Archean mantle plumes: Evidence from greenstone belt geochemistry

Abstract

Two main Archean greenstone associations and one minor one are recognized as possible mantle plume-generated sequences. The mafic-plain association, which may represent oceanic plateau remnants, comprises predominantly pillow basalts, variable amounts of komatiite, and small amounts of chert and banded iron formation (BIF). The platform association, which may represent plume-generated magmas erupted through continental crust, overlies tonalitic basement and comprises tonalite conglomerates, quartz arenites, carbonates and BIF, overlain by komatiites and basalts. A third but minor association, comprising andesitic to felsic arc rocks overlain by komatiitebasalt sequences, may represent arcs that rifted in response to mantle plumes. Basalts and komatiites from these greenstone associations have Th/Ta and La/Yb ratios similar to those of oceanic plateau basalts with plume affinities. If we employ these lithologic and geochemical screens, we find that ∼35% of Late Archean (3.0-2.5 Ga) and 80% of Early Archean (>3.0 Ga) greenstones may have plume affinities. Plume-related greenstones contaminated by continental crust acquire a "pseudosubduction" zone geochemical signature (Ta and Nb depletion and Th enrichment). For this reason, estimated percentages should be considered to be minimums. High Th/Ta ratios may reflect crustal contamination or recycling of continental crust within mantle plumes. Low Th/Ta and high La/Yb ratios may reflect melting at different depths within mantle plumes or recycling of oceanic crust within plumes. The relative abundance of plume-related greenstones in the Archean may reflect the warm, less easily subducted Archean oceanic lithosphere. However, the variety of Archean plume-generated assemblage types, combined with their abundance, may suggest instead that mantle plumes were more widespread in the Archean than subsequently.

Full text

doi:10.1130/0-8137-2352-3.341
Geological Society of America Special Papers 2001;352; 341-357
Kirsty Y. Tomlinson and Kent C. Condie
Archean mantle plumes: Evidence from greenstone belt geochemistry
Geological Society of America Special Papers
E-mail alerting services
this article
to receive free e-mail alerts when new articles citewww.gsapubs.org/cgi/alerts
click
Subscribe
Special Papers
to subscribe to Geological Society of Americawww.gsapubs.org/subscriptionsclick
Permission request
to contact GSA.www.geosociety.org/pubs/copyrt.htm#gsaclick
viewpoint. Opinions presented in this publication do not reflect official positions of the Society.
positions by scientists worldwide, regardless of their race, citizenship, gender, religion, or political
article's full citation. GSA provides this and other forums for the presentation of diverse opinions and
articles on their own or their organization's Web site providing the posting includes a reference to the
science. This file may not be posted to any Web site, but authors may post the abstracts only of their
unlimited copies of items in GSA's journals for noncommercial use in classrooms to further education and
to use a single figure, a single table, and/or a brief paragraph of text in subsequent works and to make
GSA,employment. Individual scientists are hereby granted permission, without fees or further requests to
Copyright not claimed on content prepared wholly by U.S. government employees within scope of their
Notes
Geological Society of America
on November 16, 2013specialpapers.gsapubs.orgDownloaded from on November 16, 2013specialpapers.gsapubs.orgDownloaded from

Geological Society of America
Special Paper 352
2001
341
Archean mantle plumes: Evidence from
greenstone belt geochemistry
Kirsty Y. Tomlinson*
Department of Earth Sciences, Laurentian University, Ramsey Lake Road, Sudbury, Ontario P3E 2C6, Canada
Kent C. Condie
Department of Earth and Environmental Science, New Mexico Institute of Mining and Technology,
Socorro, New Mexico 87801, USA
ABSTRACT
Two main Archean greenstone associations and one minor one are recognized as
possible mantle plume–generated sequences. The mafic-plain association, which may
represent oceanic plateau remnants, comprises predominantly pillow basalts, variable
amounts of komatiite, and small amounts of chert and banded iron formation (BIF).
The platform association, which may represent plume-generated magmas erupted
through continental crust, overlies tonalitic basement and comprises tonalite conglom-
erates, quartz arenites, carbonates and BIF, overlain by komatiites and basalts. A third
but minor association, comprising andesitic to felsic arc rocks overlain by komatiite-
basalt sequences, may represent arcs that rifted in response to mantle plumes.
Basalts and komatiites from these greenstone associations have Th/Ta and La/Yb
ratios similar to those of oceanic plateau basalts with plume affinities. If we employ
these lithologic and geochemical screens, we find that ⬃35% of Late Archean (3.0-
2.5 Ga) and 80% of Early Archean (⬎3.0 Ga) greenstones may have plume affinities.
Plume-related greenstones contaminated by continental crust acquire a “pseudo-
subduction” zone geochemical signature (Ta and Nb depletion and Th enrichment).
For this reason, estimated percentages should be considered to be minimums. High
Th/Ta ratios may reflect crustal contamination or recycling of continental crust within
mantle plumes. Low Th/Ta and high La/Yb ratios may reflect melting at different
depths within mantle plumes or recycling of oceanic crust within plumes.
The relative abundance of plume-related greenstones in the Archean may reflect
the warm, less easily subducted Archean oceanic lithosphere. However, the variety of
Archean plume-generated assemblage types, combined with their abundance, may
suggest instead that mantle plumes were more widespread in the Archean than sub-
sequently.
Tomlinson, K.Y., and Condie, K.C., 2001, Archean mantle plumes: Evidence from greenstone belt geochemistry, in Ernst, R.E., and Buchan, K.L., eds., Mantle
Plumes: Their Identification Through Time: Boulder, Colorado, Geological Society of America Special Paper 352, p. 341–357.
*E-mail: ktomlins@nrcan.gc.ca
*Present address: Continental Geoscience Division, Geological Survey of
Canada, 601 Booth Street, Ottawa, Ontario K1A 0E8, Canada
INTRODUCTION
There is considerable interest in the role that mantle plumes
may have played in Archean magma production, crustal un-
derplating, and heat liberation from the Earth. Mantle tempera-
tures must have been higher in the Archean, and hence the
sinking of oceanic lithosphere into the mantle can account for
only a small fraction of Archean heat loss (Bickle, 1986; Da-
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

K.Y. Tomlinson and K.C. Condie342
Figure 1. Primitive-mantle-normalized incompatible element distri-
butions in basalts. N-MORB—normal mid-ocean ridge basalt.
References: Sun and McDonough (1989), McCulloch and Gamble
(1991); Mahoney et al. (1993).
vies, 1992). Although plumes today account for no more than
about 10% of the Earth’s heat loss, the question of whether
mantle plumes were more important in cooling the mantle dur-
ing the Archean is an important unknown in terrestrial thermal
history (Davies, 1993; Abouchami et al., 1990; Storey et al.,
1991).
One way to track the relative importance of mantle plumes
in the Archean is from the geochemical signatures in greenstone
belts and lower-crustal xenoliths (Condie, 1994, 1997, 1999;
Campbell et al., 1989; Campbell, 1998). From the geophysical
and geochemical characteristics of lower-crustal xenoliths,
Condie (1997, 1999) suggested that plume-derived components
may be more abundant in the lower continental crust than
generally recognized (Rudnick, 1995). Lower-crustal xenoliths
from kimberlites in the Slave and Superior provinces record
zircon-growth ages that correspond to four major plume-
generated dike-swarm or magmatic events (1.27 Ga Mackenzie
swarm, 2.22 Ga Malley-McKay swarm, 2.45 Ga Matachewan
swarm, and 1.1 Ga Keweenawan magmatism; Davis, 1997;
Moser and Heaman, 1994; Neymark et al., 1994). Considered
as a whole, the mafic xenolith data suggest that a minimum of
one-third of the post-Archean continental crust comprises mafic
rocks from plume sources, either as accreted oceanic plateaus
or as mafic rocks magmatically underplated (Condie, 1999).
Because both crustal and host-magma contamination may raise
such ratios as Th/Ta and La/Nb in xenoliths, the value of one-
third is a minimum value for the post-Archean lower crust. Only
one data set is available for Archean lower-crustal xenoliths
(Huang et al., 1995), and it also suggests a plume source.
In the past few years, new geochemical data for Archean
greenstone belts have become available, mainly from the west-
ern Superior province in Canada. Thurston and Chivers (1990)
showed that Archean greenstone belts comprise several differ-
ent lithologic assemblages. The so-called “mafic-plain” succes-
sions are candidates for oceanic plateau basalts, whereas the
“platform” successions may have been erupted through, or de-
posited on, continental crust. Both of these may reflect mantle
plume sources. The Archean mafic-plain successions contain a
large proportion of pillow basalt with variable amounts of ko-
matiite and small amounts of chemical sedimentary rocks such
as chert and banded iron formation (BIF). Felsic to intermediate
volcanic rocks are uncommon in mafic-plain greenstones.
These sequences are often tectonically bounded and interpreted
to be allochthonous (e.g., Pickle Crow assemblage, western Su-
perior province). The platform successions on the other hand,
overlie a tonalitic or felsic volcanic basement. In some cases, a
basal unconformity is preserved, with tonalite conglomerates
and arkoses or arenites overlying tonalite (e.g., Steep Rock belt,
North Caribou Lake belt, western Superior province). These
rocks are overlain by carbonates and BIF, which are in turn
overlain by komatiites and basalts with minor proportions of
felsic volcanic rocks. The volcanic sequences have been shown
in some cases to contain zircons inherited from the basement
and are variably contaminated by continental crust (Tomlinson
et al., 1999b).
In this study, we summarize the lithologies and incompat-
ible element geochemistry of mafic-plain–type and platform-
type Archean greenstone belts from the Superior province
and elsewhere, and we evaluate the results in terms of plume
sources for these successions. We also include data from an-
other less common assemblage type, that of rifted arcs. We then
go on to discuss the proportion of Archean plume-generated
sequences and the relative importance of Archean mantle
plumes.
La/Yb AND Th/Ta RELATIONSHIPS
When normalized to primitive mantle and plotted on a
multi-element diagram (Fig. 1), normal mid-ocean ridge basalts
(N-MORBs) show depletion in incompatible elements, whereas
oceanic-island basalts (OIBs) show enrichment in these ele-
ments. Oceanic-plateau basalts (OPBs) show generally flat rare
earth element (REE) profiles with slight depletion in the most
incompatible elements. Arc-related basalts show anomalies, the
most prominent of which is a depletion in Ta and Nb compared
to neighboring incompatible elements and REEs (Fig. 1). The
most striking distinction between arc-related and oceanic non–
arc-related basalts is this negative Ta-Nb anomaly (McCulloch
and Gamble, 1991), which may be used to determine the tec-
tonic setting of oceanic basalts in Archean greenstone belts
(Condie, 1994). Other geochemical features, such as the Ni con-
tent at a particular Mg number (Condie, 1997), can also be
useful in distinguishing arc-related from non-arc-related green-
stone belts. In the Archean however, Ni contents were higher
in all mafic magmas because of greater degrees of melting in
the mantle (Arndt, 1991); therefore, Ni indices are less useful
in distinguishing tectonic setting.
Ta-Nb anomalies can be identified on a plot of La/Yb ver-
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

Archean mantle plumes: Evidence from greenstone belt geochemistry 343
Figure 2. La/Yb vs. Th/Ta plot showing fields of various types of basalt
and various mantle sources. Also shown for comparison are Archean
continental flood basalts from the Fortescue Group, South Africa (open
squares). OPB—oceanic plateau basalt, OIB—oceanic island basalt,
SZB—subduction zone basalt, N-MORB, normal mid-oceanic ridge
basalt, AUC—average Archean upper crust, PM—primitive mantle,
HIMU-type basalt; HIMU means high l, where the l-value is the ratio
of
238
U/
204
Pb. EM1-type basalt and EM2-type basalt—basalts from
enriched mantle sources, DM—depleted mantle, FOZO—focal zone
(not a point but an area on the graph around the word FOZO). Ref-
erences: Nelson et al. (1992); Weaver (1991); Hart et al. (1992); Tarney
et al. (1979); McCulloch and Gamble (1991); Floyd et al. (1992);
Seifert et al. (1996); and Condie (1994, 1997, and 1998 and references
cited therein).
sus Th/Ta, where arc-related basalts show Th/Ta ratios mostly
⬎2 and La/Yb ratios ⬎3 (Fig. 2). This geochemical signature
reflects an enrichment in the mantle wedge of Th and light
REEs (LREEs), but not in Ta and Nb, which, unlike most large
ion lithophile elements, appear to be insoluble in aqueous fluids
that metasomatize the mantle wedge (Hawkesworth et al.,
1993). Unlike subduction-related basalts, N-MORBs and OPBs
have relatively low Th/Ta ratios and variable La/Yb ratios (Fig.
2). Although basalts from these two tectonic settings overlap in
geochemical characteristics, N-MORBs typically have Th/Ta
ratios ⱕ1, whereas this ratio in OPBs is generally ⬎1. All of
the mantle geochemical components recognized in oceanic-
plume–related basalts (Zindler and Hart, 1986; Weaver, 1991)
fall within the OPB or OIB fields on the La/Yb versus Th/Ta
plot. Basalts representing the most common plume end mem-
ber, FOZO (Hart et al., 1992), fall near the center of the OPB
field (Fig. 2), whereas basalts representing the enriched-mantle
components (EM1, EM2, HIMU) have high La/Yb values and
define one end of the OIB field (Fig. 2).
Not all greenstone basalts are free from crustal contami-
nation. Contamination of N-MORB or plume-derived magmas
with upper continental crust raises Th/Ta and La/Yb ratios;
samples generally plot on mixing arrays between average Ar-
chean upper continental crust (AUC; Fig. 2) and a mafic end
member in the OPB or N-MORB fields. It may be possible to
identify crustal-contamination trends in Archean greenstone
belts (e.g., Condie, 1994); in these instances, it is possible to
recognize OPBs that have acquired a “pseudo-subduction” geo-
chemical signature by upper-crustal contamination. Many Ar-
chean greenstone belts have been described as arc or backarc
type on the basis of limited geochemical data. Such greenstone
belts may represent crustally contaminated plume-derived mag-
mas. To identify such cases, it is important to examine geo-
chemical trends and lithostratigraphy.
At the end of the spectrum, Archean continental flood-
basalt sequences such as the Fortescue Group of the Pilbara
craton (Western Australia) and the Ventersdorp Supergroup of
the Kaapvaal craton (South Africa) are suggested to have been
derived by the interaction of asthenospheric mantle plumes with
subcontinental lithospheric mantle and continental crust (e.g.,
Nelson et al., 1992; White, 1997; Arndt et al., this volume).
Such sequences, although plume-generated, preserve a crustal
signature in terms of Th/Ta and La/Yb ratios and plot purely in
the subduction-zone-basalt field of Figure 2 (e.g., data of Nel-
son et al., 1992). These types of sequences (subaerial, conti-
nental flood basalts) are uncommon in the Archean; they do not
preserve a plume geochemical signature and will hence not be
considered further in this paper.
PLUME-GENERATED GREENSTONE BELTS
In this section we present a compilation of data from Ar-
chean greenstone belts that are most likely to have been derived
from mantle plume sources. We begin with the Superior prov-
ince of Canada where the geochemical database is most exten-
sive; we first concentrate on the well-studied Abitibi subprov-
ince and then focus on greenstone belts in the northern Superior
province that formed in both oceanic and continental environ-
ments (Fig. 3). Last, we present examples of greenstone belts
from Archean cratons outside of the Superior province. Where
Ta data were unavailable, proxy values have been calculated
from Nb by using the primitive mantle ratio (Ta ⳱ Nb/17.4;
Sun and McDonough, 1989).
Abitibi subprovince, Canada
The Abitibi subprovince is well known for its komatiites
(e.g., Pyke Hill in Munro township) and for its abundance of
komatiite-tholeiite sequences. The subprovince has been sug-
gested by various authors to contain accreted oceanic plateau
fragments, and the importance of plume magmatism in this re-
gion has been recognized for several years (e.g., Desrochers et
al., 1993; Kimura et al., 1993; Xie et al., 1993). Four examples
of plume-generated sequences have been described by Xie et
al. (1993), Xie and Kerrich (1994), and Xie (1996) from Bos-
ton, Munro, and Tisdale townships and from the Kinojevis
Group. These sequences are thought to represent tectonically
juxtaposed remnants of oceanic plateaus derived from mantle
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

K.Y. Tomlinson and K.C. Condie344
Figure 3. Simplified geologic map of
Ontario showing the distribution of Ar-
chean greenstone belts that probably had
mantle plume sources.
plume magmatism. A fifth sequence, the Stoughton-Roque-
maure Group (which may be partially equivalent to Munro
township komatiites) is described by Dostal and Mueller (1997)
as autochthonous, produced by plume-generated rifting of an
underlying rhyolitic arc sequence (Hunter Mine Group).
Whether these sequences are allochthonous or autochthonous
and whether they represent oceanic magmatism or magmas re-
lated to rifting in a supra-subduction zone setting is the topic
of debate, but most workers seem to agree that the sequences
are plume generated. The stratigraphic features of these se-
quences, where available, are summarized in Table 1, and rep-
resentative komatiitic and tholeiitic lavas are plotted on a La/
Yb versus Th/Ta diagram in Figure 4.
La/Yb and Th/Ta ratios for basalts and komatiites from
Abitibi sequences show a greater range of values compared to
those of modern plume-generated basalts (Fig. 4). Ratios of
REEs to high field strength elements (HFSEs) are variable in
the Abitibi rocks and have been interpreted by Xie and co-
workers (Xie et al., 1993; Xie and Kerrich, 1994; Xie, 1996)
to represent (1) different depths of melting within one or more
mantle plumes, where the role of majorite garnet or perovskite
generates HFSE anomalies and (2) recycling of crustal material
within mantle plumes. For example, the Boston township lavas
with high La/Yb ratios are depleted in heavy REEs (HREEs;
Table 1), have negative Zr and Hf anomalies, and have been
shown by Xie et al. (1993) to represent melting in the majorite
garnet stability field. Th enrichment in komatiites and basalts
from Munro and Tisdale townships and in basalts from the
Kinojevis Group may represent recycling of continental crust
within a mantle plume (R. Kerrich, 1999, personal communi-
cation). Such features could also be explained by subduction-
zone interaction and/or crustal contamination, as has been sug-
gested for some Abitibi komatiites (Barrie, 1997; Dostal and
Mueller, 1997; Kerrich et al., 1998). The unusually low La/Yb
ratios in some basalts from Tisdale and Munro townships reflect
LREE depletion (Xie, 1996; Table 1) and indicate a source that
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

Archean mantle plumes: Evidence from greenstone belt geochemistry 345
TABLE 1. PLUME-GENERATED GREENSTONE BELTS OF THE ABITIBI SUBPROVINCE, SOUTHERN SUPERIOR PROVINCE, CANADA
Location and age Stratigraphy and relationships Ratios References
Boston Creek flow
Boston township, SVZ, Abitibi
subprovince, Superior province,
Ontario; 2720 Ma?
100-m-thick flow differentiated from peridotite to
diorite with a 30-m-thick spinifex-textured zone.
Underlain and overlain by Fe-tholeiite flows.
Interpreted to represent plume magmatism with
melting in majorite garnet stability field.
Th/Ta ⳱ 1.2–2.0 Data:
La/Yb ⳱ 7–11 Xie et al. (1993)
Komatiites: Xie and Kerrich (1994)
(La/Sm)
N
⳱ 2.2–2.5 Xie (1996)
(Gd/Yb)
N
⳱ 2.0–2.3 Other:
Tholeiites: Stone et al. (1995)
(La/Sm)
N
⳱ 2.1–2.8
(Gd/Yb)
N
⳱ 1.7–2.1
Stoughton-Roquemaure Group
NVZ, Abitibi subprovince, Superior
province, Quebec; 2714 Ma
Conformable contact with underlying rhyolitic
submerged arc complex (Hunter Mine Group).
Up to 2-km-thick mafic-ultramafic sequence;
mafic pillowed flows predominate with lesser
komatiite and komatiitic basalt flows (pillowed
and massive with spinifex texture). Interpreted
to represent plume-generated rifting of an arc
complex.
Th/Ta ⳱ 1.1–2.0 Dostal and Mueller (1997)
La/Yb ⳱ 0.9–3.0
Komatiites:
(La/Sm)
N
⳱ 0.6–1.1
(Gd/Yb)
N
⳱ 1.1–1.8
Tholeiites:
(La/Sm)
N
⳱ 0.8–1.0
(Gd/Yb)
N
⳱ 1.0–1.2
Kinojevis Group
SVZ, Abitibi subprovince, Superior
province, Ontario; 2701 Ma
Predominantly tholeiitic basalt sequence with
minor tholeiitic andesite to rhyolite.
Th/Ta ⳱ 3.0–4.6 Xie (1996)
La/Yb ⳱ 1.0–1.7
Interpreted to represent plume activity with melting
at shallow depth. May be tectonically juxtaposed
with adjacent sequences.
Tholeiites:
(La/Sm)
N
⳱ 0.7–1.0
(Gd/Yb)
N
⳱ 1.1–1.3
Lower Tisdale Group
Tisdale and Munro townships,
NVZ, Abitibi subprovince, Superior
province, Ontario; 2714 Ma
Possibly equivalent to Stoughton-Roquemaure
Group (described above). Tholeiitic basalt–
dominated sequence (massive flows) with lesser
komatiite flows (massive with cumulate and
spinifex textures), polysutured flow tops, and
rare pillow lavas. Classic Pyke Hill locality is part
of Munro township sequence. Interpreted by Xie
and co-workers to represent plume magmatism
with melting at different depths and accretion as
oceanic-plateau fragments.
Tisdale: Data:
Th/Ta ⳱ 1.5–3.3 Xie et al. (1993)
La/Yb ⳱ 0.4–1.2 Xie and Kerrich (1994)
Komatiites: Xie (1996)
(La/Sm)
N
⳱ 0.6–0.8 Other:
(Gd/Yb)
N
⳱ 1.1–1.4 Corfu et al. (1989)
Tholeiites: Jensen and Pyke (1982)
(La/Sm)
N
⳱ 0.8–1.1 Dostal and Mueller
(Gd/Yb)
N
⳱ 0.4–1.0 (1997)
Munro:
Th/Ta ⳱ 1.2–3.5
La/Yb ⳱ 0.3–3.0
Komatiites:
(La/Sm)
N
⳱ 0.3–0.5
(Gd/Yb)
N
⳱ 0.9–1.0
Tholeiites:
(La/Sm)
N
⳱ 0.6–1.2
(Gd/Yb)
N
⳱ 1.0–1.7
Note: (La/Sm)
N
and (Gd/Yb)
N
are primitive mantle-normalized ratios. SVZ—Southern Volcanic Zone; NVZ—Northern Volcanic Zone.
was previously depleted by melt extraction (Sun and Nesbitt,
1978).
Northern Superior province, Canada
Greenstone belts of the northern Superior province, par-
ticularly in northwestern Ontario, have received much attention
in recent years in terms of detailed geochemical studies (e.g.,
Tomlinson, 1996; Hollings, 1998). As a result, plume magma-
tism has been recognized in both oceanic and continental en-
vironments. Some sequences appear to represent fragments of
oceanic plateaus, and on the basis of lithostratigraphy, these
were called oceanic mafic-plain assemblages (Thurston and
Chivers, 1990; Thurston, 1990). Other greenstone belts repre-
sent plume-generated sequences that overlie older tonalitic or
volcanic basement, and these were called platform assemblages
by the same authors. Results from Archean greenstone belts of
the northern Superior province are summarized in Tables 2 and
3 and La/Yb versus Th/Ta relationships are shown in Figures 5
and 6.
Greenstone belts with oceanic affinities. With the excep-
tion of the Heaven Lake greenstone belt and the Wawa sub-
province greenstone belts, the oceanic plume–generated se-
quences that have been recognized in the northern Superior
province are thought to represent fragments or slivers of oce-
anic plateaus. They are no more than 5 km thick and do not
have significant lateral extent. The Heaven Lake sequence is
also ⬃5 km thick, but it is part of a more extensive plume-
generated magmatic province that also contains the adjacent
Steep Rock and Lumby Lake greenstone belts (Tomlinson et
al., 1998a, 1999a; Fig. 3). The Wawa greenstone belts are also
laterally extensive and encompass the Schreiber-Hemlo and
White River–Dayohessarah belts (Polat et al., 1999; Fig. 3). All
of the sequences are dominated by mafic lavas, but they vary
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

K.Y. Tomlinson and K.C. Condie346
Figure 4. La/Yb vs. Th/Ta diagram for plume-generated sequences of
the Abitibi subprovince, Superior province, Canada. See Table 1 for
the data sources. Other information given in Figure 2.
considerably in the proportion of preserved ultramafic units.
The Vizien greenstone belt contains an abundance of ultramafic
intrusive rocks; the Heaven Lake, Pickle Crow, and Wawa se-
quences contain a small proportion of komatiite to komatiitic
basalt lavas; and the southern Onaman–Tashota sequence con-
tains no known komatiites. Each sequence generally contains
thin and rare interbeds of BIF and/or chert.
Th/Ta and La/Yb ratios for basalts, ultramafic intrusive
rocks, and komatiites are generally similar to those of modern
oceanic-plateau basalts (Fig. 5). About 80% of the samples fall
within the oceanic plateau basalt or ocean island basalt fields.
Of the samples that fall outside these fields, some have higher
Th/Ta ratios (generally Th enrichment and/or small negative Nb
anomalies on mantle-normalized multi-element diagrams), and
some have lower La/Yb ratios (generally resulting from slight
LREE depletion, Table 2).
Greenstone belts with continental affinities. Several
greenstone belts (or assemblages within greenstone belts) have
been identified in the northwestern Superior province as rep-
resenting plume-generated sequences that erupted through con-
tinental basement, and these are generally of similar age (2.99-
2.90 Ga; Table 3). In both the Steep Rock and the North Caribou
Lake greenstone belts, an unconformity is preserved between
older tonalitic basement (or older mafic to felsic volcanic base-
ment), and slightly younger sedimentary rocks (quartz arenites,
arkoses, limestones, and BIF), which are overlain by crustally
Figure 5. La/Yb vs. Th/Ta diagram for plume-generated greenstone
belts of the Northern Superior province with oceanic affinities. See
Table 2 for the data sources. Other information given in Figure 2.
Figure 6. La/Yb vs. Th/Ta diagram for plume-generated greenstone
belts of the Northern Superior province with continental affinities. See
Table 3 for the data sources. Other information given in Figure 2.
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

Archean mantle plumes: Evidence from greenstone belt geochemistry 347
TABLE 2. PLUME-GENERATED GREENSTONE BELTS WITH PROBABLE OCEANIC SETTINGS IN THE
NORTHERN SUPERIOR PROVINCE, CANADA
Location and age Stratigraphy and relationships Ratios References
Heaven Lake greenstone belt
Central Wabigoon subprovince,
Superior province, Ontario;
2954 Ma
⬃5-km-thick sequence of tholeiitic pillowed flows
with ⬃500 m of pillowed and massive komatiitic
flows having thin interbeds of BIF and chert.
Sequence intruded by younger granitoids; hence
basal relationships are unclear. Interpreted as
plume generated but unclear if oceanic or
continental environment.
Th/Ta ⳱ 1.2–2.2 Data:
La/Yb ⳱ 1.1–2.1 Tomlinson et al. (1998a)
Komatiites:
(La/Sm)
N
⳱ 0.6–1.3 Other:
(Gd/Yb)
N
⳱ 0.9–1.0 Tomlinson et al. (1998a)
Tholeiites:
(La/Sm)
N
⳱ 0.8–1.1
Sage, 1998
(Gd/Yb)
N
⳱ 0.9–1.3
Thurston et al., 1987
Southern Onaman–Tashota terrane
eastern Wabigoon subprovince,
Superior province, Ontario;
⬎2740 Ma?
⬃5-km-thick sequence of tholeiitic pillowed and
massive flows intruded by gabbro sills and cut
by felsic dikes. Overlain by ⬃5-km-thick felsic
volcanic sequence that erupted through the
lower mafic sequence. Interpreted to represent
an accreted fragment of an oceanic plateau with
overlying arc volcanism.
Th/Ta ⳱ 1.0–1.5 Tomlinson, 1996
La/Yb ⳱ 1.6–2.0
Tholeiites:
(La/Sm)
N
⳱ 1.0–1.1
(Gd/Yb)
N
⳱ 1.1–1.2
Vizien greenstone belt
Minto block, Superior province,
Quebec; 2786 Ma
Allochthonous mafic-ultramafic sequence (⬍1km
thick) in fault contact with adjacent sequences
in the belt. Sequence comprises a basal
serpentinite unit; ultramafic schists with gabbro
pods, komatiite, and hyaloclastite; an overlying
mafic unit of pillow lavas, fragmental rocks, and
chert. Sequence is also intruded by gabbro sills
and peridotite. Interpreted to represent a sliver
of an oceanic plateau.
Th/Ta ⳱ 3.0–5.0
La/Yb ⳱ 0.8–2.1
Skulski and Percival, 1996
Ultramafics:
(La/Sm)
N
⳱ 1.3–2.3
(Gd/Yb)
N
⳱ 0.7–0.8
Tholeiites:
(La/Sm)
N
⳱ 0.8–0.9
(Gd/Yb)
N
⳱ 1.0–1.3
Wawa greenstone belts
Wawa subprovince, Superior
province, Ontario; ca. 2.7 Ga
Sequence of pillow basalts, massive flows, and
amphibolites with spinifex-textured and olivine
cumulate ultramafic flows. Interpreted as
fragments of oceanic plateau or oceanic islands
derived from a heterogeneous mantle plume.
Juxtaposed in a subduction-accretion-complex
setting.
Th/Ta ⳱ 1.2–3.0 Polat et al., 1999
La/Yb ⳱ 0.5–10
Komatiites:
(La/Sm)
N
⳱ 0.5–2.1
(Gd/Yb)
N
⳱ 0.9–2.4
Tholeiites:
(La/Sm)
N
⳱ 0.8–1.1
(Gd/Yb)
N
⳱ 0.9–1.2
Pickle Crow assemblage
Pickle Lake belt, Uchi subprovince,
Superior province, Ontario;
2.86 Ga
Sequence of mafic pillowed and massive flows
with intercalated iron formation and minor felsic
tuffs. Suggested to be juxtaposed with other
assemblages in the belt and interpreted to
represent a fragment of an oceanic plateau.
Th/Ta ⳱ 1.2–3.0 Hollings, 1998
La/Yb ⳱ 0.8–3.0
Komatiites:
(La/Sm)
N
⳱ 1.0
(Gd/Yb)
N
⳱ 1.7
Tholeiites:
(La/Sm)
N
⳱ 0.7–1.1
(Gd/Yb)
N
⳱ 0.9–1.2
Note: (La/Sm)
N
and (Gd/Yb)
N
are primitive mantle—normalized ratios.
North Caribou Lake belt, up to 2% of the Lumby Lake belt,
and less than 2% in the Red Lake and D’Alton Lake–Toronto
Lake belts. Pyroclastic komatiites are prominent lithologies in
both the Steep Rock and Lumby Lake greenstone successions.
On the La/Yb versus Th/Ta diagram (Fig. 6), 60% of the
lavas fall in the field of modern oceanic plateau basalts (or near
the boundary between the OPB and N-MORB fields), 11% fall
in the field of modern ocean-island basalts, and the remainder
have higher Th/Ta ratios but La/Yb ratios similar to those of
ocean-plateau basalts. Lavas from each of these belts have been
shown to be in part contaminated by continental crust (see ref-
erences in Table 3). The uncontaminated to least-contaminated
lavas plot in the OPB field, whereas more contaminated vol-
canic rocks have higher Th/Ta ratios and appear to fall on a
contaminated komatiites and basalts (Table 3) (Wilks and Nis-
bet, 1988; Thurston et al., 1991; Tomlinson, 1996; Tomlinson
et al., 1999a, 1999b; Hollings and Kerrich, 1999). In the Lumby
Lake greenstone belt, the komatiite-tholeiite sequence is in-
ferred to overlie tonalitic basement owing to the presence of
xenocrystic zircons (Tomlinson et al., 1999b) and crustally con-
taminated lavas (Tomlinson et al., 1999a). In the D’Alton Lake–
Toronto Lake belt and the Balmer assemblage of the Red Lake
greenstone belt, the lavas are inferred to have interacted with
older felsic crust on the basis of geochemical evidence of crustal
contamination of some of the mafic lavas (Tomlinson et al.,
1998a, 1998b). The proportion of komatiites preserved in all of
these greenstone belts is small. Komatiitic rocks represent up
to 10% of the section in the Steep Rock belt, up to 5% of the
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

K.Y. Tomlinson and K.C. Condie348
TABLE 3. PLUME-GENERATED GREENSTONE BELTS WITH PROBABLE CONTINENTAL SETTINGS
IN THE NORTHERN SUPERIOR PROVINCE, CANADA
Location and age Stratigraphy and relationships Ratios References
Steep Rock greenstone belt
Central Wabigoon subprovince,
Superior province, Ontario; ca.
2930 Ma
Tonalitic basement unconformably overlain by
platformal sedimentary rocks (quartz arenites,
limestones, iron formation), overlain by komatiitic
lapilli tuffs (Dismal Ashrock Formation) and
mafic lavas (Witch Bay Formation). Interpreted
as plume-generated magmatism in a
continental-rift environment.
Th/Ta ⳱ 1.2–2.2 Data:
La/Yb ⳱ 1.2–11.1 Tomlinson et al. (1999a)
Komatiites: Tomlinson (1996)
(La/Sm)
N
⳱ 1.3–2.6 Other:
(Gd/Yb)
N
⳱ 2.5–4.0 Wilks and Nisbet (1988)
Tholeiites:
(La/Sm)
N
⳱ 0.8–1.3
(Gd/Yb)
N
⳱ 1.1–1.4
Lumby Lake greenstone belt
Central Wabigoon subprovince,
Superior province, Ontario;
⬍2963-2898 Ma (mafic-ultramafic
sequence)
Implied tonalitic basement and early felsic volcanic
rocks (3014-2999 Ma) overlain by up to 7-km-
thick sequence of predominantly tholeiitic
pillowed and massive flows with thin intercalated
iron formation, chert and rare felsic volcanic
units. Approximately 200-m-thick komatiitic and
komatiitic basalt flows occur toward the top of
the sequence (cumulate and spinifex-textured
flows and pyroclastic units). Interpreted to
represent plume magmatism.
Th/Ta ⳱ 1.2–3.0 Tomlinson et al. (1999a)
La/Yb ⳱ 1.1–12
Komatiites:
(La/Sm)
N
⳱ 0.8–2.2
(Gd/Yb)
N
⳱ 0.9–2.4
Tholeiites:
(La/Sm)
N
⳱ 0.8–1.3
(Gd/Yb)
N
⳱ 0.9–1.3
D’Alton Lake–Toronto Lake belt
Central Wabigoon subprovince,
Superior province, Ontario; ca.
2920 Ma
Tectonic thickness of up to 14 km, sequence of
predominantly tholeiitic flows (massive and
lesser pillowed) and rare komatiitic flows, iron
formation, chert, and thin rhyolitic units. In fault
contact with 3075 Ma tonalite gneiss and older
volcanic sequence. Original relationship to older
crust is unknown. Interpreted to represent plume
magmatism possibly as a result of rifting of older
crust.
Th/Ta ⳱ 1.2–3.1 Tomlinson et al. (1998a)
La/Yb ⳱ 1.0–3.0
Tholeiites:
(La/Sm)
N
⳱ 0.7–1.4
(Gd/Yb)
N
⳱ 0.8–1.1
North Caribou Lake greenstone belt
North Caribou terrane, Superior
province, Ontario; ⬍2980-2932 Ma
Tonalitic basement implied and older mafic-felsic
volcanic sequence.
Th/Ta ⳱ 1.3–10 Data:
La/Yb-1.0–5 Tomlinson, unpublished
Keeyask assemblage: basal platformal
sedimentary rocks (⬃20 m thick) unconformably
overlie older volcanic sequence and are overlain
by a 50–1700-m-thick sequence of komatiitic
and komatiitic basalt flows (spinifex-textured
massive flows and pillowed flows). In turn
overlain by tholeiitic flows, minor komatiitic and
felsic volcanic rocks, and intercalated iron
formation and chert that compose the South
Rim and North Rim assemblages (up to 5 km
thick). Interpreted to represent plume-generated
rifting of older basement.
Komatiites: data (1998)
(La/Sm)
N
⳱ 0.9–2.9 Opapamiskan unit data
(Gd/Yb)
N
⳱ 0.9–1.4 from Hollings and
Tholeiites: Kerrich (1999)
(La/Sm)
N
⳱ 0.7–1.9 Other:
(Gd/Yb)
N
⳱ 1.0–1.3 Breaks et al. (2001)
Tomlinson et al. (1999b)
Thurston et al. (1991)
Balmer assemblage
Red Lake greenstone belt, North
Caribou terrane, Superior
province, Ontario; 2992-2964 Ma
Tholeiitic flows (pillowed and massive) with minor
spinifex-textured komatiitic flows, ultramafic sills,
iron formation, chert, felsic flows, and tuffs.
Section tectonically thickened to ⬃10 km.
Interpreted to represent plume magmatism.
Contamination in basalts also suggests older
crustal involvement.
Th/Ta ⳱ 1.6–3.0 Tomlinson et al. (1998b)
La/Yb ⳱ 0.8–2.5
Komatiites:
(La/Sm)
N
⳱ 0.5–0.7
(Gd/Yb)
N
⳱ 1.0–1.3
Tholeiites:
(La/Sm)
N
⳱ 0.8–1.4
(Gd/Yb)
N
⳱ 1.1–1.3
Note: (La/Sm)
N
and (Gd/Yb)
N
are primitive mantle—normalized ratios.
stone belt (Fig. 6). The OIB signature also occurs in komatiites
in the adjacent Steep Rock greenstone belt. Komatiites with the
OIB signature (Al-depleted komatiites) are depleted in HREEs
(Table 3), display negative Zr and Hf anomalies, and are ex-
plained by Tomlinson et al. (1999a) to represent melting in the
presence of majorite garnet. These komatiites are interpreted to
have come from the hot axial core of a mantle plume. Basalts
and Al-undepleted komatiites with the OPB geochemical sig-
nature, on the other hand, may represent melting at shallower
mixing array in which Archean upper crust represents the
crustal contaminant (Fig. 6). The La/Yb ratio is less sensitive
to crustal contamination, and therefore this ratio has not in-
creased as much as the Th/Ta ratio in the contaminated rocks.
This feature sets them apart from arc-related basalts, which
have higher La/Yb ratios and plot in the SZB (subduction-zone
basalt) field on the La/Yb versus Th/Ta diagram.
It is interesting that both OIB and OPB geochemical sig-
natures are found in komatiites from the Lumby Lake green-
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

Archean mantle plumes: Evidence from greenstone belt geochemistry 349
Figure 7. La/Yb vs. Th/Ta diagram for plume-generated greenstone
belts outside the Superior province with oceanic affinities. See Table
4 for the data sources. Other information given in Figure 2.
Figure 8. La/Yb vs. Th/Ta diagram for plume-generated greenstone
belts outside the Superior province with continental affinities. See Ta-
ble 5 for the data sources. Other information given in Figure 2.
depths and may have come from a plume head in which plume
and depleted mantle sources mixed (Tomlinson et al., 1999a,
1999b).
Other Archean provinces
Although collectively, greenstone belts of the Superior
province are the most extensively studied of all Archean green-
stone belts and they have the most complete geochemical data-
base in terms of immobile trace elements, several greenstone
belts from other Archean provinces also have sufficient geo-
chemical data to suggest plume sources. Examples of plume-
generated Archean greenstone belts from other regions are sum-
marized in Tables 4 and 5, and La/Yb versus Th/Ta data are
plotted in Figures 7 and 8.
Greenstone belts with oceanic affinities. Plume-related
greenstone belts with oceanic affinities include the Onverwacht
Group in the Barberton greenstone belt in South Africa and the
Kostomuksha greenstone belt in the northwestern Baltic Shield,
Russia. Both are well preserved, each with an ⬃3.5-km-thick
supracrustal section. The Fiskenaesset sequence in southwest-
ern Greenland and the Tungurcha greenstone belt in the western
Aldan Shield, Russia, on the other hand, are preserved as am-
phibolite layers and enclaves within tonalite gneisses.
Much of the Onverwacht Group was considered part of an
ophiolite sequence by de Wit et al. (1987). Later zircon geo-
chronology and other studies, however, have shown that this
group contains tectonically disparate components of different
ages (Kamo and Davis, 1994; de Ronde and de Wit, 1994). The
Komati Formation in the Onverwacht Group is an allochthon-
ous pile dominated by komatiites and komatiitic basalts, and
the lithologic sequence is suggestive of an oceanic plateau (Ta-
ble 4). Al-depleted komatiites with HREE depletion (Table 4)
are present, and majorite garnet fractionation was suggested by
Lahaye et al. (1995) to have played a role in the genesis of
some of the komatiites. Th/Ta and La/Yb ratios of komatiites
from the Komati Formation are similar to those of modern oce-
anic-plateau basalts (Fig. 7).
On the basis of stratigraphy and geochemistry, the Kosto-
muksha greenstone belt was interpreted by Puchtel et al. (1998)
to represent the upper part of an obducted oceanic plateau (Ta-
ble 4). The belt is dominated by mafic volcanic rocks (70%)
with lesser amounts of komatiite, gabbro, and komatiitic pyro-
clastic rocks, and the sequence is capped by chemical sediments
and felsic volcanic rocks. In terms of Th/Ta ratios, komatiitic
and basaltic lavas from the belt are similar to modern oceanic
plateau basalts (Fig. 7). La/Yb ratios are slightly lower than in
modern OPB, reflecting a more LREE-depleted mantle source
(Table 4).
Amphibolites of the Fiskenaesset sequence are not well
known geochemically. They are mafic to ultramafic in com-
position, and relict pillows are preserved in places. Weaver et
al. (1982) interpreted the amphibolites as a fragment of thick-
ened oceanic crust that was tectonically emplaced into the lower
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

K.Y. Tomlinson and K.C. Condie350
TABLE 4. PLUME-GENERATED ARCHEAN GREENSTONE BELTS OF PROBABLE OCEANIC SETTINGS FROM OTHER AREAS
Location and age Stratigraphy and relationships Ratios References
Lower Onverwacht Group
Barberton, South Africa;
3472 Ma
Komati Formation: ⬎3.5-km-thick allochthonous(?)
pile of komatiites and komatiitic basalt flows,
spinifex-textured flows, pillow lavas, ultramafic
dikes, rare volcaniclastic sedimentary units.
Th/Ta ⳱ 1.6–2.1 Data:
La/Yb ⳱ 2.2–3.5 Jahn et al. (1982)
Komatiites: Jochum et al. (1991)
(La/Sm)
N
⳱ 0.9–1.6 Lahaye et al. (1995)
(Gd/Yb)
N
⳱ 1.3–1.5 Other:
de Ronde and de Wit
(1994)
Kamo and Davis (1994)
Viljoen et al. (1982)
Armstrong et al. (1990)
de Wit et al. (1987)
Kostomuksha belt
Karelian domain, North-west
Baltic Shield, Russia;
U-Pb: ⬎2.795 Ga
Sm-Nd: 2.84 Ga
Pb-Pb: 2.81 Ga
3.5-km-thick supracrustal sequence comprising
70% mafic lavas (pillowed, variolitic, and
massive). Komatiites occur in middle and upper
parts of sequence as spinifex-textured flows,
volcanic breccias, and minor ash-flow and lapilli
tuffs. Sequence capped by ironstones and small
proportion of felsic volcanic rocks. Gabbro and
peridotite sills also intrude sequence. Interpreted
as part of an oceanic plateau.
Th/Ta ⳱ 1.1–2.2 Puchtel et al. (1998)
La/Yb ⳱ 0.6–1.0
Komatiites:
(La/Sm)
N
⳱ 0.4–0.6
(Gd/Yb)
N
⳱ 1.0–1.2
Tholeiites:
(La/Sm)
N
⳱ 0.6–0.7
(Gd/Yb)
N
⳱ 1.0–1.1
Fiskenaesset
South-West Greenland;
⬎2.85 Ga?
Amphibolites are part of gneiss complex and
possibly associated with crosscutting
anorthosite complex. Amphibolite layers up to
1 km thick are preserved and contain relict
pillows. Suggested oceanic environment of
formation.
Th/Ta ⳱ 0.7–3.3 Puchtel et al. (1993)
La/Yb ⳱ 1.0–1.3
Ultramafics:
(La/Sm)
N
⳱ 1.0
(Gd/Yb)
N
⳱ 1.0
Basalts:
(La/Sm)
N
⳱ 0.6–0.8
(Gd/Yb)
N
⳱ 0.9–1.1
Tungurcha greenstone belt
western Aldan Shield, Russia;
⬎2.8 Ga?
Enclaves of actinolite schist and amphibolite within
tonalitic gneisses. Interpreted as mafic and
komatiitic lava flows. Deep-mantle source
suggested for komatiites.
No HFSE data Weaver et al. (1982)
La/Yb ⳱ 0.9–4.5
Komatiites:
(La/Sm)
N
⳱ 0.6–1.4
(Gd/Yb)
N
⳱ 1.0–2.0
Note: (La/Sm)
N
and (Gd/Yb)
N
are primitive mantle-normalized ratios.
greenstone belts with continental affinities are tabulated in Ta-
ble 5, and Th/Ta and La/Yb ratios are plotted in Figure 8.
The ca. 3.46 Ga North Star Basalt is the lowest strati-
graphic unit in the Talga Talga Subgroup in the Pilbara craton
in Western Australia (Barley, 1993; Krapez, 1993; Thorpe et
al., 1992). The base of the 4-km-thick sequence is not exposed,
and thick units of arc-related volcanic and sedimentary rocks
overlie the Talga Talga Subgroup, which includes tholeiitic ba-
salts, diabase sills and dikes, gabbro sills and dikes, and minor
high-Mg tholeiites. The subgroup also includes minor cherty
sedimentary rocks and felsic tuffs, suggestive of shallow-water
deposition (Barley, 1993). Lavas in the North Star Basalt unit
range from massive to pillowed and are commonly associated
with hyaloclastic breccias indicating subaqueous eruption. Ko-
matiites have not as yet been identified in the Talga Talga Sub-
group, although basaltic komatiites may be present. The North
Star basalts plot in and above the OPB field on the La/Yb versus
Th/Ta diagram (Fig. 8); their mostly higher Th/Ta ratios are
suggestive of mantle plume sources that have been variably
contaminated with continental crust (Condie, 1994).
The Belingwe greenstone belt is divided into the ca. 2.7
Ga Ngezi Group and the ca. 2.9 Ga Mtshingwe Group (Table
crust. The Th/Ta and La/Yb ratios for the amphibolites show a
greater range than those of modern oceanic plateau basalts (Fig
7). This may reflect mixing of plume and depleted mantle
sources, poor analytical data, or element mobility during high-
grade metamorphism.
Actinolite schist and amphibolite that compose most of the
Tungurcha greenstone belt are compositionally komatiites and
tholeiites and contain rare relict textures to suggest that they
were submarine flows (Puchtel et al., 1993). Both Al-depleted
and Al-undepleted komatiites are present in this belt, and the
petrogenesis of the Al-depleted komatiites appears to have in-
volved garnet or majorite fractionation and partial melting at
high pressures (Puchtel et al., 1993). We suggest that this in-
terpretation is consistent with the sequence having been gen-
erated by a mantle plume. Although Th/Ta ratios are not avail-
able for Tungurcha rocks, La/Yb ratios (0.9-4.5) are similar to
those of oceanic plateau basalts.
Greenstone belts with continental affinities. Although
there is only a limited amount of published HFSE data available
for plume-generated greenstone belts with continental affinities
outside of the Superior province, some of the best-known and
well-described greenstone belts fall into this category. Archean
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

TABLE 5. PLUME-GENERATED ARCHEAN GREENSTONE BELTS WITH PROBABLE CONTINENTAL SETTINGS FROM OTHER AREAS
Location and age Stratigraphy and relationships Ratios References
Mtshingwe Group
Belingwe greenstone belt,
Zimbabwe; ca. 2900 Ma
Older gneissic crustal basement. Approximately
8-km-thick andesitic volcanic pile (Hokonui
Formation) conformably(?) overlain by 2.5–6-
km-thick Bend Formation comprising repeated
cycles of komatiites, komatiitic basalts and
basalts with interbedded ironstone and chert.
Interpreted to represent rifting of the andesitic
pile. Disconformably overlain by the 1–2-km-
thick Koovoodale Formation comprising mainly
conglomerates.
No HFSE data Orpen et al. (1993)
(La/Sm)
N
⳱ n.a. Bickle et al. (1993)
(Gd/Yb)
N
⳱ n.a.
Ngezi Group
Belingwe greenstone belt,
Zimbabwe; 2692 Ma
Unconformably overlies granite-gneiss basement.
Basal sedimentary units (up to 200 m thick:
sandstones, quartzite, banded iron formation,
limestones, and siltstones) overlain by ⬃1-km-
thick Reliance Formation (tholeiitic basalt,
komatiitic basalt, and komatiite pillow lavas,
flows, and tuffs), overlain by ⬃5.5-km-thick
Zeederbergs Formation (pillowed and massive
tholeiitic basalt).
Th/Ta ⳱ 1.2–8 Data:
La/Yb ⳱ 0.8–7 Silva, written commun.
(1998)Komatiites:
Silva (1998)(La/Sm)
N
⳱ 0.6–1.0
Other:(Gd/Yb)
N
⳱ 1.1–1.4
Bickle et al. (1993)Tholeiites:
Nisbet et al. (1993)(La/Sm)
N
⳱ 0.7–3.2
Chauvel et al., 1993(Gd/Yb)
N
⳱ 1.3–1.5
Kambalda, Norseman–Wiluna belt
Yilgarn craton, Western Australia;
ca. 2.7 Ga
Continental crustal basement. Th/Ta ⳱ 1.1–2.6 Lesher and Arndt (1995)
Lunnon Formation: ⬎2 km of tholeiitic basalt flows. La/Yb ⳱ 0.8–1.2 Jochum et al. (1991)
Kambalda komatiite: up to 1 km of komatiite flows
with cumulate and spinifex textures, rare pillows,
and interflow sedimentary units.
Komatiites:
(La/Sm)
N
⳱ 0.5–0.8
(Gd/Yb)
N
⳱ 0.9–1.1
Arndt and Jenner (1986)
Sylvester et al. (1997)
Devon Consuls basalt: up to 100-m-thick pillowed
and massive komatiitic basalt flows with basalt
to gabbro massive flows.
Lunnon basalts:
(La/Sm)
N
⳱ 0.8–0.9
(Gd/Yb)
N
⳱ n.a.
Paringa basalt: ⬎500-m-thick pillowed komatiitic
basalt flows and mafic sills and/or thick flows.
Kambalda komatiite suggested to represent
magmas derived from a plume tail.
Devon basalts:
(La/Sm)
N
⳱ 1.3–1.4
(Gd/Yb)
N
⳱ n.a.
Forrestania greenstone belt
Southern Cross province,
Yilgarn craton, Western Australia;
ca. 2.9 Ga
Predominantly mafic sequence but contains
komatiites intercalated with BIF. Komatiites
occur as large unconfined sheet flows and as
smaller channelized flows. Both spinifex and
cumulate textures are common.
Th/Ta ⳱ 1.4–10.1 Perring et al., 1996
La/Yb ⳱ 0.6–5
Komatiites:
(La/Sm)
N
⳱ 0.7–1.7
(Gd/Yb)
N
⳱ 1.0–1.5
Tholeiites:
(La/Sm)
N
⳱ 1.2
(Gd/Yb)
N
⳱ 1.2
Kam Group
Yellowknife belt, Slave province,
Canada; 2680–2660 Ma
Central Slave Cover Group unconformable overlies
Central Slave Basement Complex and is
overlain by a parautochthonous sequence
(⬃13 km thick) of massive and pillowed,
predominantly mafic tholeiitic flows. Lower Kam
is a 6.5-km-thick sequence comprising 5.9 km of
mafic volcanic rocks capped by 600 m of dacitic
flows (Townsite flows).
Th/Ta ⳱ 1.0–10 Data:
La/Yb ⳱ 1.3–4.3 Goodwin (1988)
Tholeiites: Other:
(La/Sm)
N
⳱ 0.8–2.0 Bleeker et al. (1999)
(Gd/Yb)
N
⳱ n.a.
North Star basalts
Pilbara craton, Western Australia;
ca. 3450 Ma
Talga Talga subgroup (4-km-thick base unex-
posed): Massive and pillowed tholeiitic basalt
flows, hyaloclastite, high-Mg flows, diabase and
gabbro sills and dikes, small proportion of cherty
sedimentary rocks and felsic tuffs.
Th/Ta ⳱ 1.2–5 Barley (1993)
La/Yb ⳱ 1–4 Krapez (1993)
(La/Sm)
N
⳱ n.a. Thorpe et al. (1992)
(Gd/Yb)
N
⳱ n.a.
Kolar schist belt
Dharwar craton, southern India;
ca. 2.9 Ga
Probable sialic basement. Approximately 4-km-
thick sequence of tholeiitic amphibolites folded
into a synform, with small proportion of
komatiitic amphibolites occurring near the base.
Sequence contains BIF and both amygdaloidal
and pillow structures. Suggested plume model.
No HFSE data Rajamani et al. (1985)
Komatiites: Rajamani (1990)
(La/Sm)
N
⳱ n.a.
(Gd/Yb)
N
⳱ 1.3–2.0
Tholeiites:
(La/Sm)
N
⳱ n.a.
(Gd/Yb)
N
⳱ 1.0–1.6
Prince Albert Group and
Woodburn Lake Group
Northwest Territories, Canada;
⬎2.7 Ga
The two groups are thought to be correlative.
Sequences are preserved within gneissic
terrane. Mafic volcanic rocks predominate
including pillow lavas, but komatiites, quartzite,
iron formation and lesser felsic volcanic rocks
are also widespread. Rocks erupted through or
were deposited on gneissic basement. Deep-
mantle source suggested for komatiites.
No HFSE data Data:
Fryer and Jenner (1978)
Other:
Schau (1982)
Kjarsgaard et al. (1997)
Note: (La/Sm)
N
and (Gd/Yb)
N
are primitive mantle-normalized ratios; n.a. ⳱ not available.
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

K.Y. Tomlinson and K.C. Condie352
The ca. 2.7 Ga Kam Group of the Yellowknife greenstone
belt in the Slave province of northern Canada may represent a
parautochthonous breakup sequence that overlies the ca. 2.8 Ga
Central Slave Cover Group and underlying 2.9-3.3 Ga Central
Slave Basement Complex rocks (Bleeker et al., 1999). The Kam
Group is dominated by voluminous pillowed tholeiitic flows
with intercalated banded iron formation (Table 5). The group
is unconformably overlain by 2.68-2.66 Ga calc-alkalic vol-
canic rocks of the Banting Group and ca. 2.66 Ga turbidites of
the Duncan Lake Group (Bleeker et al., 1999; Henderson, 1981;
Goodwin, 1988). Data from the lower part of the Kam Group
are plotted here (Goodwin, 1988). Goodwin (1988) suggested
a primitive (high Ni) mantle source for the lower Kam Group.
La/Yb ratios are similar to those of OPBs, whereas Th/Ta ratios
are higher, suggestive of crustal contamination of plume-de-
rived magmas (Fig. 8).
The ca. 2.9 Ga Kolar schist belt of the Dharwar craton in
southern India contains an ⬃4-km-thick sequence of tholeiitic
amphibolites with minor komatiitic amphibolits (Table 5). The
belt may have a sialic basement, and mafic and komatiitic lavas
are thought to have been plume-generated (Rajamani et al.,
1985). There are limited REE and HFSE data available for this
belt, but the tholeiites have flat to LREE-enriched trace element
patterns, and the komatiites are Ti and Zr enriched and HREE
depleted (Table 5; Rajamani et al., 1985).
Amphibolites of the ⬎2.7 Ga Prince Albert Group and the
Woodburn Lake Group in the Churchill province, Northwest
Territories, Canada, may also represent plume magmatism of
continental affinity. Both groups comprise amphibolites, thol-
eiitic pillowed flows, komatiites, quartzites, BIF, and lesser fel-
sic volcanic rocks preserved within a gneiss terrane and are
suggested to have been deposited on gneissic basement (Table
5; Fryer and Jenner, 1978; Schau, 1982; Kjarsgaard et al.,
1997). Geochemical data suggest that the komatiites are unde-
pleted in Al and have chondritic Ti/Zr and Ti/Al ratios and
unfractionated HREE profiles. Tholeiites generally show flat
REE patterns with slight LREE depletion (Fryer and Jenner,
1978).
DISCUSSION
Our results suggest that there are two main Archean green-
stone belt associations and one minor one that are recognized
as plume-generated sequences. First is the mafic-plain associ-
ation (predominantly pillowed tholeiitic flows Ⳳ komatiites Ⳳ
chert Ⳳ BIF) as exemplified by the Kostomuksha greenstone
belt in the northwestern Baltic Shield and the Heaven Lake
greenstone belt in the northwestern Superior province. The
mafic-plain association is similar to lithologic assemblages
found in young oceanic plateaus (Kerr et al., 1998), but in the
Archean, sequences are generally preserved as fragments or
slivers, and their original lateral extent and thickness are un-
known. The Archean sequences also differ from young oceanic
plateaus in that komatiites are a common, although not a major,
5). The Ngezi Group unconformably overlies granite-gneiss
basement. It is similar to platform assemblages in the Superior
province (especially the Steep Rock greenstone belt) in that it
comprises basal platform sedimentary rocks overlain succes-
sively by komatiites (Reliance Formation) and tholeiites. Lavas
from the Reliance Formation have La/Yb ratios similar to mod-
ern OPBs, but they have higher Th/Ta values that probably
reflect crustal contamination. Nisbet et al. (1993b) suggested
that the Ngezi Group may have formed as a result of extension
and adiabatic decompression-driven melting. We suggest that
like the Steep Rock greenstone belt, the Ngezi Group may
represent plume-generated magmatism in a continental envi-
ronment.
The komatiite-tholeiite Bend Formation of the Mtshingwe
Group, also in Zimbabwe, conformably(?) overlies a thick an-
desitic volcanic pile (Orpen et al., 1993) (Table 5). The ande-
sites are inferred to represent part of a continental-arc sequence
(Nisbet et al., 1993). The komatiite-tholeiite sequence has been
interpreted to have formed during rifting of this andesitic pile;
hence the sequence may represent plume-generated rifting of
arc crust. Unfortunately, HFSE and REE data are unavailable
for the Mtshingwe Group, so this interpretation is speculative.
The ca. 2.7 Ga greenstone sequence at Kambalda (Norse-
man-Wiluna belt) in Western Australia is well known for its
komatiites and komatiitic basalts that have interacted with con-
tinental basement (Arndt and Jenner, 1986; Lesher and Arndt,
1995). The ⬃4-km-thick sequence comprises komatiites, ko-
matiitic basalts, and tholeiite flows with the greatest abundance
of komatiites occurring in the ⬃1-km-thick Kambalda Koma-
tiite Formation (Table 5). These komatiitic magmas may have
been derived from a high-temperature plume tail (Lesher and
Arndt, 1995). Although a full range of REE and HFSE data are
not available, a limited number of samples are shown in Figure
8 from the Kambalda Komatiite Formation. As discussed by
Lesher and Arndt (1995), crustal contamination is minimal in
the lower parts of the section but shows a general increase with
increasing stratigraphic height. Basalts from the Lunnon For-
mation and channel-facies komatiites from the Kambalda Ko-
matiite Formation are relatively uncontaminated and show Th/
Ta ratios similar to OPBs. Basalts from the Devon Consuls
Formation have higher Th/Ta ratios (Sylvester et al., 1997).
The ca. 2.9 Ga Forrestania greenstone belt of the Southern
Cross province, Yilgarn craton, contains an ⬃5-km-thick suc-
cession of predominantly mafic volcanic rocks with a small
proportion of laterally extensive komatiite flows, BIF, and mi-
nor felsic sedimentary rocks (Perring et al., 1996). Both Al-
undepleted and Al-depleted komatiites are present and are in-
terpreted to reflect different depths of melting and variations in
degree of partial meting and proportion of residual garnet in
the source (Perring et al., 1996). Crustal contamination is also
suggested to have played an important role in the petrogenesis
of the lavas. Komatiites from the belt plot either within the
oceanic-plateau field or show higher Th/Ta ratios that suggest
variable crustal contamination of plume-derived magmas.
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

Archean mantle plumes: Evidence from greenstone belt geochemistry 353
component. The mafic-plain association is preserved through-
out the Archean to as far back as 3.5 Ga, as shown by the
Onverwacht Group in the Barberton greenstone belt in South
Africa. Although not as well documented, the Isua supracrustal
rocks in South-West Greenland (ca. 3.8 Ga) may be another,
even older example of a mafic-plain association (Rosing et al.,
1996). Numerous examples of mafic-plain greenstone belts are
found at 3.0–2.9, 2.8, and 2.75–2.7 Ga.
The second association is the platform association (older
continental crustal basement Ⳳ platform sedimentary rocks Ⳮ
komatiites Ⳮ tholeiitic pillowed flows) as exemplified by the
Steep Rock greenstone belt in the northwestern Superior prov-
ince and the Ngezi Group of the Belingwe greenstone belt in
Zimbabwe. The platform greenstone association may be anal-
ogous to modern continental flood basalts in terms of a volcanic
sequence erupted through, and deposited on, continental base-
ment, although the Archean volcanic sequences are submarine.
Alternatively, the platform greenstone belts may be analogous
to rifted volcanic margins in terms of shallow-water sedimen-
tary rocks, but komatiites have not been found in either of these
settings in the modern record; therefore, this assemblage type
may not have a direct modern analogue. The assemblage type
has also been recognized throughout the Archean from 3.46 Ga
(North Star Basalt in Western Australia) to 2.69 Ga (in Zim-
babwe), and it is particularly widespread in the 3.0–2.9 Ga
greenstone belts in the northwestern Superior province.
A third but minor lithologic association may represent arc
rifting in response to a mantle plume (andesitic to felsic arc
rocks conformably overlain by komatiite-tholeiite sequences).
Examples of this association are the ca. 2.9 Ga Mtshingwe
Group in the Belingwe greenstone belt of Zimbabwe and the
ca. 2.71 Ga Stoughton-Roquemaure Group of the Abitibi sub-
province in Canada.
Geochemical signatures of Archean plume-generated green-
stone belts may be more diverse than those of modern oceanic
plateau volcanic sequences. This diversity is partly due to crus-
tally contaminated plume magmas with elevated Th/Ta ratios,
but some of the oceanic Archean greenstone sequences show
higher Th/Ta ratios and lower La/Yb ratios than young oceanic-
plateau basalts. Workers on Abitibi greenstone sequences have
generally ascribed higher Th/Ta ratios to recycling of continen-
tal crust within mantle plumes, whereas lower La/Yb ratios may
be ascribed to LREE depletion of the source due to previous
melt extraction. High La/Yb ratios that result from HREE de-
pletion (combined with negative Zr and Hf anomalies) indicate
majorite garnet in the source and melting at depths of 400–650
km (Xie et al., 1993). The greater diversity in Archean plume-
generated sequences may also just be apparent, in that most
modern oceanic plateaus have only been sampled at their sur-
face (i.e., on the seabed), and hence, they may also be geo-
chemically more heterogeneous than currently described.
Plume-derived lavas contaminated by continental crust
may acquire a “pseudo-subduction” zone geochemical signa-
ture (Ta-Nb depletion and Th enrichment). Hence, some Ar-
chean greenstone belts that have been described as arc-type on
the basis of their geochemistry may represent crustally contam-
inated plume-type magmas. Stratigraphic relationships, particu-
larly the proportion of felsic volcanic or calc-alkalic rocks, in
addition to Nd isotopes (to evaluate crustal contamination) can
help to distinguish these two cases. Because crustal contami-
nation increases Th/Ta and La/Yb ratios, the abundance of Ar-
chean greenstone sequences with plume geochemical signatures
probably represents a minimum. In the western Superior prov-
ince, ⬃20% of the Archean greenstone belts for which there is
geochemical data appear to have plume sources. Adding to this
value the mafic-plain and platform assemblages for which there
is no geochemical data suggests that ⬃35% of greenstone belts
may have been generated by plumes. In terms of the worldwide
database for greenstone belts that have geochemical data and/
or lithologic descriptions, it would appear that about 35% of
Late Archean (3.0–2.5 Ga) greenstone belts have plume affin-
ities. To speculate, ⬃80 % of Early Archean (⬎3.0 Ga) green-
stone belts may have plume affinities, although there are rela-
tively few Early Archean examples to base this percentage on.
These proportions of Archean plume-related greenstones are
rather greater than those found in post-Archean greenstones
(Condie, 1994, 1997, 1999).
This observation that plume-related basalts and komatiites
are a major component of Archean greenstone belts and appear
to be more frequent in the Archean than subsequently rests in
part on two major assumptions: (1) Archean OPBs are truly
part of oceanic plateaus and are not oceanic crust generated at
mid-ocean ridges, and (2) the relative abundance of greenstone
belts with plume affinities is not a relict of preservational bias
in the Archean. We next explore each of these assumptions.
1. It is important to distinguish mid-ocean ridge basalts
from oceanic plateau basalts in Archean greenstone belts be-
cause the former type does not require a mantle plume com-
ponent. Because lithologic assemblages formed at mid-ocean
ridges and above oceanic mantle plumes may be similar (i.e.,
submarine basalts and intrusive equivalents together with minor
chemical sedimentary rocks), mid-ocean ridge and oceanic pla-
teau basalts may be difficult to distinguish from each other,
especially in the Archean when the production rate of oceanic
crust was probably greater than today (Abbott et al., 1994). If
Archean oceanic crust was on the order of 20 km thick (Sleep
and Windley, 1982; Galer and Mezger, 1998), the oceanic lith-
osphere would have been more buoyant and could have be-
haved much like an oceanic plateau. So where do we draw the
line between oceanic plateau and mid-ocean-ridge basalts in
greenstone successions? This is a difficult question, and with
our present database, we suggest three major observations that
favor, but do not prove plume affinities for Archean mafic-plain
greenstones. First of all, komatiites, which reflect higher mantle
temperatures than basalts and probably greater melting depths
(Herzberg, 1995), occur almost exclusively in these types of
greenstone belts. If komatiites require a mantle plume source
as most data suggest (Campbell et al., 1989; Xie and Kerrich,
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

K.Y. Tomlinson and K.C. Condie354
1994; Kerrich et al., 1998; Kerrich et al., 1999; Galer and Mez-
ger, 1998; Campbell, 1998), then komatiite-bearing greenstone
belts should reflect mantle plume sources. Second, the Th/Ta
ratios in many Archean mafic-plain basalts are greater than in
N-MORBs, and the average values of these ratios typically plot
in the OPB field (or above) rather than in the N-MORB field
on La/Yb versus Th/Ta plots (Figs. 4–8). Although some green-
stone basalts may come from depleted-mantle sources similar
to N-MORB sources, many if not most of these basalts come
from undepleted sources like those of mantle plumes. And fi-
nally, the Zr, Hf, Nb, and Th anomalies that have been con-
firmed to exist in some mafic-plain and platform komatiites and
basalts suggest a deep mantle source, one that is characteristic
of mantle plumes but not of shallow mid-ocean ridge sources
(Xie and Kerrich, 1994; Kerrich et al., 1998; Tomlinson et al.,
1999a).
2. If plume-related basalts were selectively preserved in the
Archean compared to the post-Archean, their higher frequency
in Archean greenstone belts may reflect preservational bias
rather than a greater abundance of Archean compared to post-
Archean mantle plumes. What may lead to increased preser-
vation of these greenstone belts in the Archean? The oceanic
Archean lithosphere was probably thicker and more buoyant
than the post-Archean lithosphere because of higher mantle
temperatures in the Archean (Bickle, 1986; Hoffman and Ran-
alli, 1988; Davies, 1992). Hence, it would be less easily sub-
ducted, and during plate collisions, oceanic plateaus may have
accreted to the growing continental crust (Condie, 1997), thus
selectively preserving volcanic rocks with plume sources com-
pared to volcanic rocks with arc sources. After the Archean,
when mantle temperatures fell, oceanic plateaus may have been
more easily subducted; thus, the frequency of preservation of
oceanic plateau-type greenstones decreased. Indeed, the appar-
ent higher frequency of oceanic plateau-type greenstones in the
Archean could be due solely to this factor. However, the variety
of plume-generated assemblages that are present in the Ar-
chean—i.e., mafic-plain and OPB assemblages, platform as-
semblages, rifted-arc assemblages, and continental flood-basalt
assemblages—combined with their high frequency, does seem
to suggest that mantle plumes may have been more widespread
in the Archean than subsequently.
If, as suggested, plumes and hence oceanic plateaus were
more widespread in the Archean than at present, then a number
of questions arise. What proportion of the ocean floor was
plume-generated crust and what proportion was generated at
mid-ocean spreading ridges? Was mid-ocean ridge crust pro-
portionally less abundant than today, and, if so, does this help
to explain the lack of identification of ophiolites in the Ar-
chean? As greenstone belt studies continue and the geochemi-
cal, isotopic, and geochronological tools at our disposal become
more sophisticated, we may be able to answer such questions,
but for now they remain the topics of future research.
ACKNOWLEDGMENTS
We thank K. Silva and A. Goodwin for the use of unpublished
geochemical data and N.T. Arndt, H. Helmstaedt, and R. Ernst
for reviews that improved this manuscript. Tomlinson acknowl-
edges the University of Portsmouth (UK), the Ontario Geolog-
ical Survey, the Geological Survey of Canada, and LITHO-
PROBE of Canada for supporting her work on Archean
greenstone belts. Condie acknowledges the U.S. National Sci-
ence Foundation that has funded many of his projects dealing
with Archean greenstone belts.
REFERENCES CITED
Abbott, D., Drury, R., and Smith, W.H.F., 1994, Flat to steep transition in
subduction style: Geology, v. 22, p. 937–940.
Abouchami, W., Boher, M., Michard, A., and Albare´de, F., 1990, A major 2.1
Ga event of mafic magmatism in West Africa: An early stage of crustal
accretion: Journal of Geophysical Research, v. 95, p. 17605–17629.
Armstrong, R.A., Compston, W., de Wit, M.J., and Williams, I.S., 1990, The
stratigraphy of the 3.5 to 3.2 Ga Barberton greenstone belf revisited: A
single zircon ion microprobe study: Earth and Planetary Science Letters,
v. 101, p. 90–106.
Arndt, N.T., 1991, High Ni in Archean tholeiites: Tectonophysics, v. 187,
p. 411–420.
Arndt, N.T., and Jenner, G.A., 1986, Crustally contaminated komatiites and
basalts from Kambalda, Western Australia: Chemical Geology, v. 56,
p. 229–255.
Barley, M. E., 1993, Volcanic, sedimentary and tectonostratigraphic environ-
ments of the 3.46 Ga Warrawoona Megasequence: A review: Precambrian
Research, v. 60, p. 47–67.
Barrie, C.T., 1997, Komatiite flows of the Kidd Creek footwall, Abitibi Sub-
province, Canada: Geological Association of Canada/Mineralogical As-
sociation of Canada, Annual Meeting, Ottawa, Abstracts, v. 22, p. A-9.
Bickle, M. J., 1986, Implications of melting for stabilization of the lithosphere
and heat loss in the Archean: Earth and Planetary Science Letters, v. 80,
p. 314–324.
Bickle, M.J., Arndt, N.T., Nisbet, E.G., Orpen, J.L., Martin, A., Keays, R.R.,
and Renner, R., 1993, Geochemistry of the igneous rocks of the Belingwe
greenstone belt, Zimbabwe, in Bickle M.J., and Nisbet, E.G., eds., The
geology of the Belingwe greenstone belt, Zimbabwe, a study of the evo-
lution of Archean continental crust: Rotterdam, Balkema, Geological So-
ciety of Zimbabwe Special Publication 2, p. 175–214.
Bleeker, W., Ketchum, J.W.F., Jackson, V.A., and Villeneuve, M., 1999, The
central Slave basement complex: Part 1. Its structural topology and
autochthonous cover: Canadian Journal of Earth Sciences, v. 36, n. 7,
p. 1083–1109.
Breaks, F.W., Osmani, I.A., and DeKemp, E.A., 2001, Geology of the North
Caribou Lake area, District of Kenora (Patricia portion): Ontario Geolog-
ical Survey, Open File Report 6023, 82 p.
Campbell, I. H., 1998, The mantle’s chemical structure: insights from the melt-
ing products of mantle plumes, in Jackson, I., ed., The Earth’s mantle:
Composition, structure, and evolution: Cambridge, Cambridge University
Press, p. 259–310.
Campbell, I.H., Griffiths, R.W., and Hill, R.I., 1989, Melting in an Archean
mantle plume: Heads it’s basalts, tails it’s komatiites: Nature, v. 339,
p. 697–699.
Chauvel, C., Dupre´, B., and Arndt, N.T., 1993, Pb and Nd isotopic correlation
in the Belingwe komatiites and basalts, in Bickle, M.J., and Nisbet, E.G.,
eds., The geology of the Belingwe greenstone belt, Zimbabwe, a study
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

Archean mantle plumes: Evidence from greenstone belt geochemistry 355
of the evolution of Archean continental crust: Rotterdam, Balkema, Geo-
logical Society of Zimbabwe Special Publication 2, p. 167–174.
Condie, K. C., 1994, Greenstones through time, in Condie, K.C., ed., Archean
crustal evolution: Amsterdam, Elsevier, p. 85–120
Condie, K.C., 1997, Contrasting sources for upper and lower continental crust:
The greenstone connection: Journal of Geology, v. 105, p. 729–736.
Condie, K.C., 1998, Tracking mantle plumes in the Archean: Can it be done
with greenstone geochemistry?: GSA Abstracts with Programs, v. 30, n.
7, p. A-345.
Condie, K.C., 1999, Mafic crustal xenoliths and the origin of the lower conti-
nental crust: Lithos, v. 46, p. 95–101.
Corfu, F., Krough, T.E., Kowk, Y.Y., and Jensen, L.S. 1989, U-Pb zircon geo-
chronology in the southwestern Abitibi belt, Superior province: Canadian
Journal of Earth Sciences, v. 26, p. 1747-1763.
Davies, G.F., 1992, On the emergence of plate tectonics: Geology, v. 20, p. 963–
940.
Davies, G.F., 1993, Conjectures on the thermal and tectonic evolution of the
earth: Lithos, v. 30, p. 281–289.
Davis, W.J., 1997, U-Pb zircon and rutile ages from granulite xenoliths in the
Slave province: Evidence for mafic magmatism in the lower crust coin-
cident with Proterozoic dike swarms: Geology, v. 25, p. 343–346.
de Ronde, C.E.J., and de Wit, M.J., 1994, Tectonic history of the Barberton
greenstone belt, South Africa: 490 million years of Archean crustal evo-
lution: Tectonics, v. 13, p. 983–1005.
de Wit, M.J., Hart, R.A., and Hart, R.J., 1987, The Jamestown ophiolite com-
plex, Barberton mountain belt: A section through 3.5 Ga oceanic crust:
Journal of African Earth Sciences, v. 5, p. 681–730.
Desrochers, J-P., Hubert, C., Ludden, J.N., and Pilote, P., 1993, Accretion of
oceanic plateau fragments in the Abitibi greenstone belt, Canada: Geol-
ogy, v. 21, p. 451–454.
Dostal, J., and Mueller, W.U., 1997, Komatiite flooding of a rhyolitic arc com-
plex: Geochemical signature and tectonic significance of the Stoughton-
Roquemaure Group, Abitibi greenstone belt, Canada: Journal of Geology,
v. 105, p. 545–-563.
Floyd, P.A., Winchester, J.A., and Castillo, P.R., 1992, Geochemistry and pe-
trography of Cretaceous sills and lava flows, Sites 800 and 802, in Dear-
mont, L.H., et al., eds., Proceedings of the Ocean Drilling Program, Sci-
entific Results, Volume 129: College Station, Texas, Ocean Drilling
Program, p. 345–385.
Fryer, B.J., and Jenner, G.A., 1978, Geochemistry and origin of the Archean
Prince Albert Group volcanics, western Melville Peninsula, Northwest
Territories, Canada: Geochimica et Cosmochimica Acta, v. 42, p. 1645–
1654.
Galer, S.J.G., and Mezger, K., 1998, Metamorphism, denudation and sea level
in the Archean and cooling of the Earth: Precambrian Research, v. 92,
p. 389–412.
Goodwin, A.M., 1988, Geochemistry of Slave province volcanic rocks: Yel-
lowknife belt: Yellowknife, North West Territories Geology Division,
Contribution to the Geology of the Northwest Territories, v. 3, p. 13–25.
Hart, S.R., Hauri, E.H., Oschmann, L.A., and Whitehead, J. A., 1992, Mantle
plumes and entrainment: Isotopic evidence: Science, v. 90, p. 273–296.
Hawkesworth, C.J., Gallagher, K., Hergt, J.M., and McDermott, F., 1993, Man-
tle slab contributions in arc magmas: Annual Review of Earth and Plan-
etary Sciences, v. 21, p. 175–204.
Henderson, J.B., 1981, Archean basin evolution in the Slave province, Canada,
in Kroner, A., ed., Precambrian plate tectonics: Amsterdam, Elsevier,
p. 213–235.
Herzberg, C., 1995, Generation of plume magmas through time: An experi-
mental approach: Chemical Geology, v. 126, p. 1–16.
Hoffman, P.F., and Ranali, G., 1988, Archean oceanic flake tectonics: Geo-
physical Research Letters, v. 15, p. 1077–1080.
Hollings, P.N., 1998, Geochemistry of the Uchi subprovince, northern Superior
province: An evaluation of the geodynamic evolution of the northern
margin of the Superior province ocean basin [Ph.D. thesis:], Saskatche-
wan, Canada, University of Saskatchewan, 285 p.
Hollings, P.N., and Kerrich, R., 1999, Trace element systematics of ultramafic
and mafic volcanic rocks from the 3 Ga North Caribou greenstone belt,
northwestern Superior province: Precambrian Research, v. 93, p. 257–
279.
Huang, Y., Van Calsteren, P., and Hawkesworth, C. J., 1995, The evolution of
the lithosphere in southern Africa: A perspective on the basic granulite
xenoliths form kimberlites in South Africa: Geochimica et Cosmochimica
Acta, v. 59, p. 4905–4920.
Jahn, B., Gruau, G., and Glikson, A.Y., 1982, Komatiites of the Onverwacht
Group, South Africa: REE geochemistry, Sm/Nd age and mantle evolu-
tion: Contributions to Mineralogy and Petrology, v. 80, p. 25–40.
Jensen, L.S., and Pyke, D.R., 1982, Komatiites in the Ontario portion of the
Abitibi belt in Arndt, N.T., and Nisbet, E.G., eds., Komatiites: London,
Allen and Unwin, p. 147–158.
Jochum, K.P., Arndt, N.T., and Hofmann, A.W., 1991, Nb-Th-La in komatiites
and basalts: Constraints on komatiite petrogenesis and mantle evolution:
Earth and Planetary Science Letters, v. 107, p. 272–289.
Kamo, S.L., and Davis, D.W., 1994, Reassessment of Archean crustal devel-
opment in the Barberton Mountain Land, South Africa, based on U-Pb
dating: Tectonics, v. 13, n. 1, p. 167–192.
Kerr, A. C., Tarney, J., Nivia, A., Marriner, G. F., and Saunder, A. D., 1998,
The internal structure of oceanic plateaus: Inferences from obducted Cre-
taceous terranes in western Colombia and the Caribbean: Tectonophysics,
v. 292, p. 173–188.
Kerrich, R., Wyman, D., Fan, J., and Bleeker, W., 1998, Boninite series: Low
Ti-tholeiite associations from the 2.7 Ga Abitibi greenstone belt: Earth
and Planetary Science Letters, v. 164, p. 303–316.
Kerrich, R., Polat, A., Wyman, D.A., and Hollings, P., 1999, Trace element
systematics of Mg- to Fe-tholeiite suites of the 2.9-2.7 Ga Superior prov-
ince greenstone belts: implication for multi-component mantle plumes.
Lithos, v. 46, p. 163–187.
Kimura, G., Ludden, J.N., Desrochers, J.-P., and Hori, R., 1993, A model of
ocean-crust accretion for the Superior province, Canada: Lithos, v. 30,
p. 337–355.
Kjarsgaard, B.A., Kerswill, J.A., and Jenner, G.A., 1997, Lithostratigraphy and
metallogenic implications of komatiite–banded iron-formation—felsic
volcanic rocks of the Archean Woodburn Lake group, Pipedream Lake,
central Churchill province, Northwest Territories: Geological Survey of
Canada, Current Research 1997-C, p. 101–110.
Krapez, B., 1993, Sequence stratigraphy of the Archean supracrustal belts of
the Pilbara block, Western Australia: Precambrian Research, v. 60,
p. 1–45.
Lahaye, Y., Arndt, N., Byerly, G., Chauval, C., Fourcade, S., and Gruau, G.,
1995, The influence of alteration on the trace-element and Nd isotopic
compositions of komatiites: Chemical Geology, v. 126, p. 43–64.
Lesher, C.M., and Arndt, N.T., 1995, REE and Nd isotope geochemistry, pet-
rogenesis and volcanic evolution of contaminated komatiites at Kam-
balda, Western Australia: Lithos, v. 34, p. 127–157.
Mahoney, J.J., Storey, M., Duncan, R.A., Spencer, K.J., and Pringle, M., 1993,
Geochemistry and geochronology of Leg 130 basement lavas: Nature and
origin of the Ontong-Java plateau, in Berger, W.H., et al., eds., Proceed-
ings of the Ocean Drilling Program, Scientific Results, Volume 130: Col-
lege Station, Texas, Ocean Drilling Program, p. 3–22.
McCulloch, M.T., and Gamble, J.A., 1991, Geochemical and geodynamical
constraints on subduction zone magmatism: Earth and Planetary Science
Letters, v. 103, p. 358–374.
Moser, D.E., and Heaman, L.M., 1994, Paleoproterozoic metamorphism at the
base of the North American craton revealed through U-Pb zircon dating
of granulite-facies xenoliths, in Lanphere, M.A., et al., eds., International
Conference on Geochronology, Cosmochronology, and Isotope Geology,
8
th
, [abs.]: U.S. Geological Survey Circular 1107, p. 228.
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

K.Y. Tomlinson and K.C. Condie356
Nelson, D.R., Trendall, A.F., deLaeter, J.R., Grobler, N.J., and Fletcher, I.R.,
1992, A comparative study of the geochemical and isotopic systematics
of late Archean flood basalts from the Pilbara and Kaapvaal Cratons:
Precambrian Research, v. 54, p. 231–256.
Neymark, L.A., Paces, J.B., and Zartman, R.E., 1994, Whole rock and mineral
U-Th-Pb systematics in lower crustal xenoliths from the Baltic, Anabar
and Canadian Shields: Early Archean lithospheric differentiation and evo-
lution, in Lanphere, M.A., et al., eds., International Conference on Geo-
chronology, Cosmochronology, and Isotope Geology, 8
th
, Geology, [abs.]:
U.S. Geological Survey Circular 1107, p. 232.
Nisbet, E.G., Martin, A., Bickle, M.J., and Orpen, J.L., 1993a, The Ngezi
Group: Komatiites, basalts and stromatolites on continental crust, in Bic-
kle, M.J., and Nisbet, E.G., eds., The geology of the Belingwe greenstone
belt, Zimbabwe: A study of the evolution of Archean continental crust:
Rotterdam, Balkema, Geological Society of Zimbabwe Special Publica-
tion 2, p. 121–166.
Nisbet, E.G., Bickle, M.J., Orpen, J.L., and Martin, A., 1993b, Controls on the
formation of the Belingwe greenstone belt, Zimbabwe; in M.J. Bickle
and E.G. Nisbet, eds., The Geology of the Belingwe greenstone belt,
Zimbabwe: A study of the evolution of Archean continental crust: Rot-
terdam, Balkema, Geological Society of Zimbabwe Special Publication
2, p. 215–223.
Orpen, J.L., Martin, A., Bickle, M.J., and Nisbet, E.G., 1993, The Mtshingwe
Group in the west: Andesites, basalts, komatiites and sediments of the
Hokonui, Bend, and Koodoovale Formations, in Bickle, M.J., and Nisbet,
E.G., eds., The geology of the Belingwe greenstone belt, Zimbabwe: a
study of the evolution of Archean continental crust: Rotterdam Balkema,
Geological Society of Zimbabwe Special Publication 2, p. 69–86.
Perring, C.S., Barnes, S.J., and Hill, R.E.T. 1996, Geochemistry of komatiites
from Forrestania, Southern Cross province, Western Australia: Evidence
for crustal contamination: Lithos, v. 37, p. 181–197.
Polat, A., Kerrich, R., and Wyman, D.A., 1999, Geochemical diversity in oce-
anic komatiites and basalts from the late Archean Wawa greenstone belts,
Superior province, Canada: Trace element and Nd isotope evidence for a
heterogeneous mantle: Precambrian Research, v. 94, p. 139–173.
Puchtel, I.S., Zhuravlev, D.Z., Samsonov, A.V., and Arndt, N.T., 1993, Petrol-
ogy and geochemistry of metamorphosed komatiites and basalts from the
Tungurcha greenstone belt, Aldan Shield: Precambrian Research, v. 62,
p. 399–417.
Puchtel, I.S., Hofman, A.H., Mezger, K., Jochum, K.P., Shchipansky, A.A., and
Samsonov, A.V., 1998, Oceanic plateau model for continental crustal
growth in the Archean: A case study from the Kostomuksha greenstone
belt, NW Baltic Shield: Earth and Planetary Science Letters, v. 155,
p. 57–74.
Rajamani, V., 1990, Petrogenesis of metabasites from the schist belts of the
Dharwar craton: Implications to Archean mafic magmatism: Journal of
the Geological Society of India, v. 36, p. 565–587.
Rajamani, V., Shivkumar, K., Hanson, G.N., and Shirey, S.B., 1985, Geochem-
istry and petrogenesis of amphibolites, Kolar schist belt, south India: Evi-
dence for komatiite magma derived by low percentages of melting of the
mantle: Journal of Petrology, v. 26, p. 92–123.
Rosing, M.T., Rose, N.M., Bridgwater, D., and Thomsen, H.S., 1996, Earliest
part of Earth’s stratigraphic record: A reappraisal of the ⬎3.7 Ga Isua
(Greenland) supracrustal sequence: Geology, v. 24, p. 43–46.
Rudnick, R.L., 1995, Making continental crust: Nature, v. 378, p. 571–578.
Sage, R.P., 1998, Operation Ignace-Armstrong, Obonga-Lac Des Iles area: Ge-
ology of area II: Part 1. Supracrustal rocks: Ontario Geological Survey
Open File Report No. 5978, 218 p.
Schau, M., 1982, Geology of the Prince Albert Group in parts of Walker Lake
and Laughland Lake map areas, District of Keewatin: Geological Survey
of Canada, Bulletin 337, 62 p.
Seifert, K., Gibson, I., Weis, D., and Brunotte, D., 1996, Geochemistry of
metamorphosed cumulate gabbros from Hole 900A, Iberia abyssal plain,
in Sawyer, D.S., and Whitmarch, R.B., et al., eds., Proceedings of the
Ocean Drilling Program, Scientific results, volume 149: College Station,
Texas, Ocean Drilling Program, p. 471–488.
Silva, K., 1998, Komatiites from the Belingwe greenstone belt, Zimbabwe:
Constraints on the development of Archaean greenstone belts [Ph.D. the-
sis]: London, University of London, 208 p.
Skulski, T., and Percival, J.A., 1996, Allochthonous 2.78 Ga oceanic plateau
slivers in a 2.72 Ga continental arc sequence: Vizien greenstone belt,
northeastern Superior province, Canada: Lithos, v. 37, p. 163–179.
Sleep, N.H., and Windley, B.F., 1982, Archean plate tectonics: Constraints and
inferences: Journal of Geology, v. 90, p. 363–379.
Stone, W.E., Crocket, J.H., Dicken, A.P., and Fleet, M.E., 1995, Origin of
Archean ferropicrites: Geochemical constraints from the Boston Creek
flow, Abitibi greenstone belt, Ontario, Canada: Chemical Geology, v. 121,
p. 51–71.
Storey, M., Mahoney, J.J., Kroenke, L.W., and Saunders, A.B., 1991, Are oce-
anic plateaus sites of komatiite formation?: Geology, v. 19, p. 376–379.
Sun, S.S., and McDonough, W.F., 1989, Chemical and isotopic systematics of
oceanic basalts: Implications for mantle composition and processes, in
Saunders, A.S., and Norry, M.J., eds., Magmatism in the ocean basins:
Geological Society of [London] Special Publication 42, p. 313–345.
Sun, S.S., and Nesbitt, R.W., 1978, Petrogenesis of Archean ultrabasic and
basic volcanic rocks: Evidence from rare earth elements: Contributions
to Mineralogy and Petrology, v. 65, p. 301–325.
Sylvester, P.J., Campbell, I.H., and Bowyer, D.A., 1997, Niobium/uranium
evidence for early formation of the continental crust: Science, v. 275,
p. 521–523.
Tarney, J., Wood, D.A., Varet, J., Saunders, A.D., and Cann, J.R., 1979, Nature
of mantle heterogeneity in the North Atlantic: American Geophysical
Union Maurice Ewing Series, v. 2, p. 285–301.
Thorpe, R.I., Hickman, A.H., Davis, D.W., Mortensen, J.K., and Trendall, A.F.,
1992, U-Pb geochronology of the Archean felsic units in the Marble Bar
region, Pilbara craton, Western Australia: Precambrian Research, v. 56,
p. 169–189.
Thurston, P.C., 1990, Early Precambrian basic rocks of the Canadian Shield,
in Hall, R.P., and Hughes, D.J., eds., Early Precambrian basic magmatism:
Glasgow, Blackie, p. 221–247.
Thurston, P.C., and Chivers, K.M., 1990, Secular variation in greenstone se-
quence development emphasizing Superior Province, Canada: Precam-
brian Research, v. 46, p. 21–58.
Thurston, P.C., Cortis, A.L., and Chivers, K.M., 1987, A reconnaissance re-
evaluation of a number of northwestern greenstone belts: Evidence for
an early Archean sialic crust, in Summary of fieldwork and other activities
1987: Ontario Geological Survey Miscellaneous Paper 137, p. 4–24.
Thurston, P.C., Osmani, I.A. and Stone, D., 1991, Northwestern Superior
province: Review and terrane analysis, in Thurston, P.C., et al., eds., Ge-
ology of Ontario: Ontario Geological Survey Special Volume 4, Part 1,
p. 81–142.
Tomlinson, K.Y., 1996, The geochemistry and tectonic setting of early Precam-
brian greenstone belts, northern Ontario, Canada [Ph.D. thesis]: Ports-
mouth, UK, University of Portsmouth, 287, p.
Tomlinson, K.Y., Davis, D.W., Hughes, D.J., and Thurston, P.C., 1998a, The
central Wabigoon subprovince: Geochemistry, geochronology and tec-
tonic reconstruction, in Harrap, R.M., and Helmstaedt, H., eds., Western
Superior Transect 1998 Annual Meeting, Lithoprobe Report No. 65: Van-
couver, University of British Columbia, Lithoprobe Secretariat, p. 35–47.
Tomlinson, K.Y., Stevenson, R.K., Hughes, D.J., Hall, R.P., Thurston P.C., and
Henry, P., 1998b, The Red Lake greenstone belt, Superior province: Evi-
dence of plume-related magmatism at 3 Ga and evidence of an older
enriched source: Precambrian Research, v. 89, p. 59–76.
Tomlinson, K.Y., Hughes, D.J., Thurston, P.C., and Hall, R.P., 1999a, Plume
magmatism and crustal growth at 2.9 to 3.0 Ga in the Steep Rock and
Lumby Lake area, western Superior province: Lithos, v. 146, p. 103–136.
Tomlinson, K.Y., Davis, D.W., Thurston, P.C., Hughes, D.J., Percival, J.A., and
Sasseville, C., 1999b, Geochemistry, Nd isotopes and geochronology
on November 16, 2013specialpapers.gsapubs.orgDownloaded from

Archean mantle plumes: Evidence from greenstone belt geochemistry 357
Printed in the U.S.A.
from the central Wabigoon subprovince and North Caribou terrane: Re-
gional correlations leading towards a Mesoarchean reconstruction, in Har-
rap, R.M., and Helmstaedt, H., eds., Western Superior Transect 1999 An-
nual Meeting, Lithoprobe Report #70: Vancouver, University of British
Columbia, Lithoprobe Secretariat, p. 147–152.
Viljoen, M.J., Viljoen, R.P., and Pearton, T.N., 1982, The nature and distribu-
tion of Archean komatiite volcanics in South Africa, in Arndt, N.T., and
Nisbet, E.G., eds., Komatiites: London, Allen and Unwin, p. 81–90.
Weaver, B.L., 1991, The origin of ocean island basalt end-member composi-
tions: Trace element and isotopic constraints: Earth and Planetary Science
Letters, v. 104, p. 381–397.
Weaver, B.L., Tarney, J., Windley, B.F., and Leake, B.E., 1982, Geochemistry
and petrogenesis of Archean metavolcanic amphibolites from Fiskenaes-
set, S.W. Greenland: Geochimica et Cosmochimica Acta, v. 46, p. 2203–
2215.
White, R.S., 1997, Mantle plume origin for the Karoo and Ventersdorp flood
basalts, South Africa: South African Journal of Geology, v. 100, no. 4,
p. 271–282.
Wilks, M.E., and Nisbet, E.G., 1988, Stratigraphy of the Steep Rock Group,
northwestern Ontario: A major Archean unconformity and Archean stro-
matolites: Canadian Journal of Earth Sciences, v. 25, p. 370–391.
Xie, Q., 1996, Trace element systematics of mafic-ultramafic volcanic rocks
from the Archean Abitibi greenstone belt, Canada: Implications for chem-
ical evolution of the mantle and Archean greenstone belt development
[Ph.D. thesis]: Saskatchewan, Canada, University of Saskatchewan,
222 p.
Xie, Q., and Kerrich, R., 1994, Silicate-perovskite and majorite signature ko-
matiites from the Archean Abitibi greenstone belt: Implications for early
mantle differentiation and stratification: Journal of Geophysical Research,
v. 99, p. 15799–15812.
Xie, Q., Kerrich, R., and Fan, J., 1993, HFSE/REE fractionations recorded in
three komatiite-basalt sequences, Archean Abitibi greenstone belt: Im-
plications for multiple plume sources and depths: Geochimica et Cos-
mochimica Acta, v. 57, p. 4111–4118.
Zindler, A., and Hart, S., 1986, Chemical geodynamics: Annual Review of
Earth and Planetary Sciences, v. 14, p. 493–571.
M
ANUSCRIPT
A
CCEPTED BY THE
S
OCIETY
A
UGUST
16, 2000
on November 16, 2013specialpapers.gsapubs.orgDownloaded from