0801.2534/ms.tex
1: %\documentclass{aastex}               % default manuscript style
2: %\documentclass[preprint,12pt]{aastex}               	% default manuscript style
3: %\documentclass[preprint]{aastex}
4: \documentclass{emulateapj}
5: %%\usepackage{graphicx}
6: %%\usepackage{epsfig}
7: %\DeclareGraphicsRule{.tif}{png}{.png}{`convert #1 `basename #1 .tif`.png}
8:  
9: 
10: % Manuscript-dependent macros
11: \newcommand{\km}{${\rm km\,s}^{-1}$}
12: \newcommand{\fuse}{{\em FUSE}}
13: \newcommand{\hi}{H$\;${\small\rm I}\relax}
14: \newcommand{\hii}{H$\;${\small\rm II}\relax}
15: \newcommand{\neviii}{Ne$\;${\small\rm VIII}\relax}
16: \newcommand{\ari}{Ar$\;${\small\rm I}\relax}
17: \newcommand{\ci}{C$\;${\small\rm I}\relax}
18: \newcommand{\cii}{C$\;${\small\rm II}\relax}
19: \newcommand{\caii}{Ca$\;${\small\rm II}\relax}
20: \newcommand{\ciii}{C$\;${\small\rm III}\relax}
21: \newcommand{\civ}{C$\;${\small\rm IV}\relax}
22: \newcommand{\nni}{N$\;${\small\rm I}\relax}
23: \newcommand{\nii}{N$\;${\small\rm II}\relax}
24: \newcommand{\niii}{N$\;${\small\rm III}\relax}
25: \newcommand{\nv}{N$\;${\small\rm V}\relax}
26: \newcommand{\oi}{O$\;${\small\rm I}\relax}
27: \newcommand{\oii}{O$\;${\small\rm II}\relax}
28: \newcommand{\oiii}{O$\;${\small\rm III}\relax}
29: \newcommand{\oiv}{O$\;${\small\rm IV}\relax}
30: \newcommand{\ov}{O$\;${\small\rm V}\relax}
31: \newcommand{\ovi}{O$\;${\small\rm VI}\relax}
32: \newcommand{\ovii}{O$\;${\small\rm VII}\relax}
33: \newcommand{\oviii}{O$\;${\small\rm VIII}\relax}
34: \newcommand{\pii}{P$\;${\small\rm II}\relax}
35: \newcommand{\si}{S$\;${\small\rm I}\relax}
36: \newcommand{\sii}{S$\;${\small\rm II}\relax}
37: \newcommand{\siii}{Si$\;${\small\rm II}\relax}
38: \newcommand{\siiii}{Si$\;${\small\rm III}\relax}
39: \newcommand{\Siii}{S$\;${\small\rm III}\relax}
40: \newcommand{\siiv}{Si$\;${\small\rm IV}\relax}
41: \newcommand{\siv}{S$\;${\small\rm IV}\relax}
42: \newcommand{\svi}{S$\;${\small\rm VI}\relax}
43: \newcommand{\feii}{Fe$\;${\small\rm II}\relax}
44: \newcommand{\feiii}{Fe$\;${\small\rm III}\relax}
45: \newcommand{\Niii}{Ni$\;${\small\rm II}\relax}
46: \newcommand{\lya}{Ly\,$\alpha$\relax}
47: \newcommand{\lyb}{Ly\,$\beta$\relax}
48: \def\dex#1{10$^{#1}$}
49: \def\tdex#1{$\times$10$^{#1}$}
50: 
51: % Comments, running heads, etc.
52: \slugcomment{Accepted for publication in the ApJ}
53: \shortauthors{Lehner et al.}
54: \shorttitle{Metallicity and Ionization of the Bridge}
55: %----------------------------------------------------------------------------%
56: %----------------------------------------------------------------------------%
57: \begin{document}
58: 
59: \title{Metallicity and Physical Conditions in the Magellanic Bridge\altaffilmark{1} }
60: \author{N.\ Lehner\altaffilmark{2},
61: 	J.C. \ Howk\altaffilmark{2},
62: 	F.P. \ Keenan\altaffilmark{3},
63: 	J.V. Smoker\altaffilmark{3}
64: 	}
65:    
66: \altaffiltext{1}{Based on observations made with the NASA-CNES-CSA 
67: Far Ultraviolet Spectroscopic Explorer. FUSE is operated for NASA by the Johns 
68: Hopkins University under NASA contract NAS5-32985. Based on observations made with the NASA/ESA Hubble Space Telescope,
69: obtained at the Space Telescope Science Institute, which is operated by the
70: Association of Universities for Research in Astronomy, Inc. under NASA
71: contract No. NAS5-26555.}
72: \altaffiltext{2}{Department of Physics, University of Notre Dame, 225 Nieuwland Science Hall, Notre Dame, IN 46556}
73: \altaffiltext{3}{Astrophysics Research Centre, School of Mathematics and Physics, The Queen's University of Belfast, Belfast, UK.}
74: 
75: \begin{abstract}
76: We present a new analysis of the diffuse gas in the Magellanic Bridge (RA\,$\ga 3^{\rm h}$) based 
77: on {\em HST}/STIS E140M and {\fuse}\ spectra of 2 early-type stars lying within the Bridge 
78: and a QSO behind it. We derive the column densities of \hi\ (from \lya), 
79: \nni, \oi, \ari, \siii, \sii, and \feii\ of the gas in the Bridge. 
80: Using the atomic species, we determine the first gas-phase metallicity of the Magellanic Bridge,
81: $[{\rm Z/H}] = -1.02 \pm 0.07$ toward one sightline, and $-1.7 < [{\rm Z/H}] < -0.9$
82: toward the other one, a factor 2 or more smaller than the present-day SMC metallicity. 
83: Using the metallicity and $N($\hi$)$, 
84: we show that the Bridge gas along our three lines of sight is $\sim$70--90\% ionized, 
85: despite high \hi\ columns, $\log N($\hi$) \simeq 19.6$--20.1. 
86: Possible sources for the ongoing ionization are certainly the hot stars 
87: within the Bridge, hot gas (revealed by \ovi\ absorption), and leaking
88: photons from the SMC and LMC.
89: From the analysis of \cii*, we deduce that 
90: the overall density of the Bridge must be low ($<0.03$--0.1 cm$^{-3}$).
91: We argue that our findings combined with other recent observational results 
92: should motivate new models of the evolution of the SMC-LMC-Galaxy system.
93: \end{abstract}
94: 
95: \keywords{galaxies: abundances --- Magellanic Clouds --- galaxies: interactions --- ISM: structure --- ultraviolet: ISM}
96: 
97: \section{Introduction}
98: The evolution of galaxies is closely coupled to 
99: the interactions between them. Encounters
100: between galaxies are frequent and were certainly 
101: habitual in the early epoch of galaxy formation \citep[e.g.,][]{barnes92,larson78,maller06}. These 
102: encounters reshape the galaxies and transfer mass, energy, 
103: metals between the galaxies, and may even create new sites of star
104: formation within the newly produced gaseous features between the
105: galaxies or new burst of star-formations in the colliding
106: galaxies \citep{larson78,barton07,demello07}. These galactic interactions
107: are not believed to be a major contributor to 
108: the pollution of the intergalactic medium \citep[e.g.,][]{aguirre01}, 
109: but they nonetheless affect the metal content and physics of the galaxies 
110: themselves and their halos, and therefore the evolution of the galaxies
111: and their surroundings. 
112: 
113: In our galactic neighborhood, interactions
114: between the Magellanic clouds and the Galaxy are believed to have 
115: produced several large gaseous features: The Magellanic Bridge linking the Small
116: Magellanic Cloud (SMC) and the Large Magellanic Cloud (LMC), 
117: the Magellanic Stream and the Leading Arm apparenly linking the Clouds
118: and our Galaxy \citep[e.g.,][]{mathewson74,putman00,bruns05}.  
119: The Magellanic Bridge is believed to have been produced
120: through an interaction between the SMC and LMC, while the
121: Stream and Leading arm may have arisen through an interaction
122: between the Galaxy and the Clouds \citep[e.g.][]{gardiner96}, although 
123: the origin of these latter gaseous features is far from settled
124: \citep{nidever07,besla07}. The proximity of these gaseous features 
125: and the global view of the Magellanic System with little line-of-sight confusion
126: allow us to study the result of interacting galaxies in a way not 
127: otherwise possible. The Magellanic Bridge (hereafter Bridge) is the focus of 
128: the present work. 
129: 
130: In the past few years, multi-wavelength observations of the Bridge have 
131: revealed some key characteristics. Radio \hi\ observations have 
132: shown that approximately 2/3 of the \ion{H}{1} gas surrounding the Clouds is
133: found in the Magellanic Bridge, while only 25\% and 6\% are found in
134: the Stream and the Leading Arm, respectively \citep{bruns05}.  Some of the
135: Bridge gas appears even to build-up and feed the Stream via an Interface
136: Region \citep{bruns05}. Far-ultraviolet  observations have 
137: shown multiple gas phases in the Bridge including not only the well
138: known neutral gas, but also a significant amount of ionized gas and a small
139: H$_2$ content \citep[][this paper]{lehner01,lehner02}. Dense molecular 
140: clouds were also subsequently identified in form of CO \citep{muller03,mizuno06} 
141: (although these CO surveys were realized in the SMC Wing, a much denser region
142: in stars and gas content of the Bridge).  
143: Optical studies have revealed  massive hot stars throughout the Bridge \citep{demers98}. 
144: The ages of these stars are $\sim$10 to 40 Myr, implying that star formation 
145: is ongoing within the Bridge gas since these stars could not migrate from the SMC 
146: during their lifetimes, although \citet{harris07} suggests that star formation 
147: in the Bridge may  have been more important 200--300 Myr ago. 
148: 
149: The N-body numerical  simulation of the Galaxy-SMC-LMC system by
150: \citet{gardiner96} have reproduced some 
151: of the observed characterics of the Bridge (principally the \hi\ column density 
152: and velocity distributions), along with those of the Stream.
153: In this paper, we still use their numerical simulation as a basis, although
154: we note that the hypotheses and validity of this model and other recent numerical simulations 
155: \citep[e.g.,][]{yoshizawa03} has been put recently in question in view of the 
156: new {\em Hubble Space Telescope (HST)}\ measurements of the proper motions of the LMC
157: and SMC \citep[][and see \S\ref{sec-per}]{kallivayalil06,kallivayalil06b,besla07}. 
158: Gardiner \& Noguchi's model  predicts
159: that the Bridge was formed from tidally-stripped gas pulled from the SMC 
160: during a close encounter between the SMC and the LMC some 200 Myr ago. 
161: The new calculations of the LMC and SMC orbits still suggest that the closest 
162: approach between these two galaxies occured some 200 Myr ago \citep{kallivayalil06},
163: and therefore the new proper motions may only affect the interpretation of the 
164: origins of the Stream and Leading Arm. Nevertheless, there appear to be some
165: challenges to these models. 
166: 
167: Tidal models predict that both 
168: gas and stars are pulled from the SMC, and yet search for an old stellar population in 
169: the Bridge has so far failed \citep{harris07}. Furthermore, material pulled from the SMC should have a somewhat similar 
170: metallicity that the SMC some 200 Myr ago. But this appears in conflict with the B-type star 
171: abundances in the Bridge that imply an extremely low Bridge metallicity, $-1.1$ dex 
172: from solar \citep{rolleston99,lee05}, a factor of 3 and 5 metal deficiency with respect to the SMC and LMC, 
173: respectively. The metallicity in the sparse regions of the 
174: Bridge is also at odds with those measured in the SMC wing, also known as the 
175: Western end of the Bridge, where the metallicity of B-type stars is found to be 
176: SMC-like \citep{lee05}. These metallicity estimates complicate the origin of the Bridge and 
177: suggest different evolutions or origins for the sparse and dense stellar regions
178: of the Bridge. 
179: 
180: An estimate of the present-day metallicy of the Bridge that is independent of stellar measurements 
181: is therefore critical. In this work, we present an estimate of the absolute 
182: gas-phase abundances of the Bridge toward two stars situated in relatively low \hi\ column
183: density regions. We compare the column densities of the neutral 
184: species (\nni, \oi, \ari) observed in absorption in the Space Telescope Imaging Spectrograph (STIS)
185: and {\em Far Ultraviolet Spectroscopic Explorer} ({\fuse}) spectra to those for
186: \hi\ Ly$\alpha$ lines toward two early-type stars, i.e. our metallicity estimate is mostly
187: independent of ionization correction.  The Bridge Ly$\alpha$ absorption is heavily blended
188: with the Galactic component, but we show that it is feasible to 
189: retrieve the Bridge \hi\ column density. Using the metallicity and the amount
190: of \hi\ along these sightlines, we can for the first time quantify the amount
191: of ionized gas in the Bridge; ionized gas is revealed to be a major component of the Bridge,
192: yet this gas-phase is mostly ignored in numerical simulations. 
193: 
194: 
195: Our paper is organized as follows. In \S\ref{sec-obs} we briefly discuss the 
196: previous STIS and {\fuse}\ observations, and the new {\fuse}\ observations of 
197: DGIK\,975 and DI\,1388. In \S\ref{sec-analysis} we discuss our measurements
198: of the metals and the \hi\ colum densities. We present in \S\ref{sec-result}
199: the first metallicity estimate of the Bridge gas and the fraction of neutral, ionized, and molecular
200: gas. From the \cii*\ diagnostic, we estimate the electron density and
201: cooling rate in the Bridge. In \S\ref{sec-discuss} we discuss our results (in particular 
202: address the possible origins for the low metallicity and ionization
203: sources of Bridge) and the current observational challenges to numerical simulations. 
204: Finally, in \S\ref{sec-sum} we summarize our key findings.
205: 
206: 
207: \section{Observations}\label{sec-obs}
208: In  Fig.~\ref{fig-map} we show the locations of our targets on an
209: \hi\ map of the Magellanic System from \citet{bruns05} (more properties 
210: about the stars can be found in Lehner 2002).
211: DI\,1388 and PKS\,0312--77 are situated approximately midway between the SMC and LMC, 
212: and DGIK 975 lies at the eastern end of the Bridge near the LMC halo, 
213: allowing us to probe different regions of the Bridge (see Fig.~\ref{fig-map}). 
214: All these sightlines are outside the SMC Wing that is approximately delimited
215: by the box in Fig.~\ref{fig-map}. 
216: 
217: \begin{figure*}[tbp]
218: \epsscale{0.7} 
219: \plotone{f1.eps}
220: \caption{Location of our targets superimposed on the \hi\ column density distribution from \citet{bruns05}
221: (darker regions represent higher $N($\hi$)$, see Br\"uns et al. for the detailed $N($\hi$)$ scale).
222: Our sightlines and major features of the Magellanic system are indicated. The box highlights 
223: a region known as the SMC Wing or Western end of the Magellanic Bridge.
224: \label{fig-map}}
225: \end{figure*}
226: 
227: 
228: 
229: The two stars were observed with {\em FUSE}, {\em HST}/STIS E140M, 
230: and the Anglo-Australian 3.9-m telescope (AAT). 
231: The QSO PKS\,0312--77 was observed with STIS 
232: E140M and {\em FUSE}, but the STIS and {\em FUSE}\ data are of too poor 
233: quality for a detailed metal-line absorption analysis. Unfortunately, the
234: {\em FUSE}\ program F018 (PI: Lehner) to obtain a good quality FUV spectrum 
235: of this QSO was not completed before the failure
236: of {\fuse}\ and only 16\% of the requested time was observed.  
237: However, the STIS data can be rebinned and used to study the Ly$\alpha$ absorption.  
238: For the STIS E140M and AAT observations of DGIK\,975 and DI\,1388, 
239: we refer the reader to \citet{lehner01} and \citet{lehner02}. The 
240: previous {\fuse}\ observations of these two stars are also described 
241: in \citet{lehner02}.  New {\em FUSE}\ observations 
242: were obtained for DI\,1388 (program U106 non-proprietary re-observations of science 
243: targets) and DGIK\,975 (program G050, PI: Lehner),
244: adding another 10 ks  and 29  ks exposure time, respectively. 
245: (Note that the total time for the program U106  is 
246: 30 ks, but about half the exposures have no signal.) 
247: The total {\fuse}\ exposure times are therefore 28 ks for DI\,1388
248: and 52 ks for DGIK\,975.
249: 
250: All the {\fuse}\ data were recalibrated using the newest {\sc calfuse} version 
251: (v3.2, Dixon et al. 2007). The extracted spectra associated with the separate exposures 
252: were aligned by cross-correlating the positions of absorption lines, and then co-added.
253: The oversampled {\fuse}\ spectra were binned to a bin size of 0.027 \AA\ 
254: (4 pixels), providing about three samples per $\sim$20 \km\ resolution element. 
255: In order to achieve the optimum signal-to-noise, 
256: segments with overlapping wavelengths were coadded for the DGIK\,975
257: {\fuse}\ spectra. However, we ensured before coadding the various
258: spectra that none of the absorption lines of interest was affected
259: by fixed-pattern noise, by comparing the interstellar profiles in multiple
260: detector segments. 
261: The zero point in the final {\fuse}\ wavelength scale was
262: established by  shifting the average {\fuse}\ velocity to
263: the STIS 140M velocity of the same species (e.g., \oi, \nni, \feii). 
264: STIS data reductions provide an excellent wavelength calibration, with
265: a velocity uncertainty of $\sim$1 \km. 
266: 
267: 
268: \section{Analysis}\label{sec-analysis}
269: \subsection{Metals}\label{sec-metal}
270: The continuum levels near the metal absorption lines 
271: were modeled by fitting Legendre polynomials within 
272: about $[-300,600]$ \km\ of each absorption line. Low-order polynomials 
273: were generally adopted, but in some cases high-order polynomials were
274: necessary (e.g. \feiii\ is blended with the stellar photospheric 
275: and wind lines).  For weak lines, several continuum placements were tested 
276: to be certain that the continuum error was robust \citep[see][]{sembach92}. 
277: In Figs.~\ref{fig-di} and \ref{fig-dgik}, we show the normalized spectra
278: for the interstellar metal-line transitions in DI\,1388 and DGIK\,975, respectively
279: \citep[see also][]{lehner01,lehner02}. Selected transitions have no serious blends with 
280: other features. 
281: 
282: In order to estimate the column density we adopted 
283: the apparent optical depth (AOD) method of \citet{savage91}. In this method 
284: the absorption profiles are converted into
285: apparent column densities per unit velocity $N_a(v) = 3.768\times
286: 10^{14} \ln[F_c/F_{\rm obs}(v)]/(f\lambda)$ cm$^{-2}$\,(\km)$^{-1}$, where $F_c$ is the
287: continuum flux, $F_{\rm obs}(v)$ is the observed flux as a function of
288: velocity, $f$ is the oscillator strength of the absorption and
289: $\lambda$ is in \AA\ (atomic parameters were adopted from Morton 2003).  
290: The total column density was obtained by integrating over the absorption profile  $N_a= \int N_a(v)dv$.
291: According to \citet{savage91}, this method is adequate
292: for data with $b_{\rm line} \ga 0.25$--$0.50 b_{\rm inst}$, where $b_{\rm line}$
293: is the intrinsic $b$-value of the line and $b_{\rm instr}$ is the $b$-value 
294: of the instrument. Since $b \equiv$\,FWHM$/1.667$, for STIS E140M, $b_{\rm inst} \simeq 4$ \km\ 
295: and for {\em FUSE}, $b_{\rm inst} \approx 12$ \km. Therefore, we assume that
296: a negligible fraction of the gas has $b\ll 1$ \km, an assumption made implicitly in most 
297: ISM abundance analyses using these instruments. Yet, we note that if the apparent column densities of a species with similar 
298: transitions estimated by both intruments are similar, this implies that there must be
299: little or no unresolved saturation. 
300: 
301: The reader should be aware that signatures of cold gas in the Bridge have been found 
302: toward DI\,1388 and PKS\,0312--77. The fraction of cold gas with respect
303: to the warmer gas is unknown. We note that \citet{lehner02} derived $b \simeq 2.6$ \km\ 
304: for the H$_2$ in the Bridge toward DI\,1388 using a curve of growth with a 
305: single component. If a typical temperature of the molecular
306: gas is $T\sim 100$--200 K, turbulent motions would dominate the broadening of the 
307: H$_2$ lines. Hence even though there may be cold gas, turbulent motions may be
308: important enough to keep the broadening of the atomic and ionic lines greater than 1 \km. 
309: Because the investigated gas is multiphase \citep{lehner01,lehner02}, the AOD method 
310: is favored over the curve-of-growth (COG)  or profile fitting methods because the AOD does 
311: not make any a priori on the kinematical distribution of the gas along the sightline.
312: Finally, we note that our estimates are also based on non-detection of a line: these
313: strict limits are consistent with our AOD estimates (see below), given us confidence in 
314: the results presented here. 
315: 
316: When $\tau_a \ll 1$ (which is the case for several transitions used in
317: this work), unresolved saturation should not be problematic as long as $b$ is 
318: not much smaller than 1 \km. For stronger lines, unresolved saturated structure can 
319: be identified by comparing the lines of the same species with different $f\lambda$ . 
320: Following \citet{savage91}, the difference in $f\lambda$ must be a factor of 2 (or 0.3 dex) or more to be 
321: able to detect unresolved saturation. The $f\lambda$ values are summarized 
322: in Table~\ref{t-metal}. If some moderate saturation exists, we can
323: correct for it using the procedure described in \citet{savage91}. 
324: For various cases of blending and line broadening, 
325: they found a tight relation between the difference of the true column density
326: and the apparent column density of the weaker line 
327: versus the difference of the strong line and weak line apparent
328: column densities. The correction to the apparent column density
329: of the weak line for a given difference between  the strong- and weak-line 
330: apparent column densities are summarized in their Table~4. 
331: 
332: In cases where we have only reliable information on a single line, 
333: we only quote a lower limit on the apparent column density. However,
334: we note that due to the low metallicity of the Bridge, the peak apparent optical depths 
335: of metal lines are generally less than 1. Even strong transitions such as 
336: \oi\ $\lambda$1302 and \nii\ $\lambda$1083, which have line profiles that are 
337: usually completely saturated in the diffuse gas of Galactic or SMC/LMC environments 
338: with similar \hi\ column densities, do not reach zero flux (at least toward DI\,1388). 
339: 
340: When no  absorption is observed for a given species, we measured
341: the equivalent width (and $1\sigma$ error) over the same velocity range found
342: from similar species that are detected. The 3$\sigma$ upper limit on
343: the equivalent width is defined as the 1$\sigma$ error times three.
344: The 3$\sigma$ upper limit on the column density was then derived assuming
345: the absorption line lies on the linear part of the curve of the growth.
346: 
347: 
348: 
349: 
350: \begin{deluxetable}{lccc}
351: \tabcolsep=6pt
352: \tablecolumns{4}
353: \tablewidth{0pt} 
354: %\tabletypesize{\scriptsize}
355: \tablecaption{Apparent Column Densities of the Metals \label{t-metal}} 
356: \tablehead{\colhead{Species}   &  \colhead{$\lambda$} &  \colhead{$\log f\lambda$} &\colhead{$\log N_a$}\\
357: \colhead{}   &  \colhead{(\AA)} &  \colhead{}&\colhead{}}
358: \startdata
359:   \cutinhead{DI\,1388}
360: \oi\   &  1302.168& 1.795 & $(>) 14.77 \pm 0.03$ \\
361: \oi\   &  1039.230& 0.974 & $(>) 14.97 \pm 0.06$ \\
362: \oi\   &  950.885 & 0.176 & $15.33 \pm 0.10$ \\
363: \oi\   &  936.630 & 0.534 & $15.26 \,^{+0.13}_{-0.16}$ \\
364: \oi\   &  924.950 & 0.154 & $15.32 \pm 0.13$ \\
365: \nni\  &  1200.710& 1.715 & $(>) 14.07 \pm 0.04$ \\
366: \nni\  &  1200.223& 2.018 & $(>) 13.89 \pm 0.04$ \\
367: \nni\  &  1134.980& 1.674 & $(>) 14.00 \pm 0.08$ \\
368: \nni\  &  964.626 & 0.882 & $(\le) 14.32 \,^{+0.13}_{-0.20}$ \\
369: \nni\  &  954.104 & 0.582 & $< 14.45$ \\
370: \ari\  &  1066.660& 1.857 & $< 13.17	   $ \\
371: \ari\  &  1048.220& 2.440 & $12.80 \pm 0.15$$^a$ \\
372: \siii\ &  1526.707& 2.308 & $14.14 \pm 0.07$ \\
373: \siii\ &  1304.370& 2.051 & $14.18 \pm 0.02$ \\
374: \sii\  &  1253.805& 1.136 & $14.25 \pm 0.10$ \\
375: \sii\  &  1250.578& 0.832 & $14.27 \pm 0.07$ \\
376: \feii\ &  1608.451& 1.970 & $13.95 \pm 0.02$ \\
377: \feii\ &  1144.939& 1.980 & $13.93 \pm 0.05$ \\
378: \feii\ &  1143.226& 1.341 & $13.95 \pm 0.10$ \\
379: \feii\ &  1055.262& 0.812 & $(\le)14.05 \,^{+0.15}_{-0.25}$ \\
380:   \cutinhead{DGIK\,975}
381: \oi\   &  1039.230& 0.974 & $(>) 15.20 \pm 0.04$ \\
382: \nni\  &  1134.980& 1.674 & $(\le) 13.88 \,^{+0.12}_{-0.16}$ \\
383: \ari\  &  1048.220& 2.440 & $<12.92$ 			\\
384: \siii\ &  1020.699& 1.225 & $(\le) 14.54\,^{+0.14}_{-0.20}$ \\
385: \feii\ &  1144.939& 1.980 & $14.18 \pm 0.03$ \\
386: \feii\ &  1143.226& 1.341 & $14.36 \pm 0.07$ \\
387: \feii\ &  1125.448& 1.245 & $14.44 \pm 0.10$ \\
388: \feii\ &  1093.877& 1.555 & $14.34 \pm 0.05$ \\
389: \feii\ &  1055.262& 0.812 & $(\le) 14.60 \,^{+0.10}_{-0.16}$ \\
390: \enddata
391: \tablecomments{Atomic parameters are from \citet{morton03}. $N_a$ is the apparent
392: column density in cm$^{-2}$, except when the ``$(<)$" sign is present, which is
393: a 3$\sigma$ upper limit; ``$(\le)$ indicates that the detection
394: is barely 3$\sigma$, and ``$(>)$" indicates that the line is likely saturated (see text 
395: for more details). 
396: $a$: We note that \citet{lehner02} found a column density twice smaller for \ari. 
397: N. Lehner revisited the original reduced data ({\sc calfuse} v2.0.5) and concludes
398: that the measurement was made using solely LiF\,1A, which appears to suffer from fixed-pattern
399: noise contamination (see Fig.~2 in Lehner 2002) but has been corrected with the new calibration. Indeed
400: the LiF\,1A and LiF\,2B segments reduced with {\sc calfuse} v3.2 give consistent results, 
401: and those are consistent with $N($\ari$)$ determined with LiF\,2B  {\sc calfuse} v2.0.5 data. 
402: }		
403: \end{deluxetable}
404: 
405: 
406: \begin{figure*}[tbp]
407: \epsscale{0.8} 
408: \plotone{f2.eps}
409: \caption{Normalized profiles against the LSR velocity near the Bridge velocities of  
410: neutral and singly ionized species in the STIS E140M and {\fuse}\ spectra of DI\,1388.
411: The Bridge absorption occurs between about 160 and 220 \km\ toward this line of sight. 
412: The ``x'' shows part of the spectrum that is contaminated by other absorbing features. 
413: \label{fig-di}}
414: \end{figure*}
415: 
416: 
417: \begin{figure*}[tbp]
418: \epsscale{0.8} 
419: \plotone{f3.eps}
420: \caption{Normalized profiles against the LSR velocity near the Bridge velocities of  
421: neutral and singly ionized species
422: in the STIS E140M and {\fuse}\ spectra of DGIK\,975.
423: The Bridge absorption occurs between about 100 and 200 \km\ toward this line of sight. 
424: The ``x'' shows part of the spectrum that is contaminated by other absorbing features. 
425: \label{fig-dgik}}
426: \end{figure*}
427: 
428: 
429: 
430: In Table~\ref{t-metal}, we present our estimates of the apparent column densities 
431: for DI\,1388 and DGIK\,975. Below, we review each sightline separately: 
432: 
433: {\em DI\,1388}: In Table~\ref{t-metal}, we present our raw
434: measurements of the apparent column densities for DI\,1388. 
435: Each considered species have at least 2 transitions with 
436: $\Delta \log(f\lambda) \ga 0.3$. 
437: \\
438: {\em Singly-ionized species:} In Fig.~\ref{fig-aod}, 
439: we show the apparent column density profiles of 
440: \siii\ $\lambda$$\lambda$1304, 1526 and the integrated
441: column densities agree remarkably well, especially 
442: at $\sim$198 \km\ where the absorption is the strongest. 
443: The continuum near \siii\ $\lambda$1526
444: is somewhat more complicated 
445: than for \siii\ $\lambda$1304, which results in a larger
446: uncertainty. A similar agreemeent is found for \sii, although
447: we note that  the continuum near the blue side of  
448: \sii\ $\lambda$1253 is more complicated. For \feii, the data are from {\fuse}\ 
449: and STIS. There is again an excellent agreement 
450: between the weak and strong transisitions. Furthermore, 
451: \feii\ $\lambda$$\lambda$1608 (STIS), 1144 ({\fuse}) have similar strengths
452: and apparent column densities. Therefore
453: the coarser spectral resolution of {\fuse}\ does not
454: seem to affect our apparent column estimates. 
455: Hence for the singly-ionized species, there is 
456: no evidence of saturation for the lines summarized in
457: Table~\ref{t-metal}. 
458: 
459: {\em \ari:} The transition at 1066.66 \AA\ is a 
460: 3$\sigma$ upper limit. This is a firm upper limit
461: irrespective of the intrinsic broadening of the line. 
462: \ari\ $\lambda$1048 is detected and its 
463: apparent column density is consistent with 
464: the weaker transition.  If the strong transition
465: has some unresolved saturation, it is likely less that 0.1 dex.
466: 
467: {\em \nni:} The transitions at 1200.71 and 1134.98 \AA\
468: have similar strengths, but the agreement in the $N_a$ of 
469: these two lines is not as good as for \feii, even though 
470: the $N_a$ estimates of these two lines overlap within 
471: 1$\sigma$. This suggests that there may be some unresolved
472: saturation. The apparent column density of \nni\ $\lambda$1200.22
473: is 0.18 dex smaller than \nni\ $\lambda$1200.711 (see also Fig.~\ref{fig-aod}). 
474: Therefore  $N_a($\nni\ $\lambda$1200.711$)$ needs
475: to be corrected for saturation. Using the results
476: of \citet{savage91}, we find that the correction
477: is 0.27 dex. The corrected column density of 
478: \nni\ is therefore $\log N($\nni$) = 14.34 \pm 0.08$ (where the 
479: errors include statistical and saturation
480: correction uncertainties).  This result can be directly tested with the weak
481: \nni\ $\lambda$964 transition, a 2.8--3.1$\sigma$ dectection, which
482: yields $\log N_a = 14.32 \,^{+0.13}_{-0.20}$ dex, in excellent agreement
483: with our above corrected column density (note that this \nni\ line has a similar
484: strength to \oi\ $\lambda$924 described below but
485: the continuum placement is much more straightforward for the \oi\ line, 
486: explaining the difference in the errors). 
487: The upper limit on \nni\ $\lambda$954 is also consistent with our estimate. 
488: 
489: {\em \oi:} The strong transitions at 1302.17 and 1039.23 \AA\ 
490: give much smaller apparent column densities than the weak
491: lines, and are therefore saturated. There are also 
492: two weak transitions at 924.95 and 950.89 \AA\ that
493: are 3.0--3.2$\sigma$ and 4.2--4.5$\sigma$ detections,
494: respectively.
495: Near the weakest transitions \oi\ $\lambda$924 
496: there is a contaminating feature on the blue side of 
497: the line, while there is a very weak contaminant on the
498: red side of \oi\ $\lambda$950 (see Fig.~\ref{fig-di}). 
499: However, the apparent column densities of \oi\ $\lambda$924 and 
500: $\lambda$950 are consistent with each other. We also tested this by
501: measuring half the \oi\ $\lambda$924 profile known to be 
502: free of contamination and 
503: the resulting $N_a \times 2$ is consistent with the value 
504: reported in Table~\ref{t-metal}.  The continua near these lines are  
505: well modeled with a straight line. 
506: \oi\ $\lambda$$\lambda$924,950 have  $\tau_a \ll 1$ 
507: and similar $ f \lambda$. \oi\ $\lambda$936 is about 0.3 dex stronger
508: than these lines and can be used to test for any unresolved structures. 
509: However, this line is partially blended with \hi\ $\lambda$937, 
510: complicating the continuum placement, which results in a larger
511: error. Within the errors, the apparent column densities of the 
512: three lines are consistent. Nevertheless, to be cautious,
513: we adopted the weighted average of the apparent column densities of 
514: \oi\ $\lambda$$\lambda$924,950 
515: and added an error in quadrature of $+$0.07 dex  to include any 
516: possible weak saturation.  Our adopted \oi\ column density is 
517: $\log N($\oi$) = 15.33 \,^{+0.11}_{-0.08}$. 
518: 
519: {\em DGIK\,975:} The STIS data are of lower quality than those for the DI\,1388
520: and the peak apparent optical depths are greater than 1 for all the detected species (see Fig.~\ref{fig-dgik}). 
521: We therefore use only the {\fuse} data. Unfortunately, the {\fuse}\ stellar spectrum
522: is not as well behaved as the one of DI\,1388, and we often have to rely 
523: on a single line. Stellar contamination and amount of \hi\  are more important 
524: (see \S\ref{sec-hi}), so that none of the weak \oi\ and \nni\ lines can be used. 
525: The $\lambda$1039 transition is the only available \oi\ line and is likely saturated,
526: providing a firm lower limit. (The apparent column density of \oi\ $\lambda$ cannot 
527: be estimated because the signal-to-noise is too low and its absorption reaches
528: zero-flux). \ari\ $\lambda$1048 is not detected and provides a firm 3$\sigma$ upper limit. 
529: \nni\ $\lambda$1134 is a $2.8$--$3.2 \sigma$ detection, and therefore the estimate of
530: the apparent column  should be considered as an upper limit. The same
531: applies for \siii\ $\lambda$1020 (the measurement of this line being complicated
532: by the uncertain continuum placement). Only for \feii, several transitions
533: can be measured. Following the above method, we estimate the 
534: average column density using the weak lines (i.e. excluding \feii\ $\lambda$1144), 
535: and add a systematic error to take into account possible unresolved structures. 
536: 
537: \begin{figure}[tbp]
538: \epsscale{0.8} 
539: \plotone{f4.eps}
540: \caption{Comparison of the apparent column density profiles for \siii\ 
541: and \nni\ for the DI\,1388 sightline. The \siii\ lines are essentially fully resolved while \nni\ shows 
542: some unresolved saturation. 
543: \label{fig-aod}}
544: \end{figure}
545: 
546: 
547: \begin{deluxetable}{lc}
548: \tabcolsep=6pt
549: \tablecolumns{2}
550: \tablewidth{0pt} 
551: %\tabletypesize{\scriptsize}
552: \tablecaption{Adopted column densities of the Metals \label{t-adopt}} 
553: \tablehead{\colhead{Species}    &\colhead{$\log N$}\\
554: \colhead{}   &\colhead{}}
555: \startdata
556:   \cutinhead{DI\,1388}
557: \oi\   &   $15.33 \,^{+0.11}_{-0.08}$ \\
558: \nni\  &   $14.34 \pm 0.08$ \\
559: \ari\  &   $12.80 \pm 0.15$ \\
560: \siii\ &   $14.17 \pm 0.02$ \\
561: \sii\  &   $14.26 \pm 0.06$ \\
562: \feii\ &   $13.95 \pm 0.02$ \\
563:   \cutinhead{DGIK\,975}
564: \oi\   &   $>15.20$ \\
565: \nni\  &   $\le 13.88 \,^{+0.12}_{-0.16}$ \\
566: \ari\  &   $<12.92$ \\
567: \siii\ &   $\le 14.54\,^{+0.14}_{-0.20}$ \\
568: \feii\ &   $14.38 \,^{+0.12}_{-0.05}$ \\
569: \enddata
570: \tablecomments{See \S\ref{sec-metal} for more details.}		
571: \end{deluxetable}
572: 
573: 
574: Our adopted column densities for neutral and singly-ionized species 
575: are summarized in Table~\ref{t-adopt}. There is an overall agreement
576: with previous column density estimates \citep{lehner01,lehner02}, 
577: in particular if in \citet{lehner02} the \oi\ and \nni\ column densities estimated in the DI\,1388 spectrum 
578: are systematically smaller than presently, they nonetheless overlap within $1.5\sigma$. 
579: The main difference is that he relies on a single-component curve-of-growth
580: analysis using both weak and strong lines, which depends on an assumed velocity distribution 
581: that is likely more complex than one Gaussian component. We believe 
582: using the AOD method is a better approach to determine the most 
583: reliable column densities and errors in view of the complexity of the  velocity 
584: distribution in these sightlines. 
585: 
586: 
587: \begin{figure*}[tbp]
588: \epsscale{0.9} 
589: \plotone{f5.eps}
590: \caption{{\em Left panels}: The \lya\ profiles in the STIS E140M spectra
591: of DI\,1388 ({\em top}), DGIK\,975 ({\em middle}), and PKS\,0312--77 ({\em bottom}). 
592: The blue line shows the continuum  and the yellow is the 
593: two-component fit to the interstellar \lya\ profile. {\em Right panels}: A zoom in on the core of the 
594: normalized \lya\ profiles is shown for each line of sight. The yellow line is the two-component
595: fit, while the green and red lines show only the Galactic and Bridge components, respectively. 
596: \label{fig-hi}}
597: \end{figure*}
598: 
599: 
600: \subsection{Neutral Hydrogen}\label{sec-hi}
601: 
602: \begin{deluxetable}{lcccc}
603: \tabcolsep=3pt
604: \tablecolumns{5}
605: \tablewidth{0pt} 
606: %\tabletypesize{\scriptsize}
607: \tablecaption{\hi\ Column Density \label{t-hi}} 
608: \tablehead{\colhead{Sightline/}   &  \colhead{$v_{\rm LSR}$} &  \colhead{$\log N$}&\colhead{$\log N^\star$} &\colhead{$\log N_{\rm f}$} \\
609: \colhead{Component}   &  \colhead{(\km)} &  \colhead{}&\colhead{} &\colhead{}}
610: \startdata
611: DI\,1388 G  	& $-5$  & $20.41 \pm 0.02$		& \nodata	&  $20.41 \pm 0.02$  \\
612: DI\,1388 B  	& $195$ & $19.67 \pm 0.07$		& $18.58$	&  $19.63 \pm 0.07$   \\
613: DGIK\,975 G 	& $-5$  & $20.95 \pm 0.08$		& \nodata	&  $20.95 \pm 0.08$  \\
614: DGIK\,975 B 	& $160$ & $19.95 \pm 0.30$		& $18.98$	&  $19.90 \pm 0.30$   \\
615: PKS\,0312--77 G& $5$   & $20.78 \pm 0.06$		& \nodata	&  $20.78 \pm 0.06$  \\
616: PKS\,0312--77 B& $210$ & $20.12 \pm 0.30$		& \nodata	&  $20.12 \pm 0.30$   
617: \enddata
618: \tablecomments{``G": Galactic component; ``B": Bridge component 
619: The LSR velocities are fixed and are estimated from the \oi\ absorption
620: profiles. $N$ is the observed column density, $N^\star$ is the estimated
621: stellar \hi\ contribution for the H$\beta$ profiles (see text for more
622: details), and $N_{\rm f}$ was
623: corrected for the stellar component in the Bridge component. 
624: Note that the errors include different choices of continuum placement
625: and polynomial degree for the continuum, and velocity shifts of $\pm 5$ \km.}		
626: \end{deluxetable}
627: 
628: Here we discuss our derivation of the Bridge \hi\ column based
629: on the analysis of the \lya\ absorption.  The \hi\ column density 
630: is more complicated to derive because both the Galaxy and Bridge 
631: contribute to the absorption. \hi\ 21-cm emission observations
632: toward the stars are of little use for the Bridge component because 
633: the stars are embedded in the Bridge with an unknown depth. We therefore 
634: need to derive $N($\hi$)$ from the \lya\ absorption. To do so, we
635: fit each damped \lya\ absorption line profile with two components 
636: that correspond to the Galactic and Magellanic Bridge components.  
637: The blue-ward wing of the profile is due primarily to the 
638: Galactic component, but the Galactic component cannot completely 
639: account for the absorption in the red-ward wing, indicating the need 
640: for a Bridge component.  By fixing the velocities of the Galactic and 
641: Bridge components, we can determine the \hi\ column densities.
642: We note that the amount of \hi\ gas between the Galaxy and the Bridge is 
643: negligible, since the \hi\ 21-cm data do not show any emission toward these 
644: two lines of sight at a level of $10^{18.3}$ cm$^{-2}$, too small to affect 
645: the damping wings of \lya. \oi\ is the best metal proxy for \hi\ since its 
646: ionization potential and charge exchange reactions with hydrogen ensure 
647: that the ionization of \ion{H}{1} and \ion{O}{1} are strongly coupled. 
648: We can therefore use the kinematics of \oi\ to infer those for \hi. While 
649: the \oi\ profiles show multiple structures in both the Galactic and Bridge components, 
650: we only use the average velocities to fix the values for each component. 
651: These are listed in Table~\ref{t-hi}. Below we detail our fitting results for each star. 
652: 
653: {\it DI\,1388:} This line of sight is our best case because the continuum
654: of the star is well behaved and the Galactic \hi\ column density is not
655: so large that it fills the absorption in the red part of the \lya\ profile to a large degree. 
656: In  Fig.~\ref{fig-hi}, we show the profile of \lya\ with its continuum and  the 
657: fit with the two components (top-left panel). On the top-right panel, we show 
658: part of the  normalized \lya\ profile with the two-component fit in yellow, 
659: in red  the Bridge component, and in green the Milky Way component. This shows 
660: that while the Galactic component nearly fits the blue part of
661: the \lya\ absorption, there is extra absorption in the red part 
662: that is missed by the Galactic component. 
663: In Table~\ref{t-hi}, we summarize the derived
664: column density for each component in the third column.  
665: To test the robustness of our fit, we undertook many simulations where the velocity centroids were
666: changed by $\pm 5$ \km, the placement of the continuum was varied, 
667: and the continuum was modeled by a range of polynomials of degree 
668: of 1 to 4. The errors reported in Table~\ref{t-hi} were estimated
669: to reflect the 1$\sigma$ distribution between all the trials.  
670: 
671: {\it DGIK\,975}: This line of sight is more complicated because 
672: the Galactic \hi\ component is much stronger than
673: toward DI\,1388. For the latter, the Galactic \hi\ component is about
674: a factor 5 stronger than the Bridge component, while toward 
675: DGIK\,975, we found the Galactic component is $\sim$10 times stronger
676: than that of the Bridge.  The \siiii\ $\lambda$1206 line is also very 
677: strong and removes some information in the blue part of the
678: \lya\ absorption. Finally the STIS E140M DGIK\,975 spectrum has a much lower
679: S/N; we therefore rebinned these data by 5 pixels before fitting. In Fig.~\ref{fig-hi}, 
680: we show our resulting best fit overplotted on the \lya\ 
681: profile (middle-left panel), where the continuum is well behaved and 
682: can be modeled by a straight line.  The normalized profile (middle-right panel) indeed
683: shows that the difference between the two-component fit and the single Galactic
684: component fit is small, which results in a large error in $N($\hi$)$
685: of the Bridge feature. However, as for DI\,1388, at $ 1219 \la \lambda \la 1222$ \AA,
686: where the optical depth is more important, the Galactic component does not 
687: properly fit the red part of the absorption without the Bridge component.
688: We also undertook several trials in order to derive the 1$\sigma$ error reported
689: in Table~\ref{t-hi}. 
690: 
691: The \hi\ 21-cm emission data can help to evaluate the 
692: solution from our fits, at least for the Galactic component; indeed, since 
693: the whole column of Galactic gas is probed in both emission and absorption 
694: toward the stars, the infered column densities derived from absorption
695: and emission spectra can be compared. The main uncertainty that 
696: arises from such a comparison is that the beam that collected the \hi\ 21 cm 
697: emission data is much larger than the pencil-like beam of the absorption 
698: observations. Therefore the beam of the radio telescope only represents
699: an average column density of the region near the line-of-sight.  
700: \citet{wakker01} studied this effect for low-, intermediate-, and high-velocity 
701: clouds by comparing $N($\hi$)$ measured via a beam of 36\arcmin, 9\arcmin, 1\arcmin,
702: and through \lya\ absorption. They found that the ``beam-effect'' was 
703: less important for the low-velocity gas where the range of ratio of $N($\hi$)$
704: measured with a 9\arcmin\ versus a 36\arcmin\ is  0.9 to 1.4 than for the HVCs where 
705: the range of ratio is 0.3 to 2.1. This difference is because the physical sizes probed 
706: by the \hi\ 21 cm data are naturally smaller in nearby gas and therefore less likely to 
707: have large angular/spatial variations. Hence \hi\ 21-cm emission observations
708: observed through a 15\arcmin\ beam should provide reliable $N($\hi$)$ within a
709: beam error of about $\pm 0.10$ dex (see also Lehner et al. 2004)
710: for the low-velocity Galactic gas that can be compared to the results from 
711: the profile fitting of the \lya\ absorption. 
712: 
713: 
714: With \hi\ 21-cm emission observations from the Parkes 64 m radio telescope with
715: a 15\arcmin\ beam  and 1 \km\ spectral resolution (M.E. Putman, 1999, private communication)
716: we derive the \hi\ column density for the Galactic component where we make the usual assumption 
717: that the medium is optically thin: $\log N($\hi$) = 20.44 \pm 0.11$ toward DI\,1388 and 
718: $\log N($\hi$) = 20.85 \pm 0.11$ toward DGIK\,975. Using  the 21-cm emission data 
719: from the 36\arcmin\ Leiden-Argentine-Bonn survey within $\sim$1\degr\ from the star's coordinates
720: \cite[LAB,][]{kalberla05,bajaja05}, we find 
721: Galactic \hi\ column densities that are systematically similar 
722: to those quoted above within the errors. Within $1\sigma$, 
723: the \hi\ column densities of the Galactic component derived from emission 
724: and absorption observations are consistent. 
725: 
726: Finally, we also consider  the QSO PKS\,0312--77 sightline:
727: since both the UV and radio observations probe the full depth of the Bridge, we 
728: can compare the resulting  \hi\ column densities measured with both observations with the 
729: caveat that we have to assume that small-scale variations in the 15\arcmin\ beam (corresponding 
730: to about 240 pc at 55 kpc) are small enough in the direction of  PKS\,0312--77.
731: \citet{kobulnicky99} detected \hi\ 21-cm emission for the Bridge component with the 
732: Parkes telescope toward this line-of-sight. They derived a Bridge \hi\ column 
733: density of $\log N($\hi$) = 20.09 \pm 0.13$, where the errors include 
734: statistical error and systematics associated with the beam that
735: is larger than our pencil beam for the UV observations (see above). 
736: We note that the LAB data give systematically larger $N($\hi$)$ for the Bridge 
737: but with relatively small variations within a radius (in the $l,b$ plane)
738: of $\sim$40\arcmin: the closest LAB pointing is 4\arcmin\ away where 
739: $\log N($\hi$) = 20.23 \pm 0.25$ and the range of $\log N($\hi$)$ is 
740: between 20.18 to 20.34 dex (toward the two stars, the variation in $N($\hi$)$
741: is quite larger -- a factor 2--3 -- within the same radius).   
742: 
743: The \lya\ profile of PKS\,0312--77 is shown in Fig.~\ref{fig-hi} (bottom-left panel) 
744: and is clearly more challenging to model than those for the two stars, 
745: as the S/N is lower (the data shown were rebinned by 5 pixels) and the red part
746: of the continuum is heavily affected by emission lines from 
747: the QSO (emission from \lyb, \ovi, and \oi), complicating 
748: the continuum placement. The Galactic
749: component is also strong, but because the whole Bridge gas is probed
750: along this line of sight, this results in a higher \hi\ Bridge column
751: density than toward the two stars. In Fig.~\ref{fig-hi}, we show one of our best fits, 
752: and as for the other lines of sight, the errors reflect many
753: trials. However, because of the complexity of the continuum, 
754: the degree of the polynomial was allowed to vary between 2
755: and 7; a 5$^{\rm th}$ degree polynomial is shown in the figure.  The Galactic component
756: again cannot account for absorption at $\lambda \ga 1219$ \AA.
757: Table~\ref{t-hi} summarizes the results and while the error is large, 
758: the measured value of $\log N($\hi$) = 20.12 \pm 0.30$ is quite consistent with the 
759: \hi\ 21-cm column density derived by \citet{kobulnicky99}. These authors do 
760: not provide $N($\hi$)$ for the Galactic component, but using the LAB data
761: we estimate that it is $\log N($\hi$) = 20.83 \pm 0.17$, consistent with the
762: \lya\ absorption estimate in the Galactic component. 
763: 
764: The general agreement between the $N($\hi$)$ estimates from
765: absorption and emission measurements gives further confidence
766: in our fits to the \lya\ absorption observed along each lines of sight. 
767: The error estimates were systematically estimated via several 
768: fit trials and we believe that we are conservative in our error 
769: estimates. The stellar sightlines need to be, however, corrected 
770: for  contributions from the stellar photospheres. 
771: We follow the method originally presented by \citet{diplas91} and expanded by 
772: \citet{howk99} to estimate the stellar \lya\ contribution, which involves measuring the 
773: photospheric H$\beta$ equivalent width of each star and using the
774: equation A7 in  Howk et al. to estimate $W($\lya$)$. The stellar \hi\ column density
775: density is then derived via $\log N($\hi$) = 18.27 + W($\lya$)$. 
776: Using our AAT spectra, we estimate that $W({\rm H\beta})= 2.1\pm 0.2$  \AA\ 
777: for  DGIK\,975 and  $W({\rm H\beta})= 2.1\pm 0.2$  \AA\ for DI\,1388. 
778: These correspond to stellar \hi\ logarithm column density of 
779: 18.58 and 18.98 for DI\,1388 and DGIK\,975, respectively. The correction is larger
780: for DGIK\,975 because the temperature of the star is lower than 
781: the temperature of DI\,1388.  The last column 
782: of Table~\ref{t-hi} report the Bridge \hi\ column densities corrected
783: for the stellar contribution.
784: 
785: \section{Abundances and Physical Conditions in the Magellanic Bridge}\label{sec-result}
786: \subsection{Metallicity}
787: In Fig.~\ref{fig-aodcomp}, we show the ratio of the apparent column densities
788: of \oi\ and \nni\ to $N_a($\sii$)$ normalized to the solar abundances
789: for the DI\,1388 sightline. Sulphur is our reference because it is not 
790: depleted into dust. Also note that throughout the text we use 
791: the following notation ${\rm [X^i/Y^j]} = \log N({\rm X}^i)/N({\rm Y}^j) - \log({\rm X/Y})_\odot$, and we
792: adopt the solar abundances of \citet{asplund06}.\footnote{For the species used
793: throughout this work, the followings are the adopted solar abundances $\log ({\rm X/H})$: 
794: C: $-3.61 \pm 0.05$;
795: N: $-4.22 \pm 0.06$; O: $-3.34 \pm 0.05$; Si: $-4.49 \pm 0.02$; Ar: $-5.82 \pm 0.08$; 
796: S: $-4.85 \pm 0.03$; Fe: $-4.55 \pm 0.03$ \citep{asplund06}. Note that our error estimates
797: do not take into account the solar abundance errors.}
798: We explicitly retain the ionization state information to emphasize 
799: the impact of ionized gas on these measurements. As already discussed by \citet{lehner01}, 
800: the Bridge gas toward DI\,1388 has two main components, one mostly ionized 
801: ($165 \le v_{\rm LSR} \le 193$ \km), the other one partially ionized ($193< v_{\rm LSR} \le 215$ \km).
802: The extremely ionized gas is illustrated by $[$\oi/\sii$] \simeq -1 $ for $170 \le v_{\rm LSR} \le 193$ \km\
803: (\sii\ is a important in both neutral and ionized gas, but \oi\ arises only in neutral gas).
804: Toward DGIK\,975, the normalized profiles shown in Fig.~\ref{fig-dgik} suggest a similar pattern: 
805: two main components are observed and the higher velocity component is more neutral than 
806: the lower velocity component as revealed by the stronger absorption in \oi\ and \nni\ and 
807: weaker \feii\ at 175 km and the opposite pattern is observed at $140$ \km. 
808: Therefore singly-ionized species cannot be directly used to estimate the metallicity
809: of the gas because they contain a significant contribution from ionized gas. 
810: 
811: 
812: \begin{deluxetable}{lc}
813: \tabcolsep=6pt
814: \tablecolumns{2}
815: \tablewidth{0pt} 
816: %\tabletypesize{\scriptsize}
817: \tablecaption{Gas-Phase Abundances of the Magellanic Bridge \label{t-abund}} 
818: \tablehead{\colhead{Species}  &\colhead{[X/H]} }
819: \startdata
820:   \cutinhead{DI\,1388}
821: \nni\ 	&  $-1.07 \pm 0.11 	  $	   \\
822: \oi\ 	&  $-0.96 \,^{+0.13}_{-0.11}	  $	   \\
823: \ari\ 	&  $-1.01 \,^{+0.16}_{-0.17}	 $	   \\
824:   \cutinhead{DGIK\,975}
825: \nni\ 	&  $(\le) -1.75\,^{+0.31}_{-0.38}$		    \\
826: \oi\ 	&  $>-1.36 (\pm 0.30)$		    \\
827: \ari\ 	&  $<-1.16 (\pm 0.30)$  	    
828: \enddata
829: \tablecomments{The adopted solar abundances are from Asplund et al. (2006): N: $-4.22$; 
830: O: $-3.34$; Ar: $-5.82$. The uncertainties in the solar abundances are not taken into
831: account in the errors listed in the table. A ``$<$" sign means a 3$\sigma$ upper limit. 
832: }		
833: \end{deluxetable}
834: 
835: 
836: 
837: \begin{figure}[tbp]
838: \epsscale{1.0} 
839: \plotone{f6.eps}
840: \caption{Relative gas-phase abundance of \oi\ and \nni\ with respect to 
841: \sii\  (where $\lambda$1250) for the DI\,1388 sightline. At $v_{\rm LSR} \ga 193$ \km, 
842: \oi\ and \nni\ absorption suffers from saturation. 
843: \label{fig-aodcomp}}
844: \end{figure}
845: 
846: 
847: 
848: On the other hand, the ionization of oxygen and hydrogen are strongly coupled, 
849: and  oxygen is only  mildly depleted into dust grains 
850: \citep[up to $-0.16$ dex using the updated \oi\ abundance -- see below --][]{meyer98}, implying 
851: that the \oi/\hi\ ratio is the best indicator of the metallicity of the Bridge gas.
852: We note that both the \hi\ column densities and molecular fraction 
853: in \citet{meyer98} are typically much larger than those 
854: in the Bridge gas. The oxygen measurements in the Local Bubble 
855: imply a depletion of $-0.12$ dex \citep{oliveira05}. 
856: Therefore the depletion of O in the Bridge is likely to 
857: be less than $-0.1$ dex.   We find
858: that the Bridge gas oxygen abundance is $[$\oi/\hi$] =-0.96 \,^{+0.13}_{-0.11}$ toward
859: DI\,1388. Toward  DGIK\,975,  only a lower limit is derived (see Table~\ref{t-abund}). 
860: We note that the solar abundance of O has changed quite significantly over the 
861: years and the new value may still be controversial. 
862: Prior measurements to \citet{asplund06} gave a solar O abundance $+0.17$ dex
863: larger \citep{grevesse98}, i.e  the Bridge O abundance would be further below. 
864: 
865: The solar Ar and N abundances appear more settled, with litte changes between 
866: those listed in \citet{grevesse98} and \citet{asplund06}. Furthermore,
867: Ar and N are not depleted into dust grains \citep{sofia98,meyer97}, and
868: \ari\ and \nni\  behave like \oi\ in gas with large \hi\ column density
869: \citep[see Fig.~6 in][]{lehner03}, i.e., the ionization fractions of these elements 
870: are coupled with the ionization fraction of H via a resonant charge-exchange
871: reaction. Toward DI\,1388, the values of the N and Ar abundances summarized in Table~\ref{t-abund}
872: show an excellent agreement with those of O.
873: This might appear puzzling at first glance since \oi\ is clearly detected in both components,
874: while \ari\ and \nni\ absorption is mostly absent in the 180 \km\ component. However, 
875: the neutral gas is clearly dominated by the component at 198 \km, since 
876: the column density in the 180 \km\ component is only about 10\% of the total
877: \oi\ column density (where we estimate $\log N($\oi$) \simeq 14.22$ in the 180 \km\
878: from a profile fit to the \oi\ $\lambda$$\lambda$1302, 1039, 963 lines).
879: 
880: The AOD estimate of \sii\ for $193 \le v_{\rm LSR} \le 215$ \km\ is 
881: $\log N($\sii$) = 13.98 \pm 0.08$. Assuming that 90\% of the
882: observed \hi\ is in this component (based on \oi) yields $[$\sii/\hi$] \simeq -0.8 $. 
883: Since absorption of \ciii, \nii, \siiii\ is found in this velocity range, 
884: this component is partially ionized. Therefore, the $[$\sii/\hi$]$ provides 
885: a strict upper limit to the abundance of S since the gas is partially 
886: ionized in this component. 
887: 
888: We mentioned above that O could be slightly depleted into
889: dust. If we assume an oxygen depletion of 0.05 dex (which is still consistent 
890: with the S abundance in Bridge and about 25\% ionized gas in this component), 
891: $[$\ari/\oi$] \sim -0.1$ and  $[$\nni/\oi$] \sim -0.2$. In low metallicity gas, 
892: N is generally found deficient \citep[e.g.,][]{henry07} 
893: (although we note for a metallicity of $-1$ dex, $[$N/O$]$ is more 
894: typically $-0.5$ dex), and this could explain the lower abundance of N. 
895: Nucleosynthesis history is unlikely to affect the Ar/O ratio since they
896: are both $\alpha$-elements. But \ari\ may be 
897: deficient because of photoionization \citep[see Fig.~6 in][]{lehner03}, although 
898: with $\log N($\hi$) = 19.6$ one would likely expect a deficiency less than 
899: 0.1 dex (except possibly if a hard ionization source overionizes the edge of the 
900: neutral gas). (The \citet{grevesse98} value gives a large O depletion, but 
901: the $[$\nni/\oi$]$ would be less deficient). 
902: The errors can accommodate a slight O depletion (as compared with Ar and N) or none; 
903: we therefore use a weighted average of the abundances of \oi, \nni, \ari\ 
904: to derive the metallicity of the gas within the Bridge: $[{\rm Z/H}] = -1.02 \pm 0.07$.
905: 
906: Although the errors are large toward DGIK\,975, the limits on \ari\ and 
907: \nni\ suggest an extremely low metallicity as well. Using \ari\ and \oi\ (both 
908: $\alpha$-elements), we can bracket the metallicity of the Bridge gas toward DGIK\,975: 
909: $-1.7 < [{\rm Z/H}] < -0.9$. We note that toward this sightlines N may
910: appear more defficient than the $\alpha$-elements. 
911: 
912: Our gas-phase metallicity is in agreement with the metallicity
913: derived from 3 Bridge B-type stars, where the average is $-1.1 \pm 0.1$ dex  \citep{rolleston99,lee05}.
914: Since the  interstellar and stellar abundance measurements are made using different
915: techniques and have different systematics, the present-day metallicity of the Bridge is now 
916: secure with an average value  $[{\rm Z/H}]= -1.05 \pm 0.06$ dex. This is about a factor $\sim$3 times
917: smaller than the SMC abundance of $-0.6$ dex solar (see, e.g., Welty et al. 1997, Russel \& Dopita 1992, 
918: and references therein). In \S\ref{sec-discuss} we discuss the implication
919: of the metallicity results. 
920: 
921: \subsection{Ionization and Molecular Fraction}\label{sec-ionization}
922: Using the measurement of \hi\ along our two stellar sightlines, one can estimate the 
923: fractions of neutral, ionized, and molecular gas. Assuming an average 
924: Bridge metallicity from the stellar and interstellar measurements
925: of about $-1$ dex solar, we can use the singly and doubly ionized species
926: to estimate the amount of ionized gas. Toward DI\,1388, we use sulphur 
927: since this element is not depleted into dust and has 
928: the same nucleosynthetic history as
929: O and Ar since these are all $\alpha$-elements. 
930: Using the total column densities listed in Table~\ref{t-metal}, we 
931: find that the total hydrogen column density (H\,$=$\,\hi$+$\hii$+{\rm H_2}$) is 
932: $\log N({\rm H}) = \log[N($\si$)+N($\sii$)+N($\Siii$)] - \log (Z_{\rm MB}/Z_\odot) - \log({\rm S/H}_\odot)
933: \simeq 20.2$ dex (where we estimate $\log N($\Siii$)= 13.63\,^{+0.11}_{-0.15}$ from 
934: the line at 1012.495 \AA\ using the AOD method and $N($\si$)$ is negligible).
935: \citet{lehner02} derived $\log N({\rm H}_2) = 15.45$ toward DI\,1388. The fraction 
936: of neutral gas is therefore $f($\hi$) = N($\hi$)/N({\rm H}) = 0.27$, 
937: while the fraction of molecular gas is 
938: $f({\rm H}_2) = 2 N({\rm H}_2)/N({\rm H}) = 3.6 \times 10^{-5}$.
939: Therefore, about 70\% of the gas toward DI\,1388 is ionized. As we discussed 
940: above the lower velocity Bridge gas is nearly fully ionized ($\sim$95\%) while 
941: the higher velocity Bridge is partially ionized ($\sim$53\%, assuming O is not
942: depleted into dust and the O abundance from Asplund et al. 2006).
943: 
944: The near absence of \nni\ and \ari\ in the low velocity gas of the Bridge toward DI\,1388
945: can be mostly explained by photoionization. Indeed if steady 
946: photionization dominates, \ari\  is deficient because the
947: photoionization cross section of Ar is about 10 times that of H over
948: a broad range of energy. In our Galaxy, \ari\   has  been found to 
949: be deficient with respect to \oi\  \citep{sofia98,jenkins00,lehner03}.
950: In partially ionized gas, N also behaves more like Ar than O \citep{jenkins00,lehner03}.
951: Ionization is unlikely to affect the higher velocity gas because the 
952: $N($\hi$)$ is large enough to shield the gas against ionizing photons.  
953: Since there are hot early-type stars in the Bridge, 
954: a potential source for ionization is the photoionization by these objects.
955: In \S\ref{sec-discuss-ion} we discuss further the possible sources of ionization.
956: We also note that nucleosynthesis may also lower the N abundance. Using 
957: the lower limit on \nii\ ($>14.19$ dex) and the H column density derived
958: above, $[{\rm N/H}] > -1.51$ (\niii\ is negligible, see Lehner 2002). It is 
959: therefore not clear if N is deficient or not, although it appears less deficient
960: than usually observed for gas with a metallicity of 0.09 solar \citep[][and references therein]{henry07}.  
961: 
962: 
963: Using the same arguments to that above for DGIK\,975 (and assuming a metallicity of 
964: $-1$ dex), but employing \feii\ 
965: and \feiii\ (where we estimate $\log N($\feiii$)= 14.29\pm 0.06$ from 
966: the line at 1122.524 \AA\ and assume that \feiii\ is not contaminated by the star) 
967: and correcting for the depletion of Fe (where we assume
968: it is $-0.6$ dex based on the similarity of the $[$\feii/\siii$]$ toward
969: both stellar sightlines and that the same depletion applies for 
970: \feii\ and \feiii), we find 
971: $\log N($H$) \sim 20.8$, which implies that $\sim$75--95\% of the gas is also 
972: ionized along this line of sight. 
973: No H$_2$ has been found  toward this sightline, implying 
974: $f[{\rm H}_2]\la  5\times 10^{-6}$.
975: As for DI\,1388, the low velocity component apppears 
976: more ionized than the higher velocity cloud (see above), but the Bridge 
977: gas toward DGIK\,975 appears even more ionized than along the DI\,1388 sightline. 
978: 
979: 
980: Toward PKS\,0312--77, we estimate via the AOD method
981: $\log N($\sii$) = 14.95 \pm 0.07$ using the \sii\ $\lambda$$\lambda$1250,1253 lines. 
982: From the \hi\ column density derived by \citet{kobulnicky99}, 
983: we find that  $[$\sii/\hi$] = -0.29 \pm 0.16$. 
984: Assuming that the metallicity is $-1$ dex along this sightline,
985: this implies again that about 70--85\% of the gas 
986: is ionized toward PKS\,0312--77. We note that this line 
987: of sight probes the whole depth of the Bridge and therefore 
988: shows that ionized gas appears significant everywhere. 
989: We also note that none of these estimates take into account the highly ionized
990: phase probed by \ovi, \siiv, and \civ\ absorption \citep[see][]{lehner01,lehner02}. 
991: The weakly and highly ionized gas  are unlikely to probe the same material, 
992: and therefore these ionized fractions should be considered as lower limits. 
993: 
994: 
995: \subsection{\cii*\ Diagnostic}
996: \cii*\ can be used to estimate the electron density and radiative cooling rate 
997: of the gas \citep[see e.g.][]{lehner04}. In ionized gas and cold neutral gas, 
998: the \cii\ radiation is a far more important coolant than in warm neutral gas \citep{wolfire95}.
999: Toward DI\,1388, \cii*\ is detected between 193 and 215 \km, i.e 
1000: in the partially ionized component, but is absent between 165 and 193 \km, 
1001: i.e. in the nearly fully ionized component  (see Fig.~\ref{fig-cii}).
1002: We  find $\log N($\cii*$) = 12.77 \pm 0.11$ with an average velocity of 
1003: $v_{\rm LSR} = 202.8 \pm 1.5$ \km and a $b$-value $b = 4.2\pm 1.1$ \km\ (obtained 
1004: from a single-component profile fit), which implies $T<2\times 10^4$ K. 
1005: It is therefore not clear if \cii*\ arises in some ionized gas at $T\sim 10^4$ K or
1006: in cold neutral gas where the turbulent motions may be important.  
1007: For the gas observed at $165 \le v_{\rm LSR} \le 193$ \km,  we derive 
1008: a 3$\sigma$ upper limit: $\log N($\cii*$) < 12.18$. 
1009: Toward DGIK\,975, the 3$\sigma$ upper limit is 13.4 and is not useful.
1010: 
1011: 
1012: 
1013: \begin{figure}[tbp]
1014: \epsscale{1.0} 
1015: \plotone{f7.eps}
1016: \caption{Normalized profile of \cii*\ against the LSR velocity. Note
1017: the absence of absorption between $\sim$160 and 190 \km\ where the nearly
1018: fully ionized component is observed. 
1019: \label{fig-cii}}
1020: \end{figure}
1021: 
1022: 
1023: If we assume that the electron collisions dominate the excitation of 
1024: \cii, 
1025: the electron density  within the Bridge can be written as  \citep[see][]{lehner04}:
1026: \begin{equation}\label{elredeqt}
1027: n_e  \simeq 0.183 \sqrt{T} \,\frac{N({\rm C}^{+*})}{N({\rm C}^{+})} \, {\rm cm}^{-3}\,.
1028: \end{equation}
1029: We assume a temperature of $T \sim 10^4$ K \citep[typically observed in ionized gas, e.g.,][]{howk06}.
1030: We use \sii\ as a proxy of \cii\ and assume that carbon is 0.25 dex depleted 
1031: \citep[][and using the updated Asplund et al. 2006's solar C abundance]{cardelli96}. 
1032: For gas at  $165 \le v_{\rm LSR} \le 193$ \km, we find $\log N($\sii$) = 13.94 \pm 0.08$
1033: and therefore  $n_e < 0.03$ cm$^{-3}$ (3$\sigma$) at $T = 10^4$ K. Since this gas is nearly
1034: fully ionized, $n_e \approx n_p$, and therefore the overall density of the ionized
1035: gas is quite low. From \sii, $\log N($H$) \sim 19.8$ (including \Siii\ would increase
1036: this column by at most 0.2 dex); the pathlength of probed ionized gas is $>$0.6 kpc. 
1037: Since the projected size of the Bridge is about 5 kpc, DI\,1388 is likely not 
1038: deeply embedded in the Bridge (this is consistent with the comparison of the column densities 
1039: from \hi\ emission and \hi\ absorption). For the gas at  $193 \le v_{\rm LSR} \le 215$ \km, we find  
1040: $n_e < 0.1$ cm$^{-3}$ at $T = 10^4$ K (here it is an upper limit because of the 
1041: \cii*\ absorption could possibly arise in cold gas). 
1042: 
1043: Since the H$\alpha$ intensity scales with the square of the density, it is not surprising 
1044: that most of the current H$\alpha$ observations of the Bridge outside dense \hii\ regions  yield no  
1045: H$\alpha$ detection down to $ I[{\rm H}\alpha]  < 0.5$--2 R \citep{putman03,muller07}. 
1046: \citet{lehner04} estimated for the Galactic warm ionized medium (WIM) the relationhip between \cii*\ 
1047: produced in the WIM and the intensity of H$\alpha$. Adapting their relationship (Eqt. 7 in their paper) 
1048: to the Bridge conditions,  $I[{\rm H}\alpha] \simeq 1.5 \times 10^{-13}\, N_{\rm WIM}($\cii*$)$ Rayleigh
1049: (where $N_{\rm WIM}($\cii*$)$ is in cm$^{-2}$). With our 3$\sigma$ upper limit $\log N($\cii*$) < 12.18$, 
1050: we derive $I[{\rm H}\alpha] < 0.2$ R, typically smaller than current limits from H$\alpha$
1051: observations. In the 
1052: partially ionized component $ I[{\rm H}\alpha]  < 0.9$ R. Future deep H$\alpha$ observations
1053: with the Wisconsin H-Alpha Mapper (WHAM) currently being installed in Australia should yield important 
1054: clues on  the faint, diffuse H$\alpha$ emission in the Bridge and other tidal structures around 
1055: the Magellanic Clouds. 
1056: 
1057: 
1058: The \cii\ radiative cooling rate is expressed as \citep[see][]{lehner04}:
1059: \begin{equation}\label{ecool2}
1060: l_c   = 2.89 \times 10^{-20}\frac{N({\rm C}^{+*})}{N({\rm H})}\, \,{\rm erg}\, {\rm s}^{-1}\,  ({\rm H})^{-1}\,.
1061: \end{equation}
1062: As discussed the \cii*\ absorption is only detected in the partially ionized gas. 
1063: Most of the \hi\ is contained in this component and therefore the \cii\ radiative cooling rate
1064: is $4.5 \times 10^{-27}$ erg\,s$^{-1}$\,(\hi)$^{-1}$. We can express this as the cooling 
1065: per nucleon if we use the amount of 
1066: \sii\ present in this component: we derive $l_c < 2.8 \times 10^{-27}$ erg\,s$^{-1}$\,(H)$^{-1}$
1067: (it is an upper limit because we neglected \Siii). 
1068: Since the  cooling rate is directly proportional to the metallicity of the gas for ionized
1069: material \citep[e.g.,][]{wolfire03}, 
1070: these cooling rates per metal are similar to the Galactic average rate within the 1$\sigma$ dispersion 
1071: derived by \citet{lehner04} at similar $N($\hi$)$ or $N({\rm H})$. Where we do not 
1072: detect \cii*, $l_c < 7 \times 10^{-28}$ erg\,s$^{-1}$\,(H)$^{-1}$, much smaller than 
1073: those observed in the Milky Way at low $N($\hi$)$. 
1074: 
1075: 
1076: \section{Discussion}\label{sec-discuss}
1077: In Table~\ref{t-prop}, we summarize the properties of the Bridge derived from this work. 
1078: The most recent estimate of the \hi\ mass of the Bridge was derived by \citet{bruns05}, and to estimate the mass of 
1079: H of the Bridge we assume that the large ionization fraction in the Bridge observed in 
1080: our 3 lines of sight is characteristic of the Bridge as a whole.  Since our sightlines probe very 
1081: different regions of the Bridge, the hypothesis that the ionization fraction 
1082: is large in most regions of the Bridge appears reasonable. This mass is  $\sim$5 times
1083: larger than usually estimated in current models \citep[e.g.][]{muller07a}.
1084: However, since the gas mass fraction of the SMC and LMC and the ratio of the gas disk 
1085: to the stellar disk of SMC can be adjusted in models, and since the SMC mass
1086: is not known to better than a factor two, it is possible that simulations 
1087: may be able to accommodate a larger Bridge mass. 
1088: 
1089: 
1090: 
1091: \begin{deluxetable}{lc}
1092: \tabcolsep=6pt
1093: \tablecolumns{2}
1094: \tablewidth{0pt} 
1095: %\tabletypesize{\footnotesize}
1096: \tablecaption{Summary  of the Properties in the Diffuse Gas of the Magellanic Bridge \label{t-prop}} 
1097: \tablehead{\colhead{Parameter}    &  \colhead{Value} }
1098: \startdata
1099: $[{\rm Z/H}]$$^a$		    & $-1.05 \pm 0.06 $  \\
1100: $[$\ovi$/$H$]$			    & $-3.8$ to $-3.3$  \\
1101: $[$\ovi$/$\hi$]$		    & $-3.1$ to $-2.3$  \\
1102: Depletion of Si			    & $-0.45 \pm 0.06$  \\
1103: Depletion of Fe			    & $-0.61 \pm 0.06$  \\
1104: Fraction of total \hi\		    & $\sim$20\%   \\
1105: Fraction of total \hii\		    & $\sim$80\%   \\
1106: Fraction of H$_2$		    & $\la 0.004$\% \\
1107: Density $n_e \simeq n_p$$^b$	    & $<0.03$--$0.1\sqrt{T_4}$  cm$^{-3}$ \\
1108: Total H mass$^c$		    & $\sim 9\times 10^8$ M$_\odot$
1109: \enddata
1110: \tablecomments{$a$: Average metallicity from the interstellar and B-type stellar estimates.
1111: $b$: Density of ionized gas. $T_4$ is the temperature in units of $10^4$ K. 
1112: $c$: Adopting a total \hi\ mass of $1.8 \times 10^8$ M$_\odot$ \citep{bruns05} 
1113: and assuming that the fraction of ionized is 80\% throughout the Bridge.   
1114: Note that the {\em total} mass of the Bridge should include the H$+$He mass. 
1115: }		
1116: \end{deluxetable}
1117: 
1118: 
1119: 
1120: \subsection{Sources of Ionization of the Magellanic Bridge}\label{sec-discuss-ion}
1121: One likely source of ionization of the Bridge is the stars within
1122: the Bridge. A young population of stars was discovered 
1123: throughout the Bridge \citep{irwin90,bica95,demers98}. The density of OB associations
1124: is, however, not uniform: the OB-type stars are highly concentrated 
1125: in the SMC Wing where the \hi\ column densities are the largest 
1126: (roughly delimited by the box in Fig.~\ref{fig-map}) and are far 
1127: more sparse farther away from the SMC \citep{irwin90,battinelli92}. The rectangle 
1128: shown in Fig.~\ref{fig-map} also corresponds to the CO survey undertaken 
1129: by \citet{mizuno06}, where CO emission was discovered, suggesting that
1130: conditions might be conducive to star formation.  We note that the star-formation
1131: history must be quite different from the less dense region of Bridge (see \S\ref{sec-per}), 
1132: since in the SMC Wing, an SMC-like metallicity is derived from the
1133: analysis of B-type stars \citep{lee05}.\footnote{There might be 
1134: some confusion in the literature with the nomenclature
1135: ``SMC Wing'' and ``Bridge'' (also called InterCloud Region -- ICR). In some cases, the
1136: Wing is an environment by itself (usually in the studies
1137: of stars), in other cases (usually in  the studies of CO and \hi\ emission) 
1138: it is a direct component of the Magellanic Bridge. The latter
1139: is possible and seems to be justified in view of the \hi\ column density and velocity maps. 
1140: But  the SMC Wing has a much higher concentration of OB associations \citep{irwin90}
1141: and an SMC-like metallicity \citep{lee05}, which seem to imply a different evolution from the
1142: Bridge at RA(J2000)\,$\ga 2^{\rm h}\,30^{\rm m}$ (see \S\ref{sec-per}). In particular, in view of these 
1143: differences, it does not seem to be justified to extrapolate the properties derived
1144: in the SMC Wing to the whole Bridge and vice versa, at least not before a better understanding
1145: of the connection between these two regions.} 
1146: Yet, the presence of early-type stars of age 20 Myr old or younger well inside and throughout the
1147: Bridge \citep{demers98,rolleston99} implies ongoing star-formation at some level. 
1148: 
1149: The H$\alpha$ intensity of a cloud can be used to estimate 
1150: the incident Lyman continuum flux. In our Galaxy, \citet{reynolds84,reynolds93} 
1151: has used this approach to calculate the power required to ionize
1152: the Galactic warm ionized medium up to 2--3 kpc above the plane, comparing it with the power available
1153: from OB-type stars. Following \citet{tufte98}, the incident Lyman continuum flux is
1154: $F_{\rm LC} = 2.1\times 10^5 I[{\rm H}\alpha]$ photons\,s$^{-1}$\,cm$^{-2}$, where 
1155: $I[{\rm H}\alpha]$  in Rayleigh is assumed to arise solely from photoionization. 
1156: Using the 3$\sigma$ upper limit derived from \cii*, we find for the diffuse ionized gas of 
1157: the Bridge (outside denser \hii\ regions around OB-type stars) 
1158: $F_{\rm LC}(\rm MB) < 0.4 \times 10^5$ photons\,s$^{-1}$\,cm$^{-2}$.
1159: This rate is at least 100 times smaller than the one found in 
1160: the Galaxy at high galactic latitudes \citep{reynolds84}. 
1161: This rate would increase by a factor four if we used 
1162: the \cii*\ detection but as discussed above the origin
1163: of the excitation is uncertain as both cold neutral and warm 
1164: ionized could contribute to the observed absorption. We therefore
1165: concentrate in this section on the origin of the nearly fully ionized
1166: gas. 
1167: 
1168: The flux of Lyman continuum photons from stars is highly dependent
1169: on their types \citep{panagia73}. For example, an O6-type star will produce
1170: a factor 10 more ionizing photons than an O9. But more importantly
1171: the rate of photons drops dramatically for B1 and later types: a B1 compared
1172: to an O9 produces 600 times fewer ionizing photons per second. In our Galaxy, 
1173: surveys of OB-type stars have shown that O-type stars produce 
1174: 93\% of ionizing photons, even though they are much less numerous than
1175: B-type stars \citep{terzian74,abbott82}. 
1176: 
1177: Unfortunately, little is known about the OB-type distribution in the Bridge, 
1178: especially outside the SMC Wing. \citet{irwin90} reported an O8 star outside
1179: the SMC Wing, and therefore there is likely to be O-type stars in the Bridge. 
1180: In the SMC Wing, \citet{pierre86} estimated that the number of ionizing
1181: photons over a very small area (0.1 square degree) is 
1182: $ 70 \times 10^6$ photons\,s$^{-1}$\,cm$^{-2}$. Such a value would only require
1183: 0.1\% escaping Lyman continuum photons from the SMC Wing to ionize the Bridge. 
1184: However, the sample of \citet{pierre86} is small and may be not be representative
1185: of the OB distribution in the Wing. In particular, 20\% of the stars in their sample with 
1186: types earlier than B0 are O7, which seems quite a larger number (in our Galaxy
1187: this number is about 4\%, e.g., Terzian 1974); the three O7 stars in their sample 
1188: contribute to the majority  of ionizing photons. A better characterization of 
1189: the spectral types of the stars in the SMC Wing and Bridge would clearly help 
1190: to better understand this important source of ionizing photons. 
1191: 
1192: 
1193: OB stars in the Bridge and Wing will also create an  environment that is favorable 
1194: for the escape of Lyman continuum photons and for sustaining a high level 
1195: of ionization throughout the Bridge. Indeed, OB associations through the winds
1196: and death of massive stars can produce large bubbles and chimneys that can help ionizing
1197: photons to travel large distances \citep[e.g.,][]{dove94,dove00,norman89}.
1198: Such \hi\ shells and even possibly blow-outs and chimneys were surveyed by
1199: \citet{muller03}, but only in the Wing at RA\,$< 3^{\rm h}$ and some 
1200: may be only due to projection effect \citet{muller07a}. 
1201: Furthermore, in an overall low density medium, it is likely that signatures of supernova
1202: remnants and stellar winds will be difficult to decipher since rather than forming 
1203: shells as observed in dense \hi\ regions, they might just dilute themselves in their surroundings.
1204: 
1205: Supernovae could also play a role in the ionization of the Bridge. 
1206: Since stellar formation has occured for  200--300 Myr \citep{harris07}, 
1207: it is likely that there have been several supernovae in and near the Bridge.
1208: Assuming steady ionization of H (with $\sim$13.6 eV per ionization), our limits
1209: on H$\alpha$ require $< 9 \times 10^{-7}$ erg\,s$^{-1}$\,cm$^{-2}$. Over
1210: the entire Bridge (assuming a thickness of 5 kpc), this corresponds to a constant
1211: power input to the gas of $<7\times 10^{38}$ erg\,s$^{-1}$. This power requires 
1212: a supernova rate of about one every 50,000 yr  \citep[see][]{abbott82}. The supernova
1213: rate is unknown and, as discussed by \citet{reynolds84}, only a small fraction ($\sim 17$\%) of the kinetic 
1214: energy injected in the interstellar medium may be converted into ionizing hydrogen. 
1215: Yet the presence of young, massive stars, supernovae have likely contributed to the ionization
1216: and heating of the Bridge. 
1217: 
1218: Highly ionized gas found in the Bridge \citep{lehner01,lehner02} is  an indirect signature 
1219: of hot gas that is still present in significant amounts (e.g. where the observed high ions are in an 
1220: interface between the hot and cooler gas) or that was present and is in the process of
1221: cooling. Both interfaces and cooling of hot gas can produce photons that may be important 
1222: sources of photoionization of the environment \citep{slavin00,borkowski00,knauth03}. 
1223: For example, \citet{slavin00} modeled the contribution to the photoionization from 
1224: cooling of hot supernova remnants in the disk of our Galaxy and found that this source was adequate 
1225: to account for the observed ionization of the Galactic halo. Therefore supernovae in the
1226: Bridge and in the SMC and LMC near the Bridge could not only inject energy and heat the Bridge
1227: but also their cooling remnants may provide a source of ionization. 
1228: 
1229: While the escaping of ionizing photons from the SMC Wing could
1230: be an important source of ionization, 
1231: the escape of ionizing photons from the Galactic, LMC, and SMC disks \citep[see][]{bm99,dove00} 
1232: is also likely to contribute to the ionization of the Bridge.
1233: Figure~8 in \citet{putman03} shows predicted $I[{\rm H}\alpha]$ near the LMC 
1234: caused by the escape of ionizing photons from the LMC and the Galactic stellar bulge. 
1235: The H$\alpha$ emission estimate produced from the leaking photons of the LMC (and Galaxy, 
1236: but the latter is rather negligible) could reach 0.1 to  0.25 R near the LMC. 
1237: Thus, such leakage alone is likely an  important source of ionizing photons for 
1238: the Bridge.
1239: 
1240: While it is not clear which ionizing source is 
1241: dominant in the Bridge, there appear to be enough photons to maintain
1242: the high level of ionization seen in the Bridge. The presence of massive stars in the rarefied environment
1243: of the Bridge is likely sufficient in itself for this environment to be 
1244: largely ionized. Since cooling and recombination times are 
1245: expected to be longer than in the Galactic environment because the Bridge gas has 
1246: an extremely low metallicity and density, the ionization is likely to be sustainable for 
1247: a long time. For example, for a $10^6$ K gas with a density $10^{-3}$ cm$^{-3}$, the radiative cooling is 
1248: $t_{\rm cool} \sim 2$--$3\times 10^9$ yr for a 0.1 solar metallicity \citep{gnat07}, 
1249: about 10 times longer than for a solar metallicity environment.  
1250: Furthermore, according to \citet{harris07}, star formation has occured in the Bridge for 
1251: the last 200--300 Myr, with a lower rate over the last $\sim$40 Myr. 
1252: Therefore, several generations of OB stars could have ionized
1253: the Bridge over a long time period. 
1254: 
1255: 
1256: We finally note that the ionization fraction toward DGIK\,975 is not only larger than 
1257: toward DI\,1388, but the gas is also more highly-ionized. \citet{lehner02}
1258: reported the measurements of \ovi, \siiv, and \civ\ toward these stars. In particular, 
1259: $N_{\rm DGIK\,975}($\ovi$) \simeq 13 N_{\rm DI\,1388}($\ovi$)$, while 
1260: $N_{\rm DGIK\,975}($\siii$) \simeq 2 N_{\rm DI\,1388}($\siii$)$. 
1261: One can refer to Fig.~11 in \citet{lehner02} to see the dramatic
1262: difference in the \ovi\ absorption profiles toward DI\,1388 and DGIK\,975. 
1263: Recently, \citet{lehner07} argued for the presence of outflows
1264: and even possibly a hot galactic wind from the LMC. 
1265: Since DGIK\,975 is in the outskirts of the LMC and DI\,1388 is
1266: farther away from the LMC (see Fig.~\ref{fig-map}), 
1267: the larger amount of ambient hot gas toward DGIK\,975 might be related to 
1268: the ouflows from the LMC. Perhaps the feedback processes occurring in the 
1269: LMC inject energy into the Bridge. We note that the amount of metals incorporated into the 
1270: Bridge gas from the LMC at large distance must be extremely small 
1271: since neither the interstellar nor stellar abundances 
1272: \citep[$-1.2 \pm 0.2$ dex solar for DGIK\,975, see][]{rolleston99,lee05} suggest an increase of 
1273: the metallicity in this region.  
1274: 
1275: 
1276: 
1277: \subsection{Metallicity of the Magellanic Bridge}
1278: 
1279: The metallicity of the Bridge is difficult to accommodate
1280: with most models of the SMC-LMC-Galaxy interactions (but see below). 
1281: The present-day low metallicity of the Bridge has, however, a simple explanation 
1282: in the context of the chemical evolution of the SMC
1283: if the Bridge is much older than 200 Myr. Indeed the bursting model for 
1284: the star formation rate in the SMC described by \citet{pagel98} 
1285: suggests that its metallicity only slowly increased from $-1.4$ to $-1.1$
1286: dex between about 12 and 2.5 Gyr ago. In this model \citep[see also][]{harris04,idiart07}, 
1287: it is only in the last $\sim$2.5 Gyr that the 
1288: SMC metallicity has exceeded the modern Bridge value.
1289: This is a large window for forming the Bridge from SMC gas, and suggests that the
1290: Magellanic Bridge could be 10 times older than current  N-body simulations suggest.
1291: \citet{harris07} showed that there were several bursts (2.5, 0.4, 0.06 Gyr ago) 
1292: of star formation in the SMC in the last  $\sim$3 Gyr, but the most important one being 2.5 Gyr ago. 
1293: New bursts of star-formation are believed to be often caused by the effects of close 
1294: tidal interactions between galaxies \citep[e.g.,][]{larson78,barton07}. 
1295: While \citet{harris07} noted that the bursts at 2.5 Gyr and 0.4 Gyr are temporally
1296: coincident with past perigalactic passages of the SMC and Galaxy, these epochs also 
1297: correspond to close encounters between the SMC and LMC \citep{kallivayalil06}. 
1298: Due to its proximity, the LMC tidal perturbations on the SMC are larger
1299: than those of the Galaxy \citep{lin95}. However, calculations of the orbits
1300: of the SMC and LMC with the updated proper motions of the SMC and 
1301: LMC still predict two more perigalactic approaches between the SMC and LMC since the 2.2 Gyr close encounter 
1302: at $\sim$1.4, 0.2 Gyr, the latter being the closest approach \citep{kallivayalil06}. 
1303: It is therefore unclear if the Bridge could be stable over 2--3 Gyr
1304: (Gardiner \& Noguchi 1994 argued it could not be stable over a very long time 
1305: but principally because of the Galaxy's gravitational field; yet they did not quantify it further). 
1306: 
1307: On the other hand, \citet{gardiner96} predicted from their model that the Bridge gas originated 
1308: primarily from the halo of the SMC, not its disk. Under the assumption that the tidal
1309: models are correct, this would mean that despite the steep increase of star
1310: formation rate in the SMC disk, the metallicity of the SMC halo would not have been 
1311: enhanced in the last 2.5 Gyr by any disk material via outflows or else was diluted 
1312: by an external low-metallicity source.
1313: Assuming that the initial mass of the Bridge is $M_{\rm MB}^i$ and metallicity
1314: is $Z_{\rm MB}^i$, and a ``diluting'' gas with  $M_{\rm dil}$ and $Z_{\rm dil}$, 
1315: the  metallicity of the mixed-up gas in the Bridge, $Z_{\rm MB}^f$, can be expressed as: 
1316: \begin{equation}
1317: Z_{\rm MB}^f = \frac{Z_{\rm MB}^i M_{\rm MB}^i + Z_{\rm dil} M_{\rm dil}}{M_{\rm MB}^i + M_{\rm dil}}\,;
1318: \end{equation}
1319: and assuming that $Z_{\rm MB}^i M_{\rm MB}^i \gg Z_{\rm dil} M_{\rm dil}$, we have
1320: \begin{equation}
1321: Z_{\rm MB}^f = Z_{\rm MB}^i \left(1+ \frac{M_{\rm dil}}{M_{\rm MB}^i}\right)^{-1}\,. 
1322: \end{equation}
1323: Therefore, if the metallicity of the tidally disrupted material was present-day SMC like,  
1324: $Z_{\rm MB}^i = 0.25 Z_\odot$, one would need  $M_{\rm dil} \approx 2 M_{\rm MB}^i$ 
1325: (with $Z_{\rm dil} \la 0.01 Z_\odot$) to achieve $Z_{\rm MB}^f \approx 0.09 Z_\odot$. In contrast,  
1326: $Z_{\rm MB}^i M_{\rm MB}^i \ll Z_{\rm dil} M_{\rm dil}$, $Z_{\rm dil}^i = 0.09 Z_\odot$
1327: and $Z_{\rm MB}^i = 0.25 Z_\odot$ would require $M_{\rm dil} \ga 200 M_{\rm MB}^i$ 
1328: to have $Z_{\rm MB}^f = 0.09 Z_\odot$. 
1329: A determination of the ratio the SMC halo to SMC disk particles that were pulled into 
1330: the Bridge in simulations would be valuable. This scenario might be plausible since
1331: dwarf galaxies can have the \hi\ gas occupying radii quite larger than the 
1332: stellar component \citep{salpeter96}. This scenario was also proposed 
1333: by \citet{harris07} to explain the absence of evidence of tidally stripped 
1334: stars in the Bridge. If this envelope was extended enough, it could have reduced 
1335: the impact of feedbacks from massive stars in the SMC at large 
1336: distances from the SMC for $\sim$2--3 Gyr.  
1337: 
1338: If the very low metallicity material did not come from an extended halo of gas around 
1339: the SMC stellar disk (and assuming the Bridge is young),  the Bridge could
1340: have swept up metal-poor ambient matter over time. For example, 
1341: measurements of the proper motion of the Fornax dwarf spheroidal galaxy suggest that the orbit 
1342: of the Magellanic Clouds was crossed by Fornax some 200 Myr ago \citep{dinescu04},
1343: a time coincidentally similar to the believed formation epoch of the Bridge.
1344: The Fornax has a wide range of metallicity from $-2.0$ to $-0.4$ dex solar, 
1345: and peak near $-0.9$ dex \citep{pont04}. While this is speculative, low-metallicity 
1346: material from the Fornax or other sources may have mixed-up with Bridge gas. 
1347: 
1348: Bekki \& Chiba (2007a,2007b) have recently produced self-consistent chemodynamical 
1349: models of the LMC-SMC-Galaxy interactions and found
1350: a metallicity distribution that peaks at $-0.8$ dex but could extend down 
1351: to $-0.9$ dex. To the best of our knowledge, this is the only model 
1352: that attempts to explain the lower metallicity of Bridge. In this model, a low metallicity
1353: is reached because the Bridge is mostly formed from the SMC halo gas, where the metallicity
1354: is significantly smaller because of an assumed metallicity gradient in the SMC,
1355: although it is not clear that there is or has been such a gradient since
1356: the present-day metallicity in the SMC Wing \citep{lee05} is similar to the SMC. 
1357: 
1358: \subsection{Concluding Remarks}\label{sec-per}
1359: Our analysis provides the first metallicity of the Magellanic Bridge gas
1360: and shows for the first time that the Bridge is 
1361: substantially ionized. As our sample of sightlines is small, it is, however, not  
1362: certain if large chemical and/or ionization inhomogeneties exist. 
1363: As we argued above, our three sightlines probe the low \hi\ column density
1364: regions of the Bridge (RA\,$\ga 3^{\rm h}$) at very different locations and 
1365: depths. The present-day metallicity of the Bridge derived from 3 B-type 
1366: stars (including DGIK\,975) also suggests some chemical homogeneity \citep{rolleston99,lee05}. 
1367: We note that two of the Rolleston et al. stars are close to the SMC Wing, but
1368: outside the high \hi\ column density regions. In contrast, \citet{lee05}
1369: found an SMC-like metallicity for 4 B-type stars well inside the 
1370: SMC Wing (and in the high \hi\ column density region of the Wing). 
1371: While the samples are small in each region, it seems very unlikely
1372: that current samples probe each an unrepresentative population. 
1373: 
1374: The difference of 0.4--0.5 dex in the metallicity between these two adjacent
1375: regions is another puzzle, although we note that \citet{lee05} quoted
1376: errors of $\pm0.3$ dex in their abundances and the difference
1377: may actually be smaller. These authors also used the same
1378: method for the SMC Wing and Bridge stars and found a systematic 
1379: difference of about 0.5 dex in the abundance of these stars. 
1380: This difference may be explained (assuming chemical inhomogeneities 
1381: in each region are small) if the Wing and the low \hi\ column density
1382: regions of the Bridge have a different origin or  the star-formation 
1383: rate and initial-mass function were quite different in each region, 
1384: and/or the Wing was not diluted with extremely-low metallicity gas. In the 
1385: former scenario, the Wing must be younger than the rest of the Bridge, so that it was
1386: formed recently from gas and stars stripped from the SMC that have a present-day SMC
1387: metallicity. This would imply that the Bridge itself must be much older
1388: and that the star-formation was  inefficient over a long period 
1389: of time or only started in the last 200--300 Myr as the results of 
1390: \citet{harris07} suggest. In the second scenario, at the epoch the 
1391: Bridge and the Wing were formed, the metallicity should have been 
1392: around  $\la -1.1$ dex in both regions. However, it would seem very unlikely that 
1393: the metallicity of the Wing has increased by 0.5 dex in the last 200 Myr assuming 
1394: this is the age of these structures. In the SMC where star formation is more
1395: important than in the Wing, about 2--3 Gyr were needed 
1396: to increase the metallicity from $-1.1$ dex to the present-day metallicity \citep{pagel98}.
1397: Therefore the second scenario is not likely if these structures are only 200 Myr old. 
1398: In the third scenario, the Bridge outside the Wing regions could have
1399: been diluted with low metallicity gas, while the Wing was not or at a 
1400: neglible level, possibly because the star-formation was more efficient or the
1401: Wing material comes from deeper regions of  the SMC (where the gas has present-day SMC
1402: metallicity) than the rest of the Bridge gas. 
1403: The first and third scenarios are consistent with the findings of evolved
1404: stars in the Wing and not anywhere else in the Bridge \citep{harris07}. 
1405: 
1406: Hence despite the fact that current tidal simulations reproduce well
1407: the observed \hi\ structures (i.e. the Bridge, Stream, and Leading Arm), they seem
1408: to be in conflict with the above conclusions regarding the metallicities.  
1409: As already mentioned, the velocities of the LMC and SMC adopted in the models 
1410: of \citet{gardiner96} and subsequent models differ from those derived from the recent
1411: proper motion estimates of the SMC and LMC by  over 100 \km\ \citep{kallivayalil06,kallivayalil06b}. 
1412: It is also unclear how bound the SMC and LMC are and were \citep{kallivayalil06}. 
1413: With those updated proper motions, \citet{besla07} strongly suggest that existing numerical models
1414: of the Clouds may no longer be appropriate, and in particular cannot explain anymore
1415: the Stream and Leading Arm.  As discussed 
1416: by \citet{nidever07}, current simulations may also miss important physics, including 
1417: feedback processes such as outflows from the LMC that may interact with the Bridge gas. 
1418: In particular, it is quite important to understand 
1419: feedback and mixing processes in the SMC disk and halo for the last 3 Gyr 
1420: if the Bridge is only 200 Myr old and its main origin is the halo of the SMC. 
1421: 
1422: Future theoretical investigations using the updated proper motions of the LMC 
1423: and SMC should not only try to reproduce the \hi\ properties but also accommodate the low metallicity of 
1424: the Bridge, the apparent discrepancy of the metallicity between the Bridge and the
1425: SMC Wing, its ionization structure, and the apparent lack 
1426: of old stars in the Bridge \citep{harris07}. Future models should in 
1427: particular investigage if the Bridge can be stable over a long time (2--3 Gyr) or despite an
1428: important burst of star formation within the SMC 2.5 Gyr ago, the SMC halo gas
1429: could maintain a very low metallicity. If none of these conditions are 
1430: satisfied, an important (depending on the chemical homogeneity in the Bridge)
1431: amount of matter in the Bridge must come from somewhere else. If the Bridge
1432: could survive several perigalactic approaches between the SMC and LMC, it
1433: would be interesting to know if the low density and low metallicity regions of the Bridge 
1434: could have been formed some 2--3 Gyr ago, while the SMC Wing would have been 
1435: produced at a latter time, some 200 Myr ago during the last and closest approach 
1436: between the SMC and LMC. 
1437: 
1438: From an observational point of view, a larger 
1439: stellar and interstellar sample where the abundances can be estimated in 
1440: the Bridge and the SMC Wing would be extremely valuable. The soon to 
1441: be installed Cosmic Origins Spectrograph (COS)  will be particularly useful 
1442: for the study of the gas-phases in the Bridge since fainter targets 
1443: can be observed with it. In particular, to better characterize the ionization 
1444: structure, future COS and deep H$\alpha$ observations will be invaluable. A better
1445: characterization of the stellar population in the Bridge will also help to 
1446: constrain the origin of the ionization, the star-formation history, and the inital-mass 
1447: function in the Bridge. Since the Bridge is the closest gaseous and stellar tidal remnant, and 
1448: since interactions between galaxies are rather common, future observational and 
1449: theoretical efforts offer the unique opportunity to comprehend the
1450: physical processes and origin of such structures. 
1451: 
1452: 
1453: \section{Summary}\label{sec-sum}
1454: Using far-UV spectra obtained with {\fuse} and {\em HST}/STIS E140M,
1455: we analyse the physical properties and abundances of the Magellanic Bridge gas 
1456: toward three sightlines that are situated at different locations in the Bridge
1457: and probe various depths. These observations provide access to neutral and 
1458: ionized species at sufficient signal-to-noise and resolution to estimate their 
1459: column densities and kinematics. In Table~\ref{t-prop} we summarize the 
1460: abundances and physical properties of the Magellanic Bridge determined from 
1461: our investigation.  Our main findings are summarized as follows:
1462: 
1463: \noindent
1464: 1. Toward one sigthline, we find that the Magellanic Bridge metallicity is 
1465: $[{\rm Z/H}] = -1.02 \pm 0.07$, and toward another 
1466: one $-1.7 < [{\rm Z/H}] < -0.9$.  To derive these quantities we compare 
1467: the column densities of neutral elements (\oi, \ari, \nni) to that of \hi. 
1468: The metallicity of the gas in the Bridge is in excellent agreement with the average metallicity
1469: determined in B-type stars. There is some evidence that N might be less
1470: deficient than usually observed in gas with $[{\rm Z/H}] \approx -1$ and 
1471: even possibly not deficient with respect to O. 
1472: 
1473: \noindent
1474: 2. The very low present-day metallicity in the Bridge
1475: is similar to the SMC before a burst of star formation that occured
1476: about 2.5 Gyr ago. This may not only be a pure coincidence since
1477: interaction between galaxies are believed to create bursts of star
1478: formation within the interacting galaxies. Yet, it is unclear at this 
1479: time if the Bridge could survive subsequent perigalactic passages of 
1480: the LMC with the SMC. Hence the Bridge could be much younger as currently
1481: predicted by tidal models. In this case, 
1482: it would require a high level of dilution with an unidentified component with 
1483: extremely low metallicity; this component must be dominant in order to reach
1484: the present-day Bridge metallicity.
1485: 
1486: \noindent
1487: 3. We determine that the gas of the Bridge is largely ionized toward our 3 sightlines, 
1488: with only $\sim$20\% of the gas being neutral, implying that the largest fraction of the gas mass 
1489: of the Bridge comes from the ionized gas. Toward two sightlines, we find that there are at least two main 
1490: components in absorption, and the component with the lower velocity is systematically more ionized. 
1491: We argue that possible sources for the ionization are the hot stars within and near the Bridge, 
1492: hot gas (indirectly detected via \ovi\ absorption), and photons leaking from the SMC, LMC, 
1493: and Milky Way. 
1494: 
1495: \noindent
1496: 4. From the analysis of \cii*, we find $n_e < 0.03 \sqrt{T_4}$ cm$^{-3}$ in the
1497: nearly fully ionized gas and $n_e < 0.1 \sqrt{T_4}$ cm$^{-3}$ in the
1498: partially ionized gas. Since the gas is dominantly 
1499: ionized, our analysis suggests  that  the overall density of the Bridge gas is extremely 
1500: low. This is consistent with the absence of detection of H$\alpha$ emission in 
1501: the diffuse gas of the Bridge with current observations. 
1502: Denser and less dense regions must also exist in the 
1503: Bridge on account of the multiphase (cooler and hotter gas) nature of the Bridge. 
1504: 
1505: 
1506: \acknowledgments
1507: 
1508: We thank the anonymous referee for constructive comments that improved
1509: and strengthened the content of our manuscript. As this paper 
1510: was submitted,  {\em FUSE}\ abruptely failed permanently.
1511: We want to express our gratitude to the {\em FUSE}\  science, hardware, and mission-operation 
1512: teams for providing high-quality FUV spectroscopic data for the last 8 years. 
1513: Support for this research was provided by NASA through grants 
1514: FUSE-NNX06F42G and FUSE-NNX07AK09G. F.P.K. is grateful to AWE Aldermaston for the award 
1515: of a William Penney Fellowship. This research has made use of the NASA
1516: Astrophysics Data System Abstract Service and the Centre de Donn\'ees de Strasbourg (CDS).
1517: 
1518: 
1519: \begin{thebibliography}{}
1520: 
1521: \bibitem[Abbott(1982)]{abbott82} 
1522: 	Abbott, D.~C.\ 1982, \apj, 263, 
1523: 
1524: \bibitem[Aguirre et al.(2001)]{aguirre01} 
1525: 	Aguirre, A., Hernquist, L., Schaye, J., Katz, N., Weinberg, D.~H., \& 
1526: 	Gardner, J.\ 2001, \apj, 561, 521 
1527: 
1528: \bibitem[Asplund et al.(2006)]{asplund06} 
1529: 	Asplund, M., Grevesse, N., \& Sauval, A.~J.\ 2006, Communications in Asteroseismology, 147, 76 
1530: 
1531: \bibitem[Bajaja et al.(2005)]{bajaja05} 
1532: 	Bajaja, E., Arnal, E.~M., Larrarte, J.~J., Morras, R., P{\"o}ppel, W.~G.~L., \& Kalberla, 
1533: 	P.~M.~W.\ 2005, \aap, 440, 767 
1534: 
1535: \bibitem[Barnes \& Hernquist(1992)]{barnes92} 
1536: 	Barnes, J.~E., \& Hernquist, L.\ 1992, \araa, 30, 705 
1537: 
1538: \bibitem[Barton et al.(2007)]{barton07} 
1539: 	Barton, E.~J., Arnold, J.~A., Zentner, A.~R., Bullock, J.~S., \& Wechsler, R.~H.\ 2007, \apj, in press
1540: 	 (arXiv:0708.2912)
1541: 
1542: \bibitem[Battinelli \& Demers(1992)]{battinelli92} 
1543: 	Battinelli, P., \& Demers, S.\ 1992, \aj, 104, 1458 
1544: 
1545: \bibitem[Bekki \& Chiba(2007a)]{bekki07a} 
1546: 	Bekki, K., \& Chiba, M.\ 2007a, PASA, 24, 21 
1547: 
1548: \bibitem[Bekki \& Chiba(2007b)]{bekki07b} 
1549: 	Bekki, K., \& Chiba, M.\ 2007b, MNRAS, 381, 16
1550: 
1551: \bibitem[Besla et al.(2007)]{besla07} 
1552: 	Besla, G., Kallivayalil, N., Hernquist, L., Robertson, B., 
1553: 	Cox, T.~J., van der Marel, R.~P., \& Alcock, C.\ 2007, \apj, 668, 949 
1554: 
1555: \bibitem[Bica \& Schmitt(1995)]{bica95} 
1556: 	Bica, E.~L.~D., \& Schmitt, H.~R.\ 1995, \apjs, 101, 41 
1557: 
1558: \bibitem[Bland-Hawthorn \& Maloney(1999)]{bm99} 
1559: 	Bland-Hawthorn, J., \& Maloney, P.~R.\ 1999, \apjl, 510, L33 
1560: 
1561: \bibitem[Borkowski et al.(1990)]{borkowski00} 
1562: 	Borkowski, K.~J., Balbus, S.~A., \& Fristrom, C.~C.\ 1990, \apj, 355, 501 
1563: 
1564: \bibitem[Br{\"u}ns et al.(2005)]{bruns05} 
1565: 	Br{\"u}ns, C., et al.\ 2005, \aap, 432, 45 
1566: 
1567: \bibitem[Cardelli et al.(1996)]{cardelli96} 
1568: 	Cardelli, J.~A., Meyer, D.~M., Jura, M., \& Savage, B.~D.\ 1996, \apj, 467, 334 
1569: 	
1570: \bibitem[Demers \& Battinelli(1998)]{demers98}
1571: 	Demers, S., \& Battinelli, P. 1998 \aj, 115, 154
1572: 
1573: \bibitem[de Mello et al.(2007)]{demello07}
1574: 	De Mello, D.~F., Smith, L.~J., Sabbi, E., Gallagher, J.S., Mountain, M.,
1575: 	\& Harbeck, D.~R. \ 2007, \aj, in press [arXiv:0711.2685]
1576: 	
1577: \bibitem[Dinescu et al.(2004)]{dinescu04} 
1578: 	Dinescu, D.~I., Keeney, B.~A., Majewski, S.~R., \& Girard, T.~M.\ 2004, \aj, 128, 687 
1579: 
1580: \bibitem[Diplas \& Savage(1991)]{diplas91} 
1581: 	Diplas, A., \& Savage, B.~D.\ 1991, \apj, 377, 126 
1582: 
1583: \bibitem[Dixon et al.(2007)]{dixon07} 
1584: 	Dixon, W.~V., et al.\ 2007, \pasp, 119, 527 
1585: 
1586: \bibitem[Dove \& Shull(1994)]{dove94} 
1587: 	Dove, J.~B., \& Shull, J.~M.\ 1994, \apj, 430, 222 
1588: 
1589: \bibitem[Dove et al.(2000)]{dove00} 
1590: 	Dove, J.~B., Shull, J.~M., \& Ferrara, A.\ 2000, \apj, 531, 846 
1591: 
1592: \bibitem[Gardiner \& Noguchi(1996)]{gardiner96}  
1593: 	Gardiner, L. T., \& Noguchi, M. 1996, \mnras, 278, 191
1594: 
1595: \bibitem[Gnat \& Sternberg(2007)]{gnat07} 
1596: 	Gnat, O., \& Sternberg, A.\ 2007, \apjs, 168, 213 
1597: 
1598: \bibitem[Grevesse \& Sauval(1998)]{grevesse98} 
1599: 	Grevesse, N., \& Sauval, A.~J.\ 1998, Space Science Reviews, 85, 161 
1600: 
1601: \bibitem[Harris(2007)]{harris07} 
1602: 	Harris, J.\ 2007, \apj, 658, 345
1603: 	
1604: \bibitem[Harris \& Zaritsky(2004)]{harris04} 
1605: 	Harris, J., \& Zaritsky, D.\ 2004, \aj, 127, 1531 
1606: 
1607: \bibitem[Henry \& Prochaska(2007)]{henry07} 
1608: 	Henry, R.~B.~C., \& Prochaska, J.~X.\ 2007,\pasp, 119, 962
1609: 
1610: \bibitem[Howk et al.(1999)]{howk99} 
1611: 	Howk, J.~C., Savage, B.~D., \& Fabian, D.\ 1999, \apj, 525, 253 
1612: 
1613: \bibitem[Howk et al.(2006)]{howk06} 
1614: 	Howk, J.~C., Sembach, K.~R., \& Savage, B.~D.\ 2006, \apj, 637, 333 
1615: 
1616: \bibitem[Idiart et al.(2007)]{idiart07} 
1617: 	Idiart, T.~P., Maciel, W.~J., \& Costa, R.~D.~D.\ 2007, \aap, 472, 101 
1618: 
1619: \bibitem[Irwin, Demers, \& Kunkel(1990)]{irwin90} 
1620: 	Irwin, M. J., Demers, S., Kunkel, W. E. 1990, \aj, 99, 191
1621: 
1622: \bibitem[{Jenkins} {et al.}(2000)]{jenkins00}
1623:          Jenkins, E. B., et al. 2000, \apj, 538, L81
1624: 
1625: \bibitem[Kalberla et al.(2005)]{kalberla05} 
1626: 	Kalberla, P.~M.~W., Burton, W.~B., Hartmann, D., Arnal, E.~M., Bajaja, E., 
1627: 	Morras, R., Poumlppel, W.~G.~L.\ 2005, \aap, 440, 775 
1628: 
1629: \bibitem[Kallivayalil et al.(2006a)]{kallivayalil06} 
1630: 	Kallivayalil, N., van der Marel, R.~P., \& Alcock, C.\ 2006a, \apj, 652, 1213 
1631: 
1632: \bibitem[Kallivayalil et al.(2006b)]{kallivayalil06b} 
1633: 	Kallivayalil, N., van der Marel, R.~P., Alcock, C., Axelrod, T., Cook, K.~H., Drake, A.~J., 
1634: 	\& Geha, M.\ 2006b, \apj, 638, 772 
1635: 
1636: \bibitem[Knauth et al.(2003)]{knauth03} 
1637: 	Knauth, D.~C., Howk, J.~C., Sembach, K.~R., Lauroesch, J.~T., \& Meyer, D.~M.\ 2003, \apj, 592, 964 
1638: 
1639: \bibitem[Kobulnicky \& Dickey(1999)]{kobulnicky99} 
1640: 	Kobulnicky, H. A., \& Dickey, J. M. 1999, \aj, 117, 908
1641: 
1642: \bibitem[Larson \& Tinsley(1978)]{larson78} 
1643: 	Larson, R.~B., \& Tinsley, B.~M.\ 1978, \apj, 219, 46 
1644: 	
1645: \bibitem[Lee et al.(2005)]{lee05} 
1646: 	Lee, J.-K., Rolleston, W.~R.~J., Dufton, P.~L., \& Ryans, R.~S.~I.\ 2005, \aap, 429, 1025 
1647: 
1648: \bibitem[Lehner(2002)]{lehner02} 
1649: 	Lehner, N.\ 2002, \apj, 578, 126 
1650: 
1651: \bibitem[Lehner \& Howk(2007)]{lehner07} 
1652: 	Lehner, N., \& Howk, J.~C.\ 2007, \mnras, 377, 687 
1653: 	
1654: \bibitem[Lehner et al.(2003)]{lehner03} 
1655: 	Lehner, N., Jenkins, E.~B., Gry, C., Moos, H.~W., Chayer, P., \& Lacour, S.\ 2003, \apj, 595, 858 
1656: 
1657: \bibitem[Lehner et al.(2001)]{lehner01} 
1658: 	Lehner, N.,  Sembach, K. R., Dufton, P. L., Rolleston, W. J. R., \& Keenan, F. P. 2001, \apj, 551, 781 
1659: 
1660: \bibitem[Lehner et al.(2004)]{lehner04} 
1661: 	Lehner, N., Wakker, B.~P., \& Savage, B.~D.\ 2004, \apj, 615, 767 
1662: 
1663: \bibitem[Lin et al.(1995)]{lin95} 
1664: 	Lin, D.~N.~C., Jones, B.~F., \& Klemola, A.~R.\ 1995, \apj, 439, 652 
1665: 
1666: \bibitem[Maller et al.(2006)]{maller06} 
1667: 	Maller, A.~H., Katz, N., Kere{\v s}, D., Dav{\'e}, R., \& Weinberg, D.~H.\ 2006, \apj, 647, 763 
1668: 
1669: \bibitem[Mathewson et al.(1974)]{mathewson74} 
1670: 	Mathewson, D.~S., Cleary, M.~N., \& Murray, J.~D.\ 1974, \apj, 190, 291 
1671: 
1672: \bibitem[Meyer et al.(1997)]{meyer97} 
1673: 	Meyer, D.~M., Cardelli, J.~A., \& Sofia, U.~J.\ 1997, \apjl, 490, L103 
1674: 
1675: \bibitem[{Meyer} {et al.}(1998)]{meyer98}
1676:          Meyer, D. M., Jura, M., \& Cardelli, J. A. 1998, \apj, 493, 222
1677: 	 
1678: \bibitem[Mizuno et al.(2006)]{mizuno06} 
1679: 	Mizuno, N., Muller, E., Maeda, H., Kawamura, A., Minamidani, T., Onishi, T., Mizuno, A., \& Fukui, 
1680: 	Y.\ 2006, \apjl, 643, L107 
1681: 
1682: \bibitem[Morton(2003)]{morton03} 
1683: 	Morton, D.~C.\ 2003, \apjs, 149, 205 
1684: 
1685: \bibitem[Muller \& Bekki(2007)]{muller07a} 
1686: 	Muller, E., \& Bekki, E.\ 2007, MNRAS, 381, L11
1687: 
1688: \bibitem[Muller \& Parker(2007)]{muller07} 
1689: 	Muller, E., \& Parker, Q.\ 2007, PASA, 24, 69
1690: 
1691: \bibitem[Muller et al.(2003)]{muller03} 
1692: 	Muller, E., Staveley-Smith, L., \& Zealey, W.~J.\ 2003, \mnras, 338, 609 
1693: 	
1694: \bibitem[Muller et al.(2004)]{muller04} 
1695: 	Muller, E., Stanimirovi{\'c}, S., Rosolowsky, E., \& Staveley-Smith, L.\ 2004, \apj, 
1696: 	616, 845 
1697: 
1698: \bibitem[Nidever et al.(2007)]{nidever07} 
1699: 	Nidever, D.~L., Majewski, S.~R., \&  Burton, W. B. \ 2007, \apj, submitted, arXiv:0706.1578 
1700: 
1701: \bibitem[Norman \& Ikeuchi(1989)]{norman89} 
1702: 	Norman, C.~A., \& Ikeuchi, S.\ 1989, \apj, 345, 372 
1703: 
1704: \bibitem[Oliveira et al.(2005)]{oliveira05} 
1705: 	Oliveira, C.~M., Dupuis, J., Chayer, P., \& Moos, H.~W.\ 2005, \apj, 625, 232 
1706: 
1707: \bibitem[Pagel \& Tautvaisiene(1998)]{pagel98} 
1708: 	Pagel, B.~E.~J., \& Tautvaisiene, G.\ 1998, \mnras, 299, 535 
1709: 	
1710: \bibitem[Panagia(1973)]{panagia73} 
1711: 	Panagia, N.\ 1973, \aj, 78, 929 
1712: 
1713: \bibitem[Pierre et al.(1986)]{pierre86} 
1714: 	Pierre, M., Viton, M., Sivan, J.~P., \& Courtes, G.\ 1986, \aap, 154, 249 
1715: 
1716: \bibitem[Pont et al.(2004)]{pont04} 
1717: 	Pont, F., Zinn, R., Gallart, C., Hardy, E., \& Winnick, R.\ 2004, \aj, 127, 840 
1718: 
1719: \bibitem[Proffitt et al.(2002)]{proffitt02}         
1720: 	Proffitt, C., et al. 2000, STIS Instrument Handbook, v6.0, (Baltimore:STScI)
1721: 
1722: \bibitem[Putman(2000)]{putman00} 
1723: 	Putman, M. E. 2000, PASA, 17, 1
1724: 
1725: \bibitem[Putman et al.(2003)]{putman03} 
1726: 	Putman, M.~E., Bland-Hawthorn, J., Veilleux, S., Gibson, B.~K., Freeman, K.~C., \& 
1727: 	Maloney, P.~R.\ 2003, \apj, 597, 948 
1728: 
1729: \bibitem[Reynolds(1984)]{reynolds84} 
1730: 	Reynolds, R.~J.\ 1984, \apj, 282, 191 
1731: 
1732: \bibitem[Reynolds(1993)]{reynolds93} 
1733: 	Reynolds, R.~J.\ 1993, Back to the Galaxy, 278, 156 
1734: 
1735: \bibitem[Rolleston et al.(1999)]{rolleston99}
1736: 	 Rolleston, W. R. J., Dufton, P. L., McErlean, N. D., \& Venn, K. A. 1999, \aap, 348, 728
1737: 	 
1738: \bibitem[Russell and Dopita(1992)]{russell92} 
1739: 	Russell, S. C., \& Dopita, M. A. 1992, \apj, 384, 508
1740: 
1741: \bibitem[Salpeter \& Hoffman(1996)]{salpeter96} 
1742: 	Salpeter, E.~E., \& Hoffman, G.~L.\ 1996, \apj, 465, 595 
1743: 
1744: \bibitem[Savage \& Sembach(1991)]{savage91} 
1745: 	Savage, B. D., \& Sembach, K. R. 1991, \apj, 379, 245
1746: 		
1747: \bibitem[Sembach \& Savage(1992)]{sembach92} 
1748: 	Sembach, K. R., \& Savage, B. D. 1992, \apjs, 83, 147
1749: 	
1750: \bibitem[Slavin et al.(2000)]{slavin00} 
1751: 	Slavin, J.~D., McKee, C.~F., \& Hollenbach, D.~J.\ 2000, \apj, 541, 218 
1752: 
1753: \bibitem[Smoker et al.(2000)]{smoker00} 
1754: 	Smoker, J.V., Keenan, F.P., Polatidis, A., Mooney, C.J.,  Lehner, N., \&
1755: 	Rolleston, W.R.J. 2000, \aap, 363, 451 
1756: 
1757: \bibitem[Sofia \& Jenkins(1998)]{sofia98} 
1758: 	Sofia, U. J., \& Jenkins, E. B. 1998, \apj, 499, 951  
1759: 
1760: \bibitem[Terzian(1974)]{terzian74} 
1761: 	Terzian, Y.\ 1974, \apj, 193, 93 
1762: 	
1763: \bibitem[Tufte et al.(1998)]{tufte98} 
1764: 	Tufte, S.~L., Reynolds, R.~J., \& Haffner, L.~M.\ 1998, \apj, 504, 773 
1765: 
1766: \bibitem[Wakker et al.(2001)]{wakker01} 
1767: 	Wakker, B.~P., Kalberla, P.~M.~W., van Woerden, H., de Boer, K.~S., \& Putman, M.~E.\ 2001, \apjs, 
1768: 	136, 537 
1769: 
1770: \bibitem[Welty et al.(1997)]{welty97} 
1771: 	Welty, D. E., Lauroesch, J. T., Blades, J. C.,  Hobbs, L. M., \& York, D. G. 1997, \apj, 489, 672  
1772: 	
1773: \bibitem[Wolfire et al.(1995)]{wolfire95} 
1774: 	Wolfire, M.~G., McKee, C.~F., Hollenbach, D., \& Tielens, A.~G.~G.~M.\ 1995, \apj, 453, 673 
1775: 
1776: \bibitem[Wolfire et al.(2003)]{wolfire03} 
1777: 	Wolfire, M.~G., McKee, C.~F., Hollenbach, D., \& Tielens, A.~G.~G.~M.\ 2003, \apj, 587, 278 
1778: 
1779: \bibitem[Yoshizawa \& Noguchi(2003)]{yoshizawa03} 
1780: 	Yoshizawa, A.~M., \& Noguchi, M.\ 2003, \mnras, 339, 1135 
1781: 
1782: \end{thebibliography}
1783: 
1784: 
1785: \end{document}
1786: