Studied the short-term memory for length in a paired-comparisons (AX) paradigm. The 6 retention intervals (1, 2, 3, 5, 9, and 15 sec.) and the 19 possible lengths of the standard stimulus were randomly varied within blocks. Besides equality, there were 10 possible length differences between the standard and comparison stimuli, and these differences were presented equally often within blocks except that the probability of a difference was equal to that of no difference. 2 well-trained undergraduates were asked to make a same-different recognition judgment in Part 1 and both a same-different and a longer-shorter judgment in Part 2. There was a significant decline in performance over unfilled retention intervals in both parts, but superimposed on this was a tendency for recognition accuracy to increase from 2-sec to 5-sec delay. The performances for the 2 types of decisions in Part 2 were not closely related. Thus, evidence was found for both 2 decay processes and for 2 decision processes. (PsycINFO Database Record (c) 2006 APA, all rights reserved)