Fig. 3. Plots of neutron log response vs. measured He porosity and - log response vs. measured bulk porosity. Regression fits were uniformly poor for these traditional log-to-physical property correlations. Samples were dried 2.5 by 2.3 cm cylinders taken from basalt cores, from locations in the saturated and vadose zones, respectively. Porosity was determined using a He porosimeter. Bulk density was determined from multiple measurements of physical dimensions and mass. Neutron and - log data used for the regression analyses were averaged over 30 cm. Gamma- data was from the "long-spacing" (LS) source-detector pair, recorded in counts per second. Neutron data was recorded in API-N units. Regression analyses were made for both neutron and - data, from the upper vesicular zones of basalt flows (red), lower vesicular zones of basalt flows (green), all vesicular basalt (purple), massive interiors of basalt flows (blue), the massive interiors of basalt flows with outliers points removed (brown, - only), and the near-vent facies of thinly bedded, fractured basalts and shelly pahoehoe (black). Neutron data regression coefficients for aquifer basalts were: upper vesicular, R = 0.0265; lower vesicular, R = 0.4121; all vesicular, R = 0.0200; near-vent, R = 0.9042; and massive, R = 0.4798. Neutron data regression coefficients for vadose basalts were: upper vesicular, R = 0.1233; lower vesicular, R = 0.1277; all vesicular, R = 0.0300; near-vent, R = 0.0970; and massive, R = 0.0424. Gamma- data regression coefficients for aquifer basalts were: upper vesicular, R = 0.2827; lower vesicular, R = 0.2470; all vesicular, R = 0.0332; near-vent, R = 0.1166; massive, R = 0.4592; and massive-no outliers, R = 0.7270. Gamma- data regression coefficients for vadose basalts were: upper vesicular, R = 0.0100; lower vesicular, R = 0.0906; all vesicular, R = 0.0624; near-vent, R = 0.0245; massive, R = 0.2903; and massive-no outliers, R = 0.3250.