The activity associated with the Congo earthquake of March 20 1966 (mb = 6.5 to 7) was studied with emphasis on the time and magnitude distributions. The data were recorded at the Abéché, Chad, seismograph station operated by Lamont-Doherty Geological Observatory. Over a period of about 70 days, 815 earthquakes with magnitude (mb) greater than or equal to 3.3 were recorded, and they form the basis for this study.
The aftershocks are distributed with magnitude (mb) according to the formula long n = a - bm with b = 1.05 ± 0.07 at the 95 per cent confidence limits. The foreshocks have b = 1.06 ± 0.35 at the 95 per cent confidence limits. These b values are in general agreement with b values derived from other aftershock sequences throughout the world.
Some authors have suggested that foreshocks may have a lower b value than background activity and that this difference might be used in earthquake prediction. In this paper, an evaluation is made of the limitations of this method of prediction. Assuming that such a difference in b values does exist, it is found that a closely spaced network of high-gain seismographs with wide dynamic range would be required to assure successful prediction.