Abstract It is not intended in this paper to attempt to mention all the wide variety of aspects of rubber friction which have been discussed in the literature. The objective is to draw attention to two interesting aspects, namely, that the maximum friction does not arise until some movement occurs and that, in the general case, velocity alone is not the sole deciding factor but rather that the friction developed is a function of a ratio of velocities. These phenomena appear to be fundamental to the behavior of rolling wheels and other dynamic friction problems. In special cases, more particularly those in which the relative sliding velocities are high, the frictional phenomena reduce to being velocity dependent. Experiments show that, while consistently high friction values are obtained at slow sliding speeds on rough emery cloth and similar surfaces (coefficients of 0.9– 1.5 say), the highest coefficients (2.5– 3.0 or even more) at slow sliding speeds are obtained on smooth glass and other materials with a glass-like surface (e.g. highly polished bakelite) provided the surfaces are clean and dry and the rubber surface is free from bloom or wax. This is consistent with the view that a high coefficient of friction is obtained provided the rubber itself makes contact with the other solid in some areas at least. Provided such surface films as are present are broken through, at least in some places, the necessary intimate contact is made and a high friction obtained. To ensure intimate contact between the main mass of rubber and a body with a smooth surface, both surfaces must be clean and dry or the rubber be shaped with a tread pattern capable of exerting a wiping action to remove the surface film. Not all surface films can be removed in this manner. For example, bloom on the rubber needs to be cleaned off in other ways, such as by rubbing with a rag dipped in solvent. In the case of rough or coarse textured materials the high stresses developed on the points of the surface ensures intimate contact by breaking through the surface films, even including bloom on the rubber. The presence of surface films of any kind would be expected to reduce the friction developed because in those areas in which they carry load they provide a medium of lower tangential shear strength than the bulk of the rubber and in many cases virtually act as a lubricant carrying the normal load with very low tangential drag. The suggestions above explain the small importance of cleanliness in the case of rough surfaces and the greater importance in the case of a smooth surface. The explanations agree, at least qualitatively, with the observation that the highest frictions are obtained on clean, smooth surfaces as then the maximum actual area of contact is possible.