2 Observation and data reduction

The observations presented here were carried out with the 1.56 m telescope of Shanghai Observatory (SHO) and the one-meter telescope of Yunnan Observatory (YNO) from October 1993 to April 1998. The 1.56 telescope of SHO is equipped with a direct $1024 \times 1024$ pixels CCD camera at the Cassergrain focus. The one meter telescope of YNO was equipped with a direct $512 \times 512$ pixels CCD camera at the Cassergrain focus before 1996 November. Since 1996 November the telescope has been equipped with a $1024 \times 1024$ pixels CCD camera. The exposure time for B and V filters ranges from 300 to 400 s, and for R and I filters ranges from 100 to 300 s according to the brightness of the source. The bias was taken at the beginning and the end of the observation. Sky flat field frames were taken at dusk and dawn when possible. All images have been pre-reduced with the CCDRED package and then processed with the photometry tool, APPHOT, in IRAF software package. The calibration sequences in the field are taken from Smith et al. (1985). The magnitude is calculated with respect to the brightest standard star in the image frame containing the source, and the observing uncertainty is the rms error of differential magnitude between the calibration star (star 1) and the standard star that is not brighter than the source (star 2),

$$\sigma=\sqrt{\frac{\Sigma\delta_{i}^{2}}{N-1}} , \quad i=1,2,...\,N,$$ (1)

where $\delta_{i}=(m2-m1)_{i}-\overline{(m2-m1)}$, $\overline{(m2-m1)}$ is the mean differential magnitude, N is the number of repeat observations. In the case that the source was fainter than any standard star, the observing uncertainty was greater than the $\sigma$ calculated above and was given according to the typical uncertainty of stars with brightness comparable with the source. The results are given in Table 1. In Table 1 the first column is the UT date, the second the Julian date, the third the magnitude, the fourth the rms error, and the last the filter used.