All cores drilled during Leg 176 were measured using the natural gamma ray (NGR) logger on the multisensor track system (MST) at intervals of 4 cm with a time period of 4 s. Results were output in counts per second (cps). We recorded 22,940 points with a mean value of 12.58 ± 1.52 cps. Count rates did not vary significantly between zones of intact core, zones of broken core, and empty liner sections. Thus, the native radioactivity of the gabbroic rocks recovered from Hole 735B was indistinguishable from the background radiation limits within the core laboratory of JOIDES Resolution. Nevertheless, we continued NGR recording so as not to miss any zone of markedly different radioactive properties. However, no such zone was ever encountered. The results from the NGR system therefore do not greatly assist geological interpretation of Hole 735B.
All Leg 176 cores were measured using the gamma-ray densiometry logger (GRAPE) on the MST at intervals of 4 cm with a time period of 4 s, with results output in grams per cubic centimeter (g/cm3). Again, 22,994 points were recorded with a mean of 2.20 ± 0.81 g/cm3.
The GRAPE system used on JOIDES Resolution is designed chiefly to measure sedimentary materials that fill the core liner along its length. Rotary-cored hard rocks are of a significantly narrower diameter than the core liner; this diameter also varies considerably along the length of the core. Consequently, the GRAPE system cannot compensate for the smaller-diameter core and therefore underestimates the density of the material. The changing diameter of the core is not directly recorded in any way, preventing a simple correction of the data for the smaller diameter of hard-rock cores.
GRAPE data are recorded continuously along core sections and can be acquired over solid cores, broken cores, and empty core liner. Before analysis, GRAPE data should be filtered with respect to the piece log. A comparison of the GRAPE density points to the measured densities of the nearest minicores (see "Index Properties") shows a very poor correlation (Fig. F118). There is no significant trend other than a systematic shift of ~0.5 g/cm3 between the two data sets.
Magnetic susceptibility is particularly sensitive to grains of magnetite larger than ~10 µm and can be used to identify iron-rich zones in the rock, such as oxide-rich gabbros and felsic veins. Magnetic susceptibility is also used to compute the Königsberger ratio, which is the ratio of remnant to induced magnetization in the rock. Magnetic susceptibility values were acquired on the MST at 4-cm intervals for the cored material. The Bartington response function of the MS2C meter used for measuring magnetic susceptibility is shown in Figure F119.
Analysis of the MST magnetic susceptibility data provides an interesting correspondence with the petrology, but there are some conventions that need to be mentioned. First, the readings from the MST saturate at machine values greater than 10,000. For example, a true value of 12,000 would appear as 2000 in the data set. Care should be taken in interpreting these saturated values. Second, machine values are converted to volume-normalized SI units by multiplying by 0.66 × 10-5. This is a geometrical factor that depends on the fact that the cores have a radius of 66 mm, the radius of a full core liner. Values in SI units are incorrect to the extent that the core, or rubble, in the liner is less than this diameter. Also, spot measurements of whole-round core give diameters of 58.5 to 59.5 mm, less than the nominal 66 mm. The third issue is perhaps most important. The MST takes a measurement every 4 cm from the top to the bottom of a section. Sometimes there is no rock at the measurement point. For example, there are gaps at the plastic dividers between pieces. Also, when rubble is acquired it is unevenly distributed in the liner. Finally, we measured the half-width of a point element of susceptible material (a steel nail) during the cruise on the MST, and it was about 5 cm. For a rock of even the nominal 66 mm radius to yield its true volume-normalized value requires a sample at least 10 cm long. Data listed in Table T14 (also in ASCII format in the TABLES directory) have been cleaned only to the extent that obvious errors, such as inverted pieces of rock, have been deleted. Otherwise, all data are reported in machine units and without regard to core diameter.
In Figure F120, we plot the magnetic susceptibility of the first data set (Table T14, also in ASCII format in the TABLES directory) as a function of depth. In this linear plot, clear spikes in magnetic susceptibility occur at depths where oxide minerals, notably magnetite, are present in the rock. The figure shows more than 20,000 data points acquired over the 1-km interval drilled during Leg 176. The number of spikes decreases with depth, indicating a decrease in the frequency of occurrence of oxide-rich intervals with depth (see Fig. F108). The figure also shows that the predominant mode of occurrence of oxide gabbros is as vary narrow seams, many of them a little thicker than the "point source" used to produce a 5-cm "half-width" response.
Magnetic susceptibility is a useful tool in identifying the petrological properties of the rock. With a value every 4 cm, the data density is greater than thin-section or chemical analysis of core can provide, and the resolution is better than can be conveniently handled by visual inspection. Magnetic susceptibility can help the petrologist identify thin or obscure regions of magnetite content. This is demonstrated in Figure F121, where we show the susceptibility as a function of depth for a single core about 9.5 m long. This is a preliminary study, but at first glance there is a striking correlation between spikes in the magnetic susceptibility, which exceed 2000 machine units, and both narrow intervals rich in oxide minerals and veins as identified in the vein log (see "Igneous Petrology"). In addition to the spikes, there is a deterministic low-amplitude signal (less than about 1000 machine units) with a wavelength of a few meters that may reflect small but systematic variations in low proportions of oxide minerals in such rocks as olivine gabbros and troctolites. This background level also appears to diminish with depth.
Mass and volumetric measurements were made on 218 minicores at irregularly spaced intervals (Table T15, also in ASCII format in the TABLES directory). Measurements were performed as described in "Index Properties" in the "Explanatory Notes" chapter. Measurements of wet and dry mass were begun on the Scitech balance within the physical properties laboratory. After completion of ~50 measurements, we concluded that the balance was unable to measure masses in excess of 30 g to the required precision; this limit included the majority of the minicores measured up to that point. Samples measured after this discovery were measured for both wet and dry mass using a more accurate balance. The earlier samples were also remeasured dry, resaturated, and remeasured wet. We could not demonstrate whether previously dried cores resaturated.
Figure F122 shows the full data distribution for bulk and grain density minicores measured during Leg 176. The mean bulk density for the Leg 176 minicores is 2.979 ± 0.10 g/cm3 with the mean grain density of 2.991 ± 0.107 g/cm3. The mean porosity for the Leg 176 minicores is 0.649 ± 2.884%; the population is heavily skewed by both the large number of minimal porosities and the small number of porosities significantly greater than 1%. Direct comparison with the index properties recorded during Leg 118 is not possible because the bulk densities recorded for that leg were calculated using wet volume measurement. The values are, however, very similar.
Table T16, shows the physical properties of the principal lithologies of Hole 735B computed from both Leg 118 and 176 samples. Discussion of variation of density with lithology is problematic because 219 of the 347 minicores measured were of olivine gabbros. Oxide gabbronorite and oxide gabbro have higher densities (3.09 ± 0.08 g/cm3 and 3.21 ± 0.06 g/cm3, respectively) than olivine gabbros (2.96 ± 0.10 g/cm3) owing to the high density of oxide minerals. Two samples have densities >3.4 g/cm3. They cannot be closely correlated with zones of alteration seen in thin sections, and there is no obvious reason for these very high densities recorded in the shipboard data.
Comparison of the density profile with the interpreted VSP results for Hole 735B from Leg 118 (Swift et al., 1991) shows some coincidence (Fig. F123). The interpreted reflector at 225-250 mbsf (meters below seafloor) has a significant higher density than surrounding rocks because of the abundance of oxide mineral in Unit IV, Massive Oxide Gabbro (Shipboard Scientific Party, 1989). This also occurs to a lesser degree in the amphibolite shear zone at 50-70 mbsf, where there is also a pronounced mineral fabric. However, no similar zones of mineralization could be seen at the levels of the other two reflectors encountered by Leg 176 coring (560 and 760-825 mbsf). It therefore seems likely that these two reflectors represent structural changes, such as large-scale zones of fracturing, rather than changes in mineralogy. This is further suggested by the low percentage of recovery (Fig. F124) across these two zones.
Densities in the Leg 176 cores are less variable than in the upper 500 m. The density profile (Fig. F123) for the Leg 176 cores can be divided into three main regions. Between 500 and ~780 mbsf the rocks show considerable variation in both lithology and bulk density (between 2.8 and 3.3 g/cm3). Below 780 mbsf the rock composition is largely olivine gabbro with minor gabbro and orthopyroxene-bearing gabbro. These rocks show a far more consistent range of densities (mostly between 2.8 and 2.9 g/cm3). Below 1200 mbsf the rock is almost exclusively olivine gabbro, but increasing variations in the relative proportions of plagioclase and olivine in olivine gabbros and troctolites cause a greater scatter of densities (between 2.8 and 3.0 g/cm3). The mean densities of the deepest rocks are slightly lower than those above owing to the increasing proportion of plagioclase.
Thermal conductivity measurements were made at 219 irregularly spaced intervals through the borehole section. The measurements were made using the TK04 meter (see "Physical Properties" in the "Explanatory Notes" chapter). To minimize the experimental error, five measurements were made on each sample and the mean value was recorded. Where the five values of the measurements varied by more than 0.2 W/(m·K) the sample was remeasured. The measurement needles were tightly strapped to samples during measurements to ensure a good thermal contact; the most significant source of error during these experiments was the strap coming loose during the measurement cycle.
The thermal conductivity of 219 data points over the section is 2.276 ± 0.214 W/(m·K) (Table T17, also in ASCII format in the TABLES directory). This compares with a mean value obtained during Leg 118 of 2.21 ± 0.22 W/(m·K). During Leg 118, thermal conductivity measurements were made both parallel and perpendicular to the axis of the core. We could not make similar measurements during Leg 176 because the needles used by the TK04 meter are 10 cm in length, which is greater than the width of the sawed core face.
Two felsic veins sampled for thermal conductivity had a mean value of 1.98 W/(m·K), a significant contrast to the host rock (shown in the "Others" lithology column in Table T16). Troctolites also have lower thermal conductivities than most of the rocks, with a mean conductivity of 2.100 ± 0.100 W/(m·K). Figure F125, on which thermal conductivity is plotted vs. depth, shows no evidence for the threefold division of the borehole seen in the density profile.
Seismic velocities are the basis of the first-order description of ocean crust in terms of Layer 2 (2500-5500 m/s), Layer 3 (6500-7000 m/s), and the mantle (7800-8300 m/s). Attempts to correlate the layers with lithology are constrained by observations of compressional velocities in minicores under pressure. The gabbros acquired from Hole 735B during Legs 118 and 176 clearly have velocities representative of seismic Layer 3.
During Leg 176, the compressional velocities of 217 minicores averaged 6777 ± 292 m/s (Table T18, also in ASCII format in the TABLES directory). This value is comparable to the velocities measured during Leg 118, is representative of seismic Layer 3 in the ocean crust, and is typical of gabbro and metagabbro. The velocities were measured at room temperature and pressure. Studies of the Leg 118 cores showed that increasing pressure to values appropriate for the in situ lower crust raises velocities by as much as 400 m/s (Iturrino et al., 1991). Cracking and fracturing in the gabbros at a scale larger than the minicores, as has been observed in this hole, would act to decrease the velocity that would be observed on seismic refraction experiments.
Compressional velocities measured from minicores show no marked changes in the vicinity of the seismic reflectors interpreted from the Leg 118 VSP except for the massive oxide gabbro interval Unit IV (Fig. F126). The smoothing of the cut faces of the minicores, required for accurate velocity measurements, stresses the samples and causes breakage in fractured rock. Therefore, samples are specifically chosen in cohesive rather than fractured rock to prevent this failure. The minicores taken from the zones of these reflectors, therefore, are not representative of the whole rock, but of undeformed rock, either analogous to the rocks before the deformation episode or lithologies that are resistant to deformation. The low core recovery in the reflector zones indicates that the seismic reflectors are zones of fractured rock. Once again, Figure F126 does not show the threefold division of the borehole seen in the density profile.
A knowledge of the electrical resistivity of the lithologies represented by minicores is necessary to infer porosity from electrical resistivity logs.
A few test measurements of electrical resistivity on minicores were made. These were all based on the "Meissner" cell, a device designed and built on board JOIDES Resolution. The device uses spring pressure to attach electrodes to either face of a minicore. The faces of the core are saturated with seawater to improve the contact of the electrodes. The sides of the core are wrapped in insulating Teflon tape to restrict the current from passing down the wet sides of the minicore. The resistance between the electrodes was measured on a Wayne-Kerr 6424 meter using differential inputs. Only aluminum standards were available to calibrate the device. These are inappropriate because aluminum is a near perfect conductor and dry igneous rocks are near perfect insulators. Better standards are required in the future for testing these devices
Straight DC resistance
measurements were performed. A small number of minicores (20) were tested, giving results
of about 3600 
m, comparable to
logging results from Leg 118 (Table T19, also in
ASCII format in the TABLES directory).
However, values were both time and frequency dependent. As there was considerable
variability in our results and no way of calibrating the system, there was little point in
continuing with the measurements.
