Simultaneous ASCA and EUVE Observations of Capella

We present simultaneous observations taken in Mar 1996 of the bright stellar coronal source Capella (HD 34029) with the ASCA and EUVE satellites. Previous EUVE observations of Fe emission lines (Fe VIII --- XXIV, excluding XVII) revealed a narrow emission measure feature at 6 x 10(6) K, which has proven to be remarkably stable over several years (flux from Fe XVIII and XIX has not varied by more than 30%), while lines formed at higher temperatures have shown intensity variations up to factors of 4. Furthermore, extremely high signal-to-noise spectra obtained by summing all EUVE measurements show that the Fe/H abundance ratio is consistent with solar photospheric. (See Dupree et al. 1993, ApJ, 418, L41; Brickhouse, Raymond, & Smith 1995, ApJSupp, 97, 551; Brickhouse 1996, IAU Coll. 152, Astrophysics in the Extreme Ultraviolet, Bowyer & Malina, eds (Kluwer), 141.) Meanwhile, the ASCA data of Capella have proven notoriously difficult to analyze. The performance verification (PV) phase data suggested a somewhat subsolar Fe abundance, but models were in poor agreement with the data (chi (2_red) ~ 6). (See Drake 1996, Conf. on Cosmic Abundances, U. Maryland). Since the emission lines observed by EUVE are formed at the same emitting temperatures as the X-ray spectrum (Capella is ``soft'' such that very little flux is observed above 2 keV), the emission measure distribution derived from EUVE lines should provide a direct prediction of the X-ray spectrum, with only the relative abundances of species other than Fe as free parameters. Like the PV data, the new ASCA spectrum is not well fit by any of the standard models. Applying the constraints imposed by EUVE does not make a major improvement in the fit --- multi-thermal, variable abundance models such as Raymond-Smith and MEKAL do not provide any acceptable fit (chi (2_red) > 5). We discuss our efforts to understand the X-ray spectrum, including studies of the uncertainties in the atomic data and of the underlying assumptions of the source models.