An 40Ar/39Ar study on biotite and hornblende from baked tonalite gneiss adjacent to a 14-m-wide Kapuskasing dyke in the Chapleau Block of the Kapuskasing Structural Zone employed resistance-furnace step-heating, of bulk-mineral separates, single-grain laser step-heating, and laser spot dating of individual grains in thick section. All three methods yielded concordant results for the biotite, indicating that it has been uniformly overprinted at ca. 2.05 Ga. This date is probably a close estimate of the age of the dyke intrusion. The argon data from the hornblende are more difficult to interpret. All three methods suggest that the hornblende was partially overprinted and then affected by a later influx of excess argon from the external environment. Laser spot dates record an apparent age gradient, extending radially outward from ≈2.25 Ga in the grain cores to 2.4-2.7 Ga in the grain rims. Corresponding 37Ar/39Ar values show no systematic variation across the hornblende, although 38Ar/39Ar ratios display a similar pattern to the age distribution, with lower values in the cores than in the rims. The bulk separate and single grain step-heating spectra mimic a volume diffusion Ar-loss profile that would appear to be inconsistent with the laser-spot dates. Three possible interpretations to explain this apparent discrepancy may involve either experimental difficulties, an effective radius of diffusion which is much smaller than the actual grain radius, or an effective radius of diffusion that is of the order of the grain radius. It is clear that for the 40Ar/39Ar dating of samples that have experienced complex geologic histories, it is important to utilize both resistance-furnace and laser techniques.