Résumé : Language and music are two human universals, but the precise nature of their cognitive relationships remains unanswered. We addressed this issue by examining the relationships between the deficits of these systems in children, namely between developmental dyslexia (DD, a specific disorder of reading acquisition) and congenital amusia (a lifelong disorder in music perception and production). Our double objective was (1) to estimate the extent of the comorbidity between DD and congenital amusia in children and (2) to examine, at a behavioral level, the factors that could explain their co-occurrence. To do so, we first measured the prevalence of congenital amusia in a relatively large sample of children with DD (Study 1). Then, in order to test the hypothesis of shared underlying impairments between DD and congenital amusia, we carried out a comparative behavioral examination of the phonological and pitch abilities of dyslexic children with and without congenital amusia for several cognitive functions, including auditory short-term memory, perceptual awareness, and attention (Study 2). As this does not allow to conclude whether the observed impairments merely coexist or are causally related, a second, longitudinal, approach, followed the evolution of the reading and musical skills of children at risk for DD in relation to their early abilities in these domains (Study 3). Overall, our data suggest that DD and congenital amusia coexist in a significant proportion of children (34%). This association likely results from the combined effects of a plurality of factors, among which a shared underlying impairment in auditory serial-order memory and an influence of early musical skills on later literacy acquisition via phonology. These results shed new light on the currently underexplored relationship between DD and congenital amusia, but also on understanding the broader issue of comorbidities between neurodevelopmental disorders. They are consistent with the idea that certain key skills are shared between music and speech processing. Further research will be needed to investigate the potential clinical applications of these results, both for the construction of early detection batteries for learning disabilities and for the development of remediation programs.