An unsuccessful search has been made in the directions of Sagittarius B2, Sagittarius A, and Orion A for molecular lines of formyl, pyridine, formaldehyde, acrolein, and formamide in the frequency range 8·6 to 9·2 GHz. The failure to detect the formyl line in particular provides evidence about the possible modes of formation of interstellar formaldehyde.A search in the frequency range 8·6 to 9·2 GHz has failed to detect lines of formyl (HCO), pyridine (C5H5N), formaldehyde (HzCO), acrolein (CHzCHCHO), or formamide (NHzCHO) in the directions of the molecular sources Sagittarius B2, Sagittarius A (NH 3 ), and Orion A. The observations were carried out in November 1971 using the Parkes 64 m radio telescope. At this frequency the antenna has an aperture efficiency of 38 % and a beamwidth of 2' . 5 arc. The receiver consisted of a tunable cryogenic parametric amplifier (Kerr 1972) giving a system temperature of '" 180 K, followed by a bank of 64 filters each of 100 kHz bandwidth and with approximately Gaussian passbands (Batchelor et al. 1969). The filter outputs were connected via a set of integrate-and-hold circuits to a multiplexer and on-line computer. The computer controlled the receiver, switching it between the feed horn and a cryogenic termination and switching on the calibration noise source at preselected intervals. In addition it carried out the synchronous detection, normalization, and accumulation of data.The telescope was positioned on Sgr B2 and Orion A by maximizing the continuum signal. On Sgr A (NH3) the telescope was pointed at the nominal position using corrections determined for Sgr B2. Pointing was always within 0' . 3 arc of the nominal position. Observations were made by integrating for 20 min at the line frequency while following the source. A reference spectrum was then obtained on a blank region of sky chosen so that the telescope traversed the same range of zenith and azimuth angles. Such source and reference profiles were accumulated by the computer for a total of several hours, and their difference was taken.The molecules and the transitions unsuccessfully searched for are listed in Table 1, together with the sources observed for each transition and the peak-to-peak, channelto-channel noise expressed as antenna temperature. The noise was between 0·07 and 0·11 K for each observation except that for formamide in OrionA. In this case the