Until now, with regard to the hobo system of hybrid dysgenesis, natural populations of Drosophila melanogaster have been investigated using only two criteria: at the molecular level, the presence or absence of XhoI fragments 2·6 kb long or smaller; and/or at the genetic level, the ability to induce gonadal dysgenesis sterility in crosses A (females of an E reference strain crossed with males under test) and A* (females under test crossed with males of an H reference strain). Recently, analyses of laboratory strains using these criteria as well as the mobilization of two reporter genes, the male recombination and the number of ‘TPE’ repeats in the S region, revealed a lack of correlation between the different dysgenic parameters themselves, and also between these parameters and the molecular characteristics of the strains. Thirteen current strains derived from world populations were therefore investigated with regard to all these dysgenic traits, to determine discriminating criteria providing a robust method of classifying natural populations and deducing the dynamics of hobo elements in these populations. We show, as in laboratory strains, a lack of correlation between the parameters studied. Therefore, the significance of each of them as well as the nature of hobo hybrid dysgenesis are discussed, to propose an analysis method of the hobo system applicable to natural populations. According to the geographical distribution of hobo activities in world populations and to the variable polymorphism of the number of ‘TPE’ repeats, we propose a new scenario for the invasion of D. melanogaster by hobo elements.