Source: Dix, MR and Hallpike, CS. The Pathology, Symptomatology and Diagnosis of Certain Common Disorders of the Vestibular System. Ann Otol. Rhinol. Laryngol. 6:987-1016, 1952.
"The patient is laid supine upon a couch with his head just over its end. The head is then lowered some 30' below the level of the couch and turned some 30 to 45' to one side. In taking up this position, the patient is first seated upon the couch with the head turned to one side and the gaze fixed upon the examiner's forehead. The examiner then grasps the patient's head firmly between his hands and briskly pushes the pt back into the critical position. The reaction which results calls for some detailed description.
First of all there is nearly always a marked latent period. Sometimes this is as long as five or six seconds. Occasionally it is very short and indeed the reaction may seem to come on at once. This, however, is uncommon. The onset of the nystagmus is nearly always preceded by an appearance of distress. The colour may change; the patient may close his eyes, cry out in alarm and make active efforts to sit up again. At this point it is necessary to reassure the patient and maintain the position of the head. The nystagmus is chiefly rotatory, the direction of the rotation being towards the undermost ear. (Note-In specifying the direction of the rotation reference is made to the displacement of the 12 o'clock point of the corneal circumference.) In addition to the rotatory element there is generally a horizontal component which is gain directed towards the undermost ear. The nystagmus increases in a rapid crescendo in a period which may be as short as 2/3 seconds, or as long as 10 seconds. Thereafter it rapidly declines and the patient's distress is relieved. If the patient is then allowed to sit up, a recurrence of the vertigo in a slighter form is generally noted, and if the eyes are examine at this point nystagmus can be seen, the direction of which is, on the whole, reversed. If this is allowed to disappear and the critical supine position is again assumed, the nystagmus again makes its appearance but generally in slighter form and disappears more rapidly than before. After two or three repetitions of this test it is generally found that the reaction has been eliminated altogether and cannot be elicited except, as Barany pointed out, after a period of rest."