Factors Affecting the Spread of Plum pox virus Strain M in Peach Orchards Subjected to Roguing in France

ABSTRACT We evaluated the impact of roguing on the spread and persistence of the aggressive Plum pox virus strain M (PPV-M) in 19 peach orchard blocks in Southern France. During a 7- to 10-year period, orchards were visually inspected for PPV symptoms, and symptomatic trees were removed every year. Disease incidence was low in all orchards at disease discovery and was <1% in 16 of the 19 orchard blocks. The spread of Sharka disease was limited in all 19 blocks, with an annual disease incidence between 2 and 6%. However, new symptomatic trees were continuously detected, even after 7 to 10 years of uninterrupted control measures. An extended Cox model was developed to evaluate to what extent tree location, orchard characteristics, environment, and disease status within the vicinity influenced the risk of infection through time. Eleven variables with potential effect on tree survival (i.e., maintenance of a tree in a disease- free status through time) were selected from survey data and databases created using a geographical information system. Area of the orchard, density of planting, distance of a tree from the edge of the orchard block sharing a boundary with another diseased orchard, and distance to the nearest previously detected symptomatic tree had a significant effect on the risk for a tree to become infected through time. The combined results of this study suggest that new PPV-M infections within orchards subjected to roguing resulted from exogenous sources of inoculum, disease development of latent infected trees, as well as infected trees overlooked within the orchards during visual surveys. A revision of the survey and the roguing procedures used for more effective removal of potential sources of inoculum within the orchards and in the vicinity of the orchards would improve disease control suppression of PPV.     

Temporal and spatial aspects of the epidemiology of sorghum downy mildew on maize

The development of systemic disease from primary inoculum sources of sorghum downy mildew was studied on field-grown maize in Thailand. Data were recorded five times, from the first appearance of disease until 5 weeks after plant emergence. The incidence of diseased plants decreased with increasing distance from the primary inoculum sources, and the slope of the gradient flattened as the epidemic progressed. The steepest gradient of disease incidence was observed downwind. The progress in time and spread in space of disease about primary foci is described by three non-linear models which fit the data equally well. However, the resulting gradients at wider distances are different. With two models the gradients decrease asymptotically to zero with increasing distance, whilst the other model leads to negative values above a certain distance. The rates of isopath movement of all models decrease with time, but the effect of distance on the isopathic rate is different; the rate can decrease, stay constant or increase with distance.     

Effects of initial epidemic conditions, sporulation rate, and spore dispersal gradient on the spatio-temporal dynamics of plant disease epidemics

ABSTRACT A stochastic model that simulates the spread of disease over space and time was developed to study the effects of initial epidemic conditions (number of initial inocula and their spatial pattern), sporulation rate, and spore dispersal gradient on the spatio-temporal dynamics of plant disease epidemics. The spatial spread of disease was simulated using a half-Cauchy distribution with median dispersal distance mu (units of distance). The rate of temporal increase in disease incidence (beta(I), per day) was influenced jointly by mu and by the sporulation rate lambda (spores per lesion per day). The relationship between beta(I) and mu was nonlinear: the increase in beta(I) with increasing mu was greatest when mu was small (i.e., when the dispersal gradient was steep). The rate of temporal increase in disease severity of diseased plants (beta(S)) was affected mainly by lambda: beta(S) increased directly with increasing lambda. Intraclass correlation (kappa(t)), the correlation of disease status of plants within quadrats, increased initially with disease incidence, reached a peak, and then declined as disease incidence approached 1.0. This relationship was well described by a power-law model that is consistent with the binary form of the variance power law. The amplitude of the model relating kappa(t) to disease incidence was affected mainly by mu: kappa(t) decreased with increasing mu. The shape of the curve was affected mainly by initial conditions, especially the spatial pattern of the initial inocula. Generally, the relationship of spatial autocorrelation (rho(t,k)), the correlation of disease status of plants at various distances apart, to disease incidence and distance was well described by a four-parameter power-law model. rho(t,k) increased with disease incidence to a maximum and then declined at higher values of disease incidence, in agreement with a power-law relationship. The amplitude of rho(t,k) was determined mainly by initial conditions and by mu: rho(t,k) decreased with increasing mu and was lower for regular patterns of initial inocula. The shape of the rho(t,k) curve was affected mainly by initial conditions, especially the spatial pattern of the initial inocula. At any level of disease incidence, autocorrelation declined exponentially with spatial lag; the degree of this decline was determined mainly by mu: it was steeper with decreasing mu.     

The effectiveness of consistent roguing in managing banana bunchy top disease in smallholder production in Africa

The removal of infected individuals is a common practice in the management of plant disease outbreaks. It minimizes the contact between healthy individuals and inoculum sources by reducing the infectious window of contaminated individuals. This requires early detection and consistent removal at landscape scale. Roguing of mats with symptoms of banana bunchy top disease (BBTD) in Cavendish banana production systems has been tested in Australia, using trained personnel, but has never been tested in smallholder systems. We studied the effectiveness of long-term consistent roguing in prolonging the productivity of banana orchards under smallholder farming systems in highland banana and plantain dominated production systems in Africa. We assessed the possibility of low-risk seed sourcing from the managed plots. Roguing reduced BBTD incidence to 2% in managed farmer fields and to 10% in experimental field plots, while a nonmanaged field eventually collapsed in the same period. With roguing, new infections decreased monthly compared to an exponential increase in a nonmanaged field. The emergence of new infections in both managed and nonmanaged farms followed a seasonal cycle. BBTD managed plots were a source of low-risk seed for replacing the rogued mats in the same fields, but perhaps not safe for use in nonendemic areas. We conclude that it is possible for smallholder farmers to recover and maintain banana productivity with rigorous roguing, which would entail early identification of symptoms and early removal of diseased mats. Studies are needed on the intensity of roguing under different disease and production conditions.     

Effect of host and inoculum patterns on take-all disease of wheat incidence,severity and disease gradient

The importance of the spatial aspect of epidemics has been recognized from the outset of plant disease epidemiology. The objective of this study was to determine if the host spatial structure influenced the spatio-temporal development of take-all disease of wheat, depending on the inoculum spatial structure. Three sowing patterns of wheat (broadcast sowing, line sowing and sowing in hills) and three patterns of inoculum (uniform, aggregated and natural infestation) were tested in a field experiment, repeated over 2 years. Disease (severity, root disease incidence, plant disease incidence and, when applicable, line and hill incidences) was assessed seven times during the course of each season and the spatial pattern was characterized with incidence-incidence relationships. In the naturally infested plots, disease levels at all measurement scales were significantly higher in plots sown in hills, compared to plots sown in line, which were in turn significantly more diseased than plots with broadcast sowing. Disease aggregation within roots and plants was stronger in line and hill sowing than in broadcast sowing. Analysis of the disease gradient in the artificially infested plots showed that the disease intensified (local increase of disease level) more than it extensified (spatial spread of the disease), the effect of the introduced inoculum was reduced by 95% at a distance of 15 cm away from the point of infestation. Yield was not significantly affected by sowing pattern or artificial infestation.     

Modeling and analysis of disease-induced host growth in the epidemiology of take-all

ABSTRACT Epidemiological modeling, together with parameter estimation to experimental data, was used to examine the contribution of disease-induced root growth to the spread of take-all in wheat. Production of roots from plants grown in the absence of disease was compared with production of those grown in the presence of disease and the precise form of diseaseinduced growth was examined by fitting a mechanistic model to data describing change in the number of infected and susceptible roots over time from a low and a high density of inoculum. During the early phase of the epidemic, diseased plants produced more roots than their noninfected counterparts. However, as the epidemic progressed, the rate of root production for infected plants slowed so that by the end of the epidemic, and depending on inoculum density, infected plants had fewer roots than uninfected plants. The dynamical change in the numbers of infected and susceptible roots over time could only be explained by the mechanistic model when allowance was made for disease-induced root growth. Analysis of the effect of disease-induced root production on the spread of disease using the model suggests that additional roots produced early in the epidemic serve only to reduce the proportion of diseased roots. However, as the epidemic switches from primary to secondary infection, these roots perform an active role in the transmission of disease. Some consequence of disease-induced root growth for field epidemics is discussed.     

Temporal and Spatial Dynamics of Primary and Secondary Infection by Armillaria ostoyae in a Pinus pinaster Plantation

ABSTRACT Epidemiological investigations were performed in a 3-ha maritime pine (Pinus pinaster) plantation established on a site heavily infested by Armillaria ostoyae. Geostatistics were used to examine the density and the distribution of the initial inoculum. Disease dynamics were monitored for 17 years after planting. On the whole site, the cumulative mortality rate reached 35% over this period, plateauing at 12 years. Disease progress curves differed according to the density of the initial inoculum, although in all the cases, the Gompertz model described the epidemics well. The epidemiological contributions of both primary (initially colonized stumps) and secondary inoculum (newly dead pines) were evaluated by analyzing their spatial relation to annual mortality. Newly dead pines acted as secondary inoculum from year 3 and their role increased with time. When the initial inoculum density was low, the contribution of secondary inoculum to epidemic development increased faster and halted sooner than when the density of primary inoculum was high. Regardless of its density, the primary inoculum acted throughout the dynamic phase of the epidemic. When the inoculum density was low, the probability of mortality during the first 6 years of the epidemic depended on the tree distance from the nearest stumps colonized by Armillaria sp. When the inoculum density was high, the probability of mortality was higher and not related to the distance between trees and colonized stumps.     

Coupling Disease-Progress-Curve and Time-of-Infection Functions for Predicting Yield Loss of Crops

ABSTRACT A general approach was developed to predict the yield loss of crops in relation to infection by systemic diseases. The approach was based on two premises: (i) disease incidence in a population of plants over time can be described by a nonlinear disease progress model, such as the logistic or monomolecular; and (ii) yield of a plant is a function of time of infection (t) that can be represented by the (negative) exponential or similar model (zeta(t)). Yield loss of a population of plants on a proportional scale (L) can be written as the product of the proportion of the plant population newly infected during a very short time interval (X'(t)dt) and zeta(t), integrated over the time duration of the epidemic. L in the model can be expressed in relation to directly interpretable parameters: maximum per-plant yield loss (alpha, typically occurring at t = 0); the decline in per-plant loss as time of infection is delayed (gamma; units of time(-1)); and the parameters that characterize disease progress over time, namely, initial disease incidence (X(0)), rate of disease increase (r; units of time(-1)), and maximum (or asymptotic) value of disease incidence (K). Based on the model formulation, L ranges from alphaX(0) to alphaK and increases with increasing X(0), r, K, alpha, and gamma(-1). The exact effects of these parameters on L were determined with numerical solutions of the model. The model was expanded to predict L when there was spatial heterogeneity in disease incidence among sites within a field and when maximum per-plant yield loss occurred at a time other than the beginning of the epidemic (t > 0). However, the latter two situations had a major impact on L only at high values of r. The modeling approach was demonstrated by analyzing data on soybean yield loss in relation to infection by Soybean mosaic virus, a member of the genus Potyvirus. Based on model solutions, strategies to reduce or minimize yield losses from a given disease can be evaluated.     

The effects of dispersal gradient and pathogen life cycle components on epidemic velocity in computer simulations

ABSTRACT The velocity of expansion of focal epidemics was studied using an updated version of the simulation model EPIMUL, with model parameters relevant to wheat stripe rust. The modified power law, the exponential model, and Lambert's general model were fit to primary disease gradient data from an artificially initiated field epidemic of stripe rust and employed to describe dispersal in simulations. The exponential model, which fit the field data poorly (R (2) = 0.728 to 0.776), yielded an epidemic that expanded as a traveling wave (i.e., at a constant velocity), after an initial buildup period. Both the modified power law and the Lambert model fit the field data well (R(2) = 0.962 to 0.988) and resulted in dispersive epidemic waves (velocities increased over time for the entire course of the epidemic). The field epidemic also expanded as a dispersive wave. Using parameters based on the field epidemic and modified power law dispersal as a baseline, life cycle components of the pathogen (lesion growth rate, latent period, infectious period, and multiplication rate) and dispersal gradient steepness were varied within biologically reasonable ranges for this disease to test their effect on dispersive wave epidemics. All components but the infectious period had a strong influence on epidemic velocity, but none changed the general pattern of velocity increasing over time.     

Relationships Between Verticillium dahliae Inoculum Density and Wilt Incidence, Severity, and Growth of Cauliflower

ABSTRACT Microplot and field experiments were conducted to evaluate the effects of inoculum density on Verticillium wilt and cauliflower growth. Soil containing Verticillium dahliae microsclerotia was mixed with various proportions of fumigated soil to establish different inoculum densities (fumigated soil was used as the noninfested control). Seven inoculum density treatments replicated four times were established, and the treatments were arranged in a randomized complete block design. Soil was collected from each microplot immediately after soil infestation for V. dahliae assay by plating onto sodium polypectate agar (NP-10) selective medium using the Anderson sampler technique. Five-week-old cauliflower was transplanted into two beds within each 1.2- by 1.2-m microplot. At the same time, several extra plants were also transplanted at the edge of each bed for destructive sampling to examine the disease onset (vascular discoloration) after planting. Cauliflower plants were monitored for Verticillium wilt development. Stomatal resistance in two visually healthy upper and two lower, diseased leaves in each microplot was measured three times at weekly intervals after initial wilt symptoms occurred. At maturity, all plants were uprooted, washed free of soil, and wilt incidence and severity, plant height, number of leaves, and dry weights of leaves and roots were determined. The higher the inoculum density, the earlier was disease onset. A density of 4 microsclerotia per g of dry soil caused 16% wilt incidence, but about 10 microsclerotia per g of soil caused 50% wilt incidence. Both wilt incidence and severity increased with increasing inoculum density up to about 20 microsclerotia per g of soil, and additional inoculum did not result in significantly higher disease incidence and severity. A negative exponential model described the disease relationships to inoculum levels under both microplot and field conditions. Stomatal resistance of diseased leaves was significantly higher at higher inoculum densities; in healthy leaves, however, no treatment differences occurred. The height, number of leaves, and dry weights of leaves and roots of plants in the fumigated control were significantly higher than in infested treatments, but the effects of inoculum density treatments were variable between years. Timing of cauliflower infection, crop physiological processes related to hydraulic conductance, and wilt intensity (incidence and severity) were thus affected by the inoculum density. Verticillium wilt management methods used in cauliflower should reduce inoculum density to less than four micro-sclerotia per g of soil to produce crops with the fewest number of infected plants.     

Long-lasting systematic roguing for effective management of CABMV in passion flower orchards through maintenance of separated plants

In Brazil, passion flower is grown across almost the entire country. The predominant disease of the passion flower crop is passion fruit woodiness, caused by the potyvirus cowpea aphid-borne mosaic virus (CABMV), and transmitted by aphids in a nonpersistent manner. The disease reduces the useful life of the orchard from 36 months to approximately 18 months. Up to now, there has not been an efficient method for disease management. The aim of this work was to evaluate the efficiency of systematic roguing of diseased plants through weekly inspections, for disease management in the field. The latent and incubation periods of CABMV in passion flower vines were determined in order to optimize roguing efficiency. Passion fruit plants inoculated with CABMV started to act as sources of inoculum from 3 days after inoculation (DAI), and the symptoms were expressed, on average, at 8 DAI. Five field experiments, conducted in the states of São Paulo and Bahia, Brazil, demonstrated that systematic roguing of diseased plants was significantly efficient for managing passion fruit woodiness disease. In order to facilitate identification and subsequent removal of the infected plants, they need to be grown separately. This cultural practice can be recommended for managing passion fruit woodiness disease, provided it is applied on a regional scale by all passion fruit growers. The development of some pilot plantings for the application of roguing in a passion flower-producing region is recommended to validate the use of this technique for managing passion fruit woodiness disease.     

Growth of leek rust epidemics in time in three cultivars during the early stage of the epidemic

The growth of leek rust epidemics in time under favourable conditions in three leek cultivars during two years was analysed. In both years, the highest disease levels were found on cultivar Albana, followed by Carina and Cortina. A simple model is presented to correct the results for exchange of inoculum between adjacent plots. The results of this model indicate, that the difference in rust infection between the cultivars may be due to a reduced growth of the epidemic in young plants of cultivars Cortina and Carina. In older plants, the ranking in susceptibility was reversed, causing a less pronounced difference in infection between the cultivars. The growth of leek rust epidemics during the early stage of the epidemic in isolated plots was satisfactorily described by an exponential model.     

Spread and development of quambalaria shoot blight in spotted gum plantations

The aim of this study was to examine the disease development of quambalaria shoot blight, caused by the fungal pathogen Quambalaria pitereka, in plantation‐grown spotted gum (Corymbia citriodora subsp. citriodora, C. citriodora subsp. variegata, C. henryi and C. maculata) in south‐east Queensland, Australia. The results showed that native spotted gums are a primary source of inoculum followed rapidly by the production of secondary inoculum from infected trees in the plantation. The rate of spread and development of Q. pitereka within plantations increased exponentially over time as additional trees became infected and produced secondary inoculum. Spore concentration was shown to play an important role in disease development, with disease severity increasing with increasing disease incidence on individual trees and incidence across the plantation.     

Spatiotemporal characterization of citrus postbloom fruit drop in Brazil and its relationship to pathogen dispersal

Citrus postbloom fruit drop (PFD) is caused by Colletotrichum acutatum and C. gloeosporioides. These pathogens attack the flowers and cause premature fruit drop and the retention of fruit calyces. This study was designed to characterize the spatial and temporal dynamics of PFD in commercial citrus‐growing areas to better understand the disease spread. Experiments were carried out in three young orchards (500 trees each) in two municipalities in Sao Paulo State, Brazil. Symptoms of PFD on the flowers and presence of persistent calyces were assessed in each of three orchards for three years. Logistic, Gompertz and monomolecular models were fitted to the incidence data over time from the trees with symptoms. The spatial pattern of diseased trees was characterized by a dispersion index and by Taylor′s power law. An autologistic model was used for the spatiotemporal analysis. The logistic model provided the best fit to the disease incidence data, which had a fast progress rate of 0·53 per day. During the early epidemic of PFD, the spatial pattern of diseased trees was random, which suggested that inoculum spread was due to mechanisms other than rain splash. As the disease incidence increased (up to 12·6%), the spatial pattern of diseased trees became aggregated. The rapid rate of disease progress and the distribution of PFD suggest that dispersal of the pathogen is possibly related to a mechanism other than splash dispersal, which is more typical of other fruit diseases caused by Colletotrichum spp.     

Quantification of within‐season focal spread of wheat take‐all in relation to pathogen genotype and host spatial distribution

Within season gradient of wheat take‐all was measured in field experiments according to line or random sowing host spatial distribution, and two Gaeumannomyces graminis var. tritici (Ggt) isolates (G1i and G2i), representative of the G1 and G2 genotype groups in terms of aggressiveness. Root disease incidence and severity were assessed at six dates from early March to late June on plants located at regular distances from the inoculum sources. Simple models relating disease intensity at different levels of hierarchy fitted observed data well, and indicated a strong disease aggregation both within and among plants. Disease severity on source plants placed nearby the inoculum source increased over time, ranging from 5 to 46% at the first assessment, and from 55 to 98% at the last assessment, being in general larger for G2i than G1i. In line‐sown plots, disease progressed steadily along the line but did not extend beyond 20 cm, seldom reaching the neighbour line. Disease rarely reached 25 cm in the direct‐seeded crop stands. These results indicate that Ggt intensifies but does not spread to a large extent during a cropping season. Distance from the source, pathogen genotype and assessment date had a significant effect on disease severity according to mixed model analyses, disease spread being larger for G2i than G1i. However, no significant effect of host spatial distribution could be detected. Yield loss within 20 cm of the source plant ranged between 20 and 40%, and was not significantly affected by pathogen genotype or host spatial distribution.     

Effect of crop growth and canopy filtration on the dynamics of plant disease epidemics spread by aerially dispersed spores

Most mathematical models of plant disease epidemics ignore the growth and phenology of the host crop. Unfortunately, reports of disease development are often not accompanied by a simultaneous and commensurate evaluation of crop development. However, the time scale for increases in the leaf area of field crops is comparable to the time scale of epidemics. This simultaneous development of host and pathogen has many ramifications on the resulting plant disease epidemic. First, there is a simple dilution effect resulting from the introduction of new healthy leaf area with time. Often, measurements of disease levels are made pro rata (per unit of host leaf area or total root length or mass). Thus, host growth will reduce the apparent infection rate. A second, related effect, has to do with the so-called "correction factor," which accounts for inoculum falling on already infected tissue. This factor accounts for multiple infection and is given by the fraction of the host tissue that is susceptible to disease. As an epidemic develops, less and less tissue is open to infection and the initial exponential growth slows. Crop growth delays the impact of this limiting effect and, therefore, tends to increase the rate of disease progress. A third and often neglected effect arises when an increase in the density of susceptible host tissue results in a corresponding increase in the basic reproduction ratio, R(0), defined as the ratio of the total number of daughter lesions produced to the number of original mother lesions. This occurs when the transport efficiency of inoculum from infected to susceptible host is strongly dependent on the spatial density of plant tissue. Thus, crop growth may have a major impact on the development of plant disease epidemics occurring during the vegetative phase of crop growth. The effects that these crop growth-related factors have on plant disease epidemics spread by airborne spores are evaluated using mathematical models and their importance is discussed. In particular, plant disease epidemics initiated by the introduction of inoculum during this stage of development are shown to be relatively insensitive to the time at which inoculum is introduced.     

Host mixture efficacy in disease control: effects of lesion growth analyzed through computer-simulated epidemics

Lesion growth varies among foliar parasites and in order to study the effect of lesion growth on the efficacy of host mixtures to control epidemics, we altered the epidemic simulator Epimul by integrating a lesion growth function into the model. A theoretical study was performed by simulating epidemics caused by parasites with different lesion growth rates, spore dispersal gradients and multiplication rates. We found that increases in lesion growth rates resulted in large decreases in the effectiveness of mixtures for disease control and interacted strongly with parasite multiplication rate and spore dispersal gradient. The decline in mixture efficacy for epidemics with high lesion growth rates was reduced when parasite multiplication rate was higher and spore dispersal gradient steeper. Our results suggested that the lower number of infections on susceptible plants in the mixture as a result of inoculum loss on resistant hosts was partially compensated by lesion growth.     

Spatial and temporal spread of powdery mildew (Erysiphe pisi) in peas (Pisum sativum) varying in quantitative resistance

A field experiment was conducted to assess the progress in time and spread in space of powdery mildew (caused by Erysiphe pisi ) in pea ( Pisum sativum ) cultivars differing in resistance to the disease. Disease severity (proportion of leaf area infected) was measured in 19 × 23 m plots of cultivars Pania and Bolero (both susceptible) and Quantum (quantitatively resistant). Inoculum on infected plants was introduced into the centre of each plot. Leaves (nodes) were divided into three groups within the canopy (lower, middle, upper) at each assessment because of the large range in disease severity vertically within the plants. Disease severity on leaves at upper nodes was less than 4% until the final assessment 35 days after inoculation. Exponential disease progress curves were fitted to disease severity data from leaves at middle nodes. The mean disease relative growth rate was greater on Quantum than on Pania or Bolero, but it was delayed, resulting in an overall lower disease severity on Quantum. Gompertz growth curves were fitted to disease progress on leaves at lower nodes. Disease progress on Quantum was delayed compared with Pania and Bolero. The average daily rates of increase in disease severity from Gompertz curves did not differ between the cultivars on these leaves. Disease gradients in the plots from the inoculum focus to 12 m were detected at early stages of the epidemic, but the effects of background inoculum inputs and the rate of disease progress meant that these gradients decreased with time as the disease epidemic intensified. Spread was rapid, and there were no statistically significant differences between cultivar isopathic rates (Pania 2.2, Quantum 2.9 and Bolero 4.0 m d⁻¹).     

The Effect of Irrigation Practices on the Spatio-temporal Increase of Asiatic Citrus Canker in Simulated Nursery Plots in Reunion Island

Asiatic citrus canker is a potentially severe disease of several citrus species and cultivars in many tropical and subtropical areas. In such areas, infected nursery plants constitute an important source of primary inoculum for newly established citrus groves. The influence of overhead, drip, and mist irrigation systems on the development of Asiatic citrus canker was studied in simulated, Mexican-lime nurseries in Reunion Island. Overhead irrigation exacerbated the increase of disease incidence and severity caused by a streptomycin-resistant strain of Xanthomonas axonopodis pv. citri. The temporal development of Asiatic citrus canker for overhead irrigated nursery plots was best described by an exponential model, because disease incidence in these plots did not come close to an asymptote during the experimental period. This can be explained by the continuous production of new growth, susceptible to infection by Xanthomonas axonopodis pv. citri, and splash dispersal of Xanthomonas axonopodis pv. citri associated with overhead irrigation. Based on spatial correlation and spatio-temporal analyses, aggregated disease patterns were found irrespective of the irrigation system. In overhead-irrigated plots, the spread of Xanthomonas axonopodis pv. citri lacked directionnality. Rainstorms of short duration and high intensity were apparently associated with disease increase in drip-irrigated plots. There is a need to improve cultivation practices in Reunion Island citrus nurseries to minimize Asiatic citrus canker incidence in nurseries and to minimize the introduction of Xanthomonas axonopodis pv. citri to new groves.