Knowledge of the lower oxygen limit (LOL) is critical for optimising the gaseous storage environment for fruits. The optimum storage atmosphere occurs just above the LOL at which aerobic respiration is at the lowest level which can be achieved without development of anaerobic metabolism. Measures of LOL based on a fruit's internal atmosphere, rather than external or package atmospheres, estimate the true LOL as these account more directly for differences in respiration rate and skin permeance of individual fruit. Two measures of LOL were considered: the anaerobic compensation point (ACP) and the fermentation threshold (FT). The ACP was described in terms of plots of the internal partial pressure of CO2 versus internal O-2 (p(O2)(i)) and external O-2 (p(O2)(e)) partial pressures, The FT was described in terms of plots of both the respiratory quotient (Re) and ethanol concentration versus p(O2)(i) and p(O2)(e), and occurred at higher p(O2) than the ACP. Mathematical solutions for estimating the ACP and the FT based on the Re (FTRQ) are described. A statistical 'bootstrap' procedure is described for estimating the FT based on ethanol concentration (FTEtOH) and was also suitable for estimating all other LOLs and their confidence intervals. LOLs were estimated for 'Cox's Orange Pippin' apples (Malus domestica, Borkh) at 24 degrees C using controlled atmospheres (CA). The steady-state internal partial pressures of O-2, CO2 and concentrations of acetaldehyde, ethyl acetate and ethanol were estimated non-invasively by sampling the headspace of 1000 mm(3) glass surface chambers sealed to the equatorial surface of the apples. LOLs estimated on a p(O2)(i) basis were on average 1.69 kPa, 1.94 kPa and 2.10 kPa p(O2) lower for ACP, FTRQ and FTEtOH respectively than those estimated relative to P-O2(e). The bootstrap 95% confidence limits for internal ACP (ACP(i)) were 0.70 to 0.78 kPa p(O2)(i) whilst for internal FTRQ (FTRQi) the interval was 1.04 to 1.20 kPa p(O2)(i) and for internal FTEtOH (FTEtOHi) 0.68 to 0.87 kPa p(O2)(i). Bootstrap estimates were similar, though typically higher, than mathematically fitted estimates, Determining LOLs on a steady-state internal atmosphere basis estimates the true LOL more accurately than those estimated from external or package atmospheres, and provides a more mechanistic basis for models used to predict fruit responses to CA. As the FTRQ represents the critical point at which fermentation occurs it was considered the safest estimate of the true LOL for optimising storage atmospheres.
