1 Introduction

The "stationary'' spectral features at 5780 Å and 5797 Å, observed by Heger ([1922]) at the Lick Observatory are the first discovered members of the set of Diffuse Interstellar Bands (DIBs). The interstellar origin of several strong DIBs has been proven by Merrill ([1934]), Beals & Blanchet ([1937]), York ([1971]) and Herbig ([1975]). The latter work based on photographic spectra of the S/N ratio enhanced by means of averaging, pointed the possible existence of weak 5784 and 5795 (near 5797) bands. These weak DIBs as well as some other similiar features may be related to some of the strong ones as pointed by Chlewicki et al. ([1987]) and are more evident in "zeta'' targets - i.e. those in which the strength ratio of 5797 and 5780 is relatively high (Kreowski & Westerlund [1988]). During recent years the number of known DIBs keeps growing. The paper of Smith et al. ([1981]) suggested that the feature near 6194.5 Å is a very weak DIB, well seen in spectra of dark interstellar clouds. Generally many very weak features were successfully detected by Ferlet et al. ([1983]), Herbig ([1988]) and by Herbig & Leka ([1991]) in the red and near infrared spectral ranges.

The survey of Kreowski & Sneden ([1993]) proved the existence of very weak DIBs in the vicinity of the 5780 and 5797 bands and shortward of the 6196 band. Last surveys of Kreowski et al. ([1995]) and Kreowski et al. ([1997]) proved that number of very weak features grows with resolution and signal-to-noise ratio. The most complete survey (Jenniskens & Désert [1994]) covered a broad spectral range from $\sim$3800 to $\sim$8700 Å in the spectra of heavily reddened, early type stars. Based on a very small sample (only four stars) this survey detected hundreds of weak features, some of them not known in previous works, but the resolution and signal-to-noise ratio were not high enough to discover very weak DIBs. Also the small sample has not allowed to investigate the behaviour of the weak vs. strong DIBs in different environments.

The presented survey is based on high quality spectra of moderately reddened early spectral type stars. The method of averaging spectra to get a very high signal-to-noise (Kreowski et al. [1997]) makes possible to detect very weak, shallow features never known before. Results of this survey confirm the existence of the already discovered features; additionally many new very weak ones were discovered. These are presented both in tabulated form and as the set of plots showing the spectra in the whole range of this survey.