Summary. The Atlantic segment of the Africa–Europe plate boundary has usually been interpreted as a transform boundary on the basis of the bathymetric expression of the Gloria fault and dextral strike‐slip first‐motion mechanisms aligned along the Azores–Gibraltar line of seismicity. The 1975 May 26 earthquake (Ms=7.9) was assumed to fit into this framework because it occurred in the general area of this line and has a similar first‐motion focal mechanism (strike=288°, dip=72°, slip angle=184°). However, several anomalies cast doubt on this picture: the event is abnormally large for an oceanic transform event; a sizeable tsunami was excited; the aftershock area is unusually small for such a large event; and most significantly, the epicentre is 200 km south of the presumed plate boundary. The Rayleigh wave radiation pattern indicates a change in focal mechanism to one with a significant dip‐slip component. The short duration of the source time history (20 s, as deconvolved from long‐period P‐waves), the lack of directivity in the Rayleigh waves, and the small one‐day aftershock area suggest a fault length less than 80 km. One nodal plane of the earthquake is approximately aligned with the trace of an ancient fracture zone. We have compared the Pasadena 1‐90 record of the 1975 earthquake to that of the 1941 North Atlantic strike‐slip earthquake (200 km to the NNW) and confirmed the large size of the 1941 event (M=8.2). The non‐colinear relationship of the 1975 and 1941 events suggests that there is no well‐defined plate boundary between the Azores and Gibraltar. This interpretation is supported by the intraplate nature of both the 1975 event and the large 1969 thrust event 650 km to the east. This study also implies that the largest oceanic strike‐slip earthquakes occur in old lithosphere in a transitional tectonic regime.

In order to investigate the applicability of the asperity model to intraplate earthquakes we have studied the rupture process of the great 1977 Sumba normal‐faulting earthquake (Mw=8.2–8.3), one of the largest earthquakes since 1963. This event has been variously interpreted as a plate detachment event (i.e., rupture through the entire lithosphere) or a shallow plate‐bending event. We have analyzed long‐period body waves in order to determine the spatiotemporal rupture characteristics, including the depth extent of the rupture area. Visible depth phases in the initial stage of weak radiation beginning about 15 s before the main pulse suggest that the earthquake nucleated near 29 km depth, which corresponds to the maximum depth of the aftershock sequence. The excitation of a strong tsunami implies that rupture extends to the surface, while deconvolutions of long‐period P waves suggest a relatively shallow (≤ 50 km) maximum depth extent. However, the existence of slip below this depth cannot be ruled out, particularly if it preferentially radiates frequencies below the passband of the body waves used (< 0.02 Hz). For undiffracted phases the source time function of the principal rupture, which initiated 30 km west of the epicenter of the precursor, comprises a dominant, initial short‐duration (24 s) triangular pulse followed by several smaller pulses. However, slightly diffracted phases indicate a simple basic (underlying) character for the rupture, consisting of the initial 24‐s triangle superimposed on a longer‐duration (42–44 s) trapezoid. The absence of observable directivity in the deconvolved source time functions suggests a bilateral rupture mode. Although the body waves do not require a fault length greater than 80 km, the aftershock area and surface wave results of Zhang and Kanamori [1988] imply a total length of at least 200 km. The body wave source functions are well modeled by a 200‐km‐long fault, extending to 30–50 km depth, on which slip nucleates at the center of the fault and has variable displacement along strike. The peak moment release of the deconvolved source functions indicates higher displacement in the hypocentral area than in other areas of the fault. The spatial relationship of the precursor and the main pulse is suggestive of rupture initiating at and propagating along the edge of an asperity prior to the rupture of the asperity itself. The large size of the event may be due to an unusual stress environment arising from strong lateral gradients in seismic interplate coupling along the Java‐Timor arc.

The Middle to Late Cambrian loop in the North American apparent polar wander path (APWP) has been variously attributed to tectonic rotations, remagnetizations and primary magnetizations. Although no primary thermal remanent magnetizations or primary detrital remanent magnetizations have as yet been demonstrated, the temporally self-consistent nature of the loop has been used as an argument for primary magnetizations. We have studied 535 + 5 Ma nepheline syenites and syenites of the McClure Mountain alkalic complex, as well as 495 ___ 10 Ma red trachyte dikes which intruded the complex, in an effort to find a primary TRM. Because Zijderveld analysis yielded consistent results for only one trachyte dike, remagnetization great-circle analysis was employed, giving a pole for the trachyte dikes at the tip of the loop (43°N, l14°E), while the syenites and nepheline synenites gave a pole at the base of the loop (18°N, 142°E). The magnetic carrier in the trachytes is hematite which apparently formed during a pervasive hydrothermal alteration. K-Ar whole rock dating of the trachytes suggests a Pennsylvanian age for the alteration, and thus a late Paleozoic remagnetization of the trachytes. Thus, the low-latitude Cambrian pole is confirmed, but we find no evidence in this study to support the primary nature of the Cambrian APWP.

No abstract

F‐k analyses of short‐period recordings of both U.S. and Soviet underground nuclear explosions at the Eskdalemuir (EKA), Scotland array indicate arrivals from a local source about 15 km due northwest of the array. Use of residual seismograms derived by subtracting the beamed record from each array channel confirmed the presence of the same scatterer. The secondary seismic source appears to be short‐period surface waves due to the scattering of incident P waves at a nearby deep valley known as Moffat Water, in agreement with Key's (1967) results. Similar analyses of the NORESS (Norway) array data from U.S. and Soviet nuclear shots also indicate secondary arrivals suggesting a local near‐surface scatterer about 25–30 km southwest of the array, in the region of Lake Mjosa. Thus, short‐aperture array data can be useful in identifying and locating sources of near‐receiver scattering.