This study was planned to to compare the hearing level of Kendo-Players with that of non-Kendo-Players to make clear whether or not long term practice of Kendo causes hearing loss. Pure-tone audiometry by air conduction was applied to 172 Kendo-Players and 76 non-Kendo-Players with an audiometer (Audiometer AA-69, Audiogram Recorder RE-05, Soundproof Room AT-45, Rion, Japan). Among them, the following cases were excluded from the analysis. 1) Cases who have had medication against tuberculosis. 2) Cases with perforation in the tympanic membranes. 3) Cases with occupational experience of more than a year in noise nuisance. 4) Cases with experience of head phone listening for more than one hour a day, twice a week for at least one year. The thresholds of hearing between 149 Kendo-Players (cases, age: 19-76) and 44 non Kendo-Players (controls' age: 20_78) were analysed to compare mean hearing thresholds between the age groups, and to examine covariance adjusted with age. The dose-response relationship was examined by the partial correlation coefficients controlled by age between the thresholds of hearing and the length of Kendo experience (year) or the total experience hours of Kendo (hour). The rusults showed that the hearing loss of Kendo-players was greater than non-Kendo players, especially at the frequency of 4000 and 8000 Hz in the age group of 30-59 years old. It is supposed that long term practice of Kendo may cause hearing loss. Further study, however, will be needed to re-examine the dose-response relactionship with more carefully selected samples for the epidemiological survey in the future.