I am one of a large team studying an X-ray flux limited sample of 35 AGN, at radio (Unger et al 1987 MNRAS 228 521), IR (Ward et al 1987 ApJ 315 74 and Carleton et al 1987 ApJ 318 595), optical-UV (Boisson et al in preparation), and X-ray (Turner PhD thesis, Leicester) wavelengths. A gap in the data which we have just started to fill is the millimetre region. (Lawrence, Ward, Elvis, Robson, Smith, Duncan, and Rowan-Robinson). In Jan/Feb 1988 we made measurements of twelve objects at 800 and 1100 micron, using the ROE/QMC bolometer, UKT14, on the new UK/Dutch/Canadian facility on Mauna Kea, the James Clerk Maxwell Telescope, reaching 1 sigma sensitivity of ∼15–20 mJy, an order of magnitude improvement over previous data. The four radio loud objects measured were easily detected, as expected. These all have a strong blazar component, showing smooth but curved spectra over many decades, possibly log-Gaussian in form (Landau et al 1986 ApJ 308 78), or alternatively explicable by a small number of power-law components (Robson et al 1988 MNRAS in press). In any case, other evidence points to non-thermal radiation by a relativistically moving feature (high polarization, strong variability, superluminal motion). Eight radio quiet objects were measured, and upper limits only found, except for a possible four sigma detection of N2992. In all cases, the mm limits are far below the 100 micron IRAS fluxes. In four of the nearest objects, this is not too surprising, as fluxes are rising steeply throughout 12 to 100 micron, a sign that the IRAS data is dominated by cool interstellar dust emission (“cirrus”) from the discs of the parent galaxies. However we can also say that any postulated power law component of spectral index ∼1 dominating the near-IR, must become self-absorbed around ∼200 micron if the mm limits are not to be exceeded. Four rather more interesting objects are shown in Fig. 1. Again, any underlying power-law component must be self-absorbed by ∼100 micron, but is not clear that such a power-law is needed. N5506 and IC4329A have falling optical energy distributions, and large H α/Hβ ratios; on the other hand, the IR continuum lies well above the X-ray level, so there is good argument for absorption and re-radiation by dust. N4151, while flat through the near-IR-optical, has a large hump centred at ∼25 micron. Particularly important here are further new measurements by Engargiola et al (1987, ApJ in press),and Edelson et al (1988, preprint) which show the energy distribution to be falling so steeply from 155 to 438 micron that self-absorbed synchrotron is ruled out in this region. In fact, the whole energy distribution from mm to UV can be modelled without a power law at all, as shown in Fig 2. This uses a starburst component (from Rowan-Robinson and Crawford 1988, MNRAS in press), hot dust represented by three greybodies at 200K, 500K, and 1000K, starlight from a nuclear cusp, and a blackbody at 30,000K. Even MKN590, which at first sight looks like a power-law, can be modelled by similar components (Fig. 3).