drive the ultraviolet/optical variations. However, the medium energy X-ray NVA is 2-4 times that in the ultraviolet, and the single-epoch, absorption-corrected X-ray/γ-ray luminosity is only about 1/3 that of the ultraviolet/optical/infrared, suggesting that at most ∼1/3 of the total low-energy flux could be reprocessed high-energy emission.The strong wavelength dependence of the ultraviolet NVAs is consistent with an origin in an accretion disk, with the variable emission coming from the hotter inner regions and non-variable emission from the cooler outer regions. These data, when combined with the results of disk fits, indicate a boundary between these regions near a radius of order R ≈ 0.07 lt-day. No interband lag would be expected as reprocessing (and thus propagation between regions) need not occur, and the orbital time scale of ∼1 day is consistent with the observed variability time scale. However, such a model does not immediately explain the good correlation between ultraviolet and X-ray variations.

The results of an optical monitoring campaign on the active nucleus in the luminous Seyfert 1 galaxy Fairall 9 are presented. This campaign was undertaken in parallel with ultraviolet spectroscopic monitoring with the IUE satellite which is described in a separate paper. The primary purpose of this program is to measure the response times (or "" lags ÏÏ) of the emission lines to continuum variations and thus to extend the range in luminosity of active galactic nuclei (AGNs) for which such measurements have been made. The main conclusions of this work are as follows :1. Continuum (at 5340 variations of amplitude D12% are detected on timescales as short as D20 A ) days. These variations are much larger than the typical uncertainties in the measurements, which are of order D2%. Over D94 days, a factor of 2 change in the nuclear continuum was observed.2. The optical continuum light curve resembles that of the UV continuum, showing two "" events ÏÏ of low-amplitude variations with a duration of D70 days and with no measurable lag between the UV and optical continuum light curves. The UV data show a third larger amplitude event that occurred after the optical monitoring had terminated and unfortunately went unobserved in the optical.3. The Hb emission-line Ñux also underwent signiÐcant, low-amplitude (º20%) variations. Crosscorrelation analysis reveals that Hb lags behind the UV continuum by about 23 days, a value much smaller than what was previously suggested by earlier variability studies. However, this small lag is consistent with the lags for the UV lines during this campaign in the sense that the Hb lag is approximately 50% larger than that of Lya j1216, as it has been found for lower luminosity AGNs.4. The Hb di †erence proÐle produced by subtracting the low-state from the high-state data can be described as a two-component structure with blue and red components of similar width (D2500 km s~1) and that appear to vary in phase.

Resultsof a ground-based optical monitoring campaign on 3C 390.3 in 1994-95 are presented. The broad-band fluxes (B, V , R, and I), the spectrophotometric optical continuum flux F λ (5177Å), and the integrated emission-line fluxes of Hα, Hβ, Hγ, He i λ5876, and He ii λ4686 all show a nearly monotonic increase with episodes of milder short-term variations superposed. The amplitude of the continuum variations increases with decreasing wavelength (4400 -9000Å). The optical continuum variations follow the variations in the ultraviolet and X-ray with time delays, measured from the centroids of the cross-correlation functions, typically around 5 days, but with uncertainties also typically around 5 days; zero time delay between the high-energy and low-energy continuum variations cannot be ruled out. The strong optical emission lines Hα , Hβ , Hγ , and He i λ5876 respond to the high-energy continuum variations with time delays typically about 20 days, with uncertainties of about 8 days. Thereis some evidence that He ii λ4686 responds somewhat more rapidly, with a time delay of around 10 days, but again, the uncertainties are quite large (∼ 8 days).The mean and rms spectra of the Hα and Hβ line profiles provide indications for the existence of at least three distinct components located at ±4000 and 0 km s −1 relative to the line peak. The emission-line profile variations are largest near line center.