Magnetostratigraphy, when coupled with biostratigraphic control, provides high-resolution correlation among sites and a precise calibration to the absolute timescale. Paleomagnetic vectors indicate the history of plate rotation and paleolatitude motion of the site, and magnetic characteristics provide insight into the mineralogy, preferential orientation, and grain size of remanence carriers.

Paleomagnetic investigations for Leg 207 had a shipboard and a shore-based phase. The measurements performed aboard the JOIDES Resolution concentrated on a high-resolution survey of the natural remanent magnetization (NRM) of archive-half core sections before and after alternating-field (AF) demagnetization and on low-field magnetic susceptibility (k) measurement. Approximately 2000 core sections were analyzed.

The shore-based program consisted of progressive thermal demagnetization of ~800 discrete minicores at the paleomagnetism laboratory at the University of Munich, Germany. The discrete samples required shore-based analysis because sediment sample magnetizations are too weak to be measured reliably with shipboard equipment and the extensive core flow did not allow time for these procedures. In addition to revealing a more reliable magnetostratigraphy at each site, the detailed thermal demagnetization of these discrete samples allowed insight into the magnetic properties and magnetic mineralogy of the sediments and a determination of paleolatitudes for plate motion studies. The combined results for magnetostratigraphy are incorporated in the "Paleomagnetism" sections in each site chapter.

Whenever possible, we offer an interpretation of the magnetic polarity, with the naming convention following that of correlative Late Cretaceous and Cenozoic anomaly numbers prefaced by the letter C (e.g., Tauxe et al., 1984; Cande and Kent, 1995). Normal polarity subchrons are referred to by adding suffixes (e.g., C24n.n1 and C24n.n2) that increase with age. Assignments of polarity chrons are constrained by the associated biostratigraphy and by characteristic patterns of relative thickness in the sedimentary record.

The ages of the polarity intervals used during Leg 207 are modified from the Berggren et al. (1995b) composite of magnetic polarity timescales calibrated to marine microfossil biostratigraphy (Table T7). The Leg 207 magnetic polarity timescale is summarized in Figure F5.

For each hole, tabulated and annotated paleomagnetic measurements for each discrete sample and a summary of the preliminary polarity interpretations with graphical representations are included.

Shipboard Measurements and Equipment

The standard ODP coordinate system was used, where +x is the vertical upward direction to the split surface of archive halves or 0 declination in core coordinates, +y is the direction to the left along the split-core surface when looking upcore or 90 declination in core coordinates, and +z is the downcore direction or 90 inclination in core coordinates (see Fig. F6).

Remanent magnetization was measured using the shipboard long-core cryogenic magnetometer equipped with direct-current superconducting quantum interference devices (DC SQUIDs) (2-G Enterprises model 760R) and an in-line automated AF demagnetizer capable of reaching a peak field of 80 mT. Continuous core measurements were made at 5- to 10-cm intervals. An additional 20-cm header and trailer distance allows future deconvolution of the core data. The pickup coils of the cryogenic magnetometer measure the core over an interval of a little more than 30 cm, although 85% of the remanence is sensed from a 20-cm width. Because of this window size and the 5- to 10-cm sampling intervals, adjacent measurements made on Leg 207 cores are not strictly independent. Measurements within 5 cm of core and section ends, within 5 cm of contacts between rotated blocks, and within intervals of drilling-related core disturbance were removed during data processing. In addition, the upper 20 cm of each core that commonly displayed spurious high-intensity magnetization or downhole contamination and the upper 5 cm of each section that was influenced by magnetization (blue-colored end cap) were excluded. The magnetization of the core liner itself was removed by applying a holder-tray correction to each run.

The background noise of the cryogenic magnetometer seems to be amplified by the ship's movement compared to shore-based instruments; it was measured by running empty holder trays to be ~3 x 10–5 A/m, assuming a measured volume of ~100 cm3. The relatively large volume of core material in the sensing region normally compensates for the relatively high background noise; however, for the chalks that dominated the stratigraphy at Leg 207 sites, more than one-half of the sediment recovered was measured to be <5 x 10–5 A/m and was therefore considered to be too near the instrument noise level to be useful for polarity determinations. In contrast, the shore-based cryogenic magnetometer enabled an extra order of magnitude sensitivity for the minicores of 12-cm3 volume.

Measurements of NRM and stepwise AF demagnetization were performed on all archive halves longer than 20 cm. More than 6000 runs were carried out on the shipboard magnetometer. Because AF demagnetization of the archive half was conducted mostly to remove the soft magnetic overprint that was acquired during the drilling process, field strengths were limited to 20 mT to avoid erasing the entire primary NRM component. Sections were measured at NRM, 10- and 15-mT AF demagnetization steps, with an additional 20-mT step applied if core flow permitted it. The 10-mT step appeared to be effective in removing extraneous overprints induced during the drilling process. In general, the additional 20-mT demagnetization step did not significantly alter the magnetic direction obtained at the prior 15-mT step for the majority of the sediment types.

The orientations of APC cores (only used for the uppermost cores in Hole 1257A) were recorded using the Tensor tool (Tensor Inc., Austin, Texas). The instrument has a three-axis fluxgate magnetometer that records the orientation of the double lines scribed on the core liner with respect to magnetic north. The critical parameters for core orientation are the inclination angle (typically <2) and the angle between magnetic north and the double line on the core liner, known as the magnetic tool face angle. The Tensor tool readings were recorded continuously at 30-s intervals, downloaded to a computer, and analyzed once the tool was back on deck. Orientation of core declinations is of particular importance in magnetostratigraphic interpretation of sediments from equatorial regions, where the paleomagnetic inclination is close to zero. APC coring and associated Tensor tool orientation is limited to relatively soft sediment intervals, whereas every core drilled during Leg 207, except the uppermost ones in Hole 1257A, were from nonoriented XCB and RCB cores in harder sedimentary rocks.

Magnetic susceptibility was measured for each whole-core section as part of the MST analysis (see "Multisensor Track Measurements" in "Physical Properties"). Susceptibility is measured on the MST using a Bartington Instruments MS2 susceptibility meter coupled to a MS2C sensor coil with a diameter of 8.8 cm operating at 0.565 kHz. The sensor was set on SI units, and the data were stored in the Janus database in raw meter units. The sensor coil is sensitive over an interval of ~4 cm (half-power width of the response curve), and the width of the sensing region corresponds to a volume of 166 cm3 of cored material. To convert raw instrument units to true SI volume susceptibilities, these values should be multiplied by 10–5 and then multiplied by a correction factor to account for the actual volume of material that passed through the susceptibility coils. Except for measurements near the ends of each section, the correction factor for a standard full ODP core is ~0.68. The end effect of each core section is not adequately corrected using this procedure.

Shipboard Polarity Interpretations: Data Filtering and Inclination Clusters

All recovered sediment pieces from XCB and RCB cores have unknown horizontal rotations. The usage of magnetic declination was limited to the few APC cores in the uppermost portion of Hole 1257A. Therefore, our interpretations of characteristic magnetic polarity rely on measurements obtained from within continuous sediment intervals that exceed 15 cm in length with no suspected interblock rotations. These intact segments yield at least one measurement positioned at least 5 cm from an adjacent block having relative rotation. For this purpose, each intact interval was logged in each analyzed section, and we filtered out all measurements that did not meet these criteria. All measurements from the upper 20 cm of each core were omitted because these were generally contaminated with downhole displaced sediment or anomalous intensity spikes in drill slurry.

All 15- and 20-mT measurements less than the background noise level of the cryogenic magnetometer (3 x 10–5 A/m) were filtered from the block-data file, and all measurements <5 x 10–5 A/m are considered unreliable indicators of polarity. These low-intensity filters imply that the demagnetized data from the majority of light-colored chalks, which constitute more than half of the Campanian–middle Eocene facies, yielded directions that we considered to be unreliable. Indeed, all shipboard measurements from several entire cores failed this threshold. The signal-to-noise ratio of each of the remaining intact blocks was enhanced by employing a three-point moving mean through adjacent data measurements.

In most shipboard magnetostratigraphy interpretations, polarity zones are identified by stratigraphic clusters of positive or negative magnetic inclinations. For sites located north of the paleoequator, these are respectively associated with normal polarity or reversed polarity chrons but have an opposite association for sites south of the paleoequator. The Leg 207 sites are between 9 and 10N latitude. Global plate reconstructions generally indicate that South America has experienced a small amount of northward drift with negligible plate rotation since the Early Cretaceous (C.R. Scotese, pers. comm., 2002); therefore, pre-Oligocene paleolatitudes of the Leg 207 sites were projected to be closer to the paleoequator.

During Leg 207, the filtered pass-through cryogenic measurements generally consisted of three broad classes of stratigraphic clustering of inclinations: (1) uniformly positive, (2) uniformly negative, or (3) random mixing of positive and negative. Stratigraphic intervals dominated by each class were generally consistent among the different holes at a site after correcting to composite depth (the meters composite depth [mcd] scale). For shipboard estimates of magnetic stratigraphy, we assumed a model in which the original polarity has a variable persistent overprint of present-day north-directed polarity and that clusters of negative inclination correspond to polarity zones of reversed polarity. These shipboard polarity interpretations of these three classes and the associated fitting to polarity chrons were implicitly based on an optimistic assumption that the Leg 207 sites had been north of the paleoequator during the latest Cretaceous and early Cenozoic. Therefore, Class 1 was considered either to be a primary normal polarity or a dominance by present-day overprinting, Class 2 was assigned as reversed polarity, and Class 3 was considered to indicate reversed polarity with variable overprinting by the normal polarity present-day field.

The rate of core flow coupled with the observed ineffectiveness of additional progressive demagnetization by AF methods to significantly change the declination-inclination vectors of most lithologies precluded using sample-by-sample analysis of demagnetization behavior to evaluate removal of overprints and relative rotations of magnetic vectors as clues to the underlying magnetic polarity. In general, for a site located north of the paleoequator, the magnetic vectors of a normal polarity sample would display minimal directional change during AF demagnetization and a normal polarity interval would display a high degree of inclination clustering near 0 to +20. In contrast, a reversed polarity sample might display significant directional change during partial removal of the present-day overprint and a reversed polarity interval might be characterized by a significant scatter in inclinations toward relatively steep downward to low-angle negative (upward) inclinations. This simplified model of ideal magnetic behavior upon AF demagnetization was implicit in our association of inclination classes with polarity zones.

However, the near-equatorial position and the uncertainty in the Cretaceous–Eocene paleolatitudes of the Leg 207 sites and the lack of response of some sediment lithologies to AF demagnetization precluded using magnetic inclination as a reliable indicator of polarity. Therefore, we systematically drill-pressed an extensive suite of minicores to provide stratigraphic coverage of the Campanian–Eocene at each site.

Shore-Based Measurement and Procedures

Discrete samples were collected during the cruise and at the Bremen repository from working halves of core sections by drill-pressing standard (12 cm3) minicores with the orientation arrow on the cut face of the sample pointing upcore (see Fig. F6). The shipboard sampling frequency was generally one minicore from every second section (3 m spacing) from one hole per site (a total of ~400 minicores). Additional postcruise sampling enhanced resolution of the magnetic polarity pattern in selected intervals (a total of ~400 minicores), especially in the Maastrichtian–Campanian of all sites and in the Eocene of Site 1258.

The cryogenic magnetometer at the Department of Earth and Environmental Sciences at the University of Munich is housed in a magnetically shielded room with equipment for thermal and AF demagnetization and measurement of susceptibility. The effective background noise of this three-axis 2-G Enterprises cryogenic magnetometer for a paleomagnetic minicore is ~1 x 10–6 A/m (1 x 10–3 emu/cm3), which implies that reliable polarity information can be obtained from samples with remanent magnetizations as low as 4 x 10–6 A/m. This sensitivity level is an order of magnitude greater than the operation of the shipboard pass-through cryogenic magnetometer during Leg 207. To further increase the signal-to-noise ratio, we performed duplicate measurements whenever a sample had a remanent magnetization <8 x 10–6 A/m.

A suite of pilot demagnetizations of representative sediment lithologies of different ages recovered during Leg 207 indicated that a combined AF and progressive thermal demagnetization treatment was effective in resolving characteristic polarities. All 800 minicores underwent an initial AF demagnetization of 5 mT, followed by progressive heating steps at 30–50C increments in the 150–450C range. Magnetic susceptibility was measured after each thermal demagnetization step above 300C to monitor potential formation of new magnetic minerals or other anomalous changes in magnetic characteristics.

Nearly all samples displayed NRM with positive inclinations that are generally significantly steeper than the 20 inclination expected for the present latitude of the Leg 207 sites. These inclinations typically shallowed upon application of the initial 5-mT demagnetization step. Therefore, we interpret the NRM to have a significant component of a downward overprint induced during the coring of the seafloor, as is commonly observed by most ODP paleomagnetic studies (summarized in Acton et al., 2002).

An interesting and temporary magnetodiagenetic artifact was observed after the initial thermal demagnetization step (150 or 200C) of relatively unoxidized gray-colored minicores that still retained significant moisture. The magnetic vector measured after this first heating step was often offset from the main trend from NRM to the 5-mT AF step and through the higher heating steps. The colors of the minicores after this initial heating were generally less dark or bleached, and there was no detectable change in magnetic susceptibility. Our explanation for this offset magnetic direction that appears to vanish at the next heating step (200 or 250C) is that the minicore lithologies have undergone creation of goethite-type iron clays as fine particles of iron sulfides or other reactive reduced iron phases undergo heating while surrounded by pore water. The goethite minerals acquired a weak magnetization even from the very low fields in the oven and cooling chamber, and this magnetization was erased upon next stage of heating above the Curie point of goethite (150–200C range). The lack of a significant susceptibility response by goethite minerals is consistent with these observations. This hypothesized goethite artifact was not observed in most minicores collected on board ship that had undergone unanticipated storage at room temperature for 2 months during an extended delay in customs release by Brazil, which suggests that these diagenetic alterations had already taken place prior to our measurement of NRM. Such rapid diagenetic alternation of iron minerals upon sample storage and initial heating steps was also observed when the Leg 207 cores were resampled at the ODP repository at Bremen—in just 2 months since the leg ended, many intervals in these refrigerated sediments had already experienced a change in surface coloration from the original greenish gray to a grayish tan.

Shore-Based Interpretations of Magnetostratigraphy

The polarity and characteristic magnetization for each sample were interpreted from a graphical display of the progressive variation in magnetic vectors (declination, inclination, and intensity). We used the PALEOMAG software package (freeware available from Purdue University at that has a combined graphical and analytical package designed for interpreting magnetostratigraphy of large suites of samples.

Two types of magnetic behavior were observed upon progressive thermal demagnetization (Fig. F7). The "N-type" displays a stable declination and consistently decreasing magnetic intensity as the inclination first undergoes a shallowing then stabilization at a low angle during progressive demagnetization steps (Fig. F7B). This magnetic behavior is consistent with a primary normal polarity direction that has been overprinted by a downward drilling-induced remanence and the present-day normal polarity field. South America has not had a significant rotation relative to the pole since the Cretaceous; therefore, as these overprints are progressively removed the declination remains relatively stable. In general, the NRM declination was within 20 of the characteristic declination observed at the higher steps of thermal demagnetization.

In contrast, the "Hook-type" is characterized by a significant change in declination during the first demagnetization stages and a temporary increase in magnetic intensity when the new stable declination is attained (Fig. F7A, F7C, F7D). The new magnetic vector then decreases in intensity with higher thermal demagnetization steps. We interpret this magnetic behavior as an initial NRM that contained an underlying reversed polarity vector that was partially opposed and rotated by normal polarity overprints. As these overprints are reduced during early stages of demagnetization, the reversed polarity vector is unblocked and increases in intensity until further demagnetization reduces this primary vector.

Approximately 80% of the minicore demagnetization behavior could be assigned as N-type or Hook-type, and these types were generally in stratigraphic clusters that were interpreted as polarity zones. About 20% of the samples yielded uncertain polarities for a variety of reasons, including magnetizations near the background noise level of the cryogenic magnetometer upon early stages of demagnetization, persistent steep downward overprints, or unstable magnetic directions. However, an ambiguous intermediate behavior could be displayed by a sample in which there are is significant normal polarity overprint superimposed on a primary reversed polarity direction. In this case, a misleading N-type response might be displayed. This is one caution to interpreting magnetic polarity of samples solely from either demagnetization curves or inclination data without having orientation with respect to present magnetic north.

Characteristic magnetization directions and associated variances were computed for each sample by applying the least-squares three-dimensional line-fit procedure of Kirschvink (1980), which is also called principal component analysis. The characteristic direction was visually assigned to the set of vectors that during progressive demagnetization appeared to display removal of a single component in equal-area and vector plots. The intensity of characteristic magnetization was computed as the mean of the intensities of those vectors used in the least-squares fit.

Each characteristic direction was assigned a polarity rating based on the individual demagnetization behavior: (1) well-defined N or R directions computed from at least three vectors, (2) less precise NP or RP directions computed from only two vectors or a suite of vectors displaying high dispersion, (3) NPP or RPP samples that did not achieve adequate cleaning during demagnetization but their polarity was obvious—these were omitted from computations of paleolatitudes, and (4) samples with uncertain N?? or R?? or indeterminate (INT) polarity that were not used to define polarity zones. To reduce the bias of a single observer, selection and rating of characteristic magnetization vectors and associated polarity interpretations were generally examined independently by both paleomagnetists.

When both polarity and characteristic inclination is known for a discrete sample, then its magnetic paleolatitude can be computed. In contrast to our expectations prior to Leg 207, the suites of discrete minicores revealed that the sites were at or just south of the paleoequator during most of the Eocene and Paleocene. This southern latitude implies that most of the shipboard interpretations of polarity zones, which were based only on inclination clusters as explained above, were incorrect. During the Maastrichtian and Campanian, the paleolatitudes of the array of sites were slightly north of the paleoequator; therefore, shipboard interpretations of inclination clusters were partially supported. During the Albian, the paleolatitudes of the sites appear to have been at the present latitude (10N) or even farther north. The discrepancies of these paleolatitudes obtained from the arrays of hundreds of minicores with prior estimates based on regional plate motion reconstructions has yet to be resolved.


It was possible to complete all the shore-based paleomagnetic studies (~800 minicores) for this volume only through the generosity of Professor Valerian Bachtadse, Manuela Weiss, and the Institute of Geophysics at the University of Munich, who allowed us to have 3 weeks of full-time usage of their paleomagnetics laboratory facility and adjacent living quarters. The energetic group at the ODP Repository at Bremen sent us >100 additional minicores at the beginning of our shore-based paleomagnetic analyses and then aided us for 4 days in drill-pressing an additional 300 minicores to improve resolution of magnetic reversal boundaries at the midpoint of our analysis marathon. Joint Oceanographic Institutions/US Science Advisory committee (JOI/USSAC) provided us advance funding and a special travel grant to accomplish the postcruise collection and analysis of the minicores in Germany prior to the revision of this Initial Reports volume.