On the microgenesis of illusory figures: a failure to replicate

Reynolds found in 1981 that with increased viewing time of a pattern which may or may not produce illusory contours there were: first, reports of the pattern without an illusory figure; then, at longer exposures, an increase in the frequency with which illusory figures were reported; and then, with still longer exposures, a decrease in such reports if the pattern contained elements which tended to contradict the possibility of such a figure. Unfortunately, however, three attempts to replicate these potentially very important findings--with the aid of substantially improved methodology--consistently failed to do so. It is suggested that this failure, although it is disappointing to those who subscribe to a 'problem solving' explanation of illusory contours, may not constitute a strong refutation of such a theory. Regardless, the purpose in the report is to clarify and rectify the published record concerning this part of the evidential underpinnings of that theory.