The Threshold Hypothesis states that it is students who develop a high level of competence in two languages that most benefit from their bilingualism. This is borne out by findings that show children who have added a second language to their linguistic repertoire are better at acquiring a third language. However, fostering additive bilingualism in education is a complex process. The language used at school is definitely crucial, but active use of the target language is also essential for students to attain a high level of competence. In the case of a minority language, active use can be even more important because its development often depends on school instruction alone. This paper reports on a study carried out on 155 bilingual adolescents, aged 12—16, in a school in the Basque Country(Northern Spain). The language of instruction is Basque(minority language). Spanish(majority language) is taught as a subject for 3-4hours a week as is English (foreign language). The students have been learning through the medium of Basque since the beginning of their schooling. Spanish and English were introduced in grade 3 when students were eight. For half of the students(78) their education through the medium of Basque constitutes a maintenance program whilst it is an immersion program for the other half (77). The results show that all the adolescents are highly competent in Basque and Spanish but that it is the students who use Basque in more language domains, who get the best scores in English.