2000PASP..112..177F

We explore the limits of photometric reductions of crowded stellar fields observed with the Wide Field and Planetary Camera 2 on board the Hubble Space Telescope. Two photometric procedures, based on the DoPHOT and DAOPHOT/ALLFRAME programs, are tested, and the effects of crowding, complex sky background, and cosmic-ray contamination are discussed using an extensive set of artificial star simulations. As a specific application of the results presented in this paper, we assess the magnitude of photometric biases on programs aimed at finding Cepheids and determining distances. We find that while the photometry in individual images can be biased too bright by up to 0.2 mag in the most crowded fields due to confusion noise, the effects on distance measurements based on Cepheid variables are insignificant, less than 0.02 mag (1% in distance) even in the most problematic cases. This result, which is at odds with claims that have recently surfaced in the literature, is due to the strict criteria applied in the selection of the variable stars, and the photometric cross-checks made possible by the availability of multiple exposures in different filters which characterize Cepheid observations.