We compared the results of analysis of vestibulo-ocular reflex (VOR) obtained by manual rotation in routine vestibular clinical practice with that of caloric testing, and examined the validity and limitations of VOR analysis as a test for the estimation of peripheral vestibular function and imbalance. VOR response was recorded in daily vestibular clinical examinations by manually rotating the standard clinical chair for , 30 s. VOR gain was slightly, but significantly, correlated with the peak slow phase velocity of caloric response ( r = 0.50, p < 0.001). However, 8 out of 12 patients with no caloric response failed to exceed the range of two standard deviations of the mean value of age-matched normal subjects, indicating that it is clinically difficult to use VOR gain alone as a estimate of unilateral vestibular function. VOR directional preponderance (VOR DP%) correlated well with caloric canal paresis (CP) (CP%; r = 0.89, p < 0.001). VOR DP% was within the normal range in patients with caloric CP% < 40 and exceeded the normal range in most cases with caloric CP% > 80. VOR DP% varied widely when caloric CP% ranged between 40 and 80. The effect of vestibular compensation on VOR DP% was examined by plotting VOR DP% divided by caloric CP% (DP CP) against the number of days since the onset of vertigo in patients with vestibular neuritis or sudden deafness with vertigo. DP CP was large within 50 days of the onset of vestibular damage, especially when caloric CP% was < 80, and gradually decreased with time. These results indicate that determination of VOR DP% should contribute to the early diagnosis of fresh vestibular imbalance, especially in daily clinical practice, because this type of VOR recording can be performed in < 1 min in routine vestibular clinics. The decay time constant of DP CP was larger when caloric CP% exceeded 80, indicating that vestibular compensation proceeds more slowly when the vestibular damage is severe.