1: 

2: %\documentclass[12pt,preprint]{aastex}

3: \documentclass{emulateapj}

4: 

5: %\usepackage{lscape}

6: 

7: \setcounter{totalnumber}{100}

8: 

9: \newcommand{\eps}[1]{\mbox{log~$\epsilon$(#1)}}

10: \newcommand\iso[2]{$^{\rm #1}$#2}

11: \newcommand\wave[1]{\mbox{#1\,\AA}}

12: \def\deg{{$^{\circ}$}}

13: \def\eg{\mbox{\it e.g.}}

14: \def\etal{\mbox{et al.}}

15: \def\ie{\mbox{i.e.}}

16: \def\kmsec{\mbox{km~s$^{\rm -1}$}}

17: \def\logg{\mbox{log~{\it g}}}

18: \def\Msun{\mbox{$M_{\odot}$}}

19: \def\teff{\mbox{$T_{\rm eff}$}}

20: \def\vturb{\mbox{$v_{\rm t}$}}

21: \def\rpro{\mbox{$r$-process}}

22: \def\spro{\mbox{$s$-process}}

23: \def\ncap{\mbox{$n$-capture}}

24: \def\starname{\mbox{CS~xxxxx--yyy}}

25: 

26: 

27: \shorttitle{Stellar Composition of Milky Way Halos}

28: \shortauthors{I.~U.\ Roederer}

29: 

30: 

31: \begin{document}

32: 

33: 

34: \title{Chemical Inhomogeneities in the Milky Way Stellar Halo}

35: 

36: 

37: \author{Ian U.\ Roederer}

38: \affil{Department of Astronomy, University of Texas at Austin \\

39: 1 University Station C1400, Austin, TX 78712-0259}

40: \email{iur@astro.as.utexas.edu}

41: 

42: 

43: \begin{abstract}

44: 

45: 

46: We have compiled a sample of 699 stars from the recent literature

47: with detailed chemical abundance information

48: (spanning $-4.2 \lesssim$~[Fe/H]~$\lesssim +0.3$), and we compute their

49: space velocities and Galactic orbital parameters.

50: We identify members of the inner and outer stellar halo populations 

51: in our sample based only on their kinematic properties 

52: and then compare the abundance ratios of these populations

53: as a function of [Fe/H].

54: In the metallicity range where the two populations overlap

55: ($-2.5 \lesssim$~[Fe/H]~$\lesssim -1.5$),

56: the mean [Mg/Fe] of the outer halo is lower than 

57: the inner halo by $\sim$~0.1~dex.

58: For [Ni/Fe] and [Ba/Fe], 

59: the star-to-star abundance scatter of the inner halo is consistently

60: smaller than in the outer halo.

61: The [Na/Fe], [Y/Fe], [Ca/Fe], and [Ti/Fe] ratios of both populations 

62: show similar means and levels of scatter.

63: Our inner halo population is chemically homogeneous, 

64: suggesting that a significant fraction

65: of the Milky Way stellar halo originated from a well-mixed ISM.

66: In contrast, our outer halo population is chemically diverse,

67: suggesting that another significant fraction

68: of the Milky Way stellar halo formed in remote regions where

69: chemical enrichment was dominated by local supernova events.

70: We find no abundance trends with maximum radial distance from

71: the Galactic center or maximum vertical distance from the Galactic disk.

72: We also find no common kinematic signature for groups of

73: metal-poor stars with peculiar abundance patters, such as 

74: the $\alpha$-poor stars or 

75: stars showing unique neutron-capture enrichment patterns.

76: Several stars and dSph systems with unique abundance patterns 

77: spend the majority of their time in the distant

78: regions of the Milky Way stellar halo, suggesting that the 

79: true outer halo of the Galaxy may have little resemblance

80: to the local stellar halo.

81: 

82: 

83: \end{abstract}

84: 

85: 

86: \keywords{

87: Galaxy: formation ---

88: Galaxy: halo ---

89: globular clusters: general ---

90: nuclear reactions, nucleosynthesis, abundances ---

91: stars: abundances ---

92: stars: Population~II

93: }

94: 

95: 

96: 

97: \section{Introduction}

98: \label{introduction}

99: 

100: 

101: The nucleosynthesis reactions necessary to 

102: produce metals in stars were realized nearly half a century ago

103: \citep{burbidge57,seeger65,fowler67,wagoner67}, yet the challenge to

104: interpret the wide variety of nucleosynthetic signatures 

105: observed in different stellar populations today remains as

106: intriguing as ever.

107: Metal-poor stellar populations should contain

108: recycled stellar material from fewer generations of stars 

109: than metal-rich populations, making interpretation of their

110: chemical enrichment history---\textit{in principle}---simpler.

111: Thanks to numerous large surveys over the last four decades

112: designed to identify metal-poor stars (see review by \citealt{beers05}),

113: the tally of known metal-poor stars now stretches well into the thousands.

114: Concurrently, great advances have been made in the analysis and

115: interpretation of the chemical signatures and 

116: enrichment histories revealed by stellar spectra 

117: (e.g., \citealt{audouze76}, 

118: \citealt{kraft79}, \citealt{wheeler89}, 

119: \citealt{mcwilliam97}, \citealt{gratton04}, 

120: \citealt{beers05}, and \citealt{sneden08}).

121: Furthermore, careful laboratory analysis has improved 

122: our knowledge of the relevant atomic data necessary to make

123: accurate and detailed records of the chemical composition of the

124: atmospheres of metal-poor stars.

125: The confluence of advances in each of these fields

126: has built upon the foundation of stellar nucleosynthesis

127: to greatly increase our

128: understanding of the earliest generations of stars, 

129: the chemical evolution of the various components of the Milky Way Galaxy, 

130: and the formation process of the Galaxy, to name just a few successes.

131: 

132: Correlations between metal enrichment in stellar populations 

133: and the kinematic properties of these stars have been known for

134: some time now \citep[e.g.,][]{eggen62,wallerstein62}, 

135: and through the years these relationships 

136: have been fleshed out in increasing detail.

137: Most studies of the formation of the stellar halo of the Galaxy 

138: have employed limited chemical data (e.g., [Fe/H]\footnote{

139: We adopt the usual spectroscopic notations that

140: [A/B]~$\equiv$ log$_{10}$(N$_{\rm A}$/N$_{\rm B}$)$_{\star}$~--

141: log$_{10}$(N$_{\rm A}$/N$_{\rm B}$)$_{\odot}$ and

142: log~$\epsilon$(A)~$\equiv$ 

143: log$_{10}$(N$_{\rm A}$/N$_{\rm H}$)~$+$~12.00

144: for elements A and B.}) to accompany the kinematic data.

145: This is sufficient to study the metallicity distribution

146: function (MDF) of the halo 

147: (e.g., \citealt{hartwick76,ryan91b,ivezic08,schorck08}),

148: formation and age of the halo 

149: (e.g., \citealt{eggen62,searle78,sandage86,wyse88,

150: gilmore89,preston91,ryan91a,majewski92,norris94,carney96,

151: sommerlarsen97,chiba98,chiba00,carollo07,bell08,miceli08,morrison08}),

152: or for investigating stellar streams and halo substructure

153: (e.g., \citealt{majewski96,helmi99,chiba00,gilmore02,kinman07}).

154: Iron (Fe)---or, in some cases, calcium (Ca)---alone 

155: is less useful for studies of the 

156: chemical enrichment of the halo, which can examine, e.g., 

157: supernova (SN) models and rates, 

158: stellar binary fractions, 

159: mixing processes in the interstellar medium (ISM) of the halo,

160: the Galactic potential, Galactic structure, 

161: and relationships between various substructures

162: (e.g., globular clusters, Local Group dwarf spheroidal [dSph]

163: systems, stellar streams).

164: It was not until recently that kinematic and detailed chemical data 

165: for halo stars were analyzed together

166: \citep[e.g.,][]{gratton03,simmerer04,venn04,pritzl05,font06,geisler07}.

167: 

168: Three substantial advances have been made in the short period of time since 

169: \citet{venn04} and \citet{pritzl05} performed

170: a detailed chemical comparison between field stars in the halo and

171: nearby dSph systems and globular clusters.

172: First, echelle spectrographs on multiple 

173: large (6--10~m class) telescopes have enabled

174: investigators to carry out detailed abundance analyses of large numbers

175: of (often faint) metal poor stars, including many stars with

176: [Fe/H]~$<-3.0$.

177: Second, investigators at the US Naval Observatory 

178: have released revised proper motion catalogs based on

179: longer time baselines, improved quality of astrographs, 

180: and better techniques for digitizing 

181: photographic plates \citep[e.g.,][]{zacharias04a}.

182: Improved proper motions are given in these catalogs for many of the

183: metal-poor stars investigated in recent years, 

184: enabling us to derive their full space motions through the Galaxy.

185: Finally, large numbers ($\sim$~few~$\times$~10$^{4}$) of calibration

186: stars with known stellar parameters and metallicities have been

187: observed as part of the Sloan Digital Sky Survey (SDSS) and 

188: Sloan Extension for Galactic Understanding and Exploration (SEGUE)

189: projects \citep[e.g.,][]{newberg03,allendeprieto07,lee07}.

190: This has enabled the previously known kinematic properties

191: of nearby members of the so-called inner and outer stellar halos 

192: to be assessed with a new level of detail \citep{carollo07}.

193: These advances insist on a fresh reanalysis of the existing data.

194: The goal of the present study is to interpret the wealth of

195: recent high-resolution abundance analyses of metal-poor stars

196: in light of the most recent kinematic knowledge of these two

197: major components of the Galactic halo.

198: 

199: 

200: \section{Abundance Data from the Literature}

201: \label{litdata}

202: 

203: 

204: \citet{venn04} compiled from existing literature 

205: a large sample of stellar abundances and

206: \textit{UVW} kinematic data, when available, to compare the 

207: chemical enrichment patterns of metal-poor Galactic halo stars with

208: present-day dSph systems.  

209: We adopt their data from the high-resolution abundance analyses of

210: \citet{edvardsson93},

211: \citet{nissen97}, 

212: \citet{hanson98}, 

213: \citet{prochaska00}, 

214: \citet{fulbright00,fulbright02},

215: \citet{stephens02}, 

216: \citet{bensby03}, and 

217: \citet{reddy03}.

218: The Milky Way thin and thick disks are represented in this sample

219: along with the halo, totaling 620 stars.

220: 

221: We supplement this sample with abundances derived from more

222: recent high-resolution analyses of metal-poor halo stars 

223: or older studies that did not include any kinematic analysis.

224: We add another 309 stars from the analyses of 

225: \citet{mcwilliam95a,mcwilliam95b},

226: \citet{hill02},

227: \citet{ivans03},

228: \citet{cayrel04},

229: \citet{honda04a,honda04b},

230: \citet{barklem05},

231: \citet{honda06},

232: \citet{francois07}, and

233: \citet{lai08}.

234: For stars without sufficient kinematic information to compute 

235: \textit{UVW} velocities and Galactic orbital parameters

236: (i.e., only the stellar radial velocity is published), we obtain 

237: proper motions from the catalogs listed in \S~\ref{uvw}.

238: Many of these stars are not in the

239: Hipparcos Catalog \citep{perryman97}, and for these stars we derive a 

240: photometric parallax (see \S~\ref{distances}).

241: We also require that the total proper motion 

242: is greater than 2.5 times its error, 

243: otherwise we discard the star from further kinematic analysis.

244: We insert this requirement to derive accurate

245: space velocities, yet unfortunately this will also bias our 

246: sample against stars with small proper motions;

247: see \S~\ref{caution} for closer analysis of this point.

248: This requirement culls our sample by 230 stars. 

249: Our final sample of stars is given in Table~\ref{startab}

250: along with their adopted [Fe/H] values.

251: 

252: This large dataset will allow us to probe abundance trends as a function

253: of stellar kinematics.

254: Because we mix data from different sources,

255: systematic offsets in the abundance ratios arise 

256: from to a variety of factors, including

257: (1) different spectral resolution and signal-to-noise (S/N) ratios 

258: of the data themselves, leading to potential blending of lines

259: in crowded spectral regions, 

260: since high resolution is desirable to more-fully resolve the absorption

261: line profiles yet such high resolution and high S/N 

262: is impractical for studies of very faint stars;

263: (2) different sets of absorption lines and transition probabilities

264: used in each study;

265: (3) different methods employed to determine stellar parameters;

266: (4) differences in the structure of the model atmospheres themselves;

267: and so on.

268: These sources of systematic error should

269: be no larger than 0.1--0.2~dex in [X/Fe],

270: which does limit our ability to detect subtle chemical signatures; 

271: gross trends should still be identified reliably.

272: At any given [Fe/H], our sample is comprised of stars from a variety

273: of studies, and the characteristic scatters of the inner and outer 

274: halo populations are distinct beyond this level of systematic scatter.

275: 

276: 

277: \section{Kinematic Data}

278: \label{kinematics}

279: 

280: 

281: \subsection{Distance Estimates}

282: \label{distances}

283: 

284: 

285: Hipparcos \citep{perryman97} geometric parallaxes are the preferred method

286: for deriving accurate distances for our sample (using the 

287: new reduction of the Hipparcos data described in \citealt{vanleeuwen07}).

288: For stars not in the Hipparcos catalog, we derive distance estimates

289: from the photometric parallax

290: by comparing the observed $V$ magnitude with a predicted absolute

291: magnitude $M_{V}$, derived from the dereddened $V-K$ color.\footnote{

292: This method was only used to derive distances to stars with

293: [Fe/H]~$< -1.0$.}

294: To establish the relationship between color and absolute magnitude, 

295: we use the $Y^{2}$ set of theoretical isochrones \citep{demarque04}.

296: We select the appropriate isochrone for each star, 

297: assuming [$\alpha$/Fe]~$=+0.3$ for all stars and 

298: choosing a metallicity close to the individual stellar [Fe/H].

299: We find that the choice of (old) age has negligible effect on our

300: derived photometric distances (comparing between 10, 11, 12, and 13~Gyr

301: isochrones), so we adopt an age of 12~Gyr for all stars.

302: To break the degeneracy of $M_{V}$ with $V-K$ (since any given isochrone

303: is not single-valued in $M_{V}$ for each value of $V-K$),

304: we establish the evolutionary state of each star based on the

305: surface gravity estimate given in the literature

306: (``dwarfs'': $\log g \geq 4.0$; ``giants'': $\log g < 4.0$).

307: We adopt $V$ magnitudes from SIMBAD if not given in the references

308: in \S~\ref{litdata}, $K$ magnitudes from 2MASS\footnote{

309: This publication makes use of data products from the Two Micron 

310: All Sky Survey, which is a joint project of the University of 

311: Massachusetts and the Infrared Processing and Analysis 

312: Center/California Institute of Technology, funded by the 

313: National Aeronautics and Space Administration and the 

314: National Science Foundation.} \citep{cutri03},

315: reddening estimates from the dust maps of \citet{schlegel98}

316: and the reddening laws of \citet{mccall04}:

317: $R_{V} = 3.070 \times E(B-V)$ and $R_{V-K} = 2.727 \times E(B-V)$.

318: We use the prescription given by \citet{bonifacio00b} to reduce

319: the \citet{schlegel98} estimates for $E(B-V) > 0.10$:

320: $E(B-V)_{\rm revised} = 0.10 + 0.65[E(B-V)_{\rm S98} - 0.10]$.

321: 

322: Figure~\ref{distanceplot} compares our photometric and geometric distances

323: for stars in \citet{venn04}.

324: In order to assess the reliability of our photometric distances alone,

325: we attempt to minimize the effect of other influences on the result

326: by restricting this comparison to stars with

327: $E(B-V) < 0.30$ and [Fe/H]~$< -1.0$.

328: The difference in the distance computed by the two methods is shown 

329: as a function of $V-K$ color, [Fe/H] metallicity,

330: and Hipparcos parallax.

331: While it is apparent that there

332: is poor agreement for the distance estimates for our giants, there exists

333: no trend between the differences and color or metallicity.

334: There is, however, a very clear correlation 

335: between these differences and the Hipparcos parallax.

336: In other words, 

337: the agreement is very good for stars with reliable Hipparcos

338: parallax measurements, but this correlation breaks down for stars with 

339: $\pi_{\rm Hipparcos} \lesssim 5$~mas~yr$^{-1}$. 

340: (cf.\ Figure~1 of \citealt{chiba98}, 

341: \S~3.3.5 and Figure~3 of \citealt{beers00}).

342: This preferentially affects distant stars (giants, in our sample)

343: that have parallax measurement uncertainties comparable to the

344: size of the parallaxes themselves.

345: The vast majority of stars with $\pi_{\rm Hipparcos} > 5$~mas~yr$^{-1}$ 

346: are dwarfs, but the few giants with large parallaxes exhibit generally

347: good agreement between the two distance estimates.

348: We also show the relative differences as a function of distance

349: in the bottom panel of Figure~\ref{distanceplot}.

350: The scatter in the giants ($\sigma = 0.59$, 7 stars)

351: is about twice as large as the scatter in the dwarfs 

352: ($\sigma = 0.29$, 28 stars), which is to be expected since

353: a given uncertainty in $(V-K)_{0}$ will translate to a 

354: relatively larger uncertainty in $M_{V}$ for stars on the 

355: giant branch than for stars on the main sequence or near the turn-off.

356: As we proceed, we calculate geometric distances 

357: if the Hipparcos parallax is greater than 5 times its uncertainty, 

358: otherwise we estimate distance from the photometric parallax.

359: These distances are reported in Table~\ref{startab}.

360: 

361: \begin{figure}

362: \epsscale{1.15}

363: \plotone{f01.eps}

364: \caption{

365: \label{distanceplot}

366: Comparison of distances derived from Hipparcos parallax 

367: measurements and photometric parallaxes, as a function of

368: ($V-K$)$_{0}$, [Fe/H], and Hipparcos parallax.

369: The bottom panel shows the relative differences as a function

370: of distance (as derived from the Hipparcos parallax) for 

371: stars with $\pi_{\rm Hipparcos} > 5$~mas~yr$^{-1}$.

372: Open triangles represent our dwarf sample ($\log g \geq 4.0$) 

373: and filled squares represent our giant sample ($\log g < 4.0$).

374: }

375: \end{figure}

376: 

377: 

378: \subsection{Space Velocities}

379: \label{uvw}

380: 

381: 

382: We calculate $UVW$ space velocities for the stars in our sample 

383: from measurements of the radial velocity, proper motion, and distance

384: to each star.

385: We adopt a left-handed coordinate system, where

386: positive $U$ denotes velocity towards the Galactic anti-center

387: ($\ell = 180^{\circ}$, $b = 0^{\circ}$),

388: positive $V$ denotes velocity parallel to the direction of rotation

389: of the Local Standard of Rest (LSR; $\ell = 90^{\circ}$, $b = 0^{\circ}$),

390: and positive $W$ denotes velocity perpendicular to the plane of the

391: disk of the galaxy ($b = +90^{\circ}$).

392: We correct for the motion of the Sun with respect to the LSR, adopting

393: ($U_{\rm LSR}$, $V_{\rm LSR}$, $W_{\rm LSR}$)$_{\odot}=$~($-9$, $12$, $7$) 

394: \citep{mihalas81}, 

395: where all velocities are measured in km\,s$^{-1}$.

396: We adopt $V_{\rm LSR} = 220$\,km\,s$^{-1}$ \citep{kerr86} and 

397: report the $V$ component in a Galactic reference frame (i.e., 

398: where the Sun's motion is described by [$U$, $V$, $W$] $=$

399: [$-9$, $232$, $7$]).

400: Radial velocities were taken from the literature references

401: described in \S~\ref{litdata}.

402: Proper motions were taken from the second data

403: release of the USNO CCD Astrograph Catalog (UCAC2; 

404: \citealt{zacharias04a,urban04})

405: or the USNO-B astrometry catalog\footnote{

406: Accessible through the Naval Observatory Merged Astrometric Dataset,

407: NOMAD \citep{zacharias04b}.} \citep{monet03},

408: with preference given to the UCAC2 measurements.

409: 

410: The $UVW$ velocities, measured in a Cartesian coordinate system,

411: can be transformed into a cylindrical coordinate system 

412: ($R$, $\phi$, $z$) with its origin at the Galactic center, 

413: where $R = \sqrt{(R_{\odot}+x)^{2}+y^{2}}$, 

414: $\phi = \arctan(y/(R_{\odot}+x))$, and $z=z$.

415: Our adopted and derived $UVW$ and $V_{\phi}$ velocities are 

416: given in Table~\ref{startab}.

417: 

418: 

419: \subsection{Galactic Orbit Parameters}

420: \label{orbit}

421: 

422: 

423: We use the orbit integrator developed by D.~Lin to calculate 

424: the Galactic orbital parameters for stars in our sample.

425: The gravitational potential of the Galaxy is computed by summing

426: the individual contributions of the disk, spheroid (or bulge), 

427: and halo components:

428: \begin{mathletters}

429: \begin{eqnarray}

430: \Phi_{\rm disk} & = & \frac{-GM_{\rm disk}}

431:   {\sqrt{R^{2} + (a + \sqrt{z^{2} + b^{2}})^{2}}} \\

432: \Phi_{\rm spheroid} & = & \frac{-GM_{\rm spheroid}}{r+c} \\

433: \Phi_{\rm halo} & = & v^{2}_{\rm halo} \ln{(r^{2} + d^{2})}, 

434: \end{eqnarray}

435: \label{ourpotential}

436: \end{mathletters}

437: where 

438: $R = \sqrt{x^{2} + y^{2}}$ and 

439: $r = \sqrt{x^{2} + y^{2} + z^{2}}$,

440: $M_{\rm disk} = 10^{11}$\,M$_{\sun}$,

441: $M_{\rm spheroid} = 3.4 \times 10^{10}$\,M$_{\sun}$,

442: $v_{\rm halo} = 128$\,km\,s$^{-1}$,

443: $a = 6.5$, 

444: $b = 0.26$, 

445: $c = 0.7$, and 

446: $d = 12.0$. 

447: This model is described in detail in \citet{johnston96} and \citet{johnston98},

448: who chose values for these parameters 

449: to reproduce a nearly-flat rotation curve for the Milky Way Galaxy 

450: between $r = 1$ and $r = 30$\,kpc and a disk scale height of 0.2\,kpc.

451: We require that each star complete a minimum of 20 orbits around the

452: Galactic center in 10~Gyr (at which point the orbital parameters

453: have settled to a constant value); no orbital parameters are given

454: in Table~\ref{startab} for stars that did not meet this requirement.

455: From the output we calculate 

456: $r_{\rm peri}$ and $r_{\rm apo}$, 

457: the minimum and maximum Galactocentric radii reached by each star; 

458: $|Z_{\rm max}|$, the maximum distance above (or below) the Galactic plane; 

459: and $e$, the eccentricity of the orbit, defined as 

460: $(r_{\rm apo}-r_{\rm peri}) / (r_{\rm apo}+r_{\rm peri})$.

461: These properties are also reported in Table~\ref{startab}.

462: 

463: Our kinematic results ought to be reasonably independent of our 

464: model for the Galactic potential.

465: Since we adopt the Galactic orbital properties of the inner and outer

466: halo populations from \citet{carollo07} (see \S~\ref{halokinematics}), 

467: we want to demonstrate some degree of consistency between 

468: results derived from the two different potentials.

469: To assess this, we compare the Galactic orbital properties 

470: of stars computed using Equation~\ref{ourpotential} 

471: to those computed using the analytic St\"{a}ckel-type potential 

472: \citep{sommerlarsen90,chiba98} adopted by \citet{beers00}

473: and \citet{carollo07}.

474: The parameters of their model have been tuned to reproduce a 

475: flat Galactic rotation curve beyond $R=4$~kpc and the local mass 

476: density at the Solar radius.

477: We use our three-component model of the Galactic potential to 

478: compute Galactic orbital parameters for 100 stars from the sample

479: of \citet{beers00}, strictly using their input data (distance, 

480: radial velocity, proper motion, etc.).

481: Figure~\ref{compareb00} compares the orbital parameters presented

482: in \citet{beers00} (i.e., computed with the St\"{a}ckel-type potential)

483: with those derived with our three-component model.

484: Both models generally produce similar apogalactic and perigalactic

485: distances 

486: as well as maximum vertical distance, but with a fair degree of scatter

487: in each of these quantities.

488: For $R_{\rm apo} > 10$~kpc and $|Z_{\rm max}| > 2$~kpc, the 

489: distances for individual stars are typically different by less than

490: a factor of two.

491: These systematic differences are inherent to the representations of the

492: potential.

493: 

494: \begin{figure*}

495: \epsscale{1.00}

496: \plotone{f02.eps}

497: \caption{

498: \label{compareb00}

499: Comparison of our predicted Galactic orbit properties 

500: with those of \citet{beers00}.

501: Our values have been computed using the three-component model

502: for the potential, while \citet{beers00} used the analytic 

503: St\"{a}ckel-type potential.

504: }

505: \end{figure*}

506: 

507: How worrisome are these differences?

508: Our approach to identifying the two populations isolates them

509: from one another in kinematic space by placing a buffer between 

510: them---reasonable uncertainty in the distance estimate or proper motion

511: of a star should not cause it to drift from one classification to

512: the other.

513: For example, stars in our inner halo population must not orbit more than 

514: 15~kpc from the Galactic center, while stars in our outer halo population

515: must orbit to at least 25~kpc.

516: We also adopt membership standards for the inner halo that 

517: rely on more than one kinematic selection criterion.

518: Therefore, we do not consider this systematic difference a serious problem

519: in our kinematic selection criteria.

520: 

521: 

522: \section{New Kinematic Definitions of the Inner and Outer Halo Populations}

523: \label{halokinematics}

524: 

525: 

526: Analyzing more than 10,000 calibration stars from low resolution 

527: ($R \equiv \lambda/\Delta\lambda \sim 2000$) spectra

528: obtained with the SDSS, \citet{carollo07} reported that the

529: Milky Way stellar halo is roughly divisible into two

530: broadly-overlapping components.

531: Their ``inner halo'' is composed of stars on highly eccentric orbits,

532: has a flattened density distribution,

533: exhibits a modest net prograde rotation, and

534: has a peak in the MDF at [Fe/H]~$= -1.6$.

535: Their ``outer halo'' is composed of stars with a variety of orbital

536: eccentricities, has a spherical density distribution,

537: exhibits a net retrograde rotation, and has a peak in the

538: MDF at [Fe/H]~$= -2.2$.

539: While their inner halo primarily dominates the stellar halo

540: at Galactocentric radii $\lesssim$~10--15~kpc and the

541: outer halo primarily dominates at radii $\gtrsim$~15--20~kpc,

542: these two components overlap in the Solar neighborhood,

543: disguising their distinct characteristics from being

544: recognized by earlier, smaller samples of stars.

545: 

546: We define membership in the inner and outer halo populations 

547: based \textit{only} on the kinematics of stars in our sample.

548: These new criteria, based on the classifications sketched by 

549: \citet{carollo07}, are shown in Figure~\ref{innerouterplot}.

550: For the remainder of the present study, we shall use the working

551: definition that 

552: inner halo membership is characterized by prograde rotation about

553: the Galactic center ($V_{\phi} > 0$~\kmsec), 

554: relatively small maximum vertical distance from the Galactic plane

555: ($|Z_{\rm max}| < 5$~kpc), relatively small maximum radial distance

556: from the Galactic center ($R_{\rm apo} < 15$~kpc), and

557: rather high orbital eccentricity ($e > 0.5$).

558: A star must meet each of these criteria to be classified as a 

559: member of the inner halo.

560: Outer halo membership is characterized by significant retrograde

561: rotation about the Galactic center ($V_{\phi} < -150$~\kmsec),

562: relatively large maximum vertical distance from the Galactic plane

563: ($|Z_{\rm max}| > 10$~kpc), or relatively large maximum radial 

564: distance from the Galactic center ($R_{\rm apo} > 25$~kpc).

565: We apply no explicit orbital eccentricity requirements for 

566: outer halo membership, although the criteria essentially 

567: select stars on relatively eccentric orbits since they

568: are presently in the Solar neighborhood and yet must travel to

569: great distances. 

570: In contrast to the inner halo membership, stars only need to possess

571: one of these characteristics to be included in the outer halo 

572: population.

573: While our selection criteria may eliminate 

574: genuine members of either population, they help to ensure that

575: there is minimal contamination between the two populations.

576: 

577: \begin{figure*}

578: \epsscale{1.00}

579: \plotone{f03.eps}

580: \caption{

581: \label{innerouterplot}

582: The kinematic definitions of our inner and outer halo samples.

583: The inner halo population (dark gray circles, red in the online edition) 

584: is defined by stars

585: that have $V_{\phi} > 0$~\kmsec, $|Z_{\rm max}| < 5$~kpc, 

586: $R_{\rm apo} < 15$~kpc, and $e > 0.5$.

587: The outer halo population (black triangles, blue in the online edition) 

588: is defined by stars

589: that have $V_{\phi} < -150$~\kmsec, $|Z_{\rm max}| > 10$~kpc,

590: or $R_{\rm apo} > 25$~kpc.

591: Stars that did not meet the kinematic criteria for these two 

592: populations are shown as small dots.

593: [Please see the electronic edition for a color version

594: of this figure.]

595: }

596: \end{figure*}

597: 

598: Interestingly, our inner halo criteria may allow for the inclusion of

599: a few thick disk stars on mildly eccentric orbits.

600: To eliminate these stars, we adopt one additional criterion for 

601: membership in the inner halo.

602: Assuming that the stellar thick disk and halo populations 

603: each have Gaussian velocity distributions in each of the three 

604: components (e.g., \citealt{bensby03}, eqns.\ 1 and 2), 

605: we calculate the relative probability that a star is

606: in the thick disk versus the halo.

607: We use the velocity dispersions

608: ($\sigma_{U}$, $\sigma_{V}$, $\sigma_{W}$)~=~(46, 50, 35)~\kmsec\ 

609: for the thick disk and (141, 106, 94)~\kmsec\ for the halo,

610: as well as the rotation speeds $\langle V_{\phi} \rangle = 200$~\kmsec\ and 

611: $\langle V_{\phi} \rangle = 40$~\kmsec\ for the thick disk and halo,

612: respectively \citep{chiba00}.

613: We require that inner halo members

614: are at least twice as likely to be members of the halo than the thick disk.

615: This eliminates less than 15\% of the stars that would otherwise be

616: classified as members of the inner halo.

617: 

618: Previous investigations \citep{stephens99,stephens02,fulbright02}

619: have employed very similar criteria to ours for 

620: defining outer halo membership, including stars with retrograde Galactic

621: orbits or probing to large Galactocentric distances in and above the 

622: plane of the disk.  

623: One unique feature of our study is the adoption of several selection

624: criteria for defining membership in the inner halo, rather than 

625: simply identifying stars at less extreme maximum distances as inner halo

626: members.

627: This advance has been made possible because of large surveys (e.g.,

628: SDSS/SEGUE) that have identified very large numbers of metal-poor stars 

629: in the Solar neighborhood.

630: 

631: One abundance criterion---but not [Fe/H]---is necessary to exclude 

632: stars whose abundances

633: do not reflect their initial composition.

634: When signatures of the slow neutron-capture nucleosynthesis process

635: (``\spro''; see \S~\ref{ncapstuff}) are enhanced in

636: metal-poor stars, this is regarded as evidence of past mass transfer

637: from an undetected companion star that passed through the 

638: asymptotic giant branch (AGB) phase of stellar evolution.

639: Since the metal content of these stars has changed over the

640: course of their lives, we exclude them

641: membership in either halo population.\footnote{

642: The pure-$s$-process [Ba/Eu] ratio is predicted to be 

643: [Ba/Eu]$_{\rm s} = +0.63$ \citep{simmerer04}.

644: We exclude stars with [Ba/Eu]~$>+0.5$ and [Ba/Fe]~$>+1.0$.

645: Unfortunately, Eu has been measured in only $\approx$35\% 

646: of the stars in our entire sample, so it is likely that some stars 

647: that have experienced \spro\ enrichment during their lifetimes 

648: will remain in our sample.

649: Fortunately, though, nucleosynthesis in the envelope of stars during the 

650: AGB phase has a negligible effect on most light elements, 

651: (C, N, O, F, Ne, and perhaps Na being notable exceptions; 

652: e.g., \citealt{straniero06})

653: including most of those examined in detail in our study.}

654: Later, in \S~\ref{rspro}, we examine these stars 

655: for any common kinematic signatures.

656: 

657: The members of both the inner and outer halo populations according

658: to these definitions are listed, along with the abundance ratios

659: we consider, in Tables \ref{innerabundtab} and \ref{outerabundtab}.

660: 

661: 

662: \subsection{Classification Uncertainties}

663: \label{classuncertainties}

664: 

665: 

666: In Figure~\ref{uncertainties} we examine the changes in the derived

667: orbital parameters resulting from variations of the 

668: input measurements within reasonable uncertainties.

669: We use the subsample of stars without prior kinematic analysis

670: (the 309 stars noted in \S~\ref{litdata}) for this test.

671: The changes are obtained by adjusting the distance by $\pm$20\%,

672: the proper motions by $\pm$5~mas~yr$^{-1}$ 

673: (or 20\% for small values of the proper motion), or 

674: the radial velocity by $\pm10$~\kmsec.

675: We also examine the effect of changing 

676: the mass of the Galaxy by $\pm$10\%.

677: RMS values for each set of changes are shown on the right side of 

678: each panel in Figure~\ref{uncertainties}.

679: It is clear that, for most stars, these changes do not introduce 

680: appreciable uncertainty into the derived orbital parameters.

681: For a small fraction of stars ($\approx$5\%), the changes are large enough 

682: that they could cause an

683: inner halo star to be classified as an outer halo star or vice versa.

684: While some stars scatter in and

685: out of the two populations, we confirm that \textit{no} stars 

686: in this sample actually change their classification from inner to outer

687: (or vice versa) with these variations in the input parameters.

688: This result is a consequence of the multiple requirements necessary

689: for inclusion in the inner halo population and the wide buffers placed

690: between the kinematic and orbital properties of the two populations.

691: 

692: \begin{figure}

693: \epsscale{1.15}

694: \plotone{f04.eps}

695: \caption{

696: \label{uncertainties}

697: Uncertainties in the derived orbital parameters.

698: Changes are shown for alterations in the distance

699: (dark gray circles, black in the online edition),

700: the mass of the Galaxy

701: (open four-point stars, blue in the online edition),

702: the proper motions (open squares, red in the online edition),

703: or the radial velocity (light gray ``X'' points, green in the online edition).

704: The $\Delta$s refer to (modified parameter)$-$(unmodified parameter).

705: RMS uncertainties are shown for each set of $\Delta$ values

706: on the right side of each panel.

707: [Please see the electronic edition for a color version

708: of this figure.]

709: }

710: \end{figure}

711: 

712: Confident that our definitions of the inner and outer halo populations

713: are robust against uncertainties in the measurements, we proceed

714: to compare the compositions of stars in the inner and outer halo populations.

715: 

716: 

717: \section{A Caution: the Connection between Low-Metallicity

718: Stars and Retrograde Orbits}

719: \label{caution}

720: 

721: 

722: Many authors have examined the relationship between 

723: [Fe/H] and $V_{\phi}$ to characterize the kinematic properties

724: of the thick disk and halo in an effort to better understand

725: the formation mechanisms of these components

726: \citep[e.g.,][]{norris86,sandage87,norris89,beers95,chiba98,chiba00}. 

727: Our sample is not designed to rederive the relationship between

728: [Fe/H] and $V_{\phi}$,

729: but we must assess the effect of the proper motion

730: bias---which selects against stars with small proper motions---in 

731: our sample before proceeding.

732: Examination of the upper left panel of Figure~\ref{innerouterplot}

733: reveals that many metal-poor stars in our sample have retrograde orbits and

734: vice-versa.

735: We show this more explicitly in Figure~\ref{cautionplot}. 

736: Of the 94 stars in our sample on retrograde orbits, 53

737: (56\%) also have [Fe/H]~$< -2.0$.

738: Turning the problem around reveals that, of the 93 stars in our

739: sample with [Fe/H]~$<-2.0$, 53 (57\%) are on retrograde orbits.

740: This correlation strengthens at lower metallicities:

741: 35 of the 54 stars (65\%) with [Fe/H]~$<-2.5$ are on retrograde orbits, and

742: 13 of the 17 stars (76\%) with [Fe/H]~$<-3.0$ are on retrograde orbits.

743: 

744: \begin{figure}

745: \epsscale{1.15}

746: \plotone{f05.eps}

747: \caption{

748: \label{cautionplot}

749: Rotational velocity $V_{\phi}$ versus [Fe/H] for our sample.

750: Of the $\sim$700~stars represented on this plot, 94 have retrograde

751: Galactic orbits ($V_{\phi} < 0$~\kmsec; indicated in turquoise 

752: in the online edition) and

753: 93 have [Fe/H]~$< -2.0$ (indicated in orange in the online edition).

754: A large number of stars (53) fit both categories:

755: 57\% of stars with [Fe/H]~$<-2.0$ are on retrograde orbits, and

756: 56\% of stars on retrograde orbits have [Fe/H]~$<-2.0$.

757: The large gray squares (magenta in the online edition)

758: indicate $\langle V_{\phi} \rangle$ and the associated

759: 1 standard deviation scatter for stars in 0.2~dex bins of [Fe/H], 

760: but we emphasize that our sample is biased toward metal-poor stars

761: with large space velocities, hence the disproportionately large

762: number of stars with [Fe/H]~$< -2.0$ and significant retrograde

763: velocities.

764: [Please see the electronic edition for a color version of this figure.]

765: }

766: \end{figure}

767: 

768: The mean Galactic rotational velocities, 

769: $\langle V_{\phi} \rangle$, binned in 0.2~dex intervals in [Fe/H],

770: are shown in Figure~\ref{cautionplot}.

771: We find an approximate relation between $V_{\phi}$ and [Fe/H] of

772: $\Delta\langle V_{\phi}\rangle / \Delta$[Fe/H]~$ \approx

773: -140$~km~s$^{-1}~{\rm dex}^{-1}$ over 

774: $-2.0 <$~[Fe/H]~$< -0.6$, lessening to 

775: $\Delta\langle V_{\phi}\rangle / \Delta$[Fe/H]~$ \approx

776: -90$~km~s$^{-1}~{\rm dex}^{-1}$ over 

777: $-3.0 <$~[Fe/H]~$< -2.0$.

778: This relationship is a direct consequence of our

779: proper motion bias introduced in \S~\ref{litdata},

780: which would tend to select against very metal-poor stars

781: with near-zero or prograde net rotation, and 

782: is not a contradiction to previous results such

783: as \citet{norris86}, \citet{beers95}, or \citet{chiba00}.

784: The most recent of these studies found 

785: $\Delta\langle V_{\phi}\rangle / \Delta$[Fe/H]~$ \approx 

786: 160$~km~s$^{-1}~{\rm dex}^{-1}$ in the range 

787: $-1.7 <$~[Fe/H]~$<-0.6$ and a nearly constant 

788: $\Delta\langle V_{\phi}\rangle / \Delta$[Fe/H]~$ \approx

789: 0$~km~s$^{-1}~{\rm dex}^{-1}$ for

790: $-2.6 <$~[Fe/H]~$<-1.7$ in the Solar neighborhood.

791: Presumably, if precise proper motions (even small ones) 

792: were known for all the metal-poor stars found in the literature,

793: our relationship between $\langle V_{\phi}\rangle$ and [Fe/H]

794: would flatten out for the same metallicities as found by 

795: previous studies.

796: 

797: Some metal-poor stars are identified by high proper motion

798: searches and would be expected to exhibit a kinematic bias;

799: however, a large percentage of metal-poor stars 

800: are identified with no kinematic selection criteria

801: (via objective-prism surveys or ultraviolet excess), 

802: although these surveys do avoid stars at low Galactic latitude.

803: Also, investigators performing abundance analyses 

804: tend to preferentially select the most metal-poor stars for 

805: high-resolution followup.

806: Whether because of a selection bias or a true physical preference for

807: extreme orbits (or both), a significant fraction of the

808: stars with [Fe/H]~$\lesssim -2.0$ that have been subject to

809: high-resolution abundance analyses over the last 15 years

810: have retrograde orbits.

811: It is possible that the large numbers of very metal-poor stars

812: with [Fe/H]~$\lesssim -3.0$ found in the last few years could

813: reintroduce a slope into the relationship between 

814: $\langle V_{\phi}\rangle$ and [Fe/H] at low metallicities,

815: and such a reexamination should be undertaken in the near future.

816: 

817: This relation shown in Figure~\ref{cautionplot} 

818: is important in the context of examining

819: kinematic and chemical correlations in metal-poor stars.

820: Any abundance trend that is preferentially found

821: in low-metallicity stars will also be preferentially found

822: in stars with retrograde velocities, insofar as those

823: velocities are computed based on proper motion measurements

824: with an intrinsic bias.

825: As such, we caution that any abundance trends correlated with

826: retrograde orbits alone should not be over-interpreted as 

827: signatures of past accretion events if they

828: can be attributed to nucleosynthetic patterns that

829: inherently occur at low metallicity.

830: 

831: In Figure~\ref{els62plot} we show the relationship between 

832: metallicity and orbital eccentricity for our inner and outer halo 

833: populations, similar to Figure~4 of \citet{eggen62}.

834: The shaded band in this figure represents the locus of stars

835: used by \citet{eggen62}.  

836: (We convert ultraviolet excess, $\delta (U-B)$, to [Fe/H] 

837: using the approximate relationships given in \citealt{sandage87}.)

838: Their band encompasses the disk stars at high metallicity and low

839: eccentricity as well as the greater part of our inner halo population,

840: but their sample did not extend to metallicities much lower than 

841: [Fe/H]~$= -2.5$.

842: Many studies since \citet{eggen62} have used larger, unbiased 

843: datasets to demonstrate 

844: that there is no correlation between orbital eccentricity and metallicity 

845: (e.g., \citealt{chiba00} and references therein) in the halo.

846: While our inner halo is explicitly chosen to include stars with

847: high eccentricities and our outer halo effectively selects for

848: these stars, too, Figure~\ref{els62plot} shows that our entire sample

849: does include metal-poor stars across the full range of eccentricity.

850: 

851: \begin{figure}

852: \epsscale{1.15}

853: \plotone{f06.eps}

854: \caption{

855: \label{els62plot}

856: The relationship between metallicity and Galactic orbital eccentricity

857: in our sample.

858: Symbols are the same as in Figure~\ref{innerouterplot}.

859: The shaded band (yellow in the online edition) represents the 

860: relationship identified in Figure~4 of \citet{eggen62}.

861: [Please see the electronic edition for a color version of this figure.]

862: }

863: \end{figure}

864: 

865: 

866: \section{Abundance Results}

867: \label{results}

868: 

869: 

870: \subsection{Trends with $R_{\rm apo}$ and $|Z_{\rm max}|$}

871: \label{trends}

872: 

873: 

874: In Figure~\ref{trendplot} we examine the trends of 

875: the sodium (Na), magnesium (Mg), calcium (Ca), titanium (Ti), 

876: nickel (Ni), yttrium (Y), barium (Ba), and europium (Eu) to Fe ratios

877: as a function of $R_{\rm apo}$ and $|Z_{\rm max}|$

878: for stars with [Fe/H]~$<-1.0$.

879: Linear least-squares fits are also shown.

880: No slopes are significant at the 2$\sigma$ level.

881: Using a sample of 11 stars, 

882: \citet{stephens99} found no trends between [$\alpha$/Fe] 

883: and $R_{\rm apo}$ or $|Z_{\rm max}|$, a result which we confirm.

884: Using a larger sample (56 stars) and a homogeneous abundance analysis,

885: \citet{stephens02} found a slight decrease in 

886: $\langle$[$\alpha$/Fe]$\rangle$ 

887: ($-0.0012$~dex~kpc$^{-1}$) from $R_{\rm apo} \sim 8$--100~kpc.

888: Since they used a homogeneous abundance analysis, we defer to their

889: result, but we note that their slope only represents a very subtle

890: change in [$\alpha$/Fe] of $\sim 0.1$~dex over $\sim 90$~kpc.

891: They also found an increase in the cosmic scatter 

892: of [Y/Fe] at larger $R_{\rm apo}$.

893: We propose that the complex nucleosynthetic origins of the 

894: light neutron-capture species (see \S~\ref{rspro}) must be better 

895: understood before attempting any serious interpretation of these results.

896: \citet{stephens02} reported no other dependences of 

897: abundance ratios with orbital parameters.

898: Combining the \citet{stephens02} and \citet{fulbright00,fulbright02}

899: datasets, \citet{fulbright04a} noted slight decreases in all of the light

900: element ratios with respect to Fe (i.e., Na, Mg, Al, Si) and 

901: weaker decreases of heavier elements (i.e., Ca, Ti, Ni, Y, and Zr)

902: when using Galactic rest-frame velocity as a surrogate of the

903: model-dependent $R_{\rm apo}$.

904: Our larger---but inhomogeneous---dataset does not show these trends.

905: The trends reported by \citet{nissen97}---whose dataset differed

906: significantly with regard to both kinematics and metallicity range

907: from subsequent studies---have largely not been reproduced by those

908: studies, including ours.

909: 

910: \begin{figure*}

911: \epsscale{1.00}

912: \plotone{f07.eps}

913: \caption{

914: \label{trendplot}

915: Elemental abundance trends as a function of $R_{\rm apo}$ and 

916: $|Z_{\rm max}|$ for stars with [Fe/H]~$<-1.0$.

917: Least squares fits are shown as solid gray lines (red in the online edition).

918: The solar ratios are indicated by the dotted lines.

919: [Please see the electronic edition for a color version of this figure.]

920: }

921: \end{figure*}

922: 

923: 

924: \subsection{$\alpha$ and Iron-Peak Elements in the Inner and Outer Halo

925: Populations}

926: \label{alphafe}

927: 

928: 

929: Figures~\ref{abundplot1} and \ref{abundplot2} display the 

930: logarithmic abundance ratios of Na, Mg, Ca, Ti, and Ni to Fe

931: (relative to Solar) for our entire sample 

932: and for members of the inner and outer halo populations.  

933: For [Fe/H]~$<-1.0$, 

934: both the inner and outer halo populations show super-Solar 

935: [Mg/Fe], [Ca/Fe], and [Ti/Fe] ratios 

936: (hereafter loosely defined as [$\alpha$/Fe])

937: scattered around $+0.3$~dex, 

938: with no apparent slope below [Fe/H]~$\lesssim-1.5$.

939: \citet{stephens02} reported an increase in [$\alpha$/Fe] with decreasing

940: [Fe/H], a trend not reproduced elsewhere in studies with large numbers

941: of very metal-poor stars 

942: \citep{mcwilliam95b,carretta02,cayrel04,cohen04b,arnone05,lai08}.

943: The [Mg/Fe] scatter may increase at low metallicities, and

944: \citet{stephens02} only studied 4 stars with [Fe/H]~$<-3.0$.

945: [Na/Fe] exhibits significantly larger scatter (increasing scatter with

946: decreasing metallicity, at least to [Fe/H]~$\sim -2.3$)

947: than any of the [$\alpha$/Fe] ratios does, 

948: and [Na/Fe] approximately follows the Solar ratio.

949: Ni should be produced along with Fe, and [Ni/Fe]

950: correlates with [Fe/H] at all metallicities with

951: relatively small scatter (increasing also to a maximum below

952: [Fe/H]~$\lesssim -2.0$).

953: 

954: \begin{figure}

955: \epsscale{1.15}

956: \plotone{f08.eps}

957: \caption{

958: \label{abundplot1}

959: [Na/Fe], [Mg/Fe], [Ca/Fe], and [Ti/Fe] abundance ratios for our

960: inner (dark gray circles, red in the online edition) 

961: and outer (black triangles, blue in the online edition) 

962: halo populations.

963: Stars that did not meet the kinematic criteria for these two 

964: populations are shown as small dots.

965: [Please see the electronic edition for a color version of this figure.]

966: }

967: \end{figure}

968: 

969: \begin{figure}

970: \epsscale{1.15}

971: \plotone{f09.eps}

972: \caption{

973: \label{abundplot2}

974: [Ni/Fe], [Y/Fe], [Ba/Fe], and [Eu/Fe] abundance ratios for our

975: inner and outer halo populations.

976: Symbols are the same as in Figure~\ref{abundplot1}.

977: Upper limits are not displayed.

978: [Please see the electronic edition for a color version of this figure.]

979: }

980: \end{figure}

981: 

982: One fact is readily apparent from these plots:

983: the bulk of stars in our inner halo sample only are found

984: with metallicities in the range 

985: $-2.5 \lesssim $~[Fe/H]~$\lesssim -0.5$, 

986: while most stars in our outer halo are found over the 

987: metallicity range $-3.5 \lesssim $~[Fe/H]~$\lesssim -1.5$.

988: \citet{carollo07} claimed that the metallicity distribution function

989: (MDF) of the inner halo peaks around [Fe/H]~$\sim -1.6$ and the 

990: MDF of the outer halo peaks around [Fe/H]~$\sim -2.2$; our results

991: support these assertions.

992: Nevertheless, we caution that (1) our proper motion bias likely 

993: selects against lower-metallicity stars on prograde orbits and 

994: (2) our sample is drawn from studies

995: designed to select interesting metal-poor stars for detailed 

996: abundance analyses, and therefore it should not be used for

997: any assessments of metal-poor MDF's.

998: 

999: Our sample is sensitive to chemical differences in the inner and

1000: outer halo populations for stars with the same Fe abundance. 

1001: Between $-2.3 <$~[Fe/H]~$<-1.6$, the mean [Mg/Fe] of the outer halo 

1002: ($\langle$[Mg/Fe]$\rangle = 0.30$, 

1003: $\sigma_{\rm mean} = 0.05$)

1004: is slightly lower than the mean [Mg/Fe] of the inner halo

1005: ($\langle$[Mg/Fe]$\rangle = 0.40$, 

1006: $\sigma_{\rm mean} = 0.04$), 

1007: even when the well-known $\alpha$-poor star \mbox{G~004-036} 

1008: ([Fe/H]~$=-1.93$, $R_{\rm apo} = 33^{+6}_{-4}$~kpc, \citealt{ivans03}) 

1009: is excluded.\footnote{

1010: The quoted uncertainties in the orbital parameters

1011: are computed by the same methods discussed in \S~\ref{classuncertainties}.}

1012: This difference is not obvious in [Ca/Fe] or [Ti/Fe], and the scatter

1013: in [Ca/Fe] for both populations is noticeably smaller than the 

1014: scatter in either [Mg/Fe] or [Ti/Fe].

1015: For all three of [Mg/Fe], [Ca/Fe], and [Ti/Fe], it is worth noting that

1016: the ``extreme'' [X/Fe] ratios at a given [Fe/H] are predominantly

1017: found in the outer halo population.

1018: 

1019: In Figure~\ref{abundplot2}, this interesting trend begins to emerge 

1020: more clearly with [Ni/Fe]:

1021: the inner halo abundance ratios at a given [Fe/H] appear much more

1022: tightly correlated than the outer halo abundance ratios.

1023: The scatter in inner halo [Ni/Fe] is commensurate with the typical

1024: abundance uncertainties in a given measurement, typically 0.1--0.2~dex,

1025: while the scatter in the outer halo [Ni/Fe] is typically 0.5--0.7~dex

1026: (at least for [Fe/H]~$\lesssim-1.8$).

1027: 

1028: In Figure~\ref{binnedplot} we display the abundance ratios

1029: for [Mg/Fe] and [Ni/Fe] binned as a function of [Fe/H].

1030: A boxplot is shown for each [Fe/H] bin (typically 0.4~dex wide), 

1031: displaying the median, inner quartiles, and extremes of 

1032: the entire sample, only the inner halo stars, and only the outer halo stars.

1033: For $-2.2 \lesssim$~[Fe/H]~$\lesssim -1.4$ (where 

1034: there is significant abundance overlap between the two populations),

1035: the median [Mg/Fe] ratios for the inner halo are consistently higher than the

1036: outer halo medians by 0.10--0.20~dex. 

1037: The mean [Ni/Fe] ratios of both populations trace the Solar ratio very closely.

1038: For [Ni/Fe], it is apparent that the extremes of the

1039: inner halo are significantly smaller than the extremes of the outer halo.

1040: In this case we point to the different degrees of scatter in the 

1041: inner versus outer populations 

1042: as evidence for chemical differences between the inner and outer 

1043: halo populations;

1044: no stars with significant deviations from the median ratios

1045: are members of the inner halo population.

1046: The precise amount of scatter of each population may be affected

1047: by the inhomogeneous nature of our sample, but the relative 

1048: scatter between the two populations is robust. 

1049: 

1050: \begin{figure}

1051: \epsscale{1.15}

1052: \plotone{f10.eps}

1053: \caption{

1054: \label{binnedplot}

1055: Binned abundance ratios for [Mg/Fe], [Ni/Fe], and [Ba/Fe],

1056: displayed as quartile boxplots.

1057: Black boxes represent all stars in the bin 

1058: (including the inner and outer halo populations), 

1059: light gray boxes (red in the online edition) 

1060: represent the inner halo stars in the bin,

1061: and dark gray boxes (blue in the online edition)

1062: represent the outer halo stars in the bin.

1063: The divisions in the [Fe/H] bins are indicated by vertical gray stripes.

1064: [Please see the electronic edition for a color version of this figure.]

1065: }

1066: \end{figure}

1067: 

1068: Our [Na/Fe] ratios in Figure~\ref{abundplot1}

1069: exhibit a large degree of scatter.

1070: Two of our literature sources for low metallicity stars,

1071: \citet{honda04a,honda04b} and \citet{barklem05}, did not report 

1072: [Na/Fe] ratios for their samples, thus at low

1073: metallicities ([Fe/H]~$\lesssim -2.0$) there are not enough 

1074: measurements to adequately compare the inner and outer halo populations.

1075: In Figure~\ref{naplots}, we show [Na/Fe] as a function of

1076: [Mg/Fe], [Ca/Fe], and [Ti/Fe].

1077: The line in the upper left panel of this figure shows the

1078: correlation between [Mg/Fe] and [Na/Fe] found in field halo giants

1079: by \citet{hanson98}, which generally matches our metal-poor sample.

1080: A similar trend (not shown) exists between [Ca/Fe] or [Ti/Fe] and 

1081: [Na/Fe]. 

1082: Outer halo stars generally occupy the extremes of each distribution,

1083: particularly the Na-depleted extremes.

1084: This implies that these stars formed from---at least in 

1085: part---an incompletely mixed ISM where the yields of 

1086: individual Type~II SNe events could still be ``noticed'' 

1087: against the overall chemical background of the ISM.

1088: 

1089: \begin{figure*}

1090: \epsscale{1.00}

1091: \plotone{f11.eps}

1092: \caption{

1093: \label{naplots}

1094: [Na/Fe] as a function of [Mg/Fe], [Ca/Fe], [Ti/Fe], and [Ni/Fe]

1095: for our inner and outer halo populations.

1096: Symbols are the same as in Figure~\ref{abundplot1}.

1097: The dotted line represents [X/Na]~$= +0.0$.

1098: The solid line (green in the online edition)

1099: in the top left panel is the correlation between

1100: [Mg/Fe] and [Na/Fe] reported by \citet{hanson98}.

1101: [Please see the electronic edition of the journal for a color version

1102: of this figure.]

1103: }

1104: \end{figure*}

1105: 

1106: \citet{hanson98} reported both 

1107: an increase in the [Na/Fe] scatter with decreasing [Fe/H]

1108: and an increase in the [Na/Fe] scatter when comparing stars on

1109: retrograde orbits to those on prograde orbits.

1110: To investigate this trend, we turn to a species with more 

1111: measurements in our dataset, Ni.

1112: Previous analyses have revealed correlations between

1113: [Na/Fe] and [Ni/Fe] ratios and stellar kinematic properties

1114: (e.g., \citealt{nissen97}, \citealt{hanson98}, 

1115: \citealt{shetrone03}, \citealt{venn04}).

1116: We show this relationship in our sample in Figure~\ref{naplots}.

1117: The Na-Ni relationship originates from the neutron-rich nature

1118: of the dominant isotopes of these species, which can be produced

1119: in Type~II SNe, albeit in non-Solar ratios.

1120: (See \citealt{venn04} for an extensive discussion of

1121: the Na-Ni nucleosynthesis relationship in this context.)

1122: This relationship breaks down in material

1123: enriched by Type~Ia SNe products, since 

1124: Type~Ia SNe produce very little $^{23}$Na, 

1125: so we only examine stars with [Fe/H]~$<-1.0$.

1126: Ni-poor stars ([Ni/Fe]~$\leq-0.2$)

1127: are preferentially associated with stars on retrograde orbits; however,

1128: Ni-rich stars ([Ni/Fe]~$\geq +0.2$) and Ni-normal stars

1129: ($-0.05 \leq$~[Ni/Fe]~$\leq +0.05$) also exhibit a preference 

1130: for retrograde or no net rotation orbits in our sample.

1131: The same effect occurs even if we only consider stars with 

1132: [Fe/H]~$<-2.0$, where our proper motion bias is most pronounced.

1133: Thus we are unable to confirm or refute the correlation between

1134: non-Solar [Ni/Fe] (and, by extension, {Na/Fe]) ratios and retrograde orbits.

1135: 

1136: 

1137: \subsection{Neutron-Capture Species in the Inner and Outer Halo Populations}

1138: \label{ncapstuff}

1139: 

1140: 

1141: Nuclei heavier than the iron group are formed by the addition of 

1142: neutrons to existing seed nuclei.

1143: The timescales for neutron- ($n$-) captures determine the resulting

1144: abundance patterns.

1145: If the average time between successive 

1146: neutron captures is less than the typical halflife against $\beta^{-}$ 

1147: decay, this is referred to as the slow- ($s$-) \ncap\ process.

1148: In contrast, the rapid- ($r$-) \ncap\ process occurs when many neutrons 

1149: are added before any $\beta^{-}$ decays can occur.

1150: These two processes result in very different heavy element 

1151: abundance patterns.

1152: An exact site for the \rpro\ has yet to be conclusively identified,

1153: but the short stellar timescale necessary to produce \rpro\ enrichment

1154: in stars with [Fe/H]~$\lesssim -3.0$ suggests that an association 

1155: with massive core-collapse Type~II SNe is likely.

1156: Low and intermediate mass ($\sim$~1.5--3.0~$M_{\odot}$) stars that pass

1157: through the AGB phase of evolution are 

1158: the primary source of \spro\ material \citep[e.g.,][]{busso99,straniero06}. 

1159: 

1160: In Figure~\ref{abundplot2} we examine the abundances of the \ncap\ 

1161: species [Y/Fe], [Ba/Fe], and [Eu/Fe] in our inner and outer

1162: halo populations.

1163: These ratios display very little scatter for [Fe/H]~$\gtrsim-1.8$,

1164: but they exhibit considerable scatter below [Fe/H]~$\lesssim-1.8$.

1165: A familiar trend reappears:

1166: the inner halo [X/Fe] ratios follow (roughly) a monotonic relationship

1167: with [Fe/H], while the outer halo [X/Fe] ratios scatter about 

1168: appreciably at a given [Fe/H].

1169: For example, the scatter in inner halo [Ba/Fe] is larger ($\sim$~0.5~dex)

1170: than the typical measurement uncertainty ($\sim$~0.1--0.2~dex)

1171: and follows the Solar ratio, 

1172: yet the outer halo [Ba/Fe] scatter is \textit{much} larger, 

1173: typically 1--2~dex 

1174: (see also, e.g., \citealt{mcwilliam98} and \citealt{francois07}).

1175: This pattern may also be observed with [Y/Fe], though the outer halo

1176: scatter is much less extreme.

1177: It is more difficult to discern these trends with [Eu/Fe], 

1178: for which fewer measurements exist, and 

1179: these effects may also be attributed only to increased [Eu/Fe] scatter 

1180: at the lowest metallicities.

1181: The [Ba/Fe] ratios, binned by [Fe/H], are also shown in 

1182: Figure~\ref{binnedplot}.

1183: The mean [Ba/Fe] of all stars traces the Solar value over the range

1184: $-2.0 \lesssim$~[Fe/H]~$\lesssim +0.0$, only declining at the lowest 

1185: metallicities.

1186: The majority of [Ba/Fe] ratios in both populations are similar.

1187: Stars with extreme ratios only comprise the outer halo;

1188: the inner halo ratios are remarkably similar to one another.

1189: 

1190: The top panel of Figure~\ref{innerouterncap} displays the [Ba/Y] ratio

1191: as a function of [Fe/H] for our inner and outer halo samples.

1192: $^{89}$Y ($Z=39$) contains 50 neutrons, which is one of the 

1193: magic neutron numbers that correspond to closed nuclear shells

1194: and significantly lower the nuclear cross section to further neutron capture.

1195: During the slow ($s$) nucleosynthesis reaction, this bottleneck causes 

1196: lots of nuclei with 50 neutrons to be produced.

1197: Thus Y is representative of the atomic species produced at this first

1198: abundance peak in the \spro.

1199: $^{138}$Ba ($Z=56$) is the dominant isotope of Ba produced in \spro\ 

1200: nucleosynthesis and also contains a magic number of neutrons, 82.

1201: Ba is representative of the atomic species produced at the second abundance

1202: peak in the \spro.

1203: The [Ba/Y] ratio is useful as a probe of the relative amounts of material

1204: produced at these two peaks in the \spro.

1205: The general decline in [Ba/Y] at [Fe/H]~$\lesssim -2.0$ results 

1206: from decreasing Ba contributions from the main \spro\ (see the 

1207: bottom panel of this figure), while Y production from an apparently

1208: primary (i.e., not metallicity-dependent) process remains 

1209: approximately constant to the lowest metallicities observed

1210: (see, e.g., \citealt{travaglio04}).

1211: The inner and outer halo members appear to be randomly 

1212: distributed among the normal scatter for a given [Fe/H], and there

1213: is no obvious correlation with these populations.

1214: (The greatest scatter does appear 

1215: in the outer halo population at [Fe/H]~$\lesssim -2.0$ 

1216: where there is an overall lack of inner halo stars.)

1217: 

1218: \begin{figure}

1219: \epsscale{1.15}

1220: \plotone{f12.eps}

1221: \caption{

1222: \label{innerouterncap}

1223: [Ba/Y] and [Ba/Eu] abundance ratios for our

1224: inner and outer halo populations.

1225: Symbols are the same as in Figure~\ref{abundplot1}.

1226: The lower gray dashed line (orange in the online edition)

1227: represents the pure \rpro\ ratio and

1228: the upper gray dashed line (turquoise in the online edition)

1229: represents the pure \spro\ ratio as predicted by \citet{simmerer04}.

1230: [Please see the electronic edition for a color version of this figure.]

1231: }

1232: \end{figure}

1233: 

1234: The bottom panel of Figure~\ref{innerouterncap} displays the [Ba/Eu]

1235: ratio as a function of [Fe/H] for our inner and outer halo samples.

1236: Relative to Ba, very little Eu is produced via the \spro, yet 

1237: Eu is relatively easy to produce in the rapid ($r$) nucleosynthesis

1238: reaction, so the [Ba/Eu] ratio provides a good assessment of the 

1239: relative amounts of $s$- and \rpro\ material present in a star.

1240: We also show the [Ba/Eu] ratios predicted for pure $s$- or \rpro\ 

1241: nucleosynthesis \citep{simmerer04}.

1242: (Recall that we have excluded stars with a pure-$s$-process

1243: signature from membership in our inner and outer halo populations.)

1244: The stellar data generally decline from the Solar [Ba/Eu] ratio at high 

1245: metallicity toward an \rpro\ dominant ratio at low metallicity, though

1246: a small number of stars at low-metallicity show evidence of 

1247: $s$-only enrichment.

1248: Again, the inner and outer halo members appear to be randomly 

1249: distributed among the scatter of all stars at a given [Fe/H].

1250: 

1251: 

1252: \section{Discussion}

1253: \label{discussion}

1254: 

1255: 

1256: \subsection{The Kinematically and Chemically Diverse Outer Halo}

1257: 

1258: 

1259: Is it possible that the inner halo population consistently exhibits

1260: a smaller degree of scatter than the outer halo population because of

1261: our classification process and not an astrophysical phenomenon?

1262: In other words, have we more precisely determined a kinematic population

1263: with our ``and'' selection criterion for the inner halo than with our

1264: ``or'' selection criterion for the outer halo?

1265: To address this possibility, we more closely analyze each of our 

1266: selection criteria for the outer halo population.

1267: In Figure~\ref{retroplot} we display the [Mg/Fe] and [Ba/Fe] ratios

1268: of stars on increasingly retrograde Galactic orbital velocities,

1269: in Figure~\ref{highzplot} we display these ratios for stars

1270: with increasingly higher values of $|Z_{\rm max}|$, and 

1271: in Figure~\ref{highrplot} we display these ratios for stars

1272: with increasingly larger values of $R_{\rm apo}$.

1273: Appreciable increases in the [Ba/Fe] scatter are not obvious 

1274: for stars with the most retrograde velocities,

1275: highest distances from the Galactic plane,

1276: or increasing maximum distance from the Galactic center.

1277: Discernible changes in the [Mg/Fe] ratios are not apparent either.

1278: Any set of stars sharing one of these outer halo defining characteristics 

1279: likely would exhibit considerably more abundance 

1280: scatter than the stars in our inner halo population.

1281: The stars in our outer halo population

1282: appear to be genuinely kinematically and chemically uncorrelated.

1283: 

1284: \begin{figure}

1285: \epsscale{1.15}

1286: \plotone{f13.eps}

1287: \caption{

1288: \label{retroplot}

1289: Identifying stars of increasing retrograde rotation about the Galaxy

1290: and associated [Mg/Fe] and [Ba/Fe] abundance ratios.

1291: Stars are selected based only on their rotation velocity:

1292: $-100 < V_{\phi} \leq -50$~\kmsec\ 

1293: (filled gray triangles, red in the online edition),

1294: $-150 < V_{\phi} \leq -100$ 

1295: (filled gray squares, orange in the online edition),

1296: $-200 < V_{\phi} \leq -150$ 

1297: (filled gray pentagons, turquoise in the online edition),

1298: and    $V_{\phi} \leq -200$ 

1299: (filled gray circles, blue in the online edition).

1300: The stars selected according to these definitions are highlighted in 

1301: the lower two panels.

1302: [Please see the electronic edition for a color version of this figure.]

1303: }

1304: \end{figure}

1305: 

1306: \begin{figure}

1307: \epsscale{1.15}

1308: \plotone{f14.eps}

1309: \caption{

1310: \label{highzplot}

1311: Identifying stars of increasing vertical distance from the Galactic plane

1312: and associated [Mg/Fe] and [Ba/Fe] abundance ratios.

1313: Stars are selected based only on $|Z_{\rm max}|$:

1314: $ 2.5 < |Z_{\rm max}| \leq  3.5$~kpc 

1315: (filled gray triangles, red in the online edition),

1316: $ 3.5 < |Z_{\rm max}| \leq  5$ 

1317: (filled gray squares, orange in the online edition),

1318: $ 5 < |Z_{\rm max}| \leq 10$ 

1319: (filled gray pentagons, turquoise in the online edition),

1320: and  $|Z_{\rm max}| >    10$ 

1321: (filled gray circles, blue in the online edition).

1322: The stars selected according to these definitions are highlighted in

1323: the lower two panels.

1324: [Please see the electronic edition for a color version of this figure.]

1325: }

1326: \end{figure}

1327: 

1328: \begin{figure}

1329: \epsscale{1.15}

1330: \plotone{f15.eps}

1331: \caption{

1332: \label{highrplot}

1333: Identifying stars of increasing maximum radial distance from the 

1334: Galactic center and associated [Mg/Fe] and [Ba/Fe] abundance ratios.

1335: Stars are selected based only on $R_{\rm apo}$:

1336: $20 < R_{\rm apo} \leq 25$~kpc 

1337: (filled gray triangles, red in the online edition),

1338: $25 < R_{\rm apo} \leq 30$ 

1339: (filled gray squares, orange in the online edition),

1340: $30 < R_{\rm apo} \leq 35$ 

1341: (filled gray pentagons, turquoise in the online edition),

1342: and  $R_{\rm apo} > 35$ 

1343: (filled gray circles, blue in the online edition).

1344: The stars selected according to these definitions are highlighted in

1345: the lower two panels.

1346: [Please see the electronic edition for a color version of this figure.]

1347: }

1348: \end{figure}

1349: 

1350: It is somewhat surprising that even a couple of stars in our outer halo

1351: population are found at such a high metallicity with ``standard'' 

1352: elemental abundance ratios for their metallicity.

1353: These stars, HIP~19814 and HIP~117041, with metallicities 

1354: [Fe/H]~$= -0.71$ \citep{stephens02} and $-$0.88 \citep{fulbright00}, 

1355: are on orbits extending 

1356: to $R_{\rm apo} = 27^{+4}_{-3}$~kpc and $52^{+9}_{-1}$~kpc, respectively.

1357: These stars are also on highly eccentric orbits 

1358: ($e = 0.96$ and $e = 0.95$, respectively) and

1359: may be in the metal-rich end of the (outer) halo MDF.

1360: The uncertainty in $R_{\rm apo}$ for HIP~19814 could marginally demote this

1361: star from outer halo membership; even so, its Galactic orbit would

1362: remain eccentric.

1363: 

1364: It has long been common practice to 

1365: assume that ancient metal-poor stars do not accrete any appreciable

1366: amount of metals (certainly not enough to enrich a metal-free star

1367: to [Fe/H]~$\sim -4.0$)

1368: from passage through the gas-rich Galactic 

1369: disk (see commentary on this subject by, e.g., 

1370: \citealt{yoshii81}, \citealt{iben83}, and \citealt{frebel08b}.). 

1371: If this is so, 

1372: then the composition and Galactic orbits of these stars suggest that 

1373: significant metal enrichment ($\sim 1/5$ to $1/8$ Solar Fe) may have

1374: occurred in some localized regions far from the present Galactic disk.

1375: This is in qualitative agreement (but perhaps not quantitative, 

1376: since our local sample of stars may not be representative of the 

1377: bulk of the stellar halo; see \S~\ref{kfa})

1378: with the halo chemical evolution model

1379: presented by \citet{tumlinson06}, who found that some stars in 

1380: the metal-rich end of the halo MDF were forming within the first

1381: few hundred million years after star formation began.

1382: 

1383: A wide diversity of stellar orbits and chemical compositions is

1384: found in our outer halo population, which is strong evidence that

1385: a significant fraction of the halo was formed from the conglomeration

1386: of small fragments representing a variety of nucleosynthetic

1387: enrichment scenarios \citep[e.g.,][]{searle78}.

1388: 

1389: 

1390: \subsection{Relationship to the Inner and Outer Halo Globular Clusters}

1391: \label{globular}

1392: 

1393: 

1394: Globular clusters can be classified according to their Galactic

1395: orbital parameters, traditionally defined such that 

1396: ``outer halo'' clusters have orbits that take them to much greater

1397: radii from the Galactic center than ``inner halo'' clusters.

1398: Might we learn any additional information by classifying clusters

1399: according to the inner and outer halo population definitions given

1400: in \S~\ref{halokinematics}?

1401: We use a sample of 25 globular clusters with measured distances

1402: and space velocities compiled from the literature by \citet{pritzl05},

1403: who derived mean abundance ratios for each cluster from recent

1404: high-resolution spectroscopic analyses of individual stars in each cluster.

1405: The cluster positions are taken from the most recent version

1406: of the \citet{harris96} catalog (February 2003), and the cluster 

1407: velocities are taken from a series of papers by 

1408: \citet{dinescu99,dinescu00,dinescu01,dinescu03}.

1409: We compute orbital parameters using our model for the Galactic potential.

1410: Table~\ref{gctab} displays the adopted distances and velocities

1411: and our derived orbital parameters for this set of clusters.

1412: Four clusters match our inner halo kinematic criteria 

1413: (M4, M71, NGC~6397, and NGC~6752), ranging in metallicity from

1414: $-2.0 \leq$~[Fe/H]~$\leq -0.7$.

1415: Eight clusters match our outer halo kinematic criteria

1416: (M3, M30, M68, NGC~288, NGC~362, NGC~5466, Pal~5, and Pal~12),

1417: ranging in metallicity from $-2.4 \leq$~[Fe/H]~$-0.7$.

1418: Pal~12 has been conclusively identified as having been stripped

1419: from the Sagittarius dSph;

1420: excluding Pal~12, our outer halo 

1421: globular clusters span the metallicity range $-2.4 \leq$~[Fe/H]~$-1.3$.

1422: 

1423: In Figure~\ref{globabund1} we compare the [Mg/Fe], [Ca/Fe], and

1424: [Ti/Fe] abundance ratios for our inner and outer halo globular clusters

1425: and our inner and outer halo field stars. 

1426: In all three cases, the abundance ratios of the 

1427: inner halo globular clusters obey the same trends and degree of scatter

1428: defined by the inner halo field stars.

1429: The outer halo globular clusters likewise follow the trends and scatter

1430: of the outer halo field stars, with a few exceptions.

1431: [Mg/Fe] and [Ca/Fe] are marginally low (but still super-Solar)

1432: in NGC~5466, though they

1433: have been derived from a single Cepheid variable \citep{mccarthy97} 

1434: and should be treated with some caution; even so, NGC~5466 is 

1435: within the scatter expected for an outer halo cluster.

1436: [Ti/Fe] is Solar in M68, which helped lead \citet{lee05} to postulate

1437: that M68 may have sampled an IMF biased toward higher masses, where the

1438: Mg and Si overabundances were produced by Type~II SNe.

1439: In this scenario Ti would be primarily produced by by 

1440: lower mass SN~Ia along with the Fe-peak elements.

1441: (See also the extensive discussion in \citealt{lee02}, 

1442: who identified constant [Ca/Fe] but 

1443: decreasing [Si/Fe] and increasing [Ti/Fe] ratios with 

1444: \textit{current} Galactocentric radius in their sample, and

1445: \citealt{gratton04}, who also reproduced this result when using 

1446: $R_{\rm apo}$.)

1447: Finally, M68 does not appear to be associated with the Canis Major

1448: dSph \citep{pritzl05}, as had been suggested by earlier

1449: models \citep{martin04}; however, \citet{pritzl05} have suggested

1450: that it may have an extragalactic origin based on its younger

1451: age, high [Si/Ti] ratio, and high prograde rotational velocity.

1452: 

1453: \begin{figure}

1454: \epsscale{1.15}

1455: \plotone{f16.eps}

1456: \caption{

1457: \label{globabund1}

1458: Mean [Mg/Fe], [Ca/Fe], and [Ti/Fe] abundance ratios for globular 

1459: clusters with kinematics like those that define our

1460: inner and outer halo (field star) populations.

1461: Globular clusters with kinematics like the inner halo 

1462: are indicated by the large, open gray circles (red in the online edition), and 

1463: globular clusters with kinematics like the outer halo

1464: are indicated by the large, open black triangles (blue in the online edition).

1465: The triangle marked with an ``X'' indicates Pal~12,

1466: which has been conclusively identified as a cluster

1467: accreted from the Sagittarius dSph.

1468: All other symbols are the same as in Figure~\ref{abundplot1}.

1469: The Solar ratios are indicated by the dotted lines.

1470: [Please see the electronic edition for a color version of this figure.]

1471: }

1472: \end{figure}

1473: 

1474: Figure~\ref{globabund2} displays the [Y/Fe], [Ba/Fe], and [Eu/Fe]

1475: ratios for our inner and outer halo field stars and globular clusters.

1476: The inner \textit{and} outer halo globular clusters possess 

1477: [Ba/Fe] and [Eu/Fe] ratios that very closely follow these ratios in

1478: the inner halo field stars with very small scatter.

1479: To some degree, this reflects the fact that we have represented

1480: the abundance ratios by means rather than the scatter intrinsic

1481: from one star to another within a given cluster; however,

1482: this scatter is much smaller (typically $\lesssim 0.5$~dex) 

1483: than that found for field stars ($\gtrsim 2$~dex), so it cannot

1484: tell the full story (cf., e.g., \citealt{sneden97,sneden00},

1485: \citealt{ivans01}, \citealt{gratton04}, \citealt{yong08}).

1486: This theme---also present in recent reviews of globular cluster

1487: abundances \citep{gratton04,sneden04}---suggests 

1488: that these globular cluster stars formed from a 

1489: homogenized ISM much like the field stars of the inner halo.

1490: In this sense, the abundance trends traced by our inner halo population

1491: and the globular cluster population may represent a time-averaged

1492: set of chemical yields for a metal-poor stellar population.

1493: If the earliest generations of stars 

1494: pre-enriched the ISM from which the present stars formed, 

1495: this would also explain the lack of stars with

1496: [Fe/H]~$\lesssim -2.5$ in the inner halo and globular clusters.

1497: 

1498: \begin{figure}

1499: \epsscale{1.15}

1500: \plotone{f17.eps}

1501: \caption{

1502: \label{globabund2}

1503: Mean [Y/Fe], [Ba/Fe], and [Eu/Fe] abundance ratios for globular 

1504: clusters with kinematics like those that define our

1505: inner and outer halo populations.

1506: Symbols are the same as in Figure~\ref{globabund1}.

1507: [Please see the electronic edition for a color version of this figure.]

1508: }

1509: \end{figure}

1510: 

1511: The [Y/Fe] ratios implore us to exercise some caution with this

1512: interpretation.

1513: Mean [Y/Fe] ratios for the five outer halo clusters with kinematic and

1514: abundance data are consistently lower than the mean [Y/Fe] ratios for 

1515: the three inner halo clusters

1516: ($\langle$[Y/Fe]$\rangle _{\rm outer} = -0.38$, $\sigma_{\rm mean} = 0.13$; 

1517: $\langle$[Y/Fe]$\rangle _{\rm inner} = -0.08$, $\sigma_{\rm mean} = 0.11$)

1518: These differences are less apparent in Figure~\ref{globabund3}, 

1519: where we display the [Ba/Y] and [Ba/Eu] ratios, yet four of the five

1520: outer halo clusters do have super-Solar [Ba/Y] ratios.

1521: This might suggest a decreased contribution from very massive

1522: SNe (see \S~\ref{rspro}), 

1523: although analysis of a more homogeneous dataset including

1524: abundances for more clusters with kinematic information would be necessary

1525: to draw any robust conclusions.

1526: 

1527: \begin{figure}

1528: \epsscale{1.15}

1529: \plotone{f18.eps}

1530: \caption{

1531: \label{globabund3}

1532: Mean [Ba/Y] and [Ba/Eu] abundance ratios for globular 

1533: clusters with kinematics like those that define our

1534: inner and outer halo populations.

1535: Symbols are the same as in Figure~\ref{globabund1}.

1536: [Please see the electronic edition for a color version of this figure.]

1537: }

1538: \end{figure}

1539: 

1540: Finally, in Figure~\ref{globtrends} we plot the [Ca/Fe], [Ti/Fe],

1541: [Y/Fe], and [Ba/Fe] ratios as a function of $R_{\rm apo}$, 

1542: $|Z_{\rm max}|$, and $V_{\phi}$.

1543: Excluding Pal~12, [Ca/Fe] exhibits no 

1544: trends with these kinematic properties, while earlier reports of 

1545: a [Ti/Fe] trend with $R_{\rm apo}$ appear less secure with this dataset.

1546: No obvious trends of [Ba/Fe] with kinematics are visible.

1547: Only Y-deficient ([Y/Fe]~$<-0.2$) clusters are found at 

1548: $R_{\rm apo} > 10$~kpc or $|Z_{\rm max}| > 5$~kpc, including Pal~12.

1549: Pal~12 displays the low [$\alpha$/Fe] ratios and extreme kinematics

1550: associated with being a captured cluster.\footnote{

1551: Pal~12 has low Mg, Ca, and Ti, which closely correspond to the 

1552: notoriously low Mg, Ca, and Ti abundances of the Sagittarius dSph

1553: and its associated globular cluster system 

1554: \citep{bonifacio03,tautvaisiene04,cohen04a,sbordone07},

1555: although the low-metallicity globular clusters associated with 

1556: the Sagittarius dSph have elevated [$\alpha$/Fe] ratios 

1557: \citep{mottini08}.}

1558: Could these properties be used to diagnose additional clusters that have 

1559: been captured from dSphs?

1560: Given the inhomogeneous abundance analysis methods, 

1561: wide range of number of stars examined in each cluster,

1562: abundance dispersions within clusters,

1563: and range of mean elemental abundance ratios for clusters already examined,

1564: we suspect that these diagnostics may be more useful for confirming

1565: capture proposals based on kinematic evidence rather than as 

1566: a priori criteria (such as the case made for Pal~12 by \citealt{cohen04a}).

1567: 

1568: \begin{figure*}

1569: \epsscale{1.00}

1570: \plotone{f19.eps}

1571: \caption{

1572: \label{globtrends}

1573: [Ca/Fe], [Ti/Fe], [Y/Fe], and [Ba/Fe] ratios for globular clusters as a 

1574: function of $R_{\rm apo}$, $|Z_{\rm max}|$, and $V_{\phi}$.

1575: Globular clusters are indicated by the large, open black

1576: squares (green in the online edition)

1577: and our stellar sample is indicated by the small gray dots.

1578: Pal~12 is indicated by the square marked with an ``X''.

1579: The Solar ratios are indicated by the dotted lines.

1580: [Please see the electronic edition for a color version of this figure.]

1581: }

1582: \end{figure*}

1583: 

1584: 

1585: \subsection{Is the Inner Halo a Remnant of an Accreted Dwarf Galaxy?}

1586: 

1587: 

1588: The orbital characteristics we have used to identify members of 

1589: the inner halo population are reminiscent of a clump of stars

1590: found with $e \sim 0.9$, [Fe/H]~$\sim -1.7$,

1591: and near-zero (or perhaps just slightly prograde) net rotation

1592: by \citet{chiba00}.

1593: They speculated that a significant fraction of the stars in 

1594: this clump may have formed from infalling gas with this metallicity.

1595: The very small abundance scatter for other species at a given [Fe/H] 

1596: in our inner halo population supports this interpretation.

1597: \citet{dinescu02} articulated that a slight retrograde

1598: clump of stars spanning $-2.0<$~[Fe/H]~$-1.5$ in the \citet{beers00}

1599: data (employed also by \citealt{chiba00}) could be reconciled

1600: with stellar debris associated with $\omega$~Centauri,

1601: the stripped core of an accreted dSph \citep[e.g.,][]{norris97,dinescu99}

1602: that has a similar peak in its MDF \citep{norris96}.

1603: Could our inner halo population share a similar origin with this 

1604: $\omega$~Cen debris

1605: (which could also help to place age constraints on the inner halo)?

1606: Comparing the chemical abundances of $\omega$~Cen giants \citep{norris95}

1607: with our inner halo sample reveals notable differences in 

1608: Ca, Ti, Ni, Y, and Ba trends with metallicity, 

1609: effectively dismissing this hypothesis.

1610: Furthermore, it is unlikely that a sizable fraction

1611: of the entire stellar halo of the Milky Way Galaxy (which 

1612: we associate with the inner halo population) was composed of

1613: stars accreted from a single dSph.

1614: 

1615: The inner halo abundances might still suggest that

1616: this population resembles a large-scale stellar stream

1617: \citep[e.g.,][and references therein]{kepley07,helmi08}.

1618: While the spatial density distribution of stellar streams has long

1619: since dissipated, their velocity space density remains more intact

1620: \citep{helmi99}.

1621: In Figure~\ref{angmomplot} we show our inner and outer halo populations

1622: expressed in terms of their angular momenta 

1623: $J_{X} = yW - Vz$, $J_{Y} = zU - Wx$, $J_{Z} = xV - Uy$, and

1624: $J_{\perp} = (J_{X}^{2} + J_{Y}^{2})^{1/2}$ 

1625: (per unit mass, 

1626: where [$x$, $y$, $z$] and [$U$, $V$, $W$] are the positions

1627: and velocities, respectively, of a star in a left-handed 

1628: Galactocentric frame).

1629: The components of the angular momentum for each star are 

1630: also listed in Table~\ref{startab}.

1631: The outer halo stars are preferentially found (as per our 

1632: selection criteria) at low values of $J_{Z}$ and high values of

1633: $J_{\perp}$.

1634: In contrast to the outer halo,

1635: the inner halo stars occupy a limited region of the diagram, and

1636: this region appears as part of the continuous progression from 

1637: stars on disk-like orbits to stars on orbits that take them far

1638: from the Galactic center.  

1639: In this regard the inner halo certainly looks like a more coherent

1640: kinematic component of the halo 

1641: (even if it is too large to be realistically

1642: classified as a stellar ``stream'') than the outer halo, which 

1643: looks like a smattering of stars on completely unrelated orbits.

1644: 

1645: \begin{figure}

1646: \epsscale{1.15}

1647: \plotone{f20.eps}

1648: \caption{

1649: \label{angmomplot}

1650: Inner and outer 

1651: halo members as a function of their angular momentum components.

1652: $J_{\perp}$ is defined as $(J_{X}^{2} + J_{Y}^{2})^{1/2}$.

1653: Symbols are the same as in Figure~\ref{innerouterplot}.

1654: [Please see the electronic edition for a color version of this figure.]

1655: }

1656: \end{figure}

1657: 

1658: In light of these arguments, it is perhaps even more remarkable

1659: that, of the eight [X/Fe] ratios considered

1660: for our halo populations in the previous sections, 

1661: not a single abundance ratio of any inner halo star

1662: deviates even minimally from 

1663: the well-defined mean abundance trends of the inner halo population

1664: (though [Na/Fe] shows an overall greater amount of scatter).

1665: Even the heavy \ncap\ species obey a tight correlation with [Fe/H],

1666: an abundance pattern rarely seen at [Fe/H]~$\lesssim -2.0$,

1667: even in dSphs (e.g., \citealt{shetrone03}, A.\ Frebel et al., in prep.).

1668: This is strong evidence for one of two possibilities.

1669: Either the stars in our inner halo population

1670: formed from a very homogenized ISM,

1671: or the time-averaged abundance yields of various regions of the

1672: stellar halo are nearly identical.

1673: Both scenarios could also explain the apparent lack of very

1674: low metallicity stars in the inner halo population, since 

1675: it may have been common practice for stars with [Fe/H]~$\ll -2.0$ 

1676: to form in regions of the halo where chemical enrichment was still

1677: governed by local SNe events, rather than the time-averaged

1678: yields of many SNe \citep[e.g.,][]{argast00}.

1679: 

1680: 

1681: \subsection{Chemical Signatures of Possible Accretion Events}

1682: \label{accretion}

1683: 

1684: 

1685: \subsubsection{Blue Metal-Poor Stars}

1686: 

1687: 

1688: Blue metal-poor (BMP) stars are the field analogs of the 

1689: blue straggler stars found in clusters, 

1690: except BMP's are believed to be formed by 

1691: binary mass-transfer rather than stellar mergers 

1692: \citep{preston00,sneden03b,carney05}.

1693: \citet{preston94} have suggested that BMP stars 

1694: may signify accretion events.

1695: Based on new binary orbital solutions and analysis of their 

1696: chemical compositions, however, \citet{preston00} and \citet{sneden03b} 

1697: suggested that only the radial velocity (RV) constant BMP stars

1698: may be intermediate-age stars that have been accreted, 

1699: perhaps from satellite dSph systems.

1700: 

1701: Of the 175 stars examined by \citet{preston94}, two are present 

1702: in our sample, \mbox{CS~22966--043} and \mbox{CS~22941--012}.

1703: \mbox{CS~22966--043}, 

1704: a long period binary with a nearly circular binary orbit 

1705: ($P = 317$~days, $e = 0.1$; \citealt{preston00}),

1706: possesses a heretofore unique mix of the $\alpha$ elements

1707: for a star with [Fe/H]~$=-1.91$: 

1708: [Mg/Fe]~$=-0.65$, [Ca/Fe]~$=-0.24$, and [Ti~\textsc{ii}/Fe]~$= +0.49$

1709: \citep{ivans03}.

1710: \mbox{CS~22941--012}, a RV-constant star \citep{preston00}, 

1711: was employed as a BMP, $\alpha$-normal 

1712: comparison to \mbox{CS~22966--043} by \citet{ivans03}.

1713: \mbox{CS~22966--043} barely did not pass our proper motion requirement.

1714: If we relax this requirement and naively adopt a photometric distance,

1715: \mbox{CS~22966--043} is on a very extreme orbit, only circling the 

1716: Galactic center 4--8 times over 10 Gyr

1717: ($R_{\rm apo} = 152^{+65}_{-43}$~kpc and

1718: $|Z_{\rm max}| = 75^{+26}_{-13}$~kpc).

1719: \mbox{CS~22941--012}, in contrast, has orbital parameters 

1720: $R_{\rm apo} = 26^{+3}_{-6}$~kpc and 

1721: $|Z_{\rm max}| = 4.3\pm0.6$~kpc.

1722: While the evolutionary origin of BMP binary stars is not in doubt,

1723: it is interesting that \mbox{CS~22966--043}, a star with such 

1724: a bizarre chemical composition, has an extreme orbit

1725: that is consistent with the accretion scenario.

1726: Meanwhile the RV-constant BMP star, \mbox{CS~22941--012},

1727: appears to be on a more normal outer halo orbit.

1728: 

1729: 

1730: \subsubsection{Dwarf Spheroidal Systems}

1731: 

1732: 

1733: The number of dSph systems with known Galactic orbital

1734: parameters is growing, thanks to careful measurements of their

1735: proper motions.

1736: The majority of the present-day dSphs with known proper motions and 

1737: orbital parameters do not approach closer than 

1738: a few tens of kpc from the Galactic center or the Solar neighborhood

1739: (Canis Major: $R_{\rm peri} = 10.5\pm0.9$~kpc, \citealt{dinescu05};

1740: Carina: $R_{\rm peri} = 20^{+43}_{-17}$~kpc, \citealt{piatek03};

1741: Fornax: $R_{\rm peri} = 138\pm19$~kpc, \citealt{dinescu04};

1742: $R_{\rm peri} = 118^{+19}_{-52}$~kpc, \citealt{piatek07};

1743: Sculptor: $R_{\rm peri} = 120\pm51$~kpc, \citealt{dinescu04};

1744: $R_{\rm peri} = 68^{+15}_{-37}$~kpc, \citealt{piatek06};

1745: Ursa Minor: $R_{\rm peri} = 40^{+36}_{-30}$~kpc, \citealt{piatek05}).

1746: Thus it is unlikely that these specific systems could have contributed

1747: any significant fraction of the local stellar halo.

1748: Additionally, the Leo~II dSph system has not undergone significant

1749: tidal interaction with the Milky Way Galaxy, is currently at a very

1750: great distance (218~kpc), and therefore might be on a more circular 

1751: Galactic orbit that does not approach near to the Galactic center

1752: \citep{siegel08}.

1753: The orbital periods of the dSph systems are typically several Gyr, implying

1754: that they will have reached their perigalactica no more than a few times,

1755: and even the systems that do venture near the Solar radius would

1756: only spend a small fraction of their orbital periods there.

1757: 

1758: Not coincidentally, this conclusion was also reached from 

1759: comparison of the detailed abundances of several of these same dSphs 

1760: (Carina, Fornax, Leo~II, Sculptor, and Ursa Minor) 

1761: with halo stars of the same metallicity ranges \citep{venn04,shetrone08}.

1762: One exception is the Sagittarius dSph, whose leading tidal

1763: arm has deposited debris within a few kpc of the present

1764: Solar neighborhood \citep{majewski03}.

1765: The stars that Sagittarius is presently contributing to the halo

1766: are different than the stars in its residual core

1767: \citep{bellazzini06,chou07,siegel07},

1768: signaling that dSph systems of the present day may not resemble the stars 

1769: that they have already lost to the stellar halo of the Galaxy.

1770: 

1771: While the growing consensus points to few accreted dSph stars 

1772: constituting the local metal-poor stellar halo,

1773: chemical and kinematic analysis of

1774: large samples of in situ halo stars will be needed to determine

1775: if a significant fraction of the true outer halo (e.g., stars that

1776: never approach nearer to the Galactic center than, say, 20~kpc) 

1777: is comprised of the remnants of former or present dSph systems.

1778: This is a daunting observational challenge, but the results from such

1779: a study would be extremely interesting.

1780: 

1781: 

1782: \section{Kinematics as a Function of Abundances}

1783: \label{kfa}

1784: 

1785: 

1786: \subsection{$\alpha$-poor Stars}

1787: \label{alphapoor}

1788: 

1789: 

1790: Most metal-poor stars exhibit [$\alpha$/Fe] ratios that are

1791: 0.3--0.4~dex above the Solar ratio 

1792: \citep[e.g.,][]{wallerstein62,edvardsson93,mcwilliam95b}.

1793: A handful of metal-poor stars, however, have [$\alpha$/Fe] ratios that

1794: are significantly lower than the standard plateau,

1795: and a few such stars have [$\alpha$/Fe] ratios

1796: that are well below the Solar ratio 

1797: \citep[e.g.,][]{carney97,ivans03}.  

1798: Similar $\alpha$ deficiencies have been observed in 

1799: several nearby dSph galaxies

1800: \citep{bonifacio00a,shetrone01,shetrone03,tolstoy03,fulbright04b,

1801: sadakane04,geisler05,sbordone07,koch08,shetrone08},

1802: prompting speculation that the $\alpha$-poor stars may be 

1803: signatures of past accretion of dSphs.

1804: Furthermore, stars associated with the Sagittarius dSph, 

1805: which is presently interacting with the Milky Way 

1806: \citep{ibata94,majewski03}, are very dissimilar to field stars

1807: and are decidedly underabundant in $\alpha$- and Fe-peak elements

1808: (at least in the metallicity range covered, $-1.0<$~[Fe/H]~$<+0.0$;

1809: \citealt{bonifacio00a,mcwilliam05,sbordone07}). 

1810: Only a few $\alpha$-poor field stars are known, and for this

1811: small subset of the halo the accretion hypothesis 

1812: deserves additional scrutiny.

1813: With the recent report of a handful of new $\alpha$-poor 

1814: metal-poor stars \citep{barklem05} from the stellar 

1815: content of the Hamburg-ESO Survey \citep{frebel06,christlieb08},

1816: we are in a position to reevaluate the kinematic properties 

1817: of these stars to search for clues of their origin.

1818: 

1819: In the top panel of Figure~\ref{alphapoorplot} 

1820: we identify two classes of $\alpha$-poor stars: 

1821: those with $+0.0 \leq$~[Mg/Fe]~$<+0.1$ (``$\alpha$-deficient'')

1822: and those with [Mg/Fe]~$<+0.0$ (``$\alpha$-poor'').

1823: For purposes of this analysis we also restrict our sample to 

1824: stars with [Fe/H]~$<-1.0$.

1825: In the bottom panels of Figure~\ref{alphapoorplot} we display the

1826: kinematic and orbital properties of these stars.

1827: They share no common region in phase space.

1828: Several orbits extend to large distances above the Galactic plane 

1829: or to large radii, but no kinematic signature is preferred.

1830: This suggests that

1831: at least a very small fraction of Galactic halo stars formed in

1832: chemically inhomogeneous regions that were deficient (and,

1833: allegedly, deficient by varying degrees) in the $\alpha$ elements.

1834: \citet{ivans03} performed an exhaustive comparison of the abundance patterns 

1835: of three $\alpha$-poor stars to Type~Ia SN models.

1836: They concluded that

1837: these stars seem to have increased contributions from Type~Ia yields

1838: (relative to other stars at their metallicity), yet they cautioned that

1839: no model could reproduce the overall abundance patterns of these stars.

1840: They speculated that these stars may have been among the earliest to 

1841: from from Type~Ia products, though they could also not rule out the

1842: accretion hypothesis.

1843: 

1844: \begin{figure}

1845: \epsscale{1.15}

1846: \plotone{f21.eps}

1847: \caption{

1848: \label{alphapoorplot}

1849: Defining stars of low [Mg/Fe].

1850: Restricting our Mg-poor sample to stars with [Fe/H]~$< -1.0$, 

1851: filled squares (orange in the online edition) represent stars with 

1852: $+0.0 <$~[Mg/Fe]~$<+0.1$ and 

1853: filled triangles (turquoise in the online edition) 

1854: represent stars with [Mg/Fe]~$<+0.0$.

1855: The stars selected according to these definitions are highlighted in the

1856: lower four panels.

1857: [Please see the electronic edition for a color version of this figure.]

1858: }

1859: \end{figure}

1860: 

1861: A number of studies

1862: (\citealt{unavane96}, \citealt{fulbright02}, \citealt{venn04},

1863: \citealt{pritzl05}, and \citealt{geisler07}, in addition to those listed above)

1864: have concluded that the chemical compositions of dSph stars and halo giants 

1865: are sufficiently different, and only a small fraction of the Milky Way stellar

1866: halo can be composed of stars accreted from dSphs chemically

1867: similar to those surviving to the present day.

1868: Our result does not imply that the $\alpha$-poor stars could not

1869: have formed in dSph systems that were assumed into the Milky Way.

1870: Since many of these stars are on eccentric orbits that take them 

1871: to large Galactic radii, we may surmise that either these inhomogeneous

1872: regions were located at large Galactic radii (perhaps before the 

1873: bulk of the Galaxy formed) or the stars were

1874: formed in (separate?) dSphs and accreted at some time in the past.

1875: 

1876: 

1877: \subsection{Stars with Specific \ncap\ Enrichment Signatures}

1878: \label{rspro}

1879: 

1880: 

1881: Can the \ncap\ material present in these metal-poor stars

1882: be used to identify any preferred kinematics of the parent

1883: clouds from which they formed?

1884: Given the different mass ranges of the stars commonly thought

1885: to produce \ncap\ material, this is an attractive possibility 

1886: if the IMF's of the parent clouds differ significantly.

1887: 

1888: In the top panel of Figure~\ref{rsplot}, we identify stars

1889: that have either a nearly pure \spro\ or \rpro\ [Ba/Eu] ratio.

1890: Pure \rpro\ enrichment (or, viewed another way, lack of 

1891: appreciable \spro\ enrichment) 

1892: occurs over a very wide range of metallicity

1893: ($-3.0 <$~[Fe/H]~$-0.4$, or a factor of $\approx$~400 in Fe/H---a 

1894: fact which is remarkable in its own right!).

1895: Our sample includes a rather small number of stars exhibiting

1896: a pure \spro\ signature; this is a consequence of our (somewhat

1897: arbitrary) choice of literature data to include.

1898: In the lower panels of Figure~\ref{rsplot}, we show the kinematic properties

1899: of these stars.

1900: The stars dominated by \rpro\ \ncap\ enrichment show no preferred

1901: kinematic properties---they scatter over all values of 

1902: $V_{\phi}$, $|Z_{\rm max}|$, $R_{\rm peri}$, and $R_{\rm apo}$.

1903: The three stars exhibiting a pure \spro\ signature do not 

1904: indicate any kinematic preferences, either.

1905: 

1906: \begin{figure}

1907: \epsscale{1.15}

1908: \plotone{f22.eps}

1909: \caption{

1910: \label{rsplot}

1911: Defining stars exhibiting pure \rpro\ and pure \spro\ 

1912: [Ba/Eu] ratios.

1913: The lower dashed line (orange in the online edition) 

1914: represents the pure \rpro\ ratio ($-$0.67) and 

1915: the upper dashed line (turquoise in the online edition)

1916: represents the pure \spro\ ratio ($+$0.63)

1917: as predicted by \citet{simmerer04}.

1918: Filled triangles (orange in the online edition)

1919: represent stars with [Ba/Eu]~$< -0.55$ and

1920: filled squares (turquoise in the online edition) 

1921: represent stars with [Ba/Eu]~$>+0.50$.

1922: The stars selected according to these definitions are highlighted in the

1923: lower four panels.

1924: [Please see the electronic edition for a color version of this figure.]

1925: }

1926: \end{figure}

1927: 

1928: The main \rpro\ produces a robust abundance pattern

1929: for species with $Z \geq 56$ 

1930: \citep[e.g.,][]{sneden03a,ivans06,cowan06,roederer08}, 

1931: but it has become clear over the

1932: last decade that this unique signature does not apply to 

1933: the lighter \ncap\ nuclei, such as the Sr-Y-Zr group.

1934: Close analyses of two metal-poor stars, \mbox{HD~88609}

1935: \citep{honda07} and \mbox{HD~122563} \citep{honda06}, reveal

1936: an abundance pattern that gradually declines with increasing $Z$,

1937: producing high ratios between light and heavy \ncap\ species

1938: (see also \citealt{mcwilliam95b}, \citealt{wasserburg96},

1939: \citealt{johnson02}, \citealt{aoki05},

1940: \citealt{lai07}, and \citealt{lai08}).

1941: This pattern cannot be associated with the main \rpro,

1942: the main \spro, or the weak \spro, and could be representative of---for

1943: example---a light element primary process \citep{travaglio04},

1944: a rapid proton-capture process \citep{wanajo06a},

1945: a weak \rpro\ \citep{wanajo06b},

1946: a cold \rpro\ \citep{wanajo07},

1947: hypernovae \citep{qian08}, or 

1948: a high entropy wind from Type~II SNe \citep{farouqi08}.

1949: This multiplicity of scenarios, with a variety of proposed 

1950: SN mass ranges responsible for the creation of the

1951: light \ncap\ elements, demands additional observational constraints.

1952: Once again, if the IMF of a certain proto-stellar metal-poor

1953: cloud was significantly different from other clouds, this could

1954: manifest itself in the derived [Ba/Y] ratios.

1955: In the top panel of Figure~\ref{weakrplot} we identify stars

1956: with low [Ba/Y] ratios ([Ba/Y]~$<-0.55$) and low [Ba/Fe] ratios

1957: ([Ba/Fe]~$<-0.50$), similar to the abundances found in 

1958: \mbox{HD~88609} and \mbox{HD~122563}.\footnote{ 

1959: We note, as \citet{honda07} have, that high

1960: [Sr/Ba] or [Y/Ba] ratios do not uniquely define the abundance pattern 

1961: associated with this process, since \mbox{HD~88609} was

1962: selected for analysis based on its [Sr/Ba] ratio, 

1963: identical to that of \mbox{HD~122563}.}

1964: Again, though, these stars exhibit no clear kinematic properties,

1965: although they may slightly prefer eccentric orbits, which 

1966: could point to a common---though yet poorly constrained---origin.

1967: (The common retrograde orbits are a result of the narrow

1968: [Fe/H] range of this sub-sample, and are not of physical consequence.)

1969: 

1970: \begin{figure}

1971: \epsscale{1.15}

1972: \plotone{f23.eps}

1973: \caption{

1974: \label{weakrplot}

1975: Defining stars with excesses of the light \ncap\ elements and

1976: deficiencies of the heavy \ncap\ elements.

1977: The filled squares (orange in the online edition)

1978: represent stars with [Ba/Y]~$<-0.55$ and [Ba/Fe]~$<-0.50$.

1979: (The [Ba/Y] ratio is $-$0.69 and $-$0.68 in \mbox{HD~88609} and 

1980: \mbox{HD~122563}, respectively; \citealt{honda06,honda07}.)

1981: The stars selected according to this definition are highlighted in the

1982: lower four panels.

1983: [Please see the electronic edition for a color version of this figure.]

1984: }

1985: \end{figure}

1986: 

1987: At low metallicities, \ncap\ enrichment is probably a very

1988: localized phenomenon that results in a wide distribution of

1989: \ncap\ abundances, and thus it will be extremely difficult

1990: to identify any associated large-scale kinematic behaviors

1991: of the proto-stellar clouds from which these stars formed.

1992: The potential gain from such a connection, however, justifies

1993: the continued effort toward this goal.

1994: 

1995: 

1996: \subsection{The Most Metal-Poor Stars Known}

1997: \label{ultrampstars}

1998: 

1999: 

2000: We also examine the kinematic properties of the three most

2001: metal-poor stars known, 

2002: \mbox{HE~0107--5240} ([Fe/H]~$=-5.3$; \citealt{christlieb02,christlieb04}),

2003: \mbox{HE~1327--2326} ([Fe/H]~$=-5.96$; \citealt{frebel05,aoki06,frebel08a}), 

2004: and

2005: \mbox{HE~0577--4840} ([Fe/H]~$=-4.75$; \citealt{norris07}).

2006: Proper motions have been reported for all three of these stars

2007: in the UCAC2 or NOMAD catalogs; unfortunately, for 

2008: \mbox{HE~0107--5240} and \mbox{HE~0577--4840} the measurements

2009: are no larger than their uncertainties ($\sim 6$~mas~yr$^{-1}$) and are

2010: therefore not useful for our purposes.

2011: The proper motions for \mbox{HE~1327--2326} are large:

2012: $\mu_{\alpha} = -40\pm1$, $\mu_{\delta} = 58\pm4$~mas~yr$^{-1}$

2013: \citep{zacharias04b}.

2014: We compute a photometric distance of 1160~pc for \mbox{HE~1327--2326}

2015: (assuming it is a subgiant and not a dwarf; see 

2016: \citealt{korn08} and \citealt{frebel08a}), 

2017: in fair agreement with the previous estimate of 1400~pc \citep{aoki06}.

2018: \mbox{HE~1327--2326} lies above the plane of the Galaxy

2019: ($\ell = 314^{\circ}$, $b = +38^{\circ}$) and is moving rapidly

2020: away from the Galactic plane ($U = 221$, $V = 246$, $W=325$, all in \kmsec).

2021: Using these input parameters, \mbox{HE~1327--2326} is only weakly

2022: bound to the gravitational potential of the Galaxy, completing no more than

2023: 2~orbits over 10~Gyr.  

2024: Altering the distance by $\pm$500~pc (much larger than the 20\% uncertainty

2025: assumed in \S~\ref{uncertainties}),

2026: proper motion by $\pm$5~mas~yr$^{-1}$, 

2027: radial velocity by $\pm$10~\kmsec, and 

2028: mass of the Galaxy by $\pm$10\%, we find that this star

2029: completes between 1 and 4~orbits over 10~Gyr.

2030: For these orbits, $R_{\rm peri} = 8.0\pm0.2$~kpc,

2031: $R_{\rm apo} = 405^{+265}_{-186}$~kpc, and

2032: $|Z_{\rm max}| = 160^{+97}_{-45}$.

2033: 

2034: All reasonable variations in the input parameters still point to 

2035: a very extreme orbit for \mbox{HE~1327--2326}, which is perhaps 

2036: not unexpected given the uncharacteristically low metal abundance

2037: of this interesting star.

2038: Precise proper motion measurements for \mbox{HE~0107--5240} and

2039: \mbox{HE~0577--4840} will be necessary to determine whether extreme

2040: orbits are a common characteristic of the most metal-poor stars in 

2041: the Galaxy.

2042: Furthermore, \mbox{HE~1327--2326} spends a very small fraction of

2043: its time in the inner regions of the Galaxy, suggesting, perhaps, that

2044: many more objects with compositions similar to this star

2045: currently reside in the distant realms of the stellar halo.

2046: This again hints to the possibility that

2047: objects in the true outer halo of the Galaxy may bear

2048: little resemblance to those that inhabit the inner few tens of kpc.

2049: 

2050: 

2051: \section{Conclusions}

2052: \label{conclusions}

2053: 

2054: 

2055: We have compiled a large sample of metal-poor 

2056: stars with reported abundances from the literature.

2057: For the subset of these stars with reliable proper motion measurements,

2058: we have computed space velocities and Galactic orbital parameters.

2059: We have used the combination of chemical and kinematic information 

2060: to identify any abundance patterns that are common to stars with 

2061: similar kinematics or identify any kinematic signatures that are common

2062: to stars with similar chemical enrichment patterns. 

2063: Our major results can be summarized as follows.

2064: 

2065: (1) A proper motion selection criterion for our sample 

2066: biases the very metal-poor end of our sample toward stars 

2067: with large proper motions and hence extreme orbits, 

2068: many of which are on retrograde Galactic orbits.

2069: This bias must be borne in mind when interpreting chemical

2070: enrichment patterns in very metal-poor stars.

2071: 

2072: (2) We find no abundance trends with maximum radial distance from

2073: the Galactic center or maximum vertical distance above the Galactic disk.

2074: Current high-resolution abundance analyses are limited to sampling

2075: halo stars that happen to be passing relatively near to the 

2076: Solar neighborhood.

2077: Other studies have found very slight decreases in abundance ratios

2078: with increasing distance from the Galactic center, and this issue may

2079: remain unsettled until large, deep abundance surveys and proper motion

2080: surveys can reach beyond the 

2081: Solar neighborhood into a true in situ sample of the stellar halo.

2082: 

2083: (3) We use only kinematic criteria to define our inner halo

2084: (stars on prograde, eccentric orbits that do not stray beyond 

2085: 15~kpc from the Galactic center or 5~kpc from the Galactic plane)

2086: and outer halo (stars on very retrograde orbits or stars whose

2087: orbits reach more than 25~kpc from the Galactic center or

2088: more than 10~kpc above the Galactic plane) samples, based

2089: on the kinematic properties of these populations derived from

2090: much larger datasets.

2091: In the metallicity regimes where the two populations overlap,

2092: roughly $-2.5 <$~[Fe/H]~$<-1.5$,

2093: the [Mg/Fe] ratio of the outer halo may be lower than 

2094: the inner halo by $\sim$~0.1~dex.

2095: For [Ni/Fe] and [Ba/Fe]

2096: the star-to-star abundance scatter of the inner halo is consistently

2097: smaller than the star-to-star abundance scatter of the outer halo.

2098: The [Na/Fe], [Ca/Fe], [Ti/Fe], and [Y/Fe] 

2099: ratios of both populations show similar levels of scatter.

2100: We do not have enough [Eu/Fe] measurements in our sample to draw

2101: any conclusions from this abundance ratio.

2102: 

2103: (4) Our inner halo population appears chemically homogeneous, 

2104: suggesting that a significant fraction

2105: of the Milky Way stellar halo had a common origin from a 

2106: well-mixed ISM.

2107: 

2108: (5) In contrast, our kinematically diverse outer halo population is 

2109: also chemically diverse, suggesting that another significant fraction

2110: of the Milky Way stellar halo formed in remote regions where

2111: chemical enrichment was dominated by local SN events.

2112: This component is reminiscent of the ``chaotic origin'' for the

2113: Galaxy suggested by the globular cluster data of \citet{searle78}.

2114: 

2115: (6) If we classify globular clusters by these same kinematic 

2116: criteria, all of the inner halo and most of the outer halo clusters 

2117: follow similar (mean) abundance trends with comparable degrees of scatter 

2118: to the inner halo population of field stars.

2119: The chemical similarity of the inner halo and the globular clusters 

2120: may suggest that

2121: these abundance trends represent the time-averaged mean abundances

2122: of a metal-poor ISM pre-enriched by the earliest generations of stars.

2123: We find no mean cluster abundance dependence with increasing radial

2124: distance from the Galactic center or vertically from the Galactic plane.

2125: 

2126: (7) We find no kinematic signature in common to groups of metal-poor

2127: stars with peculiar abundance patters ($\alpha$-poor stars,

2128: stars showing a pure-$s$- or pure-$r$-process \ncap\ enrichment pattern,

2129: or stars with a deficiency of heavy \ncap\ material).

2130: 

2131: (8) Our results do not exclude the possibility that any of these 

2132: individual stars

2133: were accreted by the Milky Way from, e.g., dSph systems;

2134: however, the orbits of many present-day dSphs (especially those 

2135: whose stars have been subject to high-resolution abundance analyses) 

2136: rarely bring them near the Solar radius, 

2137: so they would not be expected to contribute

2138: many stars to the local metal-poor stellar halo.

2139: 

2140: (9) Several individual stars---including the most metal-poor star 

2141: known---and dSph systems whose compositions differ greatly from 

2142: the bulk of the stellar halo in the Solar neighborhood

2143: have long orbital periods ($\sim$ few Gyr) 

2144: and extreme orbital characteristics ($R_{\rm apo} \sim$ few hundred kpc).

2145: If stars like these spend the majority of their time in the distant

2146: regions of the Milky Way stellar halo, this raises the possibility that

2147: many more stars with unusual abundance patterns may occupy the 

2148: true outer halo of the Galaxy, which may have little resemblance

2149: to the local stellar halo.

2150: 

2151: The abundance dataset used for this analysis was compiled from a variety 

2152: of literature sources, and the inherent systematic differences

2153: from one study to another limit our ability to detect more subtle

2154: chemical differences than those described above.

2155: Large, metal-poor stellar samples from which abundances are derived in a 

2156: homogeneous manner are necessary to perform a more detailed 

2157: nucleo-kinematic analysis.

2158: The construction and analysis of such datasets is presently underway.

2159: 

2160: Evidence continues to grow in support of the notion 

2161: that the chemical enrichment history 

2162: of the Milky Way stellar halo is nonuniform on both small and large scales.

2163: While it is unlikely that an exact correlation

2164: between the kinematic properties of a star and its chemical abundance

2165: pattern will ever be identified, 

2166: we may be approaching an era where it is necessary to 

2167: know the kinematic properties of a field halo star in order to 

2168: place its chemical abundance pattern in the proper nucleosynthetic context.

2169: The challenge for future studies will be to articulate the degree 

2170: of chemical dissimilarity in the Milky Way halo kinematic substructure.

2171: 

2172: 

2173: 

2174: \acknowledgments

2175: 

2176: 

2177: It is a pleasure to thank Anna Frebel and Chris Sneden 

2178: for many extensive discussions relating to this topic and 

2179: detailed comments on assorted versions of this manuscript.

2180: I.~U.~R.\ also thanks John Cowan, George Preston, and Matthew Shetrone 

2181: for their helpful suggestions and encouragement, as well as

2182: David Lambert, who originally suggested this literature study.

2183: Additional thanks go to Monique Spite for sending radial velocity

2184: measurements for the ``First Stars'' sample in advance of publication,

2185: Douglas Lin for sharing a copy his orbit integrator code, and 

2186: the anonymous referee for helpful suggestions.

2187: This research would not have been possible without the resources available

2188: from the SIMBAD and VizieR \citep{ochsenbein00} online databases, 

2189: operated at CDS, Strasbourg, France, 

2190: the Two Micron All-Sky Survey, 

2191: and NASA's Astrophysics Data System Bibliographic Services.

2192: Funding has been provided by NSF grant AST~06-07708 (to C.~Sneden).

2193: 

2194: 

2195: 

2196: 

2197: \begin{thebibliography}{}

2198: 

2199: 

2200: \bibitem[Allende Prieto et al.(2007)]{allendeprieto07} Allende Prieto, 

2201: C., et al.\ 2007, ArXiv e-prints, 710, arXiv:0710.5780 

2202: 

2203: \bibitem[Aoki et al.(2005)]{aoki05} Aoki, W., et al.\ 2005, 

2204: \apj, 632, 611 

2205: 

2206: \bibitem[Aoki et al.(2006)]{aoki06} Aoki, W., et al.\ 2006, 

2207: \apj, 639, 897 

2208: 

2209: \bibitem[Argast et al.(2000)]{argast00} Argast, D., Samland, M., 

2210: Gerhard, O.~E., \& Thielemann, F.-K.\ 2000, \aap, 356, 873 

2211: 

2212: \bibitem[Arnone et al.(2005)]{arnone05} Arnone, E., Ryan, S.~G., 

2213: Argast, D., Norris, J.~E., \& Beers, T.~C.\ 2005, \aap, 430, 507 

2214: 

2215: \bibitem[Audouze \& Tinsley(1976)]{audouze76} Audouze, J., \& 

2216: Tinsley, B.~M.\ 1976, \araa, 14, 43 

2217: 

2218: \bibitem[Barklem et al.(2005)]{barklem05} Barklem, P.~S., et al.\ 2005, 

2219: \aap, 439, 129 

2220: 

2221: \bibitem[Beers \& Sommer-Larsen(1995)]{beers95} Beers, T.~C., \& 

2222: Sommer-Larsen, J.\ 1995, \apjs, 96, 175

2223: 

2224: \bibitem[Beers et al.(2000)]{beers00} Beers, T.~C., Chiba, M., 

2225: Yoshii, Y., Platais, I., Hanson, R.~B., Fuchs, B., 

2226: \& Rossi, S.\ 2000, \aj, 119, 2866 

2227: 

2228: \bibitem[Beers \& Christlieb(2005)]{beers05} Beers, T.~C., \& 

2229: Christlieb, N.\ 2005, \araa, 43, 531 

2230: 

2231: \bibitem[Bell et al.(2008)]{bell08} Bell, E.~F., et al.\ 2008, 

2232: \apj, 680, 295 

2233: 

2234: \bibitem[Bellazzini et al.(2006)]{bellazzini06} Bellazzini, M., 

2235: Newberg, H.~J., Correnti, M., Ferraro, F.~R., \& Monaco, L.\ 2006, \aap, 

2236: 457, L21 

2237: 

2238: \bibitem[Bensby et al.(2003)]{bensby03} Bensby, T., Feltzing, S., 

2239: \& Lundstr{\"o}m, I.\ 2003, \aap, 410, 527 

2240: 

2241: \bibitem[Bonifacio et al.(2000a)]{bonifacio00a} Bonifacio, P., Hill, V., 

2242: Molaro, P., Pasquini, L., Di Marcantonio, P., \& Santin, P.\ 

2243: 2000a, \aap, 359, 663 

2244: 

2245: \bibitem[Bonifacio et al.(2000b)]{bonifacio00b} Bonifacio, P., Monai, 

2246: S., \& Beers, T.~C.\ 2000b, \aj, 120, 2065 

2247: 

2248: \bibitem[Bonifacio \& Caffau(2003)]{bonifacio03} Bonifacio, P., \& 

2249: Caffau, E.\ 2003, \aap, 399, 1183 

2250: 

2251: \bibitem[Burbidge et al.(1957)]{burbidge57} Burbidge, E.~M., 

2252: Burbidge, G.~R., Fowler, W.~A., 

2253: \& Hoyle, F.\ 1957, Reviews of Modern Physics, 29, 547 

2254: 

2255: \bibitem[Busso et al.(1999)]{busso99} Busso, M., Gallino, R., \& 

2256: Wasserburg, G.~J.\ 1999, \araa, 37, 239 

2257: 

2258: \bibitem[Carney et al.(1996)]{carney96} Carney, B.~W., Laird, 

2259: J.~B., Latham, D.~W., \& Aguilar, L.~A.\ 1996, \aj, 112, 668 

2260: 

2261: \bibitem[Carney et al.(1997)]{carney97} Carney, B.~W., Wright, 

2262: J.~S., Sneden, C., Laird, J.~B., Aguilar, L.~A., 

2263: \& Latham, D.~W.\ 1997, \aj, 114, 363 

2264: 

2265: \bibitem[Carney et al.(2005)]{carney05} Carney, B.~W., Latham, 

2266: D.~W., \& Laird, J.~B.\ 2005, \aj, 129, 466 

2267: 

2268: \bibitem[Carollo et al.(2007)]{carollo07} Carollo, D., et al.\ 

2269: 2007, \nat, 450, 1020 

2270: 

2271: \bibitem[Carretta et al.(2002)]{carretta02} Carretta, E., Gratton, 

2272: R., Cohen, J.~G., Beers, T.~C., \& Christlieb, N.\ 2002, \aj, 124, 481 

2273: 

2274: \bibitem[Cayrel et al.(2004)]{cayrel04} Cayrel, R., et al.\ 2004, 

2275: \aap, 416, 1117 

2276: 

2277: \bibitem[Chiba \& Yoshii(1998)]{chiba98} Chiba, M., \& Yoshii, Y.\ 

2278: 1998, \aj, 115, 168 

2279: 

2280: \bibitem[Chiba \& Beers(2000)]{chiba00} Chiba, M., \& Beers, T.~C.\ 

2281: 2000, \aj, 119, 2843 

2282: 

2283: \bibitem[Chou et al.(2007)]{chou07} Chou, M.-Y., et al.\ 2007, \apj, 670, 346 

2284: 

2285: \bibitem[Christlieb et al.(2002)]{christlieb02} Christlieb, N., et 

2286: al.\ 2002, \nat, 419, 904 

2287: 

2288: \bibitem[Christlieb et al.(2004)]{christlieb04} Christlieb, N., 

2289: Gustafsson, B., Korn, A.~J., Barklem, P.~S., Beers, T.~C., Bessell, M.~S., 

2290: Karlsson, T., \& Mizuno-Wiedner, M.\ 2004, \apj, 603, 708 

2291: 

2292: \bibitem[Christlieb et al.(2008)]{christlieb08} Christlieb, N., 

2293: Sch{\"o}rck, T., Frebel, A., Beers, T.~C., Wisotzki, L., \& 

2294: Reimers, D.\ 2008, \aap, 484, 721 

2295: 

2296: \bibitem[Cohen(2004)]{cohen04a} Cohen, J.~G.\ 2004, \aj, 127, 1545 

2297: 

2298: \bibitem[Cohen et al.(2004)]{cohen04b} Cohen, J.~G., et al.\ 

2299: 2004, \apj, 612, 1107 

2300: 

2301: \bibitem[Cowan \& Sneden(2006)]{cowan06} Cowan, J.~J., \& Sneden, C.\ 

2302: 2006, \nat, 440, 1151 

2303: 

2304: \bibitem[Cutri et al.(2003)]{cutri03} Cutri, R.~M., et al.\ 

2305: 2003, The IRSA 2MASS All-Sky Point Source Catalog, NASA/IPAC Infrared 

2306: Science Archive  

2307: 

2308: \bibitem[Demarque et al.(2004)]{demarque04} Demarque, P., Woo, 

2309: J.-H., Kim, Y.-C., \& Yi, S.~K.\ 2004, \apjs, 155, 667 

2310: 

2311: \bibitem[Dinescu et al.(1999)]{dinescu99} Dinescu, D.~I., Girard, 

2312: T.~M., \& van Altena, W.~F.\ 1999, \aj, 117, 1792 

2313: 

2314: \bibitem[Dinescu et al.(2000)]{dinescu00} Dinescu, D.~I., 

2315: Majewski, S.~R., Girard, T.~M., \& Cudworth, K.~M.\ 2000, \aj, 120, 1892 

2316: 

2317: \bibitem[Dinescu et al.(2001)]{dinescu01} Dinescu, D.~I., 

2318: Majewski, S.~R., Girard, T.~M., \& Cudworth, K.~M.\ 2001, \aj, 122, 1916 

2319: 

2320: \bibitem[Dinescu(2002)]{dinescu02} Dinescu, D.~I.\ 2002, Omega 

2321: Centauri, A Unique Window into Astrophysics, 265, 365 

2322: 

2323: \bibitem[Dinescu et al.(2003)]{dinescu03} Dinescu, D.~I., Girard, 

2324: T.~M., van Altena, W.~F., \& L{\'o}pez, C.~E.\ 2003, \aj, 125, 1373 

2325: 

2326: \bibitem[Dinescu et al.(2004)]{dinescu04} Dinescu, D.~I., Keeney, 

2327: B.~A., Majewski, S.~R., \& Girard, T.~M.\ 2004, \aj, 128, 687 

2328: 

2329: \bibitem[Dinescu et al.(2005)]{dinescu05} Dinescu, D.~I., 

2330: Mart{\'{\i}}nez-Delgado, D., Girard, T.~M., Pe{\~n}arrubia, J., Rix, H.-W., 

2331: Butler, D., \& van Altena, W.~F.\ 2005, \apjl, 631, L49 

2332: 

2333: \bibitem[Edvardsson et al.(1993)]{edvardsson93} Edvardsson, B., 

2334: Andersen, J., Gustafsson, B., Lambert, D.~L., Nissen, P.~E., \& 

2335: Tomkin, J.\ 1993, \aap, 275, 101 

2336: 

2337: \bibitem[Eggen et al.(1962)]{eggen62} Eggen, O.~J., 

2338: Lynden-Bell, D., \& Sandage, A.~R.\ 1962, \apj, 136, 748 

2339: 

2340: \bibitem[Farouqi et al.(2008)]{farouqi08} Farouqi, K., et al.\

2341: 2008, \apjl, submitted

2342: 

2343: \bibitem[Font et al.(2006)]{font06} Font, A.~S., Johnston, 

2344: K.~V., Bullock, J.~S., \& Robertson, B.~E.\ 2006, \apj, 638, 585 

2345: 

2346: \bibitem[Fowler et al.(1967)]{fowler67} Fowler, W.~A., Caughlan, G.~R., 

2347: \& Zimmerman, B.~A.\ 1967, \araa, 5, 525 

2348: 

2349: \bibitem[Fran{\c c}ois et al.(2007)]{francois07} Fran{\c c}ois, P., 

2350: et al.\ 2007, \aap, 476, 935 

2351: 

2352: \bibitem[Frebel et al.(2005)]{frebel05} Frebel, A., et al.\ 

2353: 2005, \nat, 434, 871 

2354: 

2355: \bibitem[Frebel et al.(2006)]{frebel06} Frebel, A., et al.\ 

2356: 2006, \apj, 652, 1585 

2357: 

2358: \bibitem[Frebel et al.(2008a)]{frebel08a} Frebel, A., Collet, R., 

2359: Eriksson, K., Christlieb, N., \& Aoki, W.\ 2008a, \apj, 684, 588

2360: 

2361: \bibitem[Frebel et al.(2008b)]{frebel08b} Frebel, A., Johnson, J.~L., 

2362: \& Bromm, V.\ 2008b, \mnras, submitted

2363: 

2364: \bibitem[Fulbright(2000)]{fulbright00} Fulbright, J.~P.\ 2000, \aj, 

2365: 120, 1841 

2366: 

2367: \bibitem[Fulbright(2002)]{fulbright02} Fulbright, J.~P.\ 2002, \aj, 

2368: 123, 404 

2369: 

2370: \bibitem[Fulbright(2004)]{fulbright04a} Fulbright, J.~P.\ 2004, 

2371: Origin and Evolution of the Elements, 4, 22

2372: 

2373: \bibitem[Fulbright et al.(2004)]{fulbright04b} Fulbright, J.~P., 

2374: Rich, R.~M., \& Castro, S.\ 2004, \apj, 612, 447 

2375: 

2376: \bibitem[Geisler et al.(2005)]{geisler05} Geisler, D., Smith, 

2377: V.~V., Wallerstein, G., Gonzalez, G., 

2378: \& Charbonnel, C.\ 2005, \aj, 129, 1428 

2379: 

2380: \bibitem[Geisler et al.(2007)]{geisler07} Geisler, D., 

2381: Wallerstein, G., Smith, V.~V., 

2382: \& Casetti-Dinescu, D.~I.\ 2007, \pasp, 119, 939 

2383: 

2384: \bibitem[Gilmore et al.(1989)]{gilmore89} Gilmore, G., Wyse, R.~F.~G., 

2385: \& Kuijken, K.\ 1989, \araa, 27, 555 

2386: 

2387: \bibitem[Gilmore et al.(2002)]{gilmore02} Gilmore, G., Wyse, 

2388: R.~F.~G., \& Norris, J.~E.\ 2002, \apjl, 574, L39 

2389: 

2390: \bibitem[Gratton et al.(2003)]{gratton03} Gratton, R.~G., Carretta, E., 

2391: Desidera, S., Lucatello, S., Mazzei, P., \& Barbieri, M.\ 

2392: 2003, \aap, 406, 131 

2393: 

2394: \bibitem[Gratton et al.(2004)]{gratton04} Gratton, R., Sneden, C., 

2395: \& Carretta, E.\ 2004, \araa, 42, 385 

2396: 

2397: \bibitem[Hanson et al.(1998)]{hanson98} Hanson, R.~B., Sneden, 

2398: C., Kraft, R.~P., \& Fulbright, J.\ 1998, \aj, 116, 1286 

2399: 

2400: \bibitem[Harris(1996)]{harris96} Harris, W.~E.\ 1996, \aj, 112, 1487 

2401: 

2402: \bibitem[Hartwick(1976)]{hartwick76} Hartwick, F.~D.~A.\ 1976, 

2403: \apj, 209, 418 

2404: 

2405: \bibitem[Helmi \& White(1999)]{helmi99} Helmi, A., \& White, S.~D.~M.\ 

2406: 1999, \mnras, 307, 495 

2407: 

2408: \bibitem[Helmi(2008)]{helmi08} Helmi, A.\ 2008, \aapr, 15, 145 

2409: 

2410: \bibitem[Hill et al.(2002)]{hill02} Hill, V., et al.\ 2002, \aap, 387, 560 

2411: 

2412: \bibitem[Honda et al.(2004a)]{honda04a} Honda, S., et al.\ 2004a, 

2413: \apjs, 152, 113 

2414: 

2415: \bibitem[Honda et al.(2004b)]{honda04b} Honda, S., Aoki, W., 

2416: Kajino, T., Ando, H., Beers, T.~C., Izumiura, H., Sadakane, K., 

2417: \& Takada-Hidai, M.\ 2004b, \apj, 607, 474 

2418: 

2419: \bibitem[Honda et al.(2006)]{honda06} Honda, S., Aoki, W., 

2420: Ishimaru, Y., Wanajo, S., \& Ryan, S.~G.\ 2006, \apj, 643, 1180 

2421: 

2422: \bibitem[Honda et al.(2007)]{honda07} Honda, S., Aoki, W., 

2423: Ishimaru, Y., \& Wanajo, S.\ 2007, \apj, 666, 1189 

2424: 

2425: \bibitem[Ibata et al.(1994)]{ibata94} Ibata, R.~A., Gilmore, 

2426: G., \& Irwin, M.~J.\ 1994, \nat, 370, 194 

2427: 

2428: \bibitem[Iben(1983)]{iben83} Iben, I., Jr.\ 1983, Memorie 

2429: della Societa Astronomica Italiana, 54, 321 

2430: 

2431: \bibitem[Ivans et al.(2001)]{ivans01} Ivans, I.~I., Kraft, 

2432: R.~P., Sneden, C., Smith, G.~H., Rich, R.~M., 

2433: \& Shetrone, M.\ 2001, \aj, 122, 1438 

2434: 

2435: \bibitem[Ivans et al.(2003)]{ivans03} Ivans, I.~I., Sneden, C., 

2436: James, C.~R., Preston, G.~W., Fulbright, J.~P., H{\"o}flich, P.~A., Carney, 

2437: B.~W., \& Wheeler, J.~C.\ 2003, \apj, 592, 906 

2438: 

2439: \bibitem[Ivans et al.(2006)]{ivans06} Ivans, I.~I., Simmerer, 

2440: J., Sneden, C., Lawler, J.~E., Cowan, J.~J., Gallino, R., 

2441: \& Bisterzo, S.\ 2006, \apj, 645, 613 

2442: 

2443: \bibitem[Ivezi\'{c} et al.(2008)]{ivezic08} Ivezi\'{c}, \v{Z}., et al.\ 

2444: 2008, \apj, 684, 287

2445: 

2446: \bibitem[Johnson \& Bolte(2002)]{johnson02} Johnson, J.~A., \& Bolte, M.\ 

2447: 2002, \apj, 579, 616 

2448: 

2449: \bibitem[Johnston et al.(1996)]{johnston96} Johnston, K.~V., 

2450: Hernquist, L., \& Bolte, M.\ 1996, \apj, 465, 278 

2451: 

2452: \bibitem[Johnston(1998)]{johnston98} Johnston, K.~V.\ 1998, \apj, 

2453: 495, 297

2454: 

2455: \bibitem[Kepley et al.(2007)]{kepley07} Kepley, A.~A., et al.\ 

2456: 2007, \aj, 134, 1579 

2457: 

2458: \bibitem[Kerr \& Lynden-Bell(1986)]{kerr86} Kerr, F.~J., \& 

2459: Lynden-Bell, D.\ 1986, \mnras, 221, 1023 

2460: 

2461: \bibitem[Kinman et al.(2007)]{kinman07} Kinman, T.~D., Cacciari, 

2462: C., Bragaglia, A., Buzzoni, A., \& Spagna, A.\ 2007, \mnras, 375, 1381 

2463: 

2464: \bibitem[Koch et al.(2008)]{koch08} Koch, A., Grebel, E.~K., 

2465: Gilmore, G.~F., Wyse, R.~F.~G., Kleyna, J.~T., Harbeck, D.~R., Wilkinson, 

2466: M.~I., \& Wyn Evans, N.\ 2008, \aj, 135, 1580 

2467: 

2468: \bibitem[Korn et al.(2008)]{korn08} Korn, A.~J., Mashonkina, 

2469: L., Richard, O., Frebel, A., Aoki, W., 

2470: \& Christlieb, N.\ 2008, First Stars III, 990, 167 

2471: 

2472: \bibitem[Kraft(1979)]{kraft79} Kraft, R.~P.\ 1979, \araa, 17, 309 

2473: 

2474: \bibitem[Lai et al.(2007)]{lai07} Lai, D.~K., Johnson, J.~A., 

2475: Bolte, M., \& Lucatello, S.\ 2007, \apj, 667, 1185 

2476: 

2477: \bibitem[Lai et al.(2008)]{lai08} Lai, D.~K., Bolte, M., 

2478: Johnson, J.~A., Lucatello, S., Heger, A., 

2479: \& Woosley, S.~E.\ 2008, \apj, 681, 1524

2480: 

2481: \bibitem[Lee \& Carney(2002)]{lee02} Lee, J.-W., \& Carney, B.~W.\ 

2482: 2002, \aj, 124, 1511 

2483: 

2484: \bibitem[Lee et al.(2005)]{lee05} Lee, J.-W., Carney, B.~W., 

2485: \& Habgood, M.~J.\ 2005, \aj, 129, 251 

2486: 

2487: \bibitem[Lee et al.(2007)]{lee07} Lee, Y.~S., et al.\ 2007, 

2488: ArXiv e-prints, 710, arXiv:0710.5645 

2489: 

2490: \bibitem[Majewski(1992)]{majewski92} Majewski, S.~R.\ 1992, \apjs, 

2491: 78, 87 

2492: 

2493: \bibitem[Majewski et al.(1996)]{majewski96} Majewski, S.~R., Munn, 

2494: J.~A., \& Hawley, S.~L.\ 1996, \apjl, 459, L73 

2495: 

2496: \bibitem[Majewski et al.(2003)]{majewski03} Majewski, S.~R., 

2497: Skrutskie, M.~F., Weinberg, M.~D., \& Ostheimer, J.~C.\ 2003, 

2498: \apj, 599, 1082 

2499: 

2500: \bibitem[Martin et al.(2004)]{martin04} Martin, N.~F., Ibata, 

2501: R.~A., Bellazzini, M., Irwin, M.~J., Lewis, G.~F., 

2502: \& Dehnen, W.\ 2004, \mnras, 348, 12 

2503: 

2504: \bibitem[McWilliam et al.(1995a)]{mcwilliam95a} McWilliam, A., 

2505: Preston, G.~W., Sneden, C., \& Shectman, S.\ 1995a, \aj, 109, 2736 

2506: 

2507: \bibitem[McWilliam et al.(1995b)]{mcwilliam95b} McWilliam, A., 

2508: Preston, G.~W., Sneden, C., \& Searle, L.\ 1995b, \aj, 109, 2757 

2509: 

2510: \bibitem[McWilliam(1997)]{mcwilliam97} McWilliam, A.\ 1997, \araa, 35, 503 

2511: 

2512: \bibitem[McWilliam(1998)]{mcwilliam98} McWilliam, A.\ 1998, \aj, 115, 1640 

2513: 

2514: \bibitem[McWilliam \& Smecker-Hane(2005)]{mcwilliam05} McWilliam, A., 

2515: \& Smecker-Hane, T.~A.\ 2005, \apjl, 622, L29 

2516: 

2517: \bibitem[McCall(2004)]{mccall04} McCall, M.~L.\ 2004, \aj, 128, 

2518: 2144 

2519: 

2520: \bibitem[McCarthy \& Nemec(1997)]{mccarthy97} McCarthy, J.~K., \& 

2521: Nemec, J.~M.\ 1997, \apj, 482, 203 

2522: 

2523: \bibitem[Miceli et al.(2008)]{miceli08} Miceli, A., et al.\ 

2524: 2008, \apj, 678, 865 

2525: 

2526: \bibitem[Mihalas \& Binney(1981)]{mihalas81} Mihalas, D., \& 

2527: Binney, J.\ 1981, San Francisco, CA, W.~H.~Freeman and Co., 608 

2528: 

2529: \bibitem[Monet et al.(2003)]{monet03} Monet, D.~G., et al.\ 

2530: 2003, \aj, 125, 984 

2531: 

2532: \bibitem[Morrison et al.(2008)]{morrison08} Morrison, H.~L., et 

2533: al.\ 2008, ArXiv e-prints, 804, arXiv:0804.2448 

2534: 

2535: \bibitem[Mottini et al.(2008)]{mottini08} Mottini, M., 

2536: Wallerstein, G., \& McWilliam, A.\ 2008, \aj, 136, 614 

2537: 

2538: \bibitem[Newberg et al.(2003)]{newberg03} Newberg, H.~J., \& 

2539: Sloan Digital Sky Survey Collaboration 2003, BAAS, 35, 1385 

2540: 

2541: \bibitem[Nissen \& Schuster(1997)]{nissen97} Nissen, P.~E., \& 

2542: Schuster, W.~J.\ 1997, \aap, 326, 751 

2543: 

2544: \bibitem[Norris(1986)]{norris86} Norris, J.\ 1986, \apjs, 61, 667 

2545: 

2546: \bibitem[Norris \& Ryan(1989)]{norris89} Norris, J.~E., \& 

2547: Ryan, S.~G.\ 1989, \apj, 340, 739 

2548: 

2549: \bibitem[Norris(1994)]{norris94} Norris, J.~E.\ 1994, \apj, 431, 645 

2550: 

2551: \bibitem[Norris \& Da Costa(1995)]{norris95} Norris, J.~E., \& 

2552: Da Costa, G.~S.\ 1995, \apj, 447, 680 

2553: 

2554: \bibitem[Norris et al.(1996)]{norris96} Norris, J.~E., Freeman, 

2555: K.~C., \& Mighell, K.~J.\ 1996, \apj, 462, 241 

2556: 

2557: \bibitem[Norris et al.(1997)]{norris97} Norris, J.~E., Freeman, 

2558: K.~C., Mayor, M., \& Seitzer, P.\ 1997, \apjl, 487, L187 

2559: 

2560: \bibitem[Norris et al.(2007)]{norris07} Norris, J.~E., 

2561: Christlieb, N., Korn, A.~J., Eriksson, K., Bessell, M.~S., Beers, T.~C., 

2562: Wisotzki, L., \& Reimers, D.\ 2007, \apj, 670, 774 

2563: 

2564: \bibitem[Ochsenbein et al.(2000)]{ochsenbein00} Ochsenbein, F., 

2565: Bauer, P., \& Marcout, J.\ 2000, \aaps, 143, 23 

2566: 

2567: \bibitem[Perryman \& ESA(1997)]{perryman97} Perryman, M.~A.~C., \& ESA 1997, 

2568: ESA Special Publication, 1200  

2569: 

2570: \bibitem[Piatek et al.(2003)]{piatek03} Piatek, S., Pryor, C., 

2571: Olszewski, E.~W., Harris, H.~C., Mateo, M., Minniti, D., 

2572: \& Tinney, C.~G.\ 2003, \aj, 126, 2346 

2573: 

2574: \bibitem[Piatek et al.(2005)]{piatek05} Piatek, S., Pryor, C., 

2575: Bristow, P., Olszewski, E.~W., Harris, H.~C., Mateo, M., Minniti, D., 

2576: \& Tinney, C.~G.\ 2005, \aj, 130, 95 

2577: 

2578: \bibitem[Piatek et al.(2006)]{piatek06} Piatek, S., Pryor, C., 

2579: Bristow, P., Olszewski, E.~W., Harris, H.~C., Mateo, M., Minniti, D., 

2580: \& Tinney, C.~G.\ 2006, \aj, 131, 1445 

2581: 

2582: \bibitem[Piatek et al.(2007)]{piatek07} Piatek, S., Pryor, C., 

2583: Bristow, P., Olszewski, E.~W., Harris, H.~C., Mateo, M., Minniti, D., 

2584: \& Tinney, C.~G.\ 2007, \aj, 133, 818 

2585: 

2586: \bibitem[Preston et al.(1991)]{preston91} Preston, G.~W., 

2587: Shectman, S.~A., \& Beers, T.~C.\ 1991, \apj, 375, 121 

2588: 

2589: \bibitem[Preston et al.(1994)]{preston94} Preston, G.~W., Beers, 

2590: T.~C., \& Shectman, S.~A.\ 1994, \aj, 108, 538 

2591: 

2592: \bibitem[Preston \& Sneden(2000)]{preston00} Preston, G.~W., \& 

2593: Sneden, C.\ 2000, \aj, 120, 1014 

2594: 

2595: \bibitem[Pritzl et al.(2005)]{pritzl05} Pritzl, B.~J., Venn, 

2596: K.~A., \& Irwin, M.\ 2005, \aj, 130, 2140 

2597: 

2598: \bibitem[Prochaska et al.(2000)]{prochaska00} Prochaska, J.~X., 

2599: Naumov, S.~O., Carney, B.~W., McWilliam, A., 

2600: \& Wolfe, A.~M.\ 2000, \aj, 120, 2513 

2601: 

2602: \bibitem[Qian \& Wasserburg(2008)]{qian08} Qian, Y.-Z., \&

2603: Wasserburg, G.~J.\ \apj\ preprint doi:10.1086/`591545'

2604: 

2605: \bibitem[Reddy et al.(2003)]{reddy03} Reddy, B.~E., Tomkin, J., 

2606: Lambert, D.~L., \& Allende Prieto, C.\ 2003, \mnras, 340, 304 

2607: 

2608: \bibitem[Roederer et al.(2008)]{roederer08} Roederer, I.~U., 

2609: Lawler, J.~E., Sneden, C., Cowan, J.~J., Sobeck, J.~S., 

2610: \& Pilachowski, C.~A.\ 2008, \apj, 675, 723 

2611: 

2612: \bibitem[Ryan \& Norris(1991a)]{ryan91a} Ryan, S.~G., \& Norris, J.~E.\ 

2613: 1991a, \aj, 101, 1835 

2614: 

2615: \bibitem[Ryan \& Norris(1991b)]{ryan91b} Ryan, S.~G., \& Norris, J.~E.\ 

2616: 1991b, \aj, 101, 1865

2617: 

2618: \bibitem[Sadakane et al.(2004)]{sadakane04} Sadakane, K., Arimoto, 

2619: N., Ikuta, C., Aoki, W., Jablonka, P., 

2620: \& Tajitsu, A.\ 2004, \pasj, 56, 1041 

2621: 

2622: \bibitem[Sandage(1986)]{sandage86} Sandage, A.\ 1986, \araa, 24, 421 

2623: 

2624: \bibitem[Sandage \& Fouts(1987)]{sandage87} Sandage, A., \& 

2625: Fouts, G.\ 1987, \aj, 93, 74 

2626: 

2627: \bibitem[Sbordone et al.(2007)]{sbordone07} Sbordone, L., 

2628: Bonifacio, P., Buonanno, R., Marconi, G., Monaco, L., \& 

2629: Zaggia, S.\ 2007, \aap, 465, 815 

2630: 

2631: \bibitem[Schlegel et al.(1998)]{schlegel98} Schlegel, D.~J., 

2632: Finkbeiner, D.~P., \& Davis, M.\ 1998, \apj, 500, 525 

2633: 

2634: \bibitem[Sch\"{o}rck et al.(2008)]{schorck08} Sch\"{o}rck, T., et al.\ 

2635: 2008, \aap, submitted

2636: 

2637: \bibitem[Searle \& Zinn(1978)]{searle78} Searle, L., \& Zinn, R.\ 

2638: 1978, \apj, 225, 357 

2639: 

2640: \bibitem[Seeger et al.(1965)]{seeger65} Seeger, P.~A., Fowler, 

2641: W.~A., \& Clayton, D.~D.\ 1965, \apjs, 11, 121 

2642: 

2643: \bibitem[Shetrone et al.(2001)]{shetrone01} Shetrone, M.~D., 

2644: C{\^o}t{\'e}, P., \& Sargent, W.~L.~W.\ 2001, \apj, 548, 592 

2645: 

2646: \bibitem[Shetrone et al.(2003)]{shetrone03} Shetrone, M., Venn, 

2647: K.~A., Tolstoy, E., Primas, F., Hill, V., 

2648: \& Kaufer, A.\ 2003, \aj, 125, 684 

2649: 

2650: \bibitem[Shetrone et al.(2008)]{shetrone08} Shetrone, M.~D., Siegel, M.~H.,

2651: Cook, D.~O., \& Bosler, T.\ 2008, \aj, in press

2652: 

2653: \bibitem[Siegel et al.(2007)]{siegel07} Siegel, M.~H., et al.\ 

2654: 2007, \apjl, 667, L57 

2655: 

2656: \bibitem[Siegel et al.(2008)]{siegel08} Siegel, M.~H., Majewski, S.~R., 

2657: Sohn, S.~T., Shetrone, M.~D., Munoz, R.~R., \& Patterson, R.~J.\

2658: 2008, \apj, submitted

2659: 

2660: \bibitem[Simmerer et al.(2004)]{simmerer04} Simmerer, J., Sneden, 

2661: C., Cowan, J.~J., Collier, J., Woolf, V.~M., \& Lawler, J.~E.\ 2004, 

2662: \apj, 617, 1091 

2663: 

2664: \bibitem[Sneden et al.(1997)]{sneden97} Sneden, C., Kraft, 

2665: R.~P., Shetrone, M.~D., Smith, G.~H., Langer, G.~E., 

2666: \& Prosser, C.~F.\ 1997, \aj, 114, 1964 

2667: 

2668: \bibitem[Sneden et al.(2000)]{sneden00} Sneden, C., Johnson, J., 

2669: Kraft, R.~P., Smith, G.~H., Cowan, J.~J., 

2670: \& Bolte, M.~S.\ 2000, \apjl, 536, L85 

2671: 

2672: \bibitem[Sneden et al.(2003a)]{sneden03a} Sneden, C., et al.\ 

2673: 2003a, \apj, 591, 936 

2674: 

2675: \bibitem[Sneden et al.(2003b)]{sneden03b} Sneden, C., Preston, 

2676: G.~W., \& Cowan, J.~J.\ 2003b, \apj, 592, 504 

2677: 

2678: \bibitem[Sneden et al.(2004)]{sneden04} Sneden, C., Ivans, 

2679: I.~I., \& Fulbright, J.~P.\ 2004, Origin and Evolution of the Elements, 170 

2680: 

2681: \bibitem[Sneden et al.(2008)]{sneden08} Sneden, C., Cowan, J.~J., 

2682: \& Gallino, R.\ 2008, \araa, in press

2683: 

2684: \bibitem[Sommer-Larsen \& Zhen(1990)]{sommerlarsen90} Sommer-Larsen, J., 

2685: \& Zhen, C.\ 1990, \mnras, 242, 10 

2686: 

2687: \bibitem[Sommer-Larsen et al.(1997)]{sommerlarsen97} Sommer-Larsen, 

2688: J., Beers, T.~C., Flynn, C., Wilhelm, R., 

2689: \& Christensen, P.~R.\ 1997, \apj, 481, 775 

2690: 

2691: \bibitem[Stanford et al.(2006)]{stanford06} Stanford, L.~M., Da 

2692: Costa, G.~S., Norris, J.~E., \& Cannon, R.~D.\ 2006, \apj, 647, 1075 

2693: 

2694: \bibitem[Stephens(1999)]{stephens99} Stephens, A.\ 1999, \aj, 117, 1771 

2695: 

2696: \bibitem[Stephens \& Boesgaard(2002)]{stephens02} Stephens, A., \& 

2697: Boesgaard, A.~M.\ 2002, \aj, 123, 1647 

2698: 

2699: \bibitem[Straniero et al.(2006)]{straniero06} Straniero, O., 

2700: Gallino, R., \& Cristallo, S.\ 2006, Nuclear Physics A, 777, 311 

2701: 

2702: \bibitem[Tautvai{\v s}ien{\.e} et al.(2004)]{tautvaisiene04} 

2703: Tautvai{\v s}ien{\.e}, G., Wallerstein, G., Geisler, D., Gonzalez, G., 

2704: \& Charbonnel, C.\ 2004, \aj, 127, 373 

2705: 

2706: \bibitem[Tolstoy et al.(2003)]{tolstoy03} Tolstoy, E., Venn, 

2707: K.~A., Shetrone, M., Primas, F., Hill, V., Kaufer, A., 

2708: \& Szeifert, T.\ 2003, \aj, 125, 707 

2709: 

2710: \bibitem[Travaglio et al.(2004)]{travaglio04} Travaglio, C., 

2711: Gallino, R., Arnone, E., Cowan, J., Jordan, F., 

2712: \& Sneden, C.\ 2004, \apj, 601, 864 

2713: 

2714: \bibitem[Tumlinson(2006)]{tumlinson06} Tumlinson, J.\ 2006, \apj, 641, 1 

2715: 

2716: \bibitem[Unavane et al.(1996)]{unavane96} Unavane, M., Wyse, 

2717: R.~F.~G., \& Gilmore, G.\ 1996, \mnras, 278, 727 

2718: 

2719: \bibitem[Urban et al.(2004)]{urban04} Urban, S.~E., Zacharias, 

2720: N., Wycoff, O.~G.~L.~U.~S.~N., 

2721: \& Washington, 2.~D.~C.\ 2004, VizieR Online Data Catalog, 1294, 0 

2722: 

2723: \bibitem[van Leeuwen(2007)]{vanleeuwen07} van Leeuwen, F.\ 2007, \aap, 

2724: 474, 653 

2725: 

2726: \bibitem[Venn et al.(2004)]{venn04} Venn, K.~A., Irwin, M., 

2727: Shetrone, M.~D., Tout, C.~A., Hill, V., 

2728: \& Tolstoy, E.\ 2004, \aj, 128, 1177 

2729: 

2730: \bibitem[Wagoner et al.(1967)]{wagoner67} Wagoner, R.~V., Fowler, 

2731: W.~A., \& Hoyle, F.\ 1967, \apj, 148, 3 

2732: 

2733: \bibitem[Wallerstein(1962)]{wallerstein62} Wallerstein, G.\ 1962, 

2734: \apjs, 6, 407 

2735: 

2736: \bibitem[Wanajo(2006)]{wanajo06a} Wanajo, S.\ 2006, \apj, 647, 1323 

2737: 

2738: \bibitem[Wanajo \& Ishimaru(2006)]{wanajo06b} Wanajo, S., \& 

2739: Ishimaru, Y.\ 2006, Nuclear Physics A, 777, 676 

2740: 

2741: \bibitem[Wanajo(2007)]{wanajo07} Wanajo, S.\ 2007, \apjl, 666, L77 

2742: 

2743: \bibitem[Wasserburg et al.(1996)]{wasserburg96} Wasserburg, G.~J., 

2744: Busso, M., \& Gallino, R.\ 1996, \apjl, 466, L109 

2745: 

2746: \bibitem[Wheeler et al.(1989)]{wheeler89} Wheeler, J.~C., Sneden, C., 

2747: \& Truran, J.~W., Jr.\ 1989, \araa, 27, 279 

2748: 

2749: \bibitem[Wisotzki et al.(2000)]{wisotzki00} Wisotzki, L., Christlieb, N., 

2750: Bade, N., Beckmann, V., K{\"o}hler, T., Vanelle, C., \& Reimers, D.\ 

2751: 2000, \aap, 358, 77 

2752: 

2753: \bibitem[Wyse \& Gilmore(1988)]{wyse88} Wyse, R.~F.~G., \& Gilmore, G.\ 

2754: 1988, \aj, 95, 1404 

2755: 

2756: \bibitem[Yoshii(1981)]{yoshii81} Yoshii, Y.\ 1981, \aap, 97, 280 

2757: 

2758: \bibitem[Yong et al.(2008)]{yong08} Yong, D., Karakas, A.~I., Lambert, D.~L.,

2759: Chieffi, A., \& Limongi, M.\ 2008, \apj\ preprint doi:10.1086/592600

2760: 

2761: \bibitem[Zacharias et al.(2004a)]{zacharias04a} Zacharias, N., Urban, 

2762: S.~E., Zacharias, M.~I., Wycoff, G.~L., Hall, D.~M., Monet, D.~G., \& 

2763: Rafferty, T.~J.\ 2004a, \aj, 127, 3043 

2764: 

2765: \bibitem[Zacharias et al.(2004b)]{zacharias04b} Zacharias, N., Monet, 

2766: D.~G., Levine, S.~E., Urban, S.~E., Gaume, R., \& Wycoff, G.~L.\ 2004b, 

2767: BAAS, 36, 1418 

2768: 

2769: 

2770: \end{thebibliography}

2771: 

2772: \clearpage

2773: \input{tab1-stub}   

2774: 

2775: \clearpage

2776: \input{tab2}

2777: 

2778: \clearpage

2779: \input{tab3}

2780: 

2781: \clearpage

2782: \input{tab4}

2783: 

2784: 

2785: \end{document}

2786: