We investigated otolith function using the subjective visual vertical (SVV) in 54 patients with persistent geotropic positional nystagmus (Light cupula), 30 patients with persistent apogeotropic positional nystagmus (Heavy cupula) and 45 healthy subjects. To compare benign paroxysmal positional vertigo (BPPV), SVV was also measured in 56 patients with posterior canal canalolithiasis (PC) and 38 patients with lateral canal canalolithiasis (LC). Patients attended the clinic within 1 week after the onset of vertigo, and the SVVs were examined prior to treatment. The affected sides of Light cupula and Heavy cupula were estimated based on the neutral position at which the horizontal nystagmus direction was reversed. The average SVV toward the affected side showed significantly higher values, in this order: Light cupula>LC/PC>healthy subjects and Light cupula>Heavy cupula. Otolith dysfunction could be demonstrated on the affected side in Light cupula, whereas Heavy cupula exhibited a lesser degree of otolith dysfunction. The average SVV value toward the affected side of 0.9 is recommended as a cut-off value for Light cupula (sensitivity: 0.75, specificity: 0.86). The average SVV value>1.4 toward the affected side is also recommended as a cut-off for the detection of otolith dysfunction in Light cupula, PC and LC (specificity>0.93).