1: %%
2: %% This is xspec_example.tex
3: %% LaTeX2e example file for the AGN SPECTROSCOPY WORKSHOP AT MPE
4: %% Needs the LaTeX2e class file agnspec.cls
5: %%
6: % -----------------------------------------------------------------------------
7: \documentclass{agnspec}
8: \usepackage{graphics}
9: \usepackage{psfig}
10: \usepackage{./natbib}
11: %This makes the citations look right.
12: \bibpunct{(}{)}{;}{a}{}{,}
13: % -----------------------------------------------------------------------------
14: %%% Personal Abbreviations
15: % ions and transitions
16: \newcommand{\Ha}{H$\alpha$}
17: \newcommand{\Hb}{H$\beta$}
18: \newcommand{\feka}{\hbox{Fe~K$\alpha$}}
19: \newcommand{\CIV}{\ion{C}{iv}}
20: \newcommand{\MgII}{\ion{Mg}{ii}}
21: \newcommand{\CIIIs}{\ion{C}{iii}]}
22: % units
23: \newcommand{\kms}{\hbox{km~s$^{-1}$}}
24: \newcommand{\cmsq}{\hbox{cm$^{-2}$}}
25: \newcommand{\cc}{\hbox{cm$^{-3}$}}
26: \newcommand{\flux}{\hbox{erg~cm$^{-2}$~s$^{-1}$}}
27: \newcommand{\eflux}{\mbox{photons~cm$^{-2}$~s$^{-1}$~keV$^{-1}$}}
28: \newcommand{\lumin}{\hbox{erg~s$^{-1}$}}
29: \newcommand{\persec}{\hbox{s$^{-1}$}}
30: \newcommand{\msun}{\hbox{${M}_{\odot}$}}
31: %\newcommand{\sun}{\hbox{$_{\odot}$}}
32: \newcommand{\lsun}{\hbox{${L}_{\odot}$}}
33: \newcommand{\nh}{\hbox{${N}_{\rm H}$}}
34: % relation markers
35: \newcommand{\gtrsim}{\lower 2pt \hbox{$\, \buildrel {\scriptstyle >}\over {\scriptstyle\sim}\,$}}
36: \newcommand{\lesssim}{\lower 2pt \hbox{$\, \buildrel {\scriptstyle <}\over {\scriptstyle\sim}\,$}}
37: % observatories
38: \newcommand{\asca}{{\emph{ASCA}}}
39: \newcommand{\xmm}{{\emph{XMM-Newton}}}
40: \newcommand{\chandra}{{\emph{Chandra}}}
41: \newcommand{\axaf}{{\emph{Chandra}}}
42: \newcommand{\sax}{{\emph{Beppo-SAX}}}
43: \newcommand{\rxte}{{\emph{RXTE}}}
44: \newcommand{\rosat}{{\emph{ROSAT}}}
45: \newcommand{\einstein}{{\emph{Einstein}}}
46: \newcommand{\heao}{\hbox{\emph{HEAO-1}}}
47: \newcommand{\hst}{{\emph{HST}}}
48: \newcommand{\iue}{{\emph{IUE}}}
49: \newcommand{\iras}{{\emph{IRAS}}}
50: \newcommand{\sirtf}{{\emph{SIRTF}}}
51: % numbers and parameter labels
52: \newcommand{\onehalf}{\mbox{$\scriptstyle \frac{1}{2}$}}
53: \newcommand{\aox}{$\alpha_{\rm ox}$}
54: \newcommand{\aoxp}{$\alpha^{\prime}_{\rm ox}$}
55: \newcommand{\aoxc}{$\alpha_{\rm ox}{\rm (corr)}$}
56: \newcommand{\balq}{BAL~QSO}
57: \newcommand{\balqs}{BAL~QSOs}
58: \newcommand{\etal}{et~al.}
59: % objects
60: \newcommand{\pgfif}{PG~1535+547}
61: \newcommand{\pgfive}{PG~1535+547}
62: \newcommand{\pgfour}{PG~1411+442}
63: \newcommand{\pgten}{PG~1011--040}
64: \newcommand{\pgbal}{PG~2112+059}
65: \newcommand{\pgoh}{PG~0043+008}
66: \newcommand{\pgsev}{PG~1700+518}
67: \newcommand{\phl}{PHL~5200}
68: \newcommand{\clover}{H~1413+117}
69: \newcommand{\apm}{APM~08279+5255}
70: \newcommand{\rxj}{RX~J0911.4+0551}
71: \newcommand{\pgone}{PG~1115+080}
72: \newcommand{\mrk}{Mrk~231}
73: % journals
74: %Journals:
75: \newcommand{\apj}{{ApJ}}
76: \newcommand{\apjs}{{ApJS}}
77: \newcommand{\apjl}{{ApJL}}
78: \newcommand{\aj}{{AJ}}
79: \newcommand{\aap}{{A\&A}}
80: \newcommand{\aaps}{{A\&A}}
81: \newcommand{\mnras}{{MNRAS}}
82: % -----------------------------------------------------------------------------
83: \begin{document}
84: \title{An Exploratory \chandra\ survey of Large Bright Quasar Survey broad absorption line QSOs}
85: \author{S.~C.~Gallagher,\inst{1} W.~N.~Brandt,\inst{1} G.~Chartas,\inst{1} \and\ R.~M.~Sambruna\inst{2}}
86: \institute{Department of Astronomy \& Astrophysics, The Pennsylvania State
87: University, 525 Davey Laboratory, University Park, PA U.S.A.
88: \and Department of Physics \& Astronomy and School of Computational Sciences,
89: George Mason University, 4400 University Drive M/S 3F3, Fairfax VA U.S.A.}
90: \authorrunning{S.~C.~Gallagher~\etal}
91: \titlerunning{Exploratory \chandra\ survey of LBQS BAL~QSOs}
92: \maketitle
93: %
94: \begin{abstract}
95: We are in the process of obtaining a complete
96: sample of exploratory X-ray observations of broad absorption
97: line (BAL)~QSOs from the Large Bright Quasar Survey (LBQS).
98: This survey is composed of short (5--7~ks) \chandra\ ACIS-S3
99: observations designed to
100: significantly constrain X-ray fluxes and hardness ratios.
101: Our sample avoids the selection biases of some previous
102: surveys while providing X-ray data of similar quality.
103: Of 17 \balqs\ observed to date, 13 are detected. In general, our results are consistent
104: with absorption as the primary cause of X-ray weakness in \balqs.
105: As found by \citet{GrEtal2001}, those \balqs\ with low-ionization absorption lines
106: are found to be notably X-ray weaker than \balqs\ with only high
107: ionization lines.
108: \end{abstract}
109: %
110: \section{Introduction}
111: \label{sec:lbqsintro}
112: From studies with \rosat\ and \asca, Broad Absorption Line (BAL) QSOs
113: are known to have faint soft X-ray fluxes compared to their
114: optical fluxes \citep[e.g.,][]{GrEtal1995,GrMa1996,GaEtal1999}
115: Given the extreme absorption evident in the UV, this soft X-ray faintness was
116: assumed to result from intrinsic absorption. At the time, the
117: X-ray data could not explicitly demonstrate this, though
118: the strong correlation found by Brandt, Laor, \& Wills (2000; hereafter
119: BLW)\nocite{BrLaWi2000} between \CIV\ absorption equivalent width (EW)
120: and faintness in soft X-rays supported this assumption.
121: The observation of \pgbal\ with \asca\ provided the first solid
122: spectral evidence
123: for intrinsic X-ray absorption and a normal underlying X-ray continuum
124: in a BAL~QSO \citep{GaEtal2001a}.
125: Subsequently, more observations of BAL~QSOs with \chandra\
126: also found signatures of intrinsic X-ray absorption \citep[e.g.,][]{GrEtal2001}.
127:
128: \citet{GaBrChGa2002} compiled the results from the first moderate-sized
129: survey of \balqs\ with enough counts for X-ray spectroscopic analysis. They
130: concluded from the spectroscopic evidence that the intrinsic
131: UV-to-X-ray spectral energy distributions of \balqs\ are
132: consistent with those of typical QSOs. Furthermore, complex, intrinsic
133: absorption with $\nh=(0.1$--$4)\times10^{23}$~\cmsq\ is generally evident in the
134: X-ray spectra.
135: At present, however, only a handful of these generally faint
136: targets have provided data of sufficient quality for spectral
137: analysis. This sample is generally diverse, and thus far there
138: is no known predictor of observed 0.5--10.0~keV flux based on
139: the spectral properties in other wavelength regimes.
140: In other words, there has been no obvious connection
141: between the characteristics of the UV and X-ray absorbers as
142: of yet, though all QSO with broad UV absorption also
143: apparently have X-ray absorption.
144:
145: In an effort to remedy this situation, we are in the process of
146: compiling the multi-wavelength properties of a complete sample of
147: \balqs\ from the Large Bright Quasar Survey
148: \citep*[LBQS;][]{HeFoCh1995}. The sample will ultimately
149: include all 37 of the \balqs\ from the LBQS with $z>1.35$, the redshift
150: at which the definitive \CIV\ BAL is shifted into the wavelength
151: regime accessible to ground-based spectroscopy.
152: The optically bright targets are
153: drawn from a homogeneous, magnitude-limited survey that has
154: effective, well-defined and objectively applied selection criteria
155: \citep*[e.g.,][]{HeFoCh2001}.
156: One of the ultimate goals of this survey is to understand the connection
157: between UV and X-ray absorption in luminous QSOs, and hence gain some insight
158: into the mechanism for launching and maintaining energetic winds. To
159: further this project, we are in the process of observing each object
160: in X-rays and in the rest-frame UV.
161: In this paper, we present preliminary results from the analysis of the
162: exploratory \chandra\
163: Advanced CCD Imaging Spectrometer (ACIS;
164: G.~P. Garmire \etal, in preparation) observations of the first 17
165: targets which have been observed to date.
166: These short (5--7~ks) observations are intended to determine the basic X-ray
167: fluxes and rough spectral shapes of the sample, and for those targets
168: that were not detected, sensitive upper limits can be
169: set with these exposure times. In addition to
170: presenting new X-ray results, we also compare these data to some of the
171: multiwavelength information available in the literature.
172: All of the QSOs in this sample were observed in the
173: spectroscopic \balq\ survey of \citet[][hereafter WMFH]{WeMoFoHe1991}, and so
174: a significant amount of data is available both from this work and
175: others based on the WMFH sample.
176: %
177: \section{Observations and X-ray Data Analysis}
178: \label{sec:lbqsobs}
179: Each target was observed at the aimpoint of
180: the back-illuminated S3 CCD of ACIS in faint mode. The data were
181: processed using the standard \chandra\ X-ray Center
182: (CXC) aspect solution and grade filtering.
183: Of the 17 observed \balqs, 13 were detected with counts
184: ranging from 4--84 photons.
185: After determining the position of the X-ray centroid, the counts in
186: the full (0.35--8.0~keV), soft (0.35--2.0~keV), and hard
187: (2.0--8.0~keV) bands were extracted from a source cell with a radius of
188: $2\farcs5$.
189: The hardness ratio (HR), defined as the number of 2.0--8.0~keV counts divided by the
190: number of 0.35--2.0~keV counts, was then calculated; this parameter
191: provides a coarse quantitative measure of the spectral shape.
192: In order to transform the HR into a physical parameter, the photon index, for
193: characterizing the X-ray spectrum, observations of power-law X-ray
194: spectra were simulated using the X-ray
195: spectral modeling tool, {\sc xspec} \citep{Arnaud1996}.
196: For reference, a typical QSO with no intrinsic absorption and a photon index,
197: $\Gamma=2.0$, would have been observed to have HR=0.13--0.15 for the
198: range of Galactic column densities of (1.6--4.8)$\times10^{20}$~\cmsq\
199: in this sample.
200: Once the value of $\Gamma$ corresponding to the observed HR and Galactic
201: \nh\ was determined, the model power-law spectrum was normalized using the
202: full-band count rate. Given a photon index and normalization, the
203: derived X-ray properties, $F_{\rm X}$ (observed-frame 2--8~keV flux) and
204: $f_{\rm 2~keV}$ (flux density at rest-frame 2~keV) could be
205: calculated. Finally, \aox, the slope of a hypothetical power law
206: connecting $f_{2500}$, the flux density at 2500~\AA, with $f_{\rm 2~keV}$, was determined.
207:
208: \section{Testing for Significant Correlations}
209: Utilizing the available UV and optical data in the literature, we
210: sought evidence for correlations
211: that might indicate a connection between the X-ray properties and
212: those in other parts of the spectrum.
213: To start, we considered the ``BALnicity index'' (BI), a conservative measure of the \CIV\
214: absorption equivalent width originally defined and tabulated by WMFH.
215: If the extent of X-ray weakness
216: is due to intrinsic absorption, \aox\ might be expected to correlate with the
217: BI. To test for such a relationship, we calculated the
218: non-parametric Kendall's $\tau$ statistic.
219: %
220: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
221: \begin{figure}[th]
222: \centerline{\psfig{figure=fig1.eps,width=8.0cm}}
223: \caption[BALnicity Index vs. \aox]{BALnicity index vs. \aox.
224: The BI values are from
225: WMFH; this absorption-line parameter does not have a statistically significant
226: correlation with weakness in X-rays.
227: The solid and dashed vertical lines represent the mean and one standard deviation of
228: \aox\ for typical radio-quiet QSOs from \citet{BrEtal2001}. The stars
229: in this and subsequent plots indicate low-ionization \balqs.
230: \label{fig:bi}
231: \vspace*{-0.5cm}
232: }
233: \end{figure}
234: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
235: %
236:
237: As can be seen in Figure~\ref{fig:bi}, there is no evident
238: correlation between the \CIV\ absorption-line BI and \aox;
239: this is supported by the high probability, $P=0.28$, that the null hypothesis
240: is consistent with these data. This result could be considered
241: surprising given that \citet{BrLaWi2000} found a highly
242: statistically significant correlation between \CIV\ absorption
243: EW and \aox. However, their sample, the Bright Quasar Survey \citep{ScGr1983}
244: objects with $z<0.5$, encompassed a much larger dynamic range of both \aox\ and
245: absorption EW; the majority of their sample exhibited neither weakness
246: in X-rays nor \CIV\ absorption. Our \balq\ sample is probing the
247: most extreme end of \CIV\ absorption, where the straightforward
248: relationship between weakness in soft X-rays and UV continuum
249: absorption apparently does not hold. Given the well-documented complexity of
250: BAL profiles, which can include contributions from scattered flux as
251: well as exhibit severe saturation
252: \citep[e.g.,][]{OgCoMiTr1999,AravEtal2001}, the observed lack of
253: correlation is perhaps not remarkable.
254:
255: We also investigated possible correlations between the UV continuum
256: shape and X-ray properties.
257: Though the slope of the continuum blueward of \CIIIs, $\alpha_B$,
258: is not correlated significantly with \aox, there are some interesting
259: trends in the data (see Figure~\ref{fig:alpha}). First, all of the
260: \balqs\ with the reddest continua (largest values of $\alpha_B$)
261: reside at the X-ray weak end of the sample. However, not all of the
262: X-ray weakest \balqs\ are red. If \aox\ can be associated with the
263: amount of intrinsic X-ray absorption, this suggests that the X-ray
264: absorbing gas does not necessarily cause continuum reddening.
265: Physically this could result from absorbing gas with very low
266: gas-to-dust ratio as has been seen in Seyfert galaxies
267: \citep[][and references therein]{MaiolinoEtal2001}.
268: Furthermore, dusty absorbers responsible for continuum
269: reddening may not be identical to those causing X-ray weakness.
270: %
271: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
272: \begin{figure}[th]
273: \centerline{\psfig{figure=fig2.eps,width=8.0cm}}
274: \caption[Blue continuum slope vs. \aox]{
275: The spectral index of the continuum blueward of \CIIIs, $\alpha_B$ for
276: $F_{\nu}\propto\nu^{-\alpha_B}$, versus
277: \aox. Note that the reddest \balqs\ (those with the largest values of
278: $\alpha_B$) are also X-ray weak. The values for $\alpha_B$ are from
279: \citet{HuLaRe1998}.
280: The vertical lines in the plot are as defined for
281: Figure~\ref{fig:bi}.
282: \label{fig:alpha}
283: \vspace*{-0.5cm}
284: }
285: \end{figure}
286: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
287: %
288:
289: The search for correlations between absorption-line and
290: continuum properties with \aox\ did not result in any obvious
291: patterns, but those with the HR proved more fruitful.
292: Though individual measurements of the HR suffer from large
293: statistical errors, the ensemble of HRs from the sample offers promise
294: for investigating the relationship of observed X-ray continuum shape
295: to other properties. Tests of HR versus \aox\
296: resulted in a statistically significant correlation
297: (see Figure~\ref{fig:hr}a).
298: According to the evaluation of Kendall's
299: $\tau$, this correlation is significant at the $\gtrsim99\%$
300: confidence level.
301: While typical QSOs have rest-frame 2--10~keV photon indices of $\Gamma=1.9\pm0.27$
302: \citep[e.g.,][]{ReTu2000}, a QSO exhibiting intrinsic absorption will
303: have an apparently harder X-ray spectrum due to the lack of low-energy
304: X-rays. Thus, a larger value of the HR might result from intrinsic
305: absorption, which would also cause more negative values of \aox.
306:
307: With these data, there is no statistically significant evidence for a correlation
308: between HR and $z$. At higher $z$, the spectral
309: signatures of intrinsic absorption are pushed to lower observed
310: energies, and therefore occupy less of the soft X-ray bandpass.
311: An increase in HR with decreasing $z$ would therefore be
312: expected if all \balqs\ had similar amounts of intrinsic absorption.
313: %
314: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
315: \begin{figure}[th]
316: \centerline{\psfig{figure=fig3.eps,width=8.0cm}}
317: \caption[Hardness ratio plots]{
318: {\bf (a)} Hardness ratio versus \aox. The weakest X-ray sources
319: (relative to their UV continuum flux) tend to have the
320: hardest spectra.
321: The star symbols and vertical gray lines are as defined in the
322: caption to Figure~\ref{fig:bi}.
323: {\bf (b)} Hardness ratio versus $z$.
324: The dashed curves represent the expected HR for a power-law continuum with
325: $\Gamma=2.0$, Galactic \nh$=2.8\times10^{20}$~\cmsq\ (the median value
326: for this sample), and an intrinsic, neutral column density with the log value
327: labeled in the figure. These curves are meant for reference only as
328: the actual absorption is likely to be complex \citep{GaBrChGa2002}.
329: \label{fig:hr}
330: \vspace*{-0.7cm}
331: }
332: \end{figure}
333: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
334: %
335: \section{Discussion}
336: The sample of \balqs\ presented in this paper offers the advantage of
337: being much more homogeneous than those in previous hard-band surveys
338: \citep[e.g.,][]{GaEtal1999,GrEtal2001}.
339: With the aim of obtaining the best X-ray constraints, these samples
340: included the optically brightest \balqs\ known, which were drawn
341: from surveys employing diverse selection criteria. As a result, these
342: objects were not necessarily representative of the global properties
343: of the \balq\ population as a whole.
344: In contrast, all of our targets are drawn from the same well-defined survey.
345: In addition, our survey objects have uniform,
346: high-quality spectroscopic data from the rest-frame UV available for meaningful comparisons
347: between properties in different wavelength regimes.
348:
349: From the days of \rosat\ and \asca, the detection fraction of \balqs\
350: has increased substantially. Of the 17 observed LBQS \balqs, 13 were
351: detected. Given the sensitivity of \chandra, meaningful upper limits
352: on X-ray flux and \aox\ could be set for those that were not
353: detected. For the observed values of \aox, our range encompasses
354: values from $-2.35$ to $-1.29$ with four upper limits and a median
355: value of $-2.06$. For comparison,
356: the mean and standard deviation for QSOs without evident UV absorption is
357: $-1.56\pm0.14$ \citep{BrEtal2001}.
358: Not all of the \balqs\ are observed to have
359: \aox\ values that would group them as X-ray weak, e.g., with
360: \aox$<-1.7$. While the calculated X-ray fluxes do suffer from large
361: errors due to the generally low count rates, \aox\ is not extremely
362: sensitive to these errors. Another possible source of the spread
363: is variability of the UV continuum since some of these fluxes were
364: last measured more than 20 years ago \citep{HeFoCh1995}.
365: However, \aox\ is a robust parameter; for reference, an increase in
366: $f_{2500}$ of a factor of 2 would result in $\Delta$\aox$\approx0.1$.
367:
368: From the X-ray data alone, we have evidence consistent with
369: the interpretation of intrinsic absorption as the cause of X-ray
370: weakness: the significant correlation of HR and
371: \aox. An X-ray spectrum exhibiting intrinsic absorption by neutral,
372: partial-covering, or ionized gas will have fewer counts in the soft
373: band, and thus a higher value for the hardness ratio.
374: At larger redshifts, the absorbed part of the spectrum will be shifted to lower
375: and lower energies thus occupying a smaller fraction of the observed
376: bandpass. Therefore, for a given column density, an absorbed QSO at a
377: lower redshift will have a higher value of HR than the same QSO at a
378: higher $z$. This pattern is illustrated with the dashed lines in Figure~\ref{fig:hr}b.
379: However, making a straight identification of \aox\ with column density
380: is problematic for two significant reasons. First,
381: \citet{GaBrChGa2002} have shown that the intrinsic absorption is
382: complex, but with the present spectral resolution and signal-to-noise
383: ratio, the specific nature of that complexity is poorly constrained. Thus the
384: assumption of neutral absorption, while useful for reference, is
385: probably not valid. The effect of ionization state, covering
386: fraction, and velocity structure on measured column density from a CCD
387: spectrum remains to be quantified. Empirically, this complexity
388: manifests itself as additional flux at soft energies over what would
389: be expected from a completely neutral absorber. Given that situation, the HR and
390: \aox\ data presented in this sample should not be translated into
391: measurements of intrinsic column density.
392:
393: As found by \citet{GrEtal2001}, the low-ionization \balqs\ in this
394: sample (indicated by stars in Figures~\ref{fig:bi}--\ref{fig:hr})
395: also inhabit the X-ray weak end of \aox. Attempting to correct
396: the UV continua for dust extinction in these red QSOs would
397: only push \aox\ to more negative values. Given this evident
398: distinction in the X-ray as well as the UV properties, it is important
399: to distinguish the samples of low and high-ionization \balqs\ when trying
400: to generalize about the population of \balqs\ as a whole.
401: %
402: \section{Future Work}
403: Many \balq\ studies have observed the most notable of objects. While
404: they frequently offer interesting case studies, this
405: situation can lead to spurious connections and inhibit
406: understanding of populations as a whole.
407: To avoid this pitfall, we are collecting exploratory X-ray observations
408: of a complete sample of \balqs\ in conjunction with spectroscopic
409: X-ray data of the brightest ones. By the end of 2002, we will have 5
410: additional exploratory observations.
411: Furthermore, we are in the process of
412: obtaining rest-frame UV photometry and spectroscopy for a
413: complete sample of LBQS \balqs, as described in
414: $\S$\ref{sec:lbqsintro}, to supplement the information in the literature as
415: well as the \chandra\ data.
416: The imaging and spectroscopy, obtained with the
417: Hobby-Eberly Telescope within one month of each \chandra\ observation,
418: will enable concurrent determinations of both $f_{2500}$ and $f_{\rm
419: 2~keV}$ for measuring \aox, as well as the absorption-line
420: parameters. Finally, these data will provide the means to investigate long-term
421: variability of the velocity structure of the BALs to test the
422: radiation-driven wind models of \citet*{PrStKa2000}. In their
423: hydrodynamical models, they predict variability on timescales of a few
424: years. The most comprehensive \balq\ variability study to date, that
425: of \citet{Barlow1993}, was not of sufficient duration to test these models. With
426: the data of WMFH and our Hobby-Eberly Telescope spectroscopy, we can
427: approximately triple the timescale of \citet{Barlow1993} to span more
428: than a decade.
429:
430: \begin{acknowledgements}
431: This work was made possible by \chandra\ X-ray Center grant GO1-2105X.
432: SCG also gratefully acknowledges support from NASA GSRP grant NGT5-50277 and
433: from the Pennsylvania Space Grant Consortium.
434: \end{acknowledgements}
435:
436: \begin{thebibliography}{}
437:
438: \bibitem[\protect\citeauthoryear{{Arav} et~al.}{{Arav}
439: et~al.}{2001}]{AravEtal2001}
440: {Arav}, N., et~al. 2001, \apj, 561, 118
441:
442: \bibitem[\protect\citeauthoryear{{Arnaud}}{{Arnaud}}{1996}]{Arnaud1996}
443: {Arnaud}, K.~A. 1996, in ASP Conf. Ser. 101: Astronomical Data Analysis
444: Software and Systems V, ed. G.~Jacoby \& J.~Barnes, Vol.~5, 17
445:
446: \bibitem[\protect\citeauthoryear{{Barlow}}{{Barlow}}{1993}]{Barlow1993}
447: {Barlow}, T.~A. 1993, Ph.D. thesis, University of California, San Diego
448:
449: \bibitem[\protect\citeauthoryear{{Brandt} et~al.}{{Brandt}
450: et~al.}{2001}]{BrEtal2001}
451: {Brandt}, W.~N. et al. 2001, \aj, 121, 591
452:
453: \bibitem[\protect\citeauthoryear{{Brandt}, {Laor}, \& {Wills}}{{Brandt}
454: et~al.}{2000}]{BrLaWi2000}
455: {Brandt}, W.~N., {Laor}, A., \& {Wills}, B.~J. 2000, \apj, 528, 637
456:
457: \bibitem[\protect\citeauthoryear{{Gallagher} et~al.}{{Gallagher}
458: et~al.}{2002}]{GaBrChGa2002}
459: {Gallagher}, S.~C., {Brandt}, W.~N., {Chartas}, G., \& {Garmire}, G.~P. 2002,
460: \apj, 567, 37
461:
462: \bibitem[\protect\citeauthoryear{{Gallagher} et~al.}{{Gallagher}
463: et~al.}{2001}]{GaEtal2001a}
464: {Gallagher}, S.~C. et al. 2001, \apj, 546, 795
465:
466: \bibitem[\protect\citeauthoryear{{Gallagher} et~al.}{{Gallagher}
467: et~al.}{1999}]{GaEtal1999}
468: {Gallagher}, S.~C., {Brandt}, W.~N., {Sambruna}, R.~M., {Mathur}, S., \&
469: {Yamasaki}, N. 1999, \apj, 519, 549
470:
471: \bibitem[\protect\citeauthoryear{{Green} et~al.}{{Green}
472: et~al.}{2001}]{GrEtal2001}
473: {Green}, P.~J., {Aldcroft}, T.~L., {Mathur}, S., {Wilkes}, B.~J., \& {Elvis},
474: M. 2001, \apj, 558, 109
475:
476: \bibitem[\protect\citeauthoryear{{Green} \& {Mathur}}{{Green} \&
477: {Mathur}}{1996}]{GrMa1996}
478: {Green}, P.~J., \& {Mathur}, S. 1996, \apj, 462, 637
479:
480: \bibitem[\protect\citeauthoryear{Green et~al.}{Green et~al.}{1995}]{GrEtal1995}
481: Green, P.~J., et~al. 1995, \apj, 450, 51
482:
483: \bibitem[\protect\citeauthoryear{{Hewett}, {Foltz}, \& {Chaffee}}{{Hewett}
484: et~al.}{1995}]{HeFoCh1995}
485: {Hewett}, P.~C., {Foltz}, C.~B., \& {Chaffee}, F.~H. 1995, \aj, 109, 1498
486:
487: \bibitem[\protect\citeauthoryear{{Hewett}, {Foltz}, \& {Chaffee}}{{Hewett}
488: et~al.}{2001}]{HeFoCh2001}
489: {Hewett}, P.~C., {Foltz}, C.~B., \& {Chaffee}, F.~H. 2001, \aj, 122, 518
490:
491: \bibitem[\protect\citeauthoryear{{Hutsem{\'e}kers}, {Lamy}, \&
492: {Remy}}{{Hutsem{\'e}kers} et~al.}{1998}]{HuLaRe1998}
493: {Hutsem{\'e}kers}, D., {Lamy}, H., \& {Remy}, M. 1998, \aap, 340, 371
494:
495: \bibitem[\protect\citeauthoryear{{Maiolino} et~al.}{{Maiolino}
496: et~al.}{2001}]{MaiolinoEtal2001}
497: {Maiolino}, R. et al. 2001, \aap, 365, 28
498:
499: \bibitem[\protect\citeauthoryear{{Ogle} et~al.}{{Ogle}
500: et~al.}{1999}]{OgCoMiTr1999}
501: {Ogle}, P.~M. et al. 1999, \apjs, 125, 1
502:
503: \bibitem[\protect\citeauthoryear{{Proga}, {Stone}, \& {Kallman}}{{Proga}
504: et~al.}{2000}]{PrStKa2000}
505: {Proga}, D., {Stone}, J.~M., \& {Kallman}, T.~R. 2000, \apj, 543, 686
506:
507: \bibitem[\protect\citeauthoryear{{Reeves} \& {Turner}}{{Reeves} \&
508: {Turner}}{2000}]{ReTu2000}
509: {Reeves}, J.~N., \& {Turner}, M. J.~L. 2000, \mnras, 316, 234
510:
511: \bibitem[\protect\citeauthoryear{{Schmidt} \& {Green}}{{Schmidt} \&
512: {Green}}{1983}]{ScGr1983}
513: {Schmidt}, M., \& {Green}, R.~F. 1983, \apj, 269, 352
514:
515: \bibitem[\protect\citeauthoryear{{Weymann} et~al.}{{Weymann}
516: et~al.}{1991}]{WeMoFoHe1991}
517: {Weymann}, R.~J., {Morris}, S.~L., {Foltz}, C.~B., \& {Hewett}, P.~C. 1991,
518: \apj, 373, 23
519:
520: \end{thebibliography}
521:
522: \end{document}
523:
524:
525:
526:
527: