Elementary geometry will be supposed to have its usual meaning as euclidean metrical real geometry of the plane. I propose to point out that, while “points at infinity” and “the line at infinity” may be properly introduced in conformity with the axioms, the “circular points” cannot be so introduced. They do not belong to this geometry; and those results which do properly belong to it, and are often made apparently to depend on the existence of the circular points, should therefore be expressed with no mention of these points.