Classical novae as a probe of the cataclysmic variable population

Classical novae ( CNe) are the brightest manifestation of mass transfer onto a white dwarf (WD) in a cataclysmic variable ( CV). As such, they are probes of the mass transfer rate, M, and WD mass, M-WD, in these interacting binaries. Our calculations of the dependence of the CN ignition mass, M-ign, on M and M-WD yields the recurrence times of these explosions. We show that the observed CNe orbital period distribution is consistent with the interrupted magnetic braking evolutionary scenario, in which at orbital periods P-orb > 3 hr mass transfer is driven by angular momentum loss via a wind from the companion star, and at P-orb < 3 hr by gravitational radiation. About 50% of CNe occur in binaries accreting at M similar or equal to 10(-9) M-circle dot yr(-1) with P-orb = 3 - 4 hr, with the remaining 50% split evenly between P-orb longer (higher M) and shorter (lower M) than this. This resolution of the relative contribution to the CN rate from different CVs tells us that 3(9) x 10(5) CVs with WD mass 1.0(0.6) M-circle dot are needed to produce one CN per year. In addition, one CN per year requires a CV birthrate of 1(2) x 10(-4) yr(-1) and likely ejects mass into the interstellar medium at a rate of M = 3(9); 10(-5) M-circle dot yr(-1). Using the K- band - specific CN rate measured in external galaxies, we find a CV birthrate of 2(4); 10(-4) yr(-1) per 10(10) L-circle dot,L- (K), very similar to the luminosity specific Type Ia supernova rate in elliptical galaxies. Likewise, we predict that there should be 60 - 180 CVs for every 10(6) L-circle dot, (K) in an old stellar population. The population of X-ray-identified CVs in the globular cluster 47 Tuc is similar to this number, showing no overabundance relative to the field. The observed CN Porb distribution also contains evidence for a CV population that has no period gap. These are likely systems with a strongly magnetic WD(polars) in which it has been suggested that the field interferes with the wind of the companion, limiting angular momentum losses to those of gravitational radiation and eliminating the period gap. With this reduced. M, polars evolve more slowly than systems that undergo magnetic braking. Using a two-component steady state model of CV evolution, we show that the fraction of CVs that are magnetic (22%) implies a birthrate of 8% relative to nonmagnetic CVs, similar to the fraction of strongly magnetic field WDs.