2003MNRAS.344..492H

The study of quasar variability has long been seen as a way to understanding the structure of the central engine of active galactic nuclei, and as a means of verifying the morphology of the standard model. Much work has already been done on the statistical properties of light curves monitored in one colour, and it is now possible to use these observations to test predictions of theoretical models. The addition of a second colour can add enormously to the power of such tests, and put tight constraints on the nature of the variations. In this paper, a yearly monitoring programme of several hundred quasars in blue and red passbands covering 21 yr is presented. The statistics of colour changes are examined for a 15-yr period of homogeneous data with Fourier power spectrum analysis, in a form suitable for testing against theoretical predictions. The results of the Fourier analysis show that there is more power in blue light on all time-scales than in red. Examination of the light curves shows several different modes of colour change. However, if allowance is made for the effects of the underlying host galaxy, the variations become close to achromatic. There are, however, structural differences between red and blue light curves which cannot be accounted for in this way, and various modes of variability including disc instability and microlensing are examined to provide explanations for these features.