During the month of October 1985 a second experiment was undertaken in the Charlevoix seismic zone to further test the hypothesis that shear-wave splitting could be observed in a seismically active region. The first experiment had been undertaken in 1984 and yielded only a very limited amount of data. Seismograms recorded by digital three-component seismographs located very close to the epicentres of seven earthquakes showed shear-wave splitting over 15 different paths. The amount of wave variation varied from about 24 to 160 ms or from 0.4 to 8.7% of the wave velocity. The largest value occurred over the shortest path of about 7 km, for which essentially the whole path may be anisotropic, leading to a crack density (ε) of less than 0.09. For the rest of the data, all with less than 3% shear-wave-velocity variation, ε varies from 0.005 to 0.03, if whole-path anisotropy is assumed. These values of ε are not significantly different from those obtained in 1984. The average azimuth of the initial shear-wave polarization is 37°, also similar to that observed in 1984. All the data in the zone can be explained by the presence of saturated vertical cracks striking 37 °east of north.