Constraints on EUV Spectral Diagnostics of Capella

The EUVE spectra of Capella are dominated by lines from highly ionized Iron (Dupree et al./ 1993, ApJ, 418, L41). For Fe IX -- XXIV, we compute line intensities for comparison with observations, by integrating the new emissivities of Brickhouse, Raymond & Smith (1994, preprint) over model emission measure distributions. We are interested in determining not only the overall reliability of the emission measure distribution model, but also confidence limits on its characteristic features, such as the minimum at 10(6) K and the bump at 10(6.7) K. Comparisons of models we have constructed for the SERTS solar active region spectrum (Thomas & Neupert 1994, ApJSupp, 91, 461) are useful for understanding some important issues in the analysis of high quality spectra, such as: the contribution to line emission at temperatures of large emission measure; the complexity introduced by density-sensitive lines; and the importance of testing for weak lines. We estimate the uncertainties in the individual line intensities, based on observational errors (statistical and calibration), atomic rate uncertainties, and, for density-sensitive lines, the range of possible intensities. These uncertainties then determine the chi (2) to be minimized by the fitting technique. Furthermore, we are interested in the effects of uncertainties on diagnostics for electron density, optical depth, abundances, and other parameters. For example, line ratio diagnostics for the electron density are highly non-linear functions of the density. We discuss the spread in densities derived from Fe XXI line ratios observed in Capella (Dupree et al./, this BAAS). Accuracies of other diagnostics relevant to the EUVE observations are discussed. This work was performed under NASA Grants NAGW-528 and NAG5-2330 to Smithsonian Astrophysical Observatory.