All Figures

$\begin{figure} \par\resizebox{14cm}{!}{\rotatebox{0}{\includegraphics{1066fg1a.p... ...x{0}{\includegraphics{1066fg1c.ps}} {\includegraphics{1066fg1d.ps}}}\end{figure}$ Figure 1: The most significant non-trivial correlations (see text). p₀ is the debiased broad-band optical polarization (in %), DI is the detachment index (unitless), C IV HREW is the rest equivalent width of the red part of the C IV BEL (in Å), BI is the balnicity index (in $10^3~{\rm km~s}^{-1}$ ), $\alpha _B$ is the slope of the continuum measured shortward of C III] (unitless) and Fe II 2400 is the rest equivalent width of the Fe II emission feature near 2400 Å. Open squares are HI BAL QSOs, filled squares are LI BAL QSOs and crosses are unclassified BAL QSOs.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=8cm,clip]{1066fg2.ps}\end{figure}$ Figure 2: Distribution of the 30 BAL QSOs with respect to PC1 and PC2. Open squares represent HI BAL QSOs, filled squares LI BAL QSOs while the cross represents the only unclassified BAL QSO considered in the PCA. Large symbols refer to bright QSOs with $M_B \leq -28$ and small symbols represent fainter QSOs with M_B > -28.
Open with DEXTER

In the text

$\begin{figure} \par\resizebox{17cm}{!}{\rotatebox{0}{\includegraphics{1066fg3a.p... ...x{0}{\includegraphics{1066fg3c.ps}} {\includegraphics{1066fg3d.ps}}}\end{figure}$ Figure 3: Spectropolarimetric data for the BAL QSOs B1246-0542 ( top left), B1011+0906 ( top right), B2225-0534 ( bottom left) and B2240-3702 ( bottom right). For each object, there are six panels; from top to bottom: total flux $F_{\lambda }$ , degree of polarization p, associated error $\sigma _{\rm p}$ , polarization position angle $\theta$ , associated error $\sigma _{\theta }$ , and polarized flux p $\times$ $F_{\lambda }$ . p has not been debiased. The main absorption lines are indicated.
Open with DEXTER

In the text

$\begin{figure} \par\resizebox{18cm}{!}{\rotatebox{0}{\includegraphics{1066fg4a.ps}} {\includegraphics{1066fg4b.ps}}}\end{figure}$ Figure 4: Spectropolarimetric data for the BAL QSOs B1331-0108 ( left) and B1413+1143 ( right). Panels are as in Fig. 3. The B300 and R300 spectra were cut on their reddest and bluest parts respectively and merged together at the wavelength indicated by the vertical dashed line. The main absorption lines are indicated as well as features due to cosmic rays hits. The asterisk in the spectrum of B1413+1143 indicates an intervening absorption system at $z \sim 1.66$ . The Earth symbol indicates a telluric absorption.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{1066fg5.ps}\end{figure}$ Figure 5: The correlation between the maximum value of the polarization in the C IV absorption line and the balnicity index BI. Open squares are HIBAL QSOs, filled squares are LIBAL QSOs and crosses are unclassified BAL QSOs. The open circle corresponds to an uncertain measurement (see text).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{1066fg6.ps}\end{figure}$ Figure 6: The correlation between the polarization degrees in the C IV and C III] emission lines relative to the polarization in the adjacent redward continuum. The dashed line refers to an equal polarization in the two lines. Symbols are identical to those used in Fig. 5.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{1066fg7.ps}\end{figure}$ Figure 7: The correlation between the detachment index DI and SI, the ratio of the C IV absorption depths in the polarized flux and in the total flux. Both quantities are unitless. Symbols are identical to those used in Fig. 5. The open circle corresponds to an uncertain measurement (see text).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=8.1cm,clip]{1066fg8.eps}\end{figure}$ Figure 8: Schematic illustration of the SI-DI relationship. The equatorial wind, represented by the curved streamlines, emerges from a geometrically thin accretion disk located in the equatorial plane. A mirror of electrons is located in the polar region and scatters some radiation coming from the central regions along the line-of-sight. This polarized component is then subsequently absorbed in the equatorial wind. When the disk is seen edge-on, P Cygni type profiles are observed since the observer sees absorption at all velocities including the low ones near the disk. For less inclined line-of-sights, detached absorption profiles are observed since the observer misses the low velocity part of the absorption.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=8cm,clip]{1066fg2.ps}\end{figure}$	Figure 2: Distribution of the 30 BAL QSOs with respect to PC1 and PC2. Open squares represent HI BAL QSOs, filled squares LI BAL QSOs while the cross represents the only unclassified BAL QSO considered in the PCA. Large symbols refer to bright QSOs with $M_B \leq -28$ and small symbols represent fainter QSOs with M_B > -28.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{1066fg5.ps}\end{figure}$	Figure 5: The correlation between the maximum value of the polarization in the C IV absorption line and the balnicity index BI. Open squares are HIBAL QSOs, filled squares are LIBAL QSOs and crosses are unclassified BAL QSOs. The open circle corresponds to an uncertain measurement (see text).
Open with DEXTER

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{1066fg6.ps}\end{figure}$	Figure 6: The correlation between the polarization degrees in the C IV and C III] emission lines relative to the polarization in the adjacent redward continuum. The dashed line refers to an equal polarization in the two lines. Symbols are identical to those used in Fig. 5.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{1066fg7.ps}\end{figure}$	Figure 7: The correlation between the detachment index DI and SI, the ratio of the C IV absorption depths in the polarized flux and in the total flux. Both quantities are unitless. Symbols are identical to those used in Fig. 5. The open circle corresponds to an uncertain measurement (see text).
Open with DEXTER