1INRA-Unitd de Recherche en Bioclimatologie, 78 850 Thiverval-Grignon, France; 2INRA-Unitd de Recherche en Pathologie Végétale, 78 850 Thiverval-Grignon, France; 3IACR-Rothamsted, Harpenden, Herts. AL5 2JQ, UK

Background and objectives
The dispersal of spores is necessary to the spread of airborne fungal epidemics. Aerial dispersal is usually divided into the three processes of removal, transport and deposition. Both rain and wind may be involved in all three processes as they are not mutually exclusive [1]. Rain either washes out spores from the air or leaves, or removes them by rain-splash and rain-puff. Spores of Puccinia recondita f. sp. tritici (brown rust) and P. ;striiformis (yellow rust) are dispersed by atmospheric turbulence, especially at scales greater than 1 ;m. However, rain may be involved at a smaller scale. From the mechanical point of view, the influence of rain is unknown. The aim of this work is to quantify possible dispersal of rust spores by rain, according to physical rain parameters (drop size, fall height, rain duration). The mechanism of rain dispersal was studied in the laboratory under conditions of simulated rain using either a single drop generator or a rain simulator.

Material and methods
The drop generator produced uniform water drops by using hypodermic needles. Drops of 2.5, 3.4, 4.2 and 4.9 ;mm diameter were released from heights of 5, 50 and 100 ;cm. For a given drop diameter-fall height combination, groups of three drops, up to a total of 18, of each diameter were released on a leaf bearing sporulating lesions. Removed spores were collected on slides laid in all directions and at distances up to 12 cm from the drop impact point. Slides were coated with a solution of naphthol green B to distinguish between dry-dispersed and splash-dispersed spores. The rain simulator allowed to create rains with various drop diameters and rain duration. The same four drop diameters as above were used. Simulated rain fell from 9 m. Three rain durations of 5, 10 and 15 min were used. Two parallel rows of sporulating wheat seedlings were used as spore source. Two rows of healthy wheat seedlings, placed perpendicularly to the spore source, were used as trap plants. After each simulated rain event, trap plants were put in controlled conditions conducive to disease development and disease severity was evaluated after incubation.

Results and conclusions
With the drop generator, the maximum number of brown rust spores was removed by the first three drops. Spore removal ceased after six drops for all combinations. The same occurred with spores of yellow rust for all combinations except for drops of 2.5, 3.4 and 4.2 ;mm released from a height of 5 ;cm. In these three cases, the maximum number of spores was removed after six drops. The highest total number of spores was removed for large drop diameters and high fall heights. The conclusion hold for both dry-dispersed and splash-dispersed spores of both fungi. For each drop diameter-fall height combination, the kinetic energy of drops was calculated. The number of removed spores can be described by a power function of drop kinetic energy. The number of removed spores increased wth increasing kinetic energy.

With the rain simulator, results obtained with both fungi were in good agreement with those obtained with the drop generator. The disease severity on trap plants decreased with increasing distance from the inoculum source. Disease severity also increased with increasing diameter but decreased with increasing duration of rain. Results of both experiments showed that rain was able to remove spores of both rusts in controlled conditions. However, the extrapolation of these results to field situations requires further investigation on the relative importance of wind dispersal and water dispersal and the influence of various biotic and abiotic factors.

1. Fitt BDL, McCartney HA, Walklate PJ, 1989. Annual Review of Phytopathology 27, 241-70.