The purpose of this paper is to evaluate the accuracy with which the location of turbulent separation can be predicted on two-dimensional and axisymmetric bodies. The evaluation was made by studying a considerable number of flows that had separation. Calculated separation points were compared with the experimentally measured location. Four methods of predicting separation in turbulent flow were evaluated. They were Goldschmied's method, Stratford's method, Head's method, and the Cebeci-Smith method. It was concluded from the study that the last three listed methods predict separation points with the reliability and accuracy needed for aerodynamic design purposes.Nomenclature c =chord c f = local skin friction coefficient, r w /(l/2)pw e 2 C p = pressure coefficient, (p -p m )l(l/2pu m 2 ) D = diameter h = total head H = shape factor, S*/d k = mixing-length constant L = reference body length p = pressure R c = chord-Reynolds number, u^cjv R D = diameter-Reynolds number, u^Djv R L = length-Reynolds number, u^L/v R x = local Reynolds number, u e x/v R 0 = Reynolds number, ujdjv u,v = x and y components of velocity, respectively #* = friction velocity, (r w /p) 1/2 x -streamwise distance y = distance normal to the surface of the body a = angle of attack 8 = boundary-layer thickness 8* = displacement thickness, J^ (1 -u/u e )dy 6 = momentum thickness, J^ u/u e (l -u/u e )dy IJL = dynamic viscosity v = kinematic viscosity p = density r = shear stress (/> = angle from stagnation point Subscripts e = edge of the boundary layer m = minimum pressure point sep = separation point tr = transition w = wall co = freestream conditions