Aims and objectives/purpose/research questions:This paper describes and analyses the speech from the Senegalese migrant community in the city of Pescara (Italy) from a contact-linguistics perspective. The focus is on the phenomena of code-mixing found when the Italian language, that of the host community, is added to the complex repertoire of Wolof/French multilingual speakers.Design/methodology/approach:The patterns of code-mixing and the composition of trilingual speech are presented and discussed following the Muysken grammatical model for classifying features of language-contact discourse.Data and analysis:Primary data have been collected recording spontaneous interactions among a group of first-generation migrants that have lived in Italy for about 10 years and are all speakers of a fluent learner-variety of Italian.Findings/conclusions:It will be shown that insertional and alternational code-mixing patterns are present in both the bilingual (French/Wolof) and the trilingual (Italian/French/Wolof) speech, but with different degrees of variation and distribution.Originality:Besides documenting the trilingual speech framed in a specific yet new migration contact setting, the author considers the role that the two languages of prestige, French and Italian, play within the structural composition of both types of code-mixing, especially with regards to their relationship with Wolof.Significance/implications:The results achieved contribute to the general theory of the grammatical model by reaffirming the properties of the patterns of insertion and alternation, but also by revealing other ambiguous forms of code-mixing that might suggest a future revision of the model.