[en] The microhardness test has been used extensively to give the relative hardness of phases in multiphase materials. However, many problems have arisen in attempting to determine the absolute hardness of the matrix in two-phase materials. For a valid microhardness indication, the indentation should not be near a surface or volume imperfection such as a grain boundary or second phase. The regions near the grain boundary give considerably different hardness values than the regions away from the boundary, as do regions near a harder second phase. Published results which show an increase in matrix microhardness in cemented W--Ni--Fe with percent W are explained as almost entirely due to taking microhardness readings in too thin layers. At high W percentages, the thin layers between the W particles are hardened by the geometrical constraint of the much harder W particles. The amount of this geometrical hardening is predicted from plastic limit analysis. Another factor which is investigated is the size scale involved in hardness testing. ''Picohardness'' (micro-microhardness) tests done recently by other investigators are discussed by considering the applicability of common theoretical models of hardness deformation to very small indenter sizes