astro-ph0107072/ms.tex
1: \documentstyle[11pt,aaspp4,psfig,lscape]{article}
2: \doublespace
3: 
4: \begin{document}
5: 
6: 
7: \title{Sulphur abundance in very metal-poor stars\footnote{
8: Based on observations  obtained with the William Herschel
9: Telescope, operated on the island  of La Palma by the Isaac Newton 
10: Group in the Spanish Observatorio  del Roque de los Muchachos 
11: of the Instituto de Astrof\'\i sica de  Canarias.}}
12: 
13: 
14: 
15: \author{Garik~Israelian and Rafael~Rebolo\altaffilmark{2}}
16: \affil{Instituto de Astrof\'\i sica de Canarias, E-38200
17: La Laguna, Tenerife, Spain} 
18: 
19: \authoraddr{Instituto de Astrof\'\i sica de Canarias, E-38200 La Laguna, 
20: Tenerife, Spain}
21: 
22: \altaffiltext{2}{Consejo Superior de Investigaciones Cient\'\i fcas, Spain}
23: 
24: 
25: \authoremail{gil@ll.iac.es, rrl@ll.iac.es}
26: 
27: 
28: 
29: \begin{abstract}
30: 
31: 
32: We have obtained high-resolution high signal-to-noise spectra of the S\,{\sc i}
33: near-infrared doublet at 8694 \AA\ in eight  metal-poor stars with metallicities
34: in the range  $-0.6\leq$[Fe/H]$\leq -3.0$. Elemental sulphur abundances were
35: derived for six targets and upper limits were set for two of the most metal-poor
36: stars in the sample. The sulphur doublet at 8694.62 \AA\ has been detected in 
37: three stars (HD\,19445, HD\,2665 and HD\,2796)  with [Fe/H]$\leq -1.9$.  
38: Our observations combined with those available in the literature indicate a monotonic
39: increase of the  [S/Fe] ratio as [Fe/H] decreases, reaching values [S/Fe]$\sim$0.7-0.8
40: below [Fe/H]=$-$2. We discuss plausible scenarios for interpretation of these
41: results. 
42: 
43: 
44: \end{abstract}
45: 
46: \keywords{Galaxy: evolution --- nuclear reactions, nucleosynthesis,
47: abundances --- stars: abundances --- stars: late-type --- stars:
48: Population II}
49: 
50: 
51: 
52: \section{Introduction}
53: 
54: 
55: The standard scenario for the chemical evolution of the Galactic Halo
56: (see e.g. Tinsley 1979) is based on the different  yields of Type Ia and
57: Type II Supernovae. It is well established that Type II supernovae mostly
58: produce the so-called $\alpha$-elements (O, Ca, Mg, S, Si and Ti) while 
59: Type Ia supernovae are responsible for the synthesis of the Fe-group
60: (Fe, Co, Ni, Zn). Given a different life-times of both types of supernovae
61: (10$^7$ years for Type II and 10$^9$ years for Type Ia), it is expected
62: that there will be overabundance of $\alpha$ elements  at [Fe/H]$< -1$. 
63: At first glance, this behavior is observed in many metal-poor halo star
64: (McWilliam 1997). The precise trends of the [$\alpha$/Fe] ratio versus 
65: [Fe/H] are still not clearly established. For instance, low [$\alpha$/Fe] ratios 
66: were found in several metal-poor halo stars (Carney et al. 1997) and a rather
67: complicated picture emerged from the comprehensive work by Idiart \& Thevenin (2000)
68: who studied  [Ca/Fe] and [Mg/Fe] ratios in  252 halo and disk stars. 
69: On the other hand, the [O/Fe] ratio in  unevolved stars show a linear rise 
70: (Israelian, Garc\'\i a L\'opez \& Rebolo 1998, Israelian et al. 2001)
71: with a slope $-$0.33$\pm$0.02 in the metallicity range 0$\leq$[Fe/H]$\leq -3$. 
72: Apparently there is no general agreement about a unique $plateau$ 
73: [$\alpha$/Fe]$\approx const$ at [Fe/H]$< -$1 for  $\alpha$-elements 
74: such O, Ca and Mg. This statement can be extended to another $\alpha$-element:
75: sulphur. 
76: 
77: In his extensive analysis Fran\c{c}ois (1987, 1988) was able to delineate the [S/Fe]
78: trend down to the metallicity [Fe/H]=$-$1.6 suggesting that 
79: the [S/Fe] ratio in the halo could be constant at a value around 0.6. The goal
80: of this letter is to investigate whether unevolved halo stars  at lower 
81: metallicities support the proposed trend.
82: 
83: \section{The Observations}
84: 
85: The observations were carried out in 2000 August 17-19 using the Utrecht 
86: Echelle Spectrograph (UES) at the Nasmyth focus of the 4.2-m WHT at the ORM 
87: (La Palma). The CCD detector was a 2148$\times$2148 pixel$^2$
88: SITE1. We obtained echelle spectra with a slit width $\sim$1\farcs{0} which 
89: cover the wavelength range between 5220 and 9150 \AA\ in 20 orders at a 
90: spectral resolving power of $R=\lambda /\Delta\lambda\sim 50,000$. 
91: All spectra were background and scattered light subtracted, flat-fielded
92: and extracted using standard {\sc iraf} \footnote{{\sc iraf} is
93: distributed by the National Optical Astronomical Observatories, which is
94: operated by the Association of Universities for Research in Astronomy, Inc.,
95: under contract with the National Science Foundation, USA.} procedures.
96: The flat-field correction was attempted with a fast rotator, tungsten
97: lamp and dome flats. However, the best correction from a severe fringing
98: in the near-infrared was achieved with high quality dome flats. 
99: The wavelength calibration was performed with a Th--Ar lamp giving a 
100: dispersion of 0.087 \AA\/pixel.
101: The final signal-to-noise ratio of our spectra is provided in Table 1.
102: 
103: 
104: \section{Spectral synthesis, stellar parameters and abundances}
105: 
106: The LTE spectrum synthesis code MOOG (Sneden 1973) and ATLAS9 model
107: atmospheres (Kurucz 1992) were used in the present study. Atomic parameters 
108: of the sulphur doublet 8693.93 \AA\ ($\log gf$=$-$0.5, $\chi$=7.87 eV) 
109: and 8694.62 \AA\ ($\log gf$=0.08, $\chi$=7.87 eV)  
110: were obtained from the VALD-2 database (Kupka et al. 1999). 
111: These $\log gf$-values combined with the equivalent widths  
112: 28 and 11 m\AA\ of the sulphur doublet derived from the high-resolution solar
113: atlas (Kurucz et al. 1984) provide a solar sulphur abundance 
114: $\log \epsilon$(S)=7.20$\pm$0.02 (on the customary scale in which 
115: $\log \epsilon$(H)=12). This is in excellent agreement with 
116: $\log \epsilon$(S)=7.21 given by Anders \& Grevesse (1989). 
117: 
118: All our targets (Table 1) are well known metal-poor stars. The IRFM based 
119: effective temperature, $T_{\rm eff}$, of two giants HD\,2796, BD$-$18\arcdeg5550 
120: and a subgiant HD\,2665 were taken from Alonso, Arribas,  \& Mart\'\i nez-Roger (1999) 
121: while the surface gravities and [Fe/H] values were estimated from Fe NLTE
122: analysis (Israelian et al., in preparation) using the same tools as 
123: in Israelian et al. (2001). In any case, the parameters of these three
124: evolved stars  (listed in Table 1) are very similar to those reported in the
125: literature (Pilachowski, Sneden \& Booth 1993, Alonso et al. 1999, 
126: Burris et al. 2000). The $T_{\rm eff}$ of five subdwarfs (Table 1) are also based 
127: on the IRFM and come from Israelian et al. (1998). Their metallicities and surface 
128: gravities were adopted from the NLTE Fe study of Thevenin \& Idiart (1999) and 
129: agree perfectly with those estimated using  {\it Hipparcos} parallaxes 
130: (Allende Prieto et al. 1999). We note that our $T_{\rm eff}$-values agree 
131: with those of Thevenin \& Idiart (1999) within 60 K.   
132: NLTE effects on sulphur are negligible (less than 0.05 dex) due to the weakness
133: of the sulphur lines in our targets (Takada-Hidai et al. 2001).  
134: 
135: 
136: Our tests show that a typical change of +100 K in $T_{\rm eff}$ increases
137: [S/H] by $-$0.05 and $-$0.09 dex for $\log g$=4.15 and $\log g$=1.5, respectively.
138: A change of surface gravity by +0.3 dex implies an uncertainty 0.11 and 0.13 dex
139: for $T_{\rm eff}$=5800 and 4900 K, respectively. The errors arising from
140: the uncertainties in the metallicity and the microturbulent velocity $V_{mic}$
141: are negligible. We set $V_{mic}$=2\,\hbox{km$\;$s$^{-1}$} in HD\,2796, HD\,2665,
142: BD$-$18\arcdeg5550 and $V_{mic}$=1\,\hbox{km$\;$s$^{-1}$} for the dwarfs. The sulphur
143: abundance for each star (Table 1) was obtained as the weighted mean of the 
144: abundances resulting for each line of the doublet (with weights proportional
145: to the equivalent widths). The final error in the sulphur abundance was
146:  derived as a quadratic sum of the errors associated with the stellar parameters and
147: the errors associated to the equivalent width measurements. The final errors for
148: the most metal-rich five dwarfs are similar to those reported by Fran\c{c}ois (1987, 1988)
149:  being  of the order of $\pm$0.1 dex. The errors are larger in the most metal-poor 
150: targets of the sample. 
151: 
152: We have reanalyzed sulphur abundances from Fran\c{c}ois (1987, 1988) using 
153: the equivalent widths provided in these papers. Effective temperatures (Table 2) 
154: have been estimated using the IRFM calibration of Alonso, Arribas,  \& Mart\'\i nez-Roger 
155: (1996) for Str\"omgren $b-y$ and c$_1$ indices taken from Hauck and Mermillod 
156: (1998) and [Fe/H] adopted from Thevenin \& Idiart (1999), Allende Prieto et al. (1999) 
157: and Allende Prieto \& Lambert (1999). No reddening corrections were applied since all 
158: these stars are nearby within 100 parsec (Perryman et al. 1997). Typical errors 
159: in $T_{\rm eff}$ are $\pm$100 K.  Allende Prieto et al. (1999) have shown that 
160: gravities derived from Fe NLTE analysis by Thevenin \& Idiart (1999) agree with 
161: the gravities inferred from accurate {\it Hipparcos} parallaxes in the metallicity 
162: range $-1 \leq [Fe/H] \leq 0$. 
163: Surface gravities of 
164: these stars were estimated using the {\it Hipparcos} parallaxes. We used $V$ 
165: magnitudes from {\it Hipparcos} Catalogue to compute $M_{\rm v}$  and 
166: to estimate absolute bolometric magnitudes $M_{\rm bol}$ from Alonso, 
167: Arribas, \& Mart\'\i nez-Roger (1995). Therefore the set of new parameters 
168: listed in Table 2 is homogeneous (three stars have not been considered
169: because their parameters were rather uncertain). The masses of stars were 
170: inferred using theoretical isochrones from evolutionary tracks of 
171: the Geneva group (Schaerer et al. 1993). Our gravities have typical errors 
172: $\pm$0.2 dex and are in good agreement with those presented by different 
173: authors (Allende Prieto et al. 1999, Allende Prieto \& Lambert 1999) using, 
174: however, different evolutionary tracks.
175:  
176: 
177: \section{Discussion and Conclusions}
178: 
179: 
180: 
181: The [S/Fe] ratios obtained for our targets are listed in Table 1 and plotted in 
182: Fig. 3 together with Fran\c{c}ois (1987, 1988) measurements.  NLTE estimates for 
183: the metallicities of his stars  have been made using the polynomial fit from
184: Israelian et al. (2001) which is based on the work of Thevenin \& Idiart (1999). 
185: Our results are consistent with those by Fran\c{c}ois in the overlapping metallicity 
186: range. Remarkably, at very low metallicities we find the highest [S/Fe] ratios supporting
187: a progressive increase of the [S/Fe] ratio as metallicity decreases. A minimum-square
188: linear fit for the points in Fig. 3 results in a slope of $-$0.46 $\pm$0.06 (1-$\sigma$) 
189: which can be compared with the slope in the [O/Fe] vs. [Fe/H] relationship of 
190: $-$0.33  $\pm$0.02 (1-$\sigma$) found by Israelian et al. (2001).
191:  
192: Sulphur yields from standard Type II supernovae have been computed by many
193: authors  (e.g., Woosley \& Weaver 1995; Thielemann, Nomoto, \& Hashimoto 1996). 
194: The [S/Fe] ratio in the ejecta is sensitive to the amount of mass going into the
195: compact object formed during the supernova explosion (i.e depend on the adopted
196: mass cut). There is a range of
197: values for this mass cut  which can perfectly provide [S/Fe] yields consistent
198: with the observations. The tentative linear trend delineated in Fig. 3 may 
199: be an indication that the explosive energy of  Type II supernovae may depend
200: on progenitor mass. Alternatively,  Nomoto et al. (2001) and Nakamura et al. (2001) 
201: have presented models of explosive nucleosynthesis in hypernovae. They 
202: consider exploding He cores with different initial masses and explosion
203: energies in the range 1--100 $\times$ 10$^{51}$ erg. Their computations show
204: that [S/Fe] ratios as those found in very low-metallicity stars can easily be
205: produced in these explosions. For instance, it can be seen from Table 2
206: of Nakamura et al. (2001) that the He core with a mass 16~M$_{\sun}$  and 
207: 10$\times$10$^{51}$ erg explosion energy is able to provide a ratio [S/Fe]$\sim$1.1 
208: in the final ejecta. It is possible that the linear trend of [S/Fe] can be explained 
209: by an increasing role of hypernovae in the early epochs of the Galaxy formation 
210: (i.e., [Fe/H] $< -1$). The estimated Galactic rate of Type Ic hypernovae is 
211: $10^{-3}$~yr$^{-1}$ (Hansen 1999) but it could have been  much higher in the
212: early Galaxy. Small mass-loss rates due to the low metallicity in the first
213: generations of massive halo stars could have led to large helium cores and
214: the conservation of the primordial angular momentum. Since hypernovae
215: are favored by stars with a large helium core mass and rapid rotation,
216: a significant sulphur production in the early Galaxy could have
217: been a rather common phenomenon. Similar ideas have been put forward by 
218: Qian and Wasserburg (2001) to explain a linear rise in the [O/Fe] ratio.
219: Extension of their model for the sulphur case would be very interesting.
220: 
221: 
222: Additional observational support for  large amounts of sulphur synthesized 
223: during Type II supernova/hypernova explosion  comes from the analysis of
224: the chemical composition of the  low-mass  star orbiting the black hole 
225: in the X-ray binary system  GRO J1655-40 (Israelian et al. 1999). 
226:  $\alpha$ elements are uniformly enhanced in the atmosphere of the secondary star, 
227: and in particular a  large amount of sulphur [S/H]$\sim$0.9  was found in the
228: secondary star. Most probably, this is a result of contamination from the matter 
229: ejected after supernova or hypernova explosion of the black hole stellar progenitor 
230: (Israelian et al. 1999, Brown et al. 2000). Very recently, Combi et al. (2001)
231: have detected a bubble in the large-scale HI distribution around GRO J1655-40
232: providing new evidence for a supernova explosion in this system.
233: 
234: 
235: Another possible explanation for the monotonic increase in [S/Fe] at low [Fe/H]
236: is related to a delayed deposition of the supernova synthesized products into
237: the ISM due to differences in transport and mixing. Choosing a short mixing 
238: time ($\sim$ 1\,Myr) for oxygen (a volatile), and longer mixing time
239: ($\sim$ 30\,Myr) for Fe (refractory), Ramaty et al. (2000, 2001) suggest that 
240: [O/Fe] should rise monotonically with a slope which is consistent with the 
241: observations of Israelian et al. (1998, 2001). A similar trend is also 
242: predicted (Ramaty et al. 2001) for a volatile element like sulphur. 
243: As dicussed by Ramaty et al. (2000, 2001), it is not necessary that other
244: $\alpha$-elements like Mg and Si follow the same trend as O and S since 
245: they are refractory elements with much longer mixing time in  the ISM.      
246: In summary, while more observations are clearly needed at very low metallicities
247: in order to confirm the trend delineated in the present work, our new measurements
248: do not seem to support a [S/Fe] $plateau$ versus [Fe/H] in halo stars. There are
249: several scenarios that can account for a monotonic increase in [S/Fe]
250: at low metallicities.
251: 
252: \vspace{0.1cm}
253: 
254: \begin{acknowledgements} 
255: We thank Y. Takeda and M. Takada-Hiadi for providing the NLTE corrections 
256: for sulphur IR doublet before the publication. This work was partially
257: funded by project DGES PB 98-0531-c02-02.
258: 
259: \end{acknowledgements}
260: 
261: \begin{thebibliography}{}
262: 
263: \bibitem[Allende Prieto et al. 1999]{Al99}
264: Allende Prieto, C., Garc\'\i a L\'opez, R. J., Lambert, D. L., \& Gustafsson, B.
265: 1999, \apj, {\bf 527}, 879
266: 
267: \bibitem[Alonso et al.\ 1995]{Alons95} Alonso, A., Arribas, S.,
268: \& Mart\'\i nez-Roger, C. 1995, \aap, 297, 197
269: 
270: 
271: 
272: \bibitem[Alonso et al.\ 1996]{Alons96} Alonso, A., Arribas, S.,
273: \& Mart\'\i nez-Roger, C. 1996, \aap, 313, 873
274: 
275: 
276: \bibitem[]{Alons96} Alonso, A., Arribas, S.,
277: \& Mart\'\i nez-Roger, C. 1999, \aaps, 139, 335
278: 
279: 
280: \bibitem[]{} Anders, E., Grevesse, N. 1989,
281: Geochim. Cosmochim. Acta, 53, 197
282: 
283: \bibitem[]{} Burris, D. L., Pilachowski, C. A., Armandroff, T. E., Sneden, C., 
284: Cowan, J. \& Roe, H. 2000, \apj, 544, 302
285: 
286: 
287: 
288: \bibitem[]{} Brown, G.~E., Wijers, R. A. M.~J., Lee, C.-H., Lee, H.-K., Israelian, G.
289: \& Bethe, H.~A. 2000, New Astronomy, 5, 191
290: 
291: \bibitem[]{} Carney, B., Wright, J., Sneden, C., Laird, J., Aguilar, L. \& Latham, D.
292: 1997, \aj, 114, 363
293: 
294: \bibitem[]{} Cayrel, R. 1988,
295: in {\em "The impact of Very High S/N Spectroscopy on Stellar Physics"},
296: Cayrel de Strobel, G. \& Spite, M. (eds.), Dordrecht, Kluwer, 345
297: 
298: 
299: \bibitem[]{} Combi,  J.A.,  Romero, G.E., Benaglia, P. \&  Mirabel, I.F.
300: 2001, \aap, 370, L5
301: 
302: \bibitem[]{} Fran\c{c}ois, P. 1987, \aap, 176, 294
303: 
304: \bibitem[]{} Fran\c{c}ois, P. 1988, \aap, 195, 226
305: 
306: 
307: \bibitem[Hansen 1999]{}
308: Hansen, B.M.S. 1999, \apj, 512, L117
309: 
310: 
311: \bibitem[]{}
312: Hauck, B. \& Mermilliord, M. 1998, \aaps, 129, 431
313: 
314: 
315: \bibitem[]{} Idiart, T. \& Thevenin, F. 2000, \apj, 541, 207
316: 
317: 
318: \bibitem[]{} Israelian, G., Garc\'\i a L\'opez, R. \& Rebolo, R.: 1998, \apj, 507, 805
319: 
320: \bibitem[]{} Israelian, G., Rebolo, R., Basri, G., Casares, J., \&
321: Martin, E.~L. 1999, Nature, 401, 142
322: 
323: \bibitem[]{} Israelian, G., Rebolo, R., Garc\'\i a L\'opez, R., 
324: Bonifacio, P., Molaro, P., Basri, G., \& Shchukina, N. 2001, \apj, 551, 833
325: 
326: \bibitem[]{} Kupka F., Piskunov N.E., Ryabchikova T.A., Stempels H.C.,
327: Weiss W.W. 1999, \aaps., 138, 119
328: 
329: 
330: \bibitem[Kurucz 1992]{cdrom93} Kurucz, R. L. 1992, private communication
331: 
332: \bibitem[]{} Kurucz, R. L., Furenlid, I., Brault, J. \& Testerman, L. 1984,
333: Solar Flux Atlas From 296 to 1300 nm. NOAO Atlas 1
334: 
335: \bibitem[]{} McWilliam, A. 1997, \araa, 35, 503
336: 
337: \bibitem[]{} Nakamura, T., N., Umeda, H., Iwamoto, K., Nomoto, K., 
338: Hashimoto, Masa-aki. H.,  Raphael Hix, W.,  Thielemann, F-K., 2001, \apj, 
339: in press (astro-ph/0011184)
340: 
341: \bibitem[]{} Nomoto, K., Mazzali,  P.A., Nakamura, T., Iwamoto, K.,Maeda,  K.,  
342: Suzuki, T., Turatto, M., Danziger,  I. J., Patat, F. 2001, in 
343: {\em "Supernovae and Gamma Ray Bursts"}, ed. M. Livio et al. (Cambridge
344: University Press), in press (astro-ph/0003077)
345: 
346: \bibitem[]{} Perryman et al. 1997, The Main Hipparcos Catalogue, ESA
347: 
348: \bibitem[]{} Pilachowski, C. A., Sneden, C. \& Booth, J. 1993, \apj, 407, 699
349: 
350: \bibitem[]{} Qian, Y.-Z. \& Wasserburg, G.~J, 2001, \apj, 549, 337
351: 
352: \bibitem[Ramaty et al. 2000]{} Ramaty, R., Scully, S.,
353: Lingenfelter, R. \&  Kozlovsky, B. 2000, \apj, 534, 747
354: 
355: \bibitem[]{} Ramaty, R., Lingenfelter, R. \&  Kozlovsky, B. 2001, in
356: \newblock in  L. da Silva, M. Spite, and J.R. de Medeiros (eds.),
357: The Light Elements and their Evolution, IAU Symposium 198, p. 51
358: 
359: 
360: 
361: %\bibitem[Takeda et al. 2000]{}  Takeda, Y., Takeda-Hidai, M., Sato, S.,
362: %Sargent, W.L., Lu, L., Barlow, T., Jugaku, J. 2000, \apj, submitted
363: 
364: 
365: \bibitem[]{} Schaerer, D., Charbonnel, C., Meynet, G., Maeder, A.
366: \& Schaller, G. 1993, \aaps, 102, 339
367: 
368: 
369: 
370: \bibitem[Sneden 1973]{sned73} Sneden, C. 1973, \apj, 184, 839
371: 
372: \bibitem[]{} Takada-Hidai, M., Takeda, Y.,  Sato, S.,
373: Honda, S., Sadakane, K., Kawanomoto, S., Sargent, W.L.W., Lu, L. and 
374: Barlow, T. A. 2001, submitted
375: 
376: \bibitem[Theilemann, Nomoto \& Hashimoto 1997]{} Theilemann, F.-K.,
377: Nomoto, K., \& Hashimoto, M. 1996, \apj, 460, 108
378: 
379: \bibitem[Thevenin \& Idiart 1999]{Thevenin+Idiart1999}
380: Thevenin, F. \& Idiart, T. 1999, \apj, 521, 753
381: 
382: \bibitem[Tinsley 1979]{tin79} Tinsley, B. M. 1979, \apj, 229, 1046
383: 
384: 
385: \bibitem[Woosley \& Weaver 1995]{wwo95} Woosley, S. E., \&  Weaver, T. A. 1995,
386: \apjs, 101, 181
387: 
388: 
389: 
390: 
391: \end{thebibliography}
392: 
393: \clearpage
394: 
395: \begin{landscape}
396: \begin{deluxetable}{lcccccccccccc}
397: \scriptsize
398: \tablecaption{Observing log, stellar parameters and sulphur abundances in the sample of metal-poor stars.
399: The 1-$\sigma$ errors in the equivalent widths are computed with the formula by 
400: Cayrel (1988). The
401: upper limits for HD\,140283 and BD-18\arcdeg555 correspond to a 3-$\sigma$ values.
402: \label{tbl-1}}
403: \tablewidth{0pt}
404: \tablehead{
405: \colhead{Target} &
406: \colhead{$V$}    &
407: \colhead{Exposure time (s)} &
408: \colhead{S/N}           &
409: \colhead{EW (m\AA)} &
410: \colhead{EW (m\AA)} &
411: \colhead{$T_{\rm eff}$}  &
412: \colhead{$\log g$}       &
413: \colhead{[Fe/H]}    &
414: \colhead{[S/H]}          &
415: \colhead{[S/Fe]}         &
416: \colhead{Ref.}           \nl
417: & & & & 8694.6 \AA\ & 8693.9 \AA\ &  &  &  &  &  &  \nl
418: }
419: \startdata
420: HD\,2665   &  7.75 & 3200    & 400   & 3.0$\pm$0.6   &  -  & 4990$\pm$66 & 2.50$\pm$0.2  & $-$2.0  & $-$1.31$^{+0.13}_{-0.14}$ & 0.69 & 1\nl
421: HD\,2796   &  8.51 & 5000    & 350   & 2.6$\pm$0.7   &  -  & 4867$\pm$58 & 1.80$\pm$0.2  & $-$2.3  & $-$1.49$^{+0.17}_{-0.18}$ & 0.81 & 1\nl
422: HD\,19445  &  8.05 & 10\,900 & 650   & 2.7$\pm$0.4 &  -  & 5810$\pm$150 & 4.46$\pm$0.2   & $-$1.88 & $-$1.23$^{+0.11}_{-0.12}$ & 0.65 & 2\nl
423: HD\,140283 &  7.24 & 2500    & 300   & $\leq$2.5   &  -  & 5550$\pm$100 & 3.74$\pm$0.2   & $-$2.21 & $\leq -1.23$ & $\leq$0.98 & 2 \nl
424: HD\,157214 &  5.40 & 150     & 350   & 20$\pm$0.7    & 8.0$\pm$0.7 & 5625$\pm$100 & 4.36$\pm$0.2   & $-$0.32 & $-$0.13$\pm$0.09  & 0.19 & 2 \nl
425: HD\,194598 &  8.36 & 5400    & 350   & 9.0$\pm$0.7   & 3.0$\pm$0.7 & 5960$\pm$120 & 4.38$\pm$0.2   & $-$1.06 & $-$0.77$\pm$0.1  & 0.29  & 2 \nl
426: HD\,201889 &  8.04 & 1500    & 250   & 14$\pm$1    & 5.0$\pm$1 & 5615$\pm$100 & 4.24$\pm$0.2      & $-$0.94 & $-$0.37$\pm$0.1  & 0.57 & 2 \nl
427: BD-18\arcdeg555 & 9.34 & 5000    & 300   & $\leq$2.5      &  -   & 4668$\pm$63 & 1.5$\pm$0.3       & $-$2.7  & $\leq -1.3$    &  $\leq$1.4 & 1   \nl
428: \enddata
429: \tablerefs{
430: (1): Alonso et al. (1999), Israelian et al., in preparation,
431: (2): Thevenin and Idiart (1999), Israelian et al. (1998).
432: }
433: \tablenotetext{a}{The  signal-to-noise ratio (S/N) is provided at 8697 \AA.}
434: \end{deluxetable}
435: \end{landscape}
436: 
437: 
438: \clearpage
439: 
440: \begin{deluxetable}{lcccccc}
441: \scriptsize
442: \tablecaption{Reanalysis of sulphur abundances from the samples of 
443: Fran\c{c}ois (1987, 1988). The metallicities were adopted from the
444: references listed in the last column. Typical error in [S/Fe] is of
445: the order of $\pm$0.1 dex.\label{tbl-3}}
446: \tablewidth{0pt}
447: \tablehead{
448: \colhead{Star}           &
449: \colhead{$T_{\rm eff}$}  &
450: \colhead{$\log g$}       &
451: \colhead{[Fe/H]}         &
452: \colhead{[S/H]}          &
453: \colhead{[S/Fe]}         &
454: \colhead{Ref.}           \nl
455: }
456: \startdata
457: HD\,24616   & 4943 & 3.04   & $-$0.88  &  $-$0.51 &    0.37  & 3  \nl
458: HD\,59984   & 5916 & 3.90   & $-$0.52  &  $-$0.70 & $-$0.18  & 1  \nl
459: HD\,63077   & 5752 & 4.17   & $-$0.53  &  $-$0.55 & $-$0.02  & 1  \nl
460: HD\,69897   & 6263 & 4.25   & $-$0.12  &  $-$0.41 & $-$0.29  & 1  \nl
461: HD\,76932   & 5857 & 4.08   & $-$0.79  &  $-$0.64 &    0.15  & 1  \nl
462: HD\,88218   & 5677 & 3.87   & $-$0.42  &  $-$0.29 &    0.13  & 3  \nl
463: HD\,91324   & 6181 & 3.98   & $-$0.23  &  $-$0.52 & $-$0.29  & 2  \nl
464: HD\,94028   & 5923 & 4.28   & $-$1.31  &  $-$1.12 &    0.19  & 1  \nl
465: HD\,102365  & 5603 & 4.40   & $-$0.08  &  $-$0.41 & $-$0.33  & 2  \nl
466: HD\,106516  & 6153 & 4.30   & $-$0.66  &  $-$0.56 &    0.1   & 2  \nl
467: HD\,121384  & 5185 & 3.39   & $-$0.34  &  $-$0.79 & $-$0.35  & 2  \nl
468: HD\,132475  & 5619 & 3.79   & $-$1.37  &  $-$1.13 &    0.24  & 1  \nl
469: HD\,136352  & 5637 & 4.31   & $-$0.21  &  $-$0.47 & $-$0.26  & 2  \nl
470: HD\,148816  & 5831 & 4.12   & $-$0.51  &  $-$0.55 & $-$0.04  & 1  \nl
471: HD\,157089  & 5770 & 4.04   & $-$0.39  &  $-$0.50 & $-$0.11  & 1  \nl
472: HD\,188376  & 5369 & 3.62   & $-$0.08  &  $-$0.30 & $-$0.22  & 3  \nl
473: HD\,193901  & 5668 & 4.50   & $-$0.97  &  $-$0.70 &    0.13  & 1  \nl
474: HD\,201891  & 5841 & 4.25   & $-$0.87  &  $-$0.88 & $-$0.01  & 1  \nl
475: HD\,203608  & 6067 & 4.29   & $-$0.57  &  $-$0.69 & $-$0.12  & 1  \nl
476: HD\,211998  & 5271 & 3.36   & $-$1.25  &  $-$1.40 & $-$0.15  & 1  \nl
477: \enddata
478: \tablerefs{
479: (1): Thevenin \& Idiart (1999),
480: (2): Allende Prieto et al. (1999),
481: (3): Allende Prieto \& Lambert (1999).
482: }
483: \end{deluxetable}
484: 
485: 
486: 
487: 
488: 
489: \clearpage
490: 
491: 
492: \begin{figure}
493: \psfig{figure=Fig1.ps,width=18.0cm,height=18cm,angle=270}
494: \caption[]{Sulphur lines in the UES spectra of five metal-poor stars.}
495: \end{figure}
496: 
497: 
498: 
499: \clearpage
500: 
501: 
502: \begin{figure}
503: \psfig{figure=Fig2.ps,width=18.0cm,height=16cm,angle=270}
504: \caption[]{Synthetic spectra fits to the sulphur lines in the UES 
505: spectra (dots) of HD\,19445  and HD\,2665. The solid lines are
506: best fits to the data corresponding to [S/Fe]=0.65 and 0.7 in 
507: HD\,19445 and HD\,2665, respectively. Thin dotted lines indicate
508: spectral synthesis with [S/Fe] shifted by $\pm$0.15 dex with respect
509: the previous best fit values in each star. A vertical shift of 0.03
510: has been applied to the spectra of HD 2665 for clarity.}
511: \end{figure}
512: 
513: \clearpage
514: 
515: 
516: \begin{figure}
517: \psfig{figure=Fig3.ps,width=19.0cm,height=14cm,angle=270}
518: \caption[]{Sulphur abundances with respect to iron, derived from 
519: the near-IR S\,{\sc i} lines. Unfilled squares are the data from 
520: Fran\c{c}ois (1987, 1988) and filled circles are new measurements.
521: The sun is marked with the standard symbol. Upper limits for two 
522: very metal-poor stars are not plotted.}
523: \end{figure}
524: 
525: 
526: \end{document}
527: 
528: 
529: 
530: