<p>In the diagnosis of horizontal canal benign paroxysmal positional vertigo, it is challenging to lateralize the affected side. The “bow and lean test,” and several methods built upon it, have been proposed. Many of these techniques are based on a model in which the horizontal canal’s putative anatomy involves an anteromedial segment. The actual anatomy has a posteromedial segment. We show that predictions of a model based on the hypothetical anatomy differ from those based on the true anatomy. We further show that the true anatomical model predicts that for a roll maneuver that begins in the supine position, going from the side-lying position to the prone position should generate different patterns of nystagmus depending on which side is affected; this should distinguish which side is affected, and thus has implications for treatment. We conclude with a data visualization method that depicts these findings compactly.</p>