4. Variability results
4.1. Optical observations
The relative values of , and are reported in Table (2) for each galaxy and comparison stars. Values of are smaller than in all cases and often smaller than which underlines the precision of the method employed. We consider a galaxy to be variable if firstly (i.e. ) and secondly at least one comparison star is stable. From Table (2), it appears that we have no clear variability detection for any of the objects of the sample, with some limited cases for Mkn 478, Mkn 684, Mkn 1392 and NGC 4151, studied further. Fig. 3 shows the typical light curves obtained by our algorithm for a non-variable (according to the previous criteria) galaxy, Mkn 590, and 5 comparison stars. The corresponding structure functions are also plotted in Fig. 4. They are all flat (the form of the structure function for is smarred by large statistical errors not plotted in the graph) meaning that no continuous trend are present in the data during the period of observations. The light curves of each galaxy and the associated comparison stars are plotted in Fig. 7 at the end of this paper.
4.2. Individual objects
We only presents results for the most interesting objects either because they have a limit variability detection or because they have been previously studied by other authors. We develop succinctly some tests used to confirm or not any variability detection.
4.2.1. Mkn 684 and Mkn 1383
These galaxies fulfill the two criteria of variability since and their comparison star 2 is non variable. Yet only one comparison star, in each case, has a flat structure function and the other ones increase with time lag. We suspect that a selection effect may occur in our algorithm (see Sect. 3.2). Thus, we start again the treatment, including the galaxy in the set of comparison stars. All the new structure functions appear finally stable for all time lags, quashing any variability detection.
4.2.2. Mkn 1392
This galaxy fullfills equally the criteria of variability and its structure function increases slightly during the run whereas the comparison star ones are stable (see Fig. 5). This trend remains even if we include the galaxy in the comparison star. It seems, thus, that Mkn 1392 may be variable but on a timescale larger than the length of the observations ( 4 hours).
4.2.3. NGC 4151
As previously said (see Sect. 3.2.3), important selection effect may exist in the treatment of this galaxy, since there is only one brigth star in the CCD field. To minimize these effects, we repeat the treatment but including the galaxy in the set of comparison stars. The structure function of the galaxy and its diffuse component become stable whereas the star one slightly increases during the run. We have thus to be very carefull when using differential photometry with this galaxy, since the nearest bright star seems to be variable on timescale of hours.
4.3. Infrared observations
Three galaxies of the sample, Mkn 478, Mkn 1392 and Mkn 1098, have been observed simutaneously in I and J bands. We treat the J band data with the same algorithm described in Sect. 3.2. At this wavelength we are clearly limited by the CCD and sky background noises. We can not obtain precision smaller than 2% and it is in the range 2-5% in most cases. The results are reported in Table 3. Only Mkn 478 fulfills the first variability criterium, since all comparison stars seem variable. But, once again, only two comparisons stars were used and a selection effect occurs. No more variability is detected when the galaxy is included in the set of comparison stars.
Table 3. Same as Table 2 but in J band for 3 galaxies of the sample
© European Southern Observatory (ESO) 1999
Online publication: February 23, 1999