1Iowa State University, Ames, IA 50011,USA; 2Carleton College, Northfield, MN 55057, USA

Background and objectives
The duration of periods of environmental wetness are a key input to many disease-warning systems. Accuracy in measuring duration of wetness is therefore crucial to the usefulness of these systems. Electronic wetness sensors measure environmental wetness as a change in electrical resistance of a circuit board. Sensor response to dew can be affected by sensor size, shape, deployment angle, compass orientation, and presence or absence of a paint coating [1, 2]. However, no consensus has emerged on the relative importance of these factors in wetness duration measurement. The objective of this study was to quantify effects of paint coating, deployment angle, and compass orientation on performance of electronic wetness sensors during dew periods.

Materials and methods
Wetness sensors (Model 237, Campbell Scientific, Logan, UT, USA) were spray painted with two or eight coats of flat latex paint. Unpainted sensors acted as controls. In mid-Jun 1997, sensors were deployed 0.9 ;m apart and 5 ;cm above a tomato-processing canopy in Iowa. The experimental design was an RCB with two replications. Treatments in one experiment were three coatings (unpainted or two or eight coats of paint) at three deployment angles (0, 30, or 45 from horizontal). Treatments in a second experiment were compass orientations (N, S, E, and W). Twelve leaflets at the top of the canopy were tagged for visual observation of dew onset and dry-off. Leaflet observations were made at intervals of 20-30 ;min, from sunset to dew onset and from sunrise to dew dry-off, during July and August.

Results and conclusions
During dew onset, unpainted sensors were much more sensitive than painted sensors to angle of sensor deployment. For 13 dew onset events, the mean time difference between start of sensor response and first visual observation of dew was significantly (P<0.001) affected by sensor coating, deployment angle, and coating-angle interaction; however, but compass angle had no significant effect. No significant mean time differences were noted for 11 dew dry-off events. As the deployment angle of unpainted sensors increased, start of sensor response occurred increasingly later than visually observed dew onset; the time lag was 4.2 ;h for the 45 sensors. The response of painted sensors preceded visual observation of dew onset by less than 1.1 ;h. Time lags between end of sensor response and visual observation of dew dry-off were much smaller than during dew onset. The results indicated that deployment angle can significantly affect the performance of unpainted wetness sensors during dew onset, and that painting sensors greatly reduces this sensitivity.

1. Gillespie TJ, Kidd GE, 1978. Canadian Journal of Plant Science 58,179-87.
2. Huband NDS, Butler DR, 1984. Proceedings Brighton Crop Protection Conference 2, 633-38.