1: %\documentstyle[emulateapj]{article}

2: \documentstyle[12pt,aaspp4]{article}

3: %\documentstyle[12pt,aasms4]{article}

4: 

5: \newcommand{\co}{{\rm $^{12}$CO}}

6: \newcommand{\cothree}{{\rm $^{13}$CO}}

7: \newcommand{\cn}{{\rm C$^0$}}

8: \newcommand{\cp}{{\rm C$^+$}}

9: \newcommand{\ci}{\mbox{\rm [\ion{C}{1}]}}

10: \newcommand{\oi}{\mbox{\rm [\ion{O}{1}]}}

11: \newcommand{\cii}{\mbox{\rm [\ion{C}{2}]}}

12: \newcommand{\oiii}{\mbox{\rm [\ion{O}{3}]}}

13: \newcommand{\hi}{\mbox{\rm \ion{H}{1}}}

14: \newcommand{\hii}{\mbox{\rm \ion{H}{2}}}

15: \newcommand{\jone}{($J=1\rightarrow0$)}

16: \newcommand{\jtwo}{($J=2\rightarrow1$)}

17: \newcommand{\jthree}{($J=3\rightarrow2$)}

18: \newcommand{\jfour}{($J=4\rightarrow3$)}

19: \newcommand{\fsone}{($^3$P$_1\rightarrow^3$P$_0$)}

20: \newcommand{\fstwo}{($^3$P$_2\rightarrow^3$P$_1$)}

21: \newcommand{\fseno}{($^3$P$_0\rightarrow^3$P$_1$)}

22: \newcommand{\fsowt}{($^3$P$_1\rightarrow^3$P$_2$)}

23: \newcommand{\fscii}{($^2$P$_{3/2}\rightarrow^2$P$_{1/2}$)}

24: \newcommand{\percmcu}{cm$^{-3}$}

25: \newcommand{\percmsq}{cm$^{-2}$}

26: \newcommand{\persr}{sr$^{-1}$}

27: \newcommand{\kmpers}{km~s$^{-1}$}

28: \newcommand{\Kkmpers}{K~km~s$^{-1}$}

29: \newcommand{\intunits}{erg~s$^{-1}$~cm$^{-2}$~sr$^{-1}$}

30: \newcommand{\intensity}{erg~s$^{-1}$~cm$^{-2}$}

31: \newcommand{\fluxdens}{erg~s$^{-1}$~cm$^{-2}$~$\mu$m$^{-1}$}

32: \newcommand{\Ico}{\mbox{\rm I$_{\rm CO}$}}

33: \newcommand{\Ici}{\mbox{\rm I$_{\rm [CI]}$}}

34: \newcommand{\Icii}{\mbox{\rm I$_{\rm [CII]}$}}

35: \newcommand{\Ifir}{\mbox{\rm I$_{\rm FIR}$}}

36: \newcommand{\citoco}{\Ici/\Ico}

37: \newcommand{\ciitoco}{\Icii/\Ico}

38: \newcommand{\ciitoci}{\Icii/\Ici}

39: \newcommand{\xco}{\mbox{\rm X$_{\rm CO}$}}

40: \newcommand{\av}{\mbox{\rm A$_{\rm V}$}}

41: \newcommand{\xuv}{\mbox{$\chi_{\rm uv}$}}

42: \newcommand{\vlsr}{\mbox{V$_{\rm LSR}$}}

43: 

44: \begin{document}

45: 

46: \title{First Detection of Submillimeter \ci\ Emission in the Small Magellanic Cloud}

47: 

48: \author{Alberto D. Bolatto\altaffilmark{1}, James M. Jackson\altaffilmark{1},

49: Kathleen E. Kraemer\altaffilmark{1},}

50: \affil{Institute for Astrophysical Research, Department of Astronomy,} 

51: \affil{Boston University, 725 Commonwealth Ave., Boston MA 02215} 

52: \and

53: \author{Xiaolei Zhang\altaffilmark{2}} 

54: \affil{Smithsonian Astrophysical Observatory, 60 Garden St. MS 78, 

55: Cambridge, MA 02138} 

56: 

57: \altaffiltext{1}{bolatto, jackson, kraemer@bu-ast.bu.edu}

58: \altaffiltext{2}{Currently at Raytheon ITSS and NASA GSFC, code 685, 

59: Greenbelt, MD 20771 (xzhang@specs1.gsfc.nasa.gov)}

60: 

61: \begin{abstract}

62: We report the first detection of \ci\ \fsone\ emission at 609 $\mu$m

63: in a region of the Small Magellanic Cloud (N27). Environments poor in

64: heavy elements and dust such as the SMC are thought to be dominated by

65: photodissociation regions. This is the lowest metallicity source where

66: submillimeter neutral carbon emission has been detected. Studying the

67: \citoco\ ratio in several sources spanning more than an order of

68: magnitude in metallicity, $Z$, we find that the \citoco\

69: ratio increases for decreasing $Z$. The existence of such a trend points to a

70: photodissociation origin for most of neutral carbon in molecular clouds,

71: in agreement with standard PDR models. We also report ISO FIR

72: spectroscopic observations of N27, and use them to derive its physical

73: properties. Comparison between the density and radiation field

74: revealed by FIR diagnostics ($n\sim300$---1000~\percmcu,

75: $\xuv\sim30$---$100\chi_0$) and those derived from millimeter and

76: submillimeter data ($n\sim10^5$~\percmcu, $\xuv\lesssim30\chi_0$)

77: suggests that the FIR lines originate in more diffuse gas,

78: and are perhaps dominated by the interclump medium. Regardless of the

79: cause, analysis of the FIR and mm-submm data produces a discrepancy of 

80: two orders of magnitude for the density of this source.

81: \end{abstract}

82: 

83: \keywords{Magellanic Clouds --- galaxies: ISM --- galaxies: irregular 

84: --- infrared: ISM: lines and bands --- submillimeter --- ISM: clouds}

85: 

86: \section{Introduction}

87: \label{intron27}

88: 

89: Studies of the interstellar medium (ISM) in the Small Magellanic Cloud

90: (SMC) allow us to investigate star forming clouds that are poor in

91: heavy elements. The SMC is indeed an excellent choice to study this

92: phenomenon, because it is nearby ($D\sim61$~kpc; Laney \& Stobie

93: 1994), has an unobscured line of sight and very small internal

94: obscuration, and possesses the lowest metallicity among the nearby

95: members of the Local Group ($Z\sim Z_\odot/10$; Dufour 1984).  Because

96: of their low metallicity, morphological irregularity, and active star

97: formation, the Magellanic Clouds resemble primeval galaxies.

98: Therefore, studies such as this will eventually allow us to draw

99: conclusions about those much more elusive targets: the high-z

100: primitive galaxies.

101: 

102: The depletion of heavy elements, and the consequent low dust-to-gas

103: ratio ($D/G_{\rm SMC}\sim1/8\,D/G_{\rm MW}$; Bouchet et al. 1985),

104: affects the ISM by enhancing photodissociation.  Because of its high

105: abundance (i.e., strong self-shielding) and cross-shielding with

106: atomic hydrogen lines, molecular hydrogen is mostly unaffected by $Z$.

107: The usual tracer molecules such as CO, however, are efficiently

108: destroyed in dust-poor systems. A much larger fraction of the

109: ISM in these systems is thus expected to be occupied by

110: photodissociation regions (PDRs; e.g., Tielens \& Hollenbach 1985;

111: Sternberg \& Dalgarno 1995). Consequently, we expect the intensities

112: of the spectral lines associated with the PDR (e.g., \cii, \ci) to be

113: enhanced when compared to molecular lines (e.g., \co\ \jone),

114: resulting in increasing \ciitoco\ and \citoco\ ratios (Bolatto,

115: Jackson, \& Ingalls 1999).

116: 

117: The immediate purpose of this work is twofold: 1) to test for the

118: existence of an increasing \citoco\ ratio with decreasing metallicity

119: $Z$ (\S\ref{disccin27}), and 2) to investigate the power of

120: far-infrared (FIR) spectroscopic diagnostics and compare their results

121: to those obtained using millimeter and submillimeter lines

122: (\S\ref{discpcn27}). An \citoco\ dependence on $Z$ could be useful for

123: determining metallicities in high-z sources using only radio

124: observations. The trend in \citoco\ with $Z$ is also a test for the

125: origin of neutral carbon inside molecular clouds. An increasing

126: \citoco\ ratio for decreasing $Z$ points to a PDR origin, while a flat

127: $\citoco$ ratio suggests that \cn\ is produced inside the CO cores by

128: means other than UV photons.

129: 

130: Our SMC source, the N27 nebula, was first

131: cataloged by Henize (1956). A later survey by Davies, Elliott and

132: Meaburn (1976) calls it DEM40. Its associated IR peak

133: is also known as LIRS49 (Schwering \& Israel 1989). This region was

134: mapped by the Swedish-ESO Submillimeter Telescope (SEST) in the \jone\

135: and \jtwo\ transitions of CO (Rubio et al. 1993), and studied in

136: detail in several molecular transitions by Heikkil\"a et

137: al. (1999). Albeit a weak emitter when compared to Galactic sources,

138: the molecular cloud associated with N27 features some of the strongest

139: CO \jone\ emission in the SMC (Israel et al. 1993).

140: 

141: \section{Observations}

142: \subsection{Far Infrared Data}

143: \label{obsfirn27}

144: We observed N27 using the Long Wavelength Spectrometer (LWS) onboard

145: the Infrared Space Observatory satellite (ISO), at

146: $R=\lambda/\Delta\lambda\sim200$. These LWS data are a 4160 second-long 

147: full-grating scan obtained on 20 December 1997 as part of the

148: LOMETPDR project. The instrumental characteristics are described by

149: Clegg et al. (1996), and the calibration of the ISO LWS instrument is

150: described by Swinyard et al. (1996). The data pipeline was version 8.7

151: and the observations were processed using ISAP version 2.0.  We

152: averaged the spectra across scans and scan direction, but not across

153: detectors, with the default wavelength spacing of 0.135 $\mu$m.  The

154: averaging was done after 2.5$\sigma$ clipping of outlying points. 

155: For the purposes of obtaining an \Ifir\ estimate only, we joined the

156: detectors together by offsetting them with the dark correction method in

157: ISAP, using detector LW1 as the reference.  Table \ref{tabiso}

158: summarizes the spectroscopic results.

159: 

160: We can assess the calibration of the LWS data by comparing it with

161: previous \cii\ measurements of N27 featuring

162: a similar beam size (Israel \& Maloney 1993). The

163: intensity measured by FIFI at the LWS pointing is

164: \Icii$\approx9\times10^{-4}$ \intunits. The LWS measured intensity

165: after applying the recommended correction factors for beam size and

166: extended source is \Icii$\approx9.5\times10^{-4}$ \intunits, in

167: remarkable agreement with the FIFI observations. The measurements

168: listed in Table \ref{tabiso} include both corrections.

169: 

170: \subsection{Submillimeter Data}

171: \label{obssubmmn27}

172: We observed the \ci\ \fsone\ transition of neutral atomic carbon at

173: $\nu\simeq492.1607$ GHz (609 $\mu$m) using AST/RO, the Antarctic

174: Submillimeter Telescope and Remote Observatory located at the

175: Amundsen-Scott South Pole base (Stark et al. 1997).  The observations

176: were obtained on 3 to 7 July of 1998, using the AST/RO SIS

177: quasi-optical receiver, with a system temperature

178: $T_{sys}\sim1800$---2400~K. The backend was the 2048 channel low

179: resolution AOS.  The spectra were observed in position switching mode,

180: chopping 15\arcmin\ in Azimuth (same as R.A. at the

181: pole). At 492 GHz the telescope beam had

182: HPBW$\sim3.8\arcmin\pm0.3$. The forward efficiency determined from

183: skydips was 70\%, and it is assumed to be identical to $\eta_{mb}$.

184: The pointing was better than $30\arcsec$.  A $6\arcmin\times6\arcmin$

185: region centered on this source was mapped on a 30\arcsec\ grid,

186: considerably oversampling the beam.  The data were calibrated using

187: the standard procedure for AST/RO, which includes sky, hot and cold load

188: measurements every 30 minutes, and processed using the COMB

189: package. The total accumulated integration time, after rejecting bad

190: spectra, was 63 hours. The spectrum obtained after averaging all the

191: observations together is shown in Fig. \ref{cispectrum}.  A gaussian

192: fit to the data gives $v_{cen}=115.2\pm1.4$~\kmpers, $\Delta

193: v=9.3\pm3.3$~\kmpers, and $T_{max}=14\pm4$~mK (main beam

194: brightness). The location of the \ci\ line agrees very well with the

195: strongest CO component, for which $v_{cen}=114.5$ \kmpers\ (Rubio et

196: al. 1993).

197: 

198: The relatively weak \ci\ emission from N27, combined with its small

199: size and consequent beam dilution, produced a low signal-to-noise

200: map. Thus the analysis in \S\ref{disccin27} will be focused on the

201: averaged spectrum for the region.

202: 

203: \section{Results and Discussion}

204: 

205: \subsection{Metallicity and the \citoco\ Ratio}

206: \label{disccin27}

207: 

208: The prime motivation for this study is to determine the \citoco\

209: intensity ratio in the SMC, an environment that because of its low

210: metallicity and low dust-to-gas ratio should be dominated by PDRs

211: (cf., \S\ref{intron27}). To do this we compare our \ci\ data with the

212: \co\ \jone\ map by Rubio et al. (1993). This map covers only a

213: relatively small fraction of the area mapped in \ci, but it probably

214: contains most of the CO emission: to obtain a \citoco\ ratio we assume

215: that there is no CO emission outside this region. The CO emission

216: associated with N27 has two components, at 114.5 and 126.9

217: \kmpers. For this analysis we have integrated the CO and \ci\ data

218: over the velocity range corresponding to the strongest component, at

219: $\sim114$ \kmpers.  In order to compute a ratio we have convolved the

220: CO \jone\ integrated intensity data to the angular resolution of the

221: \ci\ map. Subsequently we sampled the convolved CO map at the

222: positions observed in \ci, excluding those positions where \ci\

223: spectra were discarded.  These samples were averaged with the same

224: weighting scheme used for \ci.  This procedure allows us to obtain a

225: reliable average ratio for the N27 region, relatively insensitive to

226: uncertainties in the \ci\ beam size and pointing.  The resulting

227: intensities, integrated over the velocity interval 107---119 \kmpers,

228: are $\Ici\simeq0.132\pm0.027$ and $\Ico\simeq0.517\pm0.052$ \Kkmpers,

229: yielding a ratio $\citoco\approx0.26\pm0.06$. The corresponding

230: cooling ratio (the ratio of intensities in \intunits) is

231: $\citoco\approx19.9\pm4.5$.  The uncertainties are

232: $1\sigma$ and include the statistical errors as well as a 10\%

233: $1\sigma$ calibration uncertainty in the intensities, added in

234: quadrature.

235: 

236: Figure \ref{pdrci} compares the \citoco\ ratio in N27 to a

237: sample of molecular clouds associated with star-forming regions in

238: systems of different metallicities.  This plot suggests that there is

239: an increasing trend in the \citoco\ ratio with decreasing metallicity

240: $Z$, such that $\citoco\propto Z^{-0.5}$. The scatter around this line

241: is relatively large, as shown by the data for the Large

242: Magellanic Cloud. Theoretical expectations notwithstanding, this

243: scatter has made it difficult to identify a trend in past studies over

244: a narrower range of metallicities (e.g., Bolatto et al. 2000a).

245: 

246: Shown also in Fig. \ref{pdrci} is a model for carbon in unresolved PDRs

247: (Bolatto et al. 1999). This model consists of intensity

248: ratios computed over a distribution of clumps with uniform excitation,

249: where each spherical clump is divided in three concentric regions of

250: \cp, \cn, and CO. The sizes of these regions depend on the metallicity

251: (or the dust-to-gas ratio) of the ISM.  While necessarily crude this

252: model incorporates the effects of clumping, and the increase in the

253: physical size of the PDR for decreasing metallicity.  The growth of

254: the PDR is mostly caused by the diminishing attenuation of the UV

255: radiation, as the dust-to-gas ratio of the ISM decreases for smaller

256: metallicities. In this scenario, the primary reason why the \citoco\

257: ratio increases for decreasing metallicity is the reduction and

258: eventual disappearance of the CO cores in the clumps.

259: 

260: Are the models for the \citoco\ ratio as a function of $Z$

261: quantitatively consistent with the measurements?  Bolatto et al. explore two

262: possibilities: \cn\ produced in the growing PDR (model A), and \cn\

263: restricted to a region whose size does not change with metallicity

264: (model B).  This latter model predicts an essentially flat \citoco\

265: ratio ($\citoco\propto Z^{+0.1}$). A similar result (constant

266: $\citoco$) would be obtained if \cn\ was produced inside the CO cores

267: via unidentified chemical reactions.  The slope predicted by model A

268: in this metallicity range is slightly steeper than that suggested by

269: the data, $\citoco\propto Z^{-0.8}$.  Whether this difference is

270: actually significant is difficult to assert given the paucity of

271: observations and the intrinsic scatter in the data. Nevertheless,

272: several possibilities could explain this discrepancy: 1) an increase

273: in CO excitation for decreasing metallicity (i.e., hotter molecular

274: gas for diminishing $Z$), 2) clump-to-clump

275: shielding preventing the efficient photodissociation of CO, 3) a

276: dust-to-gas ratio not proportional to the oxygen abundance (but this

277: would require the dust abundance to be a slower function of $Z$,

278: e.g. D/G$\sim Z^{0.5}$, when observations suggest D/G$\sim Z^2$;

279: Lisenfeld \& Ferrara 1998), and 4) some fraction ($\sim30$\%) of the

280: \cn\ produced inside the CO cores instead of the PDR. We conclude

281: that, within the bounds imposed by the limitations of the models and

282: the paucity of the measurements, the observed trend for the \citoco\

283: ratio with metallicity shown in Fig. \ref{pdrci} agrees with

284: theoretical expectations for mostly photodissociation-produced \cn.

285: 

286: \subsection{Physical Conditions}

287: \label{discpcn27}

288: 

289: The FIR spectroscopic data can be used to determine the 

290: conditions prevalent in N27. Using Table \ref{tabiso} and the PDR models

291: by Kaufman et al. (1999) we will find the UV field, \xuv, and the gas

292: density, $n$. The important FIR ISM diagnostics

293: computed by Kaufman et al. are: 1) the \oi\ \fseno/\fsowt\ line ratio

294: (${\rm I_{145}/I_{63}}\sim0.051\pm0.006$, according to our

295: measurements), 2) the \oi\ \fsowt\ to \cii\ \fscii\ line ratio (${\rm

296: I_{63}/I_{158}}\sim0.42\pm0.01$), 3) the ratio of \oi\ \fsowt+\cii\

297: \fscii\ to the FIR continuum (using the LWS spectrum we estimate

298: $\Ifir\sim1.1\times10^{-2}$ \intunits, yielding ${\rm

299: (I_{63}+I_{158})/\Ifir}\sim0.012\pm0.002$), and 4) the \cii\ \fscii\

300: intensity itself. These joint four diagnostics agree very well,

301: placing N27 in the low density, $n\sim300$---$1000$~\percmcu, moderate

302: UV field, $\xuv\sim30$---$100\chi_0$, region of the $n-\xuv$ diagrams

303: ($\chi_0$ is the interstellar UV field in the vicinity of the Sun,

304: $\chi_0\simeq1.6\times10^{-3}$ erg~cm$^{-2}$~s$^{-1}$; Habing 1968).

305: The corresponding PDR temperature in these conditions is $T_{pdr}\sim200$~K.

306: The FIR continuum dust temperature, derived from the LWS spectrum, is

307: $T_d\sim40$~K.

308: 

309: The density and radiation field derived using millimeter and

310: submillimeter lines, however, strongly disagree with the

311: FIR results. This lack of agreement is not surprising,

312: perhaps, since these lines probably originate in different regions of the

313: ISM. A millimeter wavelength multiline excitation analysis using several

314: molecular species was performed by Heikkil\"a et al. (1999), yielding

315: $n\sim10^{5}$~\percmcu, $T\gtrsim15$~K, and visual extinction

316: $\av\sim1$ for N27. The \citoco\ ratio determined in \S\ref{disccin27} 

317: strongly constrains the density according to the calculations

318: by Kaufman et al. (1999), also resulting in $n\sim10^4$---$10^5$~\percmcu\

319: (albeit over a much larger area).  Finally, using the \co\ \jone\

320: observations by Rubio et al. (1993) convolved to the FIFI beam we

321: compute an accurate \ciitoco\ ratio at the LWS pointing, yielding

322: $\ciitoco\approx7000$. This ratio alone does not constrain either

323: parameter very well, but implies $\xuv>100\chi_0$ and when used in

324: combination with the measured \citoco\ ratio results in

325: $\xuv\sim10^4$---$3\times10^5\chi_0$.  This is not a fair comparison,

326: however, since both the \ciitoco\ and \citoco\ ratios will be affected

327: by the low metallicity and extinction of this source. Fortunately,

328: Kaufman et al. (1999) provide some calculations for small $Z$ and

329: \av. Using their model tracks for $Z=0.1$ and $\av=1$ we find

330: $n\sim10^5$~\percmcu\ and $\xuv\sim30\chi_0$ for N27, based on

331: $\Icii/\Ifir\sim10^{-2}$ and $\Ico/\Ifir\sim10^{-6}$. These

332: measurements are also consistent with $n_{\rm CO}>n_{\rm C^+}$ at a

333: lower radiation field, if we allow the model tracks to be shifted in

334: the abscissa as described by the model authors.

335: 

336: At any rate, the diagnostics including \ci\ or molecular line

337: observations are inconsistent with the density $n\sim10^3$ \percmcu\

338: implied by the FIR transitions. Because in general the FIR lines are

339: generated closer to the surface of the cloud, this discrepancy may

340: indicate a real density gradient in the gas. Another possibility is

341: that the mm-submm and FIR lines originate in different regions of the

342: cloud, with the latter dominated by material in a more diffuse

343: interclump medium, or perhaps associated with the \hii\ region rather

344: than the molecular cloud. A third possibility is that the FIR line

345: ratios are significantly affected by $Z$ or \av, and using model results

346: computed for Galactic metallicity and large \av\ is inappropriate. If

347: this is the cause of the density discrepancy, however, it is

348: remarkable that {\em all four} FIR diagnostics agree on essentially

349: the same value for the density and radiation field. Independent of

350: which explanation is correct, we conclude that: 1) the

351: FIR and mm-submm measurements roughly agree on the radiation field

352: intensity, although it is extremely important to use models with the

353: appropriate $Z$ and \av, and 2) the mm-submm data systematically

354: indicate a density about two orders of magnitude higher than that

355: derived from FIR fine structure data.

356: 

357: \section{Summary and Conclusions}

358: We report new FIR and submillimeter observations of the star forming

359: region N27 in the Small Magellanic Cloud. Using our measured

360: intensities for the FIR continuum and \oi\ and \cii\ fine structure

361: transitions, and comparing with existing model calculations (Kaufman

362: et al. 1999), we derive the density and radiation field in N27 to be

363: $n\sim300$---1000~\percmcu\ and $\xuv\sim30$---$100\chi_0$. In 

364: contrast, the density derived using millimeter and submillimeter data

365: is two orders of magnitude higher, $n\sim10^5$~\percmcu, while the

366: radiation field is similar or lower, $\xuv\lesssim30\chi_0$. The cause

367: for this discrepancy is not clear, but we suggest that the FIR and the

368: mm-submm lines arise in different parcels of gas, with the former

369: perhaps dominated by the interclump medium or the \hii\ region.

370: 

371: We have detected \ci\ \fsone\ emission from the Small

372: Magellanic Cloud for the first time. This is the lowest metallicity

373: source in which submillimeter neutral carbon emission has been detected.

374: Low metallicity, dust-poor environments are thought to be dominated by

375: PDRs.  Despite the factor of $\sim5$ in metallicity spanned by N27 and

376: Orion, however, the \citoco\ ratio in N27 is only modestly larger than

377: the \citoco\ ratio in Orion. Nevertheless, when we study the ratio in

378: several sources spanning more than an order of magnitude in

379: metallicity, we find a noisy but convincing increasing trend in the

380: \citoco\ ratio for decreasing metallicity (Fig. \ref{pdrci}). This

381: trend is somewhat flatter than, but in rough agreement with, the

382: modelling results by Bolatto et al. (1999). The existence of an

383: increasing trend points to a photodissociation origin for most of neutral

384: carbon inside molecular clouds, in agreement with standard PDR models.

385: 

386: \acknowledgements This research was supported in part by grants NSF

387: OPP 89-20223, NSF AST 98-03065, NASA JPL 96-1513, and made use of NASA

388: ADS and the CDS databases. We wish to thank M\'onica Rubio for

389: providing the calibrated and baseline-removed CO \jone\ spectra, Frank

390: Israel for the FIFI \cii\ data, and Tom Bania for his useful comments.

391: 

392: \begin{references}

393: \reference{ACW97} Afflerbach, A., Churchwell, E., \& Werner, M. W. 1997, 

394: \apj, 478, 190

395: 

396: \reference{BJI99} Bolatto, A. D., Jackson, J. M., \& Ingalls, J. G. 1999,

397: \apj, 513, 275

398: 

399: \reference{BO00a} Bolatto, A. D., Jackson, J. M., Wilson, C. D., \& 

400: Moriarty-Schieven, G. 2000a, \apj, 532, 909

401: 

402: \reference{BO00b} Bolatto, A. D., Jackson, J. M., Israel, F. P., Zhang, X.,

403: \& Kim, S. 2000b, \apj, in print (astro-ph/0007149)

404: 

405: \reference{BO85} Bouchet, P., Lequeux, J., Maurice, E., Pr\'evot, L., \&

406: Pr\'evot-Burnichon, M. L. 1985, \aap, 149, 330

407: 

408: \reference{CL96} Clegg, P. E. et al. 1996, \aap, 315, L38

409: 

410: \reference{DEM76} Davies, R. D., Elliot, K. H., \& Meaburn, J. 1976, 

411: \memras, 81, 89

412: 

413: \reference{DU84} Dufour, R. J. 1984, in Structure and Evolution of the 

414: Magellanic Clouds, ed. S. van der Bergh \& 

415: K. S. de Boer (Dordrecht:Kluwer), 353

416: 

417: \reference{HA68} Habing, H. J. 1968, \bain, 19, 421

418: 

419: \reference{HE99} Heikkil\"a, A., Johansson, L. E. B., \& Olofsson, H. 1999,

420: \aap, 344, 817

421: 

422: \reference{HE56} Henize, K. G. 1956, \apjs, 2, 315

423: 

424: \reference{IS93} Israel, F. P. et al. 1993, \aap, 276, 25

425: 

426: \reference{IM93} Israel, F. P., \& Maloney, P. R. 1993, in New Aspects

427: of Magellanic Cloud Research, eds. B. Baschek, G. Klare and J. Lequeux

428: (Berlin:Springer-Verlag), 44

429: 

430: \reference{KA99} Kaufman, M. J., Wolfire, M. G., Hollenbach, D. J., \& Luhman,

431: M. L. 1999, \apj, 527, 795

432: 

433: \reference{LS94} Laney, C. D., \& Stobie, R. S. 1994, \mnras, 266, 441

434: 

435: \reference{LE79} Lequeux, J., Peimbert, M., Rayo, J. F., Serrano, A., 

436: \& Torres-Peimbert, S. 1979, \aap, 80, 155

437: 

438: \reference{LF98} Lisenfeld, U., \& Ferrara, A. 1998, \apj, 496, 145

439: 

440: \reference{OJ00} Ojha, R. et al. 2000, \apjlett, submitted

441: 

442: \reference{RU93} Rubio, M. et al. 1993, \aap, 271, 1

443: 

444: \reference{ST97} Stark, A. A., Chamberlin, R. A., Cheng, J., Ingalls, J. G.,

445: \& Wright, G. 1997, Rev. Sci. Inst., 68 (5), 2200

446: 

447: \reference{SD95} Sternberg, A., \& Dalgarno, A. 1995, \apjs, 99, 565 

448: 

449: \reference{SW96} Swinyard, B. M. et al. 1996, \aap, 315, L43

450: 

451: \reference{TH85a} Tielens, A. G. G. M., \& Hollenbach, D. 1985, \apj, 291, 722

452: 

453: \end{references}

454: 

455: \newpage

456: \begin{deluxetable}{llcrcr@{$\pm$}lcr@{$\pm$}l}

457: \tablewidth{0pc}

458: \scriptsize

459: \tablecaption{ISO LWS Line Intensities in N27}

460: \tablehead{\multicolumn{1}{c}{Specie} &

461: \multicolumn{1}{c}{Transition}&\multicolumn{1}{c}{Wavelength}&

462: \multicolumn{1}{c}{Beam Size} &\multicolumn{1}{c}{ESCF\tablenotemark{a}}&

463: \multicolumn{2}{c}{Line Flux} &\multicolumn{1}{c}{Continuum}&

464: \multicolumn{2}{c}{Line/Cont.\tablenotemark{b}}\\

465: &			      &\multicolumn{1}{c}{($\mu$m)}  &

466: \multicolumn{1}{c}{(sr)}&                                     &

467: \multicolumn{2}{c}{(\intunits)}&\multicolumn{1}{c}{(MJy sr$^{-1}$)}&

468: \multicolumn{2}{c}{}

469: }

470: \startdata

471: \mbox{[O~III]} & \fstwo & \phn51.815 & 

472: $11.8\times10^{-8}$ & 0.85 &

473: 1.34 & $0.22\times10^{-5}$ & 191 &

474: 0.20 & 0.03 \\

475: \mbox{[O~I]} & \fsowt & \phn63.184 & 

476: $13.0\times10^{-8}$ & 0.83 &

477: 3.93 & $0.09\times10^{-5}$ & 237 &

478: 0.72 & 0.02 \\

479: \mbox{[O~III]} & \fsone & \phn88.356 & 

480: $11.8\times10^{-8}$ & 0.67 &

481: 9.64 & $0.85\times10^{-6}$ & 293 &

482: 0.13 & 0.01 \\ 

483: \mbox{[O~I]} & \fseno & 145.525 & 

484: $6.7 \times10^{-8}$ & 0.58 &

485: 2.01 & $0.23\times10^{-6}$ & 415 &

486: 0.06 & 0.01 \\

487: \mbox{[C~II]} & \fscii & 157.741 & 

488: $6.7 \times10^{-8}$ & 0.58 &

489: 9.46 & $0.08\times10^{-5}$ & 394 &

490: 3.14 & 0.04 \\

491: \enddata

492: \tablenotetext{a}{Recommended extended source correction factor 

493: (ISO LWS Handbook online v1.0).}

494: \tablenotetext{b}{Line to continuum ratio at line center (unresolved lines, 

495: $\Delta\lambda=0.29$ $\mu$m for $\lambda<80$ $\mu$m. Otherwise 

496: $\Delta\lambda=0.60$ $\mu$m).}

497: \label{tabiso}

498: \end{deluxetable}

499: 

500: \newpage

501: \begin{figure}[p]

502: \plotone{fig1.eps} 

503: \figcaption[fig1.eps]{Averaged \ci\ spectrum of N27 obtained by AST/RO

504: (Hanning-smoothed to $\Delta v\sim2.1$ \kmpers). Superimposed are 

505: the rescaled Gaussian \co\ components measured by Israel et al. (1993).

506: There is a hint of the 127 \kmpers\ component in the \ci.\label{cispectrum}}

507: \end{figure}

508: 

509: \begin{figure}[p]

510: \plotone{fig2.eps} 

511: \figcaption[fig2.eps]{The \ci/CO line intensity ratio as a function of

512: metallicity, for molecular clouds in the Galactic Center (GC), the

513: Orion region, the LMC, IC~10, and the SMC. The continuous line is a

514: linear fit to the data: $\log[\citoco]=-0.47*(12+\log[{\rm

515: O/H}])+5.0$.  The dashed line is Bolatto et al. (1999) model A

516: (PDR-produced neutral carbon): $\log[\citoco]=-0.82*(12+\log[{\rm

517: O/H}])+8.2$.  The references for metallicity and \citoco\ ratio for

518: each source indicated in the figure are: (1) Afflerbach et al. 1997,

519: (2) Ojha et al. 2000, (3) Tauber et al. 1995, (4) Dufour 1984, (5)

520: Bolatto et al. 2000b, (6) Lequeux et al. 1979, (7) Bolatto et

521: al. 2000a, and (8) this work.

522: \label{pdrci}}

523: \end{figure}

524: 

525: \end{document}

526: