The Radio Wavelength Time Delay of Gravitational Lens 0957+561

The gravitational lens 0957+561 was monitored with the Very Large Array from1979 to 1997. The 6 cm light curve data from 1995-1997 and the 4 cm data from1990-1997 are reported here. At 4 cm, the intrinsic source variations occurearlier and are twice as large as the corresponding variations at 6 cm. TheVLBI core and jet components have different magnification factors, leading todifferent flux ratios for the varying and non-varying portions of the VLA lightcurves. Using both the PRHQ and Dispersion statistical techniques, wedetermined the time delay, core flux ratio, and excess non-varying B image fluxdensity. The fits were performed for the 4 cm and 6 cm light curves, bothindividually and jointly, and we used Gaussian Monte Carlo data to estimate 68%statistical confidence levels. The delay estimates from each individualwavelength were inconsistent given the formal uncertainties, suggesting thatthere are unmodeled systematic errors in the analysis. We roughly estimate thesystematic uncertainty in the joint result from the difference between the 6 cmand 4 cm results, giving 409+-30 days for the PRHQ statistic and 397+-20 daysfor the Dispersion statistic. These results are consistent with the currentoptical time delay of 417+-3 days, reconciling the long-standing differencebetween the optical and radio light curves and between different statisticalanalyses. The unmodeled systematic effects may also corrupt light curves forother lenses, and we caution that multiple events at multiple wavelengths maybe necessary to determine an accurate delay in any lens system. Now thatconsensus has been reached regarding the time delay in the 0957+561 system, themost pressing issue remaining for determining H_0 is a full understanding ofthe mass distribution in the lens.Comment: 20 pages including 4 figures, AASTeX, accepted by Ap