2003ApJ...588..711C

We address the following question: How does one reliably distinguish rapid microlensing events from atmospheric or instrumental effects on observed brightness fluctuations of distant objects? Our approach was to do a controlled, albeit limited, experiment, the first of its kind: observe Q0957+561A,B simultaneously through filters of the same characteristics with two telescopes with comparable fields of view but separated sufficiently for atmospheric fluctuations to be independent. Over the 1998-1999 viewing season, we succeeded in obtaining simultaneous data on 14 of the 55 nights scheduled. Analysis of these data led to the following: (1) after correction for ``blending'' of the A and B images, most rapid microlensing candidate outliers disappeared, and the agreement of the photometry from the two telescopes improved to an rms difference of about 15 mmag; (2) no microlensing events of amplitude greater than 0.1 mag were observed in either data set over the entire run, nearly ruling out ∼10^–5 M_☉MACHOs as composing more than about half the dark matter in the lens galaxy, if the source size is ≲3x10¹⁴ cm.