Laboratory flume observations were conducted to examine the influence of the projecting shells of hippuritidrudists on water movement in the benthic boundary layer. Visualization of flow about fossil and model hippuritid shells revealed no significant differences; both showed similar, characteristic flow patterns dependent primarily on the orientation (vertical, inclined upstream, inclined downstream) of the cylinder. All orientations produced similar general flow patterns: (1) fluid near the top of the cylinder accelerated over the cylinder, separating from the cylinder at the highest edge; (2) fluid encountering the cylinder between the top and the base passed around the cylinder and a portion was entrained in the wake; (3) fluid near the substrate passed around the cylinder, was entrained in the wake, and then rose in a pair of alternating vortices to the top of the cylinder and was accelerated downstream. For both vertical and upstream-inclined cylinders, the top of the cylinder was bathed with water drawn from the mainstream flow. Flow around downstream-inclined cylinders, however, showed two distinct differences from the other orientations. First, flow separated from the top of the cylinder at the upstream edge so that the top of the cylinder was bathed with fluid drawn from the wake. Second, the vortices behind downstream-inclined cylinders were greatly amplified, even briefly lifting sediment grains at mainstream speeds in which the other two orientations showed no disturbance of the sediment behind the cylinder. The enhanced vortices strongly drew water from the wake and off the sediment surface which fed the flow within the separation bubble over the top of the cylinder; the top was thus bathed in a mixture of fluid from both the mainstream flow and the surface of the sediment. Flow from the excurrent openings of the models did not mix with fluid in the separation bubble. Because most water filtered by hippuritids entered the pores penetrating the upper valve, animals oriented vertically or inclined upstream would primarily sample water from the mainstream flow; animals inclined downstream would sample a mixture of water from the mainstream flow and water drawn from the sediment-water interface (which should be enriched in bacteria and detrital organic particles). A downstream-inclined posture would have been advantageous in nutrient poor waters because the quantity and variety of available organic particles would have been enhanced.