The O VIII phenomenon describes the occurrence of ultra-high ionization absorption lines of the CNO elements (e.g. O VIII, N VII, C VI, and even Ne X) in the optical spectra hot of DO WDs. This requires temperatures of almost 10 6 K. Since the first discovery of two such objects (Werner et al. 1995) we realized that 50% of all DO WDs are affected , hence it could be that all DO WDs experience this phenomenon at the beginning of the WD cooling sequence. The He II profiles are symmetric, hence

more »

... e formation occurs in hydrostatic layers. However, we cannot uniquely fit the unusually strong profiles with any model atmosphere, preventing precise parameter determinations. The optical high ionization CNO lines have pronounced blue wings, corresponding to outflow velocities of up to several thousand km/s. We stress that none of the objects was detected in the X-ray region. The inferred low X-ray luminosity confirms that we do not see absorption lines from super-hot photospheres, but that the high temperatures are possibly localized along shock fronts in the wind. In the UV range (IUE, HST) no high ionization CNO line is detectable, except possibly O VIII 2978Å (Werner et al. 19972978Å (Werner et al. , 1999. This negative result is still not understood. On the other hand we see two unidentified broad absorption features at 1420Å and 1500Å. The 1500Å feature may be the blue shifted absorption component from a P Cygni profile of C IV 1550Å. This is corroborated by a very weak and broad emission bump near the rest wavelength. The terminal wind speed is about 10 000 km/s and mass-loss could be driven by radiation pressure. The only photospheric line clearly identified is He II 1640Å. The presence of O V 1371Å is uncertain and the C IV resonance doublet could be of interstellar origin.