Increased magnitude of biomaterial surface roughness and micromachined-grooved surfaces has both been shown to stimulate osteoblast activity, but have not been compared in the same study quantitatively. A series of titanium alloy (Ti6Al4V) samples were prepared using simple machining techniques to undertake such a comparison. Samples were either grit blasted (Gb) or shot peened (Sp) to give random discontinuities, or silicon carbide ground (SiC) to produce ordered grooves. These were compared with micropolished samples (Mp). The samples were coated with a 1 mum continuous coating of hydroxyapatite to remove differences in surface chemistry. Human osteoblast-like cells were seeded onto the materials and metabolic activity, proliferation, alkaline phosphatase activity, and osteocalcin production assessed. Cell responses were highly dependent on the substrate that they were cultured on. Cells cultured on the smooth and ordered (Mp and SiC, respectively) samples had higher metabolic activity and a more elongated morphology than those cultured on the randomly structured Gb or Sp samples. Over 21 days, cell metabolic activity peaked relative to the control between 7 and 14 days on the Mp sample, and between 14 and 21 days on the Gb, Sp, and SiC samples. In common with other researchers, we note that micron scale topography may have potential for influencing osseointegration. More importantly, as the magnitude of the discontinuities on SiC, Gb, and Sp were similar, the differences in cell responses does not appear to lie with the size of the features, but whether the features showed an ordered or disordered structure.