In a previous article (J. Frohlich, M. Morgan, S. Ogawa, L. Burton, and D. Pfaff, 2001, Horm. Behav. 39, 39-47) an experiment to explore the structure of behavioral arousal in female mice was described. The present study extends this, to investigate the roles of thyroid hormone and estradiol in altering the statistical structure of arousal measures. Each of four groups of ovariectomized female mice was administered either thyroxine (T4), estradiol benzoate (EB), both (T4 + EB), or neither (control). They were then subjected to the same rigid protocol of tests bearing on arousal concepts used in our previous study. T4-treated mice manifested significantly increased freezing behavior relative to control mice in a fear-conditioning paradigm. When compared with EB mice, T4-treated mice evinced significantly increased acoustic startle and open-field behavior. T4 mice were also significantly more active in the open field than EB + T4-treated mice. Mice administered EB demonstrated significantly decreased acoustic startle and open-field performance than controls. Evidence for increased anxiety in the open-field test was obtained in the EB condition. Factor and cluster analysis indicated the statistical structure of arousal measures to be reasonably robust across hormonal conditions. Hormone effects on arousal components are of interest because of their likely contributions to emotional states and cognitive performance.