Résumé
To minimize application of nematicides in bananafields, crop systems have been developed in theFrench West Indies that combine fallow or rotation crop sand nematode-free in vitro plants. After two to four years, populations of the burrowing nematode Radopholus similes have developed enough to cause economic losses, leading banana growers to use nematicides. To understand how bananafields are recontaminated, we studied the dissemination of R. similis by water flow. At a 1-m scale, we analyzed the dispersion of R. similis under a rain fall simulator: we isolated a 1-m2 study plot, placed a R. similis suspension on the upstream soil surface, and simulated a 60mm/h rainfall for 72min. We collected soil samples every 10cm downstream after 12min of rainfall, and subsequently at 20-min intervals, and extracted the nematodes using a Seinhorst elutriator and then a Baermann funnel. Our results showed that the nematode dissemination follows an inverse exponential law, and depends more on soil moisture at the beginning of rainfall than on the length of rainfall: in fresh soil, 69–80% of the R. similis recovered were found less than 10cm down stream from the nematode inoculation line, whereas in wetted soil,76–85%oftherecovered individuals were collected in the outlet tub located downstream from the apparatus. This passive dissemination model partially explains the distance covered by individual nematodes but not the low percentage of motile nematodes recovered in the outlet tub (10% and 36% in fresh and wet soils) compared to the percentage of motile nematodes found in the soil (80% and 84% in fresh and wet soils). Indeed, water runoff is likely to disseminate R. similes over long distances only when soil moisture is close to field capacity