In low-temperature flash photolysis of NH3/O-2/N-2 mixtures, the NH2 consumption rate and the product distribution is controlled by the reactions NH2 + HO2 -> products (R1), NH2 + H (+M) -> NH3 (+M) (R2), and NH2 + NH2 (+M) -> N2H4(+M) (R3). In the present work, published flash photolysis experiments by, among others, Cheskis and co-workers, are re-interpreted using recent direct measurements of NH2 + H (+N-2) and NH2 + NH2 (+N-2) from Altinay and Macdonald. To facilitate analysis of the FP data, relative third-body collision efficiencies compared to N-2 for R2 and R3 were calculated for O-2 and NH3 as well as for other selected molecules. Results were in good agreement with the limited experimental data. Based on reported NH2 decay rates in flash photolysis of NH3/O-2/N-2 , a rate constant for NH2 + HO2 -> NH3 + O-2 (R1a) of k(1a) = 1.5(+/- 0.5) x 10(14) cm(3) mol(-1) s(-1) at 295 K was derived. This value is higher than earlier determinations based on the FP results but in good agreement with recent theoretical work. Kinetic modeling of reported N2O yields indicates that NH2 + HO2 -> H2NO + O (R1c) is competing with Rla, but perturbation experiments with addition of CH4 indicate that it is not a dominating channel. Measured HNO profiles indicate that this component is formed directly by NH2 + HO2 -> HNO + H2O (R1b), but theoretical work indicates that R1b is only a minor channel. Based on this analysis, we estimate k(1c) = 2.5 x 10(13) cm(3) mol(-1) s(-1) and k(1b) = 2.5 x 10(12) cm (3 )mol(-1) s(-1) at 295 K, with significant uncertainty margins.