Many authors have tried to give proper definitions for fossil dog remains, especially for molars. The data which are available from literature learn that it is possible that even small molars can be ascribed to some wolf-like Pleistocene dogs. Differences between genera are often small. Therefore it is suggested in this article to cluster all data of fossil dogs which have been described in the past and those who are to be found in the future into a recognisable frame, based upon the upper cheek molars: very large, large, small and very small. This has been illustrated by some finds from the Maasvlaktearea in comparison to some wellknown sites in Europe.