In this paper we discuss whether the presence of an optic nerve crescent might affect the way in which axial length and corneal curvature interact to determine refractive error. Subjective refraction, keratometry, measurement of body height, axial length of the eye, and stereophotography of the optic nerves were performed on 224 subjects, 8- to 25-years-old. Photographs were examined under magnification; optic nerve crescents, if present, were measured in the horizontal dimension. Those measurements were then corrected for magnification due to the eye and the camera. Logistic regression analysis suggested that male gender and myopic refractive error were most directly associated with the presence of a large crescent, whereas axial length, age, and horizontal keratometry reading were less directly associated with the presence of a crescent. The relation between axial length and refractive error differed among those with large crescents compared to those with small or no crescents. Simple regression showed that, for those with a crescent at least 0.2-mm wide, a 1-mm greater axial length was associated with, on the average, 1.26 D of myopia. For those with smaller or no crescents, a 1-mm greater axial length was associated with only 0.66 D of myopia. This difference was statistically significant at the 0.02 level of confidence.