Creaming, in the laboratory sense, is the migration of the dispersed phase of an emulsion, under the influence of buoyancy. The particles float upwards or sink, depending on how large they are and how much less dense or more dense they may be than the continuous phase, and also how viscous or how thixotropic the continuous phase might be. For as long as the particles remain separated, the process is called creaming. Where it is important that either the form or the concentration of the emulsion should be stable, it is desirable that the continuous and the dispersed phases should have similar densities, and it also is desirable that the continuous phase should be viscous or thixotropic. Thixotropy is particularly valuable in paints, sauces, and similar products, partly because it counteracts tendencies towards creaming. It also is important that the particles be as small as practicable because that reduces their tendency to migrate under the influence of buoyant forces. The electric charges on their surfaces should preferably tend to be uniform, so that the particles repel rather than attract each other. Creaming is usually seen as undesirable because it causes difficulties in storage and handling, but it can be useful in special cases, especially where it is desirable to concentrate an emulsion. A particular example is in the separation of dairy cream, either to achieve a desired concentration of butterfat, or to make butter. Depending on whether the dispersed particles are less dense or more dense than the continuous phase, they may move either to the top of a sample, or to the bottom. As already stated, the process of migration is called creaming while the particles of the substance remain separated. In this it differs ideally from flocculation (where particles clump) or emulsion breaking (where particles coalesce). One important difference between creaming and the other two processes; unlike flocculation and breaking, creaming of an emulsion is largely a simple process to reverse.