A number of skill researchers have suggested that the time period following knowledge of results may be one of verbal-cognitive activity by S. This is opportunity for S to organize -or recode and possibly rehearse components -of the KR presentation interval and, prior to major decisionmaking and response-selection processes, S's in formation base may differ from the KR presented. The present study was concerned with determining short term retention characteristics .of numerical information, particularly regard to varnables which might be operative within the KR presentation interval. Specifically, it .attempted to examine the verbal and motoric variables and nature of the input information .and their interactions within an information-processing context. Ss of the study were voluntarily participated 64 undergraduates with equal number of males .and females. They were randomly assigned to the treatment conditions. The device used was .a modification of the Selective Mathometer having 10 plastic switch response keys. A .single trial consisted of 10-sec. stimulus presentation, wherein each of the 10 keys was illuminated, 17-sec. immediate recall period and other 3-sec. of preparation. Each S performed 4 trials. The design was 24 factorial. Four variables such as digit input, display arrangement of key, input mode and output mode, each having 2 levels, were examined. Former 2 were between S and the latter 2 were within S variables. The findings could be summarized as follows: (1) A short-term loss of about 30~40% was shown, which was indicative of rapid short-term motor forgetting similar to that of verbal memory. (2) When digit input was presented systematically, the mean recall value was significantly superior to that of random material. Better recall of that which approximated to English or ascending/descending digit pairs indicated R as functions of such as pre-experimental stereotype. (3) When input modewas visual, better recall was found in comparing with the case of visual-motor mode. This finding was interpreted to negate the feeling in the literature that the more feedback channels the better . (4) The effectiveness of KR lay in the memory storage phase not in the recall processes. (5) The interaction effect of display arrangement of key X digit input was statistically significant. This could be suggestive of progressive simplication which could mean many things such as decreasing advantage of the pre-experimental habits or unique abilities for each task.