Experiments of diffusion are more difficult in ceramics than in metals, because of specific physical, chemical and mechanical properties of ceramic materials. Due to their high melting point, the amount of point defects is generally much less than in metals at the same temperature. That implies very low diffusion coefficients, comparatively to diffusion coefficients measured in metals. The nature of the chemical bond of ceramics gives rise to stronger bonding than in metals and less ability to solve impurities. In these conditions, more segregation and precipitation effects are to be expected in ceramics and these effects may have an influence on the values of the diffusion coefficients. The interpretation of diffusion coefficient values needs a good knowledge of the nature and amount of defects, which are not so easy to control and which depend on the way used to prepare the samples destined to diffusion tests. The aim of this paper is not to give a review of all the values of diffusion coefficients obtained in all ceramics, but to point out some difficulties encountered in determination of diffusion coefficients : penetration depths, comparatively high level of impurity or dopant content in regard with the low amount of intrinsic defects and segregation of impurities to the grain boundaries and to the surface. Examples are given for alpha alumina and chromia.