A neutron generator pulsed at 100 s-1 was suspended in an artificial borehole containing a 7.7 metric ton mixture of sand, aragonite, magnetite, sulfur, and salt. Two Ge(HP) gamma-ray detectors were used: one in a borehole sonde, and one at the outside wall of the sample tank opposite the neutron generator target. Gamma-ray spectra were collected by the outside detector during each of 10 discrete time windows during the 10 ms period following the onset of gamma-ray build-up after each neutron burst. The sample was measured first when dry and then when saturated with water. In the dry sample, gamma rays due to inelastic neutron scattering, neutron capture, and decay were counted during the first (150 ??s) time window. Subsequently only capture and decay gamma rays were observed. In the wet sample, only neutron capture and decay gamma rays were observed. Neutron capture gamma rays dominated the spectrum during the period from 150 to 400 ??s after the neutron burst in both samples, but decreased with time much more rapidly in the wet sample. A signal-to-noise-ratio (S/N) analysis indicates that optimum conditions for neutron capture analysis occurred in the 350-800 ??s window. A poor S/N in the first 100-150 ??s is due to a large background continuum during the first time interval. Time gating can be used to enhance gamma-ray spectra, depending on the nuclides in the target material and the reactions needed to produce them, and should improve the sensitivity of in situ well logging. ?? 1984.