5 Description of the atlas

The spectra presented in this atlas are composites of individual exposures, most of them on the KBr photocathode. The separation wavelengths are indicated by broken vertical lines at the bottom. Although we have used the onboard solar rotation tracker for all spectra presented here, there are clear indications of evolution in the network structures observed. This has to be taken into account, if radiances of lines from different exposures are compared. The variability of the solar atmosphere on all temporal and spatial scales is amazing. These include periodic phenomena (cf., e.g., Curdt & Heinzel 1998 and references therein) and explosive phenomena (cf., e.g., Innes et al. 1997 and references therein). To give an example, the emission measure of some lines increased by a factor of $\approx$5 during the sunspot observation as derived from a comparison of lines recorded in both orders of diffraction (i.e., at the beginning of the sequence and at the end of the sequence).

The spectra are displayed as profiles in different colours, the quiet-Sun spectrum as black line, the coronal-hole spectrum as blue line, and the sunspot spectrum as red line. In the sunspot profile the continuum background has been moderately noise-filtered. The BN/CI ratio of the quiet-Sun spectrum is overlayed as green, solid line using a separate axis scale. In Fig. 4 the spectrum is presented in portions of 42 Å per page with 2 Å of overlap. The atlas is also available in machine readable form via ftp or the internet This atlas represents our most up-to-date knowledge of the solar-disk spectrum in the given spectral range. Coronal and flare spectra have also been analyzed (Curdt et al. 2000b; Feldman et al. 2000), but are not the subject of this communication.

5.1 Line identification

Each resolved emission line is indicated by a mark, the measured wavelength in angstrom (Å), and the identification, if available. For the sake of a concise presentation, the wavelength values are restricted to three digits, one digit before and two rounded digits after the decimal point (e.g., the Ne VIII line observed at 770.423 Å reads 0.42 in first order and 0.85 in second order). The mark can be used as a cross-reference to line lists available in the literature, where additional spectroscopic information is found. Squares and circles point to line lists published on SUMER observations, circles to Curdt et al. (1997) and squares to Feldman et al. (1997), respectively. Diamonds point to the older lists of Kelly (1987), Cohen et al. (1978), Sandlin et al. (1986), and Bartoe et al. (1979). We have checked the line lists in this order and emphasize that marks are not meant in a sense of ``first identification by'', but, on the contrary, in general are pointing to the most recent list. Triangles denote lines observed for the first time or new identifications, these are mainly found in the sunspot spectrum. Open marks represent lines which are observed in second order of diffraction. Questionable identifications are marked by a question mark (?). Only 4% of the lines in the quiet-Sun spectrum remain unidentified at this stage, most of them being faint. This percentage increases to 12%, if we also consider the sunspot spectrum, which seems to be dominated by 3- to 6-fold ionized species and is more difficult to interpret. In these cases we have given the temperature classification of Feldman et al. (1997), if available (cf. Fig. 4). A list of all lines observed in the SUMER spectral range is given electronically in the Annex.

5.2 Bright network/cell interior ratio

In the quiet-Sun spectrum, the SUMER slit cuts five partly sub-structured supergranular cells. We have separated 21 pixels representative of bright network and 71 pixels of cell interior, respectively. The BN/CI ratio, thus derived from the same exposures, is $\ge$1 at all wavelengths. It shows an interesting variation, depending on the particular emission line under consideration. This feature is a direct consequence of the plasma temperature and density of the emitting source. For lines with high formation temperature, the BN/CI ratios are not far from 1, while they can reach values $>\!10$ for transition region lines.