Tropical regions harbour more species than temperate zones. This pattern, described since more than two centuries as the latitudinal gradient of diversity, remains one of the great mysteries of ecology. In this project, I choose to address three major unanswered fundamental questions regarding the construction of the latitudinal gradient of diversity: (1) Is extinction higher in temperate regions than in the tropics? It is notoriously difficult to estimate extinction rate and its role in the construction of diversity patterns remains poorly understood. Using recent birth-death diversification models, I will analyze the largest paleontological databases available to estimate the contribution of extinction to latitudinal differences in diversity. The second question is (2) Are species interactions stronger in the tropics? Interspecific interactions have often been proposed as one of the main factors favoring speciation in the tropics. Using phenotypic evolution models and global species functional trait data, I will test whether species interactions are detected more often in the tropics than in temperate zones. And finally, I will study the question: (3) What are the molecular mechanisms of adaptation to high latitudes? Using a very large genomic dataset collected in the field on a latitude gradient from California to Alaska, I will compare adaptations in two distant species: the three-spined stickleback and the chinook salmon. This comparison will aim to better understand whether adaptations to latitude are evolving de novo or using a common architecture between species. This project proposes to study the current key questions concerning the understanding of the latitudinal gradient of diversity, while using a multidisciplinary approach integrating paleontology, comparative methods, populations genomics and the most recent global data.