Because of bad weather conditions, no downhole measurements could be achieved during Leg 163. But the high recovery of basalt (70% overall core recovery) from Hole 990A enabled a dense sampling of physical properties measurements, including P-wave velocity, bulk density, natural gamma ray, magnetic susceptibility, and natural remanent magnetization. Gamma-ray attenuation porosity evaluator (GRAPE) density, natural gamma-ray, and magnetic susceptibility were measured with the shipboard multisensor track (MST) system on full-round cores with a sampling interval of 2 cm prior to sawing. As soon as possible after core retrieval, P-wave velocity was measured on seawater-saturated half-round cores with the Hamilton Frame velocimeter at an average sampling interval of 5 cm. Additionally, bulk density and velocity were measured on discrete samples (minicores with 1-in diameters and 1-in heights).

The magnetic intensities were measured every 10 cm with the shipboard cryogenic magnetometer. Prior to the measurement, an alternating field demagnetization at 30 mA was applied to remove secondary remanences, like drilling-induced magnetization (Y. Nakasa, pers. comm., 1996).

All measurements were checked by an intensive quality control, using the visual core descriptions (VCDs) (Duncan, Larsen, Allan, et al., 1996) for comparing and editing the data. During the MST measurements, data were taken at each predefined interval, whether there was full-round core or some core loss. This results in erroneous data at the measurement points where there was some core loss. By comparing the core measurements with the images from the core, it was possible to delete all those measurements where no full-round core was present. This quality control resulted in reliable physical properties values that do not show the minimum values produced by partial or total loss of core. Additionally, the shipboard GRAPE data had to be recalibrated using the results of the minicore density measurements. In contrast to the GRAPE density measurements, which are based on the attenuation of gamma rays traveling through the core, the minicore density measurements (buoyancy method) give absolute data. The Hamilton Frame Velocimeter measurements are in good agreement with the shipboard minicore measurements, whereas the swelling properties of smectite clays have permanently lowered the elastic properties of stored samples (S. Planke et al., unpubl. data). The results of the measurements after this quality control are shown in Figure 2 together with flow types and flow boundaries. The detailed structure of the volcanic pile is revealed by the very high sampling rate.

Flow textures were studied through both qualitative and quantitative methods. Qualitative core descriptions involved hand-specimen and thin-section inspection of core samples for both vesicularity and general groundmass characteristics. Selected samples from the different flow types were then chosen for quantitative analysis of both vesicle and crystal textures. Groundmass crystal textures measurements were based on backscattered electron (BSE) images collected using a JEOL 6300 SEM located at the University of Oregon. Optimal resolution was obtained with a 15-mm working distance, a 10-keV accelerating voltage, and a 5-nA beam current. Images of all samples were collected at a magnification sufficient to permit counting of 200 to 300 crystals per image. Because of the difficulties of working with holocrystalline samples, measurements were restricted to determination of crystal number density (the number of crystals per area).

Aa and pahoehoe flows are uniquely characterized by groundmass textures (Katz, 1997) that result from their very different flow transport and cooling histories (Ho and Cashman, 1997). The smaller average crystal size in the aa groundmass results from higher rates of cooling (>20º-30ºC/h) (Ho and Cashman, 1997), and thus higher rates of crystal nucleation (~1.0 × 10⁴/cm³/s) than in the better insulated, often slower moving pahoehoe flows. Groundmass textures thus provide a good diagnostic tool for identifying flow type when identification based on internal structure is ambiguous. Aa flows typically have 10 to 100 times the number of crystals in a given area as pahoehoe flows do, meaning that the average crystal size of pahoehoe flows is 3-10 times larger than that of aa.

A comprehensive petrological and structural summary of the volcanic succession is given by Duncan, Larsen, Allan, et al. (1996). The groundmass of all basaltic rocks consists of plagioclase (30%-50%), augite (20%-42%), magnetite (1%-4%), olivine (0%-4%), and mesostasis (0%-30%). All rocks from Hole 990A are moderately to completely altered by low-temperature secondary phases that replace primary minerals. The highly brecciated or vesicular flow tops especially show a high degree of alteration (see also Planke et al., Chap. 2, this volume). No significant differences in either primary mineralogy or phenocryst content were found between aa and pahoehoe flows (Duncan, Larsen, Allan, et al., 1996).

The lava flows cored at Site 990 range from aphyric to highly olivine or plagioclase-olivine-clinopyroxene phyric basalt. Most units are moderately phyric, although Units 6, 7, and 9 are aphyric olivine basalt. The flow thicknesses in general tend to increase upward, with units 1-3 reaching thicknesses of about 15 m. All recognized lava flows were subdivided into the aa, pahoehoe, and transitional categorizations described above (Duncan, Larsen, Allan, et al., 1996). Aa flows were identified by their thick brecciated flow tops (up to 0.5 m) and thin vesicular flow base (0.2-0.5 m). In contrast, pahoehoe flows showed thick upper vesicular zones and massive interiors, and they often had pipe vesicles at the base. Flows that showed some characteristics of both end-members were classified as transitional. Duncan, Larsen, Allan, et al. (1996) classified Units 1-3 and 13 as aa, 4-7 as transitional, and 8-12 as pahoehoe.

Measured physical and magnetic properties are shown in Figure 2 together with flow types, flow boundaries, and rock types determined on board ship (see Duncan, Larsen, Allan, et al., 1996). Because of the very high data density, a detailed structure of the volcanic pile can be seen.