1 Introduction

According to the unified model of active galactic nuclei (AGN), the central massive black hole, its surrounding accreting material and the broad-line region (BLR) are all embedded within a dusty region, probably a thick molecular torus. Along some directions, the dust extinction is sufficient to block all UV, optical, and near-IR radiation originating in the inner components. A review of the arguments which led to this picture is presented in Wills (1999). The presence of a universal inflection point near 1.2$\mu$m in the spectral energy distribution of radio-quiet AGN's strongly suggests that the bulk of the near IR flux arises from dust thermal emission (e.g., Barvainis 1987; Sanders et al. 1989). The corresponding color temperature, $\simeq$1500 K, matches closely the sublimation temperature of graphite, the most resilient of the grain constituents. The near IR emission can be variable, and therefore originates, at least in part, in a compact region. Furthermore, in four AGN, the near IR variations have been shown to be delayed with respect to the UV-optical variations. The time delay corresponds closely to the light-travel time to the dust sublimation radius $r_{\rm in}$; measured values of $r_{\rm in}$ are 400 light days for Fairall 9 (Clavel et al. 1989), 50 light days for NGC 1566 (Baribaud et al. 1992), 80 light days for NGC 3783 (Glass 1992) and 32 light days for Mrk 744 (Nelson 1996).

The emerging picture is one where the near to mid-IR emission arises from thermal re-radiation of UV and optical photons absorbed by the circumnuclear dust. Various models for the geometry and location of this dust have been proposed, but the exact configuration of the models remains unconstrained due to a lack of suitable observational data.

One can use variability as a tool to probe the inner few lightyears of the dusty regions. Reverberation-mapping techniques (Blandford & McKee 1982) have been used extensively to map the BLR in several AGN, on scales of lightdays to lightmonths, notably by the International AGN Watch consortium (Alloin et al. 1994). A similar approach can be used to probe the IR-emitting region, i.e. the warm dust component within the obscuring material. Given UV flux variations of sufficient amplitudes, a mid-IR monitoring campaign of sufficiently long duration and adequate sampling rate, it may in principle be possible to recover the transfer function of the dust.

The Infrared Space Observatory (ISO; Kessler et al. 1996) offered a unique opportunity to carry out such a spectrophotometric monitoring program in the mid-IR. The Seyfert 1 galaxy Mrk 279 ( z = 0.0294) was selected because its celestial position allows an uninterrupted 12-month visibility window for ISO and it has a well-documented variability history in the optical (Osterbrock & Shuder 1982; Peterson et al. 1985; Maoz et al. 1990; Stirpe et al. 1994), the UV (Chapman et al. 1985), and X-rays (Reichert et al. 1985). Balmer-line time-delays (Maoz et al. 1990; Stirpe & de Bruyn 1991; Stirpe et al. 1994) suggest a BLR size in the range 6 to 12 lightdays. A search for day-to-day variability across the Balmer-line profile was unsuccessful (Eracleous & Halpern 1993). No far-IR flux variations were detected with IRAS (Edelson & Malkan 1987).