5 Discussion

It is quite evident that any rise of the S/N ratio allows to detect new unidentified interstellar features. Apparently the absorpion spectra of interstellar clouds are very rich - they may contain hundreds of very weak features. They are generally weak and shallow, most of them, observed in averaged spectra ("sigma'' and "zeta'') and in the spectrum of HD 210839, are considered as certain. Some of the features listed in Jenniskens & Désert ([1994]) catalogue are not confirmed in this survey and seen in Table 4. The latter catalogue is based on spectra of only four stars of high reddening. Moreover it involves HD 183143 and HD 21389 - relatively cool stars which spectra may contain many weak stellar lines, unidentified yet. These were possibly listed also as interstellar features. Some of the weak diffuse bands from the Jenniskens and Désert survey differ in wavelength from the features listed in $\Delta \lambda$ = 0.3 Å ($\lambda$6284). This may be due to the fact that the survey of Jenniskens and Désert is based on heavily reddened stars towards which the interstellar spectra are formed in many clouds. In such a case it is difficult to decide which of the components of the sodium or calcium lines share the radial velocity with the unidentified features. The wavelengths, listed in Table 2 and Table 3 are corrected to the rest scale using the radial velocities based on Doppler shifts of narrow and symmetrical D₁ and D₂ lines. The line of the interstellar Lithium doublet at 6707.9 Å was seen only in some of our "zeta'' type targets with higher reddening.

   

Table 4: Features from Jenniskens & Désert catalogue not listed in our survey (regions 6295-6350 Å and 6563-6595 Å are excluded)

DIB	Remarks
5747.81	stellar
5853.95
6045.27	weak and shallow
6177.27	very shallow
6207.83
6215.71
6236.58
6278.89	noisy in our spectra
6280.52	noisy in our spectra
6281.07	noisy in our spectra
6377.87
6383.04	stellar
6451.60
6494.17
6494.91
6632.10	shallow

It is, however, quite evident in our average "sigma'' and "zeta'' objects. It seems to be correlated with narrow DIBs as well as K I and Ca I lines (Kreowski et al. [1998]). The fact that our averaging procedure makes this very weak line visible confirms its efficiency in finding weak interstellar features and the fact that the single components of interstellar sodium lines are Doppler-shifted identically to other interstellar lines such as Li I.

Table 4 contains the features from Jenniskens & Désert ([1994]) catalogue not listed in this survey excluding the region from 6295 to 6350 Å. In echelle spectra broad and shallow features, such as $\lambda$5705 and $\lambda$5844 are clearly seen, but generally spectra acquired with the aid of such spectrometers are not proper tools to detect broad spectral features. We cannot thus guarantee that all possible broad and shallow DIBs are listed in our survey. After excluding the region from 6295 to 6350 Å and the region in the vicinity of H$_{\alpha}$ line we found a total of 63 new certain features in the spectral range from 5650 to 6865 Å. Most of them are weak and narrow ones with profiles resembling those observed in the vicinity of $\lambda$5780 (DIBs at 5766.15, 5772.53 and 5785.11 Å). Their strength ratios are typically different in both kinds of clouds; very weak ones are generally stronger in "zeta'' clouds (e.g. new features shortward of $\lambda$6196). It would be important to divide these weak features into "families'' characterized by constant strength ratios to stronger ones. Such "families'' are very likely to be spectra of single carriers. Having them listed we should be able to identify the carriers. Precise measurements of rest wavelengths based on Doppler shifts of interstellar sodium lines (perspectively also lines of other interstellar atomic gases) will be helpful while trying to match the spectra acquired in laboratory experiments.

It is to be emphasized that our wavelength estimates are more reliable than those of Jenniskens & Désert ([1994]) or Ehrenfreund et al. ([1997]) - both papers based on heavily reddened objects. Such targets are useful while detecting the weak DIBs but not while determining their rest wavelengths. Observing a star through several clouds we do not know which DIB originates in which cloud (they can be either "sigma'' or "zeta''). Moreover their profiles can be Doppler-splitted. This is why we used averaged "sigma'' or "zeta'' type spectra in this survey.

Acknowledgements

The authors wish to express their gratitude to the staff members of the McDonald Observatory where the spectra have been acquired. The paper was supported financially by the II US-Poland Maria Skodowska-Curie Joint Fund under the grant MEN/NSF-94-196 and by the Polish State Committee for Scientific Research under the grant 2.PO3D.008.16.