Site 1155 basalts were recovered from two holes that sampled ~25-Ma crust formed within Segment B5 of the Australian Antarctic Discordance. Seven whole-rock powders were analyzed for major and trace elements by X-ray fluorescence (XRF) and inductively coupled plasma-atomic emission spectrometry (ICP-AES); two glasses and one plagioclase separate were analyzed by ICP-AES only. The results are shown in Table T3. There are few discrepancies between XRF and ICP-AES analyses, except that Na2O content is generally higher and Ni content is generally lower by ICP-AES.
Samples from Hole 1155A are assigned to two lithologic units based on macroscopic and microscopic examination (see "Igneous Petrology"). Unit 1 is a sparsely to moderately plagioclase ± olivine phyric basalt pillow lava flow. Unit 2 consists of aphyric basalt pebble and cobble fragments interpreted as unconsolidated rubble. Unit 1 is represented by a single whole-rock analysis, as there was insufficient glass for ICP-AES analysis. From Unit 2, one whole-rock and one glass sample were analyzed (Table T3). The Hole 1155A Unit 2 glass is higher in MgO (7.8 wt%) than both Unit 2 and Unit 1 whole rocks (Fig. F23). The Unit 1 whole rock (6.9-7.1 wt% MgO) is slightly more evolved than that of Unit 2 (7.2-7.3 wt% MgO). Higher TiO2, Zr, and Ni and lower Cr and Sr in Unit 1 suggest that Units 1 and 2 are not related by simple crystal fractionation (Fig. F24). The Unit 2 glass is also higher in Fe2O3, Na2O, and Ba and lower in Al2O3, CaO, Sr, and Ni than whole rocks from both units, indicating that they are also not related by simple low-pressure crystal fractionation. Similarities in TiO2, Zr, Y, and Cr contents suggest that Unit 2 glass is more like Unit 1 whole rocks than its spatially related Unit 2 whole rocks. Given that the glass sample from Unit 2 was a small pebble, it is possible that this piece fell into the hole from Unit 1 above. This is consistent with our observation, based on larger pieces (e.g., Section 187-1155A-5R-1 [Piece 3]) that such fall-in has occurred in this hole (see "Unit 2" in "Igneous Petrology").
Hole 1155B samples are interpreted as derived from a single, intact pillow lava sequence of moderately plagioclase ± olivine phyric basalts. One glass and a plagioclase separate from that glass were analyzed by ICP-AES. Five whole rocks were analyzed by ICP-AES and XRF. The glass is primitive in composition and is higher in MgO (9.3 wt%) than associated whole rocks (7.9-8.5 wt% MgO). The whole rocks form a coherent group that shows evidence of simple crystal fractionation, but they are not a product of differentiation from the more primitive Hole 1155B glass composition. The glass is higher in Fe2O3 and Na2O and lower in CaO, Al2O3, and Ni than Hole 1155B whole rocks, consistent with the observed phenocryst assemblage (i.e., plagioclase and olivine accumulation). The glass is also slightly higher in Zr, Y, and TiO2. As in samples from Hole 1155A and the majority of holes sampled, glasses and whole rocks recovered within a single hole usually are not easily related by simple crystal fractionation. In most cases, it appears that MgO has been selectively removed from the whole rock samples by low-temperature alteration. Other processes including mineral accumulation and nonequilibrium fractionation may also have affected the whole-rock compositions.
The plagioclase separate and Sample 187-1155B-2R-1, 125-129 cm, a highly unusual low-MgO basaltic sample, are not shown in Figures F23 and F24 because they lie out of the typical basalt range and therefore outside the figures. This basalt differs from other whole rocks by its extremely low MgO content and high Al2O3, CaO, Fe2O3, Sr, and, to some extent, Na2O. There is no evidence of extreme plagioclase accumulation or of excessive loss on ignition (LOI) values, which typically indicate intense alteration. However, thin-section examination reveals a groundmass extensively replaced by clay and Fe oxyhydroxide; therefore, this composition most likely, but not obviously, represents an extreme alteration effect, despite the low LOI (see "Alteration"). There is little evidence to suggest this composition is a product of magma evolution. Surprisingly, its Ba is indistinguishable from "unaltered" samples.
Site 1155 is the first site of Leg 187 within Segment B5. At present, Segment B5 lies toward the eastern side of the depth anomaly; it displays much of the range of axial morphology mapped in this area, and its western transform boundary is the current position of the Indian-Pacific isotopic boundary. For some elements, Site 1155 basalts show the same major and trace element variations seen on or near the B5 spreading axis (Figs. F23, F24). However, Site 1155 glasses have distinctly higher TiO2, Fe2O3, Zr, and Y and lower Al2O3 for a given MgO content. The only apparent evidence of a significantly different melt regime at 25 Ma is high Fe2O3, which, in the absence of other compositional evidence, may imply a deeper mean melting depth (Langmuir et al., 1993). More likely, the compositional variations between Site 1155 lavas and present Segment B5 axis lavas are caused by variations in mantle composition; cooler or warmer mantle melting conditions are less likely because other melting indicators, such as CaO/Al2O3 and Na2O, are the same as for younger dredge samples.
The Zr/Ba systematics of Site 1155 (Fig. F25A) suggest that both Indian- and Transitional-Pacific-type mantle were present beneath this site when it was at the spreading axis. The Hole 1155A Unit 2 glass and the Hole 1155A Unit 1 whole rock have relatively high Zr/Ba and low Na2O/TiO2 values (Fig. F25B), which indicate a Transitional-Pacific-type source (see "Barium and Zirconium" in "Geochemistry" in the "Leg Summary" chapter). On the other hand, Hole 1155B glass and whole rocks, as well as Hole 1155A Unit 2 whole rock, lie well within the range of Zr/Ba vs. Ba variations for Indian-type mantle; these samples lie along a low-Ba trend similar to that defined by Segment B5 axial samples that span the Indian- to Pacific-type ranges.
To test whether plagioclase accumulation could account for the coherent grouping of glasses and whole rocks along curved trends on the Zr/Ba vs. Ba diagram, a plagioclase separate was handpicked from Hole 1155B glass and analyzed by ICP-AES. The position of the plagioclase in Figure F25A clearly demonstrates that plagioclase accumulation cannot drive the negative Zr/Ba vs. Ba variations. In fact, this plagioclase is low in Ba content, and varying its abundance in small proportions would have little effect on the observed trends. We also investigated whether simple crystal fractionation could, for instance, move compositions from Hole 1155B glass to Hole 1155A glass on the Zr/Ba vs. Ba diagram. An arrow in Figure F25A shows ~50% fractional crystallization at low pressure. The trajectory of the trend is clearly toward Hole 1155A glass. However, to achieve this, MgO would be required to decrease to ~5.1 wt%, much lower than any Hole 1155A composition. Therefore, observed variations in Zr/Ba vs. Ba are not caused by simple low-pressure crystal fractionation or plagioclase accumulation.
Neither simple melting nor simple mixing variations can fully account for the difference between Hole 1155A and Hole 1155B compositions. Simple batch melting within either a Pacific- or Indian-type source moves values up and down along negative curves. At least two "enriched" end-members, an Indian type and one similar to propagating rift tip (PRT) lavas, and one or two "depleted" end-members are required to produce the curves. Varying the proportions of Pacific- and Indian-type source melts moves compositions between the two trends (e.g., varying the Ba composition of the source). For example, in the transitional region between the Pacific- and Indian-type fields (i.e., Zr/Ba = 10-15 and Ba = 8-15 ppm), as Ba increases, we expect mixing to shift toward a curve that includes the Pacific-type field and the Zone A PRT lavas; therefore, the rock would be of a Pacific-type affinity or, as we have defined, Transitional-Pacific type. The Indian-type curved field is offset toward lower Ba content. The way the Indian- and Pacific-type fields overlap one another in Zr/Ba at distinct Ba concentrations demonstrates that subtle shifts in the Ba concentration of mantle sources affect the position of mixing and melting curves on this diagram and therefore our designation of the mantle domain from which the basalts derived.