In general, leaf wetness duration (LWD) is a key parameter influencing plant disease epidemiology, since it provides the free water required by pathogens to infect foliar tissue. LWD is used as an input in many disease warning systems, which help growers to decide the best time to spray their crops against diseases. Since there is no observation standard either for sensor or exposure, LWD measurement is often problematic. To assess the performance of electronic sensors, LWD measurements obtained with painted cylindrical and flat plate sensors were compared under different field conditions in Elora, Ontario, Canada, and in Piracicaba, São Paulo, Brazil. The sensors were tested in four different crop environments--mowed turfgrass, maize, soybean, and tomatoes--during the summer of 2003 and 2004 in Elora and during the winter of 2005 in Piracicaba. Flat plate sensors were deployed facing north and at 45 degrees to horizontal, and cylindrical sensors were deployed horizontally. At the turfgrass site, both sensors were installed 30 cm above the ground, while at the crop fields, the sensors were installed at the top and inside the canopy (except for maize, with a sensor only at the top). Considering the flat plate sensor as a reference (Sentelhas et al. Operational exposure of leaf wetness sensors. Agric For Meteorol 126:59-72, 2004a), the results in the more humid climate at Elora showed that the cylindrical sensor overestimated LWD by 1.1-4.2 h, depending on the crop and canopy position. The main cause of the overestimation was the accumulation of big water drops along the bottom of the cylindrical sensors, which required much more energy and, consequently, time to evaporate. The overall difference between sensors when evaporating wetness formed during the night was around 1.6 h. Cylindrical sensors also detected wetness earlier than did flat plates--around 0.6 h. Agreement between plate and cylinder sensors was much better in the drier climate at Piracicaba. These results allow us to caution that cylindrical sensors may overestimate wetness for operational LWD measurements in humid climates and that the effect of other protocols for angling or positioning this sensor should be investigated for different crops.