2 Observations

The observations were obtained at the 1.52-meter telescope of the Observatoire de Haute Provence (OHP) on 1996 June 12 and from 1997 July 20 to 1997 July 27 and at the 6-meter telescope of the Special Astrophysical Observatory (SAO) of the Russian Academy of Sciences on 1997 July 23 (JD 2450653.46) and on 1997 November 22 (JD 2450775.16).

The AURELIE spectrograph (Gillet et al. 1994) was mounted at the Coudé focus of the 1.52 meter telescope (OHP). Two spectral regions, 4280 - 4720 and 6150 - 7030 Å, were observed in 1996, while in 1997 we obtained data from 4095 to 8936 Å. In 1996 the dispersion in the red region was 33 Åmm^-1 (resolution $\sim$ 1.1 Å) and 16.5 Åmm^-1 (resolution $\sim$ 0.6 Å) in the blue region. In 1997 the dispersion was 33 Åmm^-1 (resolution $\sim$ 1.2 Å) throughout the whole spectrum. A thorium-argon hollow cathode provided the wavelength calibration spectra. The flux calibration was completed by comparing observations of HR 7174 (HD 176318) to its absolute fluxes obtained by Gunn & Stryker (1983). Two spectral regions (6275-7150 and 5590-6500 Å) were observed during cloudy weather which led to a wrong estimate of the fluxes. In these cases, the spectra were shifted until their fluxes coincide to those observed in the other regions. In Table 1 we report the integrated fluxes for each line appearing in the 1997 spectra. The red part of the spectrum is strongly affected by telluric absorption bands. In view to identify possible strong transitions appearing in the polluted spectral regions, we divided the spectrum of MWC 314 by the spectrum observed for HR 7174 during the same nights. We emphasize that these "corrected'' regions allow us only to identify possible strong transitions. The derived equivalent widths and intensities have to be used very carefully.

The July 1997 spectrum was taken with the CCD equipped echelle-spectrometer LYNX (Klochkova 1995) mounted at the Nasmyth focus of the 6-meter telescope in the spectral range from 5370 to 6670 Å. A $1140\times 1170$ pixels CCD was employed. We used an echelle-grating with the blaze angle of 64.3$\hbox{$^\circ$}$ and 37.5 grmm^-1, the cross disperser has 300 grmm^-1 at first order. The July observations were obtained at the poor weather conditions (seeing approximately $3-4\hbox{$^{\prime\prime}$}$, clouds), therefore the slit size was enlarged to increase the signal-to-noise ratio. The average spectral resolution was equal to 0.4 Å. The November spectrum was obtained at the prime focus of the 6-meter telescope with the echelle-spectrometer PFES (Panchuk et al. 1998) in the range from 4700 to 8590 Å. The seeing was nearly 2$\hbox{$^{\prime\prime}$}$, while the spectral resolution was about 0.8 Å. However, due to the larger exposure time the signal-to-noise ratio was larger than that in July. This allowed us to measure the line profiles more precizely, but led to the H$\alpha$saturation.

An argon-filled thorium hollow-cathode lamp was used as the source of comparison spectrum. Control and correction of instrumental displacement of the comparison and object spectra were done in using the telluric lines of O₂ and H₂O. We tested focusing with the spectra obtained over the observing night using telluric lines. The sky spectrum did not reveal any splitting of these lines.