...1,

Postdoctoral Researcher FNRS (Belgium).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...1,

Research Associate FNRS (Belgium).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...5,6,

Member of CONICET, Argentina.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... ISGRI

As we were interested mainly in a priori rather soft faint sources, we did not consider PICsIT data in our analysis.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... sources

These sources are those that were already detected by De Becker (2005), along with IGR J21223+5105 and SS Cyg that were apparent point sources quoted as "s1'' and "s2'' by De Becker (2005) although they were not detected with a significance above 3$\sigma$.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Cyg X-1

Every science window image in the FCFOV was carefully inspected in order to reject those presenting the strongest artefacts (chessboard structures, bright ghosts, noisy structures). However, this selection (up to 60% of the data rejected) did not reduce the large scale artificial structures centered around Cyg X-1 that are still seen in the mosaic image obtained with the data set discussed in this paper (see Fig. 1).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... background

Without these structures and artefacts, the background would be dominated by a diffuse extragalactic emission in the soft part of the ISGRI bandpass, and by the internal background due to the de-excitation of nuclei produced by spallation reactions of cosmic rays on the instrument material at higher energies (Lebrun et al. 2003).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... WR 140

We note that the position of WR 140 is coincident with the error box of the EGRET source 3EG J2022+4317 (Romero et al. 1999; Benaglia & Romero 2003).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Pittard & Dougherty (2006)

The approach developed by Pittard & Dougherty (2006) consists in fitting models to radio data in order to determine the population and spatial distribution of relativistic electrons without an a priori knowledge of the magnetic field strength.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.