Though in my former papers on knots I have made but little allusion to cases in which two or more closed curves are linked together, the method I have employed is easily and directly applicable to them. I stated to the British Association that the number of intersections passed through in going continuously along a curve, from any intersection to the same again, is always even—whether it be linked with other curves or not Hence, even when a number of closed curves are linked together, the intersections may be so arranged as to be alternately over and under along each of the curves.