In the present study, which is based on Loeb's 1986 analysis, we used five levels of noise that varied in intermittency and meaningfulness, crossed with sex of subject and time of day. Memory tasks that differed in their reliance upon long-term, short-term free recall, and sequential short-term memory were used as dependent variables. A total of 160 subjects, 20 per block, participated in the 2 x 2 x 5 (Sexes x Time of Day x Noise Conditions) experimental design. Results support the prediction that white noise enhances performance on tasks with sequential short-term memory demands (anagrams: p < .05; random letter generation: p < .002). We found complex interactions by sex of subject, time of day, and type of noise for those tasks that placed a heavy demand on short-term working memory, i.e., complex sorting (Noise x Sex, p < .05) and random letter generation (Sex x Noise x Time of Day, p < .05). The predicted effects for anagrams were not supported. These results call into question previous generalizations about the effects of noise on performance (Broadbent, 1978; Dornic, Sarnelid, Larson, Svensson, & Fernaeus, 1982; Poulton, 1977), alternative interpretations are presented.