The variation in the corneal-retinal potential (CRP) was measured under clinical electronystagmography conditions. Twenty subjects were recorded for 39 minutes each. The interindividual differences are exceedingly large. The average fall during the first 9 to 12 minutes was about 40%. This error leads to an underestimation of the speed of the slow component of the nystagmus, and hence to false conclusions about directional preponderance, excitability difference, and response decline. The latter is probably largely an artifact of the falling CRP. Most standard series suffer from this basic error. It is recommended that calibration be performed at the start and finish of each caloric irrigation to lessen this variable.