A vertical-slot sand trap was constructed incorporating a device to adjust the freedom of flow through the trap. Using a dune-sand of equivalent diameter 0.25 mm, measurements of sand transport were made with this flow so adjusted that no change of sand surface geometry detectable by eye occurred upstream of the trap; this was considered to be a correct procedure. Further measurements were taken with the flow through the trap deliberately set to be (a) more vigorous and (b) less vigorous than the correct setting. In the latter case very significant differences were observed in measured transport rates of sand compared with the ‘correct’measurements. The disparities are of interest because it is commonly assumed that sand grains have enough ‘penetration’to enter a slot trap freely despite a degree of stagnation upstream of it. This does not appear to be a prudent assumption and substantial errors may result from lack of attention to the provision of free air flow through the trap.