2020MNRAS.492.3343S

I report on two new measures of the sudden change in the orbital period (P) across the nova eruption ([?]P) and the steady period change in quiescence ({dot}P) for classical novae (CNe) RR Pic and HR Del, bringing a total of six such measures for CNe, all in a final report of my large and long observing program. The fractional changes ([?]P/P) in parts-per-million (ppm) are -290.71 ± 0.28 (QZ Aur), -472.1 ± 4.8 (HR Del), -4.46 ± 0.03 (DQ Her), +39.6 ± 0.5 (BT Mon), -2003.7 ± 0.9 (RR Pic), and -273 ± 61 (V1017 Sgr). These results are in stark opposition to the Hibernation Model for the evolution of cataclysmic variables (CVs), which requires [?]P/P> + 1000 ppm to get the required drop in the accretion rate to produce hibernation. The hibernation model cannot be salvaged in any way. My program has also measured the first long-term {dot}P for CNe, with -2.84 ± 0.22 (QZ Aur), +4.0 ± 0.9 (HR Del), +0.00 ± 0.02 (DQ Her), -2.3 ± 0.1 (BT Mon), and +1.25 ± 0.01 (RR Pic) in units of 10-11 d per cycle. These can be directly compared to the predictions of the magnetic braking model, where the long-term average {dot}P is a single universal function of P. The measured values are +5.3, -94, 0.00, +6.9, and -190 times that predicted by the model, so the predictions are always greatly wrong. Further, the effects of the [?]P averaged over the eruption cycle are usually much larger than the magnetic braking effects. To get a realistic model of CV evolution, we must add the physics of the [?]P and {dot}P variations.