Published online by Cambridge University Press: 12 April 2016
To understand the structure and evolution of the cataclysmic variables, we will need accurate values for their masses, dimensions, mass transfer rates, and other physical properties. Unfortunately, despite an abundance of observational data on these systems, there is a severe dearth of reliable, quantitative information about their fundamental physical properties. Only two cataclysmic variables, U Gem and EM Cyg, are simultaneously eclipsing binaries and double-lined spectroscopic binaries, and only for these two systems can masses and dimensions be determined with a minimum of assumptions (Stover 1981a; Stover, Robinson, and Nather 1981). Even if there were more systems like U Gem and EM Cyg, it is not obvious that our information would be any more reliable, because observers are often unable to agree on the values of the directly measured quantities used to determine physical properties. Thus, the radial velocity curve of the brightest dwarf nova, SS Cyg, has been measured independently 5 times in the last 30 years. The agreement among the measurements is unsatisfactory, and the reasons for the disagreement are not completely understood (Joy 1956; Kiplinger 1979; Stover et al. 1980; Cowley, Crampton, and Hutchings 1980; Walker 1981). The physical properties may still be unreliable when the disagreements are understood and eliminated, because there is considerable uncertainty about the proper way to extract physical properties from observational data. For example, the observed radial velocity curves of cataclysmic variables are believed to be different from the true radial velocity curves of their component stars, but the amount of difference and ways to correct for the difference are unknown.