2000ApJ...540..104C

We have continued our effort to re-reduce archival Q0957+561 brightness monitoring data and present results for 1629 R-band images using the methods of galaxy subtraction and seeing correction that we reported previously. The new data set comes from four observing runs, several nights apiece, with sampling of typically 5 minutes, which allows the first measurement of the structure function for variations in the R band from timescales of hours to years. Comparison of our reductions to previous reductions of the same data and to r-band photometry produced at Apache Point Observatory shows good overall agreement. Two of the data runs, separated by 417 days, allow us for the first time to use intranight variations to estimate the time delay. However, our sharpened value of 417.4 days is valid only if the time delay is close to the now-fashionable 417 day value, and worse, it is highly subject to diurnal windowing effects: our statistical estimators have little discriminatory power where there is no overlap of data, namely, at half-day offsets.
Result. herein show no unambiguous signature of daily microlensing, though a suggestive feature is found in the data. Because both time delay measurement and microlensing searches suffer from the lack of sampling at half-day offsets, inevitable at a single observatory, we discuss the need for round-the-clock monitoring with participation by multiple observatories.