The optimum atmosphere for a crop, with respect to oxygen, lies just above the lower oxygen limit (LOL), at which maximum benefits in relation to fruit quality are achieved and below which fruit quality is compromised by fermentation. In contrast to previous work, LOLs in this study were estimated on the basis of steady-state internal atmospheres (LOL i s) as well as external atmospheres (LOL e s) as it is the internal O2. partial pressure (P i O2, Pa), close to equilibrium with the cytosol, that mediates important physiological processes. The study tested whether LOL i s of 'Cox's Orange Pippin' and 'Braeburn' apples were affected by temperature, elevated CO2, and physiological age. Two types of LOLs were identified: the anaerobic compensation point (ACP) and the fermentation threshold (FT). ACP was described in terms of plots of the internal CO2 (p i CO2) versus internal (p i O2) and external (P e O2) O2, and FT in terms of plots of both a measure of the respiratory quotient (RQia) and ethanol (EtOH) concentration versus P i O2 and P e O2. Mathematical solutions for estimating ACP and FT based on the RQ ia (FT RQ), and a statistical 'bootstrap' procedure suitable for estimating all LOLs and their bias-corrected 95% confidence intervals, are described. LOL i s of postclimacteric fruit of both cultivars tended to increase slightly between 0° and 28°C and sharply at 32°C. LOL i s ranged between 0.5 KPa and 2.2 KPa p i O2; values for FT i RQ and FT i EtOH tended to be higher than for ACP i . Elevated p e CO2 (0 to 8 kPa at 0° and 20°C) did not significantly affect LOL i s at 20°C, but increases in FT i RQ and FT i EtOH occurred for fruit at 0°C. A small decrease in O2 uptake and RQia was measured for fruit in 2 to 8 KPa P e CO2 at 20°C. No consistent changes in LOL i s were observed for either cultivar in relation to physiological age (preclimacteric, climacteric, or postclimatceric fruit at 0° or 20°C). In contrast to ACP i, ACP e increased markedly with temperature, resulting from its dependence on both skin permeance and respiration rate (both of which change with time fruit are in storage). Consequently, use of LOL i s, rather than LOL e s is recommended for optimising atmospheres for both sealed packages and controlled atmosphere storage, to minimise risk of fermentation.