Etudes epidemiologiques sur la rouille de l'arachide en Cote d'Ivoire.

The study of the pathosystem: Arachis hypogaea - Puccinia arachidis was undertaken at different scales: the leaf, the plant, the field, the region, which allow to consider the monocyclic (leaf and plant) and polycyclic (field and region) processes. Several epidemiological processes were investigated as well as their succession during the epidemic build-up. At the region scale, some other important diseases of groundnut were considered, leading to a study of the multiple pathosystem: Arachis hypogaea - Puccinia arachidis - Cercosporidium personatum - Cercospora arachidicola .A preliminary analysis of the data from a survey in the various regions of Ivory Coast where groundnut is cultivated showed that rust was omnipresent. A correspondence analysis allowed to unravel the large variations of weather, cropping, and pathological situations, and to produce a general picture of the development of an average small farmer's field in time. Every year, general epidemics of groundnut rust developped in the Northern and Central regions. These were, to a large extent, ascribed to the endemicity of groundnut rust in the Southern region, which is a permanent source of inoculum.Several inoculation methods of detached leaf lets with groundnut rust were compared under laboratory conditions. The infection efficiency is highest with dry urediniospores at low inoculum level and lowest with suspended urediniospores at high inoculum level.The effects of temperature and inoculum level on the monocyclic process were investigated. Temperatures of about 27°C appeared to be optimal for rust development: infection efficiency, infectious period, and sporulation intensity were then highest, while latency period was lowest. Temperature effects on groundnut rust were compared to those on other host-rust systems, The results suggested that temperature variation has a strong effect on the development of groundnut rust epidemics.Leaf age and plant development stage were also considered as influencing factors on rust development on a susceptible groundnut cultivar. The results, a reduction of infection efficiency with increasing leaf age and an increase of latency period with increasing leaf age and development stage, indicated that different leaf layers in the canopy could differ in their epidemiological parameters.The monocyclic studies on detached leaflets were accompanied by experiments on potted plants, which yielded the same general conclusions on the effects of some factors. Differences in values of variables were found (for the infection efficiency and latency period), which were discussed.Aerial dispersal of dry spores was studied In artificial groundnut rust foci. A significant linear regression was found of the logit of the relative spore content of the air on the logit of rust intensity expressed as the number of lesions m <sup><font size="-1">-2</font></SUP>of field. A major explanatory variable of the diurnal rhythmicity of the aerial spore density was relative humidity, and a secondary one was wind velocity. The hypothesis was put forward that these variables affect rhythmicity over the whole range of variation of the aerial spore density as measured at different rust intensities.Rain-induced dispersal of groundnut rust urediniospores was studied in artificially infected plots on which variable rainfall amounts and intensities were applied, using a rainfall simulator. Several trapping methods allowed to compare dry, splash and drip dispersal. The spore liberation mechanisms and the resulting spore flows were considered at the canopy and at the individual lesion scales. The results indicated that dry dispersal was predominant. The observed increase of the flow of spores dripping from the canopy to the soil with increasing rainfall amounts suggested that light rain showers were favourable, whereas heavy showers were unfavourable to rust dispersal.Disease dispersal from artificial foci was studied with respect of variation in canopy structure with increasing plot age, The primary gradients of rust which developped in older plots had higher mean values and slopes. The relative height of infection was also higher in older plots. These age-dependant variations were attributed to variations in dispersibility and accessibility. The results also indicated a strong ability of groundnut rust to disperse at mesoscaleThe analysis of survey data was resumed, focusing on groundnut rust epidemiology. This disease exhibited marked weather (temperature and rainfall) requirements, strong dispersal abilities at mesoscale and macroscale, and intensified quickly in well-tended crops. These characteristics differed much from those of leaf-spot diseases, especially C. arachidicola leaf-spot. The conclusions of this analysis were in agreement with previous results from monocyclic experiments, The hypothesis is forwarded that intensification in groundnut cultivation would enhance rust epidemics build-up,The results of this thesis on groundnut rust epidemiology may be assembled to build a pictorial, qualitative model or a quantitative summary model for dynamic simulation. Both approaches allow to identify gaps in the available information and to evaluate the results of these studies