A recent reexamination by Rogers and Hasegawa of the available seismic data from the June 23, 1946 Vancouver Island earthquake (MS = 7.2) indicates that the earthquake was of relatively shallow (30 km or less) focal depth and the epicenter was located in central Vancouver Island rather than beneath the Strait of Georgia some 30 km or more to the east as previously thought. We have tested the Rogers-Hasegawa solution by resurveying a triangulation network in the epicentral area which had first been surveyed in 1935. The distortion of the network was found to be greater than could be accounted for by either secular strain accumulation as indicated by measurements of a nearby network or survey error but is consistent with oblique slip on a section of the Beaufort Range fault, a prominent fault that crosses the triangulation network. The best model for slip on the Beaufort Range fault involves 1.00 ± 0.25 m right-lateral and 2.50 ± 0.65 m normal slip on a shallow (0 to 5 km) segment dipping 70° NE. However, pure right-lateral slip of about 1 m over a depth interval 0 to 20 km on a vertical fault is not excluded at the 90 per cent confidence limit. Thus the geodetic data support the conclusions of Rogers and Hasegawa that the 1946 earthquake was caused by right-lateral, normal slip on the Beaufort Range fault in the vicinity of Forbidden Plateau, central Vancouver Island.