Aerosol Time-of-Flight Mass Spectrometry (ATOFMS) instruments have been used widely to measure the size and composition of single ambient aerosol particles. ATOFMS data do not directly and quantitatively represent aerosol composition because the instruments exhibit non-linear response to particle concentration, size, and composition. Our approach is to analyze separately the components of non-linear ATOFMS response using field sampling data in order to understand ATOFMS response to ambient aerosols so that ATOFMS data can be scaled to more closely represent ambient aerosols. In this work we examine the effect of instrument busy time, mainly the time to process and save data, on ATOFMS response to ambient aerosols sampled during the 1999 Bakersfield Instrument Intercomparison Study (BIIS). During this study an ATOFMS instrument was operated alternately in normal and fast scatter data acquisition modes. In fast scatter mode, the instrument does not record mass spectra, minimizing instrument busy time; these data were used to determine particle arrival rates. Busy time in normal mode was found by a comparison of the number of particles detected to that expected for a Poisson process modified to include busy time. During the BIIS experiment, the ATOFMS instrument was busy between 5 and 95% of the nominal sampling time; thus busy time cannot be ignored for accurate quantitative analysis of ATOFMS data. ATOFMS data were scaled for on-line time and transmission efficiency, found by comparison with reference aerosol measurements, in order to estimate fine particle mass concentrations. Fine aerosol mass concentrations from scaled ATOFMS data demonstate semi-quantitative agreement with independent measurements using Beta Attenuation Monitors. We recommend that ATOFMS instruments be modified to measure busy time directly.