Astron. Astrophys. 355, 880-884 (2000)

2. Variation

2.1. Light curves

The optical data used here are from the literature: Brindle et al. (1986); Carini & Miller (1992); Courvoisier et al. (1995); Griffiths et al. (1979); Hamuy & Maza (1987); Jannuzi et al. (1993); Mead et al. (1990); Miller & McAlister (1983); Pesce et al. (1997); Smith & Sitko (1991); Treves et al. (1989); Urry et al. (1993); Xie et al. (1996) and shown in Fig. 1a-e. From the data, the largest amplitude variabilities in UBVRI bands are found: ; ; ; ; and color indexes are found: (N=140 pairs); (N=105 pairs); (N=90 pairs); (N=98 pairs), the uncertainty is 1 dispersion.

Fig. 1. a : The long-term U light curve of PKS 2155-304; b : The long-term B light curve of PKS 2155-304; c : The long-term V light curve of PKS 2155-304; d : The long-term R light curve of PKS 2155-304; e : The long-term I light curve of PKS 2155-304.

2.2. Periodicity

The photometric observations of PKS 2155-304 indicate that it is variable on time scales ranging from days to years (Miller & McAlister 1983). Is there any periodicity in the light curve? To answer this question, the Jurkevich (1971) method is used to search for the periodicity in the V light curve since there are more observations in this band.

The Jurkevich method (Jurkevich 1971, also see Fan et al. 1998a) is based on the expected mean square deviation and it is less inclined to generate spurious periodicity than the Fourier analysis. It tests a run of trial periods around which the data are folded. All data are assigned to m groups according to their phases around each trial period. The variance for each group and the sum of all groups are computed. If a trial period equals the true one, then reaches its minimum. So, a "good" period will give a much reduced variance relative to those given by other false trial periods and with almost constant values. To show the significance of the trial periodicity, we adopted the F-test (see Press et al. 1992).

When the Jerkevich method is used to V measurements, some results are obtained and shown in Fig. 2 (), which shows several minima corresponding to trial periods of less than 4.0-years and two broad minima corresponding to averaged periods of (4.16 0.2) and (7.0 0.16) years respectively.

Fig. 2. Plot of vs. trial period, P, in years.

For the periods, which are smaller than 4.0-years, we found that the decrease of the is less than 3 times the noise suggesting that it is difficult for one to take them as real signatures of periods, i.e., those periods should be discussed with more observations. For the two broad minima, the F-test is used to check their reality. The significance level is 93.8 for the 4.16-year period and 96.2 for the 7.0-year period.

© European Southern Observatory (ESO) 2000

Online publication: March 21, 2000
helpdesk.link@springer.de