Playing with binary quadratic forms I arrived at seq. A056991 I presume that the stepvector in the PARI program should read [3,3,2,1] instead of [3,3,2,2] -- Met vriendelijke groet, @@i = Arie Groeneveld