0810.0710/ms.tex
1: \documentclass{emulateapj}
2: 
3: \shorttitle{The highly unusual composition of the  Hercules dSph}
4: \shortauthors{Koch et al.}
5: \slugcomment{Accepted for publication in ApJ Letters}
6: \def\teff{T$_{\mathrm{eff}}$}
7: \def\kms{km\,s$^{-1}$}
8: 
9: \begin{document}
10: 
11: \title{The Highly Unusual Chemical Composition of the Hercules Dwarf Spheroidal 
12: Galaxy\altaffilmark{1}}
13: 
14: \author{Andreas Koch\altaffilmark{2}, Andrew McWilliam\altaffilmark{2}, 
15: Eva K. Grebel\altaffilmark{3}, Daniel B. Zucker\altaffilmark{4}, \& Vasily Belokurov\altaffilmark{4}}
16: \altaffiltext{1}{This paper includes data gathered with the 6.5 meter Magellan Telescopes located at 
17: Las Campanas Observatory, Chile.}
18: \altaffiltext{2}{Observatories of the Carnegie Institution Washington, Pasadena, CA, USA}
19: \altaffiltext{3}{Astronomisches Rechen-Institut, Zentrum f\"ur Astronomie der 
20: Universit\"at  Heidelberg, Germany}
21: \altaffiltext{4}{Institute of Astronomy, Madingley Road, Cambridge, UK}
22: \email{akoch@ociw.edu}
23: 
24: \begin{abstract}
25: We report on the abundance analysis of two red giants in the 
26:  faint Hercules dwarf spheroidal (dSph) galaxy. 
27: These stars show a remarkable deficiency in the neutron-capture elements,  
28: while the hydrostatic $\alpha$-elements (O, Mg) are strongly enhanced. 
29:  %
30: Our data indicate [Ba/Fe] and [Mg/Fe] abundance ratios of $\la$$-2$ dex and $\sim$+0.8 dex, respectively, 
31: with essentially no detection of other n-capture elements.  
32: %
33: In contrast to the only other dSph star with similar  abundance patterns, 
34: Dra~119, which has a very low metallicity at [Fe/H]=$-$2.95 dex, our objects, at 
35: [Fe/H]$\sim$$-2.0$ dex, are only moderately  metal poor. 
36: %
37: The measured ratio of hydrostatic/explosive $\alpha$-elements indicates
38: that high-mass ($\sim35$~M$_{\sun}$) Type II supernovae progenitors are the
39: main, if not only, contributors to the enrichment of this galaxy.  This
40: suggests that star formation and chemical enrichment in the ultrafaint
41: dSphs proceeds stochastically and inhomogeneously on small scales,  or that the IMF was strongly
42: skewed to high mass stars. 
43: %
44: The neutron capture deficiencies and the [Co/Fe] and [Cr/Fe] abundance ratios
45: in our stars are similar to those in the extremely low metallicity Galactic 
46: halo.  This suggests that either our stars are composed mainly of the ejecta from the first, massive,   
47: population~III stars (but at moderately high [Fe/H]), or that SN ejecta
48: in the Hercules galaxy were diluted with $\sim$30 times less hydrogen than
49: typical for extreme metal-poor stars.  
50: \end{abstract}
51: 
52: \keywords{Stars: abundances --- Galaxies: dwarf --- Galaxies: evolution 
53: --- Galaxies: individual (\objectname{Hercules}) --- nuclear reactions, nucleosynthesis, abundances }
54: 
55: \section{Introduction}
56: Chemical abundance patterns in stars reflect the star formation (SF) histories of stellar systems:  
57: whatever mechanisms contributed to the enrichment of the interstellar medium
58: will leave unique imprints in the resulting  
59: chemical abundance ratios. These  allow one not only to trace the predominant 
60: modes of SF and reconstruct galactic enrichment histories, but also to place important 
61: constraints on the nucleosynthetic production sites of the elements. 
62: %
63: In particular, the $\alpha$-elements (e.g., O, Mg, Si, Ca, Ti), produced mainly by massive stars which 
64: end as Type II supernovae (SNe) on short time scales, and the neutron-capture elements (e.g. Y, Sr, Ba, La)  from AGB
65: stars, probe rapid and slow timescales, respectively.
66: %
67: 
68: Knowledge about chemical abundances and the evolutionary histories of 
69: the more luminous dwarf spheroidal (dSph) galaxies is constantly growing (e.g., Shetrone et al. 2003; 
70: Venn et al. 2004; Koch et al. 2008). However,   
71: only little is known about the metal-enrichment of the 
72: ultra-faint, presumably low-mass, systems recently discovered in the Sloan Digital Sky Survey 
73: (SDSS;  Zucker et al. 2006a,b; Belokurov et al. 2006, 2007, 2008; Irwin et al. 2007; Walsh et al. 2007).  
74: Current low-resolution spectroscopy and photometric data
75: indicate that these galaxies are predominantly metal poor 
76: (e.g., Mu\~noz et al. 2006; Belokurov et al. 2007; Kirby et al. 2008). 
77: 
78: The discovery of only a couple of chemically peculiar stars in Galactic dSphs (Fulbright et al. 2004; 
79: Sadakane et al. 2004; Sbordone et al. 2007; Koch et al. 2008) and their similarity to a number of anomalous halo objects (Carney et al. 1997; 
80: Ivans et al.  2003) poses important questions for the general understanding of 
81: galaxy formation and chemical evolution: how common are chemically peculiar stars in the dSphs? 
82: What were the predominant enrichment processes that led to their abundance patterns? 
83: Are peculiar stars found in the Galactic halo related to those in dSphs? 
84: 
85: Hercules (hereafter Her) is one of the faintest ($M_V=-6.6$ mag) 
86: and among the very metal poor ($\sim -2.3$ dex) dSph candidates 
87: discovered in the SDSS (Belokurov et al. 2007). 
88: Its {\em stellar} mass is estimated to be 4--7$\times 10^4$~M$_{\sun}$ 
89: (Martin et al. 2008). 
90: Its extended morphology (Coleman et al. 2007) and a broad RGB 
91: render it an intrinsically peculiar galaxy that may have experienced interactions with  the Milky Way. 
92: An investigation of its chemical properties is thus a natural endeavor. 
93: 
94: In this Letter we present the discovery of unusual chemical abundance patterns in two red giants
95: in Hercules. The elemental abundances were identified in the course of 
96: a broader observational program dedicated to a high-resolution study of Hercules, and the 
97: comprehensive  abundance analysis of the full sample will be presented elsewhere 
98: (Koch et al. 2008, in prep.) 
99: 
100: \section{Observations and Abundance Analysis}
101: Targets were selected from the  Sixth Data Release of the SDSS  (Adelman-McCarthy et al. 2008). 
102: %
103: There is only a limited number of brighter red giants present in Her that are observable within 
104: reasonable integration times (V$\la$19.2 mag)  
105: and non-members were rejected from the sample in real time during the observing run, 
106: based on their highly discrepant velocities (cf. Simon \& Geha 2007).  
107: %
108: Here we present the results  for two of our red giants, which we label Her-2 and Her-3 according 
109: to their brightness (V=18.7 and 19.0 mag). The observations were carried out over two nights on 
110: 2007 July 10--11 with the Magellan Inamori Kyocera Echelle (MIKE) 
111: spectrograph at the 6.5-m Magellan2/Clay Telescope. 
112: %
113: A slit width of 1$\arcsec$ and a CCD binning of 2$\times$1 pixel resulted in 
114: a spectral resolution of 20000, with a full spectral coverage using the blue and red echelles  
115:  of 3350-9400\AA. 
116: The  stars were observed for a total of four  and six hours, respectively. 
117: Our data were reduced with the standard 
118: MIKE pipeline reduction package of Kelson (2003). 
119: %
120: The  signal-to-noise (S/N) ratio of our spectra is 32 per pixel at $\sim$6500\AA, 
121: but  it falls below 10 for the bluer orders. 
122: %
123: The  radial velocities of the two target stars confirm their membership in Her 
124: (Simon \& Geha 2007), but we note that Her-3  is probably a binary, based on its 
125: radial velocity curve (D. Ad\'en 2008, private 
126: communication), although there is no secondary flux detected from the 
127:  companion.
128: Therefore, the companion is most likely very faint and  will 
129: not affect the present analysis. 
130: 
131: The chemical abundances were determined from an equivalent width (EW) analysis, where 
132: we employed $gf$-values from the literature.  
133: Typically, 15--25 \ion{Fe}{1} lines of sufficient strength could be measured. 
134: For most of the other species, only a few absorption features were distinguishable (see 
135: Table~1), which, coupled with the slightly low S/N led to large scatter in some abundance ratios. 
136: Many absorption lines of the heavy elements are too weak to be reliably 
137: measured in our spectra, yielding only upper abundance limits.
138: 
139: Stellar abundances 
140: were computed with the program MOOG (Sneden 1973) using  
141: stellar atmospheres generated from the 
142: Kurucz LTE models\footnote{\url{http://kurucz.harvard.edu}}. 
143: %
144: Given the strong enhancements in the $\alpha$-elements O and Mg in our Her stars 
145: (Sect.~3), we chose to use the $\alpha$-enhanced Kurucz model atmospheres, AODFNEW, 
146: in our analysis.  
147: Furthermore, we adopted photometric ($V$$-$$I$) effective temperatures 
148: for our stellar atmospheres, where colors were transformed from 
149: the SDSS g,r,i system using the prescriptions of Jordi et al. (2006). 
150: A reddening of E(B$-$V)= 0.06 (Schlegel 1998) and the according extinction laws of 
151: Rieke \& Lebovsky (1985) were adopted throughout this work. 
152: %
153: We note that the photometric values agree well with the temperatures derived 
154: from excitation equilibrium.  The physical  gravities were derived from the photometric data, 
155: adopting a distance to Her of 140 kpc (Belokurov et al. 2007) and a stellar mass of 0.8 M$_{\sun}$, 
156: representative of the red giants (see also Koch \& McWilliam 2008).
157: % 
158: The Fe abundances from  the ionized and neutral lines agree marginally within the uncertainties so that 
159: ionization equilibrium holds with the adopted surface gravities. 
160: %
161: As a result, we derive stellar atmosphere parameters (T$_{\rm eff}$, log\,$g$, $\xi$) of 
162: (4270 K, 0.69, 2.5 km\,s$^{-1}$) for Her-2 
163: and (4340 K, 0.91, 2.5 km\,s$^{-1}$) for Her-3, with typical uncertainties of 
164: ($\pm$50 K, $\pm$0.15, $\pm$0.1  km\,s$^{-1}$) . 
165: Our adopted microturbulent velocity values are higher than 
166: typical red giant stars with similar parameters.  This is likely an artifact of the relatively low S/N 
167: of our spectra, due to asymmetry in abundance errors on strong
168: lines (Magain 1984); however, we stress that these high 
169: microturbulent velocity values are appropriate for our spectra. 
170: %
171: Hyperfine splitting was included for the even-Z elements Sc, Mn, Co, and Ba, but is usually 
172: negligible for the Na lines we used in our stars. 
173: % 
174: \section{Abundance results}
175: %
176: %
177: The final chemical abundance ratios derived for the red giants are listed 
178: in Table~1, relative to \ion{Fe}{1} for the neutral species and to \ion{Fe}{2} for ionized 
179: species and for 
180: the [\ion{O}{1}] lines. Solar abundances were taken from Asplund et al. (2005).  
181: Listed in  Table ~1 are 1\,$\sigma$ scatter of our 
182: measurements from individual lines and the numbers of lines employed to derive these
183: abundances. The above uncertainties on the atmospheric parameters translate into 
184: typical total systematic errors of less than 0.1 dex, based on the formalism 
185: of Koch \& McWilliam (2008; their Table 8).
186: %
187: Details on the full data set  and a thorough treatment of all 
188: systematic errors will be presented in a forthcoming paper. 
189: %
190: %
191: \begin{deluxetable}{rrrrrrrrr}
192: \tabletypesize{\scriptsize}
193: \tablecaption{Abundance results}
194: \tablewidth{0pt}
195: \tablehead {\colhead{} & \multicolumn{4}{c}{Her-2} &  \multicolumn{4}{c}{Her-3} \\  
196: \cline{2-4} \cline{6-9}                                       
197: \raisebox{1.5ex}[-1.5ex]{Ion} & \colhead{{[}X/Fe{]}} & \colhead{$\sigma$}& \colhead{N} &
198: \colhead{} & \colhead{{[}X/Fe{]}} & \colhead{$\sigma$}& \colhead{N}}
199: \startdata
200: %
201: {[}Fe/H{]}$_{\rm CaT}$\tablenotemark{a} & $-$2.39 & 0.13 & \nodata & & $-$1.72 & 0.15 & \nodata &  \\
202: {[}Fe\,I/H{]}  & $-$2.02 & 0.20 & 24 & & $-$2.04 & 0.43 & 15  \\
203: {[}Fe\,II/H{]} & $-$1.78 & 0.21 &   3 & & $-$1.84 & 0.41 &   2  \\
204: {[}O\,I{]}  &    1.12 &    0.18 &  2 & & $<$0.89 &    0.08 & 2   \\
205: {Na\,I}  &    0.78 &    0.11 &  2 & &    0.62 &    0.09 & 2   \\
206: {Mg\,I}  &    0.81 &    \nodata &  1 & &    0.77 &  \nodata  & 1   \\
207: {Al\,I}  &   $<$0.75 &  \nodata &  1 & & \nodata & \nodata & \nodata   \\
208: {Si\,I}  & $<$0.58 & \nodata &  1 & &    $<$0.63 & \nodata & 1   \\
209: {Ca\,I}  & $-$0.13 &    0.05 &  7 & &    0.19 &    0.09 & 3   \\
210: {Sc\,II} &    0.04 &    0.17 &  3 & &    0.26 &    0.20 & 2   \\
211: {Ti\,I}  &    0.17 &    0.23 &  9 & &    0.33 &    0.13 & 3   \\
212: {Cr\,I}  &    $-$0.05 &    0.32 &  3 & &    $-$0.14 &    0.04 & 2   \\
213: {Mn\,I}  &    0.23 &    0.08 &  2 & & $-$0.06 & 0.33 & 3  \\
214: {Co\,I}  &    0.53 &    0.10 &  2 & &    0.42 & \nodata & 1   \\
215: {Ni\,I}  &    0.30 &    0.10 & 10 & &    0.12 &    0.15 & 5   \\
216: {Ba\,II} & $<$$-$2.14 & 0.06 &  2 & & $<$$-$2.10 & 0.06 & 2 
217: %							      				   			      			         
218: \enddata
219: \tablenotetext{a}{Metallicity estimate based on the calcium triplet calibration of Rutledge 
220: et al. (1997a,b), on the metallicity scale of Carretta \& Gratton (1997).}
221: \end{deluxetable}
222: 
223: \subsection{Iron}
224: %
225: %
226: At an [Fe/H] of $-2.02$ and $-2.04$  dex, both stars are slightly more metal rich than the sample mean 
227: found from comparison with 
228: globular cluster fiducials (Belokurov et al. 2007) and from low-resolution 
229: data, which  yield a mean metallicity of $-2.3$ or $-2.5$ dex   
230: (Simon \& Geha 2007; Kirby et al. 2008). 
231: The latter work finds an intrinsic metallicity spread of at least 0.5 dex (r.m.s.), which would indicate that our stars
232: are well within the Her metal-rich tail.  In particular, almost all of the ultrafaint dSphs studied to 
233: date show broad metallicity ranges of this order of magnitude. 
234: %
235: In general, a zero-point difference between the CaT ``metallicities''
236: and our [Fe/H] values can be expected due to the low [Ca/Fe] ($-$0.15) found here.   
237: Na\"ively, we expect a zero-point correction to the CaT metallicities based on halo [Ca/Fe] calibrations 
238: of $+$0.50 dex (see also Koch et al. 2008). 
239: This would suggest [Fe/H]=$-$1.8 from correcting the Simon \& Geha (2007)
240:  metallicities.  We can check this result using our MIKE spectra, which 
241: contain the CaT lines; accordingly, we find CaT-based metallicities of [Fe/H]$_{\rm CaT}= -$2.39 and $-$1.72 for Her-2 and 3.  
242: Moreover, our high-resolution [Fe/H] for Her-2 agrees well to within the uncertainties with the  
243: low-dispersion value derived by Kirby et al. (2008) and with the value based on  Str\"omgren photometry
244: (D. Ad\'en; private communication). 
245: %
246: %
247: \subsection{Alpha-  and light elements}
248: %
249: As a glance at Table~1 shows, 
250: the [Ca/Fe] and [Ti/Fe] ratios of the Her stars are at most slightly to moderately enhanced, 
251: which is compatible 
252: with the abundance ratios found in the more luminous dSphs (Fig.~1; Shetrone et al. 2001,2003; 
253: Venn et al. 2004; Sadakane et al.\ 2004; Sbordone et al.\ 2007; Koch et al. 2008). 
254: %
255: %
256: \begin{figure}[ht]
257: \begin{center}
258: \vspace{0.5cm}
259: \includegraphics[angle=0,width=1\hsize]{f1.eps}
260: \end{center}
261: \caption{Hydrostatic (Mg) to explosive (Ca) $\alpha$-element abundance ratio  
262: for our two Her giants (red stars) in comparison with Galactic disk, bulge and halo stars (black dots; Venn et al. 2004, and references therein) 
263: and stars in the luminous Galactic dsphs (blue squares). Highlighted  as green symbols are the 
264: chemically peculiar 
265: dSph star Dra~119 (Fulbright et al. 2004) and the $\alpha$-depleted halo star BD+80$\degr$~245 (Carney et al. 1997). Typical 1$\sigma$ 
266: random error bars for the Her stars are indicated toward the lower left of each panel.}
267: \end{figure}
268: %
269: %
270: 
271: At [Si/Fe]$\la0.6$ dex, the Her stars have abundance ratios similar to  other 
272: metal poor dSph stars from the literature (yet consistent with the typical Galactic halo pattern). 
273: One has to keep in mind, however, that these values are upper limits based on only one line. 
274: Furthermore, the production of Si in massive stars, which are required to 
275: contribute to peculiar abundance patterns in metal-poor stars,  is not yet fully understood (e.g., Cohen et 
276: al. 2007). 
277: 
278: It is striking that both the hydrostatic $\alpha$-elements Mg and O are strongly 
279: enhanced by a factor of about 
280: 10 relative to the Solar abundance ratios found in the 
281: majority of the Local Group dSphs. 
282: %
283: The light element Na is
284: strongly enhanced in the Her stars, with typical [Na/Fe] ratios of 0.6--0.8 dex. 
285: For stars with similar atmosphere parameters to Her-2 and 3, the 
286: calculations of Takeda et al. (2003) show downward non-LTE corrections that
287: become more severe with increasing EW for our Na~I lines at 8183 and 8195\AA .
288: Given the large EWs of our 8183/8195 \AA\ Na I lines, near 160--190 m\AA ,
289: the Takeda et al. results suggest abundance corrections that reduce
290: our LTE values by 0.3 to 0.4 dex.  Thus, if non-LTE effects on Fe are 
291: ignored, the [Na/Fe] ratios in our Her stars are probably closer to 
292: $+$0.3 dex, rather than $\sim$0.7 dex indicated in Table~1.
293: %
294: The weak Al lines are 
295:  not detectable in Her-3, but Al appears to be enhanced in Her-2.  
296: This is well consistent with Al and Na production in massive SNe~II, similar to the $\alpha$-elements 
297: Mg and O  (e.g., Woosley \& Weaver 1995).
298: %
299: \subsection{Neutron-capture elements}
300: Fig.~2 shows the spectral regions around the \ion{Ba}{2} line at  6496\AA. 
301: For comparison, we overplot 
302: a spectrum of a red giant in the Carina dSph, with
303:  similar atmospheric parameters as the Her stars and with  
304: a low [Fe/H] of $-$2.72 dex (Koch et al. 2008). The Carina giant has
305: [Ba/Fe] of $\sim -0.63$ dex and is thus Ba-depleted like the metal poor halo stars. 
306: Yet, a weak Ba absorption feature is discernible. 
307: %
308: %
309: Neither of the Her stars, however,  shows any Ba absorption 
310: above the noise level. 
311: As a secondary test we synthesized the spectral regions around the Ba 6141 and 6496 lines 
312: and verified that essentially all of the visible absorption is due to noise and/or a blend with an Fe line.  
313: As a result, the [Ba/Fe] ratio is compatible with upper limits from the spectral S/N of $\sim -2$ dex in both stars (Fig.~3). 
314: %
315: %
316: \begin{figure}[tb]
317: \begin{center}
318: \includegraphics[angle=0,width=1\hsize]{f2.eps}
319: \end{center}
320: \caption{MIKE spectra of the two Hercules stars 
321: centered at the \ion{Ba}{2} 6496 line. Selected absorption lines are 
322: labelled. Shown for comparison are a UVES spectrum of a metal poor ([Fe/H]=$-2.72$) star in the Carina dSph that 
323: has stellar parameters similar to the Her stars and shows a moderate Ba depletion ([Ba/Fe]$\sim -0.63$) typical 
324: of its iron abundance, degraded to the MIKE resolution. The topmost spectrum is a synthesis
325:  using stellar parameters similar to Her-2 and a strong [Ba/Fe] deficiency of $-$2 dex.}
326: \end{figure}
327: %
328: %
329: %
330: \begin{figure}[tb]
331: \begin{center}
332: \includegraphics[angle=0,width=1\hsize]{f3.eps}
333: \end{center}
334: \caption{Same as Fig.~1 but for Ba abundances.}
335: \end{figure}
336: %
337: 
338: The same holds for every other heavy n-capture element. We do not detect the
339: usually prominent \ion{Sr}{2} line at 4077\AA\ above noise, nor are 
340: Y, La, and Eu measurable above the noise, indicating an extraordinary level of depletion.
341: %
342: %
343: \section{Inhomogeneous chemical enrichment in Hercules}
344: %
345: The only other red giant in a dSph known to date to have strong enhancements in 
346: the light elements O, Mg, Na and Al,  and  a complete n-capture deficiency is Dra~119 
347: (Shetrone et al. 1998; Fulbright et al. 2004). 
348: Similar to our Her stars, Dra~119 is characterized by no significant occurrence of  elements heavier than Ni.
349: Fulbright et al. (2004) argue that the chemical patterns seen in Dra~119 
350: are explicable with SNe II enrichment by high-mass progenitors of at least 20 M$_{\sun}$. 
351: Theoretical yields (e.g., Woosley \& Weaver 1995) indicate that 
352: the high ratio of hydrostatic to explosive $\alpha$ elements (O and Mg, relative
353: to Ca, Ti) found in our Her stars (at [Mg/Ca]= 0.94 and 0.58 dex, respectively), 
354: requires nucleosynthesis by relatively high mass Type II SNe
355: ($\sim$35 M$_{\odot}$).
356: %
357: Moreover, given the strong depletion of Ba in the two Her stars,  SNe events in this mass regime 
358: apparently do not produce any significant amount of n-capture elements.
359: 
360: Given the low luminosity of the Her dSph, we consider an incomplete sampling
361: of the high-mass end of the IMF as the most likely reason for the dominance
362: of material from $\sim$35 M$_{\odot}$ SNe~II (e.g., Carigi \& Hernandez 2008). 
363:  Accordingly, we have
364: performed stochastic chemical evolution experiments to investigate the 
365:  [O/Mg], [Mg/Ca] and [Mg/Fe] ratios, similar to McWilliam \& Searle (1999).  Our
366: calculations employed a Miller-Scalo (1979) IMF with slope of $-$2.56 and
367: element yields from Woosley \& Weaver (1995) for low-metallicity (Z=10$^{-4}$ Z$_{\odot}$) SNe~II 
368: progenitors in the mass range 12 to 40 M$_{\odot}$.  Our calculations indicate 
369: that the observed [Mg/Ca] and [Mg/Fe] ratios in Her are found in less than
370: 1\% of systems with 7 to 12 SNe~II events; approximately 10\% of systems possess
371: the observed [Mg/Ca] and [Mg/Fe] ratios after 1--3, and $\sim$5 SN~II events
372: respectively.  
373: About 10\% of systems show the observed [O/Mg] ratios after 11 SNe II;
374: this reduced to 1\% of systems after 32 SNe II. 
375: Thus, our experiment suggests that the abundance ratios can only be reasonably obtained in systems
376: which experienced {\em fewer than $\sim$11 SNe~II} enrichment events, although it is more
377: likely that only 1--3 events were responsible for the chemical composition 
378: of our stars.
379: 
380: Seemingly at odds with the idea of small numbers of massive SNe~II is the 
381: metallicity, at [Fe/H]=$\sim$$-$2.0 dex;  comparable abundance patterns 
382: in the Galactic halo stars are only found towards the very metal poor regime 
383: (e.g., McWilliam \& Searle 1999;  Cohen et al. 2007. Dra~119, with 
384: a comparably high [Mg/Ca] ratio has a very low [Fe/H] of almost $-3$ dex.  
385: %
386: Since Audouze \& Silk 
387: (1995) argue that most individual SNe produce [Fe/H]$\sim$$-$4.0 dex, we might statistically 
388: expect that $\sim$100 SNe contributed to the composition of our Her stars.
389: %
390: Other notable abundance anomalies in our Her stars include the extremely low
391: [Ba/Fe] and enhanced [Co/Fe] ratios.  Remarkably, low [Ba/Fe], high [Co/Fe] and 
392: low [Cr/Fe] ratios 
393: are typical of halo stars with [Fe/H]$\sim$$-$3.5 (McWilliam et al. 1995). It is, however, possible, that 
394: unusual [Co/Cr] abundance ratios may be partly due to NLTE effects. 
395: Overall, it is as if the processes typical of  the extremely low metallicity 
396: Galactic halo  occurred in the Her dwarf at [Fe/H]=$-$2.0 dex. 
397: We note, however, that the high [Mg/Ca] ratios are  inherent to the Her stars and not found 
398: in the metal poor halo stars.  
399: 
400: It is conceivable that our Her stars formed from SN ejecta  
401: diluted with a much smaller amount of pristine gas than typically occurred in the
402: Galactic halo.  
403:  If the Her stars  in fact  resemble  more the very metal poor 
404: Galactic halo stars, then their unusual composition could
405: reasonably be due to one, or at most  a few, SN events. 
406: %
407: This scenario requires that high mass SNe~II produce the enhanced [Co/Fe]
408: and deficient [Cr/Fe] values seen in our Her stars (typical at extremely
409: low metallicity).  While this scenario might explain our observed Her abundances,
410:  it leaves unanswered the origin of the enhanced [Co/Cr] ratios in found in 
411:  extremely metal-poor stars of the Galactic halo.
412: 
413: A viable interpretation is that our Her stars have an unusually high
414: abundance of nearly pure population III material, produced and ejected by
415: the first stars.  These  first stars are thought to be dominated by zero-metallicity high
416: mass SNe~II (e.g. Yoshida et al. 2008), which would naturally
417: explain the high [O,Mg/Fe] ratios.  This scenario would also easily explain 
418: the high [Co/Fe] and low [Cr/Fe] values, if the unusual trends of these two 
419: ratios at low [Fe/H] are due to population III SN ejecta, as suggested by
420: McWilliam  (1997).   While this idea removes the
421: aforementioned difficulties associated with the Co and Cr abundances, the
422: [Mg/Ca] ratios are unlike halo stars near [Fe/H]=$-$4.
423: This scenario leads to the question of why the Her dwarf has a concentration
424: of primordial material $\sim$30 times that of 
425: the low-metallicity halo.
426: %
427: Perhaps it is related to Her's presumably low 
428: baryonic mass, which may have formed only one relatively low-mass
429: population III star that ejected a large amount of population III Fe-peak 
430: material. 
431: 
432: Whatever the cause, it appears
433: that stochastic chemical evolution might reasonably explain the unusual
434: abundance patterns in the Her dwarf galaxy 
435: (see also Marcolini et al.\ 2006, 2008).
436: %
437: If true,  this would indicate that the
438:  chemical enrichment  proceeds very differently 
439: in very low-mass environments (see also Koch et al. 2008),  and that  low
440: luminosity dwarf galaxies would be useful for directly measuring the abundance
441: yields of low-metallicity Type II SNe.
442: 
443: 
444: \acknowledgments
445: %
446: We gratefully acknowledge  Mark Wilkinson for help with preparing the observations  
447: and  Andrea Marcolini and Daisuke Kawata  for very  helpful comments on an early version of this Letter.
448: %
449: \begin{thebibliography}{}
450: %
451: \bibitem[Adelman-McCarthy et al. 2008]{DR6} Adelman-McCarthy, J.K. et al. 2008, ApJS, 175, 297
452: %
453: \bibitem[Asplund et al.(2005)]{2005ASPC..336...25A} Asplund, M., Grevesse, N., \& Sauval, A.~J.\ 2005, 
454:  in ``Cosmic Abundances as Records of Stellar Evolution and Nucleosynthesis'', ASP Conf. Ser., 
455:  336, 25 
456: %
457: \bibitem[Audouze 
458: \& Silk(1995)]{1995ApJ...451L..49A} Audouze, J., \& Silk, J.\ 1995, \apjl, 451, L49 
459: %
460: \bibitem[Belokurov et al.(2006)]{2006ApJ...647L.111B} Belokurov, V., et  
461:  al.\ 2006, \apjl, 647, L111 
462: %
463: \bibitem[Belokurov et al.(2007)]{2007ApJ...654..897B} Belokurov, V., et 
464: al.\ 2007, \apj, 654, 897 
465: %
466: \bibitem[Belokurov et al.(2008)]{2008arXiv0807.2831B} Belokurov, V., et al.\ 2008, ArXiv e-prints, 807, arXiv:
467: 0807.2831 
468: %
469: \bibitem[Carney et al.(1997)]{1997AJ....114..363C} Carney, B.~W., Wright, 
470: J.~S., Sneden, C., Laird, J.~B., Aguilar, L.~A., 
471: \& Latham, D.~W.\ 1997, \aj, 114, 363 
472: %
473: \bibitem[Carigi 
474: \& Hernandez(2008)]{2008arXiv0802.1203C} Carigi, L., \& Hernandez, X.\ 2008, ArXiv e-prints, 802, arXiv:0802.1203 
475: %
476: \bibitem[Carretta & Gratton 1997]{Carretta1997} Carretta, E., \& Gratton, R. 1997, A\&AS, 121, 95
477: %
478: \bibitem[Cohen et al.(2007)]{2007ApJ...659..161L} 
479: Cohen, J. G., McWilliam, A., Christlieb, N., Shectman, S., Thompson, I., 
480: Melendez, J., Wisotzki, L., \& Reimers, D. 2007, ApJ, 659, L161
481: %
482: \bibitem[Coleman et al.(2007)]{2007ApJ...668L..43C} Coleman, M.~G., et al.\ 
483: 2007, \apjl, 668, L43 
484: %
485: \bibitem[Fulbright et al.(2004)]{2004ApJ...612..447F} Fulbright, J.~P., 
486: Rich, R.~M., \& Castro, S.\ 2004, \apj, 612, 447 
487: %
488: \bibitem[Irwin et al.(2007)]{2007ApJ...656L..13I} Irwin, M.~J., et al.\ 
489: 2007, \apjl, 656, L13 
490: %
491: \bibitem[Ivans et al.(2003)]{2003ApJ...592..906I} Ivans, I.~I., Sneden, C., 
492: James, C.~R., Preston, G.~W., Fulbright, J.~P., H{\"o}flich, P.~A., Carney, 
493: B.~W., \& Wheeler, J.~C.\ 2003, \apj, 592, 906 
494: %
495: \bibitem[Jordi et al.(2006)]{2006A&A...460..339J} Jordi, K., Grebel, E.~K., \& Ammon, K.\ 2006, \aap, 460, 
496: 339 
497: %
498: \bibitem[Kelson et al.(2003)]{Kelson03} Kelson, D.D. 2003, \pasp, 115, 688
499: %
500: \\bibitem[Kirby et al.(2008)]{2008ApJ...685L..43K} Kirby, E.~N., Simon, 
501: J.~D., Geha, M., Guhathakurta, P., \& Frebel, A.\ 2008, \apjl, 685, L43 
502: %
503: \bibitem[Koch et al.(2008)]{2008AJ....135.1580K} Koch, A., Grebel, E.~K., 
504: Gilmore, G.~F., Wyse, R.~F.~G., Kleyna, J.~T., Harbeck, D.~R., Wilkinson, 
505: M.~I., \& Wyn Evans, N.\ 2008, \aj, 135, 1580 
506: %
507: \bibitem[Koch  \& McWilliam(2008)]{2008AJ....135.1551K} Koch, A., \& McWilliam, A.\ 2008, \aj, 135, 
508: 1551 
509: %
510: \bibitem[Magain(1984)]{1984A&A...134..189M} Magain, P.\ 1984, \aap, 134, 189 
511: %
512: \bibitem[Marcolini et al.(2006)]{2006MNRAS.371..643M} Marcolini, A., 
513: D'Ercole, A., Brighenti, F., \& Recchi, S.\ 2006, \mnras, 371, 643 
514: %
515: \bibitem[Marcolini et al.(2008)]{2008MNRAS.386.2173M} Marcolini, A., 
516: D'Ercole, A., Battaglia, G., \& Gibson, B.~K.\ 2008, \mnras, 386, 2173 
517: %
518: \bibitem[Martin et al.(2008)]{2008ApJ...684.1075M} Martin, N.~F., de Jong, 
519: J.~T.~A., \& Rix, H.-W.\ 2008, \apj, 684, 1075 
520: %
521: \bibitem[McWilliam et al.(1995)]{1995AJ....109.2757M} McWilliam, A., 
522: Preston, G.~W., Sneden, C., \& Searle, L.\ 1995, \aj, 109, 2757 
523: %
524: \bibitem[McWilliam(1997)]{1997ARA&A..35..503M} McWilliam, A.\ 1997, \araa, 35, 503 
525: %
526: \bibitem[McWilliam \& Searle(1999)]{1999Ap&SS.265..133M} McWilliam, A., \& Searle, L.\ 1999, \apss, 265, 133 
527: %
528: \bibitem[Miller \& Scalo(1979)]{1979ApJS...41..513M} Miller, G.~E., \& Scalo, J.~M.\ 1979, \apjs, 41, 513 
529: %
530: \bibitem[Mu{\~n}oz et al.(2006)]{2006ApJ...650L..51M} Mu{\~n}oz, R.~R., 
531: Carlin, J.~L., Frinchaboy, P.~M., Nidever, D.~L., Majewski, S.~R., 
532: \& Patterson, R.~J.\ 2006, \apjl, 650, L51 
533: %
534: \bibitem[Prochaska et al. 2002]{proch02} Prochaska, J.~X., Naumov, S.~O., Carney, B.~W., 
535: McWilliam, A., \& Wolfe, A.~M. 2000, AJ, 120, 2513
536: %
537: \bibitem[Rieke 1985]{RL85} Rieke, G.~H., \& Lebovsky, M.~J. 1985, ApJ, 288, 618
538: %
539: \bibitem[Rutledge et al. 1997a]{Rutledge1997a} Rutledge, G.A.,  Hesser, J.E., Stetson, P.B., Mateo, M., Simard, L., Bolte, M., Friel, E.D., \& Copin, Y. 1997a, PASP, 109, 883 
540: %
541: \bibitem[Rutledge et al. 1997b]{Rutledge1997b} Rutledge, G.A.,  Hesser, J.E., Stetson, P.B. 1997b, PASP, 109, 907
542: %
543: \bibitem[Sadakane et al.(2004)]{2004PASJ...56.1041S} Sadakane, K., Arimoto, 
544: N., Ikuta, C., Aoki, W., Jablonka, P., 
545: \& Tajitsu, A.\ 2004, \pasj, 56, 1041 
546: %
547: \bibitem[Sbordone et 
548: al.(2007)]{2007A&A...465..815S} Sbordone, L., Bonifacio, P., Buonanno, R., Marconi, G., Monaco, L., \& Zaggia, S.\ 2007, \aap, 465, 815 
549: %
550: \bibitem[Schlegel et al. 1998]{Schlegel1998} Schlegel, D.J., Finkbeiner, D.P., \& Davis, M. 1998, ApJ, 
551: 500, 525
552: %
553: \bibitem[Shetrone et al.(1998)]{1998AJ....115.1888S} Shetrone, M.~D., 
554: Bolte, M., \& Stetson, P.~B.\ 1998, \aj, 115, 1888 
555: %
556: \bibitem[Shetrone et al.(2001)]{2001ApJ...548..592S} Shetrone, M.~D., 
557: C{\^o}t{\'e}, P., \& Sargent, W.~L.~W.\ 2001, \apj, 548, 592 
558: %
559: \bibitem[Shetrone et al. 2003]{Shetrone2003} Shetrone, M.~D., Venn, K.~A., Tolstoy, E., Primas, F., 
560: Hill, V., \& Kaufer, A. 2003, AJ, 125, 684 (S03)
561: %
562: \bibitem[Simon \& Geha(2007)]{2007arXiv0706.0516S} Simon, J.~D., \& Geha, 
563: M.\ 2007, ApJ, 670, 313
564: %
565: \bibitem[Sneden 1973]{sned73} Sneden, C. 1973, ApJ, 184, 839
566: %
567: \bibitem[Takeda et al.(2003)]{2003ChJAA...3..316T} Takeda, Y., Zhao, G., 
568: Takada-Hidai, M., Chen, Y.-Q., Saito, Y.-J., 
569: \& Zhang, H.-W.\ 2003, Chinese Journal of Astronomy and Astrophysics, 3, 316 
570: %
571: \bibitem[Venn et al. 2004]{Venn2004} Venn, K.~A., Irwin, M.~I., Shetrone, M.~D., Tout, C.~A., Hill, V., 
572: \& Tolstoy, E. 2004, AJ, 128, 1177
573: %
574: \bibitem[Walsh et al.(2007)]{2007ApJ...662L..83W} Walsh, S.~M., Jerjen, H., 
575: \& Willman, B.\ 2007, \apjl, 662, L83 
576: %
577: \bibitem[Woosley \& Weaver (1995)]{1995ApJS...662L..83W} Woosley, S.E., \& Weaver, T.A.\ 1995, 
578: \apjs, 101, 181 
579: %
580: \bibitem[Yoshida et al.(2008)]{2008Sci...321..669Y} Yoshida, N., Omukai, K., \& Hernquist, L.\ 2008, 
581: Science, 321, 669
582: %
583: \bibitem[Zucker et al.(2006)]{2006ApJ...650L..41Z} Zucker, D.~B., et al.\ 
584: 2006, \apjl, 650, L41 
585: %
586: \bibitem[Zucker et al.(2006)]{2006ApJ...643L.103Z} Zucker, D.~B., et al.\ 
587: 2006, \apjl, 643, L103 
588: %
589: \end{thebibliography}
590: %
591: \end{document}
592: 
593: