1 Introduction

Optical flux variations have been considered as a common characteristic of several types of AGNs almost since the original discovery of these objects (e.g. Smith & Hoffleit 1963). However, the existence of optical microvariability (i.e., flux changes at the level of a few percent over timescales from minutes to less than 10 hours) in some AGNs was not widely accepted until detailed monitoring campaigns were accomplished using modern CCD detectors in the late 1980s and the 1990s (e.g. Miller et al. 1989; Carini et al. 1990, 1991, 1992; Noble et al. 1997).

The origin of these microvariations, which can reach amplitudes up to 0.2 magnitudes within a single night (see, for instance, the recent review by Miller & Noble 1996), is not clearly established at present. Two broad kinds of intrinsic models have been proposed to account for the phenomenon (we notice that in some particular objects an extrinsic explanation involving superluminal gravitational microlensing is also possible, see Gopal-Krishna & Subramanian 1991 and Romero et al. 1995a). On the one hand there are models based on modified versions of the standard shock-in-jet scenario, which is widely used to interpret radio variability of blazars (Marscher & Gear 1985). In these models the microvariability is produced when a thin and relativistic shock strikes a feature (e.g. a particle density inhomogeneity or a bend) in the parsec-scale jet of the object (Marscher 1990; Qian et al. 1991; Gopal-Krishna & Wiita 1992; Marscher et al. 1992). On the other hand, there are models for microvariability that resort to the formation of instabilities on the surface of the accretion disk that is usually assumed to exist surrounding the central supermassive black hole in the AGN (e.g. Mangalam & Wiita 1993). The instabilities can generate short-lifetime perturbations like orbiting hot spots in the inner disk and Doppler and relativistic effects can induce very fast fluctuations of the observed optical flux.

High temporal resolution observations of different types of AGNs can be used for testing the models and probing the innermost regions of the sources. In particular, microvariability observations of radio-quiet quasars (RQQSOs) can be useful to determine the contribution of the accretion disk instabilities to the overall rapid variability in radio-loud quasars (RLQSOs) and BL Lac objects. The reason is that RQQSOs are thought not to eject relativistic jets like those observed in radio-loud AGNs and, consequently, the presence of microvariability in their emission clearly points to disk activity (e.g. Gopal-Krishna et al. 1993b). Comparative studies of the incidence of microvariability in samples of both RQQSOs and radio-loud AGNs can be used to enlighten questions on the frequency of shock propagation and the microstructure of relativistic jets in the latter kind of objects.

Recent searches of intranight optical variability in RQQSOs have been carried out by Gopal-Krishna et al. (1993a,b, 1995), Sagar et al. (1996), and Rabbette et al. (1998). Studies of microvariability in samples of both RQQSOs and RLQSOs have been presented by Jang & Miller (1995, 1997). These studies, based almost entirely on samples formed by northern objects, seem to imply that the duty cycle (i.e., the fraction of time when an object displays microvariability) is very different for these two types of AGNs (see, however, de Diego et al. 1998).

In order to confirm these results and to extend the sample of observed AGNs to encompass a significant number of southern objects, we have observed 23 sources with declinations $\delta <-5\hbox{$^\circ$}$. Our sample consists of both radio-quiet and radio-loud AGNs in such a way that the obtained results, along with those already gathered by Jang & Miller (1995, 1997), provide elements for a first all-sky statistics of the microvariability phenomenon. The purpose of this paper is to present our observational data as well as to discuss some general aspects of the phenomenon on the basis of the global sample. We emphasize that we deal here with microvariability (i.e., strictly intranight variability with timescales from minutes to a few hours) and not with the so-called intraday variability (IDV, timescales from hours to several days).