Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Widom, E.
Right arrow Articles by Schmincke, H.-U.
Right arrow Search for Related Content
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Journal of Petrology | Volume 40 | Number 2 | Pages 279-296 | 1999
© Oxford University Press 1999

Os Isotope Systematics in the Canary Islands and Madeira: Lithospheric Contamination and Mantle Plume Signatures

E. Widom1,*, K. A. Hoernle2, S. B. Shirey3 and H.-U. Schmincke2

1 Department of Geology, Miami University Oxford, OH 45056, USA
2 Department of Volcanology and Petrology, Geomar Wischhofstrasse 1-3, 24148 Kiel, Germany
3 Department of Terrestrial Magnetism Carnegie Institution of Washington, 5241 Broad Branch Road Nw, Washington, DC 20015, USA

Received April 18, 1997; Revised typescript accepted June 26, 1998


   ABSTRACT
TOP
 ABSTRACT
Introduction
 Geologic and Geochemical Setting
 Analytical Techniques
 Results
 Discussion
 Conclusions
 References
 
Osmium concentrations and isotopic signatures were measured in 28 primarily Holocene basalts (22 of which have been analyzed for Sr–Nd–Pb isotope composition), two carbonatites and two mantle xenoliths from the Canary Islands, Selvagen Grande and Madeira in the eastern North Atlantic. 187Os/188Os ratios in the basalts range from 0.129 to 0.183. The Os isotope systematics indicate that the basalts fall into three petrogenetic groups: (1) a ‘radiogenic’ group with high 187Os/188Os from 0.152 to 0.183; (2) an ‘unradiogenic’ group with low 187Os/188Os from 0.129 to 0.138; (3) an ‘intermediate’ group with 187Os/188Os between 0.139 and 0.151. The Os isotope systematics of the radiogenic group samples are consistent with minor contamination of the basalts by marine sediment. All samples in the unradiogenic group contain mantle xenoliths, and the unradiogenic Os can be explained by bulk assimilation of ≤ 5% mantle peridotite in the form of disaggregated xenoliths. The radiogenic and unradiogenic groups are also characterized by higher 87Sr/86Sr and 208Pb/204Pb but lower 143Nd/144Nd than samples with similar 206Pb/204Pb from the intermediate group, which is interpreted to reflect interaction of plume magmas with the lithospheric mantle. The intermediate group samples are believed to represent the isotopic signature of the mantle plume. The Os isotopic composition of the Canary plume is among the most radiogenic found in ocean island basalts, comparable with the endmember HIMU islands Mangaia and Tubuaii, but at significantly lower 206Pb/204Pb. The radiogenic Os and moderate 206Pb/204Pb signature of the Canary plume is consistent with a plume which contains 25–35% of relatively young (~1.2 Ga) recycled oceanic crust. Variable degree of mixing of the Canary Island plume source with shallow depleted asthenosphere containing a component of Paleozoic oceanic crust produces the limited range in Os isotopic signatures observed in the Madeira and Canary Island basalts despite a large range in 206Pb/204Pb isotopic composition.

KEY WORDS: Canary Islands; crustal contamination; lithospheric mantle; ocean island basalts; os sotopes


   Introduction
TOP
 ABSTRACT
 Introduction
Geologic and Geochemical Setting
 Analytical Techniques
 Results
 Discussion
 Conclusions
 References
 
The compositions of mantle sources are commonly inferred from the isotopic signatures of basalts erupted at the Earth's surface. To extrapolate from the composition of a basalt to that of the mantle source from which it was generated, the nature of interactions between the melt and overlying lithosphere must be understood in detail. However, it is often difficult to distinguish conclusively between enriched components in the plume and those incorporated by plume melts in the lithosphere. The Os isotope system is a powerful tracer for investigating both the processes which affect plume melts during ascent through the overlying lithosphere, as well as the nature of recycled components in mantle plume sources.

Distinguishing between contamination of basaltic melts in the crust vs the lithospheric mantle can be difficult, as mantle metasomatic processes can produce Sr, Nd and Pb isotopic signatures in the lithospheric mantle that are similar to those of crustal material In contrast, the Os isotope signatures of the lithospheric mantle and crust are distinct from one another. The lithospheric mantle generally preserves unradiogenic Os signatures caused by initial Re depletion, because of the relative insensitivity of the Os isotope system to overprinting by mantle metasomatic processes (Walker et al., 1989Go; Pearson et al., 1995bGo), whereas crustal material tends to be very radiogenic in Os because of the high Re/Os ratios generated during crust formation. Thus, very different Os isotope signatures might be imparted to plume melts depending on the nature of the lithospheric contaminant.

The very radiogenic nature of crustal material relative to residual mantle also allow the Os isotope system to be a useful tracer of recycling of crustal material back into the deep mantle. Because Os concentrations in the mantle are much greater than those in most crustal material, the Os isotope system is sensitive only to a fairly large percentag of recycled crustal material in a mantle source. In contrast, even minute amounts of crustal recycling, especially of sediments, can overwhelm the signatures of highly incompatible element isotope systems such as Sr, Nd and Pb. The Os isotope system therefore provides important evidence as to the percentage of recycled oceanic crust which might contribute to mantle plumes, provided that lithospheric contamination signatures can be eliminated.

Because Os is compatible during melting of the mantle, the Os concentrations of basaltic melts are typically extremely low, which enhances the utility of Os as a tracer of melt–lithosphere interactions. However, the ease with which basaltic melts may become contaminated during ascent through the lithosphere makes the interpretation of Os isotope systematics in ocean island basalts rather complex. To utilize the Os isotope system as a mantle tracer, it is necessary to filter out signatures superimposed on plume melts by the various high-level contamination processes. This requires detailed studies of individual ocean islands and island chains, where systematic variations in Os isotope signatures can be discerned and combined with other isotope systems such as Sr, Nd and Pb.

The location of the Canary Islands, Selvagen Islands and Madeira close to the African continent on old (>130 Ma) thick lithosphere with up to 10 km sediment cover makes these island chains particularly appropriate for an investigation of the effects of melt–lithosphere interaction on the Os isotope system. In this paper we report Os concentrations and isotope ratios for a suite of 28 primarily Holocene basalts from Madeira, Selvagen Grande and all of the Canary Islands (Hierro, La Palma, Gomera, Tenerife, Gran Canaria, Fuerteventura, Lanzarote), as well as two carbonatites from Fuerteventura and two peridotites from Lanzarote. This study concentrates on the Canary Islands but includes a few samples from Madeira and Selvagen Grande, which help constrain the processes occurring in the Canary Islands. In particular, Sr–Nd–Pb isotope systematics indicate that the Canary Islands and Madeira may share a common low 206Pb/204Pb mantle endmember (Hoernle et al., in preparation, a); we therefore consider these islands together in this study. We integrate the Os data with Sr–Nd–Pb isotope data on the same samples to evaluate the role of lithospheric contamination and in particular to distinguish between crustal and lithospheric mantle contamination, and to recognize recycled components in the mantle plume sources.


   Geologic and Geochemical Setting
TOP
 ABSTRACT
 Introduction
 Geologic and Geochemical Setting
Analytical Techniques
 Results
 Discussion
 Conclusions
 References
 
The Canary Islands comprise a chain of seven islands located off the northwest coast of Africa (Fig. 1). The origin of the island chain is thought to be hotspot related (Morgan, 1981Go), consistent with the geochemistry of the volcanic rocks and the east to west age progression in both the subaerial (from >20 to 1 Ma) and uplifted submarine complexes [25–70 to 4 Ma; see review by Hoernle & Schmincke (1993)Go]. Each of the Canary Islands has been active in the Holocene, except Gomera, where there has been no Quaternary activity; a 4 Ma sample is therefore taken to represent this island. Historic activity has occurred on four of the islands (Hierro, La Palma, Tenerife, Lanzarote), which span the entire Canary chain. Samples from Selvagen Grande and Madeira are believed to be <100 000 years in age.


Figure 01
View larger version (47K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. Map of the eastern North Atlantic Ocean basin between 25° and 45°N, showing the location of Canary Islands, Selvagen Islands and Madeira. Six of the seven Canary Islands have had Holocene volcanism and are represented in this study; one 4 Ma sample from Gomera is also included in this study, as are Late Pleistocene (<100 ka) samples from Selvagen Grande and Madeira.

 
Crustal ages beneath the Canary Islands are ~150–175 Ma based on the location of the M25 magnetic anomaly (156 Ma) between the westernmost islands of Hierro and La Palma and the S1 anomaly (~175 Ma) between the easternmost islands and Africa (Roeser, 1982Go; Klitgord & Schouten, 1986Go; Roest et al., 1992Go). The nature of the lithosphere beneath the Canary Islands (continental vs oceanic) has been the subject of a long-standing controversy. However, seismic refraction data (e.g. Bossard & Macfarlane, 1970Go; Banda et al., 1981Go) and the occurrence of mid-ocean ridge basalt (MORB)-like gabbros on Lanzarote, Gran Canaria and La Palma (Hoernle, 1998Go; Schmincke et al., 1999Go; and references therein) strongly argue for the presence of oceanic crust beneath all of the Canary Islands. The seismic refraction studies further indicate that a sediment cover of 5–10 km in thickness may underlie the easternmost islands.

Late Pleistocene and Holocene basalts from the Canary Islands are isotopically heterogeneous in Sr, Nd and Pb isotope ratios (Hoernle et al., in preparation, a; in preparation, b). Basanites and alkali basalts with SiO2 ≤ 46 wt % and MgO > 8 wt % occur on all studied islands and display binary mixing trends on Sr–Nd–Pb isotope correlation diagrams, which could be explained by mixing of two components: (1) the low seismic velocity component (LVC) of Hoernle et al. (1995)Go, a HIMU-type plume component (radiogenic 206Pb/204Pb; Zindler & Hart, 1986Go) found throughout the eastern Atlantic, western Mediterranean and Europe; (2) an asthenospheric component characterized by Holocene Madeira basalts, with HIMU trace element characteristics but MORB-like Sr–Nd–Pb isotopic ratios, interpreted as reflecting oceanic crust subducted in this region during the Paleozoic (Hoernle et al., in preparation, b). Evidence for the location of this component in the asthenosphere comes from studies of Jurassic ocean crust samples from Gran Canaria, some of which have HIMU-like trace element characteristics (Hoernle, 1998Go). Alkali basalts and tholeiites with SiO2 > 46 wt % occur only on the two easternmost Canary Islands of Lanzarote and Fuerteventura. These high-SiO2 basalts generally have higher 87Sr/86Sr and {Delta}8/4 [or 208Pb/204Pb for a given 206Pb/204Pb ratio relative to the Northern Hemisphere Reference Line (NHRL); Hart, 1984Go] but lower 143Nd/144Nd than low-SiO2 basalts from the same islands and even from the same eruption (e.g. 1730–1736 eruption on Lanzarote zoned from basanite to tholeiite), providing evidence for a third, enriched mantle (EMI; Zindler & Hart, 1986Go) component. The isotopic signatures of these basalts are attributed to contamination within the lithosphere.


   Analytical Techniques
TOP
 ABSTRACT
 Introduction
 Geologic and Geochemical Setting
 Analytical Techniques
Results
 Discussion
 Conclusions
 References
 
Holocene samples from the Canary Islands were collected with Os isotopic analysis in mind. Care was taken to limit exposure of the samples to metal, in order to minimize potential Re and Os processing blanks. Also, all sampled units were carefully inspected, over regions of more than 100 m in some cases, to identify the freshest portions of the volcanic unit and to avoid xenolithic material. The presence, type and abundance of xenoliths was carefully noted. Fresh sample chips of a size appropriate for the shatter box were carefully hand-picked by the authors to avoid any areas showing signs of alteration and to insure that the sample chips were free of any obvious xenolith debris. Sample chips were then ground to a flour in an agate shatter box. Most samples were digested by the Carius tube method; a few samples were dissolved by an acid digestion technique, which has been described in detail elsewhere (Walker, 1988Go; Shirey & Walker, 1995Go). In the Carius tube technique (Shirey & Walker, 1995Go), 2–3 g of sample powder are dissolved in concentrated aqua regia in a sealed glass tube by heating in an oven for 24–48 h at 200–250°C. This method is more effective than the Teflon vessel digestion technique for equilibrating spike and sample Os, and for extracting Os from sulfides, metals and spinels. Os concentrations measured in alkali basalts tend to be somewhat higher when samples are digested in Carius tubes compared with Teflon vessel digestions, but Os isotopic compositions in young samples are identical (Shirey & Walker, 1995Go). Total processing blanks are ~2 pg with 187Os/188Os of 0.17 ± 0.02. Procedures for Re separations using anion exchange columns with AG1-X8 resin have been described elsewhere (Pearson et al., 1995bGo). Re blanks ranged from 5 to 20 pg, and averaged 12 pg.

Re and Os isotope ratios were measured as the negative oxides ReO4– and OsO3– following existing techniques (Creaser et al., 1991Go). The procedures for measuring Re and Os at DTM have been described in detail by Pearson et al. (1995bGo). In-run precision was generally better than 0.5% for 187Re/185Re ratios and, with a few exceptions, between 0.16% and 1.5% for 187Os/188Os ratios. The small amount of sample Os and correspondingly small Os signal sizes produced in-run precisions for samples that were sometimes greater than the 0.2% 2{sigma} external reproducibility obtained on standards. Replicate measurements of Os concentration agreed within 2% to 29%. This level of reproducibility of Os concentration is not unexpected given the small sample sizes (mostly 2–3 g) and the very low Os concentrations. The Os is presumably hosted in trace phases and may therefore not be homogeneously distributed throughout the sample. Indeed, if we calculate the expected reproducibility of Os concentration measurements for 2 g samples, based on the ~5% worst-case reproducibility found in replicates of 50 g fusion samples (Hauri & Hart, 1993Go) and using the sampling error estimates of Kleeman (1967)Go, we would estimate worst-case reproducibility for Os concentration of ~100%. Such reproducibility is consistent with what we observe in the Canaries data set as well as a similar data set on samples from the Azores (Widom & Shirey, 1996Go). Replicate analyses of the Os isotope ratios all agree within 0.5–1.6% (Table 1).


View this table:
[in this window]
[in a new window]

  		Table 1: Sr–Nd–Pb–Os isotope and other relevant data for samples from the Canary Islands, Selvagen Grande and Madeira

 

   Results
TOP
 ABSTRACT
 Introduction
 Geologic and Geochemical Setting
 Analytical Techniques
 Results
Discussion
 Conclusions
 References
 
The basalts analyzed in this study range from tholeiite to alkali basalt to basanite. All samples are fairly mafic with MgO = 8–15 wt % and Ni = 113–437 ppm. Os concentrations range from 7 to 278 pg/g (Table 1 and Fig. 2), and display a positive correlation with both MgO and Ni content. Re concentrations in 13 basalts from six of the Canary Islands range from 41 to 705 pg/g, but do not vary systematically with MgO or Ni. Re/Os ratios, however, are negatively correlated with Os concentration, because of the relatively large variation in Os abundances.


Figure 02
View larger version (31K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2. 187Os/188Os vs Os concentration for basalts from the Canary Islands, Selvagen Islands and Madeira. The samples from this study are displayed as three groups, including the radiogenic group ({diamond}), the unradiogenic group ({nabla}) and the intermediate group (bullet, Canary Islands; {blacksquare}, Madeira). The radiogenic group includes samples which have <52 pg/g and have Os isotope signatures more radiogenic than any at higher Os concentrations. The unradiogenic group contains the samples with the least radiogenic Os and includes all samples that come from xenolith-bearing basalt flows. These samples also contain xenocrystic olivine from mantle peridotite (based on their forsterite and CaO contents; D. Graham, unpublished data, 1998). The intermediate group, which comprises the remaining samples, has a relatively limited range in 187Os/188Os from 0.139 to 0.151 over a range in Os concentration from 12 to 232 pg/g. Also included are literature data for 0–1.6 Ma basalts from La Palma (Marcantonio et al., 1995Go; gray circles), which extend the range in Os concentration of intermediate group samples to nearly 400 pg/g. The intermediate group is interpreted to represent the Os isotopic composition of the Canary Island plume, whereas the radiogenic and unradiogenic group samples are thought to be contaminated in the crust and lithospheric mantle, respectively.

 
187Os/188Os isotope ratios in basalts from the Canary Islands, Selvagen Grande and Madeira span a large range from 0.129 to 0.183 (Table 1). The Os isotope ratios do not seem to vary systematically between islands, but rather large variations occur within most of the islands. An extreme case of the large intra-island variations in Os isotope signature is exemplified by the island of Lanzarote, which exhibits the full range of Os isotope ratios found in the Canary archipelago as a whole.

The Os isotope signatures in the Canary Islands, Selvagen Islands and Madeira exhibit a complex relationship with Os concentration (Fig. 2). Samples with low Os concentration exhibit a large range in 187Os/188Os and display the most radiogenic values, a relationship that has been observed in other cases including St Helena (Reisberg et al., 1993Go), the Azores archipelago (Widom & Shirey, 1996Go) and La Palma (Marcantonio et al., 1995Go). For high Os concentration samples, however, the relationship between 187Os/188Os and Os concentration in the Canary Islands differs from that of St Helena and the Azores. In the latter islands, the high Os concentration samples yield a relatively narrow range in 187Os/188Os despite large ranges in Os concentration, whereas the high concentration samples in the Canary Islands are variable in 187Os/188Os, and extend to relatively unradiogenic Os isotope signatures (Fig. 2).

In detail, the Os isotope systematics of the Canary Islands, Selvagen Islands and Madeira samples allow a simple division into three sample groups. One group, the ‘radiogenic’ group, includes the nine most radiogenic samples (187Os/188Os = 0.152–0.183), all of which have Os concentrations <52 pg/g and 187Os/188Os moreradiogenic than any samples with high Os abundances (>52 pg/g Os). The second group, or ‘unradiogenic’ group, consists of the six basalt samples which come from mantle xenolith-bearing lava flows and contain xenocrystic olivine from mantle peridotite (D. Graham, unpublished data, 1998). These basalts have the least radiogenic Os isotope ratios, ranging in 187Os/188Os from 0.129 to 0.138. The third group, or ‘intermediate group’ consists of the remaining 13 samples, which have a relatively limited range in 187Os/188Os from 0.139 to 0.151 despite an almost 20-fold variation in Os concentration from 12 to 232 pg/g (Fig. 2). This is approximately the same range of 187Os/188Os as found previously in 0.8–1.6 Ma basalts from the island of La Palma (0.136–0.150; Marcantonio et al., 1995Go). In the following discussion we consider the implications of these variable Os isotopic signatures regarding the composition of the Canary Island plume and potential melt–lithosphere interactions.


   Discussion
TOP
 ABSTRACT
 Introduction
 Geologic and Geochemical Setting
 Analytical Techniques
 Results
 Discussion
Conclusions
 References
 
Os isotopic composition of the Canary Island plume
Detailed Os isotope studies of basalts from individual ocean islands have generally shown that low Os concentration basalts (<30–50 ppt) have highly variable Os isotope signatures extending to very radiogenic values, whereas higher Os abundance basalts display limited variability and have the least radiogenic Os signatures (Reisberg et al., 1993Go; Martin et al., 1994Go; Marcantonio et al., 1995Go; Widom & Shirey, 1996Go; Widom, 1997Go). In these cases, the high Os concentration basalts have been interpreted as representing the Os isotopic signature of the respective mantle plume, and the radiogenic Os isotope signatures of the low Os concentration basalts have been attributed to crustal contamination of plume-derived basalts (Reisberg et al., 1993Go; Martin et al., 1994Go; Marcantonio et al., 1995Go; Widom & Shirey, 1996Go; Widom, 1997Go). In the case of the Canary Islands basalts, where high Os concentration samples (>52 ppt) exhibit relatively variable Os isotopic signatures characteristic of both the intermediate and unradiogenic groups, determining the Os isotopic composition of the plume is more complex.

Conventional wisdom based on incompatible element isotope ratios would lead to the conclusion that the unradiogenic, mantle xenolith-bearing group have suffered the least shallow-level contamination and therefore are the most likely to represent the isotopic composition of the mantle plume from which they were derived. However, this is not necessarily an appropriate inference for a compatible element such as Os. In the case of Os, the relative mass-balance dictates that basalts might easily become contaminated during ascent through the overlying lithospheric mantle, which has an Os concentration one to three orders of magnitude higher than most basaltic magmas. This, however, applies only to bulk assimilation of lithospheric mantle material, as melts of the lithosphere induced by the passing plume melts will have relatively low Os concentrations. It is thus not immediately clear whether the unradiogenic group or intermediate group samples should best represent the Os isotopic composition of the Canary Island mantle plume. The implications of both possibilities are considered below.

If it is assumed that the unradiogenic group basalts represent the Os isotopic composition of the Canary Island plume, then the intermediate group samples must be attributed to crustal contamination of initially less radiogenic magmas. Whatever the precise mechanism of contamination, the process must be able to explain the fact that the intermediate group samples from this study plus high Os abundance samples from La Palma (Marcantonio et al., 1995Go) display a nearly 40-fold range in Os concentration from 12 to 390 ppt, yet range in 187Os/188Os from only 0.139 to 0.151 (Fig. 2).

To produce such a limited range in Os isotopic composition by crustal contamination of magmas with initial Os isotopic compositions similar to those in the unradiogenic group would require a substantial variability in the percentages of assimilated crust, with the low Os concentration basalts always affected by less crustal contamination than the high Os concentration samples. For example, if the least radiogenic sample in the unradiogenic group (EL10 with 187Os/188Os = 0.1288) is considered to represent the mantle plume composition, and all of the intermediate group samples are derived by contamination of plume melts with a given radiogenic crust or sediment (we assume bulk assimilation of sediment with 187Os/188Os = 1.2, Os = 0.1 pg/g for the purpose of this example), then sample ELP1 (12 pg/g Os), sample EH11 (114 pg/g Os), and the 390 pg/g sample from La Palma (Marcantonio et al., 1995Go), must have assimilated 2%, 14% and 35% sediment, respectively. In this scenario, the amount of sediment assimilated would vary among these samples by nearly 20-fold, yet each mixture must result in essentially the same 187Os/188Os signature. We consider this to be an unlikely scenario.

An alternative explanation to consider for the narrow range in 187Os/188Os over a large range in Os concentration in the intermediate group samples is the accumulation of variable amounts of Os-rich olivine into initially low Os concentration melts. However, this explanation suffers from several problems. First, there is no clear correlation between the modal olivine content of the basalts and the Os concentrations among the intermediate group samples. Second, it cannot easily explain the limited variability of 187Os/188Os, because there is no reason that all accumulated olivine should have the same Os isotopic ratio. Because the Os isotopic signature of the accumulated olivine would have to have been acquired by crystallization of crustally contaminated magmas to produce the elevated Os isotopic signatures (relative to the unradiogenic group samples), this scenario would require that all of the olivine crystallized from contaminated magmas with the same Os isotopic signature. This is clearly an unlikely scenario given the fact that low-Os, crustally contaminated magmas normally display extremely large ranges in Os isotopic signature (Fig. 2 and Reisberg et al., 1993Go; Martin et al., 1994Go; Marcantonio et al., 1995Go; Widom & Shirey, 1996Go; Widom, 1997Go).

Given the above difficulties in explaining the narrow range in 187Os/188Os over a large range in Os concentration in the intermediate group samples by crustal contamination of melts initially with Os isotopic signatures characteristic of the unradiogenic group samples, we consider it more likely that the intermediate group samples are the ones that most closely represent the Os isotopic composition of the Canary Islands mantle plume. For clarity, this sample group will be referred to henceforth as the ‘plume group’.

Contamination of magmas by the lithospheric mantle
Os isotope evidence
If the unradiogenic group basalts are ultimately derived from plume melts with more radiogenic Os signatures, then these basalts must have interacted in some fashion with a reservoir that is relatively unradiogenic in Os. The lithospheric mantle is the only reservoir with which the Canary Island plume melts might have come in contact that is known to have an appropriate unradiogenic Os isotopic signature. Abyssal peridotites range in 187Os/188Os from 0.122 to 0.127 (Luck & Allègre, 1992Go; Snow & Reisberg, 1995aGo, 1995bGo), and data from 11 Canary Island mantle xenoliths show a similar range (this study and E. Widom & E. R. Neumann, unpublished data, 1998). Because the Os concentration of mantle peridotite is typically 1–3 orders of magnitude higher than basaltic melts, minor amounts of bulk assimilation of lithospheric mantle could greatly alter the Os isotope signatures of ascending basaltic magmas.

We emphasize that all samples within the unradiogenic Os group came from lava flows containing mantle xenoliths. Although all samples were carefully taken and then hand-picked to avoid obvious xenolith debris, microprobe analyses of the olivine crystals in the unradiogenic group samples showed that the olivines were xenocrysts from mantle peridotite, on the basis of their high forsterite (>86) and low CaO (<0.15 wt %) contents (D. Graham, unpublished data). Oxygen isotope analyses of olivines from these samples yielded values within the range found in mantle peridotite xenoliths from the Canary Islands but distinct from most Canary Island olivine phenocrysts (Wiechert et al., 1997Go; in preparation), providing additional evidence for a lithospheric mantle origin of these olivines. Figure 3 shows that bulk assimilation of ≤ 5% average oceanic lithospheric mantle by typical plume group basalts with 50–75 pg/g Os can account for the relatively low 187Os/188Os ratios in most of the unradiogenic group samples. Because the incorporation of bulk mantle significantly increases the Os concentration in the basaltic melt, a melt with a low initial Os concentration is required to explain the unradiogenic sample with 21 pg/g Os by this mechanism. If the melt initially contained about 10 pg/g Os, bulk assimilation of 0.3% mantle would produce an appropriately decreased Os isotopic signature and increase the Os concentration to only 20 pg/g. Alternatively, it is possible that a higher Os concentration melt assimilated 1–2% lithospheric mantle, but underwent a subsequent decrease in Os concentration by fractional crystallization before eruption.


Figure 03
View larger version (42K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3. 187Os/188Os vs Os for plume group and unradiogenic group basalts from the Canary Islands and Madeira. Mixing curves illustrate the effect of bulk assimilation of oceanic lithospheric mantle by typical plume group basalts with 187Os/188Os = 0.146 and Os concentrations of 10, 50 and 75 pg/g. The mantle endmember has Os = 3.1 ng/g and 187Os/188Os = 0.125, the average isotopic composition of the two mantle xenoliths from Lanzarote (this study). This is also the average Os isotopic signature for abyssal peridotites (Snow & Reisberg, 1995aGo, 1995bGo). Assimilation of ≤5% lithospheric mantle in the form of disaggregated mantle xenoliths can explain the relatively unradiogenic Os isotope signatures in the mantle xenolith-bearing basalts. The low Os concentration xenolith-bearing sample may have been produced by contamination of a low Os concentration basalt or by fractional crystallization (fc) of an Os-bearing phase subsequent to contamination. Symbols as in Fig. 2.

 
The model of bulk assimilation depicted in Fig. 3 would apply equally to assimilation by bulk melting of lithospheric mantle or entrainment of mantle xenolith fragments. Bulk melting of the lithosphere is unlikely, on the basis of thermal constraints. On the other hand, the modal contents of olivine in the thin sections are consistent with up to 5% bulk incorporation of xenocrysts from mantle peridotite. Assimilation of Os-rich phases of mantle xenoliths (sulfides or olivine with sulfide inclusions) by plume melts can strongly influence the Os isotopic composition of the magma. In conclusion, despite careful hand-picking to remove as much xenolith debris as possible, the unradiogenic Os group samples reflect the presence of small amounts (≤5 wt %) of disaggregated xenolith material in these samples. These findings demonstrate the danger of bulk analyses of olivine-bearing basalts, in particular those from flows that contain mantle xenoliths. We therefore emphasize the need to carefully inspect sampled units for xenolithic material in the field, and the importance of determining whether crystals are phenocrystic or xenocrystic prior to Os isotope studies of basalts.

Effect on Pb, Nd and Sr isotope signatures
The Sr, Nd and Pb isotope data for the Canary Islands and Madeira basalts are shown in Fig. 4. It should be noted that the plume group samples, defined on the basis of their Os isotopic signatures, produce well-defined trends in Sr–Nd–Pb isotope space which are consistent with two-component mixing involving the LVC (low velocity composition—a HIMU-like plume component reflecting the composition of the low seismic velocity sub-lithospheric mantle; Hoernle et al., 1995Go) and a relatively low 206Pb/204Pb component characterized by the Madeira samples. The origins of these mixing components are discussed in subsequent sections.


Figure 04
View larger version (25K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 4. 206Pb/204Pb vs (a) 207Pb/204Pb, (b) 208Pb/204Pb, (c) 143Nd/144Nd and (d) 87Sr/86Sr for Canary Island and Madeira basalts (Hoernle et al., in preparation, a and b). The intermediate (or plume) group basalts fall along two-component mixing trends consistent with mixing between the HIMU-like LVC source (Hoernle et al., 1995Go) and a source with MORB-like Sr–Nd–Pb isotopic signatures but HIMU-like trace element composition, which is interpreted to contain a recycled Paleozoic oceanic crustal component (Hoernle et al., in preparation, b). The unradiogenic and radiogenic Os group samples have similar 207Pb/204Pb but extend to higher 87Sr/86Sr and {Delta}8/4 and lower 143Nd/144Nd at a given 206Pb/204Pb compared with the plume group samples. Mantle xenoliths from Lanzarote (Zartman & Tera, 1973Go; Vance et al., 1989Go; Siena et al., 1991Go) denoted by shaded box in (a) and (b) and shaded bar in (c) and (d) show that the lithosphere contains enriched mantle (EM) more extreme in Sr–Nd–Pb isotopic composition than the unradiogenic and radiogenic Os group basalts. Symbols as in Fig. 2 (refer to the caption to Fig. 5 for an explanation of the asterisks).

 
An additional observation from Fig. 4 is that the majority of the unradiogenic and radiogenic Os group samples fall outside the two-component mixing trends defined by the plume group samples. These outlying samples provide evidence for a third mixing component with an EM-type signature characterized by elevated 87Sr/86Sr, low 143Nd/144Nd and slightly higher {Delta}8/4 isotope signatures relative to the plume group samples. The fact that these Sr–;Nd–pb signatures are restricted to samples from the radiogenic and unradiogenic Os groups, which have been contaminated in the crust (see below) and lithospheric mantle, respectively, provides strong evidence that the EM signature is derived from the lithosphere and not a heterogeneous mantle plume source. If the EM component were part of the mantle plume source, some plume group samples should exhibit the anomalous Sr–Nd–Pb characteristics. In contrast, some of the samples recording Os isotopic contamination in the lithosphere have maintained plume-like Sr–Nd–Pb signatures, which indicates that the Os isotope system is generally more sensitive than Sr, Nd, and Pb isotopes to the operating lithospheric contamination processes.

Given the Os isotopic evidence for the unradiogenic group samples having experienced contamination by the lithospheric mantle, it is important to evaluate the possibility that the enriched Sr, Nd and Pb isotopic compositions relative to the plume group were also inherited from the lithospheric mantle. This interpretation is supported by previous studies of volcanic rocks from the eastern islands of Fuerteventura and Gran Canaria that, on the basis of experimental phase equilibria arguments, attributed enriched Sr–Nd–Pb isotopic signatures in alkali basalts and tholeiites compared with basanites and nephelinites to contamination of plume-derived melts within the lithospheric mantle (Hoernle & Tilton, 1991Go; Hoernle et al., 1991Go). In particular, the Holocene volcanic rocks may share the same EMI-type component as Pliocene volcanic rocks from Gran Canaria (87Sr/86Sr > 0.7033, 143Nd/144Nd < 0.51275, 206Pb/204Pb < 18.8, {Delta}8/4 ≥ 60, {Delta}7/4 ≤ 0). This component is observed not only in relatively evolved samples from Gran Canaria (Hoernle et al., 1991Go), but also in several mafic basanites (mg-number = 66–68) with oxygen isotope compositions within the range found in mantle xenoliths from the Canary Islands (Wiechert et al., in preparation), which further supports a mantle origin for this component (Hoernle, 1998Go).

Although only few data are available at present, the existence of enriched mantle lithosphere beneath the eastern Canary Islands is supported by studies of mantle xenoliths from Lanzarote. Spinel harzburgite and dunite xenoliths from Lanzarote display enrichments in many incompatible trace elements and have 87Sr/86Sr = 0.7030–0.7036 and 143Nd/144Nd = 0.51286–0.51295 (Vance et al., 1989Go; Siena et al., 1991Go). In particular, a dunite xenolith from the same eruptive center (Pico de Partido) as sample EL5, which belongs to the unradiogenic Os group, has 87Sr/86Sr = 0.7033 and 143Nd/144Nd = 0.51286. Only one harzburgite from Lanzarote has been analyzed for Pb isotopes and has 206Pb/204Pb = 18.9, 207Pb/204Pb = 15.56 and 208Pb/204Pb = 38.9 (Zartman & Tera, 1973Go). Despite the paucity of data from mantle xenoliths, the range in Sr and Nd isotope ratios in the xenoliths from Lanzarote overlap and extend to more enriched compositions than those in the basalts from Lanzarote including the Holocene unradiogenic Os group samples (Fig. 4).

Even though the mantle xenoliths from the eastern Canary Islands could qualitatively serve as the enriched (EMI-type) endmember for the radiogenic and unradiogenic Os groups, excessively large amounts of bulk lithosphere assimilation would be required to produce the observed Sr, Nd and Pb isotope variations, because of the relatively low abundances of these elements in the lithospheric mantle. We envisage two possible solutions to this problem. First, studies of mantle xenoliths from the eastern islands attribute the enrichment to relatively recent infiltration of the lithosphere by Fe–Ti-rich silicate melts and CO2-rich fluids associated with plume activity (Vance et al., 1989Go; Siena et al., 1991Go; Neumann et al., 1995Go). Carbonate-rich veins (Frezzotti et al., 1994Go) and pockets of silicic glass attributed to mantle–melt reactions at 8–12 kbar (Neumann & Wulff-Pedersen, 1997Go) have been reported in spinel harzburgite and lherzolite xenoliths from the Canary Islands. Assimilation of small amounts of partial melts of such enriched material (e.g. concentrated in veins or melt pockets) could have a major effect on the Sr–Nd–Pb isotopic composition of the plume melts. Metasomatism by plume melts, however, has not significantly affected the Os isotopic composition of the Canary oceanic lithospheric mantle (E. Widom & E. R. Neumann, unpublished data, 1998), and in any case assimilation of small degree melts would not be expected to significantly affect a compatible element isotope system such as Os. Therefore, assimilation of melts of the metasomatic veins or pockets could produce the observed enriched Sr, Nd and Pb isotope signatures without inducing radiogenic or unradiogenic Os isotopic signatures. Both the mineral and major element chemistry (Neumann et al., 1995Go), and the 187Os/188Os ratios measured in Canary mantle xenoliths (this study and E. Widom & Neumann, unpublished data), which are within the range of Os isotope ratios found in abyssal peridotites, are consistent with an oceanic origin for at least some of the enriched lithospheric material beneath the eastern islands.

Alternatively, more strongly enriched mantle may exist beneath the Canary Islands than that observed so far in the few analyzed xenoliths. Recently it has been shown that blocks of continental crust stranded during continental break-up occur in the Atlantic ocean basin as far as 250 km west of the Iberian Peninsula (Whitmarsh & Party, 1998Go). Assimilation of such enriched subcontinental lithospheric mantle, if it occurs beneath the Canary Islands, could therefore also produce the enriched Sr, Nd and Pb isotopic signatures observed in the unradiogenic group Os samples. Furthermore, as observed in samples of the Canaries oceanic lithospheric mantle, ancient subcontinental lithospheric mantle is typically characterized by unradiogenic Os isotope signatures despite extensive metasomatism resulting in enriched Sr, Nd and Pb isotope signatures (Walker et al., 1989Go; Pearson et al., 1995aGo, 1995bGo). If sufficiently enriched mantle exists beneath the Canary Islands, then smaller amounts of bulk assimilation might explain the observed EMI signatures in the unradiogenic Os group samples. For example, it has been proposed that blocks or rafts of subcontinental lithospheric mantle, delaminated or detached during the rifting of Pangaea, may have been incorporated into the oceanic mantle and could provide a source for some EM-type oceanic basalts (Hawkesworth et al., 1984Go). EMI signatures significantly more extreme than those measured in the Canaries mantle xenoliths have been found in MORB near the Oceanographer Transform of the Mid-Atlantic Ridge (35°N; Shirey et al., 1987Go), in basalts from Tristan da Cunha in the South Atlantic (Le Roex et al., 1990Go), and in Kerguelen Plateau basalts in the southern Indian Ocean (Mahoney et al., 1995Go); these EMI signatures have all been attributed to detached fragments of old subcontinental lithospheric mantle, by Shirey et al. (1987)Go, Hawkesworth et al. (1984)Go and Mahoney et al. (1995)Go, respectively. Furthermore, unradiogenic Os isotope signatures in peridotite xenoliths from the Kerguelen Islands provide strong evidence for the existence of subcontinental lithospheric mantle beneath the Kerguelen Plateau (Hassler & Shimizu, 1998Go). Fragments of enriched subcontinental lithospheric mantle might similarly have been incorporated within the lithospheric mantle beneath the eastern Canary Islands during the initial stages of continental rifting and seafloor spreading (Hoernle et al., 1991Go).

It is clear that additional studies of mantle xenoliths are needed to constrain the composition of the shallow mantle beneath the Canary Islands and to test between the aforementioned hypotheses. Nevertheless, regardless of the precise origin of the Canary Island lithospheric mantle, it clearly has appropriate isotopic signatures to qualitatively explain the unradiogenic Os and Nd, radiogenic Sr, and high {Delta}8/4 in the xenolith-bearing samples (Fig. 4).

Crustal contamination
The highly variable and often radiogenic Os isotope signatures found in the low Os concentration basalts are indicative of crustal contamination. There are a number of possible crustal sources containing radiogenic Os with which magmas might interact during ascent, including altered Jurassic oceanic crust, sediment, and the basalts and differentiates of the volcanic edifices through which the basalts are erupted. However, distinguishing between these contaminants is not straightforward using only Os isotope data. Nevertheless, we can place additional constraints on possible crustal components affecting the Os isotopic composition with Sr–Nd–Pb isotope data.

The old oceanic crust beneath the Canary Islands might be considered a likely source of contamination of the Canary Island basalts. The oceanic crust beneath the Canary Islands is dated at ~150–175 Ma and is therefore expected to have very radiogenic 187Os/188Os ratios. Assuming that the oceanic crust has an average 187Re/188Os of 50 (corresponding to average MORB basalts with Os concentrations of about 50 pg/g), the 187Os/188Os of this crust would be ~0.257. Although ~20% assimilation of such crust could have produced the observed radiogenic Os isotope signatures in some Canary Island basalts, it cannot explain the decrease in 143Nd/144Nd isotope ratio observed in most radiogenic group samples. The relatively high 143Nd/144Nd values of the altered oceanic crust beneath Gran Canaria (Hoernle, 1998Go) relative to most of the radiogenic Os group samples preclude the Jurassic oceanic crust as a significant contaminant (Fig. 5a). This is consistent with the observation that the contaminated basalts from the Canary Islands are no more radiogenic in Os than crustally contaminated basalts from the Azores (Widom & Shirey, 1996Go) and St Helena (Reisberg et al., 1993Go), which lie on very young oceanic crust. These observations indicate that oceanic crust is in general not the primary contaminant affecting the Os isotope signatures of ocean island basalts (Reisberg et al., 1993Go; Widom, 1997Go).


Figure 05
View larger version (21K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 5. Plots of 187Os/188Os vs (a) 143Nd/144Nd, (b) 87Sr/86Sr and (c) {Delta}8/4, the difference between sample 208Pb/204Pb and Northern Hemisphere Reference Line (NHRL) for a given 206Pb/204Pb (Hart, 1984Go), and (d) {Delta}7/4 (the difference between sample 207Pb/204Pb and NHRL for a given 206Pb/204Pb). Mixing curves represent contamination of plume-group basalts by assimilation of average local Jurassic altered oceanic crust (AOC; Hoernle, 1998Go), average local terrigenous sediment (TS; Hoernle et al., 1991Go; Hoernle, 1998Go), average carbonate-rich sediment (CS; Hoernle et al., 1991Go; Hoernle, 1998Go), and manganese oxide (MnOx; Ben Othman et al., 1989Go; Pegram et al., 1992Go; Ravizza, 1993Go; Ravizza & Esser, 1993Go), with 187Os/188Os assumed to be equal to >70 Ma seawater (Peucker-Ehrenbrink et al., 1995Go). Compositions of the mixing endmembers are as follows. Plume basalt: Os =0.05 ng/g, 187Os/188Os = 0.146, Nd = 50 µg/g, 143Nd/144Nd = 0.51296, Sr = 750 µg/g, 87Sr/86Sr = 0.703025, Pb = 2.5 µg/g, {Delta}8/4 = 8.93, {Delta}7/4 = –2.3; AOC: Os = 0.100 ng/g, 187Os/188Os = 0.257, Nd = 7.3 µg/g, 143Nd/144Nd = 0.51298; CS: Os = 0.010 ng/g, 187Os/188Os = 0.931, Sr = 1190 µg/g, 87Sr/86Sr = 0.7093; TS: Os = 0.01, 0.025 or 0.10 ng/g (denoted as TS.01, etc.), 187Os/188Os = 1.2, Nd = 24 µg/g, 143Nd/144Nd = 0.51200, Sr = 179 µg/g, 87Sr/86Sr = 0.7147, Pb = 16.9 µg/g, {Delta}8/4 = 61.62, {Delta}7/4 = 16.16; MnOx: Os = 1.0 ng/g, 187Os/188Os = 0.722, Pb = 655 µg/g, {Delta}8/4 = 46, {Delta}7/4 = 11. The starting plume basalt composition was chosen to represent that of an average eastern-group island magma with 206Pb/204Pb = 19.3, and appropriate respective 207Pb/204Pb, 208Pb/204Pb, 143Nd/144Nd and 87Sr/86Sr compositions given the plume group mixing trends in Fig. 4. It should be noted that the starting plume composition (marked by asterisks in Fig. 4) is such that the amount of assimilated sediment required to produce the EMI signatures in the radiogenic Os group samples from the eastern-group islands (light gray shaded fields) is minimized. The AOC field represents the range of values found in Gran Canaria Jurassic oceanic crust (Hoernle, 1998Go), and assuming an age of 160 Ma and 187Re/188Os = 50. Compositions of local sediment from Hoernle et al. (1991Go). Tick marks indicate percentage of assimilated material, except for AFC (assimilation–fractional crystallization) curves where F denotes the amount of liquid remaining. AFC curve (c) calculated with DOs = 20, DSr,Pb = 0.01, and r (ratio of mass assimilated to mass crystallized) = 0.1 (DePaolo, 1981Go). In (a), (b) and (d) the AFC curves for sediment with 10 pg/g Os are essentially indistinguishable from the bulk mixing curves for 100 pg/g Os terrigenous sediment. Symbols as in Fig. 2.

 
The Canary Islands are underlain by a thick sediment pile that also provides a potential source of radiogenic Os. The two easternmost islands in particular may be underlain by as much as 10 km of sediment, which includes limestones and terrigenous shales, siltstones and sandstones (Stillman et al., 1975Go). The effects of sediment contamination can be modeled using Sr, Nd, Pb isotope data for local Miocene to Cretaceous sediment from DSDP site 397, which lies ~100 km SE of Gran Canaria, and sediment xenoliths in Gran Canaria mafic volcanic rocks (Hoernle et al., 1991Go; Hoernle, 1998Go), and assuming Os isotope signatures characteristic of terrigenous and carbonate-rich sediments (Esser & Turekian, 1993Go; Peucker-Ehrenbrink et al., 1995Go).

The Pb isotope data place major constraints on the amount of possible sediment assimilation. Although the radiogenic Os group samples generally have slightly higher {Delta}8/4 than the plume group, there is almost no difference in the {Delta}7/4 between the groups (Figs 4a and b and 5c and d). In fact, all Holocene samples analyzed to date (>40) fall on or below the Northern Hemisphere Reference Line (i.e. have {Delta}7/4 ≤ 0) on the 207Pb/204Pb vs 206Pb/204Pb diagram (Hoernle et al., in preparation, a). Assuming assimilation of bulk sediment from the Canary region [Pb = 12.2 ppm, 206Pb/204Pb = 18.97, 207Pb/204Pb = 15.71, 208Pb/204Pb = 39.16; n = 7, samples range from carbonate rich to silicate rich (Hoernle et al., 1991Go; Hoernle, 1998Go)], and including the analytical error on the Pb isotope ratios based on replicate analyses, a maximum of 2% sediment assimilation is allowed. Although the variation in Os isotope ratio between the radiogenic Os and plume groups can be explained by <2% sediment assimilation with either the involvement of AFC processes or assimilation of a relatively high Os concentration sediment (100 pg/g, Fig. 5), it is clear that neither the Os vs Nd nor Os vs Sr isotope data trends can be explained by such small amounts of sediment assimilation (Fig. 5a and b). Significant shifts to low 143Nd/144Nd can be produced only by assimilation of relatively large amounts (≥ 15%) of sediment with very low Os concentration (10 pg/g; Fig. 5a). Although such low Os concentrations are within the observed range in Os concentration in sediments (Esser & Turekian, 1993Go), the large amounts of sediment assimilation are clearly unrealistic based on the Pb isotope data. We note that the 8% sediment assimilation proposed by Thirlwall et al. (1997)Go for a Mio cene tholeiite with low incompatible element abundances resulted in an increase in {Delta}7/4 of + 10, three times the range observed in our Holocene data set (Fig. 5d). The Sr–Os isotope relationships indicate that any sediment contaminant must be largely terrigenous; the high Sr concentration of the more carbonate-rich sediment from DSDP site 397 would produce a shift to radiogenic Sr that is far higher than observed before causing any significant shift to radiogenic Os (Fig. 5b). AFC processes (for a given Os concentration in the sediment) would produce significantly more vertical mixing curves on these diagrams, because of the compatible behavior of Os. Therefore, it will be difficult to produce significant shifts in Sr and Nd isotope ratios and {Delta}8/4 by AFC without inducing Os isotope signatures much more radiogenic than observed. An example of the effect of AFC is shown in Fig. 5c using DOs = 20 (Hart & Ravizza, 1996Go), DSr,Pb = 0.01, and r = 0.1. In Fig. 5a, b and d, the AFC curves for sediment with 10 pg/g Os are essentially indistinguishable from the bulk mixing curves for 100 pg/g Os terrigenous sediment.

Rocks making up the volcanic edifice through which magmas ascend are also possible sources of contamination. The volcanic edifices, however, are substantially younger than the oceanic crust beneath the Canary Islands and thus would be expected to be significantly less radiogenic than the Jurassic oceanic crust. Also, these rocks are likely to have similar Sr–Nd–Pb isotopic compositions to the Holocene plume magmas (Hoernle & Tilton, 1991Go). The oldest rocks in the Canary Islands, from the basal complex of Fuerteventura, may be as old as 30–80 Ma (Le Bas et al., 1986Go). Following the calculations used above for Jurassic oceanic crust, the 187Os/188Os signature of the oldest edifice basalts would be ~0.19. This is only slightly more radiogenic than some of the radiogenic Os group samples at 0.183 (e.g. samples EF12 and EL17), and thus 85% assimilation of even the oldest basaltic edifice material would be required to produce the radiogenic signatures in these Holocene Canary Island basalts. Although this cannot be categorically ruled out, such large amounts of bulk assimilation are rather unlikely, on the basis of thermal arguments. It is possible that Mn-oxide coatings on subaqueous basalts in the volcanic pile could contribute significant radiogenic Os because of their radiogenic Os signatures and very high Os concentrations (~1 ng/g; Luck & Turekian, 1983Go; Palmer & Turekian, 1986Go; Esser & Turekian, 1988Go; Palmer et al., 1988Go; Reisberg et al., 1993Go). However, a minimum of 0.3–0.4% Mn-oxide assimilation is required to produce the most radiogenic Os isotopic signatures found in the Canary Island basalts. This is clearly inconsistent with the measured MnO contents in these Canary volcanic rocks, which range from 0.14 to 0.22. Furthermore, there is no correlation between 187Os/188Os and MnO, as would be expected if Mn-oxide assimilation were the major source of the increase in 187Os/188Os. All samples within the radiogenic Os group have MnO < 0.19. Also, because of the very high Pb contents found in Mn-oxide nodules (Ben Othman et al., 1989Go), even such a minor amount of assimilation would be expected to have a significant effect on the Pb isotopic composition of the contaminated basalt, producing both {Delta}8/4 and {Delta}7/4 values that are far higher than observed in the Canary Island radiogenic Os group basalts (Fig. 5d).

Evolved differentiates in the volcanic pile are also unlikely contaminants, given their expected very low Os concentrations and the evidence that at some point during differentiation Re/Os ratios cease to increase, and perhaps even start to decrease (Walker et al., 1991Go; Johnson et al., 1995Go). Another possible contaminant in the volcanic edifice is carbonatite, which is exposed in the basal complex of Fuerteventura (Hoernle & Tilton, 1991Go). To address this possibility, we analyzed two carbonatites from the >20 Ma basal complex of Fuerteventura. Although both have extremely radiogenic Os (187Os/188Os = 0.614 and 0.455), the Os concentrations are very low (14.5 ppt and 4.7 ppt, respectively), severely limiting their effectiveness as contaminants. Mixing calculations demonstrate that as much as 20% carbonatite assimilation would be required to produce the range in radiogenic Os signatures in the Canary basalts. Such a large amount of carbonatite contamination is clearly prohibited, on the basis of major element (e.g. CaO) and trace element (e.g. Sr and the LREE) considerations (Hoernle & Tilton, 1991Go; unpublished data, 1991).

Despite clear evidence that oceanic crust underlies all of the Canary Islands (Hoernle, 1998Go; Schmincke et al., 1999Go, and references therein), it cannot be ruled out that blocks of continental crust may underlie at least the eastern Canary Islands, because of their proximity to the African continent. The enriched Sr, Nd and {Delta}8/4 but low {Delta}7/4 isotope signatures of the radiogenic Os group samples are inconsistent with assimilation of upper continental crust or sediment, as discussed above. However, early Proterozoic or Archean lower continental crust could have an appropriate Sr–Nd–Pb isotopic composition to serve as the contaminant (Hoernle et al., 1991Go) and could also cause the observed increases in 187Os/188Os (Esperanca et al., 1997Go). In particular, lower continental crust is typically characterized by significantly lower {Delta}7/4 than upper continental crust or sediment (Rudnick & Goldstein, 1990Go), and therefore substantially more crustal assimilation could be accommodated to explain the shifts to high 87Sr/86Sr and {Delta}8/4 and low 143Nd/144Nd, without concomitant shifts to positive {Delta}7/4. Although we find assimilation of lower continental crust to be an interesting possibility, to date there has been no direct evidence for the presence of continental material beneath the Canary Islands, e.g. its presence as xenoliths or its necessity to explain geophysical observations in the region.

The above discussion of the effects of crustal contamination on the Canary Island basalts indicates that assimilation of local altered Jurassic oceanic crust or local Canaries sediment could easily explain the observed radiogenic Os isotopic signatures; however, it is not possible to explain the observed shifts to enriched Sr, Nd and {Delta}8/4 signatures by this mechanism without producing either (1) significant shifts to positive {Delta}7/4, which are not observed, or (2) Os isotopic signatures that are much more radiogenic than observed. Although assimilation of lower continental crust could possibly explain the observed variations in Sr–Nd–Pb–Os isotope ratios, no such material has been found in the Canary Islands to date. These problems, and the fact that the radiogenic group samples display a similar range in 87Sr/86Sr, 143Nd–144Nd and {delta}8/4 to the mantle xenolith-bearing unradiogenic Os group samples, lead to the conclusion that the radiogenic Os group samples inherited their enriched Sr, Nd and Pb isotope signatures in the lithospheric mantle prior to crustal contamination, which affected only the Os isotope signatures. It is noted that the limited mantle xenolith data from the eastern Canary Islands are qualitatively consistent with the shifts in {Delta}8/4, Sr, and Nd isotope ratios and the absence of a shift to positive {Delta}7/4 observed in the volcanic rocks (Fig. 4). The existence of these xenoliths illustrates that such material exists within the shallow mantle beneath the Canaries. Additional support for existence of enriched mantle beneath at least the eastern Canary Islands comes from the Pliocene volcanic rocks on Gran Canaria, as was discussed above.

In summary, the elevated Sr, {Delta}8/4 and low Nd isotopic signatures are a common feature of both the radiogenic and unradiogenic Os group samples (Fig. 4), and are interpreted to reflect contamination within the shallow mantle. The differences in Os isotope signatures between these groups, however, provide evidence for two distinct contamination processes. The unradiogenic Os group samples reflect the presence of small amounts (≤ 4 wt %) of disaggregated mantle xenoliths in these samples, whereas the low concentration, radiogenic Os group samples are attributable to interaction of melts with a radiogenic crustal contaminant during ascent. Contamination of magmas with crust or sediment containing significantly more than 10 pg/g Os could produce substantial increases in Os isotopic signature without large changes in the Sr, Nd and Pb isotopic signatures. For example, <2% bulk assimilation of terrigenous sediment with 100 pg/g Os could produce the full range of observed radiogenic Os isotope signatures with only minor changes (largely within the analytical errors) in Sr, Nd and Pb (Fig. 5). If AFC processes are important, it is predicted that isotope ratios of elements such as Os with D >> 1 will show significant effects with much smaller percentages of assimilation, and with even less effect on the incompatible element isotope ratios. The radiogenic Os group samples are therefore believed to have attained their radiogenic Os isotope signatures by crustal contamination, but this contamination was not the primary mechanism that caused the enriched Sr, Nd and Pb isotope signatures. The latter were more probably imparted by contamination within shallow mantle sources, which did not affect the Os isotopic composition. The Os isotopes therefore record processes of both crustal and lithospheric mantle contamination, whereas only the lithospheric mantle contamination can be discerned in the Sr–Nd–Pb isotope systematics.

Os isotope systematics in the Canary Island plume
The plume group samples cover nearly the entire range in Pb isotopic composition (206Pb/204Pb = 18.8–19.9) found in the Canary Islands and Madeira, with samples from Gomera and Tenerife having the most radiogenic Pb and samples from Madeira the least radiogenic Pb; the Sr–Nd–Pb isotopic composition of the Madeira basalts is consistent with their being the unradiogenic Pb endmember for the Canary Island basalts (Hoernle et al., in preparation, a). On the 207Pb/204Pb–206Pb/204Pb diagram (Fig. 4a), the plume group is parallel to but generally falls below the NHRL; whereas on the 208Pb/204Pb–206Pb–204Pb diagram (Fig. 4b), the data are again parallel to but overlap or fall slightly above the NHRL. Plume group 143Nd–144Nd and 87Sr/86Sr show negative and slightly positive correlations, respectively, with 206Pb/204Pb. On all diagrams, the LVC plume component could serve as the endmember with radiogenic Pb and Sr and unradiogenic Nd (Hoernle et al., 1995Go). The Madeira endmember has an Sr–Nd–Pb isotopic composition similar to average Atlantic N-MORB, yet as mentioned previously these samples have HIMU-like trace element compositions and therefore cannot come from a normal DMM source. Instead, it has been proposed that these melts come from subducted Paleozoic ocean crust, which at present resides in the asthenosphere (Hoernle et al., in preparation, b).

The plume group basalts from the Canary Islands and Madeira display a limited range in 187Os/188Os between 0.139 and 0.151. These values are significantly more radiogenic than those found in the depleted upper mantle (Snow & Reisberg, 1995aGo, 1995bGo; Meisel et al., 1996Go) and are in fact among the most radiogenic values found in uncontaminated plume basalts, comparable with those of the endmember HIMU islands Mangaia, Tubuaii and Rurutu. Although the Canary Islands and the endmember HIMU basalts share Os isotope characteristics as well as some trace element characteristics such as high Ce/Pb ratios (Hoernle & Schmincke, 1993Go; Marcantonio et al., 1995Go), the Canary Islands have distinctly less radiogenic Pb isotopic signatures than do the endmember HIMU islands.

Origin of radiogenic plume Os
The radiogenic Pb isotopic signatures of the endmember HIMU plumes are commonly attributed to a component of recycled oceanic crust in the plume source (Chase, 1981Go; Hofmann & White, 1982Go), and explanations for the radiogenic Os isotope signatures in these plumes have followed suit (Pegram & Allègre, 1992Go; Hauri & Hart, 1993Go; Reisberg et al., 1993Go; Roy-Barman & Allègre, 1995Go). Indeed, the radiogenic Os and Pb isotopic signatures of these islands can be reproduced by incorporation of 20–25% of 1.8–2 Ga recycled oceanic crust with µ = 20 into FOZO-like enriched plume mantle (EPM; Fig. 6) with 187Os/188Os~ 0.130–0.1325 and 206Pb/204Pb ~19.1. The endmember HIMU plume Os and Pb isotope signatures could alternatively be produced by recycling of slightly greater percentages of oceanic crust into depleted MORB mantle (DMM). However, even low 206Pb/204Pb plumes have significantly more radiogenic 187Os/188Os than DMM, and thus there is no evidence for the involvement of pure DMM in mantle plumes (Widom & Shirey, 1996Go; Widom, 1997Go; Shirey & Walker, 1998Go). The comparable radiogenic Os isotopic compositions of the Canary Island plume and the endmember HIMU islands suggests that the former may also be attributed to a component of recycled oceanic crust, despite the less radiogenic 206Pb/204Pb signature (Marcantonio et al., 1995Go).


Figure 06
View larger version (37K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 6. (a) 187Os/188Os vs 206Pb/204Pb, illustrating the composition of the Canary Islands and Madeira plume group samples (filled symbols as in Fig. 2) as well as the endmember HIMU islands Mangaia (M; {diamondsuit}) and Tubuaii (T; {square}). Mixing curves depict mixing of FOZO-like enriched plume mantle (EPM) with 1.8 and 2 Ga recycled oceanic crust (ROC). The compositions of the mixing endmembers are as follows. EPM: 206Pb/204Pb = 19.1 (Hart et al., 1992Go), 187Os/188Os = 0.1325 (Widom & Shirey, 1996Go), Pb = 30 ng/g, Os = 3.3 ng/g; ROC: µ = 20, Pb = 250 ng/g, Os = 0.100 ng/g, and 187Os/188Os = 1.61 and 1.78 (assuming average 187Re/188Os = 50 and ages of 1.8 Ga and 2 Ga, respectively). It should be noted that EPM might have a range in 206Pb/204Pb anywhere from ~18.5 to 19.7 (Widom & Shirey, 1996Go; Shirey & Walker, 1998Go); we have picked an intermediate value for this model. The endmember HIMU islands are consistent with a plume containing 20–25% 1.8–2 Ga ROC with µ = 20. The Canary Islands/LVC plume can be explained by 25–35% 1.2 Ga ROC with µ = 20. Data for Mangaia and Tubuaii from Hauri & Hart (1993)Go. (b) 187Os/188Os vs 206Pb/204Pb for the Canary Islands and Island plume group basalts. The 187Os/188Os signature has a limited range from 0.139 to 0.151. The relatively limited range in Os cannot be explained by simple source mixing between the Canary Island plume/LVC and Paleozoic recycled oceanic crust (PROC), which would have very radiogenic Os despite its young age. Mixing between the Canary Island plume/LVC and DMM is also inconsistent with the Os isotope data, as this would produce a strong trend towards much less radiogenic Os isotope signatures than are observed. A possible explanation for the observed data trend is that the Canary Island plume/LVC is mixing with a shallow asthenospheric source which is itself a hybrid with ~50% DMM and ~50% Paleozoic recycled oceanic crust. The parameters used in the mixing models are as follows. Canary Islands/LVC plume (assumed to be a mixture of mantle with 30% 1.2 Ga recycled oceanic crust): Os = 2.325 ng/g, 187Os/188Os = 0.146, Pb = 96 ng/g, 206Pb/204Pb = 20; DMM: Os = 3.3 ng/g, 187Os/188Os = 0.125, Pb = 30 ng/g, 206Pb/204Pb = 18.0; PROC: Os = 0.100 ng/g, 187Os/188Os = 0.62, Pb = 250 ng/g, 206Pb/204Pb = 18.6; 50%–50% mix of PROC and DMM: Os = 1.7 ng/g, 187Os/188Os = 0.137, Pb = 140 ng/g, 206Pb/204Pb = 18.45.

 
The lower 206Pb/204Pb of the Canaries plume relative to endmember HIMU plumes could be explained by a 1.8–2 Ga recycled component with a lower µ of ~16, or with a comparable µ value if the age of the recycled crust is significantly younger (~1.2 Ga). Variable µ values between individual recycled slabs might arise from compositional variability in the original fresh oceanic crust, variability in extent of alteration of the oceanic crust, and/or variability in the extent of subduction zone modification. With time, these compositional differences would produce distinct isotopic signatures in the recycled slabs (Marcantonio et al., 1995Go). Another possible explanation for variable Pb isotope signatures in plumes with similar Os isotope signatures is that the Pb isotope signatures of plumes might be controlled by variable amounts of sediment subducted with the oceanic crust (Roy-Barman & Allègre, 1995Go). This model could explain the Os and Pb isotopic composition of the Canary Island plume (or LVC mantle source) if ~0.1% ancient pelagic sediment is incorporated with 1.8–2 Ga subducted oceanic crust. However, the observation that negative {Delta}7/4 OIB such as the Canary Islands has Ce/Pb and Nd/Pb higher than endmember HIMU OIB (Thirlwall, 1997Go) argues against this process; addition of sediment to an endmember HIMU source should, if anything, produce lower Ce/Pb and Nd/Pb ratios.

The most satisfactory model to explain the origin of mantle plumes such as the Canary plume, which have moderate 206Pb/204Pb, negative {Delta}7/4 and HIMU-like trace element signatures (e.g. high Ce/Pb), is that of recycling of relatively young oceanic crust (Thirlwall, 1997Go). Figure 6a illustrates that a mantle plume containing ~25–35% of 1.2 Ga recycled oceanic crust could produce the 187Os/188Os–206Pb/204Pb isotopic signature of the Canary Island/LVC plume. These estimates are consistent with estimates of 18–35% recycled ocean crust in the Canary Island plume based on oxygen isotope data (Wiechert et al., 1997Go). On the basis of this model, the similarity of the Os isotopic signature of the Canary Island/LVC plume and endmember HIMU OIB is a function of a slightly higher percentage in the plume of recycled oceanic crust with a lower 187Os/188Os signature. The relative difference in the amount of recycled oceanic crust between LVC and endmember HIMU plumes (25–35% vs 20–30%, respectively), however, is small compared with the difference in the ages of the respective recycled slabs (1.2 Ga vs 1.8–2 Ga); this implies that as long as plume sources contain similar percentages of recycled oceanic crust, the effect on the Os isotopic signature will be relatively insensitive to the age of the slab. In contrast, the strongly hyperbolic nature of the Os–Pb mixing curves is such that the 206Pb/204Pb signature of a plume is much more sensitive to the age of the recycled oceanic crust than to the amount of crust if in excess of 10–20%. More speculative is the possibility that subduction zone processes have changed as a result of cooling of the mantle with time, from slab melting in the Archean to slab dewatering in the post-Archean (Vidal, 1992Go; Thirlwall, 1997Go). Slab melting may produce a subducted slab residue with a low Re/Os because of the incompatible behavior of Re and compatible behavior of Os during melting, whereas slab dewatering may produce a higher Re/Os residue because of mobility of Os in aqueous slab fluids (Brandon et al., 1996Go). These processes may have operated to provide a higher Re/Os in younger subducted slabs, thus minimizing the differences in present-day 187Os/188O of variable age slab components of HIMU plumes.

Implications for mixing processes
The plume group basalts from the Canary Islands and Madeira display large variations in Pb isotope signatures which correlate with Sr and Nd isotopes (Fig. 4), and can be attributed predominantly to two-component mixing between the HIMU-like LVC component and a less enriched component with lower 206Pb/204Pb. Mixing between the LVC plume and DMM, which could produce this negative trend, has been eliminated as a possibility because of the HIMU-like trace element systematics in the isotopically depleted Madeira basalts (Hoernle et al., in preparation, b). Furthermore, it can be seen in Fig. 6b that plume–DMM source mixing would also be inconsistent with the relatively limited variation in Os isotope ratios in the Canary Island and Madeira basalts; this would result in Os isotope signatures significantly less radiogenic than those observed, because of the high Os/Pb of DMM relative to the LVC plume component, assuming 30% low Os/Pb recycled oceanic crust in LVC. Simple source mixing of the LVC plume and Paleozoic recycled oceanic crust is also inadequate to explain the Canary Islands–Madeira 187Os/188Os vs 206Pb/204Pb isotopic systematics, because the high Re/Os and thus very radiogenic 187Os/188Os expected even in young subducted oceanic crust would cause a relatively steep negative trend in Fig. 6b, unlike the observed data. A possible explanation for the relatively limited variation in Os isotope ratios, but with a slightly negative 187Os/188Os–206Pb/204Pb trend, is mixing between the LVC plume and shallow upper mantle which is hybridized DMM and Paleozoic oceanic crust. Approximately 50% DMM–50% Paleozoic oceanic crust would produce a source with both a similar Pb/Os ratio and similar but slightly less radiogenic Os compared with the LVC plume, such that mixing between these sources would produce basalts with a relatively linear, slightly negative trend of 187Os/188Os–206Pb/204Pb. The precise parameters of such mixing would be better constrained if 187Os/188Os data for high Os concentration Madeira samples were available; the low Os concentration of the Madeira samples provides only a maximum 187Os/188Os value for the low 206Pb/204Pb mixing component. However, the general model holds, as the high Os concentration Canary plume group samples alone provide evidence for a mixing process in which 206Pb/204Pb varies with no clear increase or decrease in 187Os/188Os.


   Conclusions
TOP
 ABSTRACT
 Introduction
 Geologic and Geochemical Setting
 Analytical Techniques
 Results
 Discussion
 Conclusions
References
 
The Sr–Nd–Pb–Os isotope systematics in basalts from the Canary Islands and Madeira provide evidence for both the origin of the Canary Island plume sources as well as the complex interactions of plume melts with the overlying lithosphere. The Os isotope system preserves a record of both crustal and lithospheric mantle contamination in some Canary Island basalts. Radiogenic Os isotope signatures in low Os concentration basalts are attributed to assimilation of sediment by plume melts during ascent. Less than 2% assimilation of terrigenous sediment can produce the observed range of radiogenic Os isotopic signatures without significantly affecting the Sr, Nd, or Pb isotopic signatures, but minor contribution of radiogenic Os because of assimilation of Mn-oxide, carbonate sediment and altered oceanic crust is permissible. Relatively unradiogenic Os isotope signatures in the xenolith-bearing unradiogenic group basalts indicate bulk assimilation by plume melts of lithospheric mantle in the form of disaggregated mantle xenoliths. Correspondingly high Sr and {Delta}8/4 and low Nd isotopic signatures in these samples indicate that assimilation of melts of the metasomatized lithospheric mantle has also occurred.

The Os isotope signature of the Canary Island plume (or LVC) is among the most radiogenic found in ocean islands to date, and is comparable with the signatures in the endmember HIMU islands of Mangaia, Tubuaii and Rurutu. The LVC plume, however, has distinctly less radiogenic Pb isotope signatures. The radiogenic Os isotopic signature and concomitant moderate Pb isotopic signature of the LVC plume is most consistent with a relatively young (1.2 Ga) recycled oceanic crustal component. Mixing between the Canary plume and depleted asthenospheric mantle which contains a component of young (Paleozoic) recycled oceanic crust explains the relatively limited range in Os isotope ratios of the plume group basalts despite a range in Sr, Nd and Pb isotopic signatures.


   Acknowledgements
 
This work was made possible by a Carnegie Fellowship, NRC Fellowship and Miami University College of Arts and Sciences Small Research Grant to E.W.; NSF Grants EAR-9005412 and EAR-9204718 to S.B.S. and Richard Carlson, and EAR-9105113 to K.A.H., Jim Gill and Ken Collerson; and a Leibniz Award to H.U.S. Thanks go to Dave Kuentz at DTM for assistance in the chemistry laboratories and with the figures, to John Emler at DTM for assistance in the mass spectrometry laboratory, to Kerstin Wolf at GEOMAR for help with sample preparation, and to Dave Graham and Thor Hansteen for help with sample collection.

* Corresponding author. Fax: 513-529-1542. e-mail: widome{at}muohio.edu


   References
TOP
 ABSTRACT
 Introduction
 Geologic and Geochemical Setting
 Analytical Techniques
 Results
 Discussion
 Conclusions
 References
 
Banda E., Danobeitia J. J., Surinach E., Ansorge J. Features of crustal structure under the Canary Islands. Earth and Planetary Science Letters (1981) 55:11–24.[Web of Science]

Ben Othman D., White W. M., Patchett J. The geochemistry of marine sediments, island arc magma genesis, and crust–mantle recycling. Earth and Planetary Science Letters (1989) 94:1–21.[Web of Science]

Bossard H., Macfarlane D. J. Crustal structure of the western Canary Islands from seismic refraction and gravity data. Journal of Geophysical Research (1970) 75:4901–4918.

Brandon A. D., Creaser R. A., Shirey S. B., Carlson R. W. Osmium recycling in subduction zones. Science (1996) 272:861–864.[Abstract]

Chase C. Oceanic island Pb: two–stage histories and mantle evolution. Earth and Planetary Science Letters (1981) 52:277–284.[Web of Science]

Creaser R. A., Papanastassiou D. A., Wasserburg G. J. Negative thermal ion mass spectrometry of osmium, rhenium, and iridium. Geochimica et Cosmochimica Acta (1991) 55:397–401.[Web of Science]

DePaolo D. J. Trace element and isotopic effects of combined wallrock assimilation and fractional crystallization. Earth and Planetary Science Letters (1981) 53:189–202.[Web of Science]

Esperanca S., Carlson R. W., Smith D. Dating crust–mantle separation: Re–Os isotopic study of mafic xenoliths from central Arizona. Geology (1997) 25:651–654.[Abstract/Free Full Text]

Esser B. K., Turekian K. K. Accretion rate of extraterrestrial particles determined from osmium isotope systematics of Pacific pelagic clay and manganese nodules. Geochimica et Cosmochimica Acta (1988) 52:1383–1388.[Web of Science]

Esser B. K., Turekian K. K. The osmium isotopic composition of the continental crust. Geochimica et Cosmochimica Acta (1993) 57:3093–3104.[Web of Science]

Frezzotti M. L., Touret J. L. R., Neumann E. R. Melt and fluid inclusions in dunite xenoliths from La Gomera, Canary Islands: tracking the mantle metasomatic fluids. European Journal of Mineralogy (1994) 6:805–817.[Abstract/Free Full Text]

Hart S. A large–scale isotope anomaly in the Southern Hemisphere mantle. Nature (1984) 309:753–757.

Hart S. R., Ravizza G. E. Os partitioning between phases in lherzolite and basalt. In: Earth Processes: Reading the Isotopic Code—Basu A., Hart S., eds. (1996) Washington, DC: American Geophysical Union. 123–134.

Hart S. R., Hauri E. H., Oschmann L. A., Whitehead J. A. Mantle plumes and entrainment: isotopic evidence. Science (1992) 256:517–520.[Abstract/Free Full Text]

Hassler D. R., Shimizu N. Osmium isotopic evidence for ancient subcontinental lithospheric mantle beneath the Kerguelen Islands, southern Indian Ocean. Science (1998) 280:418–421.[Abstract/Free Full Text]

Hauri E. H., Hart S. R. Re–Os isotope systematics of HIMU and EMII oceanic island basalts from the South Pacific Ocean. Earth and Planetary Science Letters (1993) 114:353–371.[Web of Science]

Hawkesworth C. J., Rogers N. W., van Calsteren P. W. C., Menzies M. A. Mantle enrichment processes. Nature (1984) 311:331–335.

Hoernle K., Schmincke H.-U. The role of partial melting in the 15-Ma geochemical evolution of Gran Canaria: a blob model for the Canary hotspot. Journal of Petrology (1993) 34:599–626.[Abstract/Free Full Text]

Hoernle K., Tilton G., Schmincke H. Sr–Nd–Pb isotopic evolution of Gran Canaria: evidence for shallow enriched mantle beneath the Canary Islands. Earth and Planetary Science Letters (1991) 106:44–63.[Web of Science]

Hoernle K., Zhang Y.-S., Graham D. Seismic and geochemical evidence for large–scale mantle upwelling beneath the eastern Atlantic and western and central Europe. Nature (1995) 374:34–39.

Hoernle K. A. Geochemistry of Jurassic oceanic crust beneath Gran Canaria (Canary Islands): implications for crustal recycling and assimilation. Journal of Petrology (1998) 39:859–880.[Abstract/Free Full Text]

Hoernle K. A., Tilton G. R. Sr–Nd–Pb isotope data for Fuerteventura (Canary Islands) basal complex and subaerial volcanics: applications to magma genesis and evolution. Schweizerische Mineralogische Petrographische Mitteilungen (1991) 71:3–18.

Hofmann A. W., White W. M. Mantle plumes from ancient oceanic crust. Earth and Planetary Science Letters (1982) 57:421–436.[Web of Science]

Johnson C. M., Shirey S. B., Barovich K. M. New approaches to crustal evolution studies and the origin of granitic rocks; what can the Lu–Hf and Re–Os isotope systems tell us? In: The Third Hutton Symposium on the Origin of Granites and Related Rocks—Brown M., Candela P. A., Peck D. L., Stephens W. E., Walker R. J., Zen E.-A., eds. (1995) College Park, MD: Geological Society of America; Edinburgh: Royal Society of Edinburgh. 339–352.

Kleeman A. Sampling error in the chemical analysis of rocks. Journal of the Geological Society of Australia (1967) 14:43–47.

Klitgord K. D., Schouten H. Plate kinematics of the Central North Atlantic. In: The Geology of North America: the Western North Atlantic Region—Vogt P. R., Tucholke B. E., eds. (1986) Boulder, CO: Geological Society of America. 351–378.

Le Bas M. J., Rex D. C., Stillman C. J. The early magmatic chronology of Fuerteventura, Canary Islands. Geological Magazine (1986) 123:287–298.[Abstract]

Le Roex A. P., Cliff R. A., Adair B. J. I. Tristan da Cunha, South Atlantic: geochemistry and petrogenesis of a basanite–phonolite lava series. Journal of Petrology (1990) 31:779–812.[Free Full Text]

Luck J.-M., Allègre C.-J. Osmium isotopes in ophiolites. Earth and Planetary Science Letters (1992) 107:406–415.[Web of Science]

Luck J.-M., Turekian K. K. Osmium–187/Osmium–186 in manganese nodules and the Cretaceous–Tertiary boundary. Science (1983) 222:613–615.[Abstract/Free Full Text]

Mahoney J. J., Jones W. B., Frey F. A., Salters V. J. M., Pyle D. G., Davies H. L. Geochemical characteristics of lavas from Broken Ridge, the Naturaliste Plateau and southernmost Kerguelen Plateau: Cretaceous plateau volcanism in the southeast Indian Ocean. Chemical Geology (1995) 120:315–345.[Web of Science]

Marcantonio F., Zindler A., Elliott T., Staudigel H. Os isotope systematics of La Palma, Canary Islands: evidence for recycled crust in the mantle source of HIMU ocean islands. Earth and Planetary Science Letters (1995) 133:397–410.[Web of Science]

Martin C. E., Carlson R. W., Shirey S. B., Frey F. A., Chen C.-Y. Os isotopic variation in basalts from Haleakala Volcano, Maui, Hawaii: a record of magmatic processes in oceanic mantle and crust. Earth and Planetary Science Letters (1994) 128:287–301.[Web of Science]

Meisel T., Walker R. J., Morgan J. W. The osmium isotopic composition of the primitive upper mantle. Nature (1996) 383:517–520.

Morgan W. J. Hotspot tracks and the opening of the Atlantic and Indian oceans. In: The Sea—Emiliani C., ed. (1981) New York: John Wiley.

Neumann E. R., Wulff–Pedersen E. The origin of highly silicic glass in mantle xenoliths from the Canary Islands. Journal of Petrology (1997) 38:1513.[Abstract/Free Full Text]

Neumann E.-R., Wulff–Pedersen E., Johnsen K., Andersen T., Krogh E. Petrogenesis of spinel harzburgite and dunite suite xenoliths from Lanzarote, eastern Canary Islands: implications for the upper mantle. Lithos (1995) 35:83–107.[Web of Science]

Palmer M. R., Turekian K. K. 187Os/186Os in marine manganese nodules and the constraints on the crustal geochemistries of rhenium and osmium. Nature (1986) 319:216–220.[Web of Science]

Palmer M. R., Falkner K. K., Turekian K. K., Calvert S. E. Sources of osmium isotopes in manganese nodules. Geochimica et Cosmochimica Acta (1988) 52:1197–1202.[Web of Science]

Pearson D. G., Carlson R. W., Shirey S. B., Boyd F. R., Nixon P. H. Stabilisation of Archean lithospheric mantle: a Re–Os isotope study of peridotite xenoliths from the Kaapvaal craton. Earth and Planetary Science Letters (1995a) 134:341–357.[Web of Science]

Pearson D. G., Shirey S. B., Carlson R. W., Boyd F. R., Pokhilenko N. P., Shimizu N. Re–Os, Sm–Nd and Rb–Sr isotope evidence for thick Archaean lithospheric mantle beneath the Siberian craton modified by multi–stage metasomatism. Geochimica et Cosmochimica Acta (1995b) 59:959–977.[Web of Science]

Pegram W. J., Allègre C. J. Osmium isotopic compositions from oceanic basalts. Earth and Planetary Science Letters (1992) 111:59–68.[Web of Science]

Pegram W. J., Krishnaswami S., Ravizza G. E., Turekian K. K. The record of sea water187Os/186Os variation through the Cenozoic. Earth and Planetary Science Letters (1992) 113:569–576.[Web of Science]

Peucker-Ehrenbrink B., Ravizza G., Hofmann A. The marine187Os/186Os record of the past 80 million years. Earth and Planetary Science Letters (1995) 130:155–167.[Web of Science]

Ravizza G. Variations of the187Os/186Os ratio of seawater over the past 28 million years as inferred from metalliferous carbonates. Earth and Planetary Science Letters (1993) 118:335–348.[Web of Science]

Ravizza G., Esser B. K. A possible link between the seawater osmium isotope record and weathering of ancient sedimentary organic matter. Chemical Geology (1993) 107:255–258.[Web of Science]

Reisberg L., Zindler A., Marcantonio F., White W., Wyman D., Weaver B. Os isotope systematics in ocean island basalts. Earth and Planetary Science Letters (1993) 120:149–167.[Web of Science]

Roeser H. A. Magnetic anomalies in the magnetic quiet zone off Morocco. In: Geology of the Northwest African Continental Margin—von Rad U., Hinz K., Sarnthein M., Seibold E., eds. (1982) Berlin: Springer–Verlag. 60–68.

Roest W. R., Danobeitia J. J., Verhoef J., Collette B. J. Magnetic anomalies in the Canary Basin and the Mesozoic evolution of the Central North Atlantic. Marine Geophysical Research (1992) 14:1–24.

Roy-Barman M., Allègre C. J. 187Os/186Os in oceanic island basalts: tracing crust recycling in the mantle. Earth and Planetary Science Letters (1995) 129:145–161.[Web of Science]

Rudnick R. L., Goldstein S. L. The Pb isotopic compositions of lower crustal xenoliths and the evolution of lower crustal Pb. Earth and Planetary Science Letters (1990) 98:192–207.[Web of Science]

Schmincke H.-U., Klügel A., Hansteen T., Hoernle K. A., van der Bogaard P. Samples from the Jurassic ocean crust beneath Gran Canaria, La Palma and Lanzarote (Canary Islands). In: Earth and Planetary Science Letters (1999) (in press).

Shirey S. B., Walker R. J. Carius tube digestions for low–blank rhenium–osmium analysis. Analytical Chemistry (1995) 67:2136–2141.

Shirey S. B., Walker R. J. 187Re–187Os isotopes in geochemistry and cosmochemistry. Annual Review of Earth and Planetary Sciences (1998) 26:423–500.[Web of Science]

Shirey S. B., Bender J. F., Langmuir C. H. Three–component isotopic heterogeneity near the Oceanographer Transform, Mid–Atlantic Ridge. Nature (1987) 325:217–223.

Siena F., Beccaluva L., Coltorti M., Marchesi S., Morra V. Ridge to hot–spot evolution of the Atlantic lithospheric mantle: evidence from Lanzarote peridotite xenoliths (Canary Islands). In: Journal of Petrology, Special Lherzolites Issue (1991) 271–290.

Snow J., Reisberg L. Os isotopic systematics of the MORB mantle: results from altered abyssal peridotites. Earth and Planetary Science Letters (1995a) 133:411–421.[Web of Science]

Snow J. E., Reisberg L. Erratum of ‘Os isotopic systematics of the MORB mantle: results from altered abyssal peridotites’. Earth and Planetary Science Letters (1995b) 136:723–733.[Web of Science]

Stillman C., Fuster J., Bennell–Baker M., Munoz M., Smewing J., Sagredo J. Basal complex of Fuerteventura (Canary Islands) is an oceanic intrusive complex with rift–system affinities. Nature (1975) 257:469–471.

Thirlwall M. F. Pb isotopic and elemental evidence for OIB derivation from young HIMU mantle. Chemical Geology (1997) 139:51.[Web of Science]

Thirlwall M. F., Jenkins C., Vroon P. Z., Mattey D. P. Crustal interaction during construction of ocean islands: Pb–Sr–Nd–O isotope geochemistry of the shield basalts of Gran Canaria, Canary Islands. Chemical Geology (1997) 135:233–262.[Web of Science]

Vance D., Stone J. O. H., O'Nions R. K. He, Sr and Nd isotopes in xenoliths from Hawaii and other oceanic islands. Earth and Planetary Science Letters (1989) 96:147–160.[Web of Science]

Vidal P. Mantle: more HIMU in the future? Geochimica et Cosmochimica Acta (1992) 56:4295–4299.[Web of Science]

Walker R. J. Low–blank chemical separation of rhenium and osmium from gram quantities of silicate rock for measurements by resonance ionization mass spectrometry. Analytical Chemistry (1988) 58:2923–2927.

Walker R. J., Carlson R. W., Shirey S. B., Boyd F. R. Os, Sr, Nd, and Pb isotope systematics of southern African peridotite xenoliths: implications for the chemical evolution of subcontinental mantle. Geochimica et Cosmochimica Acta (1989) 53:1583–1595.[Web of Science]

Walker R. J., Morgan J. W., Naldrett A. J., Li C., Fassett J. D. Re–Os isotope systematics of Ni–Cu sulfide ores, Sudbury Igneous Complex, Ontario: evidence for a major crustal component. Earth and Planetary Science Letters (1991) 105:416–429.[Web of Science]

Wiechert U., Hoernle K., Graham D. Oxygen isotope evidence for high temperature altered oceanic crust in the Canary plume. In: EOS Transactions, American Geophysical Union Fall Meeting (1997) 825–826.

Whitmarsh R. B., Party L. S. Drilling reveals transition from continental breakup to early magmatic crust. EOS Transactions, American Geophysical Union (1998) 79:173–181.

Widom E. Sources of ocean island basalts: a review of the osmium isotope evidence. Physica A (1997) 244:484–496.[Web of Science]

Widom E., Shirey S. B. Os isotope systematics in the Azores: implications for mantle plume sources. Earth and Planetary Science Letters (1996) 142:451–466.[Web of Science]

Zartman R. E., Tera F. Lead concentration and isotopic composition in five peridotite inclusions of probable mantle origin. Earth and Planetary Science Letters (1973) 20:54–66.[Web of Science]

Zindler A., Hart S. R. Chemical geodynamics. Annual Review of Earth and Planetary Sciences (1986) 14:493–571.[Web of Science]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 GeologyHome page
J. M.D. Day, D. G. Pearson, C. G. Macpherson, D. Lowry, and J.-C. Carracedo
Pyroxenite-rich mantle formed by recycled oceanic lithosphere: Oxygen-osmium isotope evidence from Canary Island lavas
Geology, June 1, 2009; 37(6): 555 - 558.
[Abstract] [Full Text] [PDF]

Home page
 GeologyHome page
S. Duggen, K.A. Hoernle, F. Hauff, A. Klugel, M. Bouabdellah, and M.F. Thirlwall
Flow of Canary mantle plume material through a subcontinental lithospheric corridor beneath Africa to the Mediterranean
Geology, March 1, 2009; 37(3): 283 - 286.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Widom, E.
Right arrow Articles by Schmincke, H.-U.
Right arrow Search for Related Content
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?