A comment on the locations of firms in Cournot spatial discrimination models