,
Pass-through magnetometer measurements were taken on all split-core archive sections. Sediment cores were measured at 5-cm intervals. Coherent basalt pieces that could be oriented unambiguously with respect to the top were measured at 1-cm intervals. Pass-through magnetic susceptibility measurements were taken on all unsplit core sections at 4-cm intervals.
In order to isolate the characteristic remanent magnetization (ChRM), cores were subjected to alternating-field (AF) cleaning. The number of AF demagnetization steps and peak-field intensity varied depending on lithology, the natural remanent magnetization (NRM) intensity, and the amount of time available. On average, sediment half-cores were demagnetized using three AF steps in addition to the measurement of NRM. The basalt half-cores were demagnetized using a minimum of six AF steps. The maximum applied field ranged between 25 and 80 mT. We analyzed the results in Zijderveld and stereoplot diagrams; where possible, we calculated the ChRM direction using principal component analysis (Kirschvink, 1980). Examples of the AF demagnetization of sediment and basalt samples are shown in Figure F70.
We were able to reliably identify the magnetic polarity of sediment Cores 192-1185A-2R through 6R and Core 192-1185B-2R. We used the sequence of polarities together with the available biostratigraphic information (see "Biostratigraphy") to develop a correlation with the geomagnetic polarity timescale (GPTS) of Berggren et al. (1995) (Fig. F71). Some of our correlations are tightly constrained, whereas others are more tentative. For example, the reverse polarity of Core 192-1185A-4R can be correlated reliably with Chron C16r as this interval of reverse polarity is the only such interval within nannoplankton Zone NP18. The normal polarity of Core 192-1185A-5R correlates well with the normal polarity interval (C17n.1n) that is older than C16r but which is still identified within NP18.
Both Cores 192-1185A-2R and 3R are dated within foraminifer Zones P16 and P17 and nannoplankton Zone NP 19/20. Core 2R is entirely reversely magnetized, and we observe a transition downward within Core 3R from normal to reverse polarity. Only one normal-to-reverse transition took place during the time period represented by both NP19/20 and P16; consequently, the transition in Core 3R is C15n-C15r (Fig. F71). Because only one reverse polarity interval younger than Chron C15r is present within P16, the reverse polarity of Core 2R is Chron C13r.
Biostratigraphic information suggests a possible hiatus spanning a portion of Zone P15 near the middle to upper Eocene boundary (see "Biostratigraphy"), but all or part of nannoplankton Zones NP 19/20, 18, and 17 are present. If an appreciable hiatus exists, then we cannot uniquely correlate the normal polarity of Core 192-1185A-6R with the GPTS. We have chosen to use the age controls provided by the nannoplankton zones and have therefore correlated this normal polarity zone with Chron C17n.1n and view it as a continuation from Core 5R. Nannoplankton Zone NP 17 corresponds to a time during which there were several normal polarity intervals. Therefore, the normal polarity may correspond to another normal polarity interval between C17n.1n and Chron C18n.2n. Samples from the topmost portion of Core 192-1185A-8R indicate nannoplankton Zone NP16 and an age no older than 42 Ma (see "Biostratigraphy"). Unfortunately, the top of Core 8R was unsuitable for pass-through magnetometer measurements, and the magnetic polarity of these sediments could not be determined. Based on the evidence provided by Cores 192-1185A-6R and 8R, we have tentatively correlated the mixed polarities observed within Core 192-1185A-7R with the multiple polarity transitions associated with Chrons C17n.2n and C17n.3n. This correlation is consistent with available biostratigraphic information (which indicates that Zone NP17 nannofossils are present in Core 192-1185A-7R-CC) but is not required by it, and the mixed polarity zone could correspond to other parts of the reversal scale between Chrons C17n.2n and C18r.
Core 192-1185B-2R is entirely reversely magnetized and lies within nannoplankton Zone NP16 and within foraminifer Zone P12. These zones overlap considerably and therefore provide little additional time resolution. Two periods of reverse polarities, C18r and C19r, are spanned by NP16 and P12; the correlation of Core 192-1185B-2R to the GPTS is therefore ambiguous (Fig. F71).
The results of AF demagnetization of sediments recovered from Hole 1185A did not allow a definition of the ChRM direction precise enough for paleolatitude studies. However, the AF demagnetization results for Core 192-1185B-2R (Fig. F70A) allowed principal component analysis, and we were able to determine a paleolatitude for this core. A mean inclination of -24.4° ± 2.2° was obtained, which is equivalent to a paleolatitude of 12.8° ± 1.3°S. Correlation of the reversed polarity observed in Core 2R with Chron C18r or C19r indicates an age in the 40.1- to 42.5-Ma interval (Fig. F71). This age control is sufficiently accurate to allow us to compare the paleolatitudinal position of Site 1185 with that obtained for material of similar age from Site 1183. The paleolatitude, 12.8°S, at Site 1185 agrees remarkably well with that at Site 1183 (12.5°S).
The basalt cores showed a variable degree of overprinting with a near-vertical secondary magnetization most likely induced by drilling. Although the amount of overprinting varied, in almost all cases we were able to remove it with 10- to 20-mT AF demagnetization and to isolate the ChRM direction using higher fields (e.g., Fig. F70B). In Table T9 we list the ChRM direction, NRM intensity, magnetic susceptibility, Koenigsberger ratio, and median destructive field (MDF) for all coherent basalt pieces longer than 15 cm for which a reliable ChRM direction could be defined. For coherent pieces longer than 50 cm we list data for roughly every 25 cm. The ChRM inclinations obtained from different parts of long, coherent pieces generally agree within a few degrees.
Although Holes 1185A and 1185B are only 20 m apart, we were unable to uniquely correlate the magnetic properties (Fig. F72) or the ChRM directions (Fig. F73), probably indicating that different cooling units (pillow lavas) were recovered in the two holes. In Figure F72, the downhole variation in the rock-magnetic parameters of both Holes 1185A and 1185B is plotted on a common depth scale. We observe a distinct difference between the pillow lavas, with relatively low magnetic susceptibility and high MDF, and the more massive lava flows, with high susceptibility and low MDF. This difference can be explained by either the finer grain size of the pillow lavas or their higher degree of low-temperature alteration (see "Alteration"), or a combination of these factors. Shore-based rock magnetic studies on discrete samples will investigate the cause of the downhole variation in rock-magnetic properties.
The magnetic inclination is negative for all 162 ChRM determinations (Table T9), indicating normal polarity for all basalt cores recovered. The normal-polarity magnetization is consistent with biostratigraphic analyses of limestone interbeds that indicate a late Aptian to earliest Turonian or latest Cenomanian age (see "Biostratigraphy") and, hence, basalt emplacement during the Cretaceous Normal Superchron. In order to define the downhole variation in the paleomagnetic inclination data, we combined the individual ChRM data into distinct paleomagnetic units, each paleomagnetic unit being associated with a single cooling unit. Following the method of Kono (1980a), we calculated the statistic Z:
,
where Inc1
and Inc2 are the means of two adjacent inclination groups, 
1
and 2
are their standard deviations, and n1 and n2
are the number of inclination values in each group. The data are divided so that
the inclination groups differ at the 95% confidence level (Z
> 1.96). Using this criterion, the basement section in Hole 1185A
can be divided into four paleomagnetic units and that in Hole 1185B into 18
paleomagnetic units (Table T10).
Note, however, that this statistical test could not be used on those
paleomagnetic units that are based on only one ChRM determination.
In most cases the paleomagnetic unit boundaries conform with basement unit boundaries. In some cases, however, we observe more than one paleomagnetic unit within a single basement unit. This apparent discrepancy is not surprising because individual basement units often contain several chilled margins (see "Igneous Petrology"). The only notable example of a paleomagnetic unit that includes more than one basement unit is paleomagnetic Unit B14 (see Fig. F73), which includes the lower 17 m of basement Unit 9 as well as the upper 1 m of Unit 10. The boundary between Units 9 and 10 marks a distinct change in chemical composition (see "Igneous Petrology") and possibly a major hiatus in volcanic activity (in accordance with the high degree of low-temperature alteration at the top of Unit 10; see "Alteration"). We speculate that the upper part of Unit 10 was remagnetized when the lower part of Unit 9 was emplaced, either directly by heating from the overlying flow or by accompanying chemical alteration. We note that the rock-magnetic properties in the uppermost meter of Unit 10 are slightly different from those in the lower part of the unit (see Fig. F72).
Using the statistics of Kono (1980b), we have calculated the mean inclination for all 22 paleomagnetic units (Table T11). As previously mentioned, a major hiatus in volcanic activity may be present between basement Units 9 and 10 in Hole 1185B. We have therefore also calculated the mean inclination for the paleomagnetic units above and below this suspected volcanic hiatus (Table T11). The upper group, comprising paleomagnetic Units A1-A4 and B1-B14, have a somewhat shallower inclination and correspondingly less southerly paleolatitude than the lower group, comprising Units B15-B18. Although this result supports a possible time gap between the two groups, the lower group comprises only four paleomagnetic units and therefore does not provide a statistically adequate sample to give confidence to this distinction. Shore-based studies of discrete samples are necessary in order to better define the paleomagnetic units and possibly identify new units not detected in our pass-through measurements.
A basic assumption when converting the inclination of the paleomagnetic field to paleolatitude is that secular variation has been averaged. The angular standard deviation (ASD) listed in Table T11 gives a direct measure of the paleosecular variation at the time and paleolatitude where the lavas were emplaced. The ASD values observed at Site 1185 agree well with the expected ASD value of 12° obtained from McFadden et al.'s (1991) global compilation of paleosecular variation from lavas. We therefore feel confident that the mean inclinations for the first two groups (i.e., all paleomagnetic units and Units A1-A4 and B1-B14) in Table T11 represent a time-averaged geocentric dipole field and that the paleolatitude calculation is valid. Finally, we note that the sediment-basement seismic reflection is horizontal and there is no evidence for local tectonic rotation of the basalts at Site 1185 (see "Background and Objectives").
A comparison between Site 1185 and previously obtained paleolatitudes is difficult because of the poor age control of the Site 1185 basalts. We note, however, that, if the younger limit of the age estimate for the upper basalt sequence (latest Cenomanian) is correct, the Site 1185 paleolatitude would be in good agreement with the paleolatitude obtained from slightly younger sediments from Site 1183 (see "Paleomagnetism" in the "Site 1183" chapter).
