"The present study extends and analyzes the previously reported finding that one form of redundant metric figures (made redundant by the sampling restriction that a particular contour detail could occur only once) was more difficult to identify among a sample of similar figures than were random (nonredundant) metric figures…. Ss will find ways of utilizing only the amount of information called for by a perceptual task. Thus, an identification, classification, and learning tasks they will use only distinctive details of a figure…. Only in the reconstruction task is it necessary to respond to all of the information (all details) in a figure. The results suggest certain interesting differences between the data obtained for phenomeno-logical and performance criteria." (PsycINFO Database Record (c) 2006 APA, all rights reserved)