Superhumps in Cataclysmic Binaries. XXV. q_crit, epsilon(q), and Mass-Radius

We report on successes and failures in searching for positive superhumps in cataclysmic variables, and show the superhumping fraction as a function of orbital period. Basically, all short-period systems do, all long-period systems don't, and a 50% success rate is found at P_orb=3.1+-0.2 hr. We can use this to measure the critical mass ratio for the creation of superhumps. With a mass-radius relation appropriate for cataclysmic variables, and an assumed mean white-dwarf mass of 0.75 M_sol, we find a mass ratio q_crit=0.35+-0.02. We also report superhump studies of several stars of independently known mass ratio: OU Virginis, XZ Eridani, UU Aquarii, and KV UMa (= XTE J1118+480). The latter two are of special interest, because they represent the most extreme mass ratios for which accurate superhump measurements have been made. We use these to improve the epsilon(q) calibration, by which we can infer the elusive q from the easy-to-measure epsilon (the fractional period excess of P_superhump over P_orb). This relation allows mass and radius estimates for the secondary star in any CV showing superhumps. The consequent mass-radius law shows an apparent discontinuity in radius near 0.2 M_sol, as predicted by the disrupted magnetic braking model for the 2.1-2.7 hour period gap. This is effectively the "empirical main sequence" for CV secondaries.