Subject to the above qualifications, CLmax is not greatly reduced until the front edge of the roughness extends forward of 0.1 chord, at which stage also the drag increment due to the roughness rapidly becomes much larger. No upper-surface, flow separation was found. Both sparse and dense distributions were used and they were intended to simulate the hoar frost deposit remaining after partial cleaning of the forward part of the aerofoil. The 13.1-percent-thick, UI-1720 airfoil was found to achieve the predicted maximum lift coefficient of nearly 2.0. Distributed roughness was progressively applied from the trailing edge forwards over the upper surface of the aerofoil. Boundary-layer control by suction in the vicinity of the wing/wall junction is shown to improve the two-dimensionality in the early stages of separation, but does not inhibit the appearance of three-dimensional flow at and beyond the stall. On the other, it was found that the flow broke down into a three-dimensional pattern when an appreciable extent of separation was present.

On the one hand, a laminar separation bubble disappears intermittently for reasons that are not fully understood. The measurement of CLmax at a Reynolds number of 2.88 x 10power6 is uncertain for two reasons. References 3 to 12 document the research effort on NASA low- and medium-speed airfoils. These airfoils provided higher cruise Mach numbers than the low-speed airfoils while retaimng good high-lift, low-speed characteristics.

Results are presented for the aerodynamic characteristics of NACA 0012 aerofoil section at Reynolds numbers of 2.88 x 10power6 and 1.44 x 10power6 with some indications of scale effect at other Reynolds numbers. between the low-speed airfoils and the supercritical airfoils for application on light executive-type air-planes.

Low speed airfoil