Site 1152 basalts were recovered from ~25-Ma seafloor formed within spreading Segment B4 of the AAD. Four whole-rock powders were analyzed for major and trace elements by XRF and ICP-AES and three samples of fresh glass chips by ICP-AES only. The results are shown in Table T5. At this early stage in Leg 187, the ICP-AES results were poor for some elements because of dissolution problems and inexperience with the new JY2000 instrument. The major element data are the most compromised; SiO2 could only be estimated by difference from 100% total oxides. This problem affects only the whole-rock ICP-AES analyses in Table T5; the glasses from Hole 1152 were reanalyzed during later ICP-AES runs. Only the XRF whole-rock data and ICP-AES glass data are shown on plots and considered in the discussion.
Samples from Hole 1152A are assigned to a single aphyric basalt unit based on macroscopic and microscopic examination. These basalts are moderately evolved (i.e., <7.5 wt% MgO). The glass from Hole 1152A is distinctly higher in MgO (7.3 wt%) than the whole rocks (6.0-6.5 wt%; Fig. F12). The other major and trace element concentrations are offset from the glass compositions in directions that are generally orthogonal to expected low-pressure crystal fractionation trends. For example, whole-rock Al2O3 is higher for a given MgO content relative to associated glass rather than lower as removal of plagioclase and olivine would dictate. In addition, Cr, Zr, and Y contents of whole rocks are offset both from the glasses and from melt evolution trends defined by younger Segment B4 glass samples (Fig. F13). These observations suggest that alteration, particularly the selective removal of MgO and/or nonequilibrium differentiation, have affected the whole-rock compositions (see "Alteration").
Samples from Hole 1152B are assigned to two lithologic units: Unit 1 is an aphyric pillow basalt, and Unit 2 is a sparsely to moderately plagioclase-clinopyroxene phyric pillow basalt. The glasses from both units (undifferentiated in Figs. F12 and F13) are higher in MgO (7.6-8.0 wt%) than the whole rocks (6.2-6.9 wt%). The plagioclase-clinopyroxene phyric Unit 2 basalt has higher Al2O3, CaO (Fig. F12), and Sr (Fig. F13) contents than the Unit 1 aphyric basalt, consistent with its phenocryst assemblage, which suggests plagioclase accumulation. Hole 1152B glasses appear to be genetically related by a typical low-pressure crystal fractionation path of decreasing Al2O3, Cr, and Ni and increasing Fe2O3, Na2O, TiO2, Zr, and Y. Glass compositions from Holes 1152A and 1152B are distinctly different on the CaO/Al2O3 vs. MgO diagram, with those from Hole 1152B having higher CaO/Al2O3 values for a given MgO content than those from Hole 1152A (Fig. F12). As for samples from Hole 1152A, the relationships between glass and whole rocks from each unit cannot be attributed to simple low-pressure fractional crystallization and are more likely controlled by alteration of the whole rock and/or nonequilibrium differentiation, perhaps including variable amounts of plagioclase accumulation.
Whole rocks from Site 1152 lie beyond the more evolved end of the compositional range of 0- to 5-Ma mid-ocean-ridge basalt (MORB) from Segment B4. The glasses have MgO contents comparable to those of younger B4 glasses; however, they differ by having significantly lower Al2O3, marginally lower Ba, and higher TiO2, Fe2O3, Zr, and CaO/Al2O3 (Hole 1152B only) for a given MgO content. The shifts in some of these geochemical characteristics could indicate a temporal change in melting parameters beneath the Segment B4 spreading axis. Additional support for this can be seen in the off-axis dredge compositions, particularly the generally higher CaO/Al2O3 values relative to B4 axial lavas. Higher Fe2O3 and CaO/Al2O3 values and lower Ba content suggest a higher degree of melting beneath this segment in the past. This would be consistent with the inference that Segment B4 was shallower at 25 Ma than it is today because it was located west of the depth anomaly at that time. The mantle source must also differ in composition, mineralogy, and/or melt extraction process to explain the higher concentrations of moderately incompatible elements TiO2, Zr, and Y, since a simple increase in melting would be expected to lower the concentrations of these elements.
Site 1152 was selected as our initial site to provide a baseline Indian-type composition at ~30 Ma because no evidence exists to suggest Pacific-type mantle was ever west of the B5 spreading axis. This may prove to have been an erroneous assumption. The Zr/Ba systematics of Site 1152 basalts do not unequivocally assign this site to either the Indian or the Pacific mantle domain. All three glasses plot within a small region that lies outside our predefined Indian and Pacific fields (Fig. F14A); they lie on the Pacific side of the trend defined by present-day transitional lavas from Segment B5 at lower Zr/Ba ratios than the predefined Pacific field. The whole-rock samples have more Ba and consequently lower Zr/Ba ratios, which could result from excess Ba introduced by seawater alteration, although results from subsequent sites suggest that there are consistent trends defined by spatially related glasses and whole rocks that cannot be explained solely by alteration.
The Na2O/TiO2 vs. MgO diagram (Fig. F14B), taken at face value, indicates a Pacific-type MORB source. On balance, it appears from the available data that Site 1152 lavas have at least some degree of Pacific affinity, if not truly a Pacific-type source. Little evidence, other than our initial expectations, suggests an Indian-type MORB source. As a general rule, where these two diagrams conflict, we consider that the Zr/Ba vs. Ba systematics are more reliable in determining source characteristics, as these trace elements are much less likely than TiO2 and Na2O to have been affected by crustal processes (e.g., boundary-layer fractionation) (Langmuir, 1989; Nielsen and Delong, 1992). We also recognize that, although Zr/Ba is demonstrably more reliable than Na2O/TiO2 for axial lavas, it is not unequivocal. Our shallow drill holes are sampling only the very last volcanic events before crust was transported off axis, and we may even be sampling off-axis magmatism. Lavas from such eruptions have been shown to be more diverse in some regions (Reynolds et al., 1992) and might be expected to have higher TiO2 and, therefore, lower Na2O/TiO2 ratios, giving the impression of a Pacific-type affinity. The mantle domain for this site will not be resolved without isotopic data.