Résumé
The Algerian localities of the Gour Lazib area (Early or early Middle Eocene) have yielded an important mammalian fauna. The Hyaenodontida are well represented in this fauna: three species-two are new- are reported. The genus Glibzegdouia, which has been previously described as a possible Carnivora, is now clearly referred to the Hyaenodontida. It appears morphologically close to Masrasector and Dissopsalis. A new genus, Furodon, is described. It appears morphologically close to the oldest Pterodon species. This discovery supports an African origin for the hyainailourine genus Pterodon and related genera (e.g. Hyainailouros, Akhnatenavus). Two very small lower molars are referred to a new genus Parvavorodon, which is also referred to Hyainailourinae. The localities of the Gour Lazib area therefore show important hyaenodontid diversity for the Early or early Middle Eocene. We performed a new phylogenetic analysis to question the relationships between the African, Asian, North American and European hyaenodontidans. Our study supports the endemism and originality of the Asian proviverrines' Indohyaenodon, Paratritemnodon, Kyawdawia and Yarshea; we propose a new subfamily: Indohyaenodontinae. The African proviverrines' (e.g. Masrasector, Anasinopa, Dissopsalis and Glibzegdouia), which are notably characterized by large premolars and the presence of a wide talonid on the molars, are close to the enigmatic African Teratodon. We therefore propose to refer them to Teratodontinae. The Hyainailourinae, which include the new genus Furodon, are characterized by the presence of secant dentition related to a hypercarnivorous diet. They appear phylogenetically close to the African Koholiinae. The genus Metapterodon is referred to Koholiinae based on the phylogenetic analysis. The African origin of the Teratodontinae and Hyainailourinae is supported by Glibzegdouia and Furodon. The origination of several subfamilies in Africa supports the hypothesis of an African origin for the Hyaenodontida. The origination of the Teratodontinae in Africa contradicts the previous hypotheses of Afro-Asian proviverrines'.