DESCRIPTION OF THE METAMORPHOSED MAFIC ROCKS
Petrology
The Hole 900A metamorphosed mafic rocks consist of discontinuous bands of recrystallized plagioclase and clinopyroxene that surround large isolated porphyroclasts of strained plagioclase and
clinopyroxene. The alternating discontinuous felsic and mafic bands define a poorly developed and highly variable characteristic foliation
(Fig. 4, Fig.
5, Fig. 6). Shipboard core descriptions were difficult because the first samples were highly altered and green whereas later samples were fine-grained and gray with a poorly developed foliation. Thin section study showed that the foliation was caused by discontinuous bands of recrystallized plagioclase and clinopyroxene
(Figs. 7A and B). A few large strained plagioclase and clinopyroxene porphyroclasts were observed within the finer grained recrystallized bands
(Fig. 7C), revealing a coarse-grained precursor, although there is no way to know if the porphyroclasts represent primary minerals. Recrystallized bands are typically monomineralic
(Fig. 7D), except for later alteration effects. X-ray diffraction
(XRD) study indicated that even finer material in the mafic bands consists largely of amphibole. All of the Hole 900A mafic rocks are brecciated and veined by iron oxides, chlorite,
epidote, clinozoisite, chalcedony, and lastly by calcite in places.
Our Hole 900A samples consist of between 40% and 70% plagioclase, with most of the remainder being
clinopyroxene. All samples contain small amounts of retrograde minerals and some small veins. The composition of recrystallized plagioclase is typically labradorite between An53 and An67, although it ranges from albite in more altered samples to calcic bytownite
(Table 1). Adjacent bands of recrystallized plagioclase typically have similar compositions and reported microprobe compositions represent an average of 10 or more crystals from at least two different bands. Samples 149-900A-86R-1, 51-55 cm and 900A-82R-4, 64-69 cm contain unusually calcic plagioclase of An89 and An78 respectively, perhaps reflecting variations in the bulk composition of the magma from which primary crystals precipitated. Albite (An6) occurs in Sample 149-900A-82R-3, 2-7 cm and sodic oligoclase (An11) coexists with labradorite (An43) in Sample 149-900A-81R-1, 96-101 cm, indicating local variations in retrograde metamorphism within the sample. The few plagioclase porphyroclasts analyzed are similar in composition to adjacent recrystallized calcic plagioclase, with the difference ranging from less than An1 to almost An9 in a sample where the plagioclases come from different parts of a thin section. No porphyroclasts were observed in the proximity of
albite.
The only pyroxene observed in Hole 900A samples is clinopyroxene with a relatively constant composition ranging from En42 to En46, Fs6 to Fs13, and Wo43 to Wo49
(Table 2). Recrystallized clinopyroxene in Sample 149-900A-86R-1, 51-55 cm, with the most calcic plagioclase (An89), also has the highest Wo value (Wo49). Two analyzed clinopyroxene porphyroclasts have compositions similar to adjacent recrystallized
clinopyroxenes. The clinopyroxene retrogrades to finegrained amphibole, chlorite and other minerals to varying degrees. We selected the least altered samples available; those with the least retrograded
clinopyroxene, therefore less amphibole and other retrograde minerals, and core relatively free of veins.
Major element ICP analyses of the Hole 900A metamorphosed mafic rocks confirms shipboard XRF analyses indicating that the original protolith of these rocks was either basalt or gabbro
(Table 3). The variations in major element composition with depth in the core, calculated on an anhydrous basis, show some significant correlations
(Fig. 8). The large TiO2 and FeOt peaks in Sample 149-900A-81R-1, 96-101 cm, correlate with the occurrence of opaque minerals; Core 149-900A-81R is the only part of the section in which opaque minerals were observed in thin section (Sawyer,
Whitmarsh, Klaus, et al., 1994). Shipboard major and trace element data support such a limited occurrence for the opaque minerals
(Gibson, et al., this volume). The Mg# is considerably lower in Sample 149-900A-81R-1, 96-101 cm, than in other samples because of the increase in FeO relative to
MgO. One of the more obvious correlations, in samples other than the oxide sample and uppermost Sample 149-900A-80R-2, 107-112 cm, which is altered, is MgO with FeOt and their anticorrelation with Al2O3
(Fig. 8). With the exception of Sample 149-900A-83R-3, 10-15 cm, Cr2O3 values correlate with FeOt values as expected because Cr substitutes for trivalent Fe in
clinopyroxene. In the lower part of the section, from Sample 149-900A-84R-3, 124-125 cm and below, there is a correlation between LOI and the oxidation index Fe3/Fe2 and an anticorrelation between LOI and
MgO# suggesting that alteration is greatest where the Mg/Fe ratio is lowest. This is opposite to the pattern expected because rocks with a high Mg/Fe ratio should be further from equilibrium near the surface. In addition, there are mineralogically unexplained peaks in SiO2 and Cr2O3, unexplained scatter in
LOI, an unexplained anticorrelation between CaO and Fe3/Fe2, and gradual upward increases in Mg# that are not understood.
The abundance of Al2O3 correlates with plagioclase content better than
CaO, as expected, because CaO content is influenced by small amounts of the Ca-rich secondary vein minerals calcite,
epidote, and
clinozoisite. In addition, it is not clear what happens to Ca during albitization of
labradorite; perhaps part of the Ca is captured by newly forming Ca-rich minerals while the remainder escapes from the immediate region by movement along vein fractures. In either case, the CaO will not correlate with the abundance of albitic plagioclase whereas most of the Al2O3 will remain within
albite. The CaO peak in Sample 149-900A-82R-4, 64-69 cm, may, however, correlates with the calcic plagioclase (An78) in that sample
(Table 1). The anticorrelation between Al2O3 and
MgO, FeOt and Cr2O3 probably relates to the relative abundance of plagioclase versus clinopyroxene in these samples. The large Al2O3 peak in Sample 149-900A-84R-3, 124-125 cm, may be an extreme case of such variation since it correlates with low
MgO, FeOt, and Cr2O3 and a high abundance of plagioclase (70%) in thin section. The low value for SiO2 in Sample 149-900A-86R-1, 51-55 cm, anticorrelates with Al2O3 and
CaO, most likely reflecting the calcic nature of the plagioclase (An89) in that sample
(Table 1).
Incompatible trace element contents
(Table 3) are low, which indicates that the protolith is a cumulate gabbro rather than a basalt or noncumulate (isotropic)
gabbro. The chondrite-normalized rare earth element (REE) abundances are even closer to chondrite than
N-MORB and, except for Sample 149-900A-81R-1, 96-101 cm with an N-MORB REE pattern, all samples have similar flat REE patterns near 2X-4X chondrite and positive europium
(Eu) anomalies (Fig. 9). The positive Eu anomalies can be attributed to Eu in the cumulus plagioclase. The total range of REE values in the Hole 900A samples is small, La varies from 0.47 to 1.03, with all normalized values between 1X to 10X
chondrite, and similar to cumulate gabbros from 26°N along the Mid-Atlantic Ridge
(Tiezzi and Scott, 1980). However, Sample 149-900A-81R-1, 96-101 cm has a REE pattern distinct from the other Hole 900A metamorphosed mafics with heavy rare earth element
(HREE) values near 10X chondrite and a strong light rare earth element (LREE) depletion
(Fig. 9). Its REE pattern is similar to
N-MORB basalts from normal segments of the Mid-Atlantic Ridge (Schilling et al., 1983), which indicates retention of considerably more magma than the typical Hole 900A cumulate gabbro and suggests a MORB parental magma.
The typical Hole 900A metamorphosed mafic rock spider diagram
(Fig. 10) supports the REE plot analysis indicating a cumulate origin. Furthermore the spider diagram suggests that the large ion lithophile elements
(LILE) K, Rb, Ba, and Sr have been enriched by hydrous solutions. The LILE, REE, and high field strength elements
(HFSE) represent incompatible element groups that travel together in magma and should be expected to increase or decrease together producing smooth spider diagram patterns
(Fig. 11). If the LILE have a greater or lesser abundance than the other incompatible element groups, hydrous solutions are most likely the transporting medium, rather than magma, because the LILE are far more soluble in hydrous fluids than the REE and
HFSE. Spider diagrams include some of all three incompatible element groups. The less mobile REE and HFSE in the Hole 900A rocks have an abundance lower than that of any basalt, even
N-MORB, but typical of cumulate gabbros from 26°N on the Mid-Atlantic Ridge
(Tiezzi and Scott, 1980). The low abundance of the less mobile elements contrasts with the high abundance of the more mobile LILE which have anomalously high peaks on the typical Hole 900A spider diagram
(Fig. 10). The elements which appear to be enriched on the spider diagram are K,
Rb, Ba, Sr, U, and to a lesser extent Eu, although the K, Sr, and Eu peaks are probably enlarged because of the plagioclase-rich nature of these cumulates. The relatively low abundance of the immobile REE and HFSE indicates that little magma was retained by the Hole 900A cumulate gabbros.
Three large powdered samples were prepared for isotopic analysis by combining the smaller Samples 149-900A-81R-1, 96-101 cm and 900A-81R-2, 82-87 cm into Sample 81, Samples 149-900A-82R-1, 128-134 cm and 900A-82R-2, 106-111 cm into Sample 82, and Samples 149-900A-85R-1, 26-30 cm and 900A-85R-5, 51-56 cm into Sample 85. The
Sr, Nd, and Pb isotopic ratios and Rb, Sr, Sm, Nd, Pb, and U elemental concentrations for these three Hole 900A metamorphosed mafic rocks are given in
Table 4. All of the Nd and Pb isotopic ratios, except 207Pb/204Pb, fall within oceanic fields similar to
MORB. The Sr isotopic ratios are far too high and reflect seawater alteration (Fig.
12, Fig. 13). The generally high and variable
U/Pb ratio, 0.10 to 0.81, and the generally low and variable Th/U ratio, 0.02 to 1.0, of the Hole 900A mafic rocks indicate secondary addition of highly soluble U from seawater
(Michard and Albarede, 1985; Chen et al., 1986). This enrichment of U was revealed previously by the typical Hole 900A spider diagram
(Fig. 10).
Our attempt to date the Hole 900A metamorphic mafic rocks by 40Ar/39Ar was unsuccessful, but an 40Ar/39Ar age of 136.4 Ma has been obtained by Feraud et al. (this volume). Consequently all isotopic ratios have been corrected for in situ decay of 87Rb, 147Sm, 235U, and 238U using that age value for Hole 900A rocks. The relatively high 143Nd/144Nd ratios correspond to initial positive
Nd values greater than six at 136.4 Ma, which indicate a MORB source
(Fig. 12). The
initial lead isotope ratios appear to allow an OIB origin, but have been displaced because of seawater alteration
(Fig. 13). The large correction required for in situ decay for the Hole 900A rocks is caused by high
µ (Tatsumoto, 1978) for these rocks. Both the higher 87Sr/86Sr ratios (>.7045) and the high 238U/204Pb ratios (29.7 to 69.4) indicate seawater alteration. Alteration of the Sr isotopics without significant resetting of the Nd isotopics limits the water/rock ratio during seawater alteration to less than 108, and suggests a value closer to 105 (McCulloch et al., 1980). The 18O/16O composition of the three Hole 900A samples lie between 4.22 and 3.48 per mil, while the typical magmatic values are at 5.5 per mil; again, this clearly reflects seafloor alteration.
