Systematic errors in one-dimensional light-curve convolution for extended sources

One-dimensional contour-following methods have proved an effective means of studying the statistics of microlensing light curves for point sources. Fo r extended sources, however, convolving a point source light curve with a o ne-dimensional source profile results in systematic deviations from the tru e light curve. This paper demonstrates that these effects are generic (rega rdless of source size), and attempts to quantify the effects so that proper caution may be taken when interpreting previous results.