Research supervisors' different ways of experiencing their supervision of doctoral students are analysed in terms of the students' questions and problems as they relate to the supervisor's research, and what consequences this connection, or non-connection, to the supervisor's research has for supervision and the role of supervisor. Thirty supervisors of doctoral students at different faculties at a university in Sweden were interviewed. The results illustrate two supervision structures, called "research practice-oriented" and "research relation-oriented supervision". The principal differences between these two ways of structuring supervision consist of whether the supervisor and the doctoral student participate in a common research practice and share objects of research with the same or a related research approach, or whether the doctoral student's research problems and research objects lack a clear connection with the supervisor's research.