This thesis is concerned with adult L1 attrition in the case of highly-educated long-term Polish immigrants in Sweden. The study sheds light on two classical issues pertaining to L1 attrition, namely what happens to a fully developed mature language system in an immigrant context, and why it happens. Specifically we aim to answer the following questions: (1) Are Polish speakers in Sweden different from comparable individuals in Poland with respect to (i) judgement and use of a number of Polish linguistic features (se keywords below), and (ii) hesitation phenomena, i.e. ability to be quick and easy and linguistic insecurity? (2) Is the variation in linguistic results dependent on how often and in what context the Polish language is used and/or which attitudes the individuals have towards it as well as how long they have been living in Sweden? One of the main contributions of the study regards methodology. The data is analysed in three steps: an initial focus on group comparisons shifts to the analysis of individual results in relation to the variation observed in the comparison group, and finally to a holistic view of the attrition effects or their absence. It is suggested that in studies on adult L1 attrition, where the effects are expected to be relatively cosmetic, the range of the linguistic behaviour in the comparison group as a reference point as well as a holistic perspective on individual results gives a more truthful picture of the attrition process. In addition the study shows that attrition effects are present in some individuals (60 %) to a different degree. The most common effect overall is linguistic insecurity followed by the overuse of 1st person pronouns as explicit subjects and to a lesser degree by the overuse of 3rd person pronouns. Surprisingly there is a scanty effect on the other hesitation phenomenon, i.e. the ability to be quick and easy. No or limited effects are observed in other structural areas which basically supports previous findings on L1 adult attrition. Finally, the presence versus absence of the attrition effect can not be straightforwardly related either to language use, attitudes or length of residence, with one exception being linguistic insecurity.