The dynamic tensile strengths of four rocks have been determined. A fiat plate impact experiment is used to generate -• 1-/•s-duration tensile stress pulses in rock samples by superposing rarefaction waves to induce fracture. A gabbroic anorthosite and a basalt were selected because they are the same rock types as occur on the lunar highlands and mare, respectively. Although these have dynamic tensile strengths which lie within the ranges 153-174 MPa and 157-1.79 4VIPa, whereas Arkansas novaculite and Westerly granite exhibit dynamic tensile strengths of 67-88 MPa and 95-116 MPa, respectively, the effect of chemical weathering and other factors, which may affect application of the present results to the moon, have not been explicitly studied. The reported tensile strengths are based on a series of experiments on each rock where determination of incipient spallation is made by terminal microscopic examination. These data are generally consistent with previous determinations, at least one of which was for a significantly chemically altered (hydroxylated) but physically coherent rock. The tensile failure data do not bear a simple relation to compressive results and imply that any modeling involving rock fracture consider the tensile strength of igneous rocks under impulse loads distinct from the values for static tensile strength. Generally, the dynamic tensile strengths of nonporous igneous rocks range from -• 100 to 180 MPa, with the more basic, and even amphibole-bearing samples, yielding the higher values.

Half‐space modeling, in the time domain, of broadband seismic records from the Oroville aftershock sequence demonstrates that these events have simple sources. Systematic scaling between the far‐field source duration and seismic moment for the aftershocks studied here and other California earthquakes confirms Brune's (1970) source model and shows that stress drops for California earthquakes range between 10 and 100 bars. These events also show no trend in stress drop as a function of size for 3.0 ≤ ML ≤ 6.4. Analysis of static deformation associated with the main shock shows that slip occurred on a fault 81 km2 in area corresponding to the zone defined by the first 3 days of aftershocks. Teleseismic studies indicate that the seismic source area was 50 km2, corresponding to a zone completely free of aftershocks in the middle of the aftershock zone. Both results imply a stress drop of 29–39 bars and seismic moment between 5.7 and 8.7×1024 dyn cm. The evidence suggests that faulting occurred seismically on the central portion of the fault and propagated out slowly, enlarging the fault plane. Complicated waveforms recorded for the same events at nearby sites on alluvium indicate the critical role that propagation through complex structure plays in determining the character of the observed seismic signal at close‐in stations. Modeling of these records shows that dipping structure may account for the observed waveform complexities.

Summary. The Philippine earthquake of 1976 August 16, is one of the largest to have occurred world‐wide in recent years (Mw=8.8; Ms=7.8; seismic moment, Mo= 1.9 × 1028dyne‐cm). It is, however, associated not with the Philippine Trench, which is the dominant tectonic feature along the eastern Philippine Islands, but with a much less prominent trench system in the Moro Gulf, North Celebes Sea, south of Mindanao. In this area most of the seismicity is at depths greater than 500 km, associated with the westward dipping Benioff zones of the Sangihe and Mindanao arc systems. This event, however, has a shallow focus and caused a locally destructive tsunami. The focal mechanism of the mainshock determined in this study from long‐period surface and body waves indicates a predominantly thrust mechanism with strike N 33°W, dip 22° NE and rake + 68°. A significant amount of directivity, which can be seen in the observed surface wave seismograms, is explained very well if the source rupture propagates 160 km unilaterally in an azimuth of 300° from the mainshock hypocentre, with rupture velocity 2.5 km/s. The largest aftershock (Ms= 6.8) occurred outside the main aftershock area 12 hr following the mainshock and apparently resulted from motion on a subsidiary fault since the P‐wave first motion data indicate strike‐slip motion for this event. Bathymetric data indicate the presence of a trench striking north‐south in the region of the Moro Gulf, and seismic reflection profiling indicates disturbed sediments east of the trench showing evidence for subduction. In addition, the geological structures mapped on the island of Mindanao are consistent with this mode of deformation. The only other known large earthquake in the region on 1918 August 15 (Ms=8.0) probably occurred along the same subduction zone, on an adjacent segment, to the south of the recent event. The 1976 August 16 Philippine earthquake thus represents the first clear seismic evidence for a north‐east dipping subduction zone beneath Mindanao in the Moro Gulf, North Celebes Sea.

Measurements of slip on major faults in southern California have been performed over the past 18 yr using principally theodolite alignment arrays and tautwire extensometers. They provide geodetic control within a few hundred meters of the fault traces, which complements measurements made by other techniques at larger distances. Approximately constant slip rates of from 0.5 to 5 mm/yr over periods of several years have been found for the southwestern portion of the Garlock fault, the Banning and San Andreas faults in the Coachella Valley, the Coyote Creek fault, the Superstition Hills fault, and an unnamed fault 20 km west of El Centro. These slip rates are typically an order of magnitude below displacement rates that have been geodetically measured between points at greater distances from the fault traces. Exponentially decaying postseismic slip in the horizontal and vertical directions due to the 1979 Imperial Valley earthquake has been measured. It is similar in magnitude to the coseismic displacements. Analysis of seismic activity adjacent to slipping faults has shown that accumulated seismic moment is insufficient to explain either the constant or the decaying postseismic slip. Thus the mechanism of motion may differ from that of slipping faults in central California, which move at rates close to the plate motion and are accompanied by sufficient seismic moment. Seismic activity removed from the slipping faults in southern California may be driving their relatively aseismic motion.