We performed chemical analyses on 16 harzburgites, 4 dunites, and 7 gabbros, selected by the shipboard scientific party, using inductively coupled plasma–atomic emission spectrometry (ICP-AES) for determining major and trace element concentrations and gas chromatography for H2O, CO2, and sulfur. These 27 samples are representative of the highly altered material recovered from Hole 1268A (see "Igneous and Mantle Petrology" and "Metamorphic Petrology"). The results for the major and trace elements are reported on a volatile-free basis for both ultramafic and mafic rocks (Table T5). In order to determine the optimum analysis conditions for elements at low concentration in ultramafic rocks (Ti, Al, K, Na, Sr, V, Cu, Co, Ni, Cr, Sc, Y, and Zr), the first six samples were analyzed at three different dilutions (4000-, 2000-, and 1000-fold). Although detection limits and precision for elements present in low abundance should have been improved at lower dilution factors, instrument performance and stability degraded significantly. This compromised our ability to efficiently perform high-quality shipboard measurements at the higher analyte concentrations present in the 1000-fold and 2000-fold dilutions. Maximum analytical efficiency was achieved by using shorter elemental menus for each run, operating at 4000-fold dilution, and running major and trace elements as separate analytical routines.
Bulk-rock analyses of peridotites from Site 1268 show that their compositions were significantly modified by hydrothermal alteration, leading to the addition of variable amounts of volatile constituents. The ultramafic rocks with predominant serpentine alteration, based on the visual core descriptions and XRD results (see "Metamorphic Petrology"), display the highest loss on ignition ([LOI] = 11.8–13.1 wt%). The five ultramafic samples showing evidence for a further stage of alteration to talc display significantly lower LOI values (4.7–6.3 wt%). LOI is correlated with water content in the peridotites, with H2O contents of 12.6–15.2 wt% where serpentine is predominant and 4.3–7.1 wt% where talc is predominant. LOI slightly underestimates the total volatile content because of the conversion of FeO to Fe2O3 during heating of the sample powders to ~1000°C (see "Geochemistry," in the "Explanatory Notes" chapter).
Talc-altered rocks have CO2 of <0.04–0.11 wt%, whereas the serpentine-altered rocks show a CO2 range of 0.05–0.20 wt%. In the peridotites that have not been altered to talc, elevated CO2 tends to occur in samples with elevated H2O, leading to a positive correlation between CO2 and H2O. The overall scatter, however, precludes making any conclusions about the origin of CO2 variations in these rocks. Sulfur contents in the peridotites range 1,100–16,000 ppm (Table T5). Based on our initial results, there does not seem to be a systematic covariation between sulfur and other elements.
Bulk rock compositions primarily reflect the proportions of minerals present. Pervasively serpentinized peridotites are composed of SiO2 (44–50 wt%), MgO (40–45 wt%), and Fe2O3 (6.7–9.5 wt%). These values correspond to the compositions of serpentinites from Leg 153 Hole 920B (Dilek et al., 1997a). The Site 1268 peridotites that are altered to talc are significantly enriched in SiO2 (60–66 wt%) and depleted in MgO (30–32 wt%) and Fe2O3 (3.7–8.2 wt%) relative to serpentinized peridotites at Site 1268. The composition of these five rocks approaches talc end-member compositions in SiO2, Fe2O3, MgO, and H2O (Fig. F80). Site 1268 peridotites are characterized by high and variable Mg# (molar Mg/[Mg + Fe]) (89%–94%). Although some of Site 1268 peridotites have Mg#s close to those of tectonically emplaced peridotites and abyssal peridotites (Fig. F81), others with high Mg# probably reflect depletions in iron relative to magnesium during alteration (see "Hydrothermal Alteration" in "Metamorphic Petrology" for more details about hydrothermal alteration mineralogy). Notably, all the peridotites are strongly depleted in CaO (relative to Al2O3), suggesting that serpentinization at Site 1268 removed almost all of the calcium that was originally present in the peridotite protolith.
Site 1268 peridotites are depleted in trace elements that are considered to be mobile during alteration, such as Sr (<5 ppm) relative to peridotites from Leg 109 Site 670 and Leg 153 Site 920. Like calcium, these elements appear to have been leached from the peridotites during alteration. The Site 1268 peridotites are highly depleted in TiO2 (<0.01 wt%), Y (<2 ppm), and Zr (<2–8 ppm) and, to a lesser extent, V (11–38 ppm) and Sc (<4–9 ppm) compared to other Mid-Atlantic Ridge peridotites. They display high values in Cr (1920–3850 ppm) and Ni (2200–3500 ppm). All these elements are considered to be relatively immobile during alteration (e.g., Hébert et al., 1990). Therefore, they can be used as indicators of protolith composition, although it should be noted that the talc-altered ultramafic samples have the lowest Cr (830–1100 ppm) and Ni (1750–2200 ppm) concentrations, probably indicating that leaching of some of these "immobile" elements occurs during talc alteration. V, Sc, and Cr variations and, to a lesser extent, Zr, Ti, and Y variations correlate with Al2O3 contents (Fig. F82). Zr, V, Sc, Ti, and Y behave as moderately incompatible elements (preferentially partitioning into the liquid during partial melting) whereas Cr and Ni behave as compatible elements. Al2O3 is mainly concentrated in pyroxenes and may be used as an indicator of the protolith pyroxene content. The trace element composition of Site 1268 peridotites suggests that the protolith was composed of depleted, highly refractory peridotite, except for Sample 209-1268A-13R-1, 33–37 cm. This sample has higher TiO2 (0.22 wt%), Y (9 ppm), and Zr (50 ppm) concentrations, suggesting that this peridotite protolith had a higher pyroxene fraction and was less depleted than other peridotites from Hole 1268A, although the visual core and thin section descriptions do not show evidence of former lithologic variations in the altered peridotites at this location.
We analyzed seven gabbro samples from Site 1268. Sample 209-1268A-16R-4, 9–13 cm, is a gabbro from Unit III (see "Lithology and Stratigraphy" in "Igneous and Mantle Petrology"), where gabbro is subordinate (9%) to harzburgite and dunite. Six additional samples were analyzed from Unit IV (104.4–147.6 mbsf), where gabbro and gabbronorite are dominant. In this latter group, the four lowermost samples are variably textured gabbronorites, one from Subunit IVA above an ultramafic portion of the core in Subunit IVB and the three others from Subunit IVC.
With the exception of two gabbros from the upper part of Unit IV (Samples 209-1286A-21R-1, 31–35 cm, and 21R-2, 21–24 cm), the analyzed gabbroic rocks from Site 1268 are relatively uniform in major and trace element contents of SiO2 (49.9–51.4 wt%), MgO (10.5–12.7 wt%), and CaO (8–13 wt%) (Fig. F83). In contrast, the two upper gabbros from Unit IV have higher MgO concentrations (20.8–22.3 wt%) and Al2O3 contents (19–23 wt%) but lower CaO (~2 wt%). In addition, they have high water contents (11.13–13.10 wt%) and LOI values (8.80–9.30 wt%) compared to other Site 1268 gabbros. These high values are associated with their high degree of alteration, as revealed by visual core and thin section description (see "Metamorphic Petrology"). Sample 209-1268A-16R-4, 9–13 cm, from Unit III appears less altered, yet it has also high water content (4.31 wt%) and LOI (3.89 wt%). H2O in Unit IV gabbronorite reaches the lowest concentrations observed at Site 1268 (2.8–4.1 wt%). CO2 in the gabbro shows limited variability (0.07–0.11 wt%), and there is no covariation with H2O content. Sulfur in the gabbroic rocks from this section is below the limit of detection (~400 ppm).
The variation in major element geochemistry can be related to the alteration observed within the gabbros of Site 1268. The two uppermost samples of Unit IV (Samples 209-1268A-21R-1, 31–35 cm, and 21R-2, 21–24 cm) have undergone alteration of plagioclase and clinopyroxene, producing significant amounts of chlorite and leading to relative increases in MgO and H2O and a decrease CaO. Significantly, the possible removal of calcium from the gabbros during serpentinization suggests that the calcium from the ultramafic rocks is not being redistributed within the system but is almost completely removed. The remaining gabbro samples from both Units III and IV appear to be significantly less altered (see "Metamorphic Petrology"), even though they have higher LOI values than the variety of gabbros recovered during Leg 153 (Fig. F83).
The trace element geochemistry of the Site 1268 gabbros shows that, like the peridotites, the gabbros are depleted in mobile elements such as Sr (~120 ppm). The two gabbros from the top of Unit IV that have undergone plagioclase alteration contain lower Sr (50–65 ppm) compared to all but one of the Leg 153 gabbros (100–400 ppm) (Agar et al., 1997). As discussed previously, the immobile trace elements such as Ni and Cr may still yield information about the original protolith (e.g., Hébert et al., 1990). When compared to gabbros from Leg 153, the Site 1268 gabbros are depleted in Cr (less than the detection limit) and other immobile compatible elements, illustrating that the original gabbro protolith was probably more evolved than the comparable compilation of gabbros recovered during Leg 153.
The compositional variations in major element, trace element, and volatile contents of the Site 1268 peridotites are controlled by the style of alteration. Where visual and XRD evidence (see "Metamorphic Petrology;" Table T2) points to serpentine as the dominant alteration phase, the serpentinized harzburgites and dunites approach the composition of a pure serpentine end-member (Fig. F80). Similarly, where such evidence points to the dominant presence of talc, the serpentinized harzburgites and dunites approach the composition of pure talc. In comparison to other localities where significant serpentinite has been recovered, such as Sites 670 and 920, the alteration in Hole 1268A peridotites is even more extensive. At Site 1268, all the geochemical evidence, as well as evidence from the visual core and thin section descriptions of the petrology groups (see the "Supplementary Material" contents list), points to a marked silica metasomatism at this locality. This style of alteration appears to be characterized by strong open-system behavior for key elements, leading to a notable loss of Ca and perhaps also Sr.
Although it is possible that rodingitization and sequestering of Ca occurred below the depth sampled at the bottom of Hole 1268A, it seems more likely that alteration led to nearly complete removal of Ca by hydrothermal fluids that were ultimately vented into the deep ocean. This is consistent with the observation that areas actively undergoing serpentinization, such as the Rainbow and Logatchev hydrothermal systems on the Mid-Atlantic Ridge, discharge hydrothermal vent fluids having the highest Ca/Cl ratios observed along the global mid-ocean-ridge system (Charlou et al., 2002; Douville et al., 2002).
Some of the gabbroic rocks at Site 1268 are also highly altered, especially the two rocks analyzed from the top of Unit IV. These two rocks may show significant Ca loss, together with possible Mg uptake, and one of them (Sample 209-1268A-21R-1, 21–24 cm) shows significant depletion in SiO2 (down to 42.6 wt%), consistent with metasomatic alteration to chlorite. At deeper levels in the section, the extent of alteration in the gabbroic rocks is significantly lower and the rocks appear to have retained much of their original chemical composition.
The Fe2O3 and H2O concentrations measured on individual samples were compared with magnetic susceptibility determined by multisensor track (MST) (see "Magnetic Susceptibility" in "Physical Properties") for the same depth level at Site 1268 (Fig. F84). As expected for their Fe2O3 concentrations of 6–8.5 wt%, the gabbroic rocks show uniformly low magnetic susceptibilities (between 32 x 10–5 and 50 x 10–5 SI) and there is no correlation with Fe2O3. However, there is a scattered positive correlation between Fe2O3 concentration and magnetic susceptibility in the peridotites. This trend is largely related to the style of alteration. The peridotites with the lowest Fe2O3 and H2O contents, indicative of their pervasive alteration to talc, have the lowest magnetic susceptibilities. Some of the serpentinized ultramafic rocks with high Fe2O3 and H2O have high magnetic susceptibilities. However, within this high-H2O group, samples with the highest H2O contents have intermediate Fe2O3 and lower magnetic susceptibilities, so a 1:1 correlation between magnetic susceptibility and any simple combination of chemical indicators appears to be absent.
The chemical composition of the Site 1268 peridotites has been almost entirely modified by pervasive alteration. Previous studies of altered peridotites from Sites 670 and 920 along the Mid-Atlantic Ridge revealed that the original protolith can potentially be characterized through bulk-rock chemistry, when elements that are relatively immobile during alteration are evaluated. The trace element composition of Site 1268 ultramafic rocks suggests that the protolith was composed of depleted and highly refractory peridotites. The low chromium contents of gabbros suggest that the original protolith was depleted in chromium and was therefore more evolved than the gabbros recovered during Leg 153.