The H_alpha absorption line

The broad H_alpha absorption line is expected to originate on the white dwarf. In the presence of the numerous absorption features of the secondary star, and in view of the emission component, it is very difficult to measure its radial velocity reliably. Attempts to fit a Gaussian profile or to calculate cross-correlation functions (masking the emission component in both cases) yielded velocity variations in phase with the secondary, albeit with a much smaller (and different for the two attempts!) amplitude. This reflects the strong influence of the late type absorption features.

To alleviate this problem it was attempted to subtract the spectrum of the red dwarf. To this end the spectrum of the M5 V standard star Yale 1755 ([Jacoby et al. 1984]), the closest available match to the RR Cae secondary spectrum (see Paper I), was ``diluted'' by a black body spectrum with a temperature of 7005 K (the temperature found for the white dwarf in RR Cae by [Bragaglia et al. 1995]) and subtracted from the RR Cae spectrum. The best result, i.e. that with the least amount of residual absorption features, is achieved assuming a veiling factor (defined as the fractional contribution of the primary to the total light) of 0.6 at 6700 14#14 Å.

Measuring the radial velocity of the residual H_alpha absorption again, using cross-correlation techniques, now resulted in a radial velocity curve shifted by about 180 degrees in phase with respect to the secondary. However, Yale 1755 being far from a perfect match to the red dwarf in RR Cae (which has a later spectral type; see Paper I) even the difference spectrum contains many residual absorption features which severely distort the H_alpha absorption profile. These doubtlessly influence the measurements considerably, making the measured values of the radial velocities unreliable. Therefore, it can only be confirmed that the H_alpha absorption component moves in counter-phase to the red dwarf, as expected.