2 Observations and data reduction

We present observations performed from January 1994 to July 1999 with the Effelsberg 100 m radio telescope, at wavelengths $\lambda$ = 21, 18, 11, 6, 3.6, 2.8, 2, 1.3cm, 9 and 7mm. One should note, that both time and frequency coverage are rather irregular, depending on the availability of time slots and receivers. Most observations have been made at wavelengths of 11, 6, 2.8, and 1.3cm.

All sources monitored are listed in Table 1 with their optical identifications, galactic latitudes, and redshifts (taken from Kühr et al. 1981 and Stickel et al. 1994). Further, we give the estimated spectral indices $\alpha^{11}_6$ and $\alpha^6_{2.8}$, and - if available from the literature - the superluminal velocity for the fastest component (see individual references). We also note whether a source showed Intraday variability (IDV, e.g. Wagner & Witzel 1995; Kraus 1997) at least once or not.

Since all target sources are point-like and moderately strong ( $\mathrel{\mathpalette\SchlangeUnter>}0.2$Jy) at the frequencies of observation, we were able to perform the measurements with cross-scans (in Azimuth and Elevation). As a first step of the analysis, a Gaussian was fitted to every cross-scan. The average amplitude of these Gaussians (which are the result of the convolution of the point-like source with the antenna-beam) is a measure of the source flux density. Subsequently, we corrected for the elevation-dependence of the gain of the telescope and systematic time-dependent effects. To determine those, we frequently observed steep-spectrum sources (and the primary calibrators) which were known to show no variability on short timescales (cf. Quirrenbach et al. 1992; Kraus 1997; Kraus et al., in preparation).

Finally, we linked our observations to an absolute flux density scale (Baars et al. 1977) by observing primary calibrators such as 3C 286 and 3C 48. The measurement errors are composed of the statistical errors from averaging the individual samples in a scan, and a contribution from the calibration errors, which are reflected by apparent residual fluctuations of the non-variable sources.