AO 0235+164 is one of the first sources for which correlated radio and optical variations have been claimed (MacLeod et al. 1976; Ledden et al. 1976; Rieke et al. 1976; Balonek & Dent 1980). The main evidence consists of two sharp optical flares, or spikes, in 1976 and 1979, coincident with radio outburst maxima (Balonek 1982). Recently, Clements et al. (1995) have performed a DCF analysis between Florida optical and Michigan radio monitoring data, finding a positive correlation with a delay of 0-2 months from optical to radio. In order to study possible correlations, we have plotted the historical optical light curve together with the combined 22 GHz and 37 GHz Metsähovi fluxes in Fig. 6 (click here). For completenes, we have included in this plot also the R-band data from Schramm et al. (1994) and from this paper. These R-band magnitudes have been converted to B-band magnitudes using a B-R value of 1.2. This B-R give consistent B-magnitudes for the flare during JD 48000. This method can produce a small error to the obtained B-magnitudes, but this will not affect the overall behaviour of the light curve.
Figure 6: The historical optical (dots) and radio (solid line) light curves. Here the dots represent the historical B-band data, the diamonds the data from Schramm et al. and the sguares data from these observations (see text for details). Both curves are plotted in logarithmic scale. Notice the similarity of the base levels in the brightness
Figure 7: Examples of the correlation between the flaring behaviour in optical and radio bands, see text for details. Here open circles are the optical, black dots the 22 GHz, black squares the 37 GHz and black diamonds the 90 GHz data, respectively
Although AO 0235+164 is very variable also in the radio, it is clear that there are no sharp spikes comparable to those in the optical. The time coverage in the radio bands is much better than in the optical bands, so such sharp radio spikes should be noticeable in the radio light curves (see also Teräsranta et al. 1992). Sharp optical flares are seen quite frequently, occurring at least once a year (and perhaps more often, considering the gaps in the data). While some of these spikes appear to be coincident with radio flares (e.g., the 1987 events, Fig. 7 (click here)a), others have no radio counterparts (e.g., the 1990-1991 optical flaring and the 1996 spike shown in Fig. 7 (click here)b). The scarcity of data especially in the optical bands does not allow a more detailed comparison. The well-documented lack of correlations, especially during our 1995-1996 monitoring, indicates either that there are at least two different mechanisms producing optical spikes, one also producing radio emission and the other not, or that the above cases of simultaneous radio and optical flares are just coincidental occurrences. Although neither hypothesis can at present be proved, we suggest that the latter is more likely.
However, a closer inspection of the historical light curves of Fig. 6 (click here) reveals a new feature, which is not very apparent unless both the optical and the radio fluxes are plotted in logarithmic scale: the overall, average behaviour is very similar in both the frequency regimes. In general, when AO 0235+164 is bright in the radio frequencies, it is also bright in the optical bands. Accordingly, periods of low radio flux seem to correspond to low levels in the optical bands (e.g., the lowest ever recorded optical and radio fluxes during 1995-1996). If we exclude the optical spikes, the "base'' emission in the optical (the lower envelope of the optical data points) corresponds quite well to the 22/37 GHz radio flux emission. This suggests that the "base'' optical emission is due to the same mechanism as the radio emission, and is therefore likely to be synchrotron emission from the relativistic jet. The optical spikes without corresponding radio flares could be caused by another mechanism, for example by microlensing or by flares in the accretion disk. Even though gravitational lensing is achromatic, the simultaneous optical and radio flares are unlikely to be caused by this mechnisim, since the shocked radio emitting regions are too large to be microlenced (e.g. Gear 1991). However optical radiation from more compact regions in the accretion disk or in other parts of the jet could be microlenced by stars in an intervening galaxy. More data are needed in order to draw a more detailed picture.