P31 phosphor screens are frequently used for short-term presentation of dot and grating patterns, but experimental data obtained with this technique have been criticized because of possible parasitic effects of phosphor persistence on subjects visual performance. Recently, this issue provoked a controversial discussion in Vision Research (Groner et al., 1993; Westheimer, 1993, 1994; Irwin, 1994; Di Lollo et al., 1994) which was concerned with persistence effects of P31 screens for dot patterns. Supplementing this discussion, the present work deals with the effects of different types of patterns (dot pattern vs. gratings) and background mean-luminance levels (scotopic vs. phototopic) on phosphor persistence. Physical measurements of P31 persistence occurring with grating patterns of a mean luminance of 20 cd m-2 (i.e. photopic range) were obtained by using an extremely linear photometer with high temporal resolution. Under this photopic condition, the measurements demonstrate a fast decay of residual grating contrast to 1.4% of its original value within 50 ms after pattern offset. This phosphor behavior must be considered when designing an experiment with a P31 screen though it certainly embodies no problems in many applications.