Description: Publication date: 15 September 2018 Source: Earth and Planetary Science Letters, Volume 498 Author(s): M. Heath, D. Phillips, E.L. Matchan Since its inception in the mid-1960s, the 40 Ar/ 39 Ar dating technique has been the premier method for determining the eruption ages of basaltic rocks, providing valuable insights into a plethora of terrestrial and planetary processes. Advances in multi-collector mass spectrometry and improved sample preparation procedures are enabling ever-improving analytical precision and clearer evaluation of the isotopic disturbances that affect many basaltic samples and cause discordant 40 Ar/ 39 Ar age spectra. Here, we present 40 Ar/ 39 Ar step-heating data for multiple samples from two Quaternary basalt flows (0.8038 ± 0.0017 and 2.309 ± 0.009 Ma) of the intraplate Newer Volcanic Province, southeast Australia. A small proportion of these samples give concordant 40 Ar/ 39 Ar results, but most are variably discordant. The factors controlling these disturbances and implications for accurate age determination are examined and modelled in both step-heating spectra and inverse isochron space. We demonstrate that the proportion of radiogenic 40 Ar ( 40 Ar ⁎ ) present in these samples strongly influences the nature of the discordance reflected in 40 Ar/ 39 Ar data. Mass-dependent fractionation appears to have a major influence on low- 40 Ar ⁎ samples, whereas 39 Ar recoil loss/redistribution effects are evident in samples with higher 40 Ar ⁎ proportions. The impact of mass fractionation is quantified via step-heating analyses of unirradiated basalt, whereby a ∼4% difference in 38 Ar/ 36 Ar ratios is observed between low- and high-temperature heating steps. On an inverse isochron plot ( 39 Ar/ 40 Ar vs 36 Ar/ 40 Ar), isotopic disturbance for groundmass samples primarily manifests as isochron rotation, leading to a negative correlation between initial 40 Ar/ 36 Ar ([ 40 Ar/ 36 Ar] i ) values and associated 40 Ar/ 39 Ar ages. We propose a new framework for the interpretation of 40 Ar/ 39 Ar step-heating data for basaltic samples, through judicious evaluation of inverse isochron data, ( 40 Ar/ 36 Ar) i ratios and inverse isochron ages. Results from this study suggest that only samples exhibiting both flat 40 Ar/ 39 Ar age spectra and atmospheric ( 40 Ar/ 36 Ar) i ratios yield accurate eruption ages; in the case of more discordant age spectra, intermediate temperature steps with atmospheric ( 40 Ar/ 36 Ar) i ratios may provide the closest approximation of the eruption age. Graphical abstract