Deciphering the colonization processes by which introduced pests invade new areas is essential tolimit the risk of further expansion and/or multiple introductions by increasing vigilance against theidentified key source populations. It also helps defining the ecological characteristics of introducedpopulations and predicting the potential extent of their distribution areas. In some cases, it can helpchoosing strains of potential auxiliary agents to develop biological control strategies. Yet, historicaland observational data often provide incomplete, sparse or even misleading information on invasivepopulations’ history.In the last decade, population genetics has been used as an indirect tool to reconstruct routes ofintroduction, highlighting the complexity and the sometimes counterintuitive nature of the true story.The recent development of new model-based methods, such as approximate Bayesian computation(ABC), has allowed quantitative inferences in case of the complex evolutionary scenarios typicallyencountered during biological invasions. It specifically allows to compare alternative scenariosregarding the number and genetic composition of sources and to explore the number of successiveintroduction events from each source, the number of introduced individuals and the dynamics ofdemographic expansion after each introduction.We describe the principles of the ABC analyses, here applied to microsatellite data and mitochondrialsequences of populations sampled within the native and the invasive range of a species. We will usethree main examples in forest entomology, from which we obtained valuable information about thecolonization routes and dispersal patterns, namely the cedar seed chalcid Megastigmus schimitscheki,the maritime pine bast scale Matscucoccus feytaudi and the Western conifer seed bug Leptoglossusoccidentalis.< Réduire