1: \documentclass[11pt,preprint]{aastex}

2: \usepackage{natbib}

3: \usepackage{iondefs}

4: 

5: 

6: \newcommand{\myemail}{blawton@nmsu.edu}

7: 

8: \begin{document}

9: 

10: \title{Limits on Reddening and Gas-to-Dust Ratios for Seven \\ Intermediate Redshift Damped {\Lya} Absorbers \\ from Diffuse Interstellar Bands}

11: 

12: \author{\sc 

13: Brandon Lawton\altaffilmark{1}, 

14: Christopher W. Churchill\altaffilmark{1}, 

15: Brian A. York\altaffilmark{2}, 

16: Sara L. Ellison\altaffilmark{2}, \\ 

17: Theodore P. Snow\altaffilmark{3}, 

18: Rachel A. Johnson\altaffilmark{4},

19: Sean G. Ryan\altaffilmark{5}, and

20: Chris R. Benn\altaffilmark{6}

21: }

22: 

23: \altaffiltext{1}{Department of Astronomy, MSC 4500, New Mexico State

24: University, P.O. Box 30001, Las Cruces, NM 88003; blawton@nmsu.edu,

25: cwc@nmsu.edu}

26: 

27: \altaffiltext{2}{Department of Physics \& Astronomy, University of

28: Victoria, 3800 Finnerty Rd., Victoria, V8W 1A1, British Columbia,

29: Canada; briany@uvic.ca, sarae@uvic.ca}

30: 

31: \altaffiltext{2}{Center for Astrophysics and Space Astronomy,

32: University of Colorado at Boulder, 389 UCB, Boulder, CO 80309;

33: Theodore.Snow@colorado.edu}

34: 

35: \altaffiltext{4}{Oxford Astrophysics, Denys Wilkinson Building, Keble

36: Road, Oxford OX1 3RH, UK; raj@astro.ox.ac.uk}

37: 

38: \altaffiltext{5}{Centre for Astrophysics Research, University of

39: Hertfordshire, College Lane, Hatfield AL10 9AB, UK;

40: s.g.ryan@herts.ac.uk}

41: 

42: \altaffiltext{6}{Isaac Newton Group, Apartado 321, E-38700 Santa Cruz

43: de La Palma, Spain; crb@ing.iac.es}

44: 

45: 

46: 

47: \begin{abstract}

48: 

49: We present equivalent width measurements and limits of six diffuse

50: interstellar bands (DIBs, $\lambda$4428, $\lambda$5705, $\lambda$5780,

51: $\lambda$5797, $\lambda$6284, and $\lambda$6613) in seven damped

52: {\Lya} absorbers (DLAs) over the redshift range $0.091 \leq z \leq

53: 0.524$, sampling $20.3 \leq \log N({\HI}) \leq 21.7$. DIBs were

54: detected in only one of the seven DLAs, that which has the highest

55: reddening and metallicity.  Based upon the Galactic DIB--$N({\HI})$

56: relation, the $\lambda$6284 DIB equivalent width upper limits in four

57: of the seven DLAs are a factor of 4-10 times below the $\lambda$6284

58: DIB equivalent widths observed in the Milky Way, but are not

59: inconsistent with those present in the Magellanic Clouds. Assuming the

60: Galactic DIB--$E(B-V)$ relation, we determine reddening upper limits

61: for the DLAs in our sample.  Based upon the $E(B-V)$ limits, the

62: gas-to-dust ratios, $N({\HI})/E(B-V)$, of the four aforementioned DLAs

63: are at least $\sim 5$ times higher than that of the Milky Way ISM.

64: The ratios of two other DLAs are at least a factor of a few times

65: higher.  The best constraints on reddening derive from the upper

66: limits for the $\lambda$5780 and $\lambda$6284 DIBs, which yield

67: $E(B-V) \leq 0.08$ for four of the seven DLAs.  Our results suggest

68: that, in DLAs, quantities related to dust, such as reddening and

69: metallicity, appear to have a greater impact on DIB strengths than

70: does {\HI} gas abundance; the organic molecules likely responsible for

71: DIBs in DLA selected sightlines are underabundant relative to

72: sightlines in the Galaxy of similarly high $N({\HI})$.  With regards

73: to the study of astrobiology, this could have implications for the

74: abundance of organic molecules in redshifted galaxies.  However, since

75: DLAs are observed to have low reddening, selection bias likely plays a

76: role in the apparent underabundance of DIBs in DLAs.

77: 

78: \end{abstract}

79: 

80: \keywords{

81: dust, extinction --

82: galaxies: ISM and abundances --

83: quasars: absorption lines and individual (AO 0235+164, Q0738+313, B2

84: 0827+243, PKS 0952+179, PKS 1127--145, Q1229--020) --

85: techniques: spectroscopic

86: }

87: 

88: \section{Introduction}

89: 

90: Since their discovery by \citet{hege22}, several hundred diffuse

91: interstellar bands (DIBs) have been studied

92: \citep{gala00,jenn94,tuai00,wese00,hobbs07}, and yet no positive

93: identifications of the carriers have been made.  The DIBs span the

94: visible spectrum between 4000 and 13,000~{\AA}.  Despite no positive

95: identifications, several likely organic molecular candidates have

96: emerged as the sources of the DIBs, including polycyclic aromatic

97: hydrocarbons (PAHs), fullerenes, long carbon chains, and polycyclic

98: aromatic nitrogen heterocycles (PANHs) \citep[e.g.,][]{herb95, snow01,

99: cox06a, hudg05}.  The organic-molecular origin of the DIBs may give

100: them an importance to astrobiology; PAHs are now considered an

101: important early constituent to the inventory of organic compounds on

102: Earth \citep{bada02}.  Via their infrared emitting vibrational bands,

103: PAHs have been observed in high redshift dusty ultra-luminous infrared

104: galaxies \citep[ULIRGs, e.g.,][]{yan05}.  Searching for DIBs using the

105: technique of quasar absorption lines provides a different approach for

106: charting the presence of possible organics to high redshift.  As such,

107: observing DIBs in high redshift galaxies may offer an independent

108: method for constraining the environmental conditions in early-epoch

109: galaxies governing the abundances of organic molecules, determining

110: the cosmic epoch at which organic molecules first formed, and

111: ultimately charting their evolution with redshift.

112: 

113: Aside from the hundreds of detections within the Galaxy

114: \citep[e.g.][]{gala00,jenn94,tuai00,wese00,hobbs07}, DIBs have been

115: detected in the Magellanic Clouds \citep{welt06,cox06b,cox07}, M31

116: \citep{cord08}, seven starburst galaxies \citep{heck00}, active galaxy

117: Centaurus A via supernova 1986A \citep{rich87}, spiral galaxy NGC 1448

118: via supernovae 2001el and 2003hn \citep{soll05}, one damped {\Lya}

119: absorber (DLA) at $z=0.524$ toward the quasar (QSO) AO~0235+164

120: \citep{junk04, lawt06}, and one $z=0.157$ {\CaII} selected absorber

121: toward QSO J0013--0024 \citep{elli07}.

122: 

123: There are several environmental factors, such as {\HI} column density

124: \citep{herb95,welt06}, reddening \citep{welt06}, and metallicity and

125: ionizing radiation \citep{cox07}, that are related to DIB strengths.

126: In the Galaxy, DIB absorption strengths correlate strongly with

127: $N({\HI})$ \citep{herb95,welt06}.  However, in the Magellanic Clouds,

128: DIBs are weaker by factors of 7-9 (LMC) and $\sim20$ (SMC) compared to

129: those observed in the Galaxy with similar $N({\HI})$ \citep{welt06}.

130: It is not known whether other galaxies in the Local Group and beyond

131: obey the Galactic DIB--$N({\HI})$ relation, or, like the Magellanic

132: Clouds, show departures from this relation.  Galaxies with high

133: $N({\HI})$ observed in absorption (i.e., DLAs) that reside at low to

134: intermediate redshifts (where the prominent DIBs fall in the optical

135: region) provide excellent astrophysical laboratories with which to

136: investigate this issue.

137: 

138: In this paper we search for $\lambda$4428, $\lambda$5780,

139: $\lambda$5797, $\lambda$6284, and $\lambda$6613 DIB absorption in

140: seven low to intermediate redshift DLAs.  In \S~\ref{sec:sample}, we

141: give a brief summary of each intervening DLA in our sample.  In

142: \S~\ref{sec:obs}, we discuss the spectroscopic observations and data

143: reduction of the background QSOs.  In \S~\ref{sec:analysis}, we

144: explain the procedure of our analysis, and the resulting spectra.  In

145: \S~\ref{sec:results}, we present our results and compare our data to

146: the Galactic DIB--$N({\HI})$ relation, deduce upper limits for the

147: reddening, $E(B-V)$, determine lower limits on the gas-to-dust ratios,

148: and discuss the role of metallicity for our sample of DLAs.  We

149: conclude in \S~\ref{sec:conclusions}.

150: 

151: \section{DLA Sample}

152: \label{sec:sample}

153: 

154: To potentially maximize our chances of detecting DIBs, and to test the

155: Galactic DIB--$N({\HI})$ relation in redshifted galaxies, we selected

156: DLAs toward background QSOs having the highest $N({\HI})$ in

157: absorption.  We limited the redshift range such that the strongest

158: DIBs would fall in the optical region.  Thus, we selected the highest

159: $N({\HI})$ DLA galaxies in the redshift range $0.09<z<0.52$.  DLAs, by

160: definition, have a large neutral gas column density ($N({\HI})\geq 2

161: \times 10^{20}$ cm$^{-2}$).  However, they are observed to have low

162: reddening ($E(B-V)<0.04$) \citep{elli05} and low metallicity

163: (typically $Z\sim0.1 Z_{\odot}$).  The low metallicity and reddening

164: of DLAs suggest that the gaseous environments selected by DLA

165: absorption differ from the Galactic ISM in which DIBs are observed.

166: As we discuss in \S~\ref{sec:reddening}, the low reddening of DLAs is

167: not merely a bias arising from the necessity of low extinction if the

168: background quasar is to be detected at all, since DLAs in

169: radio-selected quasars have similarly low reddenings.

170: 

171: Below, we describe the basic properties of each of the seven DLAs

172: comprising our sample.  Table~\ref{tab:DLAs} summarizes the main

173: parameters, including abundance information.  The columns, from left

174: to right, list the QSO, absorption redshift of the DLA, column density

175: of neutral hydrogen, the zinc abundance, the iron abundance, and the

176: associated references.  The {\Lya} line is used to measure the DLA

177: redshift.  All impact parameters have been converted to a $\Lambda$CDM

178: cosmology ($H_0=70$~km~s$^{-1}$~Mpc$^{-1}$, $\Omega_{m}=0.3$,

179: $\Omega_{\Lambda}=0.7$). 

180: 

181: The $z_{abs}=0.524$ DLA ({\#1}) toward AO~0235+164 is unique in our

182: sample for many reasons. It is the only DLA known to have DIB

183: absorption\footnote{The DIB-bearing {\CaII} absorber reported by

184: \citet{elli07} is likely to be a DLA.  However, this cannot be

185: confirmed without a spectrum of the {\Lya} absorption.}, the $\lambda

186: 4428$ DIB \citep{junk04} and the $\lambda 5705$ and $\lambda 5780$

187: DIBs \citep{lawt06}.  The $\lambda 5797$, $\lambda 6284$, and $\lambda

188: 6613$ DIBs have limits that are not as sensitive due to sky lines.

189: The associated optical galaxy with the smallest impact parameter,

190: $6.7~h_{70}^{-1}$~kpc, is a late-type spiral and is assumed to be the

191: absorber \citep{raot03,burb96,yann89}.  However, \citet{chen03} argue

192: that many of the galaxies in the optical field have the same redshift

193: and may collectively be responsible for the DLA absorption.  The

194: adopted neutral hydrogen column density is from \citet{junk04} and is

195: consistent with $N(\HI)=(4.5\pm0.4)\times 10^{21}$~{\cmsq} from

196: \citet{turn03}.  The measured metallicities \citep{turn03,junk04}

197: assume solar abundance ratios and solar metallicity for the foreground

198: Galactic absorption.  \citet{junk04} state that the difference between

199: their metallicities and those of \citet{turn03} are likely due to

200: different realizations of the noise in the X-ray data and/or the

201: variability \citep{riek76} of this QSO.  Despite the differences in

202: these measurements, this DLA has the highest metallicity in our

203: sample.  The DLA also has significant reddening with $E(B-V)=0.23$,

204: $R_{v}=2.5$, and a strong graphitic dust feature at 2175~{\AA}.

205: 

206: The $z=0.091$ DLA ({\#2}) toward Q0738+313 (OI363) is probably a low

207: surface brightness galaxy with an impact parameter of

208: $<3.5~h_{70}^{-1}$~kpc \citep{turn01}.  This DLA is one of two found

209: along the QSO sightline of Q0738+313 and was first reported by

210: \citet{raot98}.

211: 

212: The other DLA, $z_{abs}=0.221$ ({\#3}), toward the QSO Q0738+313 is

213: probably a dwarf spiral with an impact parameter of

214: $20.5~h_{70}^{-1}$~kpc \citep{turn01}.

215: 

216: The $z_{abs}=0.518$ DLA ({\#4}) toward B2 0827+243 is likely a

217: disturbed spiral galaxy with extended gas that produces the observed

218: hydrogen absorption at an impact parameter of $38.2~h_{70}^{-1}$~kpc

219: \citep{raot03}.  \citet{khare04} note that it appears the DLA requires

220: a significant radiation field, similar to the radiation within the

221: dense clouds in our Galaxy, to create a low $N(\FeI)/N(\FeII)$ upper

222: limit of 10$^{-3}$.

223: 

224: The absorbing galaxy giving rise to the $z_{abs}=0.239$ DLA ({\#5})

225: toward PKS~0952+179 has not been confirmed.  However, two candidates

226: lie within $<4.6~h_{70}^{-1}$~kpc \citep{raot03}.  The authors note

227: that these two galaxies appear to be nearly edge-on and are classified

228: as dwarf low surface brightness galaxies. The DLA appears to have

229: multiple {\Lya} components around the central line at $z=0.239$

230: \citep{raot00}.  The adopted hydrogen column density is measured from

231: the central {\Lya} feature.  However, the velocity structure suggests

232: a possible clustering.

233: 

234: The $z_{abs}=0.313$ DLA ({\#6}) toward PKS~1127--145 is possibly the

235: remnant of a dwarf low surface brightness galaxy tidally disturbed by

236: more massive spiral galaxies in the same field \citep{raot03}.  The

237: absorber appears to overlap the QSO point spread function, thus,

238: \citet{raot03} find an upper limit for the impact parameter of

239: $6.9~h_{70}^{-1}$~kpc.  \citet{turn03} discuss the difficulty of

240: determining a metallicity for this system due to uncertainties in

241: their X-ray spectrum.  We have estimated a lower limit of the iron

242: abundance from a VLT/UVES spectrum kindly donated for this work by

243: Dr.\ M. Murphy.  The system has a complicated velocity structure.

244: However, fixing the Doppler parameter, redshift, and number of Voigt

245: profiles yields a lower limit on the column density of $\log

246: N(\FeII)>15.7$~atoms~cm$^{-2}$.  Repeating this analysis using the

247: apparent optical depth \citep{sava91} gives a slightly more

248: conservative limit of $\log N(\FeII)>15.2$~atoms~cm$^{-2}$, which we

249: adopt for the column density of iron.  Assuming the \citet{lodd03}

250: solar abundances, we deduce a lower limit of $\hbox{[Fe/H]}>-2$.

251: 

252: The absorbing galaxy responsible for the $z_{abs}=0.395$ DLA ({\#7})

253: along the Q1229--029 sight--line is a low surface brightness galaxy

254: with an impact parameter of $8.2~h_{70}^{-1}$~kpc \citep{lebr97,

255: stei94}.

256: 

257: \section{Observations and Data Reduction}

258: \label{sec:obs}

259: 

260: Observations of the seven DLAs were obtained with seven facilities

261: toward six QSO sightlines between July 2002 and October 2005.  The S/N

262: of the QSO spectra range from 5--150~pixel$^{-1}$.  All quoted S/N

263: measurements are calculated in the regions of the expected locations

264: of the redshifted $\lambda4428$, $\lambda5780$, $\lambda5797$,

265: $\lambda6284$, and $\lambda6613$ DIBs.  The journal of observations is

266: presented in Table~\ref{tab:obs}.  Cataloged are the QSO, facility and

267: instrument used in the observation, the grating/grism, the slit width,

268: the UT date of the observation, the total exposure time in seconds,

269: and the wavelength coverage of each spectrum in angstroms.

270: 

271: \subsection{Observations}

272: 

273: QSO spectra covering the $z=0.524$, $z=0.239$, and $z=0.395$ DLAs

274: along the sightlines AO~0235+164, PKS~0952+179, and Q1229--020 were

275: obtained with the FORS2 spectrograph on the Very Large Telescope

276: (VLT).  All resolutions are obtained by measuring unresolved sky

277: emission lines.  Observations of A0~0235+164 and Q1229--020 use the

278: same grating and the resolutions agree within reasonable

279: uncertainties.

280: 

281: A QSO spectrum covering the $z=0.091$ and $z=0.221$ DLAs along the

282: Q0738+313 sightline was acquired with the DIS spectrograph on the

283: Apache Point Observatory (APO) 3.5~m telescope.  DIS is configured

284: with a dichroic that splits the light to a blue chip and a red chip at

285: ${\sim}5500$~{\AA}.  The resolution is measured directly from

286: unresolved sky emission lines.

287: 

288: A Keck/HIRES spectrum of B2 0827+243 covering the DLA at $z=0.518$ was

289: kindly provided by Dr.\ W. Sargent.  The resolution is measured by

290: unresolved atmospheric absorption lines (we did not have access to sky

291: emission line data for this object).

292: 

293: QSO spectra covering the $z=0.239$ and $z=0.395$ DLAs along the

294: sightlines toward PKS~0952+179 and Q1229--020 were obtained using the

295: ISIS spectrograph on the 4~m William Herschel Telescope (WHT).  The

296: resolutions were measured directly from unresolved sky emission lines;

297: the uncertainties are relatively large because of the low S/N in these

298: data.

299: 

300: A UVES/VLT spectrum covering the $z=0.313$ DLA toward PKS~1127--145

301: was kindly provided by Dr.\ M. Murphy.  The resolution is taken from

302: \citet{dekk00}.  We did not have access to sky data for this spectrum

303: nor were there any unresolved atmospheric absorption lines, so we

304: could not estimate the resolution directly.

305: 

306: A QSO spectrum covering the $z=0.313$ DLA along the sightline toward

307: PKS~1127--145 was obtained with the GMOS spectrograph on the 8.1~m

308: Gemini South telescope.  The resolution is taken from the online

309: Gemini Science Operations GMOS Instrument Manual.  We had limited sky

310: data, so we did not estimate the resolution directly.  However, due to

311: large equivalent width limits of the GMOS data, the effect of

312: uncertainties in resolution should be minimal and not affect results

313: in this paper (see Table~\ref{tab:EWs}).

314: 

315: \subsection{Data Reduction}

316: 

317: With the exception of the VLT/UVES and Gemini/GMOS spectra of

318: PKS~1127--145, the data were reduced using standard IRAF\footnote{IRAF

319: is distributed by the National Optical Astronomy Observatories, which

320: are operated by the Association of Universities for Research in

321: Astronomy, Inc., under cooperative agreement with the National Science

322: Foundation.} packages.  The IRAF reduction process involved bias

323: subtraction, flat-fielding, spectrum extraction, and wavelength

324: calibration using standard lamps.  Once wavelength calibrated, each

325: spectrum was continuum fit manually using \textit{sfit} to achieve the

326: lowest residuals in the regions of the DIBs where no detections are

327: observed.  Near telluric lines or problematic sky subtraction, the

328: continuum was fit using regions bracketing these features.  The flux,

329: uncertainty, sky (when acquired), and continuum spectra are normalized

330: and then the individual spectra are optimally combined (using an

331: algorithm of our own design).  For the Gemini/GMOS-S spectrum, data

332: reduction was performed using the IRAF Gemini tools in the

333: \textit{gmos} package.  The IRAF task, \textit{telluric}, was also

334: used on the Gemini/GMOS-S spectrum \citep{lawt06}.  The UVES spectrum

335: was reduced using the standard ESO pipeline and a custom code called

336: the UVES Post--Pipeline Echelle Reduction \citep[{\sc uves

337: popler},][]{popler}.

338: 

339: \section{Data Analysis and Spectra}

340: 

341: \label{sec:analysis}

342: 

343: \subsection{Analysis of DIBs}

344: 

345: A modified method of \citet{schn93} was developed and used to search

346: for the presence of $\lambda$4428, $\lambda$5780, $\lambda$5797,

347: $\lambda$6284, and $\lambda$6613 DIB absorption.  The Schneider et

348: al.\ technique is optimized for objectively locating unresolved

349: features in spectra.  However, some DIBs are resolved in our spectra.

350: Thus, we modified the \citet{schn93} method to be optimized for both

351: unresolved and resolved features by combining the DIB's intrinsic full

352: width at half maximum (\hbox{\small FWHM}) with the instrumental

353: spread function (ISF).  The procedure is outlined in Appendix A.1.

354: The calculation used by \citet{schn93} (given in

355: Eq.~\ref{EQ:dpsewlim}) transforms to Eq.~\ref{EQ:EWlimit}.

356: 

357: In addition, we replaced the normalized flux error of \citet{schn93}

358: by the residuals of the data in pixels where the normalized flux

359: deviates significantly from the continuum (see

360: Eqs.~\ref{EQ:flux_error}, \ref{EQ:residual}, and

361: \ref{EQ:norm_flux_error}).  This results in a more conservative

362: detection threshold (equivalent width limit) in the cases of

363: problematic sky subtraction or continuum fits.  As an example of this

364: method, Fig.~\ref{p:kspace}$a$ displays the relative flux of Q0738+313

365: in the region of the expected positions of the $\lambda$5780 and

366: $\lambda$5797 DIBs for the $z=0.091$ DLA.  Fig.~\ref{p:kspace}$b$

367: contains the sky data.  Fig.~\ref{p:kspace}$c$ contains the associated

368: 3~$\sigma$ rest-frame detection thresholds for a Gaussian profile with

369: the expected \hbox{\small FWHM} of the redshifted $\lambda$5780 DIB.

370: Both the $\lambda$5780 and $\lambda$5797 DIBs are unresolved.

371: However, if they had been resolved they would have yielded different

372: equivalent width limits because the limits depends upon the redshifted

373: \hbox{\small FWHM} for resolved DIBs.  The problematic sky subtraction

374: at $6300$~{\AA} and the problematic continuum fit at $6314$~{\AA}

375: result in conservative limits due to large residuals (see

376: Eqs.~\ref{EQ:flux_error}, \ref{EQ:norm_flux_error}, and

377: \ref{EQ:EWlimit}).  The region around the $\lambda$5797 DIB is

378: ``clean'' in that no residuals are used in the equivalent width limit

379: calculations.  The equivalent width limits of the DIBs in the

380: AO~0235+164 DLA measured by \citet{lawt06} are less stringent but not

381: inconsistent with those measured by this method.

382: 

383: Possible detections were visually inspected to determine if they were

384: a DIB or perhaps another feature or sky line residual.  Checking each

385: candidate feature is essential because of a small uncertainty in the

386: rest wavelength of each DIB as well as a small uncertainty in the

387: redshift of DIB absorbing gas relative to the {\Lya} determined

388: redshift of the DLA.  For most DIBs, we were able to measure only

389: equivalent width limits.

390: 

391: To quantify confidence levels in the measured detection thresholds

392: determined with our modified method, we have estimated the $1~\sigma$

393: uncertainties in the measured equivalent width limits.  There is some

394: uncertainty in the resolution (ISF) of each spectrum.  We estimate the

395: uncertainty in the resolution as the standard deviation of the

396: $\lambda/\hbox{\small FWHM}$ ratios of unresolved sky absorption lines

397: or emission lines, where the $\lambda$ are the line centers.  If sky

398: data are not available, we excluded the uncertainty in resolution.  In

399: addition, there is uncertainty in the accuracy of the continuum fits,

400: which we estimated using the technique of \citet{semb91}.  Of

401: importance, is the fact that DIBs have measured uncertainties in their

402: rest-frame wavelengths and \hbox{\small FWHMs} \citep{jenn94}.  The

403: uncertainty in the equivalent width limit due to uncertainty in the

404: wavelength is the standard deviation of the individual equivalent

405: width limits computed over the range of the redshifted DIB wavelength

406: uncertainty.  The full explanation of ``error'' propagation to obtain

407: uncertainties in the equivalent width limits is given in Appendix A.2

408: (see Eq.~\ref{EQ:var}).

409: 

410: Contained within Fig.~\ref{p:kspace} is an example of our equivalent

411: width limit analysis including information on the effects of

412: uncertainties in the wavelength of the band center.  If a redshifted

413: DIB wavelength is near a problematic region then an uncertainty in

414: rest wavelength can introduce a large uncertainty in a measured

415: equivalent width limit.  This is a more substantial issue for the

416: $\lambda$4428 DIB, which has the largest uncertainty in its rest-frame

417: wavelength ($\sim1.4$~\AA).  For the majority of the equivalent width

418: limits, the largest uncertainty is due to the continuum fit.

419: 

420: Only three DIBs have been detected in DLAs, the $\lambda$4428 DIB

421: reported in \citet{junk04}, and the $\lambda$5705 and $\lambda$5780

422: DIBs reported in a companion paper of this work by \citet{lawt06}.

423: All three arise in the $z=0.524$ DLA toward AO~0235+164.  The

424: equivalent widths were calculated by summing the individual equivalent

425: widths of each pixel.  Both of the detections from \citet{lawt06} were

426: at least one resolution element away from sky lines.  The undetected

427: $\lambda$5797 and $\lambda$6284 DIBs toward the $z=0.524$ AO~0235+164

428: DLA were located within strong sky lines.  \citet{lawt06} estimated

429: the equivalent width limits for these DIBs using synthetic absorption

430: features with varying \hbox{\small FWHMs}.  The undetected

431: $\lambda$6613 DIB is near a sky line; \citet{lawt06} estimated its

432: equivalent width limit directly from the signal-to-noise, dispersion

433: of the chip, and number of pixels.

434: 

435: \subsection{Measurements and Spectra}

436: 

437: The measured equivalent widths, equivalent width limits (3~$\sigma$),

438: and their 1~$\sigma$ uncertainties are listed in Table~\ref{tab:EWs}.

439: Tabulated are the DLA (by number), the corresponding QSO, the

440: absorption redshift of the DLA, the facility and instrument, and the

441: rest-frame equivalent widths or limits, with uncertainties (m{\AA}) of

442: each of the DIBs if they were observable.  Spectra from which the most

443: stringent limits were obtained are shown in

444: Figs.~\ref{s:0235}--\ref{s:1229_VLT}\footnote{The VLT/UVES spectrum is

445: not shown because only the $\lambda$4428 DIB is covered.  The UVES

446: spectrum has a high resolution that makes identifying the broad

447: $\lambda$4428 DIB very difficult.}.  Each panel ($a$)--($e$) displays

448: the region around a redshifted DIB.  The upper sub-panels display the

449: normalized flux (histogram) with the expected positions of the DIBs

450: (marked by ticks) based upon the {\Lya} redshift.  With the exceptions

451: of Figs.~\ref{s:0827} and \ref{s:1127_Gem}, the center sub-panels

452: display the uncertainty spectra of the associated normalized fluxes,

453: and the lower sub-panels display the sky counts normalized by the

454: continuum.  Sky data were unavailable for B2~0827+243 at $z=0.518$ and

455: PKS~1127--145 at $z=0.313$.

456: 

457: The smooth thin curves through the data are the expected observed DIB

458: absorption profiles based on the measured $N({\HI})$ for the DLAs,

459: where the band centers and \hbox{\small FWHM} of each DIB are taken

460: from \citet{jenn94}.  These models are not computed for the

461: $\lambda$4428 and $\lambda$6613 DIBs since these DIBs have no

462: published $N({\HI})$ relationships.  In the following sections we will

463: discuss limits on reddening and gas-to-dust ratios.  Thus, we also

464: illustrate (thick curves) the expected observed DIB absorption

465: profiles assuming an $E(B-V)=0.04$, the upper limit for high redshift

466: DLAs assuming SMC-like extinction \citep{elli05}.  For the AO~0235+164

467: DLA, we adopted the measured $E(B-V)=0.23$ \citep{junk04}).  The

468: computations of the reddening models are discussed in

469: \S~\ref{sec:reddening}.

470: 

471: The equivalent widths of the DIBs estimated from the Galactic

472: DIB--$N({\HI})$ and Galactic DIB--$E(B-V)$ models

473: \citep[see][]{welt06} are presented in Table~\ref{tab:Models}.

474: Columns 1--2 list the DLA number and the QSO with associated DLA

475: redshift.  Columns 4--8 provide the observed equivalent widths and

476: equivalent width limits for the $\lambda$4428, $\lambda$5780,

477: $\lambda$5797, $\lambda$6284, and $\lambda$6613 DIBs.  Also listed are

478: the predicted equivalent widths of the DIBs, where EW[$N$(HI)] denotes

479: the Galactic $N({\HI})$ scaling and EW[$E(B-V)$] denotes the reddening

480: scaling in {m\AA}.  $E(B-V)_{\rm lim}$ is the reddening upper limit for

481: each DLA based upon the observed DIB equivalent width limits and the

482: reddening relation (see \S~\ref{sec:reddening}). 

483: 

484: \section{Results and Discussion}

485: 

486: \label{sec:results}

487: 

488: In this section, we examine the DIB strengths in DLAs compared to the

489: Galactic DIB--$N({\HI})$ relation and the Galactic $E(B-V)$

490: relation. The former provides information on the gas content and the

491: latter provides information on the reddening (an indirect indicator of

492: dust content).  Our observations allow us to estimate lower limits on

493: the gas-to-dust ratios, $N({\HI})/E(B-V)$, of the DLAs in our sample.

494: 

495: \subsection{Gas Content}

496: \label{sec:gas}

497: The widely observed Galactic DIB--$N({\HI})$ relation describes the

498: correlation of the equivalent width of the $\lambda$5780,

499: $\lambda$5797, and $\lambda$6284 DIBs with the column density of

500: neutral hydrogen along the same line of sight \citep{herb95,welt06}.

501: \citet{welt06} extend the Galactic work by including Magellanic Cloud

502: sightlines.

503: 

504: Plotted in Fig.~\ref{p:NHI_known} are the DIB--$N({\HI})$ relations

505: for the $\lambda$5780 (panel $a$), $\lambda$5797 (panel $b$), and

506: $\lambda$6284 DIBs (panel $c$), where $\log N{(\HI)}$ [cm$^{-2}$] is

507: plotted against the logarithm of the DIB equivalent widths [m\AA].

508: Also plotted are the (1~$\sigma$) error bars for measured values or

509: downward arrows representing upper limits.  The vertical error bars

510: for the $\lambda$5780 DIB detection in the AO~0235+164 DLA are smaller

511: than the point size.  The solid lines represent the weighted best-fits

512: to the Galactic data from \citet{welt06}.  The equivalent widths

513: predicted at a given $N({\HI})$, the EW[$N({\HI})$] listed in

514: Table~\ref{tab:Models}, are computed using this best-fit line and the

515: $N({\HI})$ of each DLA (Figs.~\ref{s:0235}--\ref{s:1229_VLT} contain

516: these models superimposed on the data).  The dotted lines in

517: Fig.~\ref{p:NHI_known} roughly enclose the regions containing the

518: Galactic data.  The dashed lines roughly enclose the regions

519: containing the LMC data, and the dot-dashed lines roughly enclose the

520: regions containing the SMC data.

521: 

522: The DIBs in the DLAs whose equivalent width limits in

523: Fig.~\ref{p:NHI_known} lie below the Galactic best-fit lines are

524: robust enough that we should have detected them if they followed the

525: same dependence on $N({\HI})$ as Galactic sightlines. In several DLAs,

526: the limits are inconsistent with the expectations from the Galactic

527: sightlines; the DIB strengths are weaker than expected.  The DIB

528: limits are consistent with the strengths of DIBs in the LMC or SMC.

529: However, higher signal-to-noise data are required to determine whether

530: the DIB strengths are actually consistent with or are even weaker than

531: those observed in the LMC and SMC.  The $\lambda$6284 DIB, panel ($c$)

532: of Fig.~\ref{p:NHI_known}, provides the most stringent evidence that

533: the DIB strengths in four DLAs are not consistent with DIB strengths

534: in the Milky Way.  The four DLAs toward AO~0235+164 ($z=0.524$),

535: Q0738+313 ($z=0.091$), PKS~0952+179 ($z=0.239$), and PKS~1127--145

536: ($z=0.313$) are underabundant in their $\lambda$6284 DIB strengths by

537: factors of 4-10 times expected from the Galactic DIB--$N({\HI})$

538: relation.

539: 

540: A number of environmental factors such as reddening, metallicity, and

541: ionizing radiation may be responsible for DLAs not following the

542: Galactic DIB-$N({\HI})$ relation.  Whether it is one or a combination

543: of these factors, the data suggest that the environments probed by DLA

544: sightlines differ from those within the Milky Way in which DIBs are

545: present.  The environments that give rise to the DIBs are likely very

546: localized.  Galactic properties vary even on small scales, which is

547: why \citet{cox07} find varying differences and DIB strengths along

548: different sightlines within the larger confines of the SMC.  QSO

549: sightlines probe relatively small transverse spatial scales in DLA

550: galaxies.  Therefore, our results do not eliminate the possibility

551: that DIBs follow the Galactic DIB--$N({\HI})$ relationships elsewhere

552: in these galaxies.  We are less inclined to suggest that our results

553: indicate redshift evolution of organics, since infrared emission from

554: PAHs has been observed in the extremely dusty environments of ULIRGs

555: as high as redshift $z \sim 2$ \citep{yan05}.

556: 

557: \subsection{Reddening}

558: \label{sec:reddening}

559: 

560: There has been a long history of investigating DIB dependence on color

561: excess \citep{merr38, herb93}.  Whereas Galactic sight-lines along the

562: disk (low Galactic latitudes) have typical reddening values of 0.1 to

563: 1.0, DLAs typically have lower reddening.  \citet{murp04} use Sloan

564: Digital Sky Survey (SDSS) QSO spectra to estimate $E(B-V)<0.02$ mag

565: (3~$\sigma$) along DLA sightlines at $z \sim 3$, assuming SMC-like

566: extinction.  \citet{elli05} use radio selected QSOs from the Complete

567: Optical and Radio Absorption Line System (CORALS) survey to estimate

568: reddening along DLA sightlines.  The CORALS survey avoids the

569: potential problem of bias from optical luminosity selected samples

570: that may inhibit the detection of more reddened QSOs.  They find

571: $E(B-V)<0.04$ (3~$\sigma$) assuming SMC-like extinction for $1.9 \leq

572: z \leq 3.5$ DLAs.  Using a sample of $0.8 \leq z \leq 1.3$ DLAs

573: selected by {\CaII} absorption, \citet{wild05} estimate an average

574: reddening of $E(B-V)=0.06$ assuming LMC and SMC extinction curves.

575: 

576: The correlations between reddening and the $\lambda$5780,

577: $\lambda$5797, and $\lambda$6284 DIB strengths are tight among Galactic

578: and Magellanic Cloud sightlines, with best fits \citep{welt06},

579: \begin{equation}\label{EQ:5780_reddening}

580: \log\mbox{EW}[E(B-V)]_{\lambda5780} = 0.99 \log E(B-V) + 2.65,

581: \end{equation}

582: \begin{equation}\label{EQ:5797_reddening}

583: \log\mbox{EW}[E(B-V)]_{\lambda5797} = 0.99 \log E(B-V) + 2.26,

584: \end{equation}

585: and

586: \begin{equation}\label{EQ:6284_reddening}

587: \log\mbox{EW}[E(B-V)]_{\lambda6284} = 0.80 \log E(B-V) + 3.08.

588: \end{equation}

589: The rms scatter of the measured equivalent widths about these

590: relationships are less than 0.15 dex; in other words, the

591: $\lambda$5780, $\lambda$5797, and $\lambda$6284 DIBs fairly equally

592: obey the DIB--$E(B-V)$ relation in the Galaxy.

593: Combining Galactic and extragalactic data, \citet{elli07} find a

594: $\lambda$5780 DIB--$E(B-V)$ relation with a slightly larger slope,

595: 1.27, and intercept, 2.77.  Compared to \citet{welt06}, the

596: \citet{elli07} results yield a 10\% difference in

597: $\log\mbox{EW}[E(B-V)]_{\lambda5780}$ for unit reddening.

598: 

599: Correlations with the $\lambda4428$ DIB equivalent width are

600: notoriously problematic to compute because of the difficulty in

601: measuring the DIB's broad width.  Previous work has noted a

602: correlation of the $\lambda6613$ DIB equivalent width with $E(B-V)$

603: \citep{megi05,thor03,wese01}.  In the case of both the $\lambda$4428

604: and $\lambda$6613 DIBs, we employ the reddening relations provided by

605: T.P. Snow (unpublished, private communication).  Using the best-fit

606: for 75 Galactic sightlines for $\lambda4428$ and 123 Galactic

607: sightlines for $\lambda6613$,

608: \begin{equation}\label{EQ:4428_reddening}

609: \mbox{EW}[E(B-V)]_{\lambda4428} = 2093.99 E(B-V),

610: \end{equation}

611: and

612: \begin{equation}\label{EQ:6613_reddening}

613: \mbox{EW}[E(B-V)]_{\lambda6613} = 217.06 E(B-V).

614: \end{equation}

615: The scatter about these relations is relatively large when compared to

616: the other DIBs studied in this work.  The $\lambda4428$ DIB--$E(B-V)$

617: relation has a Pearson correlation coefficient of 0.825, while the

618: $\lambda6613$ DIB--$E(B-V)$ relation has a Pearson correlation

619: coefficient of 0.831 (unpublished, T.P. Snow, private communication).

620: 

621: Assuming the upper limit measured for high redshift DLAs of

622: $E(B-V)=0.04$ from \citet{elli05}, we computed the expected DIB

623: equivalent widths using Eqs.~\ref{EQ:5780_reddening},

624: \ref{EQ:5797_reddening}, \ref{EQ:6284_reddening},

625: \ref{EQ:4428_reddening}, and \ref{EQ:6613_reddening}.  The only

626: exception is the AO~0235+164 DLA, which has a measured reddening of

627: $E(B-V)=0.23$ \citep{junk04}.  The expected DIB equivalent widths are

628: listed in Table~\ref{tab:Models}.  Assuming these reddening values,

629: the expected DIB equivalent widths are systematically smaller than

630: those predicted by the DIB--$N({\HI})$ relationship.  In

631: Figs.~\ref{s:0235}--\ref{s:1229_VLT} we show the expected DIB profiles

632: (thick curves).  With the exception of the $\lambda6284$ DIB limit

633: measured from the spectrum of the DLA toward AO~0235+164 and the

634: $\lambda4428$ DIB limit measured from the spectrum of the DLA toward

635: PKS~1127--145, the expected DIB equivalent widths for adopted

636: reddening values are smaller than our measured equivalent width

637: limits; thus, if these reddening values are appropriate for

638: intermediate redshift DLAs, we would not have detected these DIBs.  As

639: discussed in \S~\ref{sec:analysis}, the $\lambda6284$ DIB limit for

640: the AO~0235+164 system is problematic due to a large atmospheric

641: feature and the $\lambda4428$ DIB limit for the PKS~1127--145 system

642: is unreliable due to difficulties in continuum fitting the broad DIB

643: within the high resolution VLT/UVES data.

644: 

645: Assuming the Galactic DIB--$E(B-V)$ relations are valid for DIBs in

646: DLAs, we used our measured equivalent width limits for the

647: $\lambda$4428, $\lambda$5780, $\lambda$5797, $\lambda$6284, and

648: $\lambda$6613 DIBs to compute upper limits on the reddening for the

649: DLAs in our sample directly from the weighted slopes of the

650: DIB--$E(B-V)$ relationships (Eqs.~\ref{EQ:5780_reddening},

651: \ref{EQ:5797_reddening}, \ref{EQ:6284_reddening},

652: \ref{EQ:4428_reddening}, and \ref{EQ:6613_reddening}) from

653: \citet{welt06} and T.P. Snow (unpublished, private communication). The

654: reddening limits, $E(B-V)_{\rm lim}$, are listed in

655: Table~\ref{tab:Models} for each DIB.

656: 

657: In Fig.~\ref{p:reddening_law}, we illustrate the sensitivity of this

658: method for $E(B-V)$ values of 0.02, 0.04, 0.10, 0.20, 0.40, and 1.0,

659: respectively.  Plotted are $\log (\hbox{EW Limit}/\hbox{EW}[E(B-V)])$

660: for each DLA for the $\lambda$4428, $\lambda$5780, $\lambda$5797,

661: $\lambda$6284, and $\lambda$6613 DIBs (panels $a$, $b$, $c$, $d$, $e$,

662: respectively).  For the condition $\log (\hbox{EW

663: Limit}/\hbox{EW}[E(B-V)]) \leq 0$, the reddening predicted equivalent

664: width for a given DIB is greater than or equal to the upper limits

665: afforded by our data.  A value of $\log (\hbox{EW

666: Limit}/\hbox{EW}[E(B-V)])=0$ corresponds to our computed reddening

667: upper limit, $E(B-V)_{\rm lim}$, at which the DIB should just become

668: detectable in our data.  The vertical error bars define the 1~$\sigma$

669: uncertainties in $\log (\hbox{EW Limit}/\hbox{EW}[E(B-V)])$ evaluated

670: at our computed $E(B-V)_{\rm lim}$.  The uncertainties are calculated

671: using standard error propagation taking into account the uncertainties

672: in both the W~Limit and the slope of the EW[$E(B-V)$] relations from

673: \citet{welt06} and T.P. Snow (unpublished, private communication).

674: The $\lambda4428$ and $\lambda5780$ DIBs for the DLA toward

675: AO~0235+164 are left blank because they are confirmed detections

676: \citep{junk04,lawt06} with a known reddening of $E(B-V)=0.23$

677: \citep{junk04}.

678: 

679: Note that in Fig.~\ref{p:reddening_law}, the $\lambda6613$

680: DIB--$E(B-V)$ relation is the least constraining of the DIBs in

681: determining the upper reddening limit, $E(B-V)_{\rm lim}$.

682: Furthermore, due to the scatter in the $\lambda4428$ and $\lambda6613$

683: DIB--$E(B-V)$ relations, the errors in the best-fit slope in

684: Eqs.~\ref{EQ:4428_reddening} and \ref{EQ:6613_reddening} are

685: relatively large which results in the large errors displayed in

686: Fig.~\ref{p:reddening_law}$a$ and \ref{p:reddening_law}$e$.  For these

687: reasons, we do not adopt upper reddening limits, $E(B-V)_{\rm lim}$,

688: from the $\lambda4428$ or $\lambda6613$ DIBs.

689: 

690: Our final adopted reddening limit for a given DLA is determined from

691: either the $\lambda$5780 or $\lambda$6284 DIBs.  The adopted limits

692: are listed in Table~\ref{tab:adoptedreddening}. Two of the DLAs have

693: limits of $E(B-V) \leq 0.05$ (Q0738+313 $z=0.091$ and PKS~0952+179),

694: both from the $\lambda$6284 DIB.  The DLA toward B2~0827+243 has an

695: upper limit to the reddening of $E(B-V) \leq 0.07$ from the

696: $\lambda$5780 DIB (the $\lambda$6284 DIB is not covered).  The

697: $\lambda$6284 DIB provides upper reddening limits for the remaining

698: DLAs toward Q1229--020 ($E(B-V) \leq 0.08$), Q0738+313 $z=0.221$

699: ($E(B-V) \leq 0.14$), and PKS~1127--145 ($E(B-V) \leq 0.21$).  If we

700: were to apply constraints from the $\lambda$4428 DIB, we would infer

701: upper reddening limits of 0.07, 0.04, and 0.03 for the $z=0.221$ DLA

702: toward Q0738+313, and the DLAs toward B2~0827+243 and PKS~1127--145,

703: respectively.

704: 

705: For the DLA at $z=0.524$ toward AO~0235+164, there are two detections

706: from \citet{lawt06} (the $\lambda$5705 and $\lambda$5780 DIBs) and one

707: detection from \citet{junk04} (the $\lambda$4428 DIB).  The equivalent

708: widths measured from the $\lambda$4428 and $\lambda$5780 DIBs are

709: larger than expected from the measured $E(B-V) = 0.23$ \citep{junk04}.

710: The $\lambda$5797 and $\lambda$6613 DIBs give limits consistent with

711: this, but the $\lambda$6284 DIB suggests an $E(B-V) \leq 0.06$

712: \textit{by this technique}.  Essentially, we should have discovered

713: the $\lambda$6284 DIB assuming our technique is correct because

714: \citet{junk04} measures an $E(B-V)=0.23\pm0.01$.

715: 

716: A plausible reason for our non-detection of the $\lambda$6284 DIB in

717: the AO~0235+164 DLA is that the expected position of the $\lambda$6284

718: DIB at $z=0.524$ resides directly in a broad atmospheric absorption

719: band (see panel $b$ in Fig.~\ref{s:0235}).  \citet{lawt06} attempted

720: to overcome this by modeling the $\lambda$6284 DIB with various

721: equivalent widths and convolving it with the atmospheric band.  The

722: limit they achieve (see Table~\ref{tab:EWs}) is based on the minimum

723: equivalent width needed to separate the DIB out of the atmospheric

724: band.  This is a robust way to measure the limit.  However, there is

725: some inherent difficulty in this method because it has to be assumed

726: \textit{a priori} that the atmospheric band is not contaminated by DIB

727: absorption.  A second potential reason that we do not detect the

728: $\lambda$6284 DIB in the AO~0235+164 DLA, given its high reddening, is

729: that this DIB may not follow the Galactic DIB--$E(B-V)$ relation in

730: DLAs.  Indeed, in the Magellanic Clouds, $\lambda$6284 DIBs are weaker

731: than their Galactic counterparts by a factor of two relative to the

732: Galactic $\lambda$6284 DIB--$E(B-V)$ relation \citep{welt06}.  Given

733: that the $\lambda$6284 DIBs in four of our DLAs are constrained to be

734: 4--10 times weaker than the Galactic $\lambda$6284 DIB--$N(\HI)$

735: relation, it is reasonable that DIB relations in DLAs could be more

736: Magellanic Cloud--like than Galactic--like.

737: 

738: %The detection of the $\lambda$5780 DIB in the AO~0235+164 DLA provides

739: %some evidence that DIBs in DLAs do not follow the Galactic $\lambda$5780

740: %DIB--$E(B-V)$ relation.  \citet{elli07} reports a new fit to the

741: %DIB--$E(B-V)$ relation with all known extragalactic points included.

742: %Their best-fit line is slightly steeper ($\log \hbox{EW}=1.27 \log

743: %E(B-V)+2.77$) than the Galactic best-fit line from \citet{welt06}

744: %($\log \hbox{EW}=0.99 \log E(B-V)+2.65$).  The two relations do not

745: %produce a large difference in the $E(B-V)_{\rm lim}$ for the $\lambda$5780

746: %DIB.  From \citet{welt06}, we infer $E(B-V)_{lim}=0.48$, whereas, we

747: %infer $E(B-V)_{lim}=0.45$ from \citet{elli07}.  Both of these values

748: %exceed the measured reddening of $E(B-V)=0.23$ from \citet{junk04}.

749: %Unfortunately, since we have only one $\lambda$5780 DIB detected in a

750: %DLA, it is not yet possible to deduce a $\lambda$5780 DIB--$E(B-V)$

751: %relation for DLAs.

752: 

753: Our observations have provided reddening constraints of $0.05 \leq

754: E(B-V) \leq 0.08$ for four DLAs at $0.09 \leq z \leq 0.52$.  This

755: extends to lower redshift the work of \citet{elli05}, who report

756: $E(B-V) \leq 0.04$ for $1.9 \leq z \leq 3.5$ DLAs, and of

757: \citet{murp04}, who report $E(B-V) \leq 0.02$ for $z \sim 3$ DLAs (all

758: limits are $3~\sigma$).  Reddening is an indirect measure of dust

759: content.  The slope of any DIB--$E(B-V)$ relation likely has a strong

760: dependence on the nature of the dust (particle size, abundance, and

761: composition).  \citet{junk04} argue that the dust in the AO~0235+164

762: DLA is more Galactic-like than Magellanic-like, but with fewer small

763: particles.  Because the organics responsible for the DIBs may depend

764: on dust for formation and/or survival, the abundance and nature of the

765: dust in the DLA toward AO~0235+164 might be responsible for the

766: presence of DIBs observed in this DLA.

767: 

768: \subsection{Gas-to-Dust Ratios}

769: 

770: Assuming that our method of estimating $E(B-V)_{\rm lim}$ realistically

771: reflects the upper limit on reddening in our DLAs, we can estimate

772: lower limits in their gas-to-dust ratios, $N({\HI})/E(B-V)$.  We

773: present the gas-to-dust ratio lower limits in

774: Table~\ref{tab:adoptedreddening}; they are computed using the measured

775: $N({\HI})$ from the literature and our adopted upper $E(B-V)$ limits.

776: The lower limits range from 2.9~$\times$~10$^{21}$ to

777: 42~$\times$~10$^{21}$~cm$^{-2}$~mag$^{-1}$.  Note that the gas-to-dust

778: ratio 19.2~$\times$~10$^{21}$~cm$^{-2}$~mag$^{-1}$ for the $z=0.524$

779: DLA toward AO~0235+164 was measured by \citet{junk04}.

780: 

781: \citet{bouc85} measure lower and upper values, 37~$\times$~10$^{21}$

782: and 52~$\times$~10$^{21}$~cm$^{-2}$~mag$^{-1}$, respectively, for SMC

783: gas-to-dust ratios.  \citet{gord03} determined a gas-to-dust ratio of

784: 19.2~$\times$~10$^{21}$~cm$^{-2}$~mag$^{-1}$ for the LMC (from their

785: LMC-2 data).  The average LMC sample of \citet{gord03} yields

786: 11.1~$\times$~10$^{21}$~cm$^{-2}$~mag$^{-1}$.  A linear fit by

787: \citet{cox06b} to their LMC data gives a similar gas-to-dust ratio of

788: 14.3~$\times$~10$^{21}$~cm$^{-2}$~mag$^{-1}$.  For the Milky Way,

789: \citet{bohl78} find a gas-to-dust ratio of

790: 4.8~$\times$~10$^{21}$~cm$^{-2}$~mag$^{-1}$; whereas, \citet{cox06b}

791: find a ratio of 4.03~$\times$~10$^{21}$~cm$^{-2}$~mag$^{-1}$ from the

792: fit to their Galactic data.

793: 

794: Although the SMC, LMC, and Milky Way appear to have distinct ranges of

795: gas-to-dust ratios, in reality, there is a continuum that likely

796: reflects the variation of local environments in each.  \citet{gord03}

797: measure $E(B-V)$ and $N({\HI})$ toward four stars in the bar of the

798: SMC and one star in the wing of the SMC.  From these data, the

799: gas-to-dust ratios in the SMC bar span the range 17--$51 \times

800: 10^{21}$~cm$^{-2}$~mag$^{-1}$.  These values overlap with both the LMC

801: gas-to-dust ratios of \citet{gord03} and \citet{cox06b} and the SMC

802: gas-to-dust ratios of \citet{bouc85}.  The SMC wing data yield

803: 15.2~$\times$~10$^{21}$~cm$^{-2}$~mag$^{-1}$, which falls very near

804: the \citet{cox06b} LMC fit.

805: 

806: \citet{cox07} searched for DIBs toward six lines of sight in the bar

807: and wing of the SMC.  The sightline AzV456, which is in the wing of

808: the SMC, is the only sightline in which they detected DIBs.  From

809: their $E(B-V)$ and $N({\HI})$ data (Table~9 on--line material), the

810: gas-to-dust ratio in the SMC wing is $\sim 7$ times lower ($\sim 7.4

811: \times 10^{21}$ cm$^{-2}$ mag$^{-1}$) than the gas-to-dust ratio in

812: the SMC bar ($\sim 52.4 \times 10^{21}$ cm$^{-2}$ mag$^{-1}$).  This

813: SMC wing value lies between the \citet{bohl78} and \citet{cox06b}

814: Milky Way and the \citet{gord03} and \citet{cox06b} LMC gas-to-dust

815: ratios.  The SMC bar value lies just above the upper range of the

816: \citet{bouc85} SMC gas-to-dust ratio.  It is of interest to note that

817: \citet{cox07} detected DIBs only in the sightline with the smallest

818: gas-to-dust ratio in their sample and that this ratio approaches the

819: range observed for the Milky Way.

820: 

821: \citet{cox07} argue that the SMC wing is quiescent while the star

822: formation regions of the SMC bar are turbulent and exposed to larger

823: UV fluxes.  \citet{welt06} calculate an increase of $\sim28$--83 times

824: the average interstellar radiation field (ISRF) for sightlines in the

825: SMC bar near star-forming {\HII} regions; whereas, the SMC wing is

826: more similar to the Galaxy with a radiation level of $\sim0.6$ times

827: the average ISRF.  The variation in the gas-to-dust ratios within the

828: SMC are likely a reflection of the balance of dust formation

829: processes, such as accretion, with dust destruction \citep{cox07}.

830: Thus, the dust, and possibly the organics, may be destroyed in the SMC

831: bar explaining why \citet{cox07} could not detect DIBs in their sample

832: of SMC bar stars.

833: 

834: In Fig.~\ref{p:Gas_Dust} we plot $\log N({\HI})$ versus $\log

835: E(B-V)_{\rm lim}$ for the DLAs in our sample.  The $N({\HI})$ data

836: points are taken from Table~\ref{tab:DLAs} and the $E(B-V)_{\rm lim}$

837: are taken from Table~\ref{tab:adoptedreddening}.  The $\log N({\HI})$

838: error bars are $1~\sigma$.  The leftward arrows represent reddening

839: upper limits for our sample.  For Q0738+313 at $z=0.221$, B2~0827+243,

840: and PKS~1127--145, the $\lambda$4428 DIB provides more stringent

841: limits on $E(B-V)_{\rm lim}$.  However, they should be considered less

842: robust due to the large errors as shown in Fig.~\ref{p:reddening_law},

843: panels $a$ and $e$.

844: 

845: Although the gas-to-dust ratios in the SMC, LMC, and Milky Way exhibit

846: some overlap, in general the Milky Way ISM values tend to be lower

847: than those of the LMC which tend to be lower than those of the SMC

848: \citep{bohl78,bouc85,gord03,cox06b}.  In Fig.~\ref{p:Gas_Dust}, the

849: dot-dashed lines illustrate the upper and lower SMC values of

850: \citet{bouc85}.  The middle three lines in Fig.~\ref{p:Gas_Dust}

851: represent gas-to-dust ratios for the LMC, and the bottom two lines

852: represent those for the Galaxy.  The long-dashed line gives the LMC-2

853: gas-to-dust ratio derived from \citet{gord03}.  The \citet{cox06b}

854: ratio for the LMC is given by the dot-dot line.  The short-dashed line

855: is the \citet{gord03} ratio from their average LMC sample.  The dashed

856: MW line is the \citet{bohl78} ratio, and the dotted MW line is the

857: \citet{cox06b} ratio.

858: 

859: As shown in Fig.~\ref{p:Gas_Dust}, the DLA toward PKS~0952+179 has a

860: gas-to-dust ratio that is consistent with or greater than the SMC.

861: Two of the DLAs in our sample, Q0738+313 ($z=0.091$) and

862: PKS~1127--145, have ratios consistent with or greater than the LMC

863: gas-to-dust ratios. Two, Q0738+313 ($z=0.221$) and Q1229--020, are

864: consistent with or greater than the Galactic ratios.  The ratio for

865: B2~0827+243, which has the lowest $N({\HI})$ in our sample, is not

866: well constrained.  The measured ratio for the $z=0.524$ DLA toward

867: AO~0235+164 is $N({\HI})$/$E(B-V)=19.2 \times 10^{21}$ cm$^{-2}$

868: mag$^{-1}$ \citep{junk04}, which is consistent with the LMC-2 ratio of

869: \citet{gord03} and four times larger than the Galactic ratio from

870: \citet{bohl78}.  This DLA has a very large $N({\HI})$, and the largest

871: $E(B-V)$ \citep{junk04} of the DLAs presented here.

872: 

873: Three DLAs are constrained to have slightly larger gas-to-dust ratios

874: than the $z=0.524$ DLA toward AO~0235+164, the $z=0.091$ DLA toward

875: Q0738+313 ($\simeq 1.5$ times greater), $z=0.239$ DLA toward

876: PKS~0952+179 ($\simeq 2.2$ times greater), and the $z=0.313$ DLA

877: toward PKS~1127--145 ($\simeq 1.2$ times greater).  However, for the

878: DLA toward PKS~1127--145, if we apply $E(B-V)_{\rm lim}$ from the

879: $\lambda$4428 DIB, the gas-to-dust ratio is $\simeq 7$ times greater

880: than that of the AO~0235+164 DLA and would be more consistent with

881: gas-to-dust ratios greater than measured in the SMC.  Again, we

882: caution that this more stringent ratio is not adopted due to large

883: errors from the scatter in the $\lambda4428$ DIB--$E(B-V)$ relation

884: (see Eq.~\ref{EQ:4428_reddening}).

885: 

886: The sightlines toward DLAs are certainly gas rich, which they share

887: with Galactic sightlines that contain DIBs.  However, their

888: environments, as probed by the background QSO, may have large

889: variations in $E(B-V)$, metallicity, and radiation.  Understanding the

890: magnitudes and properties of the dust, radiation, {\HI} and H$_{2}$

891: content, and metallicity in DLA samples holds promise of revealing

892: whether the environments of DLAs are conducive for the organics that

893: give rise to DIBs.  Reddening is seemingly low in the sightlines

894: probing the galaxies in this sample, which may play a pivotal role in

895: inhibiting DIB strengths.  The strengths of DIBs in DLAs do not follow

896: the Galactic $N({\HI})$ relation (see Fig.~\ref{p:NHI_known}).

897: However, the $\lambda$5780 detection in the AO~0235+164 DLA is

898: consistent with expectations from the DIB--$E(B-V)$ relation for

899: Galactic and extragalactic points \citep{elli07}.  Thus, gas-to-dust

900: ratio may not be as good a predictor as reddening for the presence of

901: DIBs in DLAs.  This would suggest that DIBs might be present in DLAs

902: with a wide range of gas-to-dust ratios, but that they lie within the

903: right-hand region of the $\log N({\HI})$--$\log E(B-V)$ plane of

904: Fig.~\ref{p:Gas_Dust}.

905: 

906: \citet{elli05} conclude that their sample of $1.9 \leq z \leq 3.5$

907: high redshift DLAs does not follow a Galactic gas-to-dust relation,

908: which is consistent with what we find for four of the DLAs in this

909: work (AO~0235+164 at $z=0.524$, Q0738+313 at $z=0.091$, PKS~0952+179

910: at $z=0.239$, and PKS~1127--145 at $z=0.313$).  However, as found by

911: \citet{elli05}, we cannot conclude whether the gas-to-dust ratios in

912: our lower redshift DLAs are consistent with the SMC gas-to-dust

913: ratios.  The remaining three DLAs (Q0738+313 at $z=0.221$,

914: B2~0827+243, and Q1229--020) do not have sufficient reddening limits

915: to compare with the high redshift sample of \citet{elli05}.

916: 

917: \subsection{Metallicity versus Reddening}

918: 

919: \citet{cox07} give evidence that metallicity plays an important role

920: in DIB strength since a high carbon abundance is necessary for the

921: creation of the organic molecules and the dust grains on which they

922: may be formed.  In Table~\ref{tab:DLAs}, we list the zinc and iron

923: abundances from the literature for our DLA sample.  Zinc is a good

924: metallicity indicator because it traces iron-group abundances and does

925: not readily deplete on dust.  Unfortunately, only upper limits on zinc

926: abundances have been measured, with the exception of the Q1229--020

927: DLA, where $\mbox{[Zn/H]}=-0.47$ \citep{bois98}.  Thus, we cannot

928: address any trends of DIB strengths with metallicity.

929: 

930: The Q1229--020 DLA has a metallicity roughly 0.5 dex above the average

931: metallicity of DLAs, and its metallicity is consistent with that of

932: the AO~0235+164 DLA.  Since it has a relatively stringent upper limit on

933: the reddening, $E(B-V)_{lim} \leq 0.08$, much below the $E(B-V) =

934: 0.23$ of the AO~0235+164 DLA, it may be a promising candidate for

935: contributing to an understanding of the role of metallicity in

936: determining DIB strengths in DLAs.  If metallicity plays a role, it is

937: expected that DIB strengths in DLAs would to some degree scale with

938: $N(\HI)$.  The $N(\HI)$ of the Q1229--020 DLA is a full dex below that

939: of the AO~0235+164 DLA.  Thus, that our upper limit on the equivalent

940: width of the $\lambda$5780 DIB in the Q1229--020 DLA is a factor of

941: 1.7 below that detected for the AO~0235+164 DLA is not constraining.

942: A very deep spectrum of the DIBs in the Q1229--020 DLA would be very

943: interesting.

944: 

945: Selecting DLAs by metal absorption would represent another approach to

946: searching for DIBs in external galaxies, since these DLAs may

947: represent a more reddened population.  \citet{wild06} propose that

948: {\CaII} absorbers may represent an intermediate link between the

949: quiescent, metal-poor and dust-poor DLAs and the intermediate

950: redshift, star-forming, and metal-rich Lyman break galaxies with

951: typical reddening of $E(B-V)=0.15$ to $0.20$.  \citet{elli07}

952: searched for DIBs in nine $0.07 \leq z \leq 0.55$ {\CaII}-selected

953: absorbers, and detected the $\lambda$5780 DIB in only one.  At this early

954: juncture, the {\CaII} selection success rate is comparable to that of

955: {\HI}-selected DLAs.  {\MgII} absorbing DLAs may also represent a

956: slightly more reddened population of DLAs \citep{mena07}.

957: 

958: \section{Conclusions}

959: 

960: \label{sec:conclusions}

961: 

962: In this paper we employ a generalized method of the \citet{schn93}

963: technique to find lines and determine conservative equivalent width

964: limits for DIBs in seven DLAs along with an assessment of

965: uncertainties in these limits.  We find:

966: 

967: (1) The $\lambda$6284 DIB in four of the DLAs in our sample have

968:     equivalent width upper limits that are 4--10 times lower then

969:     expected for similar $N({\HI})$ relative to Galactic sight lines.

970:     These limits are not inconsistent with the $\lambda$6284 DIB

971:     strengths found in the LMC and SMC.

972: 

973: (2) Assuming the $\lambda$5780 and $\lambda$6284 DIB--$E(B-V)$

974:     relations hold for DLAs, as it does for Galactic and Magellanic

975:     Cloud sightlines, we estimated upper limits on the reddening for

976:     our sample.  In four of our DLAs we estimate $E(B-V) \leq 0.08$

977:     with two DLAs having an $E(B-V) \leq 0.05$.  These results are

978:     consistent with high redshift DLA samples, which have $E(B-V) <

979:     0.04$ \citep{elli05} and $E(B-V) < 0.02$ \citep{murp04}.

980: 

981: (3) Applying the $E(B-V)$ limits, one of our DLAs is consistent with

982:     having the same or larger gas-to-dust ratio as the SMC. Two of our

983:     DLAs are consistent with having gas-to-dust ratios at least as

984:     large as sightlines in the LMC.  Three of our DLAs have less

985:     stringent limits that give lower limit gas-to-dust ratios

986:     consistent with Galactic, LMC, or SMC sightlines.  The AO~0235+164

987:     DLA has a measured $E(B-V)$ and $N({\HI})$ that puts its

988:     gas-to-dust fraction on the high end of the LMC sightlines as

989:     stated in other work \citep{junk04}.  The $E(B-V)_{lim} \leq 0.06$

990:     constrained from the $\lambda$6284 DIB is inconsistent with the

991:     known $E(B-V) = 0.23$ measured in \citet{junk04}.  We should have

992:     been able to detect the $\lambda$6284 DIB in the AO~0235+164 DLA

993:     given the limit adopted by \citet{lawt06}.  Three possibilities

994:     are that our method of determining the $E(B-V)_{\rm lim}$ does not

995:     apply to DLAs, our limits do not adequately take into account the

996:     large atmospheric absorption band (see Fig.~\ref{s:0235}$b$), or

997:     that conditions are not favorable to the formation or survival of

998:     this DIB carrier.

999: 

1000: It is interesting to speculate whether ionization conditions may be an

1001: important factor in inhibiting or enhancing DIB strengths.

1002: \citet{welt06} test the ionization effects of DIBs along Galactic,

1003: LMC, and SMC lines of sight.  However, they do not find any

1004: significant trends.  On the contrary, \citet{cox07} measure the UV

1005: radiation field along lines of sight toward the SMC and find UV

1006: radiation is an important environmental factor in DIB strengths.

1007: Laboratory spectroscopists claim that the DIBs may be due to partially

1008: ionized PAHs because they produce a wealth of absorption features in

1009: the optical spectrum \citep{snow01}.  If this is the case then a

1010: significant UV radiation may be required.  However, UV radiation that

1011: is too high will dissociate the molecules.  In this work, we are

1012: unable to explore the affects of ionization conditions; knowledge of

1013: ionization conditions in specific DLAs is difficult to obtain.

1014: 

1015: It is also interesting to speculate whether metallicity may be an

1016: important factor in inhibiting or enhancing DIB strengths.  However,

1017: since we have robust metallicity measurements for only two of the DLAs

1018: in our sample, we cannot directly address the affects of metallicity.

1019: We do point out that obtaining deep spectra of DIBs in DLAs with known

1020: metallicity could be a fruitful future research direction, especially

1021: if reddening were also known. A direct comparison between the low

1022: metallicity Q1229--020 DLA and the high metallicity AO~0235+164 DLA

1023: might be fruitful as a first examination of the affects of metallicity

1024: in determining what governs DIB strengths in DLAs.

1025: 

1026: \subsection{Implications of our Results}

1027: 

1028: Our results imply that reddening is a more crucial indicator of DIB

1029: strengths than is {\HI} content for DLAs; the low reddening in DLA

1030: selected galaxies inhibits the presence of DIBs.  However, our sample

1031: is small and additional observations are required to ascertain the

1032: strength of this statement (metallicity and ionization conditions

1033: likely play a role as well).

1034: 

1035: The weakness of DIB strengths in our sample hints that the

1036: environments of high $N({\HI})$ DLA-selected galaxies may be less

1037: suitable to create and/or sustain the organic molecules than those of

1038: the Galaxy.  Not only is the immediate solar environment beneficial

1039: for sustaining life, but it may be that the Galaxy is a more

1040: hospitable location for the survival of organic molecules that may

1041: have been the precursors to biology on Earth.  If these molecules are

1042: important as precursors to life in the universe, charting their

1043: presence to high redshift places constraints on how long ago and in

1044: which environments life could have potentially formed.  The presence

1045: of DIBs in the AO~0235+164 DLA demonstrates that organic molecules

1046: existed in at least one DLA selected environment at a redshift of $z

1047: \sim 0.5$, or some 5 Gyrs ago \citep{lawt06}.

1048: 

1049: Due to their weakness relative to DIBs in the Galaxy, observing DIBs

1050: in the general population of high redshift galaxies (as opposed to

1051: ULIRGs and star bursting galaxies) remains a challenge.  Selecting

1052: galaxies by DLA absorption may yet prove to be a lucrative method for

1053: detecting DIBs at high redshifts, provided the detection sensitivity

1054: can be increased.  However, most known DLAs reside at high redshift

1055: where the DIBs move into the near-IR, where high sensitivity

1056: spectroscopy is time intensive.

1057: 

1058: \acknowledgments

1059: 

1060: BL acknowledges support via NASA's Graduate Student Researchers

1061: Program (GSRP grant NNG04GN55H).  CWC acknowledges NSF grant

1062: AST-0708210.  BAY acknowledges the support of NSERC via their

1063: post-graduate scholarship (PGS) program, and TPS acknowledges NASA

1064: grant NNG04GL34G for partial support of this work.  Some data were

1065: obtained from the Gemini telescope\footnote{Based on observations

1066: obtained at the Gemini Observatory, which is operated by the

1067: Association of Universities for Research in Astronomy, Inc., under a

1068: cooperative agreement with the NSF on behalf of the Gemini

1069: partnership: the National Science Foundation (United States), the

1070: Science and Technology Facilities Council (United Kingdom), the

1071: National Research Council (Canada), CONICYT (Chile), the Australian

1072: Research Council (Australia), CNPq (Brazil) and SECYT (Argentina)}

1073: under program GS-2004A-Q-32.  We thank M. Murphy (Swinburne University

1074: of Technology) for the VLT/UVES spectrum and W.L.W. Sargent (Caltech)

1075: for the Keck/HIRES spectrum used in this project.  We also acknowledge

1076: Glenn Kapcrzak (NMSU) for contributed computations.

1077: 

1078: Facilities: \facility{VLT(FORS2, UVES)}, \facility{APO(DIS)},

1079: \facility{Keck(HIRES)}, \facility{WHT(ISIS), \facility{Gemini(GMOS)}}.

1080: 

1081: \appendix

1082: 

1083: \section{Appendix}

1084: 

1085: \subsection{Measuring Equivalent Width Detection Thresholds}

1086: 

1087: A straight forward method for determining equivalent width limits was

1088: presented by \citet{lanz87}.  The uncertainty (or error) spectrum is

1089: used to compute the equivalent width limit in each individual pixel,

1090: $i$,

1091: \begin{equation}

1092: \sigma _{w_i} = D_i \frac{E(\lambda _i)}{I_{c}(\lambda _i)} ,

1093: \end{equation}

1094: where $D_i$ is the pixel dispersion in angstroms, $E(\lambda _i)$ is

1095: the uncertainty in the flux at $\lambda _i$, and $I_{c}(\lambda _i)$

1096: is the estimated continuum flux at $\lambda _i$.  The $1~\sigma$ equivalent width

1097: detection threshold (limit) centered at $\lambda_i$ is then obtained

1098: by summing the individual equivalent width limits of adjacent pixels

1099: over a selected aperture,

1100: \begin{equation}

1101: \sigma _{EW} (\lambda _i) = \left[ \, \sum _{j=j_1}^{j_2} \sigma ^2_{w_j} \right] ^{1/2},

1102: \label{EQ:ltwewlim}

1103: \end{equation}

1104: where $j_1 = i-m/2$, $j_2=i+m/2$, and where $m$, the aperture size, is

1105: an even number of pixels over which the absorption feature is

1106: anticipated.  For unresolved features, $m$ can be taken to be roughly

1107: two resolution elements (i.e., roughly twice the number of pixels of

1108: the full--width half maximum of an unresolved line).  Clearly,

1109: Eq.~\ref{EQ:ltwewlim} can be generalized for resolved, broader

1110: features by increasing $m$ to span roughly twice the full--width half

1111: maximum of the anticipated broad line.  The aperture summation method

1112: of \citet{lanz87} assigns equal weight to all pixels included in the

1113: summation around $\lambda _i$.  This can result in an overestimate of

1114: the equivalent width detection threshold of the data.

1115: 

1116: \citet{schn93} properly treat the relative weighting of pixels

1117: adjacent to $\lambda _i$ by weighting these pixels by the instrumental

1118: spread function, ISF.  The ISF is then a probability weighting

1119: function of the flux centered at $\lambda _i$ with pixel weighting,

1120: $P_j$, defined such that

1121: \begin{equation}

1122: \sum _{j=1}^{m} P_j = 1 ,

1123: \end{equation}

1124: where $m = 2j_o + 1$, and where $j_o$ is a positive integer.  For

1125: unresolved features, the ISF, and therefore the relative values of the

1126: $P_j$, can be taken as a Gaussian function characterized by the

1127: Gaussian width $\sigma _{\rm ISF} = \Delta \lambda _i / 2.35$,

1128: where $\Delta \lambda _i = \lambda _i/R$ is the full--width at half

1129: maximum of the ISF centered at $\lambda _i$ for a spectrograph with

1130: resolution $R$.  The value of $j_o$ is chosen appropriately for the

1131: ISF such that $P_1 = P_m \simeq 0$ (effectively making sure that the

1132: tails of the probability function are sampled out to where the

1133: probabilities vanish).

1134: 

1135: The $1~\sigma$ equivalent width detection threshold for unresolved

1136: features is then computed using

1137: \begin{equation}

1138: \sigma _{EW}(\lambda _i) = D_i 

1139: \displaystyle \left[ \, \sum _{j=1}^{m} P^2_j E^2(\lambda _k)/I_{c}^{2}(\lambda _k) \right]^{1/2} 

1140: \displaystyle \left[ \, \sum _{j=1}^{m} P^2_j \right]^{-1}  ,

1141: \label{EQ:dpsewlim}

1142: \end{equation}

1143: where $k = i + (j-1) - j_o$ is the ``convolution index''.

1144: The elements of the probability weighting function are given by

1145: \begin{equation}

1146: P_j = \frac{\Phi _j} {\displaystyle \sum _{j=1}^{m} \Phi _j } ,

1147: \end{equation}

1148: where 

1149: \begin{equation}

1150: \Phi _j = \exp \left\{ - \frac{ \left( \lambda_k - \lambda _i \right)

1151: ^2 }{2\sigma _{\rm ISF}^2} \right\}  .

1152: \label{EQ:phij}

1153: \end{equation}

1154: If the feature is redshifted, then Eq.~\ref{EQ:dpsewlim} must be divided by

1155: the factor $1+z$ to obtain the rest--frame equivalent width limit.

1156: 

1157: We have generalized Eq.~\ref{EQ:dpsewlim} to (1) account for resolved

1158: features of known FWHM, and (2) account for regions in which

1159: problematic sky--line subtraction in the vicinity of $\lambda _i$

1160: renders the data less certain than quantified by the uncertainty

1161: spectrum, $E(\lambda)$.  We also explicitly include redshift

1162: dependence.

1163: 

1164: To account for resolved features, the $\sigma _{\rm ISF}$ in

1165: Eq.~\ref{EQ:phij} is replaced with the Gaussian width of the line

1166: spread function (LSF) of the redshifted resolved feature, which is

1167: \begin{equation}\label{EQ:sigma}

1168: \sigma _{\rm LSF} = \left[ \left( \frac{\hbox{\small FWHM}(1+z)}{2.35} \right) ^2 +

1169: \sigma ^2_{\rm ISF} \right] ^{1/2} ,

1170: \end{equation}

1171: where FWHM is the rest--frame full--width at half maximum of the

1172: feature, and $z$ is the redshift.  We have assumed that both the

1173: intrinsic line shape and the ISF are well approximated by Gaussian

1174: functions.

1175: 

1176: To account for large residuals from problematic sky--line subtraction,

1177: the error in the flux, $E(\lambda_{k})$, in Eq.~\ref{EQ:dpsewlim} is

1178: replaced by $\hat{E}(\lambda_k)$, which is determined by the quality

1179: of the data at pixel $k$,

1180: \begin{equation}\label{EQ:flux_error}

1181: \hat{E}(\lambda_{k}) = \left\{

1182: \begin{array} {c@{\quad \hbox{for} \quad}l}

1183: |r(\lambda_{k})| & |r(\lambda_{k})| \geq 3 E(\lambda_{k}),\\[6pt]

1184: E(\lambda_{k}) & |r(\lambda_{k})| < 3 E(\lambda_{k}). 

1185: \end{array}

1186: \right.

1187: \end{equation}

1188: where $|r(\lambda_{k})|$ is the residual of the flux,

1189: $I(\lambda_{k})$, with respect to the continuum,

1190: \begin{equation}\label{EQ:residual}

1191: |r(\lambda_{k})| = |I(\lambda_{k}) - I_{c}(\lambda_{k})|.

1192: \end{equation}

1193: Using the residual gives a more conservative estimate of the

1194: equivalent width detection threshold.  The residual of a given pixel

1195: is used whenever the flux significantly deviates from the continuum.

1196: In addition to poor sky--line subtraction, sources of this

1197: deviation can be large telluric features, blending with absorption

1198: features, or a problematic continuum fit.  Using the following

1199: substitution for the normalized flux error simplifies the equations

1200: for further analysis,

1201: \begin{equation}\label{EQ:norm_flux_error}

1202: Y(\lambda_{k}) = \frac{\hat{E}(\lambda_{k})}{I_{c}(\lambda_{k})}.

1203: \end{equation}

1204: 

1205: Applying these criteria transforms Eq.~\ref{EQ:dpsewlim} from

1206: \citet{schn93} into the rest--frame equivalent width detection

1207: threshold calculation used in this work,

1208: \begin{equation}\label{EQ:EWlimit}

1209: \sigma_{EW}(\lambda_i) = \frac{D_{i}}{(1+z)}

1210: \left[ \, \displaystyle \sum_{j=1}^{m} P_{j}^{2}\, Y^{2}(\lambda_{k})\right]^{1/2}

1211: \left[ \, \displaystyle \sum_{j=1}^{m} P_{j}^{2}\right]^{-1} .

1212: \end{equation}

1213: The rest--frame equivalent width limits presented in

1214: Table~\ref{tab:EWs} are quoted at the $3~\sigma$ level.

1215: 

1216: \subsection{Uncertainty Assessment}

1217: 

1218: To quantify the quality of the equivalent width limits, we estimated

1219: the uncertainties in the $\sigma_{EW}(\lambda_i)$.  The equivalent

1220: width limit in Eq.~\ref{EQ:EWlimit} explicitly includes the

1221: spectrograph resolution, $R$, the absorption line FWHM, the continuum

1222: fit, and the central wavelength of the absorption line, $\lambda_l$.

1223: Assuming that the uncertainties in these quantities can be estimated

1224: and are normally distributed, the variances in the equivalent width

1225: limits are obtained from

1226: \begin{equation}

1227: \label{EQ:var}

1228: V_{\sigma_{EW}} = 

1229:   \left[\frac{\partial \sigma_{EW}}{\partial R}\delta R\right]^{2} 

1230: + \left[\frac{\partial \sigma_{EW}}{\partial \hbox{\small FWHM}}\delta \hbox{\small FWHM}\right]^{2} 

1231: + \left[\frac{\partial \sigma_{EW}}{\partial I_{c}(\lambda_{k})}\delta I_{c}(\lambda_{k})\right]^{2} 

1232: + \sigma_{\lambda_{l}}^{2}(\sigma_{EW}),

1233: \end{equation}

1234: where $\delta R$ is the uncertainty in the resolution,

1235: $\delta\hbox{\small FWHM}$ is the uncertainty in the FWHM of the line,

1236: $\delta I_c$ is the uncertainty in the continuum, and

1237: $\sigma_{\lambda_{l}}(\sigma_{EW})$ is the standard deviation in the equivalent width

1238: limit due to uncertainty in the wavelength center of the line.  The

1239: terms in Eq.~\ref{EQ:var} for which explicit indices appear are

1240: computed at the center pixel, $i$, of the feature.

1241: 

1242: \subsubsection{Uncertainty in Resolution}

1243: 

1244: The uncertainty in the resolution, $\delta R$, can be obtained by

1245: measuring the full--width half maximum of unresolved sky lines and

1246: determining the standard deviation.  Applying the chain rule, the

1247: partial derivative of the equivalent width limit with respect to

1248: resolution is

1249: \begin{equation}\label{EQ:Rerror}

1250: \frac{\partial \sigma_{EW}}{\partial R} = 

1251:       \frac{\partial \sigma_{EW}}{\partial P_{j}}\, 

1252: \frac{\partial P_{j}}{\partial \sigma_{\rm LSF}}\, 

1253: \frac{\partial \sigma_{\rm LSF}}{\partial R},

1254: \end{equation}

1255: where,

1256: \begin{eqnarray}

1257: \label{EQ:parPj}

1258: \frac{\partial \sigma_{EW}}{\partial P_{j}} &=& 

1259: \frac{D_{i}}{(1+z)}\left[\frac{C^{1/2}}{G^{2}}\right]\left[\frac{1}{2}\frac{A\, G}{C}-B\right], \\[6pt]

1260: \label{EQ:parsigma}

1261: \frac{\partial P_{j}}{\partial \sigma_{\rm LSF}} &=& P_{j}\left(Q_{j} - T/S\right), \\[6pt]

1262: \frac{\partial \sigma_{\rm LSF}}{\partial R} &=& -\frac{1}{\sigma_{\rm LSF}\, R^{3}}\left(\frac{\lambda_{i}}{2.35}\right)^{2},

1263: \end{eqnarray}

1264: and where,

1265: \begin{equation}

1266: \begin{array}{rclrcl}

1267: \vspace{6pt}

1268: A &=& \displaystyle\sum_{j=1}^{m}2P_{j}\, Y^{2}(\lambda_{k}) \qquad & 

1269: Q_{j} &=& \displaystyle\frac{(\lambda_k - \lambda_i)^{2}}{\sigma_{\rm LSF}^{3}},\\\vspace{6pt}

1270: B &=& \displaystyle\sum_{j=1}^{m}2P_{j} \qquad  & 

1271: T     &=& \displaystyle\sum_{j=1}^{m}\frac{(\lambda_k - \lambda_i)^{2}}{\sigma_{\rm LSF}^{3}} 

1272:           \exp \left\{-\frac{(\lambda_k - \lambda_i)^{2}}{2\sigma_{\rm LSF}^{2}}\right\},\\\vspace{6pt}

1273: C &=& \displaystyle\sum_{j=1}^{m}P_{j}^{2}\, Y^{2}(\lambda_{k}) \qquad  &

1274: S     &=& \displaystyle\sum_{j=1}^{m}\exp \left\{-\frac{(\lambda_k - \lambda_i)^{2}}{2\sigma_{\rm LSF}^{2}}\right\},\\

1275: G &=& \displaystyle\sum_{j=1}^{m}P_{j}^{2}.

1276:   & & 

1277: \end{array}

1278: \end{equation}

1279: Again, the terms $P_j$ and $Q_j$ are evaluated at the center pixel,

1280: $i$, of the feature, $j=j_o$.  Note that $Q_j$ vanishes at the line

1281: center.

1282: 

1283: \subsubsection{Uncertainty in FWHM}

1284: 

1285:  In the case where the $\hbox{\small FWHM}$ has a known uncertainty,

1286:  $\delta\hbox{\small FWHM}$, the affect on our equivalent width limits

1287:  can be calculated.  Again, applying the chain rule, the partial

1288:  derivative of the equivalent width limit with respect to FWHM is

1289: \begin{equation}\label{EQ:FWHMerror}

1290: \frac{\partial \sigma_{EW}}{\partial \hbox{\small FWHM}} = 

1291: \frac{\partial \sigma_{EW}}{\partial P_{j}}\,

1292: \frac{\partial P_{j}}{\partial \sigma_{\rm LSF}}\,

1293: \frac{\partial \sigma_{\rm LSF}}{\partial \hbox{\small FWHM}},

1294: \end{equation}

1295: where $\partial \sigma_{EW}/\partial P_{j}$ and $\partial

1296: P_{j}/\partial\sigma_{\rm LSF}$ are given by Eqs.~\ref{EQ:parPj} and

1297: \ref{EQ:parsigma}.  From Eq.~\ref{EQ:sigma},

1298: \begin{equation}

1299: \frac{\partial\sigma_{\rm LSF}}{\partial \hbox{\small FWHM}} =

1300: \left(\frac{1+z}{2.35}\right)^{2}\frac{\hbox{\small FWHM}}{\sigma_{\rm LSF}}.

1301: \end{equation}

1302: 

1303: \subsubsection{Uncertainty in Continuum Placement}

1304: 

1305: We adopt the method of \citet{semb91} to calculate the uncertainty in

1306: the continuum, $\delta I_{c}(\lambda_k)$, which is obtained directly

1307: from the data by taking the rms of the residuals about the continuum,

1308: $\sigma_c$, from $j=1$ to $m$ and multiplying by 0.5. Numerical

1309: simulations suggest that this provides a conservative estimate of the

1310: errors associated with the continuum \citep{semb91}.  The resulting

1311: uncertainty is

1312: \begin{equation}

1313: \delta I_{c}(\lambda_{k}) = 0.5\, \sigma_{c}(\lambda_{k}),

1314: \end{equation}

1315: The partial derivative of the equivalent width limit with respect to

1316: the continuum is

1317: \begin{equation}\label{EQ:Cerror}

1318: \frac{\partial \sigma_{EW}}{\partial I_{c}(\lambda_{k})} = \frac{1}{2}\frac{D_{i}}{(1+z)}\, G^{-1}\, U^{-1/2}\, X,

1319: \end{equation}

1320: where,

1321: \begin{equation}

1322: \begin{array}{rcl}

1323: \vspace{6pt}

1324: G &=& \displaystyle\sum_{j=1}^{m}P_{j}^{2},\\\vspace{6pt}

1325: U &=& \displaystyle\sum_{j=1}^{m}P_{j}^{2}\, Y^{2}(\lambda_{k}),\\\vspace{6pt}

1326: X &=& \displaystyle\sum_{j=1}^{m}\left\{

1327: \begin{array} {l@{\quad \hbox{for} \quad}l}

1328: -2P_{j}^{2}\, Y^{2}(\lambda_{k})\, I_{c}^{-1}(\lambda_{k}) & \hat{E}(\lambda_{k}) = E(\lambda_{k}),\\[3pt]

1329: |2P_{j}^{2}\, I(\lambda_{k})\, I_{c}^{-2}(\lambda_{k})\left(1-I(\lambda_{k})\, I_{c}^{-1}(\lambda_{k})\right)| 

1330:  & \hat{E}(\lambda_{k}) = |r(\lambda_{k})|.

1331: \end{array}

1332: \right.

1333: \end{array}

1334: \end{equation}

1335: The $j$ elements that are selected for the computation of $X$ depend

1336: on whether one uses the flux error, $E(\lambda_{k})$, or the residual,

1337: $r(\lambda_{k})$, (see Eq.~\ref{EQ:residual}) for a given pixel.  The

1338: continuum contributes a large fraction of the error in the equivalent

1339: width limit measurements in regions where the residual is large, as

1340: can be the case with uncorrected telluric features or problematic sky

1341: subtraction.  In cases such as these we select portions around the

1342: feature where the continuum estimate is reflected more accurately.

1343: 

1344: \subsubsection{Uncertainty in Central Wavelength}

1345: 

1346: If there is a known uncertainty in the wavelength center of the line,

1347: $\delta\lambda_l$, its effect on our equivalent width limits can be

1348: measured directly from the data.  The variance of the equivalent width

1349: limit with respect to the wavelength center of the line is obtained by

1350: calculating the rest equivalent width limits over the range $\Delta

1351: \lambda_i = \pm \delta \lambda_l(1+z)$.  Eq.~\ref{EQ:EWlimit} is used

1352: for the rest equivalent width limit calculations as before.  The

1353: resulting variance is

1354: \begin{equation}\label{EQ:wave_var}

1355: \sigma_{\lambda_{l}}^{2}(\sigma_{EW}) = 

1356: \frac{1}{N_{\sigma_{EW}}-1}

1357: \displaystyle\sum_{n=\lambda_{i}^{-}}^{\lambda_{i}^{+}}\left(\sigma_{EW_{n}}-\left<\sigma_{EW}\right>\right)^{2},

1358: \end{equation}

1359: where $\lambda_{i}^{+}$ and $\lambda_{i}^{-}$ are the upper and lower

1360: wavelengths set by $(1+z)(\lambda_l \pm \delta\lambda_{l})$,

1361: respectively.  $N_{\sigma_{EW}}$ is the total number of equivalent

1362: width limit calculations, and $\left<\sigma_{EW}\right>$ is the mean

1363: equivalent width limit,

1364: \begin{equation}\label{EQ:wave_mean}

1365: \left<\sigma_{EW}\right> = \frac{1}{N_{\sigma_{EW}}} \displaystyle\sum_{n=\lambda_{i}^{-}}^{\lambda_{i}^{+}}\sigma_{EW_{n}}.

1366: \end{equation}

1367: This technique more accurately reflects the errors due to

1368: uncertainties in wave center because it takes the actual data into

1369: account.  If there is a problematic sky subtraction blended with the

1370: expected position of the feature the calculation will reflect this

1371: with a noticeably higher uncertainty.  Also, if there is an

1372: uncertainty in the redshift, $z$, this can be incorporated into the

1373: calculation.

1374: 

1375: 

1376: 

1377: \begin{thebibliography}{}

1378: 

1379: \bibitem[Bada \& Lazcano(2002)]{bada02} Bada, J. L. \& Lazcano, A. 2002, Science, 296, 1982

1380: \bibitem[Bohlin et al.(1978)]{bohl78} Bohlin, R. C., Savage, B. D., \& Drake, J. F. 1978, \apj, 224, 132

1381: \bibitem[Boiss\'{e} et al.(1998)]{bois98} Boiss\'{e}, P., Le Brun, V., Bergeron, J., \& Deharveng, J. M. 1998, \aap, 333, 841

1382: \bibitem[Bouchet et al.(1985)]{bouc85} Bouchet, P., Lequeux, J., Maurice, E., Pr\'{e}vot, L., \& Pr\'{e}vot-Burnichon, M. L. 1985, \aap, 149, 330

1383: \bibitem[Burbidge et al.(1996)]{burb96} Burbidge, E. M., Beaver, E. A., Cohen, R. D., Junkkarinen, V. T., \& Lyons, R. W. 1996, \aj, 112, 2533

1384: \bibitem[Chen \& Lanzetta(2003)]{chen03} Chen, H.-W. \& Lanzetta, K. M. 2003, \apj, 597, 706

1385: \bibitem[Chengalur \& Kanekar(1999)]{cheng99} Chengalur, J. N., \& Kanekar, N., 1999, \mnras, 302, L29

1386: \bibitem[Cordiner et al.(2008)]{cord08} Cordiner, M. A., Cox, N. L. J., Trundle, C., Evans, C. J., Hunter, I., Przybilla, N., Bresolin, F., \& Salama, F. 2008, \aap, 480, L13

1387: \bibitem[Cox et al.(2005)]{cox05} Cox, N. L. J., Kaper, L., Foing, B. H., \& Ehrenfreund, P. 2005, \aap, 438, 187

1388: \bibitem[Cox \& Spaans(2006a)]{cox06a} Cox, N. L. J., \& Spaans, M. 2006a, \aap, 451, 973

1389: \bibitem[Cox et al.(2006b)]{cox06b} Cox, N. L. J., Cordiner, M. A., Cami, J., Foing, B. H., Sarre, P. J., Kaper, L., \& Ehrenfreund, P. 2006b, \aap, 447, 991

1390: \bibitem[Cox et al.(2007)]{cox07} Cox, N. L. J., Cordiner, M. A., Ehrenfreund, P., Kaper, L., Sarre, P. J., Foing, B. H., Spaans, M., Cami, J., Sofia, U. J., Clayton, G. C., Gordon, K. D., \& Salama, F. 2007, \aap, 470, 941

1391: \bibitem[Dekker et al.(2000)]{dekk00} Dekker, H., D'Odorico, S., Kaufer, A. Delabre, B. \& Kotzlowski H. 2000, SPIE, 4008, 534

1392: \bibitem[Ellison et al.(2005)]{elli05} Ellison, S. L., Hall, P. B., \& Lira, P. 2005, \aj, 130, 1345

1393: \bibitem[Ellison et al.(2008)]{elli07} Ellison, S. L., York, B. A., Murphy, M. T., Zych, B. J., Smith, A. M., \& Sarre, P. J. 2008, \mnras, 383, L30

1394: \bibitem[Galazutdinov et al.(2000)]{gala00} Galazutdinov, G. A., Musaev, F. A., Krelowski, J. \& Walker, G. A. H. 2000, \pasp, 112, 648

1395: \bibitem[Gordon et al.(2003)]{gord03} Gordon, K. D., Clayton, G. C., Misselt, K. A., Landolt, A. U., \& Wolff, M. J. 2003, \apj, 594, 279

1396: \bibitem[Heckman \& Lehnert(2000)]{heck00} Heckman, T. M. \& Lehnert, M. D. 2000, \apj, 537, 690

1397: \bibitem[Heger(1922)]{hege22} Heger, M. L. 1922, Lick Observatory Bull. 10, 337, 146

1398: \bibitem[Herbig(1993)]{herb93} Herbig, G. H. 1993, \apj, 407, 142

1399: \bibitem[Herbig(1995)]{herb95} Herbig, G. H. 1995, \araa, 33, 19

1400: \bibitem[Hobbs et al.(2007)]{hobbs07} Hobbs, L. M., York, D. G., Snow, T. P., Oka, T., Thorburn, J. A., Bishof, M., Friedman, S. D., McCall, B. J., Rachford, B., Sonnentrucker, P., \& Welty, D. E. 2007, \apj, in preparation

1401: \bibitem[Hudgins et al.(2005)]{hudg05} Hudgins, D. M., Bauschlicher, C. W., Jr., \& Allamandola, L. J. 2005, \apj, 632, 316

1402: \bibitem[Jenniskens \& Desert(1994)]{jenn94} Jenniskens, P., \& Desert, F. -X. 1994, \aap, 106, 39

1403: \bibitem[Junkkarinen et al.(2004)]{junk04} Junkkarinen, V. T., Cohen, R. D., Beaver, E. A., Burbidge, E. M., \& Lyons, R. W. 2004, \apj, 614, 658

1404: \bibitem[Kanekar \& Chengalur(2001)]{kane01} Kanekar, N., \& Chengalur, J. N. 2001, \aap, 369, 42

1405: \bibitem[Khare et al.(2004)]{khare04} Khare P., Kulkarni, V. P., Lauroesch, J. T., York, D. G., Crotts, A. P. S. \& Nakamura, O. 2004, \apj, 616, 86

1406: \bibitem[Kulkarni et al.(2005)]{kulk05} Kulkarni, V. P., Fall, S. M., Lauroesch, J. T., York, D. G., Welty, D. E., Khare, P., \& Truran, J. W. 2005, \apj, 618, 68

1407: \bibitem[Lanzetta et al.(1987)]{lanz87} Lanzetta, K. M., Turnshek, D. A., \& Wolfe, A. M. 1987, \apj, 322, 739

1408: \bibitem[Le Brun et al.(1997)]{lebr97} Le Brun, V., Bergeron, J., Boiss\'{e}, P., \& Deharveng, J. M. 1997, \aap, 321, 733

1409: \bibitem[Lodders(2003)]{lodd03} Lodders, K. 2003, \apj, 591, 1220

1410: \bibitem[Megier et al.(2005)]{megi05} Megier, A., Krelowski, J., \& Weselak, T. 2005, \mnras, 358, 563

1411: \bibitem[M\'{e}nard et al.(2008)]{mena07} M\'{e}nard, B., Nestor, D., Turnshek, D., Quider, A., Richards, G., Chelouche, D., \& Rao, S. 2008, \mnras, 385, 1053

1412: \bibitem[Merrill \& Wilson(1938)]{merr38} Merrill, P. W. \& Wilson, O. C. 1938, \apj, 87, 9

1413: \bibitem[Moutou et al.(1999)]{mout99} Moutou, C., Krelowski, J., d'Hendecourt, L., \& Jamroszczak, J. 1999, \aap, 351, 680

1414: \bibitem[Murphy \& Liske(2004)]{murp04} Murphy, M. T. \& Liske, J. 2004, \mnras, 354, L31

1415: \bibitem[Murphy(2006)]{popler} Murphy, M. T. 2006, {\sc uves popler}, {\it http://www.ast.cam.ac.uk/$\sim$mim/UVES\_popler.html}

1416: \bibitem[Rieke et al.(1976)]{riek76}Rieke, G. H., Grasdalen, G. L., Kinman, T. D., Hintzen, P., Will, B. J., \& Wills, D. 1976, \nat, 260, 764

1417: \bibitem[Rao \& Turnshek(1998)]{raot98} Rao, S. M. \& Turnshek, D. A. 1998, \apj, 500, L115

1418: \bibitem[Rao \& Turnshek(2000)]{raot00} Rao, S. M. \& Turnshek, D. A. 2000, \apjs, 130, 1

1419: \bibitem[Rao et al.(2003)]{raot03} Rao, S. M., Nestor, D. B., Turnshek, D. A., Lane, W. M., Monier, E. M., \& Bergeron, J. 2003, \apj, 595, 94

1420: \bibitem[Rich(1987)]{rich87} Rich, R. M. 1987, \aj, 94, 651

1421: \bibitem[Savage \& Sembach(1991)]{sava91} Savage, B. D. \& Sembach, K. R. 1991, \apj, 379, 245

1422: \bibitem[Schneider et al.(1993)]{schn93} Schneider, D. P., Hartig, G. F., Jannuzi, B. T., et al. 1993, \apjs, 87, 45

1423: \bibitem[Sembach et al.(1991)]{semb91} Sembach, K. R., Savage, B. D., \& Massa, D. 1991, \apj, 372, 81

1424: \bibitem[Snow(2001)]{snow01} Snow, T. P. 2001, Spectrochimica Acta Part A, 57, 615

1425: \bibitem[Sollerman et al.(2005)]{soll05} Sollerman, J., Cox, N., Mattila, S., Ehrenfreund, P., Kaper, L., Leibundgut, B., \& Lundqvist, P. 2005, \aap, 429, 559

1426: \bibitem[Steidel et al.(1994)]{stei94} Steidel, C. C., Pettini, M., Dickinson, M., \& Persson, S. E. 1994, \aj, 108, 2046

1427: \bibitem[Thorburn et al.(2003)]{thor03} Thorburn, J. A., Hobbs, L. M., McCall, B. J., Oka, T., Welty, D. E., Friedman, S. D., Snow, T. P., Sonnentrucker, P., \& York, D. G. 2003, \apj, 584, 339

1428: \bibitem[Tuairisg et al.(2000)]{tuai00} Tuairisg, S. O., Cami, J., Foing, B. H., Sonnentrucker, P., \& Ehrenfreund, P. 2000, \aap, 142, 225

1429: \bibitem[Turnshek et al.(2001)]{turn01} Turnshek, D. A., Rao, S. M., Nestor, D., Lane, W., Monier, E., Bergeron, J., \& Smette, A. 2001, \apj, 553, 288

1430: \bibitem[Turnshek et al.(2003)]{turn03} Turnshek, D. A., Rao, S. M., Ptak, A. F., Griffiths, R. E., \& Monier, E. M. 2003, \apj, 590, 730

1431: \bibitem[Welty et al.(2006)]{welt06} Welty, D. E., Federman, S. R., Gredel, R., Thorburn, J. A., \& Lambert, D. L. 2006, \apjs, 165, 138

1432: \bibitem[Weselak et al.(2000)]{wese00} Weselak, T., Schmidt, M., \& Krelowski, J. 2000, \aap, 142, 239

1433: \bibitem[Weselak et al.(2001)]{wese01} Weselak, T., Fulara, J., Schmidt, M. R., \& Krelowski, J. 2001, \aap, 377, 677

1434: \bibitem[Wild \& Hewett(2005)]{wild05} Wild, V. \& Hewett, P. C. 2005, \mnras, 361, L30

1435: \bibitem[Wild et al.(2006)]{wild06} Wild, V., Hewett, P. C., \& Pettini, M. 2006, \mnras, 367, 211

1436: \bibitem[Wszolek \& Godlowski(2003)]{wszo03} Wszolek, B. \& Godlowski, W. 2003, \mnras, 338, 990

1437: \bibitem[Yanny et al.(1989)]{yann89} Yanny, B., York, D. G., \& Gallagher, J. S. 1989, \apj, 338, 735

1438: \bibitem[Yan et al.(2005)]{yan05} Yan, L., Chary, R., Armus, L., Teplitz, H., Helou, G., Frayer, D., Fadda, D., Surace, J., \& Choi, P. 2005, \apj, 628, 604

1439: \bibitem[York et al.(2006)]{lawt06} York, B. A., Ellison, S. L., Lawton, B., Churchill, C. W., Snow, T. P., Johnson, R. A., \& Ryan, S. G. 2006, \apj, 647, L29

1440: 

1441: 

1442: \end{thebibliography}

1443: 

1444: 

1445: \clearpage

1446: 

1447: \begin{deluxetable}{crlcclll}

1448: \tablewidth{0pt}

1449: \tabletypesize{\scriptsize}

1450: \tablecaption{DLAs in Sample\label{tab:DLAs}}

1451: \tablehead{

1452: \colhead {} & \colhead{}  &     \colhead{}                      & \colhead{}          &

1453: \colhead{$N$(\HI)/10$^{20}$}    &

1454: \colhead{}       & \colhead{}  &  \colhead{}       \\

1455: \colhead{DLA} & \colhead{}   & \colhead{QSO$\phantom{<<<}$}                   & \colhead{$z_{abs}$} &

1456: \colhead{[atoms cm$^{-2}$]}     &         

1457: \colhead{[Zn/H]} & \colhead{[Fe/H]} &  \colhead{Refs.}}

1458: \tablecolumns{8}

1459: \startdata

1460: 1 & AO  & 0235+164$^{a}$  & 0.524 & \phantom{..}50$\pm$10 &  \phantom{......}\nodata & \phantom{.......}\nodata & 1,2 \\

1461:   & & & & & & & \\

1462: 2 & Q   & 0738+313  & 0.091 & 15$\pm$2  & $<-1.14$ & \phantom{...<}$-1.63^{+0.13}_{-0.18}$ & 3,4,8,9   \\

1463:   & & & & & & & \\

1464: 3 &   &           & 0.221 & $\phantom{1}$7.9$\pm$1.4 &  $<-0.70^{+0.14}_{-0.17}$ & \phantom{.......}\nodata & 3,4,8,9   \\

1465:   & & & & & & & \\

1466: 4 & B2  & 0827+243  & 0.518 & $\phantom{1}$2.0$\pm$0.2 & $<+0.30$ & \phantom{...<}$-1.02\pm0.05$ & 4,5,8,9   \\

1467:   & & & & & & & \\

1468: 5 & PKS & 0952+179  & 0.239 & 21$\pm$3  & $<-1.02$ & \phantom{.......}\nodata & 4,5,9   \\

1469:   & & & & & & & \\

1470: 6 & PKS & 1127--145 & 0.313 & 51$\pm$9  &  \phantom{......}\nodata & $>-2$ & 1,6,9   \\

1471:   & & & & & & & \\

1472: 7 & Q   & 1229--020 & 0.395 & $\phantom{1}$5.6$\pm$1.0 &  \phantom{...<}$-0.47$ & $<-1.32$ & 7   \\

1473: \enddata

1474: \tablecomments{All upper limits are calculated to 3~$\sigma$.}

1475: \tablenotetext{\mbox{a}}{Metallicity for AO~0235+164 is estimated via X-ray spectroscopy to be $Z=0.24\pm0.06Z_{\odot}$ \citep[\textit{Chandra},][]{turn03} and $Z=0.72\pm0.28Z_{\odot}$ \citep[\textit{ASCA \& ROSAT},][]{junk04}.}

1476: \tablerefs{

1477: 1. Turnshek et al.\ (2003); 2. Junkkarinen et al.\ (2004); 3. Rao \& Turnshek (1998);

1478:  4. Kulkarni et al.\ (2005); 5. Rao \& Turnshek (2000); 6. This work; 

1479:  7. Boiss\'{e} et al.\ (1998);  8. Khare et al.\ (2004); 9. Kanekar \& Chengalur (2003)}

1480: \end{deluxetable}

1481: 

1482: \clearpage

1483: 

1484: \begin{deluxetable}{rlccclrllrc}

1485: \tablewidth{0pt}

1486: \tabletypesize{\scriptsize}

1487: \rotate

1488: \tablecaption{Journal of Observations\label{tab:obs}}

1489: \tablehead{

1490: \colhead{}                 & \colhead{}            &

1491: \colhead{}                 & \colhead{Grating/}    &

1492: \colhead{Slit}             & \colhead{}            &

1493: \colhead{Exposure}         & \colhead{S/N}         &  

1494: \colhead{Wavelength}       & \colhead{}            &

1495: \colhead{Dispersion} \\

1496: \colhead{}                 & \colhead{QSO$\phantom{<<<}$}         &

1497: \colhead{Facility}         & \colhead{Grism}       &

1498: \colhead{Width}            & \colhead{Date [UT]$\phantom{<<<}$}   &

1499: \colhead{Time [s]}         & \colhead{[pixel$^{-1}$]}             &

1500: \colhead{Coverage [\AA]}   & \colhead{Resolution}                 &

1501: \colhead{[\AA~pixel$^{-1}$]}}

1502: \tablecolumns{11}

1503: \startdata

1504: AO  & 0235+164  & VLT/FORS2         & GRIS 600z     & 1.0''      & 2005 Jul 20       & 8400$\phantom{<}$ & 60--150 & 7318--10,744 & 1880 & 1.59 \\   

1505:     &      &                   &               &            & 2005 Sep 6        &            &           &              & & \\

1506:     &      &                   &               &            & 2005 Oct 1        &        &               &              & & \\

1507: Q   & 0738+313  & APO/DIS           & HIGH          & 1.5''      & 2004 Dec 15/19    & 40,600$\phantom{<}$ & 57--84 & 4367--7817 & 2040 & 0.62/0.84 \\

1508:     &      &                   &               &            & 2005 Feb 4        &        &                 &                  &  & (Blue/Red) \\

1509: B2  & 0827+243  & Keck/HIRES  &               & C2/0.861'' & 1998 Dec 22       & 22,500$\phantom{<}$ & 27--114$^{a}$ & 5185--9234 & 43,000 & 0.04  \\

1510: PKS & 0952+179  & VLT/FORS2         & GRIS 600RI    & 1.0''      & 2005 Apr 4/5/7/8  & 9000$\phantom{<}$        & 57--95 & 5298--8622   & 1650 & 1.63 \\

1511: PKS & 0952+179  & WHT/ISIS          & R600R         & 1.0''      & 2004 May 20       & 4500$\phantom{<}$        & 20--22 & 6312--8114   & 3790 & 0.45 \\

1512: PKS & 1127--145 & VLT/UVES    & 346,564       & 1.0''      & 2002 Jul 17/18    & 24,900$\phantom{<}$            & 38$^{b}$ & 3041--6809 & 45,000 & 0.05  \\

1513: PKS & 1127--145 & Gemini/GMOS-S     & R400          & 1.0''      & 2004 Jun 19       & 3600$\phantom{<}$        & 38--84   & 5962--9998 & 960 & 2.75 \\ 

1514: Q   & 1229--020 & VLT/FORS2         & GRIS 600z     & 1.0''      & 2005 Apr 9/13     & 9800$\phantom{<}$        & 50--62   & 7464--10,000 & 1860 & 1.59 \\      

1515: Q   & 1229--020 & WHT/ISIS          & R600R         & 1.0''      & 2004 May 20       & 3600$\phantom{<}$        & 5--13   & 7212--9018    & 4980 & 0.45     

1516: \enddata

1517: \tablenotetext{\mbox{a}}{The low S/N is measured in the red, and the high S/N is measured in the blue.}

1518: \tablenotetext{\mbox{b}}{The S/N is measured at the location of the $\lambda4428$ DIB.  The other DIBs are not covered in this spectrum.}

1519: \end{deluxetable}

1520: 

1521: \clearpage

1522: 

1523: 

1524: \begin{deluxetable}{crlclllllll}

1525: \tabletypesize{\scriptsize}

1526: \tablecolumns{11}

1527: \tablewidth{0pt}

1528: \rotate

1529: \tablecaption{Equivalent Width Detections \& Limits of DIBs in DLAs\label{tab:EWs}}

1530: \tablehead{

1531: \colhead{}                           &

1532: \colhead{}                           &

1533: \colhead{}                           &

1534: \colhead{}                           &

1535: \colhead{}                           &

1536: \multicolumn{6}{c}{Rest--Frame EW Detections \& Rest--Frame 3~$\sigma$ EW Limits [m\AA]}\\

1537: \colhead{DLA}                        &

1538: \colhead{}                           &

1539: \colhead{QSO$\phantom{<<<}$}         & 

1540: \colhead{$z_{abs}$}                  &

1541: \colhead{Facility}                   &

1542: \colhead{$\lambda$4428}         &

1543: \colhead{$\lambda$5705}         &

1544: \colhead{$\lambda$5780}         &

1545: \colhead{$\lambda$5797}         &

1546: \colhead{$\lambda$6284}         & 

1547: \colhead{$\lambda$6613}}

1548: \startdata

1549: 1 & AO  & 0235+164 & 0.524 & VLT/FORS2 & $\phantom{<}$741.5$\pm$26.2$^{a}$ & 63.2$\pm$8.7$^{b}$ & $\phantom{<}$216$\pm9^{b}$ & $<$118$^{bc}$ & $<$128$^{bc}$ & $<$95$^{bc}$ \\

1550: 2 & Q   & 0738+313  & 0.091 & APO/DIS  & $<$115(8) & \nodata  & $<$88(6)$^{c}$ & $<$74(3) & $<$106(3)  & $<$414(51) \\

1551: 3 &     &           & 0.221 & APO/DIS  & $<$151(11) & \nodata  & $<$85(3)  & $<$78(3) & $<$240(16)$^{c}$ & \nodata          \\

1552: 4 & B2  & 0827+243  & 0.518 & Keck/HIRES & $<$94(14)$^{a}$ & \nodata  & $<$32(1) & $<$21(1) & \nodata & \nodata \\

1553: 5 & PKS & 0952+179  & 0.239 & VLT/FORS2 & $<$259(19) & \nodata &  $<$90(4)  & $<$289(14)$^{c}$ & $<$102(4) & $<$108(6)    \\

1554:   &     &           & 0.239 & WHT/ISIS & \nodata    & \nodata &  $<$158(4) & $<$237(18) & $<$187(3) & \nodata                   \\

1555: 6 & PKS & 1127--145 & 0.313 & VLT/UVES & $<$68(5)$^{d}$  & \nodata & \nodata & \nodata  & \nodata  & \nodata             \\

1556:   &     &           & 0.313  & Gemini/GMOS       & \nodata  & \nodata & $<$6106(77)$^{cd}$ & $<$3162(116)$^{cd}$ & $<$341(1)$^{d}$ & $<$380(1)$^{d}$\\

1557: 7 & Q   & 1229--020 & 0.395 & VLT/FORS2 & \nodata    & \nodata & $<$129(5)  & $<$115(6) & $<$162(5) & $<$135(7)  \\

1558:   &     &           & 0.395     & WHT/ISIS & \nodata    & \nodata & $<$292(10) & $<$198(8) & $<$627(14)$^{c}$ & \nodata  \\

1559: 

1560: \enddata

1561: \tablenotetext{\mbox{a}}{Spectrum provided by Junkkarinen et al.\ (2004).}

1562: \tablenotetext{\mbox{b}}{Spectrum provided by York et al.\ (2006).}

1563: \tablenotetext{\mbox{c}}{Problematic sky subtraction or large atmospheric band.}

1564: \tablenotetext{\mbox{d}}{Uncertainty in resolution is unknown and not included in error assessment.}

1565: \end{deluxetable}

1566: 

1567: 

1568: \clearpage

1569: 

1570: 

1571: \begin{deluxetable}{crlllllll}

1572: \tabletypesize{\scriptsize}

1573: \tablecolumns{9}

1574: \tablewidth{0pt}

1575: %\rotate

1576: \tablecaption{Model Predictions of DIBs in DLAs\label{tab:Models}}

1577: \tablehead{

1578: \colhead{}                           &

1579: \colhead{}                           &

1580: \colhead{QSO$\phantom{<<<}$}         &

1581: \colhead{Observed/}                    &

1582: \multicolumn{5}{c}{Rest--Frame W Detections, Limits (3~$\sigma$), \& Model Predictions [m\AA]}\\

1583: \colhead{DLA}                        &

1584: \colhead{}                           &

1585: \colhead{$z_{\rm DLA}\phantom{<<<}$}     & 

1586: \colhead{Models$^{c}$}               &

1587: \colhead{$\lambda$4428}         &

1588: \colhead{$\lambda$5780}         & 

1589: \colhead{$\lambda$5797}         &

1590: \colhead{$\lambda$6284}         & 

1591: \colhead{$\lambda$6613}}

1592: \startdata

1593: 1 & AO  & 0235+164  &  Observed & $\phantom{<}$741.5$\pm26.2^{a}$ & $\phantom{1<}$216$\pm9^{b}$   & $\phantom{.}$$<$118$^{b}$ & $\phantom{<}$$<$128$^{b}$ & $\phantom{<}$$<\phantom{.}$95$^{b}$ \\

1594:   &     & $z=0.524$ &  EW[$N$(HI)]     &  $\phantom{<}$\nodata & $\phantom{11}$1059    & $\phantom{1..}$317     & $\phantom{11}$1572 & $\phantom{<}$\nodata\\

1595:   &     &      &  EW[$E(B-V)=0.23$]    &  $\phantom{<}$482     & $\phantom{1<}$104     & $\phantom{<>}$43       & $\phantom{1<}$368  & $\phantom{1111}$50     \\

1596:   &     &      &  $E(B-V)_{\rm lim}$ &  $\phantom{1111}$0.35 & $\phantom{11111}$0.48 & $\phantom{111..}$0.64  & $\phantom{11111}$0.06 & $\phantom{11111}$0.44 \\

1597:   &     &      &                 &                       &                       &                        &            &             \\

1598: 2 & Q   & 0738+313  &  Observed         &  $<$115(8)            & $\phantom{<}$$<$88(6) & $\phantom{.1}$$<$74(3) & $\phantom{<}$$<$106(3) & $\phantom{<}$$<$414(51)  \\

1599:   &     & $z=0.091$ &  EW[$N$(HI)]     &  $\phantom{<}$\nodata & $\phantom{1..}$247    & $\phantom{<>}$89       & $\phantom{1<}$532 & $\phantom{<}$\nodata  \\

1600:   &     &      &  EW[$E(B-V)=0.04$]    &  $\phantom{<1}$84     & $\phantom{<>}$18      & $\phantom{1111}$8      & $\phantom{1111}$91  & $\phantom{11111}$3     \\

1601:   &     &      &  $E(B-V)_{\rm lim}$ &  $\phantom{1111}$0.05 & $\phantom{1111}$0.20  & $\phantom{111..}$0.40  & $\phantom{11111}$0.05 & $\phantom{11111}$1.91  \\

1602:   &     &      &                 &                       &                       &                        &            &             \\

1603: 3 & Q   & 0738+313  &  Observed         & $<$151(11)            & $\phantom{<}$$<$85(3) & $\phantom{.1}$$<$78(3) & $\phantom{<}$$<$240(16) & $\phantom{<}$\nodata     \\

1604:   &     & $z=0.221$ &  EW[$N$(HI)]     &  $\phantom{<}$\nodata & $\phantom{1..}$114    & $\phantom{<>}$46       & $\phantom{1<}$299 & $\phantom{<}$\nodata   \\

1605:   &     &      &  EW[$E(B-V)=0.04$]    &  $\phantom{<1}$84     & $\phantom{<>}$18      & $\phantom{1111}$8      & $\phantom{1111}$91 & $\phantom{<}$\nodata  \\

1606:   &     &      &  $E(B-V)_{\rm lim}$ &  $\phantom{1111}$0.07 & $\phantom{1111}$0.19  & $\phantom{111..}$0.42  & $\phantom{11111}$0.14 & $\phantom{<}$\nodata\\

1607:   &     &      &                 &                       &                       &                        &            &             \\

1608: 4 & B2  & 0827+243  &  Observed         & $\phantom{<}$$<$94(14)& $\phantom{<}$$<$32(1) & $\phantom{.1}$$<$21(1) & $\phantom{<}$\nodata    & $\phantom{<}$\nodata     \\

1609:   &     & $z=0.518$ &  EW[$N$(HI)]     &  $\phantom{>}$\nodata & $\phantom{<>}$22      & $\phantom{<>}$11       & $\phantom{<}$\nodata    & $\phantom{<}$\nodata     \\

1610:   &     &      &  EW[$E(B-V)=0.04$]    &  $\phantom{<>}$84     & $\phantom{<>}$18      & $\phantom{1111}$8      & $\phantom{<}$\nodata    & $\phantom{<}$\nodata     \\

1611:   &     &      &  $E(B-V)_{\rm lim}$ &  $\phantom{1111}$0.04 & $\phantom{1111}$0.07  & $\phantom{111..}$0.11  & $\phantom{<}$\nodata    & $\phantom{<}$\nodata     \\

1612:   &     &      &                 &                       &                       &                        &            &             \\

1613: 5 & PKS & 0952+179  &  Observed         & $<$259(19)            & $\phantom{<}$$<$90(4) & $\phantom{.}$$<$237(18)& $\phantom{<}$$<$102(4)  & $\phantom{<}$$<$108(6)   \\

1614:   &     & $z=0.239$ &  EW[$N$(HI)]     &  $\phantom{<}$\nodata & $\phantom{1...}$370   & $\phantom{1..}$127     & $\phantom{1<}$720 & $\phantom{<}$\nodata   \\

1615:   &     &      &  EW[$E(B-V)=0.04$]    &  $\phantom{<1}$84     & $\phantom{<>}$18      & $\phantom{1111}$8      & $\phantom{1111}$91  & $\phantom{11111}$3     \\

1616:   &     &      &  $E(B-V)_{\rm lim}$ &  $\phantom{1111}$0.12 & $\phantom{1111}$0.20  & $\phantom{<>.}$1.29    & $\phantom{11111}$0.05 & $\phantom{11111}$0.50  \\

1617:   &     &      &                 &                       &                       &                        &            &             \\

1618: 6 & PKS & 1127--145 &  Observed         & $\phantom{<}$$<$68(5) & $<$6106(77)           & $<$3162(116)           & $\phantom{<}$$<$341(1)  & $\phantom{<}$$<$380(1)   \\

1619:   &     & $z=0.313$ &  EW[$N$(HI)]     & $\phantom{<}$\nodata  & $\phantom{<}$1086     & $\phantom{1..}$324     & $\phantom{11}$1601 & $\phantom{<}$\nodata  \\

1620:   &     &      &  EW[$E(B-V)=0.04$]    &  $\phantom{<>}$84     & $\phantom{<>}$18      & $\phantom{1111}$8      & $\phantom{1111}$91 & $\phantom{11111}$3           \\

1621:   &     &      &  $E(B-V)_{\rm lim}$ &  $\phantom{1111}$0.03 & $\phantom{<>}$14.13   & $\phantom{<>}$17.73    & $\phantom{11111}$0.21 & $\phantom{11111}$1.75     \\

1622:   &     &      &                 &                       &                       &                        &            &             \\

1623: 7 & Q   & 1229--020 &  Observed         & $\phantom{<}$\nodata  & $\phantom{.}$$<$129(5)& $\phantom{.}$$<$115(6) & $\phantom{<}$$<$162(5)  & $\phantom{<}$$<$135(7)   \\

1624:   &     & $z=0.395$ &  EW[$N$(HI)]     & $\phantom{<}$\nodata  & $\phantom{<>}$75      & $\phantom{<>}$32       & $\phantom{1<}$219        & $\phantom{<}$\nodata     \\

1625:   &     &      &  EW[$E(B-V)=0.04$]    & $\phantom{<}$\nodata  & $\phantom{<>}$18      & $\phantom{1111}$8      & $\phantom{1111}$91 & $\phantom{11111}$3 \\

1626:   &     &      &  $E(B-V)_{\rm lim}$ & $\phantom{<}$\nodata  & $\phantom{111..}$0.29 & $\phantom{111..}$0.62  & $\phantom{11111}$0.08 & $\phantom{11111}$0.62  \\

1627: \enddata

1628: \tablenotetext{\mbox{a}}{Spectrum provided by Junkkarinen et al.\ (2004).}

1629: \tablenotetext{\mbox{b}}{Spectrum provided by York et al.\ (2006).}

1630: 

1631: \tablenotetext{\mbox{c}}{The limits are those that are the best

1632: constrained from Table~\ref{tab:EWs}.  The model EW[$N$(HI)] refers to

1633: the predicted equivalent widths (m\AA) from the Galactic best--fit

1634: lines in Figure~\ref{p:NHI_known}.  The values used for the $N$(HI)

1635: are taken from Table~\ref{tab:DLAs}.  The model EW[$E(B-V)=0.23$]

1636: refers to the predicted equivalent widths (m\AA) from the Galactic

1637: DIB--$E(B-V)$ best--fit lines with the known $E(B-V)=0.23$

1638: \citep{junk04}.  The model EW[$E(B-V)=0.04$] refers to the predicted

1639: equivalent widths (m\AA) from the Galactic DIB--$E(B-V)$ best--fit

1640: lines with the upper limit of $E(B-V)=0.04$ as measured in high-z DLAs

1641: by \citet{elli05}.  For all systems, the $E(B-V)=0.04$ except for

1642: AO~0235+164 which has a measured $E(B-V)=0.23$ \citep{junk04}.  The

1643: value $E(B-V)_{\rm lim}$ is the $E(B-V)$ upper limit inferred from our

1644: equivalent width limits and the Galactic DIB--$E(B-V)$ best-fit lines

1645: for the $\lambda$5780, $\lambda$5797, and $\lambda$6284 DIBs

1646: \citep{welt06} and the $\lambda$4428 and $\lambda$6613 DIBs

1647: (unpublished, T.P. Snow, private communication).}

1648: \end{deluxetable}

1649: 

1650: 

1651: \clearpage

1652: 

1653: 

1654: \begin{deluxetable}{crlccrc}

1655: \tabletypesize{\scriptsize}

1656: \tablecolumns{8}

1657: \tablewidth{0pt}

1658: \tablecaption{Adopted Reddening and Gas-to-Dust Ratios\label{tab:adoptedreddening}}

1659: \tablehead{

1660: \colhead{DLA}                        &

1661: \colhead{}                           &

1662: \colhead{QSO$\phantom{<<<}$}         & 

1663: \colhead{$z_{abs}$}                  &

1664: \colhead{$E(B-V)$}             &

1665: \colhead{gas/dust}                   &

1666: \colhead{Contraint}                        \\

1667: \colhead{}                        &

1668: \colhead{}                           &

1669: \colhead{}         & 

1670: \colhead{}                  &

1671: \colhead{[mag]}             &

1672: \colhead{[{\cmsq}~mag$^{-1}$]}                   &

1673: \colhead{DIB}                        

1674: }                       

1675: \startdata

1676: 1 & AO  & 0235+164  & 0.524\tablenotemark{\mbox{a}} &  \phantom{$<$}$0.23$  & 19.2$ \times10^{21}$ & \nodata \\

1677: 2 & Q   & 0738+313  & 0.091 &  $<0.05$  & $>$ 30\phantom{.0}$ \times10^{21}$ & $\lambda$6284 \\

1678: 3 &     &           & 0.221 &  $<0.14$  & $>$ \phantom{1}5.6$ \times10^{21}$ & $\lambda$6284 \\

1679: 4 & B2  & 0827+243  & 0.518 &  $<0.07$  & $>$ \phantom{1}2.9$ \times10^{21}$ & $\lambda$5780 \\

1680: 5 & PKS & 0952+179  & 0.239 &  $<0.05$  & $>42$\phantom{.0}$ \times10^{21}$ & $\lambda$6284 \\

1681: 6 & PKS & 1127--145 & 0.313 &  $<0.21$  & $>24$\phantom{.0}$ \times10^{21}$ & $\lambda$6284 \\

1682: 7 & Q   & 1229--020 & 0.395 &  $<0.08$  & $>$ \phantom{1}7.0$\times10^{21}$ & $\lambda$6284 \\

1683: \enddata

1684: \tablenotetext{\mbox{a}}{Values quoted from Junkkarinen et al.\ (2004).}

1685: \end{deluxetable}

1686: 

1687: \clearpage

1688: 

1689: \begin{figure}

1690: \includegraphics[angle=-90,width=1.0\textwidth]{f1.eps}

1691: \caption{Equivalent width limit analysis of the Q0738+313 $z=0.091$

1692: DLA. ---($a$) Relative flux in the region of the $\lambda$5780 and

1693: $\lambda$5797 DIBs (vertical long dashed line and vertical long

1694: dashed-dotted line, respectively).  The histogram is the relative flux

1695: data, and the short dashed lines are the associated $\pm3~\sigma$ flux

1696: errors.  ---($b$) Relative sky flux.  ---($c$) The $3~\sigma$ rest

1697: equivalent width limits in m{\AA}.  The two peaks in panel ($c$) are

1698: due to the problematic sky subtraction from the [OI] sky emission line

1699: at $\lambda\sim 6300$~{\AA} and the poor continuum fit at $\lambda\sim

1700: 6314$~{\AA} where the residuals are used, as explained in

1701: Eq.~\ref{EQ:flux_error}.\label{p:kspace}}

1702: \end{figure}

1703: 

1704: \begin{figure}

1705: \includegraphics[angle=-90,width=1.0\textwidth]{f2.eps}

1706: \caption{VLT/FORS2 spectrum of the $z=0.524$ DLA toward AO~0235+164.

1707: Plotted are the expected positions of the ($a$) $\lambda 5780/5797$,

1708: ($b$) $\lambda 6284$, and ($c$) $\lambda 6613$ DIBs.  The upper

1709: sub-panels are the normalized flux of the data and models.  The center

1710: sub-panels are the sigmas of the associated data fluxes normalized by

1711: the continuum.  The lower sub-panels are the sky fluxes normalized by

1712: the continuum.  The smooth curves are model predictions (see text).

1713: The thin curves are the expected DIB profiles given the measured DLA

1714: $N({\HI})$ and the Galactic DIB--$N({\HI})$ relation.  The thick

1715: curves are the expected DIB profiles using the measured $E(B-V)=0.23$

1716: from \citet{junk04}.\label{s:0235}}

1717: \end{figure}

1718: 

1719: \begin{figure}

1720: \includegraphics[angle=-90,width=1.0\textwidth]{f3.eps}

1721: \caption{Same as Fig.~\ref{s:0235}, but for the APO/DIS spectrum of

1722: the $z=0.091$ DLA toward Q0738+313.  Plotted are the expected

1723: positions of the ($a$) $\lambda 4428$, ($b$) $\lambda 5780/5797$,

1724: ($c$) $\lambda 6284$, and ($d$) $\lambda 6613$ DIBs.  The smooth

1725: curves are model predictions (see text).  The thin curves are the

1726: expected DIB profiles given the measured DLA $N({\HI})$ and the

1727: Galactic DIB--$N({\HI})$ relation.  The thick curves for this DLA are

1728: the expected DIB profiles using the $E(B-V)=0.04$ upper limit for

1729: high-z DLAs from \citet{elli05}.\label{s:0738_091}}

1730: \end{figure}

1731: 

1732: \begin{figure}

1733: \includegraphics[angle=-90,width=1.0\textwidth]{f4.eps}

1734: \caption{Same as Fig.~\ref{s:0738_091}, but for the APO/DIS spectrum of

1735: the $z=0.221$ DLA toward Q0738+313.  Plotted are the expected

1736: positions of the ($a$) $\lambda 4428$, ($b$) $\lambda 5780/5797$, and

1737: ($c$) $\lambda 6284$ DIBs.\label{s:0738_221}}

1738: \end{figure}

1739: 

1740: \begin{figure}

1741: \includegraphics[angle=-90,width=1.0\textwidth]{f5.eps}

1742: \caption{Same as Fig.~\ref{s:0738_091}, but for the Keck/HIRES spectrum of

1743: the $z=0.518$ DLA toward B2~0827+243.  Plotted are the expected

1744: positions of the ($a$) $\lambda 4428$ and the ($b$) $\lambda

1745: 5780/5797$ DIBs.\label{s:0827}}

1746: \end{figure}

1747: 

1748: \begin{figure}

1749: \includegraphics[angle=-90,width=1.0\textwidth]{f6.eps}

1750: \caption{Same as Fig.~\ref{s:0738_091}, but for the VLT/FORS2 and WHT/ISIS

1751: spectra of the $z=0.239$ DLA toward PKS~0952+179.  Plotted are the

1752: expected positions of the ($a$) $\lambda 4428$, ($b$) $\lambda

1753: 5780/5797$, ($c$) $\lambda 5780/5797$ (WHT/ISIS), ($d$) $\lambda

1754: 6284$, and ($e$) $\lambda 6613$ DIBs.  The WHT/ISIS spectrum provides

1755: the adopted limit for the $\lambda$5797 DIB; the VLT/FORS2 spectrum

1756: provides the adopted limits for the remaining DIBs.\label{s:0952_VLT}}

1757: \end{figure}

1758: 

1759: \begin{figure}

1760: \includegraphics[angle=-90,width=1.0\textwidth]{f7.eps}

1761: \caption{Same as Fig.~\ref{s:0738_091}, but for the Gemini/GMOS-S spectrum

1762: of the $z=0.313$ DLA toward PKS~1127--145.  Plotted are the expected

1763: positions of the ($a$) $\lambda 5780/5797$, ($b$) $\lambda 6284$, and

1764: ($c$) $\lambda 6613$ DIBs.\label{s:1127_Gem}}

1765: \end{figure}

1766: 

1767: \begin{figure}

1768: \includegraphics[angle=-90,width=1.0\textwidth]{f8.eps}

1769: \caption{Same as Fig.~\ref{s:0738_091}, but for the VLT/FORS2 spectrum of

1770: the $z=0.395$ DLA toward Q1229--020.  Plotted are the expected

1771: positions of the ($a$) $\lambda 5780/5797$, ($b$) $\lambda 6284$, and

1772: ($c$) $\lambda 6613$ DIBs.\label{s:1229_VLT}}

1773: \end{figure}

1774: 

1775: \begin{figure}

1776: \includegraphics[angle=0, width=1.0\textwidth]{f9.eps}

1777: \caption{The DIB equivalent width--$N$(HI) relations \citep{welt06}

1778: with our DLAs added.  ---($a$) $\lambda$5780 DIB.  ---($b$)

1779: $\lambda$5797 DIB.  ---($c$) $\lambda$6284 DIB.  The solid lines are

1780: the best-fit weighted Galactic lines.  The region enclosed by the

1781: dotted lines contains the Galactic data.  The region enclosed by the

1782: dashed lines contains the LMC data.  The region enclosed by the

1783: dot-dash lines contains the SMC data.  Error bars are 1~$\sigma$, and

1784: upper limits are marked with arrows.  The vertical error bars for

1785: AO~0235+164 in panel ($a$) are smaller than the point size and all

1786: values for this DLA are from \citet{lawt06}.}\label{p:NHI_known}

1787: \end{figure}

1788: 

1789: \begin{figure}

1790: \includegraphics[angle=0, width=1.0\textwidth]{f10.eps}

1791: \caption{The logarithm ratio of the measured equivalent width limits

1792: to the equivalent width predicted for multiple values of reddening for

1793: each DLA.  ---($a$) $\lambda$4428 DIB.  ---($b$) $\lambda$5780 DIB.

1794: ---($c$) $\lambda$5797 DIB.  ---($d$) $\lambda$6284 DIB. ---($e$)

1795: $\lambda$6613 DIB.  The equivalent widths of the $\lambda$5780,

1796: $\lambda$5797, and $\lambda$6284 DIBs, for each reddening, were

1797: computed using the best-fit Galactic lines from \citet{welt06}.  The

1798: equivalent widths of the $\lambda$4428 and $\lambda$6613 DIBs, for

1799: each reddening, were computed from the relations in

1800: Eqs.~\ref{EQ:4428_reddening} and \ref{EQ:6613_reddening} (unpublished,

1801: T.P. Snow, private communication).  The DLAs labeled are as follows:

1802: (1) AO~0235+164, (2) Q0738+313, $z=0.091$, (3) Q0738+313, $z=0.221$,

1803: (4) B2~0827+243, (5) PKS~0952+179, (6) PKS~1127--145, (7) Q1229--020.

1804: We leave the panels of the $\lambda$4428 and $\lambda$5780 DIBs for

1805: AO~0235+164 blank because they are detections \citep{junk04,lawt06}

1806: with known reddening, $E(B-V)=0.23$ \citep{junk04}.  The zero point

1807: (horizontal dotted line) on each plot marks the estimated maximum

1808: reddening of the DLA determined from the measured equivalent width

1809: limit, or $E(B-V)_{\rm lim}$ (see Table~\ref{tab:Models}).  The

1810: vertical error bars are the 1~$\sigma$ errors for the $E(B-V)_{\rm

1811: lim}$ based on the equivalent width limit uncertainties and the

1812: uncertainties in the DIB-reddening law slopes in

1813: Eqs.~\ref{EQ:5780_reddening}, \ref{EQ:5797_reddening},

1814: \ref{EQ:6284_reddening}, \ref{EQ:4428_reddening} and

1815: \ref{EQ:6613_reddening}.  The equivalent width limits for the

1816: $\lambda$4428 DIB measured in DLA (5) are unconstraining, thus, we

1817: leave off the error bars in this case.  The 1~$\sigma$ errors in

1818: panels ($b$) and ($c$) are smaller than the point

1819: sizes.\label{p:reddening_law}}

1820: \end{figure}

1821: 

1822: \begin{figure}

1823: \includegraphics[angle=0, width=1.0\textwidth]{f11.eps}

1824: \caption{The adopted upper limits on the gas-to-dust ratios of the

1825: DLAs in our sample.  For comparison, the various lines provide ranges

1826: of the measured values (given in the text) for the Milky Way (MW),

1827: Large Magellanic Cloud (LMC), and Small Magellanic Cloud (SMC).  The

1828: limits are illustrated with solid leftward arrows, based upon the

1829: $\lambda$5780 and $\lambda$6284 DIBs.  The A0~0235+164 reddening

1830: measurement (star datum point) is from

1831: \citet{junk04}.}\label{p:Gas_Dust}

1832: \end{figure}

1833: 

1834: 

1835: \end{document}

1836: 

1837: 

1838: