Temporal sampling helps unravel the genetic structure of naturally occurring populations of a phytoparasitic nematode. 2. Separating the relative effects of gene flow and genetic drift

Abstract Studying wild pathogen populations in natural ecosystems offers the opportunity to better understand the evolutionary dynamics of biotic diseases in crops and to enhance pest control strategies. We used simulations and genetic markers to investigate the spatial and temporal population genetic structure of wild populations of the beet cyst nematode Heterodera schachtii on a wild host plant species, the sea beet (Beta vulgaris spp. maritima), the wild ancestor of cultivated beets. Our analysis of the variation of eight microsatellite loci across four study sites showed that (i) wild H. schachtii populations displayed fine‐scaled genetic structure with no evidence of substantial levels of gene flow beyond the scale of the host plant, and comparisons with simulations indicated that (ii) genetic drift substantially affected the residual signals of isolation‐by‐distance processes, leading to departures from migration–drift equilibrium. In contrast to what can be suspected for (crop) field populations, this showed that wild cyst nematodes have very low dispersal capabilities and are strongly disconnected from each other. Our results provide some key elements for designing pest control strategies, such as decreasing passive dispersal events to limit the spread of virulence among field nematode populations.


O R I G I N A L A R T I C L E
Temporal sampling helps unravel the genetic structure of naturally occurring populations of a phytoparasitic nematode. 2. Separating the relative effects of gene flow and genetic drift Cécile Gracianne 1,2 | Pierre-Loup Jan 1,3 | Sylvain Fournet 1 | Eric Olivier 1 | Jean-François Arnaud 4 | Catherine Porte 1 | Sylvie Bardou-Valette 1 | Marie-Christine Denis 1 | Eric J. Petit 3

| INTRODUCTION
Agrosystems are highly homogenous artificial environments that are particularly amenable to the emergence and development of pathogens (Stukenbrock & McDonald, 2008). In such artificial, disturbed habitats, gene flow is a crucial parameter that determines the adaptive value of pathogen populations and the risk they represent for crops, especially through the evolution of virulence and resistance breakdown (McDonald & Linde, 2002). This is particularly true for pathogens with high dispersal abilities that have alternative wild hosts outside the cropping area even in the absence of crops (Burdon & Thrall, 2008). There is increasing evidence that crop pathogens can develop on wild host species, related or unrelated to the usual cultivated host (e.g. Lebeda, Petrželová, & Maryška, 2008;Monteil et al., 2013;Rouxel et al., 2014). Therefore, wild populations of pathogens may act as reservoirs of genetic diversity and initiate local crop epidemics (Burdon & Thrall, 2008;Leroy, Le Cam, & Lemaire, 2014). In this respect, a few studies on wild plant pathogen populations have investigated patterns of gene flow between wild and cultivated hosts (see examples in Stukenbrock & McDonald, 2008). However, there are no published reports on the patterns of gene flow among wild plant pathogen populations, which are nonetheless an important determinant of plant pathogen population structure and thus of its potential role as a virulence reservoir.
Studying wild plant pathogen populations offers the opportunity to better understand the evolutionary dynamics of crop pathogen populations and can provide clues on the influence of human activities on the genetic structure of pathogen populations in agrosystems (Lebarbenchon, Brown, Poulin, Gauthier-Clerc, & Thomas, 2008;Morgan, Clare, Jefferies, & Stevens, 2012). Although advocated, this approach has received very little attention thus far, with only one study on readily identifiable zoopathogens (see Morrison & Hoglund, 2005).
In the case of soilborne plant diseases, there is no information about wild populations of pathogens, probably because they are difficult to diagnose. Among soilborne pathogens, plant-parasitic nematodes are major crop pests of agrosystems that can cause severe economic losses annually (Jones et al., 2013). However, there are few detailed investigations of their population genetic structure, and even scarcer information on their spatial genetic structure in the wild (Gilabert & Wasmuth, 2013;van der Putten et al., 2006). All available data come from two studies on the pinewood nematode, Bursaphelenchus xylophilus, and four crop field surveys on the dagger nematode Xiphinema index and three cyst nematodes Globodera pallida, Globodera tabacum, and Heterodera schachtii (Alenda, Montarry, & Grenier, 2014;Mallez et al., 2013Mallez et al., , 2015Picard & Plantard, 2006;Plantard & Porte, 2004;Villate, Esmenjaud, Van Helden, Stoeckel, & Plantard, 2010). The latter three plant-parasitic nematodes cause severe damage to vineyards and potato fields in Europe, tobacco fields and sugar beet fields worldwide, respectively. Bursaphelenchus xylophilus damages pine forests around the world and differs from cyst nematode species in its life history, requiring an intermediary insect species to disperse from one pine tree to another. The population genetic structure of this nematode indicates that B. xylophilus populations can be differentiated among individual trees at a very small spatial scale, according to the dispersal of the insect vector (Mallez et al., 2013). In contrast, cyst nematodes are soilborne endoparasitic nematodes with direct life history cycles, in which only two free-living developmental stages (males and juveniles) can disperse actively in the soil. Due to their small size (<1 mm), nematodes can travel over short distances only (Norton & Niblack, 1991;Wallace, 1968).
Interestingly, populations of G. pallida and H. schachtii exhibit low genetic differentiation among cultivated fields located 50 and 150 km apart, respectively, suggesting the occurrence of gene flow over large spatial scales. This pattern was attributed to large nematode population sizes in cultivated fields and passive dispersal of cysts in agricultural areas via natural factors and human activities (Alenda et al., 2014;Picard & Plantard, 2006;Villate et al., 2010). However, there have been no investigations on wild populations of crop plant-parasitic nematodes to date, and the question remains as to whether plant-parasitic nematodes also passively disperse over large distances in wild ecosystems.
In this study, we examined the genetic structure of wild populations of a cyst nematode, H. schachtii, using both empirical and simulated datasets. Sugar beet (Beta vulgaris spp. vulgaris) is the usual cultivated host of H. schachtii which can also develop on wild relative of the sugar beet, the sea beet B. vulgaris spp. maritima (Subbotin, Mundo-Ocampo, & Baldwin, 2010). Sea beets are common along the European Atlantic and Mediterranean coastlines (De Cauwer, Dufay, Cuguen, & Arnaud, 2010;Hautekèete, Piquot, & Van Dijk, 2002). Wild populations of H. schachtii have been found on wild sea beet populations located along the Atlantic coastline from Spain to Denmark (Gracianne et al., 2014). This wide geographical occurrence provides the opportunity to explore the patterns of gene flow over different spatial scales in the wild. Specifically, (i) we investigated the levels of genetic differentiation in local wild populations of H. schachtii to examine what hierarchical scale should be considered to define a population in the wild, (ii) we examined the spatial genetic structure of nematode populations to identify signatures of dispersal events and the scale over which they occur, and (iii) we separated the respective effects of gene flow and genetic drift in observed population genetic structure.  Jan et al., 2016) to mimic the geographical sampling scheme that was performed in a previous investigation on H. schachtii conducted in sugar beet fields (Plantard & Porte, 2004). Populations of the host plant, the sea beet, are often composed of individuals clustered in geographically and genetically distinct patches (De Cauwer et al., 2010). Thus, a maximum of 10 sea beet plants were sampled in three to five arbitrarily defined patches of 3 m in diameter randomly distributed on each beach (Fig. S1). Those patches do not have any biological meanings related to the nematode biology and were only used for delineating sampling areas. In fall 2012, 120 plants were sampled and marked with plastic tags. Due to their small size (<1 mm), active dispersal of H. schachtii individuals occurs over very short spatial distances (Plantard & Porte, 2004;Wallace, 1968;Westphal, 2013).

| Biological material and sampling design
Thus, the soil sampled around the roots of one plant was considered as being one nematode population for subsequent analyses.

| Molecular characterization and genotyping
As a soil sample can contain cysts from different Heterodera species, we used restriction profiles of the ITS sequence for species identification. DNA extraction, PCR amplification of ITS sequence and digestion of PCR products were performed as described in Amiri, Subbotin, and Moens (2002). Polymorphic mitochondrial markers are not yet described in H. schachtii, but we used nuclear microsatellite loci that are markers of choice for studying neutral genetic structure. In all, 1,754 H. schachtii individuals were identified and successfully genotyped at eight microsatellite loci, named Hs33, Hs36, Hs55, Hs56, Hs68, Hs84, Hs111, and Hs114 and described in Montarry et al. (2015). Microsatellites PCR products were analyzed on an ABI Prism ® 3130xl sequencer (Applied Biosystems, Foster City, CA, USA). Allele sizes were identified using the automatic calling and binning procedure of GeneMapper v4.1 (Applied Biosystems) and completed by a manual examination of irregular results. Samples with dubious genotypes were reamplified.

| Dataset and population partitioning
In our sampling scheme, populations of H. schachtii were a priori defined at the scale of a single host plant because active dispersal of H. schachtii is considered to be spatially restricted and because it is very difficult to perform fine-scaled spatial sampling over the root system of the host plant. However, we assessed the scale of population boundaries using three different analyses. We first used a spatial principal component analysis (sPCA) to investigate the spatial distribution of genetic diversity within each surveyed beach using the R package adegenet (Jombart, 2008). This method is not based on any population genetic hypothesis and summarizes the genetic variation into synthetic variables maximizing the product of the variance taking into account the spatial autocorrelation between sampling locations using Moran's I, which was calculated using a network based on Delaunay triangulation. Synthetic components can be positive or negative reflecting, respectively, global or local structure (i.e. positive or negative spatial autocorrelation among genetic units). The two-first principal component scores were simultaneously represented into a channel of color to draw a comprehensive synthetic representation of sPCA scores, as described in Menozzi, Piazza, and Cavalli-Sforza (1978). We also used pairwise kinship coefficients F ij among nematode individuals to identify the scale over which a spatial genetic structure may appear among host plants (Loiselle, Sork, Nason, & Graham, 1995). This analysis describes kinship variation among individuals over spatial distances which can indicate the scale over which a significant spatial genetic structure occurs. Standard errors of F ij were estimated using a jackknifing procedure among loci implemented in the software SPAGeDi version 1.5 (Hardy & Vekemans, 2002). We defined eight to eleven geographical distances classes for each dataset considering an even distribution of pairwise individuals and computed 95% confidence intervals using 10,000 permutations of individual locations to test whether average kinship coefficients significantly departed from zero. The spatial scale of positive autocorrelation, defining genetic neighborhood in the broad sense, was considered as the distance value for which F ij coefficients dropped under zero (see Favre-Bac, Mony, Ernoult, Burel, & Arnaud, 2016;Sokal & Wartenberg, 1983).
Finally, we performed a nonspatially explicit Bayesian genetic clustering on the 34 populations sampled in 2012 and in 2013 to detect potential substructuring both at the beach and at the host plant scale, using STRUCTURE (Pritchard, Stephens, & Donnelly, 2000). Each K value, ranging from 1 to 25, was tested with 30 replicated runs consisting in a burn-in period of 100,000 iterations followed by 2.  (Table 1). Both kinds of analyses led to the same conclusions. In the following, we will present results based on the restricted 2013 dataset (n = 20) and will include results from the unrestricted dataset (n = 34) when they provide useful additional information.

| Characterization of basic genetic parameters
Hardy-Weinberg (HW) equilibrium is required in population assignment-based methods (see below). Therefore, single and multilocus departures from HW equilibrium were tested by estimating F IS values for all populations on each beach. Statistical significance of F IS values was assessed using 10,000 permutations of alleles among individuals, adjusted for multiple tests with Bonferroni corrections, as implemented in the software FSTAT version 2.9.3 (Goudet, 1995).
Similarly, linkage disequilibrium among loci was also assessed using permutation tests adjusted with Bonferroni corrections implemented in FSTAT. Genetic diversity of nematode populations was evaluated through the estimation of expected heterozygosity (H e ) and allelic richness (A r ) using FSTAT. Allelic richness was estimated using the rarefaction method as described in El Mousadik and Petit (1996).

| Levels of genetic differentiation
Three hierarchical levels of population structure were considered to explain the partitioning of genetic differentiation: the level of the host plant, considered as hosting one nematode population, as described above; the level of the patch, defined as a geographical clustering of host plants within a beach; and the level of the beach, comprising all the set of host plants surveyed for nematode population sampling.
We used the R package hierfstat (Goudet, 2005) to estimate variance components of each hierarchical level and to test their statistical significance, as described in De Meeûs and Goudet (2007). To test for isolation by distance (IBD), that is a gradual increase of genetic differentiation with increasing geographical distance among populations, we estimated pairwise population F ST and performed regression analyses as described in Rousset (1997) using the R package adegenet (Jombart, 2008). Significance of the relationship between the two variables was tested using classical Mantel tests (Smouse, Long, & Sokal, 1986). The same analyses were performed on both (2012 and 2013) datasets.

| Migrant detection: empirical data
At migration-drift equilibrium, a pattern of IBD depicted through variation in genetic differentiation among populations reflects spatially restricted gene flow. However, such population structure can be a residual signal of ancestral gene flow that no longer occurs or can be the result of diverse nonequilibrium situations such as colonization (Barker, 2013), secondary contact between allopatric populations (Petrou et al., 2013), or population expansion (Awad, Fady, Khater, Roig, & Cheddadi, 2014). Thus, to help us distinguish equilibrium from nonequilibrium situations in IBD patterns, we used assignment tests to detect real-time (i.e. first-generation) migrants, which are assumed to T A B L E 1 Genetic diversity and summary statistics for nematode populations sampled in 2012 and 2013 reflect current gene flow (Broquet & Petit, 2009;Manel, Gaggiotti, & Waples, 2005). The detection of first-generation migrants and population assignment of individuals were conducted using Bayesian criteria based on the computation of the probability of observing a given genotype in each population (see Rannala & Mountain, 1997 for details).
This approach is generally used when all population sources of the detected first-generation migrants are not known. Using GeneClass 2 (Piry et al., 2004), this approach was applied to define the statistical criteria that estimate the likelihood that an individual originates from a given population. The detection of first-generation migrants was performed by computing the likelihood of the individual genotype within the population where the individual was sampled. For each individual, the probabilities of belonging to each sampled population were estimated by simulating 10,000 multilocus genotypes using a Monte Carlo resampling procedure, as described in Paetkau, Slade, Burden, and Estoup (2004). Individuals with a probability lower than .01 of occurring in the sampled population were considered as potential migrants. This threshold corresponds to the minimal tolerable type I error expected from assignment tests using this procedure . Assignments of individuals to a population were based on computation of individual exclusion probabilities. In this case, the population exhibiting the highest membership probability was considered as the population of origin of the individual. Individuals assigned to another population from which they were sampled were considered as migrants. Migrant detection was thus probabilistic and did not aim at identifying the origin of migrants, because it is not possible to sample all populations in our system Piry et al., 2004 Wang (2014) recently showed that assignment-based methods can greatly overestimate migration rates in IBD contexts. We thus decided to evaluate the rate of false-positive migrants that assignment tests may generate in our analyses by simulating datasets that mirror our empirical population genetic data. To do so, we simulated a grid of 9 × 9 (81) populations with balanced sex ratio using the software EasyPop version 2.0.1 (Balloux, 2001). This grid is a very simplified representation of the spatial structure of nematode populations, but we chose to use it because (

| Population definition
Spatial principal component analysis analyses performed over the four surveyed beaches and the two sampling years failed to show a global pattern but showed significant local structure, that is repulsive structure with negative spatial autocorrelation, suggesting that nematodes located on different plants are genetically distinct (Fig. S2). In the same way, spatial genetic structure was only detected among indi- Beyond this level, average kinship estimates did not depart from spatial randomness. These results suggested a very short-distance spatial autocorrelation with a lack of detectable IBD within the four studied beaches for both sampling years. For 82% of the 68 tested datasets using Bayesian genetic clustering, ∆K were <10 which is uninformative and prevent reliable assignment of individuals. In the resting 12%, ∆K were higher, but in all cases, the most probable number of genetic clusters was 1. This analysis was thus not able to detect any genetic structuring at a beach scale but also at the host plant scale and results were not included in this study.

| Genetic data
No significant linkage disequilibrium was detected, and microsatellite loci were consequently considered as independent. Multilocus genotypic data showed contrasting levels of genetic diversity among sampled beaches (Table 1) (Table 1). Some (20% in 2012 and 65% in 2013) were significantly positive, suggesting heterozygote deficiencies (Table 1). with the restricted dataset may be due to a lack of statistical power.

| Detection of migrants
We

| Genetic drift
Isolation by distance patterns differed among sampled beaches, but they remained quite stable between 2012 and 2013 (Fig. 1).
Similarly, mean and variance of pairwise F ST values did not change between the two sampling sessions (Fig. 2). Simulations showed that the slopes of IBD patterns observed in 2013 were in the 95% confidence interval of what could be expected under pure drift, after four or ten generations, in all beaches (Fig. 3). This result was independent of effective population sizes except for Granville Sud,

| IBD patterns: the question of migration-drift equilibrium
The genetic differentiation among wild H. schachtii populations showed contrasting patterns of population structure among beaches, sampled years, and spatial scales. IBD patterns that are unstable across spatial scales suggest that the population structure is not at migration-drift equilibrium in H. schachtii (Hutchison & Templeton, 1999). Therefore, observed IBD patterns may not result from current gene flow.
This hypothesis is consistent with spatial autocorrelation results, which did not support the occurrence of IBD process within the four beaches. This hypothesis is also well in line with our results on first-generation migrant detection and simulations which showed that most of the (few) migrants detected in our empirical dataset were probably false positives. It is also worth mentioning that even if they are true migrants, they may not all contribute to maintain IBD, depending on their origin (see section "Migrant detection: empirical data"). This small number of detected migrants may be related to the low genetic variability of the microsatellite markers used, which is known to limit the statistical power of assignment tests (Waples & Gaggiotti, 2006). However, the concordance of results from observed population genetic differentiation with simulated data under pure genetic drift suggests that the low observed number of migrant was not just an artifact. According to our results, there were no or, at most, four migrants observed in sampled wild H. schachtii populations, which is not enough to maintain the IBD patterns observed in Granville Sud and Saint Léonard. One may argue that only a few migrants are required to prevent population differentiation and counteract the influence of genetic drift (Slatkin, 1985). However, simulated data under genetic drift provided further support that genetic drift alone can explain the observed variation in population genetic structure between the two sampling years for at least three of the four surveyed beaches. This result would not be expected with ongoing gene flow (Hutchison & Templeton, 1999). Our results thus suggest that wild H. schachtii populations were not at migration-drift equilibrium and probably exchange few or no genes at scales ranging from 10 cm to 150 km.
In nonequilibrium situations, patterns of genetic structure are dominated by historical factors, such as colonization history (Austerlitz, Mariette, Machon, Gouyon, & Godelle, 2000;Ibrahim, Nichols, & Hewitt, 1996). This may even be more pronounced in populations that are founded by seeds (plants) or cysts (nematodes), which are forms that can survive belowground until local conditions allow their development. Once a potential host plant grows in a place where cysts are present, the local multiplication of nematodes may generate a patchy distribution of populations and heterogeneity in infestation levels among hosts, as observed in several nematode species (Gavassoni, Tylka, & Munkvold, 2001;Jan et al., 2016;Villate et al., 2008). Most nematode movements in coastal populations may be thus related to exceptional events such as high tides during storms.
This hypothesis may explain the results observed in the fourth beach, Granville Sud, where it seems unlikely that the loss of the IBD pattern between 2012 and 2013 results from regular gene flow, because we did not detect more migrants on this beach than on the other beaches and because the slope of IBD pattern observed in 2013 in this beach is smaller than expected under pure genetic drift. However, further studies with an exhaustive sampling of all host plants of a beach conducted over several years are required to gain further insight into the influence of colonization processes on the observed patterns of genetic structure.

| Definition of wild H. schachtii populations
Our results suggest that passive dispersal through cysts movement is actually limited in the wild. This is consistent with the small effective population sizes (N e ranging from 50 to 400) observed in wild H. schachtii populations (Jan et al., 2016). Small effective sizes are one consequence of limited gene flow among populations and support the hypothesis that H. schachtii is a species constituted of genetically disconnected populations in the wild, even at very small spatial scale. Interestingly, a recent study documented a Wahlund effect in some wild H. schachtii populations sampled at the host scale, suggesting that active dispersal of this species can also be extremely limited even between subpopulations that occupy the same host plant (Montarry et al., 2015). These subpopulations may thus correspond to the smallest genetic unit at which individuals interbreed, that is the actual boundaries of populations in H. schachtii (Waples & Gaggiotti, 2006).
Strong population genetic structure has already been observed in other parasites that have limited or no active dispersal abilities (Blouin, Liu, & Berry, 1999). Actually, small isolated populations are usually expected in parasites (Price, 1980), particularly in wild plant pathosystems characterized by low host densities, but was barely observed in plant parasite species. In these pathosystems, genetic drift likely has a strong influence and wild parasite populations are more vulnerable to local extinction and/or are maladapted to their local hosts due to their lower dispersal capabilities compared with that of their hosts (Gandon, Capowiez, Dubois, Michalakis, & Olivieri, 1996;Morgan, Gandon, & Buckling, 2005;Thrall & Burdon, 2002).

| Human-mediated gene flow and consequences for resistant varieties management
Field populations of H. schachtii have a similar substructuring patterns than wild populations, suggesting that active dispersal capabilities of nematodes are equivalent in wild and field conditions (Montarry et al., 2015). By contrast, in agrosystems, low genetic differentiation suggests that nematode populations are connected through gene flow among fields separated by distances that reach 150 km. Indeed, maximum pairwise F ST reached .1 (Plantard & Porte, 2004;Porte et al., 1999), which is three times less than what we observed across a similar geographical scale in wild populations. These studies attributed the scale of gene flow in agrosystems to soil transport during sugar beet crops harvesting which is characterized by the loss of several tons of soil per hectare and per harvest (Ruysschaert, Poesen, Wauters, Govers, & Verstraeten, 2007). The inadvertent transport of a massive quantity of soil may passively disperse large numbers of cysts, resulting in higher levels of gene flow among cultivated fields than among wild H. schachtii populations. Anthropogenic influence on pathogen dispersal has already been reported for several species (Lebarbenchon et al., 2008;Morgan et al., 2012), including nematodes. Patterns of population genetic structure matched those of agricultural practices within and among field populations of G. tabacum and X. index (Alenda et al., 2014;Villate et al., 2010), and potato tuber trade favors the spreading of G. pallida and G. rostochiensis worldwide (Boucher et al., 2013;Plantard et al., 2008).
Passive dispersal of cysts in fields results in a higher dispersal capability of nematodes than of their crop hosts. In agrosystems, gene flow is thus a source of genetic variation that facilitates the adaptation of pathogen populations to their local hosts (Gandon et al., 1996;Greischar & Koskella, 2007;Morgan et al., 2005;Thrall & Burdon, 2002). This phenomenon can be strengthened by agrosystems characteristics, such as large-scale host uniformity (Montarry, Glais, Corbiere, & Andrivon, 2008). Human-mediated gene flow among fields may thus reduce the durability of resistant varieties used to control crop pathogens.
Limiting accidental gene flow among parasite populations may thus be a worthwhile improvement of the management of field nematode populations because it (i) would isolate avirulent from virulent populations and (ii) reduce the effective population size of field nematode populations. The influence of genetic drift would be favored within populations and may lighten the selective pressure imposed by sugar beet cultivars and thus decrease the ability of nematode populations to overcome resistance over large spatial scales. Such confinement strategies have already been strongly advocated by several authors (Burdon & Thrall, 2008;Stukenbrock & McDonald, 2008) but represent a real challenge. Indeed, they would require a rigorous monitoring of all fields belonging to the same farm to detect infested ones and to prevent soil mixing among fields by, for example, cleaning agricultural equipment. Similarly, in case of machine sharing among agricultural exploitations, a tight collaboration between farmers of the same production area would be helpful to manage parasite dissemination at higher spatio-temporal scales. These approaches have already been proposed for aerial fungi and would be suitable to maintain the whole pathogen metapopulation maladapted to its hosts and to prevent epidemics (Bousset, 2014;Bousset & Chèvre, 2013). However, all these examples are mostly time-consuming for farmers which may make them difficult to implement in practice, or they still have to be developed. Thus, additional experimentations and developments will be necessary to fully integrate our conclusions in sugar beet production systems.

| CONCLUSION
This study used temporal sampling to investigate the genetic structure of wild populations of the beet cyst nematode. Wild populations of H. schachtii were characterized by a nonequilibrium population structure, weak levels of gene flow beyond the scale of the host plant, and a non-negligible impact of genetic drift. This pattern appeared stable over a short period of time, suggesting no isolated disturbance.
A wild population of H. schachtii appears thus to be defined under the host plant scale, which suggests that human activities strongly influence passive dispersal among field nematode populations. Thus, the management of durable crops' genetic protection against telluric pathogens may gain from developing methods that combine passive dispersal limitation with other methods that help reduce parasite effective population sizes. Finally, this study illustrates how worthy would be data on the genetic structure of wild and field populations of other plant-parasitic nematodes species which are currently still lacking.