It was hypothesized that when the habit strength was kept constant and drive state reduced, the response tendency would be the function of intelligence. In this experiment, the habit strength as a controlled variable was kept equal by the subjects' reasoning with each of four figure-moods without error as the criterion, and the independent variables-anxiety and intelligence-were measured with the MAS and the Revised AGCT, respectively. Anxiety as an emotional reaction was aroused under the instruction that the result of valid syllogistic arguments would be used as a midterm's evaluation; however, it was relieved by permitting the subjects to consult the text during the course of reasoning. 　　It was found that the valid syllogistic arguments made were no significant interaction between anxiety and intelligence, and no significant differences between the high anxiety and the low anxiety subjects with the same range of intelligence. But there were highly significant differences between the high intelligence and the middle range of intelligence subjects, and between the high intelligence and the low intelligence subjects, irrespective of their anxiety levels. As predicted, in the condition of anxiety relieving, the deductive reasoning was determined by the subjects' intelligence when the habit strength was under control.