Source of Material

Hole 1085A is situated at the African rim of the Cape Basin at a depth of 1713.1 m on the continental slope. The position is given as 29░2.5S, 13░9.4E (Shipboard Scientific Party, 1998). Sediments of early Quaternary age, according to shipboard results, have accumulation rates between 35 and 100 m/m.y. We analyzed samples from Cores 175-1085A-7H through 10H. Sample spacing was 20-cm intervals; the total number of samples analyzed was 196.

Sediments of Quaternary age at Site 1085 are just over 90 m thick (average sedimentation rate near 5 cm/k.y.) and consist of greenish gray nannofossil-foraminifer ooze (Shipboard Scientific Party, 1998). Finely disseminated pyrite is common. Gas concentrations are moderate at Site 1085, and cores show much less expansion than for the immediately preceding sites occupied closer to the upwelling centers. Quaternary and Pliocene sediments were retrieved by hydraulic piston coring in advance of the drill bit.

General Strategy

We selected Cores 175-1085A-7H through 10H for study to investigate variations in productivity within the early Quaternary. Our samples span the period from ~1.9 to 1.1 Ma. We are only interested, in this context, in the phase of productivity-related indices relative to glacial-interglacial climatic fluctuations. Our purpose is not to provide a complete isotope stratigraphy for this interval, a task made difficult by the gaps we suspect to be present between the cores (cf. composite depth assignments in Shipboard Scientific Party, 1998). Our age assignments, consequently, are tentative. They might be helpful in work by others on these cores and are given for this reason only.

In pursuing the question of phasing of productivity, we use the fact that the early Quaternary is dominated by the 41-k.y. cycle in the global ocean 18O record (Shackleton et al., 1990; Berger and Jansen, 1994). We assume that the 41-k.y. cycle, likewise, is important at Site 1085 within the early Quaternary. Based on this assumption, we take the dominant cycle closest to 40-k.y. length within each core (as found by spectral analysis of single-core series dated by shipboard age assignments) as, in fact, representing the 41-k.y. cycle. Also, we restrict our analysis to the one core (Core 175-1085A-7H) where this cycle apparently is best represented. We present the data from the other cores, as well, to justify the selection of Core 175-1085A-7H. However, we largely ignore the data from Cores 175-1085A-8H through 10H in terms of providing evidence bearing on the question addressed. These data might prove useful in additional work on Site 1085 and in contexts not here contemplated.

Proxies were selected for (inferred) relevancy to productivity reconstruction and for ease of determination. The proxies here used are 18O of Cibicidoides wuellerstorfi, 13C of C. wuellerstorfi, benthic foraminifers per gram (BF/g), Uvigerina spp. per gram (U/g), visually estimated abundance (EDA) of biogenic silica (mainly diatom debris), and visually estimated abundance of iron (hydr)oxides and sulfides (SOA) and similar particles (mainly pyrite). The indices EDA and BF/g are taken to reflect diatom flux and carbon flux, respectively. BF/g is assumed to depend on the amount of carbon that settles from the productive surface waters to the seafloor. The higher the surface productivity, the more carbon reaches the seafloor, and the more benthic foraminifers can thrive (Herguera and Berger, 1991). In addition, U/g and 13C of C. wuellerstorfi are used to estimate relative changes in conditions regarding oxygen content and nutrient supply. To complement the data presented here, we also use the data available from shipboard analyses as published in the Initial Reports volume (Wefer, Berger, Richter, et al., 1998).

Sample Preparation

For this study, 196 20-g samples were obtained from Cores 175-1085A-7H, 8H, 9H, and 10H. The sampling was at 20-cm intervals, starting from the top of each core. The approximate sedimentation rate (based on shipboard data) is 4 to 5 cm/k.y. Thus, the 20-cm sampling interval typically has a resolution of between 4 and 5 k.y. For each sample, half the original amount was washed to obtain the coarse fraction containing the foraminifers, using standard laboratory techniques (details in Anderson, 2000). The average sand fraction was typically ~10% by weight. The coarse-fraction samples were then split into halves, one for reference, the other for use. The used half was further sieved into three size fractions: >250, 150-250, and 63-150 Ám. A "coarse" sand fraction (>250 Ám) was measured as the percent coarse sand of total sand.

Measurements and Indices

The isotopic data generated for this project were obtained for two species of benthic foraminifers: Cibicides, an epifaunal species, and Uvigerina peregrina, an infaunal species. Both 18O and 13C were measured using the Finnigan stable isotope mass spectrometer (model number MAT252) at Scripps Institution of Oceanography. The 18O index is made up of C. wuellerstorfi 18O measurements. Where there were no actual measurements from C. wuellerstorfi the measurement of Uvigerina - 0.92 was used in its place. All of the 18O Fourier analyses are based on the merged 18O series. The shift of 0.92 in oxygen isotopes of Uvigerina spp. for these early Quaternary sediments is larger than that found for more recent sediments off Peru (~0.6) (Dunbar and Wefer, 1984) but close to the one given by Woodruff et al. (1980) for recent benthic foraminifers of the deep-sea environment.

Data series were constructed for 13C of the two species in the same manner as for the 18O. Similar to 18O, an offset proved to be just as accurate as the regression equation. The difference between 13C of C. wuellerstorfi and 13C of Uvigerina is typically >1, which may be compared with the difference of 0.8 stated in Woodruff et al. (1980). No attempt was made to ascribe special significance to the 13C values of Uvigerina as an infaunal taxon (Zahn et al., 1986; McCorkle et al., 1990; Wefer and Berger, 1991; Mackensen and Bickert, 1999). A systematic error is likely to be present (increased organic carbon supply resulting in more negative delta values in Uvigerina), and results should be read with the appropriate caution.

U/gram was calculated by counting the number of specimens of Uvigerina in one-half of the washed split >250 Ám. This number was doubled to account for the reference half. This new value was divided by the dry weight of the entire sample before washing to obtain the U/g values.

BF/g denotes benthic foraminifers per gram of total dry sample. The foraminifers were noted as a fraction of total sand, whose ratio to total sediment was determined before splitting and counting.

EDA is based on visual inspection of smear slides made from the (unwashed) reference half of the original sample. The manner of determination was suggested by Carina Lange and is described in Wefer, Berger, Richter, et al. (1998), where smear slide diatom abundances are listed for Site 1085 and other Leg 175 sites. Lange et al. (2000) have shown that smear slide estimates of diatom abundance correspond closely to measured opal content in these sediments.

SOA abundance (sulfide and oxide aggregates) was recorded in nearly all of the samples in the >250-Ám size fraction as aggregates of iron sulfide and iron oxide-hydroxide, such as fillings, casts, and similar precipitates. Both SOA and EDA indices are subjective estimates of ranked abundance (from "barren" to "extremely abundant" [0 to 6]) based on microscope observations. No chemical analyses were performed on the materials.

Percent sand is the dry weight of the residue after washing on a 63-Ám sieve as a ratio to the total dry weight of the sample before washing. Percent coarse sand is the percentage of the sand fraction >250 Ám. Within calcareous ooze, sand content may be taken as a clue to preservation of foraminifer shells, that is, as a dissolution index (Johnson et al., 1977).

Depth and Age Assignments

Depth and age assignments are listed in the "Appendix," where the method leading from driller's depth to final depth is described. We used spectral analysis to estimate the in situ sedimentation rate based on the assumption that 41-k.y. cycles are present in these early Quaternary sediments. Our age assignments do not differ substantially from interpolations of shipboard tie points. In fact, they are not crucial to the arguments regarding the Walvis Paradox. These arguments deal with the phase of various proxy parameters within a typical obliquity cycle whose precise position within the early Quaternary is immaterial. We offer the depth and age assignments here because they might prove useful in other contexts.

The sedimentation rates arrived at by the procedure outlined in the "Appendix," range from 3.61 cm/k.y. for Core 175-1085A-7H to 9.81 cm/k.y. for Core 10H (see Fig. F2). They decrease upward, most rapidly in the early portion of the interval studied, presumably as a result of changes in sediment supply by the Oranje River or from an increase in bottom current activity, or both.

Core 175-1085A-7H has a strong, well-defined 41-k.y. signal showing a dominant obliquity cycle in the oxygen isotope record. Only minor adjustment was necessary to bring the peak into focus (Table T1). The other cores (Cores 175-1085A-8H, 9H, and 10H) have spectra that are less clearly tied to obliquity (or other Milankovitch forcing). Core 175-1085A-7H, having the strongest signal, is used as a template for examining the phase between proxies.