Legibility of new reflective signs made from representative combinations of background and legend components was estimated in a quasi-natural environment by measuring subjects' ability to resolve elements of a test pattern applied uniformly to each sign. Silkscreened signs and freeway guide signs with demountable copy were treated in separate full-scale, controlled experiments. Sign performance in both cases was usually better than one arc minute resolution, with an experimental precision of 0.1 arc minute. Variation among brands in the silkscreened group was almost negligible despite the range of sign luminance obtained. The freeway type signs showed generally better performance than the others. Unit reflector legend was superior to cut-out legend against any background. Reflective backgrounds performed slightly better at high levels of illumination than at low levels. Non-reflective backgrounds performed better at low levels of illumination than at high levels. Cost projections based on current materials prices and an assumed longer life for high intensity sheet revealed that even a conservatively-estimated ten percent rate of sign destruction by vandalism and crashes would eradicate and even reverse expected savings arising out of conversion from engineer grade to high intensity sheet. /Author/.
Abstract