Abstract: On the basis of various observations, we argue that there is spatial variation of the time-averaged wall shear rate in arteries, both overall and locally. From our own observations, and those of others, we show that the distribution of early atheroma in man is coincident with those regions in which arterial wall shear rate is expected to be relatively low, while the development of lesions is inhibited or retarded in those regions in which wall shear rate is expected to be relatively high. Such a correlation is inconsistent with a proposal, made by several workers, that there is a causative relation between arterial blood mechanics and the development of atheroma, i.e. that atheroma is associated with wall damage due to the motion of blood. Instead it immediately suggests that the process is associated with shear dependent mass transport phenomena. It has been demonstrated by others that mass transport, in the inner part of the arterial wall, is dominantly to and from blood flowing within the lumen. We review theory relevant to diffusional mass transport across such a sheared interface, and examine available experimental evidence, relating to normally occurring (quasi-steady state) and experimentally induced (transient-type) atheroma, as well as the distribution of cholesterol in arteries. These results are considered in the light of simple theoretical schemes which we develop for the movement of cholesterol, in particular, although the arguments may also be relevant to other diffusing species. Shear enhances mass transport by means of a steepening effect on the concentration gradient, thus diffusion of material from a wall is promoted when material which has already diffused is swept rapidly away, so that the concentration gradient leading to further diffusion remains steep. However, the influence of shear on the diffusion of a species, say, from just within the wall of an artery to fluid in the main stream, depends upon the relative resistances to its diffusion from within the wall to surface fluid (wall phase) and from surface fluid to fluid in the main stream (blood phase); diffusion is not appreciably shear dependent if the latter resistance is small compared with the former. Assuming simplified flow conditions and that as suggested by others cholesterol is transported in blood in association with plasma protein, we can estimate resistance for diffusion of this species in the blood phase, for different stations in the arterial system. However, we possess no definite comparable information for the wall phase; we conjecture that this resistance is relatively small, and assume shear dependence of diffusional transport of cholesterol between arterial walls and intraluminal blood. We find that a net flux of cholesterol from blood to wall, as has been suggested by others, cannot account, in terms of the proposed schemes, for the observed normally occurring (quasi-steady state) distribution of atheromatous lesions in man and in animals; mass transport is inhibited in low shear regions by the thick diffusional boundary layer. Instead it appears that cholesterol, which has been shown by others to be synthesized in arterial walls, accumulates in low shear regions because its local diffusional efflux from wall to blood is inhibited by the reduced concentration gradient. Given suitable values for relevant parameters, the theoretical schemes are also able to account for adequacy of supply of precursor to the wall for cholesterol synthesis, for the preferential occurrence that we now recognize of lesions in high shear regions in response to sudden natural or experimental elevation of blood cholesterol, and for the responses to administration of labelled cholesterol (transient type phenomena); it appears therefore possible, in terms of these schemes, to unify naturally occurring and experimentally induced atheroma. It is reported by others that platelets are associated only with advanced lesions; the correlation of naturally occurring atheroma with low shear regions, and transient type lesions with high shear regions, with the fluid mechanics being unaltered in the two situations, provides no support for the implication of platelets in the development of early atheroma. It appears that wall shear rate may be a major controlling factor in the development of atheroma, i.e. that high shear, such as is associated for example with increased cardiac output in exercise, will retard progression of the process. Its progression will also be retarded by any means which reduces the accumulation of atheromatous material, by influencing its rate of net production or diffusion.