non-Abelian groups

Concerning Colin McLarty's questions: The category of groups should really be seen as a 2-category, in fact as part of the 2-category of groupoids.