Among the 164 observed stars we have found: 5 Li-rich giants; 6 with a moderate Li line; 20 normal giants showing a weak Li line, and 56 showing no Li line; 54 stars present strong molecular TiO bands; and 23 remaining ones correspond to other kinds of objects. The different groups of objects are described in Appendix A (click here).

We estimate the strength of the Li line relative to the Ca I 671.768 nm line. The candidates were classified in three groups, according to the the ratio between the Equivalent Width (W) of these lines:

, and

are considered strong, moderate and weak, respectively.

**Table 1:** List of observed stars which show high and medium equivalent width of the Li 670.782 nm line
IRAS name			identification	S.T.	m_V	W(Li)	W(Li)/W(Ca)
07559-5859	07 55 54.5	-58 59 22	HD 65750, SAO 235638	M0III	6.3	0.31	1.41
16161-1445	16 16 11.4	-14 45 10	HD 146850, SAO 159846	K3III	7.2	0.37	1.32
18241-1443	18 24 06.6	-14 43 41	GCSS557	SRa	10.4	0.21	0.95
18585-0430	18 58 30.6	-04 30 39	HD 176588, SAO 142919	K2III	7.1	0.27	1.00
19012-0747	19 01 15.6	-07 47 06				0.41	1.71
00483-7347	00 48 18	-73 47 00				0.10	0.53
06215-0902	06 21 33.9	-09 02 26	HD 44889, SAO 133209	K0	8.2	0.18	0.60
10204-6135	10 20 26.0	-61 35 51	HD 90082, SAO 250932	M3	7.5	0.20	0.83
11024-6241	11 02 27.0	-62 41 26	HD 96195, SAO 251247	K5	8.0	0.17	0.61
19038-0026	19 03 49.9	-00 26 04				0.13	0.77
19049-0234	19 04 59.5	-02 34 32	HD 178168	K5	9.0	0.12	0.53

3.1. Li-rich giants

The stars showing strong and moderate Li line are listed in Table 1 (click here), where are given: the IRAS name; equatorial coordinates; other identification; the visual magnitude and the spectral type when available; the equivalent width (in Å) of the Li I 670.782 nm doublet

; and the ratio

. Comments on individual objects are presented in Appendix A (click here).

It should be noted that the equivalent width of the

line for all these stars is

, in agreement with results found for

red giants. In Figs. 1 (click here) and 2 (click here) we show the spectra obtained for stars in Table 1 (click here). For one of them (HD 146850) we carried out a detailed analysis based on high resolution spectroscopy (Castilho et al. 1995). Files containing the spectra in the Li I line region for all these stars, and an échelle spectrum of HD 146850, are available in CD ROM (Castilho et al. 1997).

Figure 1: Spectra of the five giants showing large Li-line. Differences in line widths are due to spectral resolution, see Sect. 2

Figure 2: Spectra of the six giants showing mid strength Li-line. Differences on line widths are due to spectral resolution, see Sect. 2

The lithium is expected to be strongly diluted in giant stars according to the standard models. Observations show that in disk stars (BSLD) and globular clusters (e.g. Pasquini & Molaro 1996) the Li abundance for most of the giants is even lower than expected from the models.

Li abundances 200 times larger than the mean value indicate that, either (a) in some stars the original Li is not destroyed, or (b) there should be a process of Li production during the evolution of the star.

Smith & Lambert (1989, 1990) have shown that evolved (AGB) red giants could produce Li during particular phases of their evolution. Theoretical predictions for such production of Li in intermediate mass stars (

) were presented by Sackmann & Boothroyd (1992). Sackman & Boothroyd (1997) using deep circulation computations extend the mass range of Li production to stars of

Some groups of LRG are found in other locations of the IRAS colours diagram, suggesting that they are following tracks in this diagram, probably caused by mass loss (Castilho 1995; de la Reza et al. 1996). In order to test this hypothesis, other regions in the IRAS diagram should be investigated such as the region between the locus studied by us and the T Tauri locus (Gregorio-Hetem et al. 1992). A tentative list of IRAS sources in this region has 400 sources, but the number of identified stars is very poor.

Except for the high Li abundance, the LRG are normal red giants. No correlation was found between mass, rotation or

ratio and Li abundance (da Silva et al. 1995; de Medeiros et al. 1996). This fact and their far infrared emission features suggest that we may be observing not a unique class of stars, but a short phase of the stellar evolution of ordinary stars, when Li is created (Sackmann & Boothroyd 1997).

Detailed analyses for the present sample of LRGs, based on high resolution spectroscopy, will be presented elsewhere. Also, the measurement of B and Be would be of great interest (Sackmann & Boothroyd 1997).

We are grateful to Suzanne Huille for kindly carrying out the observations during two of our runs at OHP. This work benefited from the SIMBAD database operated by Centre des Données Stellaires (CDS) in Strasbourg. We acknowledge partial financial support by CNPq, Fapesp (Brazil) and CNRS (France). BC and JGH acknowledge the CNPq PhD (Proc. No. 840417/97-7) and Post-Doc (Proc. No. 300267/92-4) fellowships.

3. Results

3.1. Li-rich giants