Results are presented from two case-studies during the 1998 Cloud Lidar And Radar Experiment (CLARE'98) in which mixed-phase clouds were observed by a multitude of ground-based and airborne instruments. In both cases supercooled liquid water was present in the form of highly reflective layers in lidar imagery, while the radar echo was dominated by the contribution from the much larger ice particles. In the first case-study, four individual liquid-water layers were observed by an airborne nadir-pointing polarimetric lidar at temperatures between −7°C and −15°C, embedded within a warm-frontal ice cloud. Their phase was confirmed by the in situ measurements and by their very low depolarization of the lidar signal. The effective droplet radius ranged from 2 to 5 µm. Simultaneous temperature and vertical-wind measurements by the aircraft demonstrated that they were generated by a gravity wave with a wavelength of around 15 km. Thin sector plates grew rapidly in the high-supersaturation conditions and were responsible for the high values of differential reflectivity measured by the ground-based radar in the vicinity of the layers. In the second case-study a liquid-water altocumulus layer was observed at −23°C, which was slowly glaciating. Profiles of liquid and ice extinction coefficient, water content and effective radius were derived from the remote measurements taken in both cases, using radar–lidar and dual-wavelength radar techniques to size the ice particles; where in situ validation was available, agreement was good. Radiative-transfer calculations were then performed on these profiles to ascertain the radiative effect of the supercooled water. It was found that, despite their low liquid-water path (generally less than 10 – 20 g m−2), these clouds caused a significant increase in the reflection of solar radiation to space, even when cirrus was present, above which the long-wave signal dominated. In the cases considered, their capacity to decrease the net absorbed radiation was at least twice as large as that of the ice. The layers were typically 100–200 m thick, suggesting that they are unlikely to be adequately represented by the resolutions of current forecast and climate models. These results suggest that a spaceborne lidar and radar would be ideally suited to characterizing the occurrence and climatological importance of mixed-phase clouds on a global scale. © Royal Meteorological Society, 2003. P. N. Francis's contribution is Crown copyright.