3. The observational material
For the present study only archival light curves of the four systems HT Cas, V2051 Oph, IP Peg and UX UMa observed at the SAAO, South Africa, at the MacDonald Observatory, USA, and at the Wise Observatory, Israel, were used. Most of the data of HT Cas and V2051 Oph were already studied with respect to other properties than the flickering. With the exception of two so far unpublished light curves the data of HT Cas were published by Patterson (1981) and Zhang et al. (1986). Of V2051 Oph, 6 light curves are unpublished; the others can be found in Warner & Cropper (1983) and Warner & O'Donoghue (1987). The quoted references also contain details about the observations. In contrast, none of the light curves of IP Peg and UX UMa used here was published to my knowledge.
The large majority of the data refers to white light, but a few light curves were observed with the four channel Stiening photometer (Horne & Stiening 1985). In these cases only the B band was employed. The brightness of the stars is given as counts per integration time, corrected for extinction and sky background. The latter features imply that only an approximate correction of the scatter curves for Poisson noise is possible. However, this does not affect the results seriously (see Bruch 1996 for a discussion of this point). For the three dwarf novae in the present sample (HT Cas, V2051 Oph and IP Peg) only light curves obtained during quiescence are used.
The amount of available data differs very much for the four stars investigated here. Whereas data of 66 orbital cycles were used in the case of V2051 Oph, the corresponding numbers for HT Cas, UX UMa and IP Peg are 22, 9 and 3. Since the dependence of the scatter in the light curves is to be studied here as a function of orbital phase all light curves were phase folded (after splitting those light curves containing more than one cycle into separate curves). Table 1 contains an overview: The individual columns list the cycle number, the heliocentric Julian Date of the corresponding eclipse, the civil date, and the time resolution of the light curves in seconds. The phase coverage of the individual cycles is not always complete. Whereas for HT Cas and V2051 Oph in most cycles the complete phase range around the eclipses is covered, in the longer period system UX UMa in general only a restricted phase range centered on the eclipse was observed.
Table 1. Eclipse numbers, Julian dates, civil dates and time resolution for the light curves used in this study
The cycle numbers and the corresponding eclipse numbers quoted in Table 1 are based on the ephemeris of Horne et al. (1991) in the case of HT Cas, Echevarria & Alvarez (1993) for V2051 Oph and Baptista et al. (1995) for UX UMa. The orbital period of IP Peg is not stable as was first pointed out by Wood et al. (1989b). Wolf et al. (1993) later found the period variations to be cyclic. Here, the ephemeris of Wolf et al. (1993) are adopted. However, in two cases none of the published ephemeris predicts the epochs of the minima observed in the present light curves even approximately. There may be a timing error. Therefore, no cycle numbers and eclipse epochs are given for these cases in Table 1.
Not in all cases the eclipse epochs predicted from the ephemeris matched well the observed eclipses. Instead of trying to trace the reasons for the inconsistencies (difficult in view of the limited information available about the observational details of the archival data) a pragmatic approach was adopted: In such cases, the epoch of the eclipse centre was visually measured in the light curve (sufficient to ensure an accuracy which is higher than the phase resolution of the final scatter curves), and the phase-folding was done using that epoch as zero-point.
© European Southern Observatory (ESO) 2000
Online publication: July 13, 2000