Isolated tsetse populations constitute a target for tsetse control programmes in endemic countries, since their isolation, if demonstrated, allows control without reinvasion risk from neighbouring populations. Population genetic parameters, such as the fixation index, have proven useful to assess isolation status, and should also give important information on the divergence time since isolation. We gathered results obtained from different datasets regarding several examples of putatively totally isolated tsetse populations of different tsetse species: Glossina palpalis gambiensis in Guinea, in the Niayes of Senegal, and in the sacred wood of Bama in Burkina Faso; G. tachinoides from Bitou and Pama in South-East Burkina Faso. The different levels of isolation were compared to differentiation between the two subspecies G. p. gambiensis and G. p. palpalis which both occur allopatrically along the Comoe River in Ivory Coast. We also use some historical evidence to calibrate differentiation speed and give estimates of time since separation for the different cases studied. Discrepancies mostly come from underestimate of effective population sizes, and we propose improving sampling design and genetic markers quality to circumvent such caveats. (Résumé d'auteur)