Astron. Astrophys. 347, 419-423 (1999)

3. Discussion

Blazars are variable at all wavelengths. Infrared observations have been available for more than 20 years, but no long-term variation is available for AGNs except for the works of Neugebeuer et al. (1979) and Litchfield et al. (1994), in which they presented infrared observations of about 8-year for some selected objects. Recently, we found that variations in the optical and infrared closely associated for PKS 0735+178 (Lin & Fan 1998) and OJ287 (Fan et al. 1998c) indicating these two bands coming from the same mechanism. But, it is reasonable that other nonvariable or slowly varying near-IR component, such as the stars in the parent galaxy, is present in the spectrum of AGNs. In this sense, when the source is bright, the spectrum is observed to steepen when the source dims, as expected from a synchrotron component which experiences radiative energy losses, but when the source dims further, because of the presence of the underlying near-IR emission, the spectrum will flatten with the source getting faint as pointed out by Brown et al. (1989). Because the underlying near-IR emission affects J more serious than other two bands, we would expect that there is a clear tendency of spectral flattening with J when J is fainter than a certain magnitude.

3C 279 is a well known OVV with a large optical variation of mag (Eachus & Liller 1975) and a highly optical polarization of (see Scarpa & Falomo 1997). It has shown a violent optical brightness increase of 2.0 mag during an interval of 24 hours (Webb et al. 1990). The largest amplitude optical variation is greater than the largest amplitude infrared variation. The straight-line fitting gives a very significant linear correlation between (H-K) and J, but Fig. 1g indicates that (H-K) decreases with J when J is fainter than 14 mag, indicating the spectrum flattens when the source dims, but this tendency does not show up obviously in Fig. 1h or Fig. 1i. The difference between the fitting and the plot is from the faint J points with large error bars, which play a less important role on the fitting. If we remove the two points with large error bar at the bottom right corner, the tendency is not very clear. Nevertheless, it is worthy of notice and being discussed with more data. If this tendency is real, it is perhaps from the affection of the underlying galaxy as discussed above. The near-IR spectrum of 3C 279 consists of at least two components. The underlying galaxy affects J more serious than H or K, so the tendency appeared in Fig. 1g does not show up clearly in Fig. 1h or 1i.

4C 29.45 also shows large amplitude variation (, Branly et al. 1996), high and variable polarizations (, =28%, Holmes et al. 1984; Mead et al. 1990). The largest amplitude variations in the infrared are smaller than, but comparable with, those in the optical band. The infrared light curves show two one-year-separating-double-peaked outbursts with an interval of 3.2 years. During the simultaneous infrared observations, 4C 29.45 showed no significant spectral changes when the source was relative bright ( when the source dimmed by 0.9 mag from K = 9.87 to 10.76 during April 5-8, 1981). But the spectral index obtained at the end of April () showed spectral steepening when the source was about 2.5 magnitude fainter than it was on April 5 (Glassgold et al. 1983). There are no continuous observations between them. From the compiled long-term data, this association is complex and some data (Smith et al. 1987) have relatively large uncertainties, , (see Fig. 2g-i). From the data, we found that the largest variation in J is smaller than that in K. The reason is that there are fewer points in J than H and K in the literature. Besides, a weak correlation of (J-K) increasing with H () can be obtained. For this object, the data are sparse, its variability properties should be discussed with more observations.

For the color indices, there are correlations between (J-K) and (H-K) and (J-H) as well, but there is almost no correlation between (J-H) and (H-K). We think the reasons are perhaps due to the facts that (1) (J-K) has wider distribution than (J-H) and (H-K) so that (J-H) and (H-K) concentrate in a narrow region diluting the correlation, and (2) the spectrum deviates from the power law (Fan et al. 1998d).

We have both compiled the infrared light curves for 3C 279 and 4C 29.45, and investigated the largest amplitude variation, color-color relation and color-magnitude relations. The largest variation in the infrared is smaller than that in the optical band, (J-K) is strongly associated with (J-H) for the two objects while a color-magnitude relation is only found for 3C 279, the (H-K)-J plot suggests that the spectrum of 3C 279 consists of at least two components. No similar results are found for 4C 29.45.

© European Southern Observatory (ESO) 1999

Online publication: June 30, 1999
helpdesk.link@springer.de