LATE EVOLUTION OF CATACLYSMIC VARIABLES

We study the evolution of hydrogen-rich cataclysmic variables (CVs) ne ar minimum orbital period at similar to 78 minutes. As has been known for many years, these are among the most intrinsically common CVs, but they hide fairly well because of their faintness and low incidence of eruptions. We discuss their number and observational signatures, payi ng special attention to those that may have passed minimum orbital per iod-the ''period bouncers.'' The status of binaries near minimum perio d is best determined by the mass ratio, and this is best constrained b y measuring the accretion disk precession frequency, because that freq uency is readily accessible to observation and proportional to the sec ondary star's mass. This method reveals four stars that are good candi dates to have survived period bounce; two appear to have secondaries a s puny as 0.02 M-.. But each star can have bounced only recently if at all. There is still no strong evidence of any long era of evolution i n a state of increasing period. This conflicts sharply with discussion s of observational data that have identified dozens of known CVs with this state. The total space density of cataclysmic variables is simila r to 10(-5) pc(-3), with short-period systems constituting similar to 75% of the total. Both estimates are far less than predicted by simple theories of evolution. It is probably necessary to have some means of destroying CVs before they reach the predicted very high space densit ies. This can be done by invoking an angular momentum loss mechanism t hat does not quickly subside as the mass ratio becomes very low.