PRINCIPAL RESULTS

In the seismic-reflection record for this site (Fig. 16), a sedimentary megasequence laps onto the upper surface of the fault block that forms acoustic basement. Basement reflection character differs from that of basaltic basement on the main plateau (e.g., Site 1183;Fig. 7). Parallel to subparallel, high-frequency, slightly dipping reflections of limited and variable continuity persist to depths as great as 1.0 s of two-way (P-wave) traveltime beneath the surface of the fault block.
We cored early Miocene pelagic calcareous ooze and chalk (Unit I) from 134.4 mbsf to acoustic basement at 201.1 mbsf and volcaniclastic rocks (Unit II) from 201.1 mbsf to the base of the hole at 538.8 mbsf. A 1-cm-thick ferromanganese oxide crust represents the contact between the two units. Paleontological data suggest that deposition of the volcaniclastic succession occurred during the middle Eocene (principally nannofossil Zone NP16) and that deposition of the calcareous ooze began during the earliest Miocene. Little, if any, sedimentary record of events during the late Eocene or Oligocene is preserved.

Unit I is dominated by nannofossil foraminifer ooze with as much as 10% siliceous microfossils; volcanic ash is a minor component. Paleodepths appear to have been bathyal. Grain densities generally lie between 2.3 and 2.6 g/cm³, with a mean of 2.5 g/cm³; porosity averages 66.1%, and the mean bulk density is 1.5 g/cm³. The ooze is weakly magnetic and was badly disturbed by drilling; consequently, we were unable to obtain reliable paleomagnetic data. The basement volcaniclastic sequence of Unit II consists of coarse lithic vitric tuff, lapilli tuff, and lapillistone, most of which have a massive texture. Several thin beds of fine ash are also present, but we recovered no pelagic or neritic interbeds. Grain densities in Unit II are significantly more variable than those in Unit I, with a mean of 2.4 g/cm³; bulk densities maintain a nearly constant value of ~1.9 g/cm³, and porosities cluster between 31% and 37%. This unit exhibits normal polarity magnetization and a continuous record of paleosecular variation. The mean inclination (-54°) is much steeper than the expected Eocene inclination and indicates a paleolatitude (35°S) significantly different from that expected for this area in the Eocene (~15°-20°S). Tectonic rotation of the volcaniclastic beds may have taken place after the magnetic remanence was acquired, but a sufficiently large amount of rotation in the direction required appears unlikely to have occurred.

We divided Unit II into five subunits on the basis of changes in grain size, sorting, and sedimentary structures (Fig. 17; Table 2). Wood fragments (Fig. 18) and organic-rich layers were found at the boundaries between four of the subunits (B, C, D and E) and at the base of the cored part of Subunit IIE, perhaps indicating lulls in volcanic activity. Subunit IID contains numerous thin-bedded intervals with inclined bedding. At about 305 mbsf, where a sharp increase in lapilli size marks the boundary between Subunits IIB and IIC, magnetic susceptibility and P-wave velocity increase abruptly, and mean thermal conductivity decreases slightly. Below 380 mbsf, where a reduction in lapilli size marks the top of Subunit IID, both magnetic susceptibility and velocity decrease and mean thermal conductivity increases slightly.

All five subunits of Unit II consist predominantly of coarse ash to lapilli-sized glass and volcanic lithic fragments (Fig. 19), with less abundant accretionary and armored lapilli, set in a fine ash matrix (Fig. 20, Fig. 21). In most of the sequence, glass fragments are more abundant than lithic fragments. However, both the abundance of lithic fragments relative to glass and the proportion of red, oxidized lithic fragments are greatest in the lapilli tuff and lapillistone of Subunit IIC. Oxidation of lapilli probably accounts for the distinctively high magnetic susceptibility of Subunit IIC, and the presence of both hematite and magnetite has been confirmed by X-ray diffraction analysis.

Glass shards in Unit II volcaniclastic rocks range from <0.1 to ~10 mm and are predominantly subangular, blocky, and nonvesicular. Slightly to highly vesicular glass shards are relatively rare. Tachylite clasts are found throughout Unit II and form the main component of the upper and lower parts of Subunit IIC. Lithic fragments are mainly subround and subequant to subelongate and principally comprise nonvesicular and vesicular basalt (generally <10 mm), ranging from partly glassy to microcrystalline and fine grained, with rare fragments of diabase (≤20 mm). Rip-up clasts of tuff (≤65 mm) are also common. Plagioclase and clinopyroxene grains are present as phenocrysts in basaltic lithic fragments and as discrete clasts; as clasts, they are generally anhedral, showing signs of mechanical transport and/or fracturing. Accretionary (Fig. 20) and armored lapilli (≤15 mm) are present in all the subunits and are sometimes concentrated in bands.

We interpret the accretionary and armored lapilli, together with abundant blocky glass shards, to indicate that these deposits were formed by explosive hydroclastic eruptions in a shallow-water to emergent eruptive setting (Fig. 22). The presence of nannofossils in finer-grained intervals of tuff suggests primary deposition or reworking in a marine environment, and wood fragments and organic-rich layers indicate proximity to a vegetated island. Several features indicate that a component of the volcaniclastic material was derived from subaerial eruptions. These include the presence of vesicular tachylite lapilli throughout the volcaniclastic sequence, two intervals of well sorted lapillistone (consisting almost entirely of nonvesicular tachylite at the top and bottom of Subunit IIC), and the abundant red, oxidized lithic fragments in Subunit IIC.

The entire 337.7-m volcaniclastic sequence cored at Site 1184 is altered to varying extents, and the uppermost 8 m are completely altered to pale brown Fe oxyhydroxide, indicative of weathering in an oxidizing (subaerial?) environment. Except for plagioclase and clinopyroxene, almost all of the volcanic components and matrix are heavily altered to smectite, analcime, celadonite, calcite, zeolite, pyrite, and Fe oxyhydroxide. Unaltered glass is present in several cores (most commonly below ~470 mbsf); individual shards are typically rimmed by brown smectite (Fig. 23). From rim to center, the most commonly observed assemblage of secondary minerals in individual glass fragments follows the progression: smectite; analcime and/or other zeolites; rare calcite. The cement between individual clasts is predominantly composed of the same minerals as those replacing glass. Rare pleochroic, blue-green celadonite is also tentatively identified in the cement, filling vesicles in glass and partly replacing individual glassy fragments. The zeolites identified by X-ray diffraction are gmelinite, chabazite, levyne, and natrolite, an assemblage rarely found in submarine basalt but common in subaerial environments. Several generations of white, hairline to >5-mm-wide veins cut the cores; these veins are filled with analcime ± other zeolites ± calcite and lined with minor smectite and/or celadonite. Haloes in the groundmass adjacent to veins are rare, diffuse, and poorly developed; if present, they typically extend <1 cm into the wall rock and contain smectite and bluish celadonite or brown Fe oxyhydroxide.

Despite the middle Eocene biostratigraphic age of the volcaniclastic rocks, their chemical compositions are similar to those of the 122-Ma Kwaimbaita-type basalt flows and many of the 90-Ma lavas, such as those at Site 803 (Fig. 11, Fig. 12, Fig. 13). This result suggests that a fertile portion of the distinctive mantle source from which much of the Ontong Java Plateau appears to have been derived remained beneath this part of the eastern salient for 50-80 m.y. In the light of shipboard inductively coupled plasma-atomic emission spectroscopy (ICP-AES) analyses, it now seems unlikely that the Samoan hot spot provided much, if any, material for volcanism at Site 1184, although it potentially could have provided a source of heat for melting.

The major results of drilling at Site 1184 are summarized as follows:

1. Nannofossil evidence suggests a middle Eocene age for the volcaniclastic sequence drilled; much of the sequence appears to have been deposited within zone NP16. The volcanism at Site 1184 could be contemporaneous with the major change in Pacific plate motion at about 43 Ma (e.g., Duncan and Clague, 1985). The volcaniclastic sequence is much younger than the main phase of construction of the Ontong Java Plateau but similar in age to the 44-Ma alkalic Maramasike Formation in Malaita (Tejada et al., 1996). However, shipboard elemental data show that the volcaniclastic rocks at Site 1184 are composed of tholeiitic basalt clasts, the bulk composition of which resembles that of the widespread 122-Ma Kwaimbaita-type basalt or the similar basalts erupted at 90 Ma. A contribution of Samoan hot-spot mantle in Eocene magmatism at Site 1184 appears unlikely.

2. The abundance of blocky, vesicle-free glass shards implies that the volcaniclastic deposit was formed by hydroclastic eruptions, through the interaction of magma with shallow water. The abundant accretionary lapilli support this conclusion because they form only in steam-rich, subaerial eruption columns.

3. The virtual absence of lapilli larger than 20 mm suggests that the tuffs could not have accumulated close to the center of eruption. Deposition on the margin of a shoaling submarine volcano provides the most likely explanation for the volcaniclastic sequence.

4. Deep subaerial weathering at the top of the basement section, coupled with the nonmarine zeolite assemblage, indicates that this part of the eastern salient was above sea level initially. Proximity to land also is suggested by wood fragments found in organic-rich ash layers in the volcaniclastic sequence. However, the presence of nannofossils shows that at least some of the tuff was deposited in seawater.