It has been reported that errors which occur in attempts to align the arms of two acute angles differ in size and direction from errors which occur when the angles are obtuse (Restle, 1969). Other experiments have qualified this conclusion, and it has been suggested that Restle's result was dependent on the use of a forced-choice method with a dual-angle display (Day, 1973). Data are reported here on dot-to-line misalignments using staircase techniques both at the free end and at the vertex end of single acute or obtuse angles. The results essentially agree with Day's method-of-adjustment results, thus implicating the display rather than the method as the cause of the original negative acute-angle effect. However, Day found no difference between the acute-angle effect and the parallelless Poggendorff illusion, while we did obtain a difference under some conditions. Consideration of the magnitudes and directions of errors at both ends of the angles' arms gives pause to any possible accounts of the errors in terms of simple or single whole-of-line effects.