The old nova DQ Herculis (N1934) is an eclipsing binary star with an orbital period of 4h 39m consisting of a cool star with a mass near 0.4 Msun, presumed to be on the main sequence, and a white dwarf with a mass near 0.6 Msun (Horne et al 1993, ApJ, 410, 357). The cool star is transferring mass from the inner Lagarangian point through an accretion disk onto the white dwarf. The light curve of DQ Her is notable for having a sinusoidal modulation with a period of 71.07 seconds and an amplitude near 1%. The rate of change of the period is small but not zero, Pdot = -6.4 x 10^(-13)s/s (Patterson et al 1978, ApJ, 224, 570).

In the most widely, if not universally accepted model for the oscillation, the white dwarf is magnetized mu ~ 3 x 10^32 G cm^3 and rotating. Gas from the inner edge of the accretion disk threads onto the magnetic field lines and falls onto the white dwarf at the magnetic poles, where the release of kinetic energy forms luminous spots. The energy from the spots is reprocessed in the accretion disk, producing a spoke-like pattern rotating around the disk, which is the source of the observed modulation. In the most recent version of this model there are two spots, two spokes, and the rotation period of the white dwarf is 142 seconds (Zhang et al 1995, ApJ, 454, 447). Although the two spokes must have nearly the same luminosity to produce a sinusoidal modulation at 71 seconds, it is unlikely that they have precisely the same luminosity and, therefore, this model predicts that the light curve should also be weakly modulated at a period of 142 seconds. The 142-second modulation has never been observed and the upper limit to its amplitude is 6% of the modulation amplitude at 71 seconds (Kiplinger & Nather 1975, Nature, 255, 125).

Although this model is widely accepted (indeed, DQ~Her has donated its name to the class of cataclysmic variables containing rapidly-rotating, accreting, magnetic white dwarfs) the two crucial signatures of magnetic accretion -- pulsed X-ray emission and circular polarization -- are both absent. It is, therefore, worth considering alternative models based on stellar pulsations. To have a long enough period and to form the rotating, two-spoke pattern observed in DQ~Her, any such pulsation would most likely be a non-radial g-mode pulsation with l = 2 and m = +/- 2. Theoretical periods for this mode in realistic models of 0.6 Msun white dwarfs can be startlingly close to 71 seconds (e.g. P = 69 seconds in Table 5 of Brassard et al 1992, ApJSup, 81, 747). Furthermore, the periods of white dwarf pulsations are just as stable as the period of the 71-second modulation in DQ~Her (e.g., Pdot < 10^(-14)s/s for G117-B15A [Kepler 1993, Baltic Astronomy, 2, 444]). This quantitative agreement with the observations demands respect. The pulsation model is testable. With no known exceptions, every pulsating white dwarf pulsates in two or more modes simultaneously, and the ratio of the amplitudes of the strongest to the next strongest pulsation is always greater than ~5%. If the 71-second modulation is a non-radial pulsation, other pulsation modes must be present in DQ Her.

Thus, magnetic accretion models and pulsation models both predict additional periodicities in the light curve of DQ Her, the subharmonic period at 142 seconds if the magnetic accretion models are correct, a multiplicity of periods consistent with a pulsation spectrum if the pulsation models are correct. We propose, therefore, to search for these additional periodicities in the light curve of DQ~Her. If the 142-second subharmonic is present, the magnetic accretion models must be correct and the rotation period of the white dwarf must be 142 seconds. If periods consistent with white dwarf pulsations are present, the pulsation models must be correct, with an additional payoff that the standard tools of asteroseismology can be applied to deduce the structure of the white dwarf.

The Whole Earth Telescope (WET) network is the ideal tool for this search. For telescopes with moderate-sized apertures, the dominant source of noise in the power spectrum of white-light light curves of DQ Her is the intrinsic flickering in the light curve. The detection limit for periodic modulations in the presence of this noise improves linearly with time, not as the square root of time. The WET network consists of many telescopes spread around the Earth and produces nearly continuous light curves many days long, so it takes full advantage of this linear improvement with time. Based on previous measurements of the power spectrum of DQ Her and similar cataclysmic variables, we will achieve a detection limit of delF/F < 1.0 x 10^(-4) in ~ 100 hours of continuous WET data. Since the present amplitude of the 71 second modulation is delB/B = 0.012 +/- 0.001 (Zhang et al 1995, ApJ, 454, 447), we will detect any modulation with an amplitude greater than 1% of the amplitude of the 71-second modulation.

RA=18 06 05.28, DEC=45 51 02.2, EPOCH=1950, Mv=14.2