Periodicities and regulatory mechanisms in beta Lyrae

We utilize 210 years of beta Lyrae eclipse timings [(O-C)s] and 100 years o f radial velocities to search for eccentricity and apsidal motion, and for periodic variations. No (O-C) periods pass standard significance tests, but we find eight significant radial velocity periods by independent methods o f time series analysis. Waveforms of the eight periodicities are readily se en in phase plots, with most not being simple sinusoids, The causes are not known, but could be orbiting stars. Inspection of the few existing polarim etric observations at two well-separated epochs does not find the radial ve locity periods reproducibly. We looked thoroughly for direct evidence of ec centricity, but without success. Eccentricity that is approximately constan t and as large as 0.010 seems ruled out, but e almost as large as this is p ossible, especially if it is variable on short time-scales. Apsidal motion could be a sensitive indicator of small e, but large gaps in the radial vel ocity record make apsidal motion detection difficult. We searched by three methods, but found only a few doubtful possibilities for apsidal periods, W e also investigated several mechanisms for the previously known 283-d light -curve period by estimating ranges of plausible apsidal and precession peri ods. Apsidal periods are much too long to account for the 283-d variation, and precession is incompatible with existing photometry. Disc pulsation is so far not ruled out. Taking note of the uniquely constant orbital period c hange that has persisted at a consistent +19s per year for 210 yr, we propo se a feedback mechanism to regulate period change via impulsive mass transf er near periastron. The mechanism involves a generalization of the theory o f period changes induced by lobe overflow so as to include impulsive overfl ow, and operates through a (perhaps unobservably) small and variable eccent ricity.