Summary: Existing descriptions of the biaxial ray surface are, in general, misleading. A more accurate description, stressing the simplicity of the surface, is proposed and this is supported by numerous sections drawn through or near the singular points.