Site 1156 basalts were recovered from two holes that sampled ~22-Ma crust formed within Segment B5 of the Australian Antarctic Discordance (AAD), ~87 km south of Site 1155. Five whole-rock powders were analyzed for major and trace elements by X-ray fluorescence (XRF) and inductively coupled plasma-atomic emission spectrometry (ICP-AES), and three glasses were analyzed by ICP-AES only. Three lithologic units, two in Hole 1156A and one in Hole 1156B, are represented by glass and whole-rock analyses (see Table T3). ICP-AES Ni and Cr results are consistently lower than those from XRF. There is also a significant analytical discrepancy in MgO and K2O for Sample 187-1156A-2R-3 (Piece 17, 129-133 cm). The reason for the latter discrepancies is unknown; however, we conclude that the ICP-AES analysis is aberrant because the MgO and K2O contents for Sample 187-1156A-2R-3 (Piece 17) differ from those in the remaining three Unit 1 whole-rock compositions.
Hole 1156A basalts are derived from two lithologic units, based on macroscopic and microscopic examination (see "Igneous Petrology"). Unit 1 is interpreted as a talus breccia composed of basalt clasts and palagonitized glass in a matrix of micritic limestone. Unit 2 is a moderately to sparsely plagioclase-olivine phyric pillow basalt underlying Unit 1. One glass and two whole rocks from Unit 1, as well as one glass and one whole rock from Unit 2 were analyzed, and the results cover a substantial MgO range (6.5-8.3 wt%). As observed at previous sites, the glasses have higher MgO contents than the associated whole rocks, although the difference is only ~0.5 wt% in most cases. The two units differ in their whole-rock compositions; however, the glasses from both units are virtually identical in composition. Glass within the Unit 1 breccia may, therefore, have been derived from Unit 2 pillow rims. Unit 1 and Unit 2 whole rocks do not lie on simple fractional crystallization trends that stem from the Hole 1156A glasses. For example, Fe2O3 decreases with decreasing MgO; Al2O3, CaO, and Ni contents are high in Unit 2, consistent with the observed presence of plagioclase and olivine phenocrysts. Unit 1 whole rocks are similar to the glasses, but Fe2O3 and Na2O contents should be higher in these samples if they were related by simple low-pressure crystal fractionation to a parental magma represented by Hole 1156A glass compositions.
Hole 1156B basalts are assigned to a single lithologic unit of moderately to highly plagioclase-olivine phyric pillow basalt with calcareous sediment locally filling interpillow spaces (see "Igneous Petrology"). The Hole 1156B glass is compositionally the same as Hole 1156A glasses within analytical uncertainties. Two different whole-rock compositions are evident in Hole 1156B: one similar to Hole 1156A (Unit 1) and one that could easily be related to Hole 1156A (Unit 2) by simple low-pressure crystal fractionation (i.e., decreasing Al2O3, CaO, CaO/Al2O3, Ni, and Cr and increasing Na2O, Fe2O3, Sr, and Ba).
The compositional characteristics of Site 1156 lavas are compared with those of the present Segment B5 axis and off-axis lavas in Figures F21 and F22. Site 1156 glasses lie at the high-MgO end of the range for 0- to 7-Ma AAD mid-ocean-ridge basalts (MORBs). For most elements, Site 1156 glasses overlap younger Segment B5 compositions. Notably, however, Na2O content is slightly lower in Site 1156 glasses and appreciably lower in whole rocks relative to younger lavas. The Na2O variations in the whole rocks are orthogonal to those observed in the glass compositions and decrease with decreasing MgO. All other trends appear reasonably similar to younger Segment B5 glasses. Progressive loss of MgO from the whole rocks through alteration would be the simplest explanation for this unusual Na2O variability.
Overall, no clear difference in source conditions (i.e., melting or composition) can be determined between Site 1156 and the near-axis Segment B5 compositions. In contrast, comparison with the data from Site 1155 (see "Geochemistry"in the "Site 1155" chapter), which is also located in Segment B5, shows that the Site 1156 glasses have lower Na2O, slightly lower TiO2 and Fe2O3, as well as higher Al2O3 and CaO for a given MgO content. Therefore, melting conditions and/or the source composition that existed beneath the Southeast Indian Ridge (SEIR) when Site 1155 lavas erupted must have changed to the parameters that now exist beneath 0- to 7-Ma crust by the time Site 1156 lavas erupted (i.e., over a period of ~3 m.y.).
The Zr/Ba systematics of Site 1156 suggest an Indian-type to transitional mantle source. Glass and whole-rock compositions lie well within the range of known Indian-type MORB glasses from the SEIR (Fig. F23A). They are very similar to the Segment B5 axial samples, which are transitional between present-day Pacific type (eastern Segment B5) and Indian type (western Segment B5). As in Hole 1155B (see Fig. F23), the glasses and associated whole rocks define a coherent negative Zr/Ba vs. Ba trend. Site 1156 and Hole 1155B lavas have the same Zr/Ba range, but Site 1156 compositions have slightly higher Ba contents at a given Zr/Ba, suggesting a slightly greater contribution of Pacific-type mantle to 1156 lavas.
Similar conclusions can be drawn from the Na2O/TiO2 vs. MgO diagram (Fig. F23B). Most Site 1156 lavas have a slightly lower Na2O/TiO2 value for a given MgO content relative to the Segment B5 axial lavas, which could indicate a slightly greater Pacific-type mantle contribution. Site 1156 whole rocks have higher Na2O/TiO2 relative to their associated glasses, similar to the whole rocks from Site 1155. They overlap the Pacific-type Segment B5 on-axis mixing line, shown by the gray shaded line (Fig. F23B), but in contrast show little variation in Na2O/TiO2 with decreasing MgO. Although Na2O/TiO2 suggests a Pacific-type parentage for Site 1156 lavas, this suggestion would be more convincing if these samples had either higher Ba contents and/or higher Zr/Ba values. Note that lavas with Zr/Ba >14 at Ba <10 ppm can be Pacific type, transitional type, or Indian type but that no 0- to 7-Ma Pacific-type lavas have Zr/Ba <14 in this Ba-concentration range. Based on their low Zr/Ba and low Ba content, we can best conclude that the basalts from Site 1156 are derived from Indian-type mantle with some tendency toward transitional characteristics.