The thesis focuses on gender in primary school and the aim was to study how girls and boys construct their subjectivities in accordance with current gender discourses and how they take up those discourses in school practices. Special attention has been paid to the students' fluid subjectivities. The theoretical frameworks used are Bronwyn Davies' postructural subjectivity theory, Robert W Connells' structural concepts and cultural-sociological research studies concerning multicultural identity. The study was carried out at two schools, one ethnically homogeneous and one ethnically heterogeneous, in six classes encompassing years 0-2 and 5-6. It is ethnographic in nature and includes classroom observations, diaries, biographies, drawings, interviews with students and schoolteachers/headteachers and videotaping on a restricted scale. Data was collected over a period of approximately two school years. The results consist of local gender discourses emanating from the datamaterial and also of poststructural analyses of protocols from lessons etc. Five feminine and six masculine gender discourses, named student types, have been diagnosed: Sporty girl, Barbie, Feminist, Academic girl and Motherly girl and Macho boy, Honourable boy, Academic, Joker, Gentle boy and Ken. These student types are abstract discursive constructions developed from positions the student took up in a more or less repetitive way. They apperar in all classes but with varying frequency due to the influence of the schools' interest profile, leading teachers or leading students. Certain gender discourses are influenced by commercial trends in society, others are characterized by reactions towards the school's academic discourses. Students from working class backgrounds often take up positions as Macho boy or Sporty girl while middle class students dominate the type Academic boy/girl. Otherwise the positions are independent of social class. Immigrant students take up the most common discourses, probably an effect of ambitions to normalise to the majority culture. The analysis reveals that a dualistic and hierarchical gender structure, with male superiority was developed in all school classes and also among the boys in their own gendergroups, and among the girls but in a lesser degree. Teachers' discourses, education strategies, group size and the student's ages influence the gender order during lessons but less so during breaks. Both girls and boys, and some teachers, shift positions and even cross gender boundaries and the younger students (year 0-2) are more flexiable as also are the girls. This is considered to provide openings for changes in gender patterns. Consistently taking up equality discourse in practice influenced the gender order in one class. Some boys showed multiple subjectivities free from desire for power and some girls also wanted to break the gender barrier. Ideas about innate equalities between the genders were common and these circumstances might provide good resources for work aimed at changing gender structures. Macho and Barbie discourses ought to be questioned from the perspective of power. The results also show that cultural meetings in the classroom are characterised by the dominance of the majority culture. Immigrant students in accordance with the curriculum should experience integration taking place from two directions, enriching and strengthing their subjectivity process and also that of their fellow students. / digitalisering@umu