4 Discussion

OJ 287 is one of the most extensively studied objects. Its optical light curve covers a period starting at the end of last century, and shows obviously the outbursts with a period of 12 years, which has been explained by the double black hole model (Sillanpaa et al. 1988a). The detail detection from the OJ-94 project also show clearly the double peak structure of the outburst, tentatively explained by Lehto & Valtonen (1996) and Sillanpaa et al. (1996a). Besides, the observations from the OJ-94 project has aroused much interest in the light curve explanation (Lehto 1996; Sadun 1996; Valtonen et al. 1996; Sundelius et al. 1996; Sillanpaa et al. 1996a). During the project OJ 287 was detected for the first time by EGRET (Pian et al. 1996; Webb et al. 1996), but its radio emission was in a low state during the first optical burst (Valtaoja et al. 1996).

4.1 Spectral index

OJ 287 is violently variable in spectral indices as well as in the flux in the optical and infrared bands. The long-term analysis of this object showed that the spectra steepened after 1971 in the optical bands (Takalo & Sillanpaa 1989) and the infrared spectra steepened during the 1975 to 1990 period (Zhang & Xie 1996). Its spectral indices changed with the brightness of the source. Takalo & Sillanpaa (1989) found a strong association between B-V and V magnitude using the available optical data, but not for U-B and V. In the present paper, we did not find correlation for U-B and V either. There is no correlation for B-V and U-B, or V-I and I.

For infrared data, Gear et al. (1986a) and Kidger et al. (1994) found a correlation between J-K and J based on the limited data, this tendency has also been reported by Zhang & Xie (1996). But during the OJ-94 project, the optical spectral index kept unchanged even during the outburst period (Sillanpaa et al. 1996a), the color indices also stayed constant during the observations of Arimoto et al. (1997). It maybe that the correlation between the brightness and the spectral index does not hold during bursts. From our limited observation, there is a close correlation between B-V and V (see Fig. 3). From the available data, B-V and B-I show the indications of a decrease of spectral index with the time after the 1972 burst; color indices show that the spectrum flattened during the 1983/1984 outbursts. But it is interesting to note that there is also clear spectral flattening during the time of about JD 2448000 when there was no large outburst (see Fig. 5).

  

Figure 5: a) Evolution of color index of B-V. b) Evolution of color index of B-I

4.2 Comparison of observations

In Autumn 1994, Fioruci & Tosti (1994a,b) announced that OJ 287 was at its high level. Arimoto et al. (1997) also obtained its optical data in the V, R and I bands covering a period of October 13, 1994 through May 25, 1996. They obtained that the object was at its high level of $V=14\hbox{$.\!\!^{\rm m}$}03$ in the first half of November 1994 and fainted out to $V=16\hbox{$.\!\!^{\rm m}$}5$ within 150 days. Comparing our data with theirs, we can see clearly that our data of $V=14\hbox{$.\!\!^{\rm m}$}85 \pm 0.09$ on Dec. 6 of 1994 is in agreement with their result of $V=14\hbox{$.\!\!^{\rm m}$}84 \pm 0.03$ on Dec. 6.79 of 1994, and our data of $V=15\hbox{$.\!\!^{\rm m}$}32 \pm 0.05$ on January 27 of 1995 is also in agreement with their result of $V=15\hbox{$.\!\!^{\rm m}$}50 \pm 0.04$ on January 26.72 of 1995. For other data we can not compare with theirs, because they got one point every week. During our observation, V changed from $15\hbox{$.\!\!^{\rm m}$}62$ (Feb. 8, 1994) to $14\hbox{$.\!\!^{\rm m}$}85$ (Dec. 6, 1995), which is also in good agreement with the results obtained by Sillanpaa et al. (1996a), who showed that the range of V was from $14\hbox{$.\!\!^{\rm m}$}0$ to $16\hbox{$.\!\!^{\rm m}$}5$. But there are no observations reported in the B band in other literature for the outburst. For our observations, the B magnitudes changed from $16\hbox{$.\!\!^{\rm m}$}07$ (Feb. 7, 1994) to $15\hbox{$.\!\!^{\rm m}$}49$ (Jan. 01, 1995). From Fig. 1 it is clear that the 1994 optical outburst was fainter than the previous two (1972, 1983). But in infrared bands, the peaks for the three (1972, 1983, 1995) outbursts are comparable. Comparing the peaks in the light curve of OJ 287 in the paper of Sillanpaa et al. (1996a), we can see that the peak in the 1910s is comparable with the peaks in the 1980s, and the peak in the 1920s is comparable with the peaks in the 1990s. So, we might have missed a large peak in the 1900s which should be comparable with the peaks of the 1970s. If this is true, then there should be a slow variation over about 70 years and we would expect that the following outburst (in 2006) should be brighter than the 1990s outburst.

4.3 Infrared observations

There is no report in the infrared bands from the object during the OJ-94 project. We had the opportunity to observe it with the 1.26-m infrared telescope during the middle of November 1995. The data shown in Table 2 indicate that the infrared emission of the source was at a high level with $K=9\hbox{$.\!\!^{\rm m}$}13 \pm 0\hbox{$.\!\!^{\rm m}$}13$ on Nov. 17, 1995. It is comparable with the previous observation obtained during the optical outbursts: $K=8\hbox{$.\!\!^{\rm m}$}8$ in 1972 (Soifer & Neugebauer 1980) and $K=9\hbox{$.\!\!^{\rm m}$}25 \pm 0\hbox{$.\!\!^{\rm m}$}03$ in 1983 (Holmes et al. 1984a). From the three peaks, we can get that the intervals between the successive peaks is about 12.0 years, which is consistent with the period derived from the optical light curves (Sillanpaa et al. 1988a). From the figure in the paper of Sillanpaa et al. (1996a) it is clear that our infrared observations correspond to the second peak during the outburst of OJ 287. So, there should be a missed peak which occurred in the end of 1994!

4.4 Variability correlation

In order to investigate the emission mechanism from OJ 287, variations in the radio, optical and infrared bands have been discussed. Kinman et al. (1974) reported that there are some indications of a radio flux increase after the optical and infrared outburst, O'Dell et al. (1978) also reported that the 3 mm flux was related with that in the optical and infrared. But it is strange that no corresponding outburst was observed in radio band during OJ-94 project optical outburst. When it was at high level in optical band its radio emission was in a low state (Valtaoja et al. 1996). For the optical and infrared bands, a good correlation was obtained between the optical flux and that at 10 $\mu$m (Rieke & Kinman 1974) based on the limited data over limited period. The results shown in this paper indicate that the two bands are strongly correlated with almost no time delay. The reason for that the peak in the DCF is not so sharp is from the fact that the infrared data are fewer than the optical ones and that the clear double-peak in the optical is missed in the infrared observations.