Figure F1 shows the chromatograms of total lipid fractions (Sample 184-1148A-8H-2, 56–57 cm; 59.86 meters below seafloor). The compounds mainly consist of n-alkanes, fatty acids, alcohols, long-chain alkenones, diols, and keto-ols. Several of these compounds from both terrestrial plants and marine sources will be discussed according to compound classes.
n-Alkanes in the studied downhole sediments range in carbon number from 15 to 35; C25–C33 n-alkanes are the most dominant homologues (Fig. F1B). An odd-over-even carbon number predominance, indicated from average carbon preference index (CPI) CPI24–34 = 2.6 (ranging between 0.3 and 18.5) and odd-even predominance (OEP) = 3.3 (ranging between 0.1 and 12.4), was observed in nearly all samples, indicating a predominantly terrigenous origin of the long-chain n-alkanes (Fig. F2). These distributions resemble those of n-alkanes from leaf waxes of higher plants (Kolattukudy, 1976; Eglinton and Hamilton, 1967) and in eolian dust samples (Gagosian et al., 1981, 1987; Simoneit et al., 1977), supporting a terrigenous origin.
The average chain length changes from 26.4 to 31.6 (average = 29.2) (Fig. F2). It has been suggested that longer-chain compounds are produced by plants in warmer climates (e.g., Poynter et al., 1989) or are derived from grasslands, which may have, on average, longer chain lengths than leaf lipids from plants in rainforests (Cranwell et al., 1973). The observed shift in our data set will be addressed in future research.
Fatty acids in the downhole sediments range in carbon number from C14 to C32 (Fig. F1C) and n-alcohols occur in the carbon number range of C16–C32 (Fig. F1D). The downhole distributions of C22–C32 fatty acid and n-alcohol concentrations are similar to that of long-chain n-alkanes. Even carbon numbered long-chain compounds dominate the distributions of both compound groups. The dominance of the long-chain compounds, together with the observed strong even-over-odd carbon number predominance (for fatty acid and n-alcohols, average CPI22–30 = 9.7 and 13.1, respectively), indicates a terrigenous origin for fatty acids and n-alcohols. Series of long-chain n-fatty acids and n-alcohols with a strong even carbon number predominance are characteristic constituents of surface waxes of higher plants (Kolattukudy, 1976). Therefore, the biological source of these compounds is closely related to that of the leaf wax–derived n-alkanes. An algal origin of the long-chain fatty acids, however, cannot be completely excluded (Volkman et al., 1998). Marine microalgae are not a major source of long-chain n-alcohols but may contain minor amounts of these compounds (for a review, see Volkman et al., 1998). However, the strong covariance of the concentrations of long-chain fatty acids and n-alcohols with the concentrations of n-alkanes, together with their high CPI values, indicate allochthonous sources of fatty acids and n-alcohols. In other words, higher plant leaf waxes are the main source.
Total concentrations of C27, C29, and C31 n-alkanes in sediments (average = 0.47 µg/g dry sediment) vary in a large range between 0 and 12.33 µg/g dry sediment (Table T1). Three sections have relatively high values. They are the sections below 495 meters composite depth (mcd), 202–245 mcd, and 0–166 mcd. The highest values occur in the section between 202 and 245 mcd (Fig. F3). The accumulation rate variation pattern of n-alkanes shows the feature more pronouncedly (Fig. F4). Sectional variations of total long-chain n-fatty acids and n-alcohols are quite similar to those of n-alkanes (Figs. F3, F4).
The downhole variations of leaf wax components possibly result from changes in the different oceanographic and climatic settings, which determined the inputs of terrigenous lipids to Sites 1147 and 1148. The section below 495 mcd was deposited with a significant supply of terrigenous fine-grained clasts in the late Oligocene, which is the SCS-floor spreading phase (Briais et al., 1993). At that time, the newly opened, narrow basin, with restricted ocean waters and the proximity of continental runoff, led to an increase in terrigenous input. The much higher concentrations of terrigenous lipids in the section between 202 and 244 mcd, deposited at ~6–8 Ma when the eolian component of the loess red-clay sequence first appeared at the Chinese Loess Plateau (Ding et al., 1998), may be related to the development of the East Asian monsoon. And the section above 166 mcd with high, fluctuating concentrations of terrigenous lipids may indicate the enhanced variations of the developed monsoon since ~3.2 Ma.
Sediment samples above 305.37 mcd contain C37 and C38 di- and triunsaturated methyl and ethyl ketones. High concentrations of these compounds were found in samples at the top of interval 0.56–61.76 mcd. Concentrations of those compounds in sediments in the interval 61.76–162.08 mcd are much lower and are trace in the interval 164.07–186.87 mcd (Table T1; Fig. F5). Some samples in the interval 202.57–305.37 mcd contain only C37:2 ketone, but its concentrations are even higher than those at the top of interval 0.56–61.76 mcd.
Long-chain alkenones are exclusively biosynthesized by haptophyte algae like Emiliania huxleyi and Gephyrocapsa spp. (Volkman et al., 1980, 1995; Marlowe et al., 1984) and were detected in numerous marine sediments (e.g., Brassell et al., 1986), including the SCS (Pelejero and Grimalt, 1997). The Uk´37 index, the ratio of the diunsaturated to the sum of the di- and triunsaturated C37 alkenones, is extensively used in paleoceanography as a temperature proxy. In this work, downhole Uk´37 values in the alkenone-containing interval 0–186.7 mcd show a decreasing trend upward, which may indicate a general cooling process since ~5 Ma. In addition, high C37:2 alkenone concentrations with the absence of C37:3 alkenone in the interval 202.57–305.37 mcd may suggest higher sea-surface temperature in the late Miocene than the following period. However, although it has been shown that the Uk´37 index is well correlated to 
18O signal for Globigerinoides sacculifer for the past 550 k.y., the indexes show little agreement in samples older than that age (Brassell et al., 1986). For Sites 1147 and 1148, only the top section to the depth of ~40 m approximately corresponds to the past 550 k.y.; therefore, the downhole shift of Uk´37 and Uk´37-based sea-surface temperature, using the equation of Pelejero and Grimalt (1997), is tentatively given here and not discussed further in this report (Fig. F6).
Alkyl diols have been reported to occur in sediments from various areas (de Leeuw et al., 1981; see review in Versteegh et al., 1997). Their major sources are probably microalgae of the class Eustigmatophyceae (Volkman et al., 1992, 1999). Long-chain saturated C30 alkyl diols were detected in some samples, mainly above 244.77 mcd and below 519.35 mcd. Its downhole variation is similar to that of C37:2 ketone (Table T1; Fig. F5).
C30 and C32 alkyl keto-ols are present in samples above 241.17 mcd. A deceasing trend is obvious in C30 alkyl keto-ols from 0 to 195.07 mcd, and a sharp increase occurs in some samples in the next 50-m interval (Table T1; Fig. F5). For C32 alkyl keto-ols, the pattern is similar, but these were not detected in many samples. Versteegh et al. (1997) pointed out that both compound classes are diagenetically independent, but their sources are as yet unidentified.
From the accumulation rate of marine microalgae biomarkers, variation in marine paleoproductivity can be illustrated. All the sectional accumulation rate fluctuations of the C37 + C38 alkenones, C30 alkyl diols, and C30 + C32 alkyl keto-ols are quite similar, with high values in intervals 202–245 and 0–162 mcd (Fig. F7). As discussed above, the pattern may be related to East Asian monsoon development and variation. Lipids from marine sources are almost absent from the deeper part of the core (from 250 m onward), indicating low productivity and/or severe diagenetic degradation.
