4 Discussion

Three arguments support the interpretation of the absorption feature detected on the blue wing of some of the higher members of the H I Lyman series, Ly$\epsilon$, Ly8, Ly10 and Ly12, as the D I absorption with the same two main components as detected in the metal lines of the DLA system.

First, the absorption in the blue wing of the Lyman series lines could also be explained by a hydrogen interloper associated with the damped Ly$\alpha$ system with a column density between 10^15.7 and 10^16.2, a b-value between 15 and 25 km s^-1 and placed between -80 and -100 km s^-1. However, from the density distribution of the Ly$\alpha$ clouds in the forest at $z\sim 3$ (see e.g. Kim et al. 2000), the probability to have such a cloud at that position is smaller than 1/1000. The lack of any metal component at this velocity in the strong and saturated lines of the DLA system (see Fig. 1) provides an additional evidence for discarding this possibility.

Secondly, the contamination on the blue wing of the Lyman series lines by different H I interlopers at different redshifts which would mimic the same deuterium abundance as derived from the D I Ly$\epsilon$, Ly8, Ly10 and Ly12 lines is even more unlikely.

Finally, the D and H column densities and broadening parameters b resulting from the fits are consistent one relatively to other in the two fitted components (2 and 3): the derived D/H ratios are the same and the b(D I)/b(H I) ratios are close to what is expected in the thermally dominated case.

This measurement of D is the first made in a DLA system. It shows that the DLA systems with their low metallicity ISM are a very promising class of absorbers for measurements of the D/H ratios at high redshifts, when it is possible to measure the higher members of the Lyman series. A systematic program of measurements using UVES data is under way.

The derived D/H ratio of $(2.24\pm 0.67)~ 10^{-5}$ is close to the low values obtained by Burles & Tytler in Lyman limit systems (1998a, 1998b) and it makes the claim of the primordial low D/H ratio more robust. Taken at face this ratio gives a baryon to photon ratio, $\eta = n_b/n_\gamma$, of $\approx$ 6.3  10^-10 (Burles et al. 2000). This $\eta$ implies a helium abundance (in mass) of $Y_{\rm p} = 0.2480$ and a lithium abundance of $\frac{^7 \rm Li}{\rm H} \approx 5~ 10^{-10}$ which are both larger than presently allowed by observations of He in extragalactic H II regions and of Li in halo stars (Izotov & Thuan 1998; Bonifacio & Molaro 1997). On the other hand an $\eta$ of $\approx$ 6.3  10^-10 corresponds to a present-day baryon density of $\Omega_{\rm b} h^2 = 0.023$ which remains significantly lower than the $\Omega_{\rm b} h^2 = 0.032^{+0.009}_{-0.008}$ derived from CMB anisotropy (Jaffe et al. 2000).

Acknowledgements

We are indebted to the UVES project team for the high quality of the spectra obtained early in the operation of the instrument. We like to thank J. X. Prochaska for making available the profiles of the stronger metal lines and for comments on an earlier version of the manuscript.