1: \documentclass[12pt,preprint]{aastex}
2:
3:
4: %Macros
5: \newcommand{\arcsecpt}{\makebox[0pt]{ .}^{\prime\prime}}
6: \newcommand{\arcminpt}{\mbox{\,\makebox[0pt]{.}^{\prime}}}
7: \newcommand{\etal}{\mbox,{\it et al.}}
8: \newcommand{\msolar}{\mbox{\,M_{\odot}}}
9: %\newcommand{\h100}{\mbox{\,$h^{-1}$ kpc}}
10: \newcommand{\HI}{\mbox{\normalsize H\thinspace\footnotesize I}}
11: \newcommand{\DI}{\mbox{\normalsize D\thinspace\footnotesize I}}
12: \newcommand{\sigbarW}{\mbox{\,$\bar\sigma_W$}}
13: \newcommand{\sigbar}{\mbox{\,$\bar\sigma$}}
14: \newcommand{\lya}{\mbox{\,Ly$\alpha$}}
15: \newcommand{\lyb}{\mbox{\,Ly$\beta$}}
16: \newcommand{\tenO}{\mbox{\,10$^{20}$ cm$^{-2}$}}
17: \newcommand{\tenOn}{\mbox{\,10$^{21}$ cm$^{-2}$}}
18: \newcommand{\hkpc}{\mbox{\,$h^{-1}$ kpc}}
19: \newcommand{\kmsec}{\mbox{\,km s$^{-1}$}}
20: \newcommand{\NV}{\ion{N}{5}}
21: \newcommand{\CIV}{\ion{C}{4}}
22: \newcommand{\MgII}{\ion{Mg}{2}}
23: \newcommand{\CIII}{\ion{C}{3}}
24: \newcommand{\OVI}{\ion{O}{6}}
25: \newcommand{\OVII}{\ion{O}{7}}
26: \newcommand{\OVIII}{\ion{O}{8}}
27:
28:
29:
30: \begin{document}
31:
32:
33: \title{Discovery of Associated Absorption Lines in
34: an X-Ray Warm Absorber:
35: {\it HST-FOS} Observations of MR2251$-$178\footnote{Based on Observations with the NASA/ESA {\it Hubble Space Telescope}, obtained at the Space Telescope Science Institute, which is operated by AURA, Inc., under NASA contract NAS 5-26555.}}
36:
37: \author{Eric M. Monier\footnote{monier@astronomy.ohio-state.edu}\.\ and Smita Mathur\footnote{smita@astronomy.ohio-state.edu}}
38: \affil{Department of Astronomy, The Ohio State University}
39: \affil{ 140 W. 18th Ave., Columbus, OH 43204}
40: \and
41: \author{Belinda Wilkes and Martin Elvis}
42: \affil{Harvard-Smithsonian Center for Astrophysics}
43: \affil{60 Garden Street, Cambridge, MA 02138}
44:
45:
46: \begin{abstract}
47:
48: The presence of a ``warm absorber'' was first suggested to explain
49: spectral variability in an X-ray spectrum of the radio-quiet QSO MR2251$-$178.
50: A unified picture, in which
51: X-ray warm absorbers and ``intrinsic'' UV absorbers are the same,
52: offers the opportunity to probe the nuclear environment
53: of active galactic nuclei. To test this scenario and understand the
54: physical properties of the absorber, we obtained a UV spectrum of
55: MR2251$-$178 with the Faint Object Spectrograph (FOS) onboard
56: the {\it Hubble Space Telescope (HST)}. The {\emph HST}
57: spectrum clearly shows absorption due to Ly$\alpha$, \ion{N}{5}
58: and \ion{C}{4}, blueshifted by 300 \kmsec\ from the emission redshift
59: of the QSO. The rarity of both X-ray and UV absorbers in radio-quiet
60: QSOs suggests these absorbers are physically related, if not identical.
61: Assuming the unified scenario, we place constraints on the
62: physical parameters of the absorber and conclude the mass outflow
63: rate is essentially the same as the accretion rate in MR2251$-$178.
64:
65: \end{abstract}
66:
67:
68: \keywords{quasars: absorption lines ---
69: quasars: individual (MR2251$-$178) -- ultraviolet: galaxies}
70:
71: \newpage
72:
73: \section{Introduction}
74:
75: Absorption due to \OVII\ and \OVIII\ in the X-ray and \OVI\
76: in the UV can in some cases be modeled
77: by a single UV/X-ray absorber (Mathur 1994; Mathur, Wilkes, \& Elvis
78: 1998), offering a unique way to determine the physical conditions
79: of the absorbing material in active galactic nuclei (AGNs).
80: As more X-ray/UV absorbers have been found (e.g., Mathur 1994; Mathur, Elvis,
81: \& Wilkes 1995; Mathur, Elvis, \& Singh 1995; Mathur, Wilkes, \& Aldcroft
82: 1997; Shields \& Hamann 1997; Mathur et al. 1998; Hamann, Netzer,
83: \& Shields 2000) it has become clear these share common characteristics.
84: Though not all of these absorbers can be readily modeled by a single zone,
85: all are composed of highly ionized, low density, high
86: column density, outflowing gas sitting outside the broad emission line region
87: (BELR). The X-ray/UV absorbers are therefore an important nuclear component
88: of AGNs, representing a wind or outflow carrying away significant kinetic
89: energy at a mass-loss rate comparable to the accretion rate needed
90: to power the AGN (Mathur et al. 1995). The properties of these absorbers
91: can be invoked in AGN models developed to explain such outflows
92: (see Elvis 2000 and references therein), and will aid in expanding
93: such models into a wider unification scenario (Elvis 2000).
94:
95:
96: It remains to be determined how common X-ray/UV absorbers are. In a sample of
97: eight Seyfert galaxies, Crenshaw et al. (1999) found six of them showed
98: both UV and X-ray absorption, strengthening the UV/X-ray connection.
99: The question remains whether X-ray absorbers are physically
100: related to UV absorbers in all AGN (two of the Seyferts of the Crenshaw et al.
101: sample were not X-ray warm absorbers and showed no UV absorption), or, if not,
102: in what fraction of cases does the unified model apply.
103: To help answer these questions, we initiated an {\it HST} program
104: to search for UV absorption in two radio-quiet QSOs with
105: known X-ray warm absorbers.
106:
107: Mathur et al. (1998) reported UV absorption in the first of these
108: objects, PG1114+445. Here we present the detection of UV absorption
109: in our second candidate, MR2251$-$178, a low-redshift ($z_{em} = 0.0640$)
110: radio-quiet QSO.
111: In addition to being the first X-ray selected QSO (Ricker et al. 1978),
112: MR2251$-$178 was the object for which Halpern (1984) originally suggested
113: the presence of partially ionized, optically thin material: the X-ray
114: warm absorber.
115:
116: In {\S 2} we describe the HST observations and the analysis of the UV absorption
117: lines. Section 3 relates the UV absorption to the X-ray absorption seen in
118: ASCA data analyzed by Reynolds (1997) and presents some simple physical quantities
119: based on the assumption that the UV and X-ray absorbers are the same. We present
120: our conclusions in {\S 4}.
121:
122: \section{The {\it HST} spectra}
123:
124: On 1996 February 2, MR2251$-$178 was observed using three gratings of
125: the {\it Faint Object Spectrograph} onboard {\it HST}. The target
126: was centered in the 1'' circular aperture through a four step peak-up
127: sequence to ensure a pointing accuray of 0$\arcsecpt$12. Two exposures
128: totalling 5740s were obtained using the G130H grating and blue side of the
129: detector. The exposures using the red side and the G190H and G270H gratings
130: were 1730s and 300s, respectively.
131:
132: The data were reduced using the standard pipeline processing and calibration
133: files, and the individual G130H spectra were combined. The final spectra
134: are shown in Figure 1. Absorption lines of \lya, \CIV\ and \NV\ can be clearly
135: seen, as predicted from the X-ray/UV models. The strong absorption lines of
136: \lya\ and \CIV\ are superposed on the emission lines of \lya\ and \CIV. Weaker
137: absorption due to the \NV\ doublet is seen against the continuum; no
138: measurable \NV\ emission is present.
139: The spectra do not show absorption due to low-ionization \MgII\,
140: indicating the X-ray absorber is highly ionized.
141:
142: There is an inherent uncertainty involved in measuring an absorption line
143: lying on top of a broad emission line profile because the unabsorbed shape
144: of the emission line is unknown. We therefore used two methods to
145: measure the line equivalent widths (EWs) as in Mathur et al. (1998).
146: We made an initial measurement using the
147: IRAF SPLOT task to fit Gaussians across the absorption in the emission line
148: profiles. A single Gaussian was used for \lya\ and \NV\ lines, and
149: the \CIV\ doublet is deblended into two Gaussians. Results for the EWs
150: and FWHMs are listed in Table 1. We estimate the errors on the EWs for
151: \lya\ and \CIV\ to be $\sim$40\%. The continuum is more readily determined
152: on either side of the \NV\ lines, hence the error on the SPLOT measurement
153: is estimated at $\sim$10\%. The \CIV\ and \NV\ doublet ratios are both
154: $\approx$1.7, so the doublets may be partially saturated, though
155: not as severely as in PG1114+445. Higher resolution spectra would be needed
156: to address the degree of saturation. Also, though it is probable the individual
157: lines will break into multiple components at higher resolution
158: (Ly$\alpha$ in fact has two components; see \S 3.1), we treat them as single lines here.
159:
160: To obtain a more accurate estimate of the absorption line EWs, we used the SPECFIT
161: (Kriss 1994) task under STSDAS to simultaneously fit the underlying continuum
162: and the emission and absorption line profiles. The continuum was
163: characterized by a simple power law, the emission lines were fit with
164: multiple Gaussians, and each absorption line was fit with a single Gaussian.
165: First, the power-law continuum was fit to featureless parts of the spectrum
166: on either side of the emission lines and the absorption to be measured.
167: The parameters of the power-law continuum -- the slope and the normalization
168: -- were then fixed in subsequent fits. The parameters of the Gaussians
169: used to fit the emission lines (flux, centroid, FWHM and skew) and the
170: absorption lines (flux, centroid and FWHM) were allowed to vary freely.
171: The fits to the \lya, \NV\ and \CIV\ emission/absorption profiles are
172: shown in Figure 2, and the corresponding EW and FWHM of the absorption
173: fits are given in Table 1. The results of the SPLOT and SPECFIT methods
174: are qualitatively similar, although the SPECFIT values are systematically
175: larger due to the higher continuum provided by fitting the emission
176: lines. These differences are unimportant to the discussion.
177:
178: The observed FWHM of the absorption lines is 1.3 - 3.0 \AA\
179: ($>$300 km s$^{-1}$) with
180: SPLOT and 1.4 - 3.3 \AA\ ($>$320 km s$^{-1}$) with SPECFIT. For each ion the total
181: absorption is likely resolved given the $\approx$230 \kmsec\ resolution of FOS, so
182: the absorber is dispersed in velocity space. In a higher resolution spectrum,
183: however, the absorption lines may split into multiple components.
184: The average absorption redshift
185: for the five lines is $<$z$>$=0.063, blueshifted by 300 \kmsec\ from the
186: QSO emission redshift of $z_{em} = 0.0640$ (Bergeron et al. 1983).
187: We expect that \OVI\ absorption is also present, but it lies below
188: the blue cutoff of the G130H grating at the redshift of MR2251$-$178.
189:
190: MR2251$-$178 was also observed in 1998 Dec
191: by J. Stocke with HST-STIS over the wavelength range 1250--1300\AA.
192: The \lya\ absorption line breaks into a broad component and a narrow component
193: at the $\approx$12 \kmsec\ resolution of the STIS G140M grating.
194: Figure 3 shows the \lya\ region and a fit to the absorption profile
195: produced with SPECFIT as described previously.
196: The resulting line centers, EWs and FWHMs of the two components are
197: listed in Table 1.
198:
199:
200: \section{The X-Ray/UV Absorber}
201:
202: High-ionization UV absorption lines due to \lya, \NV, and
203: \CIV\ are seen in the {\it HST} spectrum of MR2251$-$178, as predicted
204: by UV/X-ray models. Mathur et al. (1998) estimate the chance probability that
205: a radio-quiet QSO will have both associated
206: UV and X-ray ionized absorption to be $\sim 1.7 \times 10^{-3}$.
207: This strongly suggests the two absorption systems are physically
208: related.
209:
210:
211: \subsection{Physical Properties of the Ionized Gas}
212:
213: To explore the possibility that the UV and X-ray absorption lines arise
214: in the same component of the nuclear material of the AGN, we need to compare
215: the properties of the UV and X-ray absorbers.
216:
217: We took the measurements of the X-ray absorber in MR2251$-$178
218: from the analysis of ASCA data (obtained 6 Nov 1993) by Reynolds (1997; Komossa 2001 obtains
219: a similar result using ROSAT data from the same epoch). The total equivalent
220: hydrogen column density of the X-ray absorber is $N_{\rm H} =
221: 5 \times 10^{21}$ atoms cm$^{-2}$, determined
222: by fitting the warm absorber with a one-zone model using the
223: CLOUDY photoionization code (Ferland 1996). Reynolds constrained the
224: optical depths of the \OVII/\OVIII\ edges in a simple two-edge model with
225: a power-law continuum ($\alpha = 1.73$).
226: We used these values as an input to CLOUDY
227: to derive the column densities of the UV ions seen in absorption in the {\it HST} data.
228: The standard ``Table AGN'' (Mathews \& Ferland 1987) continuum
229: (with a resulting ionization parameter of $U = 1.0.)$\footnote{{\it U} is defined as the dimensionless ratio of
230: ionizing photon to hydrogen number density.}
231: was used with the assumption of solar abundances. The shape of the observed
232: UV continuum of MR2251$-$178 (Figure 1) is consistent with that in "Table AGN."
233: Table 1 lists the CLOUDY predictions for the ionic column densities of \ion{H}{1},
234: \ion{N}{5}, and \ion{C}{4}, as derived from the X-ray data.
235:
236: To compare these predictions to what is observed in the UV, it is necessary
237: to convert the equivalent widths to column densities. As discussed
238: in $\S$2, the lines may be somewhat saturated, thus the column
239: densities can be calculated only if the velocity dispersion parameter
240: ($b$ parameter) of the lines is known.
241: If the lines are resolved, then $b{\approx}$ 450 \kmsec\ [$b = $FWHM/(4ln2)$^{1/2}$;
242: the FWHM is given in Table 1] for
243: \lya\ and $\approx$ 200 - 270 \kmsec\ for \CIV\ and \NV.
244: The X-ray prediction for the column density of hydrogen, $N_{\rm HI} = 8 \times 10^{14}$ cm$^{-2}$,
245: is therefore about a factor of 2 larger than the minimum value measured in the FOS data
246: ($N_{\rm HI} \approx 4 \times 10^{14}$ cm$^{-2}$).
247: Agreement between the UV and X-ray values is achieved
248: for $b \approx 200$ \kmsec, implying the line breaks into at least two components at
249: higher resolution, as is shown in the STIS data (\S 2).\footnote{A weak feature at $\approx$ 1295.5 \AA\
250: may also be \lya\ redshifted relative to the quasar, as has been
251: observed in \ion{C}{4} lines of NGC 5548 (Mathur, Elvis, \& Wilkes 1999);
252: this feature would not contribute significantly to $N_{\HI}$.}
253: A curve-of-growth analysis using the new $b$-values
254: ($\sim$65 and $\sim$280 \kmsec) results in a combined
255: total \lya\ column density of $\sim7 \times 10^{14}$, in
256: good agreement with the X-ray prediction for \ion{H}{1}.
257:
258: The X-ray predicted values for the \ion{N}{5} and \ion{C}{4} are
259: smaller than those calculated from the UV equivalent width measurements.
260: For \ion{N}{5}, the X-ray prediction is $N_{\rm NV} = 1.6 \times 10^{14}$ cm$^{-2}$,
261: while the UV equivalent width measurement gives $N_{\rm HI} = 8 \times 10^{14}$ cm$^{-2}$,
262: assuming the lines are resolved. Similarly for \ion{C}{4}, the X-ray prediction
263: of $N_{\rm CIV} = 1.6 \times 10^{13}$ cm$^{-2}$ is more than an order of magnitude
264: smaller than the UV measurement of $N_{\rm CIV} \ge 7 \times 10^{14}$ cm$^{-2}$.
265: The X-ray predictions for $N_{\rm NV}$ and especially $N_{\rm CIV}$
266: lie on the linear part of the curve of growth.
267: If the UV and X-ray absorbers are the same, we predict these absorbers should
268: break into multiple components -- each with lower $b$ -- in UV data of higher resolution,
269: leading to values in closer agreement with the X-ray predictions.
270: Kinematically complex UV absorption lines have already been seen in high-resolution
271: data of Mrk 509 (Crenshaw, Boggess \& Wu 1995), NGC 4151
272: (Weymann et al. 1997), NGC 3516 (Kriss et al. 1996, Crenshaw et al. 1999),
273: and NGC 5548 (Mathur et al. 1999; Crenshaw \& Kraemer 1999).
274:
275: Also, note the column densities derived here from the
276: X-ray data are based on the Table AGN continuum. The shape of the actual MR2251$-$178
277: continuum in the unobserved EUV region may greatly affect the predicted
278: column densities (Mathur et al. 1995). A better match between the observed
279: \& predicted values may be obtained if the spectral energy distribution of
280: MR2251$-$178 has a different shape.
281: In any case, the X-ray absorber is a large contributor to the
282: absorption measured for \CIV\ and \NV\ in the UV, providing support
283: for the idea that the UV and X-ray absorption is occurring in the same material.
284:
285: \subsection{Variability}
286:
287: Halpern (1984) found that the warm absorber in MR2251$-$178 was highly variable.
288: ASCA data of MR2251$-$178 obtained in 1996 -- near the epoch of the HST-FOS data -- show
289: a decrease in X-ray intensity from log $\xi$ = 1.35 in 1993 to log $\xi$ = 0.96 in 1996
290: (Otani, Kii, \& Miya 1998), where $\xi$ is an ionization parameter defined as $\xi \equiv L/nR^2$.
291: If the ionization parameter $U$ varied similarly,
292: then in 1996 $U = 0.61$ and the X-ray-predicted column densities would be
293: $N_{\rm CIV} = 3.9 \times 10^{14}$ cm$^{-2}$ and
294: $N_{\rm NV} = 2.1 \times 10^{15}$ cm$^{-2}$, closer to the values measured in
295: the UV data. In this case, $N_{\rm HI}$ grows to $3.4 \times 10^{15}$ cm$^{-2}$.
296:
297: \subsection{Physical Implications}
298:
299: As noted, the UV absorption indicates the material is flowing
300: outward from the UV/X-ray absorber at $\approx$300 \kmsec. With the assumption
301: that the UV and X-ray absorbers are the same we can determine the physical conditions
302: and mass outflow rate of the absorbing material.
303: We use the total hydrogen
304: column of $N_{\rm H}$ = 5$\times 10^{21}$ cm$^{-2}$ and the ionization
305: parameter $U = 1.0$ from the CLOUDY modeling of the X-ray data. The distance of the absorber from
306: the nucleus is constrained by $U = (Q/4{\pi}r^2n_Hc)^{1/2}$, where $Q$ is
307: the rate of ionizing photons. We used the HST data at 1339 \AA\ rest wavelength
308: to scale the standard Table AGN continuum to the luminosity of MR2251$-$178
309: ($L_{bol} = 5 \times 10^{45}$ ergs s$^{-1}$).
310: Integrating $L_{\nu}/(h{\nu})$ over ${\nu} \ge 13.6$eV
311: gives $Q = 6h^{-2} \times 10^{54}$ s$^{-1}$, where $h$ is the Hubble
312: constant in units of 100 \kmsec\ Mpc$^{-1}$. This puts the distance of
313: the absorber from the nucleus at $r = 3.0 \times 10^{18}n_5^{-1/2}$ cm (where $H_{\rm o} = 75$
314: \kmsec\ Mpc$^{-1}$ and $n_5$ is the density in units of $10^5$ cm$^{-3}$).
315:
316:
317: The absorption extends below the continuum level (Figure 1), and so must cover a
318: substantial fraction of the BELR as well as the continuum source.
319: The absorber may lie outside the BELR or it
320: could also be cospatial with it if most of the emission is beamed
321: from the far side of the BELR. The size of the BELR in MR2251$-$178
322: can be estimated by scaling the BELR of NGC 5548 ($L_{bol} = 5 \times
323: 10^{44}$ ergs s$^{-1}$) by $L^{1/2}$ (Davidson \& Netzer 1979; but see
324: Kaspi et al. 2000). The distance -- determined from reverberation mapping --
325: of the BELR from the central continuum in NGC 5548 is $r \simeq 10$
326: light-days (Clavel et al. 1991), so the absorber in MR2251$-$178 must be at least $\sim$32
327: light-days, or $r \ge 8.2 \times 10^{16}$ cm from the central source.
328: The corresponding upper limit on the density is then $n < 1.34 \times 10^8$
329: cm$^{-3}$. To check the radius estimate, we specified as an input to CLOUDY the luminosity
330: of MR2251$-$178 at a rest wavelength of 1339 \AA, log(${\nu}L_{\nu}$) = 43.78,
331: and varied the input radius to obtain the ionization parameter output
332: previously. The radius at which $U = 1.0$ was $r = 8.76 \times 10^{16}$ cm.
333: For a covering factor of $f = 0.1$, the column density and particle density
334: of the aborber imply a mass of $M = 47f_{0.1}n_5^{-1} M_{\odot}$.
335: The mass outflow rate (Mathur et al. 1995) would then be
336: $\dot{M}_{out} = 0.9f_{0.1} M_{\odot}$ yr$^{-1}$,
337: the same rate needed to power MR2251$-$178 at 10\% efficiency.
338: The rate of kinetic energy carried away in the flow is $\dot{M}v_{out}^2/2 =
339: 8 \times 10^{40}$ ergs cm$^{-2}$ s$^{-1}$.
340:
341:
342: \section{Conclusions}
343:
344: The {\it HST} FOS UV spectrum of MR2251$-$178 exhibits associated
345: high-ionization, UV absorption lines, as predicted from models of the
346: X-ray ionized absorber. The X-ray/UV absorber is situated outside
347: the BELR or cospatial with it and outflowing with a line-of-sight
348: velocity of $\approx$300 \kmsec. The mass outflow rate of 0.9 M$_{\odot}$
349: yr$^{-1}$ for a 10\% covering factor is equivalent to the accretion rate
350: onto the nuclear black hole.
351:
352: We expect future higher resolution observations
353: will reveal each of the \CIV\ and \NV\ absorption lines
354: breaking into at least two components.
355:
356: The rarity of both UV and X-ray absorbers individually in radio-quiet QSOs
357: virtually requires that the X-ray and UV absorbers are closely
358: physically related. The consistency of the column densities obtained
359: from both UV and X-ray data suggests the absorbers are perhaps identical.
360: At a minimum, the X-ray absorber makes a substantial contribution to the
361: absorption seen in the UV. Thus, the absorber in MR2251$-$178 satisfies
362: both statistical as well as physical tests of our X-ray/UV absorber
363: model.
364:
365: \begin{references}
366:
367: \reference{}
368: Bergeron, J., Boksenberg, A., Dennefeld, M., \& Tarenghi, M. 1983, \mnras,
369: 202, 125
370:
371: \reference{}
372: Clavel, J., et al. 1991, \apj, 366, 64
373:
374: \reference{}
375: Crenshaw, D.M., Boggess, A., \& Wu, C.-C. 1995, \aj, 110, 1026
376:
377: \reference{}
378: Crenshaw, D.M., \& Kraemer, S.B. 1999, \apj, 521, 572
379:
380: \reference{}
381: Crenshaw, D.M., Kraemer, S.B., Boggess, A., Maran, S.P., Mushotzky, R.F., \&
382: Wu, C.-C. 1999, \apj, 516, 750
383:
384: \reference{}
385: Davidson, K., \& Netzer, H. 1979, {\it Rev. Mod. Phys.}, 51, 715
386:
387: \reference{}
388: Elvis, M. 2000, \apj, 545, 63
389:
390: \reference{}
391: Ferland, G. 1996, Univ. Kentucky, Dept. of Physics \& Astronomy Internal Report
392:
393: \reference{}
394: Halpern, J.P 1984, \apj, 281, 90
395:
396: \reference{}
397: Hamann, F.W., Netzer, H., \& Shields, J.C. 2000, \apj, 536, 101
398:
399: \reference{}
400: Kaspi, S., Smith, P.S., Netzer, H., Maoz, D., Jannuzi, B.T., \& Giveon, U. 2000, \apj, 533, 631
401:
402: \reference{}
403: Komossa, S. 2001, \aap, in press
404:
405: \reference{}
406: Kriss, G.A. 1994, in ASP Conf. Proc. 61, Astronomical Data Analysis Software
407: and Systems III, ed. D.R. Crabtree, R.J. Hanisch, \& J. Barnes (San Francisco:
408: ASP), 437
409:
410: \reference{}
411: Kriss, G., Espey, B.R., Krolik, J.H., Tsvetanov, Z., Zheng, W., \& Davidsen,
412: A.F. 1996, \apj, 476, 622
413:
414: \reference{}
415: Mathews, W.G., \& Ferland, G.J. 1987, \apj, 232, 456
416:
417: \reference{}
418: Mathur, S. 1994, \apjl, 431, 75
419:
420: \reference{}
421: Mathur, S., Elvis, M., \& Singh, K.P. 1995, \apjl, 455, 9
422:
423: \reference{}
424: Mathur, S., Elvis, M.S., \& Wilkes, B.J. 1995, \apj, 452, 230
425:
426: \reference{}
427: Mathur, S., Wilkes, B.J., \& Aldcroft, T. 1997, \apj, 478, 182
428:
429: \reference{}
430: Mathur, S., Wilkes, B.J., \& Elvis, M. 1998, \apjl, 503, 23
431:
432: \reference{}
433: Mathur, S., Elvis, M., \& Wilkes, B. 1999, \apj, 519, 605
434:
435: \reference{}
436: Otani, C., Kii, T., \& Miya, K. 1998, in IAU Symp. 188, The Hot Universe, ed.
437: K. Koyama, S. Kitamoto, \& M. Itoh (Kluwer: IAU), 436
438:
439: \reference{}
440: Ricker, G.R. et al. 1978, \nat, 271, 35
441:
442: \reference{}
443: Reynolds, C.S. 1997, \mnras, 286, 513
444:
445: \reference{}
446: Shields, J., \& Hamann, F. 1997, \apj, 481, 752
447:
448: \reference{}
449: Weymann, R., Morris, S., Gray, M., \& Hutchings, J. 1997, \apj, 483, 717
450:
451: \clearpage
452:
453: \end{references}
454:
455:
456: \figcaption[hst.ps]{{\it HST} spectrum of MR2251$-$178 showing absorption
457: lines of Ly$\alpha$, \ion{N}{5}, and \ion{C}{4}. The lower line on each
458: panel is the error spectrum.}
459:
460: \figcaption[fits.ps]{SPECFIT fits to the absorption lines of Ly$\alpha$ and \ion{N}{5}
461: (upper panel) and \ion{C}{4} (lower panel). Results of the fits are shown in
462: Table 1.}
463:
464: \figcaption[stis.ps]{SPECFIT fit to the {\it HST-STIS} spectrum of the
465: \lya\ absorption line of MR2251$-$178. Results of the fit are shown in Table 1.}
466:
467: \begin{figure}
468: \plotone{f1.eps}
469: \end{figure}
470:
471: \begin{figure}
472: \epsscale{0.65}
473: \plotone{f2.eps}
474: \end{figure}
475:
476: \begin{figure}
477: \epsscale{0.70}
478: \plotone{f3.eps}
479: \end{figure}
480:
481: \clearpage
482:
483: \input{tab1}
484:
485: \end{document}
486: