FIG 7.

The relationship between distractor-induced threshold elevation (i.e. crowding) and salience, for four observers. In each panel, each point represents one combination of target and distractor frequencies. Conditions in which target and distractor frequencies were equal have be excluded from this figure. Log threshold elevation (i.e. the ratio between acuities for targets with and without distractors) has been regressed against the effects of set size (N) on reaction time percent correct (PC). Regression lines are shown. There was no significant relationship between threshold elevation and reaction-time gradient for any observer (not shown), but threshold elevation did increase significantly with the percent-correct gradient for observers CG,GW and JAS.