The family Ctenizidae is a worldwide-distributed trapdoor spider group, with a modest number of genera and species but interesting biogeography. Its monophyly has been questioned on the basis of both morphological and molecular evidence. The family is represented in the Mediterranean basin by three genera and nine species: Cteniza and Cyrtocarenum, mostly endemic to the region, andUmmidia, long considered an anthropogenic introduction to the Mediterranean because the bulk of its diversity is in the New World. The taxonomic status of some of the species and genera (e.g. Mediterranean Ummidia species or Cteniza and Cyrtocarenum) has been called into question due to their close morphological affinities. Here, we use a multilocus approach that combines DNA sequence data from three nuclear genes 28S rRNA, EF1γ and H3 to investigate the origins and phylogenetic position of the Mediterranean taxa within the context of ctenizid generic-level diversity. For the first time, all known ctenizid genera are included in a phylogenetic analysis. Additionally, Bayesian relaxed clock methods and specific substitution rates are used to infer the timing of the group's diversification. Our results disagree with the traditional division of the family Ctenizidae into two subfamilies finding them polyphyletic and stress the need for re-evaluating the morphological characters that have been used in the group's classification. Time estimates indicate an ancient origin and long history of Mediterranean ctenizids. The present day disjunct distribution of Ummidia seems to be the result of the opening of the Atlantic Ocean, suggesting a former Laurasian distribution of the genus that is further supported by Baltic amber fossils. Similarly, the opening of the western Mediterranean basin has likely played a key role in the diversification of both Ummidia and Cteniza, whereas the origin of Cyrtocarenum species preceded the breakup of the former continuous landmass that formed the Aegean region. Deep divergence times and reciprocal monophyly support the status of Cteniza and Cyrtocarenum as independent evolutionary lineages. Alternatively, the taxonomic status of Ummidia with regard to the closely related genus Conothele remains unclear; a more thorough sampling of the latter is needed to evaluate whether the synonymy of the two genera is necessary.