Just noticeable differences in slant of test lines as a function of retinal eccentricity

Perceived orientation ¿ of a line stimulus presented in a frontal plane deviates systematically from the geometrical slant a of the line stimulus. In fact, ¿ is biassed a few degrees towards the horizontal H or vertical V, whichever is nearest (Bouma and Andriessen, 1968a). If a second line stimulus is present, an extra bias occurs, usually towards the perpendicular of the second line (Bouma and Andriessen, 1970). In the measurements referred to above, the subjects adjusted a dot in line with the test line. The accuracy of observation is reflected in the standard deviation of the settings (of the order of 10 ). It is relatively low for H and V and high for obliques (Bouma and Andriessen, 1968a). In working towards a theory of slant perception (Bouma and Andriessen, 1968b), we aimed at including systematic bias effects as well as accuracies of observation. Since in the method of adjustment each setting might have been based on more than one stimulus presentation, we switched for measuring accuracies to a method in which each response can be attributed to a single presentation. We chose a sequential up-and-down method (Cardozo, 1966) in which the subject had to report on the slant difference of two successive test line presentations. Roughly, we found results similar to those of the method of adjustment. In the meantime we got interested in eccentric vision, where a starting point has been found for processes of visual search (Engel, 1969) and of form interaction (Bouma, 1970; Beerens and Bouma, 1970). This made us wonder how much the accuracy of slant observation decreases with increasing eccentricity of the line stimulus. We report here on some exploratory measurements on the latter question.