In our previous study using an improved elevated plus-maze in mice, the oriental herbal medicine Saiboku-to prolonged the time spent in open arms, showing an anxiolytic effect, and the effect was mainly caused by honokiol derived from magnolia. This study was carried out to compare the anxiolytic potentials of honokiol and water extracts of three magnolia samples; two being Kara-koboku (Magnolia officinalis) (KA: from Zhejiang-sheng, China; honokiol 0.25% and magnolol 1.16%, and KB: from Sichuan-sheng, China; honokiol 1.72% and magnolol 1.71%), and one being Wa-koboku (Magnolia obovata) (WA: from Iwate-ken, Japan; honokiol 0.32% and magnolol 0.81%). Seven daily treatments with 0.1-1 mg/kg honokiol, but not 0.2 and 1 mg/kg magnolol, revealed an anxiolytic effect with the peak potential at 0. 2 mg/kg. The anxiolytic potentials of 40 and 80 mg/kg KA, which contained the highest amount of magnolol, were almost equivalent to those of 0.1 and 0.2 mg/kg honokiol, respectively. KB, at 11.6 mg/kg, and 62.5 mg/kg WA resulted in almost the same anxiolytic potential as that of 0.2 mg/kg honokiol. No significant change in the ambulatory activity was produced by any drug treatment. These results suggest that honokiol is the chemical responsible for the anxiolytic effect of the water extract of magnolia and that the other chemicals including magnolol in magnolia scarcely influence the effect of honokiol. It is also considered that the elevated plus-maze test is applicable for evaluation of the content of honokiol in magnolia.