlier
draft and for important comparative data.
Supporting Online Material
www.sciencemag.org/cgi/content/full/314/5797/285/DC1
SOM Text
Table S1
References
26 June 2006; accepted 5 September 2006
10.1126/science.1131722
Gold in Magmatic Hydrothermal
Solutions and the Rapid Formation
of a Giant Ore Deposit
Stuart F. Simmons1
* and Kevin L. Brown2
The Ladolam hydrothermal system, on Lihir Island, Papua New Guinea, hosts one of the youngest
and largest gold deposits in the world. Several deep (more than 1 kilometer) geothermal wells
were drilled beneath the ore bodies to extract water at 9275-C and to facilitate open-pit mining. Using
a titanium down-hole sampler, we determined that the deep geothermal brine of magmatic origin
contains È15 parts per billion gold. At the current gold flux of 24 kilograms per year, this deposit could
have formed within È55,000 years. The combination of sustained metal flux and efficient metal
precipitation led to the formation of a giant hydrothermal gold deposit in a short period.
T
he origins of giant hydrothermal gold
deposits are enigmatic (1). This is because
the concentrations of precious
metals and flow rates of ore-forming fluids are
poorly quantified, and the origins of the metals
are unclear. These aspects can be clarified
with direct analyses of fluids from modern
hydrothermal systems. The only known active
hydrothermal gold deposit is at Ladolam, on
Lihir Island, Papua New Guinea, and this
deposit is one of the largest in the world with
È1300 tons of gold. Geothermal drilling at
Ladolam has been ongoing since before the
start of open-pit mining in 1997, and the wells
provide access to the deep fluids upstream of
the ore zone. Because precious metals precipitate
in the wells during fluid ascent due to
boiling (2), we used a titanium down-hole
sampler to obtain deep fluids for analyses
(3). Any metals that precipitate in the
sampler during its return to the surface can
be incorporated into the water sample by
rinsing the titanium sampler with strong acid.
Thus, the deep metal concentrations can be
directly measured and used to constrain the
rate at which the gold is transported and then
deposited in order to form this giant deposit.
Because the isotopic compositions of the deep
hydrothermal waters at Ladolam are predominantly
magmatic in origin (4), our data cover
not just the origin of the metals, but also the
concentrations of gold and related metals in
magmatic hydrothermal solutions, which are
thought to be important to the formation of ore
deposits (5, 6).
The Ladolam gold deposit occupies the
center of the extinct Luise volcano on Lihir
Island (Fig. 1). The two major tabular ore
zones cover È2 km2 and extend from the
surface to 400 m below sea level. They lie in
the middle of a breached crater that formed in
response to sector collapse and unroofing
of the volcanic edifice È400,000 years ago
(4, 7, 8). The resulting explosive depressurization
of the magmatic-hydrothermal system
produced a diatreme breccia complex and
highly permeable rocks, which now host the
ore. Mineralogical, fluid inclusion, and isotope
studies (4) show that the gold was
deposited in two stages between 150- and
250-C, from solutions of magmatic origin, when
they mixed with other fluids or boiled. The
magmatic gold-bearing solutions were near
neutral to slightly alkaline pH and contained 5
to 10 wt % equivalent NaCl. The host rocks
comprise alkalic mafic to intermediate volcanic
and intrusive rocks (9) that have been hydro-
1
Geology Dept., Univ. of Auckland, Private Bag 92019,
Auckland Mail Centre, Auckland 1142, New Zealand. 2
GEOKEM,
P. O. Box 95-210, Swanson, Waitakere 0653, New Zealand.
*To whom correspondence should be addressed. E-mail:
sf.simmons@auckland.ac.nz
REPORTS
13 OCTOBER 2006 VOL 314 SCIENCE www.sciencemag.org288
onJanuary30,2007www.sciencemag.orgDownloadedfrom