This study highlight once more the importance to explicitly test for meta-population structure before interpreting the demographic signals detected by unstructured models, similarly to what advocated previously by (Maisano Delser et al., 2019; Rodríguez et al., 2018). If the meta-population structure hypothesis is rejected, the variation of Ne through time can be directly interpreted as the demographic history of the population under investigation, such as the case of tiger shark. Otherwise, this variation is still related to demographic events, but it has to be explained in the light of population structure and its consequence on the rate of coalescence events. We showed by coalescent simulations how to interpret such variation: the recent bottleneck detected by the stairwayplot in demes belonging to a meta-population is a consequence of the coalescence process. In other words, any inferential method implementing an unstructured model will detect such decline (if enough data is available) since it is a property of the gene genealogy. Importantly, the gene genealogy is only slightly affected by recent changes in connectivity if the time of this change in generations is of the same order of the size of the deme.