12Inferring the dispersal processes of vector-borne plant pathogens is a great challenge because the 13 plausible epidemiological scenarios often involve complex spread patterns at multiple scales. 14 European stone fruit yellows (ESFY), a disease caused by 'Candidatus Phytoplasma prunorum' 15 and disseminated via planting material and vectors belonging to the species Cacopsylla pruni, is a 16 major threat for stone fruit production throughout Europe. The spatial genetic structure of the 17 pathogen was investigated at multiple scales by the application of a combination of statistical 18 approaches to a large dataset obtained through the intensive sampling of the three ecological 19 compartments hosting the pathogen (psyllids, wild and cultivated Prunus) in three Prunus-growing 20 regions in France. This work revealed new haplotypes of 'Ca. P. prunorum', and showed that the 21 prevalence of the different haplotypes of this pathogen is highly uneven between all regions, and 22 within two of them. In addition, we identified a significant clustering of similar haplotypes within 23 a radius of at most 50 km, but not between nearby wild and cultivated Prunus. We also provide 24 evidence that the two species of the C. pruni complex are unevenly distributed but can spread the 25 pathogen, and that infected plants are transferred between production areas. Altogether, this work 26 supports a main epidemiological scenario where 'Ca. P. prunorum' is endemic in, and mostly 27 acquired from, wild Prunus by immature C. pruni (of both species) who then migrate to "shelter 28 plants" that epidemiologically connect sites less than 50 km apart by later providing infectious mature C. pruni to their "migration basins", which differ in their haplotypic composition. We argue 30 that such multiscale studies would be very useful for other pathosystems. 3132 Keywords: C. pruni, epidemiology, reservoir, join count. 33 34 processes has been a central question for decades in ecology 11-14 and population genetics 15,16 , the 48 development of large-scale pattern-oriented approaches to understand the processes shaping the 49 genetic structure of a population is recent 17-21 . The basic idea of these approaches is to estimate the 50 distance at which two samples become genetically independent by relating genetic data and spatial 51 information obtained for a set of samples, through combining methods of geostatistics 22 and 52 population or landscape genetics 23 . In particular, approaches used for qualitative data comprise 53 join-counts 24 and permutation tests to identify distances at which genetic dissimilarity between 54 pairs of individuals is significantly lower than expected 25-28 . 55