RESULTS 

Sand Description

In general, sand recovered at the Leg 149 sites is composed dominantly of quartz, feldspar, and metamorphic lithic fragments, with locally abundant mica (Appendix; Pl. 1, Pl. 2). The quartz grains exhibit slightly undulose to straight extinction and occur as individual monocrystalline grains and sand-sized components of lithic fragments (Pl. 1, Pl. 2). Polycrystalline quartz grains, including chert, make up less than 5% of the framework grains (Appendix; Pl. 1). Both potassium and plagioclase feldspar are present in moderate to minor amounts, and most occur as individual monocrystalline grains, rather than within feldspathic (plutonic and high-grade metamorphic) lithic fragments (Pl. 1, Pl. 2). Feldspar grains range from fresh to altered, with alteration products including sericite and kaolinite (Pl. 1, Pl. 2). Unstained albitic feldspar is rare, and generally very altered (Pl. 2, Fig. 2). Muscovite and biotite also make up a significant proportion of the framework grains (Table 2; Pl. 1, Fig. 6), with lesser amounts of chlorite. Biotite flakes exhibit brown to green pleochroism and are commonly altered to opaque minerals, clay minerals, and chlorite.

The lithic component is dominantly fragments of low-grade metamorphic and metasedimentary rocks. The metamorphic lithic fragments are mostly quartz-mica tectonite and polycrystalline mica, with lesser amounts of quartz-feldspar-mica aggregate (Pl. 1, Pl. 2). Sedimentary lithic fragments, such as shale/argillite, sedimentary or dirty chert, and micritic/microsparite carbonate, are significant components in some samples (Pl. 1, Pl. 2). In contrast, fragments of variably altered microlitic, vitric, and felsitic volcaniclastic debris are generally rare (Pl. 2).

Bioclastic material is present in each sample. A wide variety of calcareous (e.g., foraminifers, shell fragments) and siliceous (predominantly sponge spicules) bioclasts are present in varying amounts. Because of their dominance in a few samples ("nc" in the Bio column of the Appendix), they were not tallied into the total points counted.

A few samples have a significant percentage of opaque grains, but surprisingly few nonopaque dense minerals are present (Appendix). The most common nonopaque dense minerals are green hornblende, zircon, and garnet (Pl. 1), with lesser amounts of augite, zoisite, epidote, and sphene. Rare, isolated grains of tremolite, tourmaline, enstatite, and rutile are also present. The heavy mineral fraction is generally finer grained than the associated siliciclastic components. Other minor components include glauconite (Pl. 1), plant fragments, carbonate (authigenic?), and serpentine.

Sand Detrital Modes

The range of sand compositions within Leg 149 cores is shown in Figure 5 and Figure 6. In general, most samples are quartzofeldspathic with only a minor lithic component. Sand from Sites 899 and 901 is more enriched in lithic fragments, and a few samples from Sites 897 and 898 are very feldspathic (Fig. 5). In terms of monomineralic components, quartz and potassium feldspar are the most common, but a number of samples from Sites 897 and 898 are significantly enriched in plagioclase feldspar (Fig. 6). Although lithic proportions (LmLvLs) were calculated (see Table 1, Table 2), these are not presented in ternary format because of the dominance of metamorphic lithic fragments in most samples.

Relationship of Composition to Grain Size

The Gazzi-Dickinson method of point counting was used in this study to minimize grain-size effects on sand composition (Dickinson, 1970; Ingersoll et al., 1984). When sand samples are grouped by grain size, however, there appears to be a direct relationship between grain size and the QFL and QmKP detrital modes. The Cenozoic medium sand samples are much more quartzose than the fine to very fine sand samples, and there is a progressive trend in mean QFL and QmKP compositions toward more feldspathic compositions with decreasing grain size (Fig. 7). The distinction between medium, fine, and very fine sand compositions is even more pronounced for the Pleistocene samples, which exhibit a similar trend toward more feldspathic compositions with decreasing grain size (Fig. 8). Also, Pleistocene very fine sand is slightly enriched in plagioclase feldspar and lithic fragments with respect to the fine to medium sand (Fig. 8).

In their analysis of the traditional vs. the Gazzi-Dickinson petrographic point-counting methods, Ingersoll et al. (1984) counted unsieved sand samples and, for comparison, various sieved fractions from these same samples. A similar analysis of Leg 149 samples would be useful in deciphering the nature of the relationship between feldspar content and grain size as outlined in Figure 7 and Figure 8; however, further subdivision (sieving) of these samples is generally precluded by the small sample volume (~5 cm3) and high matrix to sand ratio.

Relationship of Composition to Sediment Age

Because of the grain-size correlations outlined above, sand samples were first grouped by grain size prior to further subdivision by age. The large number of very fine sand samples provides the best statistical sample (four Pleistocene, 10 Pliocene, six Miocene, five Oligocene, and four Eocene) for determining age relationships. The means and fields of variation for these age subsets exhibit a large amount of overlap (Fig. 9), but the Oligocene and Miocene samples show a higher range in QFL lithic proportions, as compared with the Eocene, Pliocene, and Pleistocene samples. In terms of monomineralic proportions, the Eocene and Oligocene samples are significantly enriched in potassium feldspar with respect to younger samples, and there appears to be a shift toward more plagioclase-rich sand from the Miocene to the Pleistocene (Fig. 10).

The limited number of fine sand samples allows for only a general comparison of detrital modes, but across a greater time span (Mesozoic to Cenozoic). A crude QFL compositional trend can be outlined, with lithic-rich Jurassic sand (stone) and lithic-poor Pleistocene sand as end-members (Fig. 11). In terms of QmKP proportions, again, some crude temporal trends are apparent (Fig. 11): the Mesozoic samples are the most quartzose with lesser plagioclase, the Eocene to Oligocene samples have the highest K-feldspar content, and the youngest Miocene to Pleistocene sediments exhibit compositions between these two end-members.

NEXT