1 Introduction

We have published multifrequency radio light curves of 12 extragalactic radio sources in the northern sky (Reich et al. 1993, Paper I), which have been detected by the Energetic Gamma Ray Experiment Telescope (EGRET, Kanbach 1988) on board the Compton Gamma Ray Observatory (CGRO). The data presented in Paper I cover the period from 1991 until the end of 1992 and represent peak flux densities measured in the cm-range and mm-range with the Effelsberg 100-m telescope and the IRAM 30-m telescope. From these data it was noted that enhanced radio emission is delayed up to several months to the $\gamma$-ray emission for the majority of the observed sources. Meanwhile more $\gamma$-ray detections of extragalactic sources have been reported and improved $\gamma$-ray flux densities have been derived (Fichtel et al. 1994; Thompson et al. 1995, 1996; Kanbach 1996; Mukherjee et al. 1997). We have continued the monitoring of extragalactic sources with the Effelsberg telescope. The relation of $\gamma$-ray activity to radio activity is of considerable interest for various reasons:

(1) First, in some $\gamma$-ray selected blazars it has been established that $\gamma$-ray flaring coincides with the launching of a new superluminal VLBI-jet component (e.g. Pohl et al. 1995; Krichbaum et al. 1995), and measuring the time delay between $\gamma$-ray flaring and radio appearance provides valuable information about the jet component formation, collimation and acceleration processes.

(2) Secondly, $\gamma$-ray loud AGN could contribute a large fraction of the diffuse extragalactic $\gamma$-ray background radiation due to the superposition of unresolved discrete sources. Most approaches (Erlykin & Wolfendale 1995; Stecker & Salamon 1996) relate the unknown $\gamma$-ray luminosity function of blazars to their radio luminosity function, and the justification of such proportionalities can only be tested by observing individual bright $\gamma$-ray loud and quiet radio AGN.

(3) Different theoretical models of the time and spectral evolution of flaring blazars make definite predictions about the onset of flares in different frequency bands, so that multifrequency broadband modelling provides an excellent test data set to discriminate quantitatively between different models.

A statistical analysis of the radio properties of EGRET sources and specifically the relation between radio and $\gamma$-ray emission has been published elsewhere (Mücke et al. 1996, 1997). It was found that those flat-spectrum radio sources which have been detected by EGRET also show more activity at cm radio wavelengths than similar sources which remain $\gamma$-ray quiet during the last years. However, there is no direct correlation between the radio and $\gamma$-ray light curves, neither in flux density nor in luminosity. Previous findings of a strict relation in luminosity can be entirely explained by the limited dynamical range of the EGRET data and selection effects inherent in the method of identification. The fact that about one third of the sources listed in Thompson et al. (1995) can be identified with radio-loud AGN with catalogue flux densities of $S(5\,{\rm GHz}) \ge 1\ {\rm Jy}$ indicates that there is a noisy luminosity relation which is further washed out by the strong variability both in $\gamma$-rays and at radio frequencies.

A time lag between $\gamma$-ray and radio outbursts was suggested for some sources on the basis of their light curves (e.g. Paper I) and the backextrapolation of the apparent motion of VLBI components. It should, however, be pointed out that there is no statistical evidence for this as a class property of all sources, which is mainly due to the limited sampling of the $\gamma$-ray light curve. Many of the promising candidates for a time lag in Paper I showed only one clear outburst in each wavelength regime (for example the BL Lac 0235+164, see Fig. 1). When adding more data from subsequent observations we get a less clear picture for some sources. An example is PKS 0528+134 which showed a $\gamma$-ray outburst in 1991 followed by a radio outburst, another bright $\gamma$-ray outburst in 1993 again followed by a radio outburst, and then nothing peculiar at $\gamma$-rays in 1995 while subsequently the brightest ever recorded radio outburst was noted (Pohl et al. 1996). Our findings indicate that if there is a time lag between $\gamma$-ray and radio outbursts then it has to be different from outburst to outburst, at least for the well-sampled sources 0528+134, 3C 273, 3C 279, and 3C 454.3 (Mücke et al. 1998).

We have continued the observations presented in Paper I with the Effelsberg 100-m telescope of EGRET detected $\gamma$-ray sources and in addition we made observations of a few sources which have been expected to show up in the $\gamma$-ray range. For most of the sources observations have been made at irregular intervals starting in 1991 until mid of 1995. For a few sources flux density monitoring was continued until February 1996 when all observations stopped due to the track replacement of the telescope. The results of the observations are presented in tabulated form and for a few sources light curves are given in addition. For some sources radio light curves or data from Table 1 have already been published. We have included the references in Table 1.

  

Table 1: List of monitored radio sources