This study concerned the hypothesis presented by Lindsley in 1961 that human sensorimotor performance should vary with the phase of the alpha cycle. Although there have been a number of studies which have supported this hypothesis, there has been no work from a modality other than the visual modality. Since eye tremor is correlated in phase and frequency with the alpha rhythm, these visual results might be explained by the peripheral eye tremor and not necessarily by the central alpha rhythm. The present study measured human auditory signal-detection performance at four different phases of the temporally (T5-linked mastoids) measured alpha rhythm. These four different phases were defined on an a priori basis by a computer algorithm. Detection performance was significantly better at the negative peak of this alpha cycle than at the positive peak, but there was no significant difference between the positively and negatively going zero cross performances. These results are consistent with the Lindsley hypothesis.