Apr 10, 1992 - unnamed transform at 6°N, the Doldrums transform system at around 8°N .... and by analyzing the center rather than the edge of grains. ... Plot of 100 Mg/(Mg + Fe 2+) versus 100 Cr/(Cr + A1) of spinel from .... plagioclase plus Cr-rich spinel. ... S7-27-2. S7-27-3. S7-27-6. S7-59-1A. S7-15-1. S7-26-2. S7-27-6.
JOURNAL
OF GEOPHYSICAL
RESEARCH,
VOL. 97, NO. B4, PAGES 4461-4476, APRIL
10, 1992
Upper Mantle Heterogeneity Below the Mid-Atlantic Ridge, 0ø-15øN E. BONATTI, 1,2A. PEYVE, 3 P. KEPEZHINSKAS,4 N. KURENTSOVA,3 M. SEYLER,l S. SKOLOTNEV,3 AND G. UDINTSEV3 Small-scale variations in composition of mantle-derived peridotites have been investigated in the 0ø-15øN portion of the Mid-Atlantic Ridge (MAR), thanks to a relatively close-spaced peridotite sample coverage achieved by combining samples collected by Russian and U.S. expeditions. Areal variations in the composition of mantle-equilibrated minerals olivine, orthopyroxene, clinopyroxene, and spinel have been interpreted as due primarily to regional variations in the initial composition, degree of partial melting, and thermal structure of the upper mantle. Mantle rocks from the eastern part of the Romanche transform frequently contain a trapped fraction of basaltic melt, while undepleted mantle prevails in the western part of the Romanche, suggestinga "cold" upper mantle thermal regime in this region, which prevented significantmelting. Immediately to the north, the St. Paul Fracture Zone (FZ) upper mantle shows intermediate degrees of melting, except for St. Peter-Paul Island which exposes metasomatized mantle rocks chemically and isotopically different from other oceanic peridotites. Between St. Paul FZ and 4øN (Strakhov FZ) we have an area of strongly depleted upper mantle. Farther north the Doldrums FZ area (-8øN) appears to be underlain by moderately depleted upper mantle with some melt entrapment. The Vema FZ (11øN) is underlain by relatively homogenousupper mantle which has undergonea rather low degree of melting. The Mercurius and Marathon transforms (between 12ø and 13øN) expose moderately depleted peridotites. Finally, the 15020' FZ area shows relatively undepleted upper mantle on the northern side of the transform and at sites distant from the MAR axis and strongly depleted mantle south of the transform. The strongly depleted mantle from the 2ø-3øN and 14ø-15øNregions is associated spatially with light rare earth element enriched mid-ocean ridge basalt showing a "hot spot"-type geochemical signature. The areal association of refractory peridotites with enriched basalt and with zero-age topographic highs in the 2ø-3øNand 14ø-15øNregions can be explained either by the influence of mantle thermal plumes or by the presencein the mantle of metasomatized,H20-rich domains which would cause enhanced melting and provide a source for basalt enrichment. These mantle domains might be relicts of an originally subcontinental mantle.
INTRODUCTION
which concentrate refractory elements). Thus, for instance, peridotite modal and mineral chemistry suggests that the The objective of this paper is to investigate the extent and upper mantle in the 34ø-45øN region of the Mid-Atlantic scale of upper mantle compositional heterogeneity in a Ridge (MAR) has been subjected to a higher degree of limited portion of the Mid-Ocean Ridge system, i.e., from 0ø melting than elsewhere in the North Atlantic [Dick et al., to 15øNin the Atlantic. The propertiesof the upper mantle in 1984; Michael and Bonatti, 1985]. The chemistry of basalts this region will be inferred from the study of peridotites from the same region also indicates that they originated from recovered from the ocean floor. a high degree of melting of their mantle peridotite source Previous studies of North Atlantic peridotites have re- [Klein and Langmuir, 1987]. One of the objectives of this paper is to investigate vealed broad, long-wavelength (-1000 km) regional variawhether or not mantle heterogeneity and the spatial corretions in their modal and mineral chemical composition [Dick lation between peridotite and basalt properties extend to et al., 1984; Michael and Bonatti, 1985]. These variations can be spatially correlated with variations in chemistry of smaller wavelength (-100 km or less). This has implications basalts [Schilling et al., 1983; Klein and Langmuir, 1987] and for ideas such as temporal cyclicity of melt injection and with variations of zero-age seafloor depth and of residual amagmatic extension at mid-ocean ridges, subseafloor flow geoid anomaly (W. F. Haxby et al., manuscript in prepara- of melt, and mechanisms of segmentation of mid-ocean ridges. These objectives have been tackled in a study of tion, 1991) along the axial zone of the North Atlantic. The regional correlation of peridotite and basalt chemistry mantle-derived peridotites in a region where sample coververifies the expected complementarity in composition be- age is relatively dense, i.e., 0ø-15øN in the Atlantic. The tween the melt fraction extracted from the upper mantle relatively high density of peridotite samplesavailable in the (basalt, where incompatible elements are concentrated) and 0ø-15øN region has been achieved by putting together samthe solid residue left behind from partial melting (peridotites, ples obtained by Russian and U.S. expeditions. Accordingly, this paper is the first resulting from an informal collaboration between groups from the Institute of Geology ILamont-Doherty Geological Observatory of ColumbiaUniverand the Institute of the Lithosphere, both of the Russian sity,,Palisades, NewYork. qnstituto Geologia Marina CNR, Bologna, and Dipartimento Academy of Sciences in Moscow, and Lamont-Doherty Scienze della Terra, Pisa, Italy. Geological Observatory of Columbia University.
3Institute of Geology,Russian Academyof Sciences, Moscow. 4Instituteof the Lithosphere, RussianAcademyof Sciences,
SAMPLE COVERAGE
Moscow.
Copyright 1992 by the American Geophysical Union.
The MAR is offset in the 0ø-15øN region by a series of major transform zones (Figure 1). These include, from south to north, the Romanche Fracture Zone (FZ) [Heezen et al.,
Paper number 91JB02838. 0148-0227/92/91 JB-02838505.00 4461
4462
BONATTIET AL.' MANTLEHETEROGENEITY BELOWMID-ATLANTICRIDGE
15o20 '
El26
15øN
25 25 33
15'
MARATHON
40
3• 0
VEMA iO ø
37
I0 ß
•8
4:5
26
19
DOLDRUMS :>5 24
SIERRA
LEONE
28
STRAKHOV
051 59
r-129 ST. PAUl
0 ß
SPINEL
A 40
.•'
TRAPPED
4 5ø
44
PETER-
ROMANCHE
0ß
PAUL 28
ISLAND
CHAIN
59
40 ø
13 58
MELT
55
35 ø
30 ø
?.5'
20 ø
15ø
IO ß
Fig. 1. Distribution ofthespinel 100Cr/(Cr+ A1)ratioinperidotites fromthe0øto I$øNarea.Thenumbers indicate theratiosaveraged foreachsite.Smallarrows pointto siteswhereperidotites showing meltentrapment werefound.
spinel. Locally, plagioclase(pl) and high-temperatureamphibolealso are present.Our work has concentratedon unnamed transform at 6øN, the Doldrums transform system determiningthe major elementchemistryof theseprimary at around 8øN [Pushcharovskyet al., 1989], the Verna FZ relict phases,therebyavoidingthe effectsof serpentiniza[Van Andel e! al., 1971], the Marathon and Mercurius tion, which make the whole rock compositionof these sampleshard to interpretin termsof upper mantlepropertransforms between 12øN and 14øN and the 15020' FZ (or Cape Verde) [Roestand Collette, 1986;Pushcharovsky et ties. The criteria for sample selectionare similar to those
1964], St. Paul FZ [Heezen et al., 1964; Udintsevet al., 1989], the Strakhov FZ at 4øN [Udintsevet al., 1989], an
al., 1988].
describedby Michael and Bonatti [1985].
Most of the peridotitesamplesutilized in this work were recoveredby dredgingfrom the wallsof the deepvalleysthat
The analyseswere carried out mostly at LDGO with a Camebax4-spectrometer,fully automatedelectron microprobe, usingnatural standardsand employinga Cameca
characterize
the transform
zones or their fracture
zone
for data correction,and extensions.A numberof sampleswere obtainedfrom reliefs program(basedon Bence-Albee) locatedparallelto the MAR axis away from fracturezones. partlyat the Instituteof Geologyof the USSRAcademyof This has been achieved particularly at 2ø-3øN, in an area Science in Moscow, with a Camebax, using the same probetween the St. Paul FZ and the Strakhov FZ at 4øN (Figure 1).
The locationsof a seriesof samplescollectedby the R/V
gram.Fe3+in spinelwascalculated witha Camecaprogram accordingto the following equation:
Akademik Nikolai Strakhov of the Russian Academy of
Fe3+ = 2/3[Fe- (A1+ Cr)/2]- 2[Ti- {(Mn + Mg)/2}]
Sciencesare givenin Table 1. Thesesamplesarefromthe St. Paul FZ, the 2ø-4øN area, the Doldrums FZ at 8øN, the
The results of the mineral chemistry determinationsfor the peridotitesfrom the St. Paul FZ and the 2ø-4øNarea are
Marathon and Mercurius FZs between 12øN and 14øN, and
reportedin Table 2. The resultsfor the samplesfrom the
the 15020' FZ (Figure 1). The locationsof samplesfrom the RomancheFZ, 6øN, and Vema FZ, and additionalsamples from the St. Paul FZ, recovered from vesselsof the University of Miami and Lamont-DohertyGeologicalObservatory (LDGO), are given by Bonatti et al. [1971, 1976, also manuscriptin preparation, 1992].
Doldrums FZ are in Tables 3 and 4, from the Marathon FZ and Mercurius FZ are in Table 5, and from the 15020' FZ in
METHODS
OF STUDY AND RESULTS
Most oceanicperidotites,includingthe samplesutilizedin this work, are serpentinizedto various degrees.However,
they containrelicts of mantle-equilibrated phasessuchas olivine(ol), orthopyroxene(opx), clinopyroxene(cpx), and
Table 6. Spinelstructuralformulaeare basedon 32 oxygens. Additional results for 15020' FZ samples are reported by Peyve et al. [1989].The resultsfor the Romanche,Vema, 6øN and St. Peter-PaulIsland peridotitesare tabulatedby E. Bonatti et al. (manuscriptin preparation, 1992),where some additionaldata for St. Paul FZ samplesare also presented. MINERAL CHEMISTRY OF THE 0ø-15øN PERIDOTITES
Even a cursoryexaminationof Tables2-5 showsthat the mineralchemistryof the 0ø-15øNperidotitesis quitevariable
BONATTI
TABLE
1.
ET AL.' MANTLE
HETEROGENEITY
Locations in the Central Atlantic From 0ø to 15øN,
Where Peridotites Were Recovered by the R/V Akade/nik Nicolai Strakhov of the Geology Institute, Russian Academy of Sciences, Moscow Station
Latitude
Longitude
S3-12 S3-22 S3-33 S 3-40 S3-55 S3-63 S3-64 S6-2 S9-8
14ø51.1 'N 15ø13.2'N 15009.8'N 15o13.4' N 15ø19.7'N 15ø21.5'N 15023.2'N 15ø03.9'N 15006.5 'N
S9-36 S9-41
12ø58.0'N 12ø57.0'N
S9-98
12ø07.2'N
S6-44
Depth, m
15ø20'FZ
Marathon
Mercurius
42ø30.4'W 45ø14.2'W 44049.0'W 44046.2 'W 46ø46.5'W 46ø24.0'W 46ø24.8'W 45ø45.9'W 44051.0'W
4800-4250 4500-4100 5075-4100 4670-4480 3980-3700 4200-3800 3800-3500 2800-2700 4250-3800
FZ 44ø39.0'W 44ø57.0'W
3120-2800 2750-2200
FZ
43ø31. I'W
3400-3000
8ø07.8'N
FZ 40ø34.7'W
3780-3750
S6-48
8ø13.5'N
37ø26.8'W
3910-3760
S6-58 S6-59 S6-63 S6-64 S9-61 S9-65 S9-66 S9-69 S9-70 S9-71 S9-73 S9-76 S9-81
8ø17.7'N 8 ø15.0'N 7ø44.7'N 7ø47.8'N 7o46.4' N 7056.1 'N 8o04.8'N 8ø11.9'N 8ø16.5 'N 8ø08.4'N 8 ø18.0'N 8 ø10.1 'N 8 ø10.8 'N
38ø04.2'W 37056.4' W 37ø43.5'W 37ø44.7'W 37046.3 'W 37ø57.6'W 38006.0'W 38ø17.9'W 38ø19.3 'W 38ø18.3 'W 38022.0'W 38o22.0'W 38ø21.8 'W
4600-4500 4650-4450 4700-4400 2070-2020 2800-2600 2900-2600 3300-2400 3425-3325 3900-3800 4000-3900 3850-3800 3900-3850 3500-3400
Doldrums
2ø-4øN
S7-39 S7-43 S7-54
2ø54.9'N 2 ø15.5'N 2ø05.8'N
0ø58.7'N 0ø59.3'N 0ø57.2'N 0ø28.1 'N 0ø32.8'N 0ø22.0'N
FZ 29ø21.8'W 29ø23.7'W 29ø04.5'W 25ø47.2'W 25ø47.7'W 25ø02.2 'W
2840-2350 2930-2070 3030-2690
RIDGE
4463
melting of which the samples are residual. This explanation is essentially the one adopted by Dick et al. [1984] and Michael and Bonatti [1985] when discussingthe regional long-wavelength variations of mineral chemistry in the North Atlantic peridotites. However, a number of additional factors can complicate this simple picture. One factor is that the composition of the parental mantle may be different in different portions of the MAR. In addition, the chemistry of mantle-equilibrated phases can be affected by secondary processes such as (1) low-pressure, subsolidus recrystallization; (2) reactions related to entrapment of melt in the peridotites; (3) lowtemperature alteration. Problem 3 has been tackled by avoiding minerals showing alteration under the microscope and by analyzing the center rather than the edge of grains. The presence of plagioclase in the samples may signify that they have been affected by processes 1 and/or 2, which may also have modified pyroxene and spinel chemistry. Let us now evaluate the extent to which melt entrapment has affected our samples. MELT
CONTAMINATION
IN THE PERIDOTITE
A plot of Mg # (Mg # = 100Mg/(Mg + Fe2+)) versus Cr # in spinel is shown in Figure 2. The ratio of spinel MgO to FeO* (total iron) and to TiO2 is shown in Figure 3. While most spinels of the 0ø-15øN peridotites contain 40), indicating that they are residual after high degrees of melting (>20%). No fresh opx were observed in these samples. Of the samplesfrom these sites, only one shows anomalously high spinel Ti and Fe content (Figure 3) which might indicate contamination by melt. We note that all three sites are on basement ridges located east of and parallel to the axial segment of the MAR, away from fracture zones. The extent to which the high degree of depletion shown by these peridotites is caused by their being originally upper mantle from an axial segment of spreading far from a transform is not clear. Previous attempts to detect a "transform effect" in the composition of peridotites failed
0.20 0.06 99.42
1.874 0.005 0.227 0.032 0.085 0.002
0.58 0.03 99.73
0.30 0.05 99.91
0.23 0.05 99.90
1.896 0.018 0.164 0.027 0.090 0.002
1.919 0.008 0.124 0.032 0.079 0.002
1.907 0.002 0.144 0.036 0.081 0.002
1.861 0.001 0.222 0.043 0.091 0.002
1.908 0.003 0.159 0.033 0.074 0.002
1.870 0.004 0.226 0.037 0.086 0.002
1.883 0.004 0.203 0.027 0.104 0.002
0.972 0.805 0.013 0.002 4.012
0.918 0.891 0.009 0.002 3.999
0.890 0.865 0.004 0.002 3.986
0.936 0.832 0.004 0.002
Mg
0.879
0.891
0.928
0.972
Ca Na Ni Total
0.881 0.014 0.002 4.001
0.883 0.041 0.001 4.013
0.891 0.022 0.002 4.007
0.847 0.016 0.002 4.009
Mg #
91.18
90.67
92.15
92.31
91.44
Wo En Fs
47.73 47.64 4.63
47.33 47.76 4.91
46.94 48.92 4.14
44.56 51.17 4.27
453.05 52.08 4.87
3.997
92.54
91.19
90.00
47.33 48.75 3.92
46.98 48.34 4.68
44.43 50.00 5.57
BONATTI ET AL.' MANTLE
HETEROGENEITY
to show a systematic difference in chemistry of peridotites from fracture zones and those away from fracture zones within the same region [Michael and Bonatti, 1985; Komor et al., 1990]. 4ø-8øN area. We have a gap in sample coverage between the 4øN FZ and the Doldrums FZ at about 8øN, except for peridotites recovered from one locality on the western wall of the MAR axial valley; near its intersection with an unnamedfracture zone at about 6øN. The peridotites contain spinel with intermediate Cr # (28 _+ 2) and opx A1203 content (close to 4%), suggestingmoderate (10-15%) degree of melting. Doldrums
FZ.
Peridotites
recovered
from six sites in the
Doldrums FZ area (Figure 6) display a wide range of spinel Cr # (range from 19 to 58) and of opx Al:O3 content (range from 2 to 6%). Samplesfrom two locations outside the active transform (one about 50 km east of the eastern MAR/ transform intersection, the other about 100 km west of the western intersection) give moderately low (24-25) spinel Cr # values and moderately high opx A1203 content (4.66.0%), correspondingroughly to 10-15% melting. However, samples from the transform/MAR intersection and from the transform fault immediately to the south of Doldrums (at 7ø40'N) are moderately depleted with spinel Cr # from 35 to 52 and some even up to 58. Some of the sampleswith high Cr # spinel values show also enrichment of Ti and Fe in the spinel (Figure 3). Thus melt entrapment was probably widespread in the Doldrums upper mantle. Evidence for basaltic intrusive magmatism is given by sampling in this area of gabbroic complexes [Kepezhinskas, 1989]. Xenoliths of spinel lherzolite composition have been found in some samples of mid-ocean ridge basalt (MORB) recovered from the Doldrums FZ [Kepezhinskas, 1989].
BELOW MID-ATLANTIC
RIDGE
4469
Based on mineral chemistry, two types of xenoliths have been identified. Type 1 inclusions contain cpx with moderate Na20 content (40) Cr #. Two of these are from the eastern
MAR/transform
intersection.
The lowest
ratio
(26) is found in peridotites from a site on the nonactive part of the fracture zone, about 200 km from the eastern MAR/
Clinopyroxene in Peridotites From the Doldrums FZ S6-63-46
S6-64-10
S9-61-4a
S9-61-4b
S9-65-1
S9-69-2
52.14 0.51 3.30 1.15 2.91 0.06 16.95 22.06 0.34 0.04 99.45
51.97 0.17 3.41 1.05 2.79 0.06 17.47 21.96 0.11 0.04 99.03
51.50 0.33 4.24 1.89 2.31 0.07 16.30 22.13 0.59 0.02 99.36
52.53 0.30 2.63 1.41 2.17 0.05 17.29 22.43 0.52 0.02 99.35
53.36 0.20 1.43 0.59 6.72 0.23 15.70 21.91 0.46 0.02 100.64
50.96 0.39 5.13 1.18 3.13 0.08 16.63 22.63 0.17 0.04 100.33
S9-71- la 51.25 0.36 5.36 1.39 2.93 0.09 16.38 22.15 0.46 0.04 100.41
S9-71-1b 51.20 0.36 5.30 1.51 3.23 '" 17.25 21.51 0.47 0.03 100.86
1.908 0.014 0.142 0.033
1.906 0.005 0.148 0.030
1.887 0.009 0.183 0.055
1.922 0.008 0.114 0.041
1.959 0.006 0.062 0.017
1.854 0.011 0.220 0.034
1.860 0.010 0.229 0.040
1.850 0.010 0.226 0.043
0.089 0.002 0.924 0.864 0.024 O.O01 4.001
0.086 0.002 0.955 0.863 0.005 O. O01 4.001
0.071 0.002 0.890 0.868 0.042 O. 000 4.007
0.067 0.002 0.943 0.879 0.036 O.O01 4.013
0.209 0.007 0.858 0.861 0.033 0.001 4.012
0.095 0.003 0.902 0.882 0.012 0.001
0.089 0.003 0.885 0.861 0.032 0.001
4.014
4.010
"' '" 0.929 0.833 0.031 0.001 3.957
91.05
91.74
92.61
93.36
80.61
90.47
90.86
"'
46.00 49.16 4.83
45.33 50.18 4.49
47.40 48.60 4.00
46.50 49.90 3.60
44.54 44.40 11.06
46.88 47.93 5.19
46.90 48.20 5.00
44.80 50.00 5.26
4470
BONAT'FI ET AL.' MANTLE
HETEROGENEITY
BELOW MID-ATLANTIC
RIDGE
TABLE 4. Selected ElectronProbeAnalyses of Olivine(ol),Orth0pyroxene (opx),andClinopyroxene (cpx)in a Lherzolite Xenolith (Sample S6-48-44)in Basalt From the Doldrums FZ
SiO2 TiO2 A1203 Cr20 3 FeO MnO
cpx
cpx
cpx
cp•
opx
opx
opx
ol
ol
53.09 0.13 4.28 1.30
53.18 0.09 4.00 1.22
52.63 0.12 4.20 1.31
53.29 0.11 4.20 1.28
56.25 0.02 2.42 0.48
55.61 0.03 2.61 0.50
55.81 0.03 2.62 0.51
40.54 0.00 0.00 0.03
40.45 0.00 0.00 0.02
2.34 0.08
2.28 0.10
2.34 0.09
2.30 0.09
5.45 0.14
5.40 0.14
5.47 0.14
8.77 0.15
8.74 0.13
34.83
35.24
34.98
51.42
51.11
0.65
0.72
0.76
0.06
0.06
0.00
MgO
16.70
16.75
16.86
16.84
CaO
20.46
20.54
20.49
20.40
Na20
1.49
1.50
1.5!
1.47
0.08
0.08
0.07
0.00
K20
0.00
0.00
0.01
0.00
0.01
0.01
0.00
0.00
Total
99.87
99.66
99.56
99.98
100.33
100.34
100.39
100.97
0.00 100.52
Si Ti AI Cr
1.922 0.004 0.182 0.037
1.929 0.003 0.171 0.035
1.914 0.003 0.180 0.038
1.925 0.003 0.179 0.037
1.909 0.000 0.098 0.061
1.890 0.000 0.106 0.061
1.901 0.001 0.106 0.061
0.981 0.000 0.000 0.001
0.983 0.000 0.000 0.001
Fe 2+
0.07i
0.069
0.071
0.070
0.153
0.153
0.157
0.177
0.178
Mn
0.003
0.003
0.003
0.003
0.002
0.002
0.004
0.003
0.003
Mg
0.901
0.906
0.914
0.907
1.762
1.784
1.772
1.855
1.851
Ca Na
0.793 0.105
0.798 0.106
0.798 0.107
0.790 0.103
0.022 0.008
0.024 0.008
0.028 0.005
0.002 0.000
0.002 0.000
K Total
0.000 4.018
0.000 4.020
0.000 4.028
0.000 4.017
0.000 4.015
0.000 4.028
0.000 4.035
0.000 3.019
0.000 3.018
Mg/(Mg+ Fe)
0.93
0.93
0.93
0.93
0.92
0.92
0.92
0.91
0.91
transform intersection. Available opx A1203 data correlate inversely with the spinel Cr # values. Only one sample shows evidence
of melt contamination.
TABLE 5a. Representative Electron Probe Analyses of Orthopyroxene (opx) and Clinopyroxene (cpx) From the
SiO2 TiO 2
and Mercurius
FZs
S9-41-31
S9-98-4
S9-98-4
opx
opx
cpx
55.32 0.02
55.12 0.03
52.59 0.07
4.34
AI2O3
3.73
3.27
Cr•O3
1.11
0.91
1.45
FeO
6.00
6.58
3.48
MnO
0.12
0.12
0.05
MgO
32.87
33.52
20.24 19.18
CaO
1.23
1.12
Na20
0.01
0.14
NiO Total
Si Ti AI Cr Fe Mn
0.10 100.52
1.906 0.001 0.151 0.030 0.173 0.004
and to the east of the MAR
axis below
older c•'ust.
These data could
suggestfor the 15020' FZ a rather strongly depleted mantle (>20% melting) in areas close to the present MAR axis mostly at the southern side of the fracture zone. Less
Marathon
depletedmantleis at the northernside of the fracture zone
0.10 100.90
1.899 0.001 0.133 0.025 0.190 0.003
0.06 0.07 101.53
1.873 0.002 0.182 0.041 0.104 0.002
DISCUSSION
Upper mantle heterogeneity. Ultramafic sample coverage is not yet dense enough along the mid-ocean ridge to detect short-wavelength heterogeneities of upper mantle composition. However, we attempted a first step towards that goal in the 0ø-15øN portion of the MAR, where sample coverage is denser than elsewhere. The results are consistent with an upper mantle which is heterogeneous on a -•100-km scale, with adjacent areas of relatively depleted and undepletedupper mantle compositions.The upper mantle is strongly refractory in two areas, i.e., the 2ø-4øN area
TABLE 5b.
Representative Electron Probe Analyses of Spinel From
A1203 Cr20 3 Fe20 3
the Marathon
and Mercurius
FZs
S9-41-30
S9-41-31
S9-98-4
S9-98-6
S9-98-8
32.18 37.59 1.61
34.40 36.62 0.49
39.70 31.86 0.89
36.83 32.93 ......
33.82 33.39
FeO
15.05
14.58
15.58
15.87
18.25
MgO
14.69
15.28
15.68
15.08
13.91
TiO2
0.03
0.03
0.05
0.22
NiO Total
0.13 101.29
0.11 101.52
0.14 103.87
8.769 6.878
9.267 6.619
Fe
0.285
0.084
0.147
......
Fe
2.915
2.788
2.863
3.029
3.600
Mg
5.063
5.206
5.134
5.130
4.892
1.689
1.723
1.075
0.046 0.001 0.003 4.003
0.041 0.009 0.003 4.026
0.732 0.004 0.002 4.015
Mg #
90.54
89.92
91.08
Total
23.942
Wo En
2.39 88.38
2.11 88.02
38.27 56.21
Cr#
Fs
9.23
9.87
5.51
Mg #
9.907 5.942
0.32 0.00 98.70
A1 Cr
Mg Ca Na Ni Total
10.280 5.535
0.10 101.02
9.402 6.227
Ti
0.006
0.004
0.008
0.043
0.063
Ni
0.025
.0.021
0.025
0.018
0.000
23.990
23.992
24.130
24.184
43.96
41.67
35.00
37.50 39.80
63.46
65.12
64.20
64.00
60.20
BONATTI ET AL.' MANTLE
TABLE 6a.
S3-22-20d
HETEROGENEITY
BELOW MID-ATLANTIC
RIDGE
4471
Representative Electron Probe Analyses of Spinel in Peridotites From the 15o20' FZ
S3-64-11 a
S3-64-11 d
S3-64-17
S6-2-3
S9-8-99
S9-8-102
S9-8-116
Al20 3 Cr20 3 Fe20 3
39.57 31.96 0.01
43.00 24.44 3.59
43.76 24.31 2.94
43.21 25.56 2.12
27.29 43.47 1.22
24.15 45.87 ......
24.15 46.26
26.07 44.85 0.47
FeO
13.82
12.86
13.29
12.90
14.96
15.88
14.94
15.68
MgO TiO2 NiO
16.32 0.07 0.14
17.23 0.06 0.30
17.09 0.05 0.26
17.21 0.05 0.24
14.29 0.06 0.09
13.46 0.00 0.02
14.33 0.00 0.0J
13.56 0.04 0.16
Total
101.88
A1 Cr
101.48
10.358 5.614
Fe3+ Fe 2+
11.116 4.240
0.001 2.567
0.592 2.359
101.69
101.31
11.270 4.200
101.37
11.116 4.431
0.483 2.429
0.349 2.366
99.87
100.76
100.91
7.589 8.113
6.950 8.856
7.054 8.800
7.368 8.505
0.219 2.954
...... 3.243
3.005
0.085 3.146
Mg
5.409
5.635
5.565
5.624
5.027
4.900
5.138
4.847
Ti
0.012
0.010
0.009
0.008
0.011
0.000
0.002
0.006
Ni
0.025
0.053
0.045
0.042
0.017
0.005
0.005
Total
23.986
24.005
24.001
23.986
Cr #
35.13
27.62
27.15
28.40
51.66
Mg #
67.75
70.50
69.63
70.34
63.03
and the 14øN area. The possible significanceof these regions of strongly depleted upper mantle is discussedin the next section. In contrast, two areas are characterized by excep-
tionally fertile upper mantle compositions.One, the St.
23.930
23.954
0.030
24.004
24.022
56.03
55.50
53.59
61.80
64.20
60.62
is similar in compositionto the estimatedparental mantle for MORB suchas pyrolite, or Zabargadspinellherzolite, and to fertile continental nodules of Jagoutz et al. [ 1979]. No major element compositionof basalts from this MAR segmenthas
Peter-Paul massif, where the mantle rocks are not only fertile but also strongly metasomatized, will be discussed below. The other is the western part of the Romanche
transform,where samplesfrom four differentsitesall within
I
I
I
ST
PAUL
FZ
ROMANCHE
SiO2 TiO2 AI20 3 Cr20 3 FeO MnO
MgO
S3-64-1 ld
S3-64-17
S9-8-116 6O
54.49 0.04 4.50 0.94
54.62 0.05 4.64 0.96
55.71 0.03 2.31 1.10
5.38 0.09
6.96 0.10
6.01 0.12
5.96 0.13
5.42 0.11
•
8
31.01
31.76
31.6•
33.82
3.86
1.54
2.26
2.54
1.82
Na20
0.06
0.00
0.0!
0.01
0.03
NiO Total
7O
56.56 0.07 3.20 0.93
31.05 0.11 100.40
'" 100.37
0.10 100.24
0.12 100.62
0.08 100.43
Si Ti AI Cr Fe Mn
1.904 0.001 0.155 0.030 0.156 0.003
1.953 0.002 0.131 0.025 0.201 0.003
1.889 0.001 0.184 0.026 0.174 0.004
1.887 0.001 0.189 0.026 0.172 0.004
1.923 0.001 0.094 0.030 0.153 0.003
Mg
1.605
1.596
1.641
1.628
1.738
Ca Na Ni
0.144 0.004 0.003
0.057 0.000 -"
0.084 0.001 0.003
0.094 0.001 0.003
0.068 0.002 0.015
Total
4.01
3.97
4.01 •
4.005
4.027
Mg #
91.01
88.67
90.24
90.25
Wo En Fs
7.59 84.09 8.32
3.07 86.08 10.84
4.43 86.24 9.34
4.95 8.58 9.27
FZ
,s-zo'•z
54.89 0.05 3.80 1.09
CaO
EAST
FZ
DOLDRUMS
TABLE 6b. Representative Electron Probe Analyses of Orthopyroxene in Peridotites From the 15o20' FZ S3-63-50
(STRAKHOV)
ROMANCHE WEST VE MA
S3-22-20d
I
ST PETER-PAUL ISLAND
100 km of the MAR/transform intersection give consistantly undepleted compositions. One sample (AT196AE) with •15% modal cpx, opx A1203 •6%, and spinel Cr # = 10.1
R
V
5o
g
• ,•o
(•)P 20
V
R
IO
I
92
i
91
I
90
819
I
88
I
87
i
86
Fo % OLIVINE
3.70 88.30 8.00
Fig. 5. Fo content of olivine versus 100 Cr/(Cr + AI) ratio of spinel in peridotites of the 0ø-15øNregion of the Mid-Atlantic Ridge. Small arrows indicate samples affected by melt entrapment.
4472
BONATTI ET AL.: MANTLE HETEROGENEITY BELOW MID-ATLANTIC
RIDGE
15ø20' FZ Z6
:35(
t "-I00 km
i
MARATHON
0•8)
FZ
MERCUR IUSFZ
ED 25 (5.i)
DOldrUms
fZ
Cr*SPINEL
x NO SPINEL
• < ?_0 0 ?_0-:30
0 30-40 ß
>'40
ANALYSIS ?_5Cr*SPINEL
(2.5) %AI?_O 3 OPX /
TRAPPED
MELT
Fig. 6. Schematic representationof the 15020', Marathon, Mercurius and Doldrums transforms, showing the locationof peridotite samples,the ratio 100Cr/(Cr + AI) of spinel, and the AI203 contentof opx (valuesin parentheses). Small arrows point to sites where peridotiteswith sign of melt entrapmentwere found.
been published as yet to confirm the low degree of melting suggestedby the peridotite data, although Schilling et al. [1990] suggesta low degree of melting in the MAR segment between
the Romanche
FZ and the St. Paul FZ.
The presence of such fertile mantle close to the axis of the MAR is puzzling. It could simply represent a fragment of originally subcontinental upper mantle, as inferred for the St. Peter-Paul ultramafic body [Bonatti, 1990b]. Another possibility is that the upper mantle in this area is exceptionally cold, and thus it has not undergone significant melting during upwelling beneath the MAR segmentextending north of the Romanche FZ. This cold upper mantle might be related to a "cold edge effect" created by the long-offset Romanche transform on the MAR segment intersecting it. However, an equivalent area with undepletedmantle has not been observed
near the eastern MAR/Romanche
transform
intersection, where instead we found evidence of widespread melt entrapment in the mantle rocks. Melt entrapment could imply mantle ascending into relatively cold lithosphere where heat loss by conduction leads to in situ melt crystallization [Boudier and Nicolas, 1986]. Accord-
ingly, the eastern side of the Romanchetransform could also be underlain by a relatively cold upper mantle, though not as cold as that inferred for the western Romanche region. Peridotite/basalt relationships. Available data on basalts from the axial segmentsof the MAR in the 0ø-15øN region show the presence of enriched MORB (EMORB) in the 2ø-3øN area [Schilling et al., 1989] and at around 14øN up to the southern side of the 15020' FZ [Bougault et al., 1988; Peyve et al., 1989]. The transition from depleted normal MORB (NMORB) in the 0ø-2øN region to EMORB in the 2ø-4øN region [Schilling et al., 1989] parallels the transition from relatively undepleted peridotite to strongly depleted peridotite in the same two regions (Figure 7). In the 14ø-15øN area NMORB are found mainly at the northern
side of the 15020 ' FZ and at a distance to the east
from the MAR axis. EMORB (La/Sm > 1) are found mostly within
the intersection
of the fracture
zone with the south
segmentof the MAR and along the ridge axis to the south [Bougault et al., 1988; Peyve et al., 1988a, b]. The highly depletedperidotitesspatially coincidewith EMORB and the undepleted peridotites with NMORB (Figure 7). It appears,
BONATTI ET AL.: MANTLE
HETEROGENEITY
therefore, that an areal correlation exists between enriched basalt and depleted peridotites. This areal correlation mimics on a small scale (-100 km) the broader (--• 1000 km) areal correlation found for the axial North Atlantic by Dick et al. [1984] and Michael and Bonatti [1985]. The strongly refractory upper mantle of the 2ø-4øN and 14øN areas could be caused by enhanced melting related to mantle temperature anomalies or hot spots. The 2ø-4øN
anomaly has been linked by Schilling et al. [1990] to the influence of the off-ridge Sierra Leone hot spot. An alternative possibility is that the two areas with enhanced melting are caused not by a mantle thermal anomaly but by heterogeneity in the initial composition of the upper mantle. The light rare earth element (LREE) enrichment of the basalt found in the 34ø-45øNportion of the Mid-Atlantic Ridge (Azores hot spot region) has been explained by the presencein that area of a metasomatized,H20- and LREEenriched upper mantle [Schilling et al., 1980]. The high degree of melting inferred for the 34ø-45øN upper mantle from peridotite [Dick et al., 1984;Michael and Bonatti, 1985] and basalt data [Klein and Langmuir, 1987] has been explained by Bonatti [ 1990b] as causednot by abnormally high mantle temperatures but by enhanced melting of H20-rich mantle domains. In fact, the solidus temperature of "wet" peridotite is substantially lower then that of"dry" peridotite [Kushiro et al., 1968; Kushiro, 1969; Wyllie, 1971, 1988; Green, 1973; Mysen and Boettcher, 1975]. This permits the achievement of a higher than normal degree of melting even in zones with normal sub-mid-ocean ridge upper mantle I
9-__3He/4He N-MORB
I
....
oCrO ....
8-
?
i•
i
,
I
I
N-MORB
i
©l
BASALT
3'- Le/Sm 2_.o
o
o
i
C
6o -
PERIDOTITE
IOOCr 50 ,-,SPINEL (Cr+AI) øo 4O i,
d
,o1'
I 2oooJZERO-AGE DEPTH
6000 25"N
•"
RIDGE
4473
temperatures. Thus, according to this hypothesis, the presence in the 34ø-45øN upper mantle of metasomatized H20rich domains could explain both the high degree of melting and the enriched nature of the basalt [Bonatti, 1990b]. A similar hypothesis could be applied to the spatial association of highly depleted peridotites with LREE-enriched basalts in the 2ø-4øN and 14ø-15øN regions. More work is necessary to decide whether the 2ø-4øN and 14øN anomalies are better explained by an unusually hot or by an unusually wet upper mantle, or by a combination of the two.
He isotopic relationships. The He isotopic composition of a number of zero-age basaltic glasses from the 14ø-15øN area has been reported by Staudacher et al. [1989]. The LREE-enriched basalts found in the 14ø-15øN zero-age to-
pographic highshow3He/4HeratioslowerthanNMORB, while typical NMORB ratios (ranging between times 8 to 9 atmospheric) were found in the other samples (Figure 7).
ThelowerthanMORB 3He/4Heanomalyimpliesthe existence in the 14ø-15øN "mini-hot spot" of a mantle component
whichhaseitherlost 3He (relativeto normalMORB mantle) by degassing,or to which U and Th have been added. One
possibilityis 3He lossrelatedto degassing inducedby a mantle metasomatic event, which could also explain the enrichment shown by the 14ø-15øN"mini-hot spot" basalts. We note that lower than MORB 3He/4He values similar to those of the 14ø-15øN "mini-hot spot" have been found in the 34ø-45øN, Azores hot spot region of the North Atlantic
[Kurzet al., 1982].TheselowerthanMORB 3He/4Heratios contrast with higher than MORB ratios found in a number of other hot spots such as Hawaii and Iceland [Kurz et al., 1982]. We note also that a sample of peridotite from St.
............ i
BELOW MID-ATLANTIC
I
Fig. 7. Profiles along the Mid-Atlantic Ridge from 0ø to 25øN showing, from bottom to top: (d) zero-age depth below sea level, from Bougault et al. [1988] and Vogt [1986]; (c) spinel 100 Cr/(Cr + A1) ratio of peridotites. Spinel data for the Kane FZ and adjacent areas (-24øN) are from E. Bonatti et al. (manuscript in preparation, 1992), Komor et al. [1990], and Dick and Bu!!en [1984]. (b) Basalt
La/Sm ratio from Bougault et al. [1988] and Schilling et al. [1988]. Circled P indicates peridotites from St. Peter-Paul Island. (a)
3He/4Heratio(relativeto theatmospheric value)of basaltsfromthe 0ø-15øN area. A value for a St. Peter-Paul Island peridotite is also shown. Data are from Staudacher et al. [1989].
Peter-PaulIslandgavea low 3He/4Heratio [Staudacher et al., 1989], similar to those found in the 14ø-15øN and 34ø-45øN so-called hot spots. Origin of the St. Peter-Paul Island mantle anomaly. The peridotites exposed on St. Peter-Paul show the following characteristics [Bonatti, 1990a]: (1) Judging from their mineral chemistry, they are rather undepleted. (2) Application of Wells [1977] and Lindsley and Anderson [1983] geothermometers suggeststhat SPP peridotites equilibrated at temperatures significantly lower than temperatures esti-
mated for peridotitesfrom the Mid-Atlantic Ridge. (3) A1203 spinel/opx partition (Figure 4) and spinel 100 Cr # versus Mg # (Figure 2), both bulk composition and temperaturedependent [Obata, 1976; Presnall, 1976], show that SPP peridotites fall outside the mid-ocean ridge peridotite array. This is consistent with a lower temperature of equilibration of SPP peridotites relative to mid-ocean ridge peridotites. (4) SPP peridotites have been subjected to intense metasomatism at mantle depth, which has resulted in the presence of H20-rich phases such as amphiboles, and in an enrichment in large ion lithophile elements (LILE) such as light REE [Frey, 1970; Roden et al., 1984]. This intense mantle metasomatism is generally not observed in mid-ocean ridge peridotites. (5) Oxygen fugacity estimates of mantle peridotites suggest that the SPP mantle is outside the range of oceanic peridotites but within that of continental peridotites (L. T. Bryndzia and B. J. Wood, Oxygen thermobarometry of abyssal spinel peridotites: The redox state and C-O-H volatile compositionof the Earth' s suboceanicupper mantle, submitted to Journal of Geophysical Research, 1991). Points 1, 2, and 3 above suggestthat the SPP mantle block
4474
BONATTI ET AL.: MANTLE
HETEROGENEITY
was never caught in a sub-mid-ocean ridge thermal regime, which would have resulted in a higher degree of partial melting and in higher equilibration temperature, as observed generally in mid-ocean ridge peridotites. The low degree of melting, the low temperature of equilibration, and the H20rich, metasomatized composition of SPP peridotites makes them different from mid-ocean ridge peridotites but similar to metasomatized, amphibole peridotites found in the Zabargad Island (Red Sea) mantle peridotite body. Zabargad exposes upper mantle from a preoceanic rift [Bonatti et al., 1986]. Moreover, SPP peridotites are in terms of Sp-Nd-Pb isotopes [Roden et al., 1984] similar to the inferred mantle source of East African Rift basalts [Norry et al., 1980]. Based on all these data, Bonatti [1990a, b] suggestedthat SPP does not expose oceanic mantle but a fragment of originally subcontinental rift mantle, left behind during the opening of the equatorial Atlantic. This hypothesis is consistent with a probable age of--• 155 m.y., obtained by Nd-Sr isotopic data, for the metasomatic event which affected the SPP mantle [Roden et al., 1984]. That age correspondsto the time when the equatorial Atlantic was in a preoceanic, continental rift stage. The SPP metasomatized,H2 O- and LILE-enriched mantle peridotites are a suitable source for the LILE-enriched basalts found in the 34ø-45øN segment of the MAR, which has led to the suggestionthat the 34ø-45øNmantle contains domains of the H20- and LILE-enriched metasomatized mantle similar to SPP [Bonatti, 1990a, b]. By analogy, the "mini-hot spot" type anomalies found at 2ø-3øN and 14ø15øN could be caused by enhanced melting due not to anomalous mantle temperatures but to the presence of metasomatized,H20-rich upper mantle domains, similar in elemental chemistry to the mantle body exposed at SPP, although Schilling et al. [1989] report that as far as Pb isotopic chemistry is concerned, basalts from the 2ø-3øN anomaly are distinct from SPP peridotites. If the SPP body represents a relict fragment of subcontinental rift mantle, then we should consider the hypothesis that the "mini-hot spot" type anomalies found at 2ø-3øN and at 14ø-15øNare causedby relict bodies of enriched, originally subcontinental mantle embedded passively in a depleted NMORB-type upper mantle. Peridotite chemistry/zero-age depth correlation. A broad positive anomaly of zero-age depth and of the residual geoid (W. F. Haxby et al., manuscript in preparation, 1992) has been detected in the 34ø-45øNregion of the Mid-Atlantic Ridge, where peridotites are highly depleted [Dick et al., 1984; Michael and Bonatti, 1985; Bonatti, 1990b]. A similar spatial correlation of highly depleted peridotite and LREEenriched basalt with positive zero-age topographicanomaly is observed also in the two "mini-hot spots" at 2ø-3øN and 14ø-15øN(Figure 7). Zero-age depth depends on density down to a compensating depth, assumed to be somewhere between 150 and 250 km [Klein and Langmuir, 1987]. Density dependson crustal thickness, mantle residue thickness, and temperature. Crustal thickness depends on the amount of melt extracted from the mantle (i.e., degree of partial melting and thickness of the mantle column subjected to melting). The thickness of the mantle residue is a factor because residual harzburgite is less dense than fertile lherzolite [O'Hara, 1975; Boyd and McCallister, 1976]. Klein and Langmuir [1987] have estimated that if we
BELOW MID-ATLANTIC
RIDGE
assume a degree of partial melting of sub-mid-ocean ridge mantle peridotite ranging between 8% and 25% and if this range were due solely to temperature differences at similar mantle depths, these differences would amount to --•200ø250øC. These temperature differences by themselves could account for about 25% of the observed differences in ridge axis depth. However, if the enhanced melting of the 2ø-3øN and 14ø-15øNanomalies is not due to higher than normal mantle temperatures,then the zero-age topographicanomalies must be due not only to the increased thickness of the crust
and of the residual
mantle
column
but also to the
inferred presence of metasomatized H20-rich mantle domains. These mantle domains, before they are subjected to partial melting, contain relatively low-density phasessuchas pargasiticamphibole (d = 3.0-3.1) and phlogopite (d = 2.8). Judgingfrom the St. Peter-Paul mantle body, amphiboles could constitutea very significantfraction (by volume) of the metasomatized mantle domains, thus lowering their bulk densityand contributingto the zero-age positive topographic anomaly. It may not be a coincidence that the St. Peter-Paul peridotite body, which is unique among peridotites from the oceanbasinsin being strongly metasomatized, is also unique in constituting a major topographic anomaly that projects above sea level.
SUMMARY AND CONCLUSIONS
1. The mineral chemistry of mantle-derived peridotites sampledfrom the 0ø-15øN region of the MAR shows systematic areal variations. These variations can be interpreted as indicating a small scale (< 100 km) heterogeneity of the initial composition and/or thermal structure of the upper mantle beneath the MAR, which give rise to different degrees of partial melting. 2. Areal mantle heterogeneity was observed within a singletransform zone, as, for instance, in the Romanche FZ. Peridotites
from
the
area
close
to
the
western
MAR/
transform intersection imply an upper mantle which has undergone very little melting. This region of fertile upper mantle in the vicinity of the MAR axis is unusual; it implies some sort of "cold spot" and the noninvolvement of the upper mantle in the "hot" thermal regime typical of midocean ridges. Peridotites from the central and eastern parts of the Romanche transform are frequently contaminated with a trapped melt fraction; this may also imply a relatively "cold" upper mantle thermal regime. 3. Melt entrapment inferred to be common in peridotites from the eastern part of the Romanche Transform and, to a lesser extent, from the Verna and Doldrums fracture zones is
heterogeneouson a small scale, and it occurs probably in veins, layers, and lenses scattered within the peridotitic rock.
4. The mantle peridotites exposed on St. Peter-Paul Island are very different from peridotites from elsewhere in the MAR. In addition to being strongly metasomatized, they are relatively undepleted and they equilibrated at lower temperature than other North Atlantic peridotites. This implies that they were never caught in a sub-mid-oceanridge thermal regime, which would have resulted in significant melting and higher equilibration temperature. 5. Two areas within the 0ø-15øN region show strongly depleted peridotites, implying a high degree of partial melting of the upper mantle. These two areas are at 2ø-4øN
BONATTI ET AL.: MANTLE
HETEROGENEITY BELOW MID-ATLANTIC
(between the St. Paul and the Strakhov transforms) and at 14ø-15øN(just south of the 15020' Transform). The strongly
depleted peridotites are spacially associated in these two areas with enriched MORB, in contrast with normal MORB
present elsewhere along the MAR in the 0ø-15øN region. These two areas show a positive zero-age topographic anomaly, which is particularly clear in the 14ø-15øNarea. The spacial associationof highly depleted mantle, enriched MORB, and zero-age positive topographic anomaly mimics what is observed elsewhere at hot spots. A conventional explanation of this association is the existence of temperature anomalies in the mantle, possibly related to thermal plumes. An alternative explanation calls for the presencein the upper mantle of blobs of metasomatized, H20-rich mantle similar in elemental chemistry to that exposed at St. Peter-Paul Island, which could cause enhanced melting and provide the source of basalt enrichment. These enriched mantle domains could be old subcontinental mantle fragments.
RIDGE
4475
Green, D. H., A. E. Ringwood, N. G. Ware, and W. O. Hibberson, Experimental petrology and petrogenesis of Apollo 14 basalts, Proc. Lunar Sci. Conf., 3rd, 197-206, 1972. Hamlyn, P. R., and E. Bonatti, Petrology of mantle-derived ultramarlcs from the Owen Fracture Zone, NW Indian Ocean: Implications for the nature of the oceanic upper mantle, Earth Planet. Sci. Lett., 48, 65-79, 1980. Heezen, B. C., J. B. Bunce, and M. Tharp, Chain and Romanche fracture zones, Deep Sea Res., 11, 11-18, 1964. Irvine, T. N., Chromian spinel as a petrogenetic indicator, Can. J. Earth Sci., 4, 71-103, 1967. Jagoutz, E., H. Palme, H. Baddenhausen, K. Blum, M. Cendales, G. Dreibus, B. Spettel, V. Lorenz, and H. Wanke, The abundance of major, minor and trace elements in the earth's mantle as derived from primitive ultramafic nodules, Proc. Lunar Planet. Sci. Conf., loth, 2031-2050, 1979. Jaques, A. L., and D. H. Green, Anhydrous melting of peridotite at 0-15 kb pressure and the genesis of tholeiitic basalts, Contrib. Mineral. Petrol., 73,287-310, 1980. Kepezhinskas, P. K., Ultramafic mafic inclusions in basalts from the Doldrums Fracture Zone, central Atlantic: Mantle composition and magma chamber evolution in oceanic transform faults, paper presented at 28th International Geological Congress, Washington, D. C., 1989.
Acknowledgments. We are grateful to the captain, officers, and crew of the R/V Akademik
Nikolai
Strakhov
for their collaboration
during the field work. One of us (E. Bonatti) is particularly thankful to Chief Scientists G. Udintsev and Y. Raznitsin, to the scientific group, and to officers and crew of Strakhov cruises 7 and 9 for the warm hospitality extended to him on the ship. We thank C. DeCleofis, D. Breger, P. Catanzaro, and V. Costello for help in putting together the manuscript. Research supported in part by NSF.
Contribution
4886 from
LDGO.
REFERENCES
Bonatti, E., Subcontinental mantle exposed in the Atlantic Ocean on St. Peter-Paul Islets, Nature, 345, 800-802, 1990a. Bonatti, E., Not so hot "hot spots" in the oceanic mantle, Science, 250, 107-111, 1990b. Bonatti, E., and J. Honnorez, Sections of the Earth's crust in the equatorial Atlantic, J. Geophys. Res., 81, 4104-4116, 1976. Bonatti, E., J. Honnorez, and G. Ferrara, Peridotite-gabbro basalt complex from the equatorial midAtlantic Ridge, Philos. Trans. R. Soc. London, Ser. A, 268, 385-401, 1971. Bonatti, E., G. Ottonello and P. R. Hamlyn, Peridotites from the island of Zabargad (St. John), Red Sea: Petrology and geochemistry, J. Geophys. Res., 91,599-631, 1986. Boudier, F., and A. Nicolas, Harzburgite and lherzolite subtypes in ophiolitic and oceanic environments, Earth Planet. Sci. Lett., 76, 84-92, 1986. Bougault, H., L. Dmitriev, J. G. Schilling, A. Sobolev, J. L. Joron, and H. D. Needham, Mantle heterogeneity from trace elements: MAR triple junction near 14øN, Earth Planet. Sci. Lett., 88, 27-36, 1988. Boyd, F. R., and R. B. McCallister, Densities of fertile and sterile garnet peridotites, Geophys. Res. Lett., 3, 509-512, 1976. Church, W. R., and R. K. Stevens, Early Paleozoic ophiolite complexes of the Newfoundland Appalachians as mantle-oceanic crust sequences, J. Geophys. Res., 76, 1460-1466, 1971. Dick, H. J. B., Abyssal peridotites, very slow spreadingridges and ocean ridge magmatism, Geol. Soc. Spec. Publ. London, 42, 71-105, 1989. Dick, H. J. B., and T. B. Bullen, Chromian spinel as a petrogenetic indicator in abyssal and alpine-type peridotites and spatially associated lavas, Contrib. Mineral. Petrol., 86, 54-76, 1984. Dick, H. J. B., R. L. Fisher, and W. B. Bryan, Mineralogic variability of the uppermost mantle along mid ocean ridges, Earth Planet. Sci. Lett., 69, 88-106, 1984. Frey, F. A., Rare earth and potassium abundances in St. Paul's rocks, Earth Planet. Sci. Lett., 7, 351-360, 1970. Green, D. H., Experimental melting studies on a model upper mantle composition at high pressure under water saturated and water undersaturated conditions, Earth Planet. Sci. Lett., 19, 37-53, 1973.
Klein, E. M., and C. H. Langmuir, Global correlations of ocean ridge basalt chemistry with axial depth and crustal chemistry, J. Geophys. Res., 92, 8089-8115, 1987. Komor, S.C., T. L. Grove, and R. Hebert, Abyssal peridotites from ODP Hole 670A: Residues of mantle melting exposed by nonconstructive axial divergence, Proc. Ocean Drill. Program Sci. Results, 106/109, 85-101, 1990. Kurz, M.D., W. J. Jenkins, J. G. Schilling, and S. R. Hart, Helium isotopic variations in the mantle beneath the central north Atlantic Ocean, Earth Planet. Sci. Lett., 58, 1-14, 1982. Kushiro, I., The system forsterite-diopside-silica with and without water at high pressure, Am. J. Sci., 267A, 269-294, 1969. Kushiro, I., Y. Syono, and S. Akimoto, Melting of a peridotite nodule at high pressures and high water pressures, J. Geophys. Res., 73, 6023-6029, 1968. Lindsley, D. H., and D. J. Anderson, A two-pyroxene thermometer, Proc. Lunar Planet. Sci. Conf. 13th, Part 2, J. Geophys. Res., 88, suppl., A887-A906, 1983. Melson, W. G., S. R. Hart, and G. Thompson, St. Paul's Rocks, equatorial Atlantic: Petrogenesis, radiometric ages and implications on sea floor spreading, Mem. Geol. Soc. Am., 132,241-272, 1972.
Michael, P. J., and E. Bonatti, Peridotite composition from the North Atlantic: Regional and tectonic variations and implications for partial melting, Earth Planet. Sci. Lett., 73, 91-104, 1985. Mysen, B. O., and A. L. Boettcher, Melting of anhydrous mantle: Phase relations of natural peridotite at high pressuresand temperatures with controlled activities of water, carbon dioxide and hydrogen, J. Petrol., 16, 520-548, 1975. Mysen, B. O., and I. Kushiro, Compositional variations of coexisting phases with degree of melting of peridotite in the upper mantle, Am. Mineral., 82, 52-67, 1977. Norry, M. J., P. H. Truckle, S. J. Lippard, C. J. Hawkesworth, S. D. Weaver, and G. F. Marriner, Isotopic and trace element evidence from lavas bearing on mantle heterogeneity beneath Kenya, Philos. Trans. R. Soc. London, Ser. A, 297, 259-271, 1980.
Obata, M., The solubility of A120 3 in orthopyroxenes in spinel and plagioclase peridotites and spinel pyroxenite, Am. Mineral., 61, 804-816, 1976. O'Hara, M. J., Is there an Icelandic mantle plume? Nature, 253, 708-710, 1975. Peyve, A. A., and S. A. Shcherbakov, Ultrabasic rocks, in Structure of the Cape Verde Fracture Zone, Central Atlantic (in Russian), pp. 106-117, Nauka, Moscow, 1989. Peyve, A. A., Yu. N. Raznitsin, S. M. Lyapunov, and S. G. Skolotnev, Mantle heterogeneity in the area of Cape Verde Fracture Zone in the central Atlantic according to the data on basalts (in Russian), Dokl. Akad. Nauk SSSR, 301(1), 165-168, 1988a.
Peyve, A. A., N.M. Sushevskaya, S. M. Lyapunov, and N. N. Kononkova, Peculiarities of tholeiitic magmatism in the area of
4476
BONATTI ET AL.: MANTLE
HETEROGENEITY
Cape Verde Fracture Zone in the Atlantic (13ø-15øN) (in Russian), DoM. Akad. Nauk SSSR, 302(5), 1174-1178, 1988b. Presnall, D.C., Alumina content of enstatite as a geobarometer for plagioclase and spinel lherzolites, Am. Mineral., 61, 582-588, 1976.
Pushcharovsky, Yu. M., A. A. Peyve, Yu. N. Raznitsin, S. G. Skolotnev, S. M. Lyapunov, and N. N. Tyrko, Cape Verde Fracture Zone: Rock composition and structures (central Atlantic) (in Russian), Geotektonika, 6, 18-31, 1988. Pushcharovsky, Yu. M., Yu. N. Raznitsin, A. O. Mazarovich, N. N. Tyrko, S. G. Skolotnev, S. M. Lyapunov, P. K. Kepezhinskas, D. A. Dmitriev, A. N. Didenko, V. M. Golod, and A. G. Popov, Geology of the Doldrums Fracture Zone (in Russian), Dokl. Akad. Nauk SSSR, 302(1), 167-170, 1989. Roden, M. K., S. R. Hart, F. A. Frey, and W. G. Melson, Sr, Nd and Pb isotopic and REE geochemistry of St. Paul's Rocks: The metamorphic and metasomatic development of an alkali basalt mantle source, Contrib. Mineral. Petrol., 85, 376-390, 1984. Roest, W. R., and B. J. Collette, The 15020' Fracture Zone and the North American/South American plate boundary, J. Geol. Soc. London, 143, 833-843, 1986. Schilling, J. G., M. B. Bergeron and R. Evans, Halogens in the mantle beneath the north Atlantic, Philos. Trans. R. Soc. London, Set. A, 297, 147-178, 1980. Schilling, J. G., M. Zajac, R. Evans, T. Johnson, W. White, J. D. Devine, and R. Kingsley, Petrologic and geochemical variations along the Mid-Atlantic Ridge from 29ø to 73øN, Am. J. Sci., 283, 510-586, 1983. Schilling, J. G., B. B. Hanan, and B. McCully, MAR volcanism from 3øSto 5øN: Pb isotopic evidence, Eos Trans. AGU, 69, 1426, 1988.
Schilling, J. G., D. Fontignie, B. McCully, and B. B. Hanan, Hot spot influence along the equatorial MAR: Nd-Sr-Pb isotopic evidence, Eos Trans. AGU,
70, 1350, 1989.
Schilling, J. G., A. N. Davis, D. Fontagnie, and B. McCully, Equatorial MAR segmentation from 3ø5 to 5øN: petrological and geochemical relationships, Eos Trans. A GU, 71,624, 1990. Sigurdsson,H., and J. G. Schilling, SpineIs in Mid-Atlantic Ridge
BELOW MID-ATLANTIC
RIDGE
basalts: Chemistry and occurrence, Earth Planet. Sci. Lett., 29, 7-20, 1976.
Sinton, J. M., Ultramafic inclusions and high pressure xenocrysts in submarine basanitoid, equatorial Mid-Atlantic Ridge, Contrib. Miner. Petrol., 70, 49-57, 1979. Staudacher,T., P. Sarda, S. H. Richardson, C. J. Allegre, I. Sagna, and L. V. Dmitriev, Noble gases in basalt glassesfrom a MidAtlantic Ridge topographic high at 14øN: Geodynamic consequences, Earth Planet. Sci. Lett., 96, 119-133, 1989. Udintsev, G. B., G. V. Agapova, M. G. Antipov, V. V. Eremeev, N. A. Kurentsova, and V. V. Lubimov, The equatorial segmentof the Mid-Atlantic Ridge: A structural barrier between the North and South Atlantic?, paper presented at First Congress of the Brazilian Geophysical Society, Rio de Janieiro, 1989. Van Andel, T. H., J. B. Collins, and V. T. Bowen, The intersection between the Mid-Atlantic Ridge and the Vema Fracture Zone in the North Atlantic, J. Mar. Res., 25, 343-351, 1967. Vogt, P. R., Map, in The Geology of North America, vol. M, The Western North Atlantic Region, Geological Society of America, Boulder, Colo., 1986. Wells, P. R. A., Pyroxene thermometry in simple and complex systems, Contrib. Mineral. Petrol., 62, 129-139, 1977. Wyllie, P. J., The role of water in magma generation and initiation of diapiric uprise in the mantle, J. Geophys. Res., 76, 1328-1338, 1971.
Wyllie, P. J., Solidus curves, mantle plumes and magma generation beneath Hawaii, J. Geophys. Res., 93, 4171--4181, 1988.
E. Bonatti and M. Seyler, Lamont-Doherty Geological Observatory of Columbia University, Palisades, NY 10964. P. Kepezhinskas, Institute of the Lithosphere, Russian Academy of Sciences, Moscow, Russia. N. Kurentsova, A. Peyve, S. Skolotnev, and G. Udintsev, Institute of Geology, Russian Academy of Sciences, Moscow, Russia.
(Received December 19, 1990; revised October 2, 1991; accepted November 12, 1991.)
