An experimental and theoretical investigation into the deposition of eolian sand near porous fences, and their use in creating and stabilising sand dunes, is reported. Measurements in a wind tunnel of sand deposition near small porous fences of "infinite" length set normal to the flow show that a dune is formed with its crest about two fence heights downstream of the fence and that the optimum porosity for maximum deposition is approximately 38%. The shape of the deposits compare well with those measured at full size fences in the field. For the same miniature porous fences, Preston tubes are used to measure boundary shear stress. By assuming that this boundary shear stress is similarly distributed in sand-laden flow, and by using a uniform sand transport rate formula, a basic numerical model describing the deposition is developed. The model shows reasonable success in predicting dune growth at miniature and full size fences. Better agreement is found when an experimentally derived transport rate formula is used. The results of the basic model can be improved by incorporating a modification which accounts for "jetting" through fence members. Further improvement in prediction of the size of deposit upstream of the fence is demonstrated by postulating that velocity profiles measured upstream of a fence can be used as a substitute for shear velocity. A hill-shape introduced into the wind tunnel causes an increase in shear stress on its upstream surface. Using this result and a simple modification of existing theory, the importance of including the surface curvature effect in the numerical model is shown. Other factors which could influence the transport of sand and its deposition at porous fences are examined but not resolved. It is concluded that there is scope for the application of the present numerical prediction methods in practical situations and recommendations are made for their use, although these are limited to circumstances where the flow is sensibly two-dimensional.