This research study follows a previous paper on the performance of adult aphasics tested with a modified form of the Wisconsin Card Sorting Task. It investigates the effects of varying the order of presentation of the three sorting concepts or problems on the overall task-learning rate, the percentage of error responses and the percentage of perseverative error. It is found that varying the order of the concepts does not effect the task difficulty, as measured by these indices. Statistical and interpretative analyses of the data also investigate the nature of performance improvement patterns. It is found that any significant improvement is due to the elimination of nonperseverative errors, while the proportion of perseverative responses remains fairly constant. The aphasics' inability to generalize problem solution methods and possible techniques for reducing perseveration, in addition to an evaluation of this task as an aphasia research instrument, are discussed.