4 Variation in color

Since the 1991 observations were made using different comparison star, only the 1992 and 1995 observations in which the same instruments and the same comparison star were used are directly comparable. The light curves in these two years are shown in Fig. 4, where observed minima were marked with phase zero, without any displacement along the ordinate. We find that the largest variation appears around phase zero, especially for blue light.

  

Figure 4: The color curves in 1970, 1992 and 1995 and the light curves in 1992 and 1995 without any displacement along the ordinate. All the primary minima were moved to phase zero for comparison

Unlike magnitude difference, the difference in color between program and comparison stars is less affected by the instruments. The color curves of 1970, 1992 and 1995 observations are also shown in Fig. 4 (top). A common feature of them is that there are deeper or shallower dips around phases of 0 and 0.5. The depths of the dips have been changing with time. In 1995 the depth at phase zero reaches 0.2 mag. More details can be seen in a close inspection, e.g. the color of AU Ser in 1970 is a little redder at MAX II than MAX I, the reverse situation takes place in 1992, and the star in 1995 is more or less severely "reddened" except for the phase range from 0.25 to 0.4.

The variation in color directly reflects the variation in effective temperature. It is very unlikely that a component in this system changes much of its temperature as a whole in only three years. A possible explanation is that the appearance of a big starspot or relatively stable group of spots can reduce the star's effective temperature and produce such a effect, which is supported by the fact that light curves (Figs. 2 and 4) are not very smooth. In Fig. 5 the separation of unfolded light curves for V and B bandpasses vanishes slightly before (not at!) the primary minimum on May 22, 1995 (upper panel), which can be naturally attributed to the distortion of a starspot. A qualitative analysis of intensive photometry and spectrophotometry will be able to provide a priori knowledge for modelling about the property, number and location of the starspot(s).

Although it is well-known that virtually all well observed W UMa type stars do show some changes in their light curves, the theoretical explanation is not exactly known. As far AU Ser, the heights of maxima and the depths of minima have been radically changing. We have not accumulated enough observations to study period variation, to infer its undisturbed light curves and then to derive reliable orbital elements. For this reason, no modeling was attempted in this paper and much more extended observations will be continued.

  

Figure 5: The observational time series (unfolded light curves) of AU Ser in 1995, which show fluctuation, especially for the yellow on May 22 (lower curve in upper panel), with an amplitude much larger than the observational error (0.016 mag)