In a number of SO galaxies, rings of gas, dust and stars are observed to lie roughly perpendicular to the galactic disc (e.g. Whitmore, these proceedings). The material sometimes forms a fairly broad annulus, and the ring appears nearly flat. Simple estimates suggest that differential precession will destroy a polar ring in much less than a Hubble time, implying that the observed structures have been formed only recently. It is then surprising that some of them are so regular, unless something acts to prevent their disruption.