All Figures

$\begin{figure} \par\includegraphics[width=14.3cm,clip]{8820fig1.eps} \end{figure}$ Figure 1: Normalised spectra around the TiII 3384 Å (solid lines) and 3242 Å (dashed lines) transitions for our target stars. The abscissa is the heliocentric velocity in km s^-1.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=6.8cm,clip]{8820fig2.eps} \end{figure}$ Figure 2: Illustration of fringe correction to the data: the extreme case of $\theta$ Car.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=12.5cm,clip]{8820fig3.eps} \end{figure}$ Figure 3: 3394-3242 Å simultaneous line-fitting used as a consistency checked: $\theta$ Car, JL9, and LS1274. The blue-shifted absorption at 3384 Å in LS1274 has no counterpart at 3242 Å and is very likely stellar (see text).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=12.8cm,clip]{8820fig4.eps} \end{figure}$ Figure 4: Same as Fig. 2: $\gamma ^{2}$ Vel, $\zeta$ Pup, and $\mu$ Col. The broad feature in $\mu$ Col cannot be fitted satisfyingly (see text).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=12.3cm,clip]{8820fig5.eps} \end{figure}$ Figure 5: Ionized titanium to neutral hydrogen column ratio vs. DI/HI with ( left) and without ( right) the star $\mu$ Col. The method used to take into account non symmetric error bars for the abscissa is illustrated in the left figure. For the data point 1 (resp. 2), because it is located on the left (resp. right) side of the fitted line, the selected error bar is P1 (resp. N2). The same method is applied for the errors in the ordinate parameter. The corresponding linear fit for FeII is shown as a dotted line in the right figure.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=6.7cm,clip]{8820fig6.eps} \end{figure}$ Figure 6: Titanium vs. deuterium columns. The star $\mu$ Col is clearly an ``outlier''.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=6.45cm,clip]{8820fig7.eps} \end{figure}$ Figure 7: Ionized iron abundance normalised to (Fe/H)₀ = -4.55 (Asplund et al. 2005) vs. DI/HI. This figure is identical to Fig. 3 from Linsky et al. (2006) except for the linear scale. The best fit linear correlation using the Orthogonal Distance Regression (ODR) is shown as a solid line, and for comparison the titanium gradient obtained with the same method (Fig. 5) is shown as a dashed line after the appropriate scaling to obtain a similar ordinate for D/H = 20. It can be seen that the dispersion is significantly larger for FeII than for TiII. Lines-of-sight filled with strongly ionized gas (G191-B2B, 36 Oph) are positive extreme ``outliers'', while those corresponding to very large columns tend to produce negative ``outliers''.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=6cm,clip]{8820fig8.eps} \end{figure}$ Figure 8: Left: ionized titanium vs. DI/OI. The corresponding linear fit for FeII (Fig. 5) is shown for comparison.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=12.7cm,clip]{8820fig9.eps} \end{figure}$ Figure 9: Iron gaseous abundance (normalised to the Asplund et al. abundance) and deuterium to oxygen ratio. a) Published DI/OI results. b) After division by a factor of 2 of the three DI/OI ratios for the three distant stars (see text).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=11.3cm,clip]{8820fig10.eps}\end{figure}$ Figure 10: TiII/DI ratio as a function of the HI column: all available measurements. The linear fit using the Orthogonal Distance Regression method is shown as a solid line. The slopes found by Linsky et al. (2006) for the FeII/DI ratio are shown as dashed and dotted lines (resp. all data, and data with log N(HI) $\geq$ 10¹⁹ cm^-2).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14.3cm,clip]{8820fig1.eps} \end{figure}$	Figure 1: Normalised spectra around the TiII 3384 Å (solid lines) and 3242 Å (dashed lines) transitions for our target stars. The abscissa is the heliocentric velocity in km s^-1.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=6.8cm,clip]{8820fig2.eps} \end{figure}$	Figure 2: Illustration of fringe correction to the data: the extreme case of $\theta$ Car.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=12.5cm,clip]{8820fig3.eps} \end{figure}$	Figure 3: 3394-3242 Å simultaneous line-fitting used as a consistency checked: $\theta$ Car, JL9, and LS1274. The blue-shifted absorption at 3384 Å in LS1274 has no counterpart at 3242 Å and is very likely stellar (see text).
Open with DEXTER

$\begin{figure} \par\includegraphics[width=12.8cm,clip]{8820fig4.eps} \end{figure}$	Figure 4: Same as Fig. 2: $\gamma ^{2}$ Vel, $\zeta$ Pup, and $\mu$ Col. The broad feature in $\mu$ Col cannot be fitted satisfyingly (see text).
Open with DEXTER

$\begin{figure} \par\includegraphics[width=6.7cm,clip]{8820fig6.eps} \end{figure}$	Figure 6: Titanium vs. deuterium columns. The star $\mu$ Col is clearly an ``outlier''.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=6cm,clip]{8820fig8.eps} \end{figure}$	Figure 8: Left: ionized titanium vs. DI/OI. The corresponding linear fit for FeII (Fig. 5) is shown for comparison.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=12.7cm,clip]{8820fig9.eps} \end{figure}$	Figure 9: Iron gaseous abundance (normalised to the Asplund et al. abundance) and deuterium to oxygen ratio. a) Published DI/OI results. b) After division by a factor of 2 of the three DI/OI ratios for the three distant stars (see text).
Open with DEXTER

$\begin{figure} \par\includegraphics[width=11.3cm,clip]{8820fig10.eps}\end{figure}$	Figure 10: TiII/DI ratio as a function of the HI column: all available measurements. The linear fit using the Orthogonal Distance Regression method is shown as a solid line. The slopes found by Linsky et al. (2006) for the FeII/DI ratio are shown as dashed and dotted lines (resp. all data, and data with log N(HI) $\geq$ 10¹⁹ cm^-2).
Open with DEXTER