Abstract
A detailed discussion of the requirements of spin and orbital symmetry in the pair‐correlation theory of Sinanoğlu is presented. The reasons why the theory of Sinanoğlu does not satisfy these requirements a priori are exposed, and it is shown that if some modifications to it are made then the symmetry properties can be incorporated in his scheme without undue complications. This is illustrated by a number of examples, and the relation to other recent work on this problem is pointed out.