⁴⁰Ar/³⁹Ar age spectra and ⁴⁰Ar/³⁶Ar vs ³⁹Ar/³⁶Ar isochrons were determined by incremental heating for 11 terrestrial rocks and minerals whose geology indicates that they represent essentially undisturbed systems. The samples include muscovite, biotite, hornblende, sanidine, plagioclase, dacite, diabase and basalt and range in age from 40 to 1700 m.y. For each sample, the ⁴⁰Ar/³⁹Ar ratios, corrected for atmospheric and neutron-generated argon isotopes, are the same for most of the gas fractions released and the age spectra, which show pronounced plateaus, thus are consistent with models previously proposed for undisturbed samples. Plateau ages and isochron ages calculated using plateau age fractions are concordant and appear to be meaningful estimates of the crystallization and cooling ages of these samples. Seemingly anomalous age spectrum points can be attributed entirely to small amounts of previously unrecognized argon loss and to gas fractions that contain too small (less than 2 per cent) a proportion of the ³⁹Ar released to be geologically significant. The use of a quantitative abscissa for age spectrum diagrams is recommended so that the size of each gas fraction is readily apparent. Increments containing less than about 4–5 per cent of the total ³⁹Ar released should be interpreted cautiously. Both the age spectrum and isochron methods of data reduction for incremental heating experiments are worthwhile, as each gives slightly different but complementary information about the sample from the same basic data. Use of a least-squares fit that allows for correlated errors is recommended for ⁴⁰Ar/³⁶Ar vs ³⁹Ar/³⁶Ar isochrons. The results indicate that the ⁴⁰Ar/³⁹Ar incremental heating technique can be used to distinguish disturbed from undisturbed rock and mineral systems and will be a valuable geochronological tool in geologically complex terranes.