2 Observations

The H$\alpha$ observations presented here were made at La Silla with a small telescope (36 cm diameter) equipped with a scanning Fabry-Perot interferometer and a photon counting camera (le Coarer et al. 1992). The data cubes (x, y, $\lambda$)obtained have a spatial resolution of $9'' \times 9''$ (total size of each field $= 38' \times 38'$) and a spectral resolution depending on the interference order of the Fabry-Perot: $\sim$5 km $\rm s^{-1}$ for order 2604 and $\sim$ 16 km $\rm s^{-1}$ for order 796.

The data reduction method is described in Georgelin et al. (1994). The interference filter, centered at 6562 Å and with FWHM 11 Å, unfortunately transmits not only the H$\alpha$ nebular emission but also two well identified night-sky lines: geocoronal H$\alpha$ ($\lambda_{\rm 0}$ = 6562.8 Å) and OH ($\lambda_{\rm 0}$ = 6568.8 Å).

In order to decompose the observed profiles into elementary components, the night-sky line and the nebular ones will be respectively modeled by a pure instrumental profile, and an instrumental profile convolved with a gaussian.

The decomposition of high emission zones allows us to identify the different components encountered along the line of sight. However, the decomposition becomes trickier when the emission zones are fainter, even if each component generally has about the same velocity across the entire field.

The actual fields observed in H$\alpha$ were selected on the basis of large scale photographic data (Georgelin & Georgelin 1970) and the 5 GHz radio emission survey (Haynes et al. 1978). They include the regions previously studied by Georgelin et al. (1987, 1988); most of these were observed with a fixed order Fabry-Perot interferometer, the others with a scanning Fabry-Perot interferometer with a very small field (diameter $\sim$12'). We note that for some regions previously observed we do not find exactly the same velocity. It is obvious that the old observations made with a fixed Fabry-Perot did not allow any profile decomposition, thus preventing night-sky lines subtraction and faint component separation, so the velocities corresponded to a complex blend.