Paleomagnetic and rock magnetic measurements at Site 1204, on the summit region of Detroit Seamount, were aimed at assessing the natural remanent magnetization (NRM) of the basement rocks recovered. Data from the site were sought for comparison with results from Site 884 (Tarduno and Cottrell, 1997), located to the northeast on the eastern flank of the seamount. Basalt and sediment discrete samples (minicores) were used for measurements. In addition, the magnetization of sediment recovered above basement in Hole 1204A was measured to define geomagnetic polarity. For magnetostratigraphic analysis, half-round core sections were analyzed.
The sedimentary rocks recovered from Hole 1204A are mostly chalk and altered volcanic ash (see "Lithostratigraphy"). A zone of resedimented chalk is present in Core 197-1204A-3R. NRMs from archive half-round core sections (Cores 197-1204A-1R to 6R) were measured using the shipboard 2-G Enterprises superconducting quantum interference device (SQUID) magnetometer. The measurement interval was 5 cm. Progressive alternating-field (AF) demagnetizations were applied to peak fields of up to 40 mT.
For many intervals this demagnetization treatment was successful in removing secondary magnetizations and in defining a characteristic remanent magnetization (ChRM) of normal and reversed polarity (Fig. F39). From some intervals, the NRM was found to be composed of an apparent present-day (Barton et al., 1995) or Brunhes-age component overprinting a clearly defined reversed polarity ChRM (Fig. F39B). However, for other intervals the direction defined during progressive AF demagnetization deviated significantly from the origin of orthogonal vector plots (Fig. F40). This departure from ideal behavior signifies the presence of a high-coercivity magnetic component. For such intervals, the AF treatment was ineffective in isolating the ChRM and the data cannot be used to assign geomagnetic polarity.
We base the preliminary geomagnetic polarity column on demagnetization data showing a linear decay to the origin of orthogonal vector plots. Specifically, inclination values derived from vector end points after AF demagnetization were used to assign polarity (Fig. F41).
Although rotary drilling can sometimes result in severe disruption of sedimentary rocks, the cores recovered from Hole 1204A generally consist of long consolidated sections with only minor disturbance. Artifacts caused by edge effects of adjacent core pieces, however, are to be expected in the raw data. Our polarity analysis follows the conventions applied in our investigation of sediment cores recovered at Site 1203. A polarity was assigned only if several consecutive measurement intervals demonstrated a consistent polarity; single-point estimates suggesting a polarity opposite to that of adjacent intervals were ignored in this preliminary analysis.
Using the available nannofossil data (see "Biostratigraphy"), we have drawn preliminary correlations of the polarity intervals identified to the geomagnetic polarity timescale. Core 197-1204A-1R has been assigned to the middle Eocene NP15 nannofossil zone. This assignment suggests that the normal polarity interval identified in the core probably corresponds to polarity Chron C21n, which is assigned an age of 47 Ma (Cande and Kent, 1995; Berggren et al., 1995). The majority of Core 197-1204A-2R has been assigned to the NP14 nannofossil zone; the base of the recovered interval is assigned to Zone NP13. These assignments suggest that the normal-reversed-normal polarity succession observed in Core 197-1204A-2R (Fig. F41) corresponds to the C22n-C21r-C21n polarity chron sequence, which has an age between 49 and 47 Ma (Cande and Kent, 1995; Berggren et al., 1995). Sediment recovered from Core 197-1204A-3R represents a slide or slump; because of the likelihood of magnetic resetting resulting from sedimentary processes, we have not considered correlations of the observed magnetic record with the geomagnetic polarity timescale.
Cores 197-1204A-4R and 5R, assigned to nannofossil Zone NP12, correlate well with normal polarity Chron C23, (~51 Ma) (Cande and Kent, 1995). Nannofossil analyses indicate that Core 197-1204A-6R represents a condensed section spanning the late Paleocene to Late Cretaceous. Nannofossil Zones NP8 and NP7 are assigned to the top of the core. A significant hiatus is present at the base of Section 197-1204A-6R-1; sediment lower in the core is assigned to the late Campanian CC22-CC23 nannofossil zone. This assignment suggests that the normal polarity interval identified at the base of Core 197-1204A-6R corresponds to Chron C33n.
The basement section at Site 1204 is dominated by basaltic lava flows (see "Physical Volcanology and Igneous Petrology" and "Discussion"). The lowermost core recovered consisted of volcaniclastic and calcareous marine sediment. All samples for paleomagnetic and rock magnetic analyses were collected as drilled minicores.
Low-field volume-specific magnetic susceptibility (K) was measured with a Kappabridge KLY-2 magnetic susceptibility meter. Magnetic susceptibility values in Hole 1204A range from 0.2 x 10-3 to 26 x 10-3 SI, with an arithmetic mean value of 6.9 x 10-3 SI. For Hole 1204B, magnetic susceptibility values range from 0.06 x 10-3 to 14 x 10-3 SI, with an arithmetic mean value of 3.9 x 10-3 SI. The Koenigsberger ratio (Q) ranges from 5.6 to 45.0 in Hole 1204A (mean = 20.6). In Hole 1204B, Q values range from 0.4 to 31.0 (mean = 7.4). Median destructive fields (MDFs) of NRM were calculated based on stepwise AF demagnetization curves. However, data from some samples did not show a monotonic decrease in intensity, suggesting the presence of multiple NRM components (see Fig. F42; "Discussion"). In particular, the demagnetization data from some samples appear to show a low-coercivity component of magnetization oriented roughly antiparallel to that carried at higher coercivities (see Fig. F43). Because of this complex directional structure, we chose not to calculate MDFs for such samples. The MDF values that could be calculated range from 7.1 to 33.9 mT (mean = 16.4 mT) for Hole 1204A and 7.0 to 47.5 mT (mean = 22.9 mT) for Hole 1204B. Downcore changes in these rock magnetic parameters are shown in Figures F44 and F45. In general, susceptibility and Q-ratio values are lower than those derived from Site 1203 basalt samples.
The Lowrie-Fuller test (Lowrie and Fuller, 1971) was performed on five basalt samples from Hole 1204A and nine basalt and diabase samples from Hole 1204B. In interpreting these data in terms of magnetic domain state, we rely on the relative shapes of the anhysteretic remanent magnetization (ARM) and saturation isothermal remanent magnetization (SIRM) demagnetization curves. For a few samples (e.g., Sample 197-1204A-8R-2, 132-134 cm) (Fig. F46), the two demagnetization curves are similar and clear constraints on magnetic domain state are not possible. In other samples, the decay of ARM relative to SIRM allowed an interpretation, but it should be emphasized that this interpretation also relies on the assumption that the samples contain a simple magnetite or titanomagnetite mineralogy (see "Discussion"). Given this assumption, the rapid decay of ARM relative to SIRM for some samples suggests the presence of multidomain grains (e.g., Samples 197-1204A-107R-1, 125-127 cm, and 197-1204B-4R-2, 63-65 cm) (Fig. F46). In other samples, the resistance of ARM to demagnetization relative to SIRM suggests the presence of single-domain to pseudo-single-domain grains (e.g., Samples 197-1204A-10R-4, 138-140 cm; 197-1204B-9R-2, 8-10 cm; and 13R-3, 33-35 cm) (Fig. F46).
Isothermal remanent magnetization (IRM) acquisition and backfield IRM acquisition were measured on the same samples used for the Lowrie-Fuller tests. These data were used to derive values of coercivity of remanence, which is an estimate of magnetic hardness (or the resistance and, hence, stability of the remanent magnetization to change by external stimuli).
Stepwise IRMs were imparted in the +x-direction (sample coordinates) using an IM-10 impulse magnetizer until the saturation remanence (Mrs) was reached. Then the samples were direct-current demagnetized by giving them a stepwise IRM in the -x-direction. The remanence decreases to zero at the coercivity of remanence (Hcr) and culminates in -Mrs, the negative saturation remanence (Fig. F47). Results of these measurements are summarized in Table T9. Mrs values range from 118 to 365 A/m (mean = 260 A/m) for Hole 1204A and from 115 to 311 A/m (mean = 179 A/m) for Hole 1204B. Hcr values range from 6.0 to 31.5 mT (mean = 20.5 mT) for Hole 1204A and from 4.6 to 34.5 mT (mean = 19.2 mT) for Hole 1204B.
Interestingly, samples that have single domain-like behavior according to the Lowrie-Fuller test have a low coercivity of remanence (e.g., Samples 197-1204A-10R-4, 138-140 cm; 197-1204B-9R-2, 8-10 cm; and 197-1204B-13R-3, 33-35 cm) (Fig. F47). Samples having a multidomain-like behavior according to the Lowrie-Fuller test have higher coercivity of remanence values (e.g., Samples 197-1204A-107R-1, 125-127 cm, and 197-1204B-4R-2, 63-65 cm) (Fig. F47). This pattern is opposite from that expected for a simple magnetite or titanomagnetite remanence carrier.
Basaltic rocks from Holes 1204A and 1204B can be categorized into two types based on visual core descriptions, the rock magnetic properties described above, and observations from reflected-light microscopy (see "Physical Volcanology and Igneous Petrology"). Most samples are light brown and have natural remanent magnetizations that are occasionally resistant to the highest alternating field applied (80 mT) (see Figs. F42, F48, F49). Sometimes brown samples show irregular directional changes during AF demagnetization. Microscopic observation under reflected light reveals that the primary titanomagnetite grains in this type of sample have been partially to completely altered to titanomaghemite. Exsolution lamella of ilmenite were observed only occasionally (see Table T7). Groundmass and vesicles are generally filled with iron oxyhydroxides, including goethite. These lines of evidence suggest that the brown basalt samples were subjected to relatively extensive low-temperature oxidation (see "Alteration and Weathering"). Diffusion of ferric ions from the primary titanomagnetite during low-temperature oxidation (Furuta, 1993) might have contributed to the formation of goethite and amorphous iron oxyhydroxides.
In general, goethite is highly resistant to AF magnetization. Titanomaghemite has a resistance to AF demagnetization that is higher than that of unoxidized titanomagnetite. But a considerable overlap in coercivity between titanomagnetite and titanomaghemite is expected in a natural sample because the coercivities of each mineral phase also depend on grain size. The presence of goethite and titanomaghemite, together with the possibility that the former mineral can carry a magnetization component different from the characteristic remanent magnetization (see "Characteristic Remanent Magnetization Directions and Inferred Paleolatitudes"), explains the inadequacy of AF demagnetization in producing a smooth decrease in NRM for some samples. It also explains the somewhat contradictory results of the Lowrie-Fuller test and the coercivity of remanence data, as the former relies on a simple, one-component magnetite (or titanomagnetite) composition for meaningful assignment of magnetic domain state. Similarly, coercivity of remanence data are difficult to interpret uniquely when multiple rock magnetic components are present.
The second type of basalt sample identified is green gray. These samples show a rapid and smooth demagnetization behavior with the application of progressively higher peak alternating fields (see Figs. F50D, F51B, F51C, F52B, F52D, F52F). The MDFs of the NRM and the coercivity of remanence are relatively low. In Figure F53, a clear contrast of rock magnetic properties between brown and green-gray samples is shown. The green-gray basalt samples are also altered (see "Physical Volcanology and Igneous Petrology" and "Alteration and Weathering") but apparently under reducing conditions. Groundmass and vesicles are filled with blue-green clay minerals, and secondary pyrite is common. Importantly, iron oxyhydroxide is absent.
Reflected-light microscopy (see "Physical Volcanology and Igneous Petrology") suggests that the green-gray samples contain, on average, unaltered titanomagnetite that is larger in grain size than that observed in brown samples. To explain the relatively low MDF and Hcr values from such samples, we speculate that some of the primary fine-grained titanomagnetite (of single-domain size) might have been dissolved under the reducing conditions signified by the presence of secondary pyrite. Accordingly, the magnetic properties of this type of basalt sample might be dominated by the larger multidomain titanomagnetite that survived reductive alteration.
In general, the multiple magnetic minerals present in the basalt samples from Holes 1204A and 1204B highlight the need for detailed thermal demagnetization studies to separate overprints and characteristic remanent magnetizations. Nevertheless, AF demagnetization data from select samples (especially less oxidized samples) provide a first-order means of estimating the potential time sequence recorded by the basalt sections in Holes 1204A and 1204B.
Minicore samples of basalt from Holes 1204A and 1204B were measured using the 2-G Enterprises SQUID magnetometer. A few volcaniclastic samples from Hole 1204B were also studied. After the measurement of the untreated NRM, samples were progressively demagnetized by AF demagnetization. A 5-mT field increment was used between 5 and 50 mT, and a 10-mT step was used after 50 mT to a peak applied field of 80 mT. The demagnetization data from the samples were plotted on orthogonal vector plots to examine the stability and component structure of the NRM.
Approximately 70% of the minicore samples from Holes 1204A and 1204B showed simple demagnetization patterns, with linear vector decay to the origin of orthogonal vector plots after the removal of a viscous overprint (which was generally removed after demagnetization to 5-10 mT) (see Figs. F50, F53). Samples with such stable behavior included the diabase recovered from Hole 1204B (Fig. F51). In ~30% of the samples measured, however, a high-coercivity component was not completely demagnetized by the highest peak AF demagnetization applied. For some of these samples, the demagnetization data defined a linear component (see Figs. F48, F52), which could be fit with principal component analysis (Kirschvink, 1980). Approximately one-half of the samples that showed a high-coercivity component also showed a more complex overall directional pattern during demagnetization. These patterns included the definition of an overprint of apparent opposite polarity to the (generally poorly defined) ChRM and unstable directional behavior during the demagnetization (Figs. F42, F49).
We observed a general correspondence between the directional behavior described above and the rock magnetic sample types discussed above. In general, samples showing a more complex magnetic behavior were more oxidized. The NRM of reduced green-gray samples had a simpler apparent magnetic component structure.
Somewhat surprisingly, we note that two hyaloclastite lapilli breccia samples from Hole 1204B basement Subunit 2d (see "Physical Volcanology and Igneous Petrology") carry a stable remanent magnetization. This magnetization must postdate emplacement of the breccia (Graham, 1949) if temperatures were below the Curie temperatures of potential primary magnetic carriers, as seems to be the case from the overall alteration state of the adjacent units (see "Alteration and Weathering"). Whereas the coercivity range and demagnetization range of basalt samples from Holes 1204A and 1204B are consistent with a thermoremanent origin of the ChRM, we relate the directions isolated from the breccia to a chemical remanent magnetization process. Goethite may be an important remanence carrier in the breccia. We further note that the chemical event responsible for the magnetization of this breccia may also be responsible for the complex magnetic behavior of some basalt samples from both Holes 1204A and 1204B.
All samples analyzed were of normal polarity; a few apparent reversed polarity samples were attributed to accidental inversions of core pieces during sample handling after recovery. ChRM directions for basalt and sediment were fit using principal component analysis (Kirschvink, 1980). In general, the characteristic remanent directions were defined between 20 and 70 mT for basalt samples. Maximum angular deviations of line fits to the demagnetization data were generally <5°-10°. ChRM inclinations from each basement unit and subunit were averaged using the method of McFadden and Reid (1982) to obtain inclination averages.
Nine samples from Hole 1204A yielded an inclination unit average of 55.5° (95% confidence interval = 7.5°) (Table T10). This value yields a nominal paleolatitude of 36.0°N. The paleolatitude is a nominal value because secular variation of the geomagnetic field will not be averaged by a single lava flow unit if cooling is rapid and the magnetization is imparted by a thermoremanent magnetization process.
Three relatively thick lava flow units were described from the Hole 1204B core recovered. Basement Unit 2, however, was further divided into four subunits (see "Physical Volcanology and Igneous Petrology"). Here, we treat each of the subunits as an independent time unit. The exceptionally large vertical extent of Unit 2 (and the associated duration of cooling) (Jaeger, 1964) suggests that the magnetization acquired by the subunits may be separated on a timescale of decades to a few hundred years. Using six independent paleomagnetic units for Hole 1204B, we derive a mean inclination of 58.9° (95% confidence interval = 6.1°) (Table T10). This inclination suggests a paleolatitude of 39.7°N (95% confidence interval = +7.4°/-6.3°).
The overall estimated dispersion of the paleomagnetic data (S = 8.4) from Hole 1204B indicates that some time has been averaged. However, the value is less than that predicted from global lava flow data spanning the 45- to 80-Ma interval (McFadden et al., 1991). This suggests that the sequence has not sampled the full range of geomagnetic secular variation at the site. This conclusion and the estimated paleolatitude rely on the accuracy of the AF demagnetizations applied. Because of the alteration of the basalt section (and especially the potential presence of goethite) in Holes 1204A and 1204B, however, thermal demagnetization studies will be required to confirm the ChRM directions, paleolatitude, and angular dispersion reported here.
The presence of ferromagnetic alteration minerals, however, may offer another opportunity to address the paleolatitude of Detroit Seamount. If overprints can be isolated from data generated by detailed stepwise thermal demagnetizations and if this alteration occurred soon after emplacement of the basalt (as suggested by our preliminary measurements), it might be possible to derive a paleolatitude estimate from the overprint directions. Such overprints often span enough time to average secular variation (e.g., Van der Voo et al., 1978).
Although the data from Holes 1204A and 1204B must be confirmed by shore-based analyses, they yield paleolatitude estimates (36.0° and 39.7°N, respectively) that are compatible with the preliminary data from Site 1203 and the paleolatitude estimate (36°; 95% confidence interval = ±7°N) obtained from analyses of older basalt (81 Ma) recovered at Site 884 (Tarduno and Cottrell, 1997) that average secular variation. The difference between the Site 884 data and the latitude of Hawaii stimulated the paleomagnetic investigation of the fixed hotspot hypothesis during Leg 197.