0707.2099/ms.tex
1: 
2: \documentclass[12pt,preprint]{aastex}
3: % \documentclass{article}
4: % \usepackage{emulateapj}
5: %\usepackage{float,epsfig,psfig}
6: %\usepackage{natbib}
7: %\bibpunct{}{}{;}{a}{}{,}
8: 
9: \newcommand{\LA}{\mbox{\raisebox{-0.6ex}{$\stackrel{\textstyle<}{\sim}$}}}
10: \newcommand{\GA}{\mbox{\raisebox{-0.6ex}{$\stackrel{\textstyle>}{\sim}$}}}%%%%%%
11: \newcommand{\cxo}{{\sl Chandra}}
12: \newcommand{\ngc}{{NGC~7331}}
13: \newcommand{\msun}{M$_{\odot}$}
14: \newcommand{\ergl}{ergs~s$^{-1}$}
15: \newcommand{\ctss}{cts~s$^{-1}$}
16: %\newcommand{\cxou}{CXOU~J223706.6+342620}
17: \newcommand{\cxou}{CXOU~J223706}
18: \newcommand{\ergcms}{ergs~cm$^{-2}$~s$^{-1}$}
19: \newcommand{\ha}{H$\alpha$}
20: \newcommand{\hb}{H$\beta$}
21: \newcommand{\hei}{He~{\sc i}}
22: \newcommand{\oiii}{[O~{\sc iii}]}
23: \newcommand{\oi}{[O~{\sc i}]}
24: \newcommand{\sii}{[S~{\sc ii}]}
25: \newcommand{\siii}{[S~{\sc iii}]}
26: \newcommand{\hii}{H~{\sc ii}}
27: \newcommand{\nii}{N~{\sc ii}}
28: \newcommand{\mdot}{$\dot{M}$}
29: \newcommand{\hst}{{\sl Hubble}}
30: \newcommand{\kms}{$km\,s^{-1}\,$}
31: \newcommand{\etal}{et al.}
32: \newcommand{\Myr}{Myr}
33: 
34: \renewcommand{\mag}[1]{^{\rm m}\!\!\!#1\,}
35: \newcommand{\AAA}{ \AA \/}
36: 
37: \slugcomment{Submitted to Astrophysical Journal}
38: 
39: \begin{document}
40: 
41: \title{Optical Spectroscopy of the environment of a ULX in NGC 7331}
42: 
43: \author{
44: Pavel K. Abolmasov\altaffilmark{1}, 
45: Douglas A. Swartz\altaffilmark{2},
46: S. Fabrika\altaffilmark{1,3},
47: Kajal K. Ghosh\altaffilmark{2},
48: O. Sholukhova\altaffilmark{1}, and
49: Allyn~F.~Tennant\altaffilmark{4}
50: }
51: \altaffiltext{1}{Special Astrophysical Observatory, Nizhnij Arkhyz 
52: 369167, Russia}
53: \altaffiltext{2}{Universities Space Research Association, NASA Marshall Space Flight Center, VP62, Huntsville, AL 35805, USA}
54: \altaffiltext{3}{University of Oulu, P.O. Box 3000, 90014, Finland}
55: \altaffiltext{4}{Space Science Department,
56:     NASA Marshall Space Flight Center, VP62, Huntsville, AL, USA}
57: 
58: \begin{abstract}
59: 
60: Optical photometric and spectroscopic data are presented that
61:  show an association of an ultraluminous X-ray source
62:  in \ngc\ with a young star cluster of
63:  mass $M = (1.1\pm0.2)\times 10^5$~\msun\ and age
64:  $t_c=4.25\pm0.25$ Myr.
65: %
66: If the ULX is part of the bright stellar cluster, 
67:  then the mass of the progenitor of the compact accretor must have been $\gtrsim$40-50~\msun\
68:  in order to already have evolved through the supernova stage 
69:  to a compact object. 
70: The companion star is also likely an evolved massive star.
71: The emission line spectrum of the nebula surrounding the cluster can be 
72:  interpreted as a result of photoionization
73:  by the cluster OB stars with an additional source of shock excitation
74:  producing strong \sii, \oi\ and \nii\ lines. 
75: This additional source appears to be as much as five times more powerful than the supernovae and
76: stellar winds in the cluster can provide.
77: Additional mechanical energy input associated with the ULX itself 
78: can help explain the residual shock excited
79: line luminosities of the emission region. 
80: 
81: \end{abstract}
82: 
83:  \keywords{galaxies: individual (NGC 7331) --- galaxies: star clusters --- line: formation --- X-rays: galaxies --- X-rays: binaries --- X-rays: individual (CXOU~J223706)}
84: 
85: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
86: \section{Introduction}
87: 
88: Ultraluminous X-ray sources (ULXs) are point-like non-nuclear sources in external galaxies
89:  with apparent X-ray luminosities greater than the Eddington limit of stellar mass
90:  black holes, i.e., $L_{\rm X}$\GA$10^{39}$~\ergl.
91: The existence of ULXs has been known since 
92:  the first {\sl Einstein} observations of nearby galaxies
93:  \citep{lvs83, helfand84,fabbiano89}
94:  but only with the sub-arcsecond 
95:   imaging capability of the \cxo\ X-ray Observatory 
96:  have investigations at other wavelengths of the crowded 
97:  fields of nearby galaxies become viable.
98: 
99: Multi-wavelength investigations enable a much broader 
100:  range of diagnostic techniques in the study of ULXs than can
101:  be achieved by X-ray data analysis alone.
102: In particular, traditional methods in 
103:  optical imaging and spectroscopy can be used to constrain 
104:  the age, metallicity and physical state of the local environments of ULXs.
105: This, in turn, can help to determine the origin and history of the ULXs themselves
106:  and lead to a better understanding of the influence of ULXs on their surroundings.
107: 
108: Many of the models for ULXs favor young high-mass X-ray binary systems
109:  (see, e.g., the reviews by \citet{fw2006}; \citet{fabbiano2006}).
110: These objects should be preferentially associated with
111:  massive star forming regions such as OB associations and,
112:  at the high-mass limit, superstar clusters.
113: These environments should be strong sources of line emission
114:  through photoionization by massive stars, supernovae (SNe) and by the ULX itself
115:  and may display enhanced forbidden line emission from
116:  shock excitation by stellar winds, SNe, and by energetic outflows and jets that
117:  could accompany the ULX X-ray emission.
118: Nebulae such as those associated with the ULXs in Ho~IX 
119:  \citep{GPM1},
120:  NGC 5408 \citep{soria5408}, NGC 5204 \citep{goad5402},  and Ho~II
121:  \citep{hoii_kaaret,Leh_ea05} are among the best-studied
122:  examples of this intricate ULX/star-formation connection.
123: 
124: As part of a program of spectroscopic studies of the 
125:  local envrironments of ULX candidates, we have obtained a moderate-resolution
126:  long-slit spectrum of a ULX in \ngc.
127: The X-ray characteristics of this ULX and the optical photometric properties of 
128:  its environment are described in \S~\ref{s:OBS}. 
129: The ULX lies roughly in the center of an extended \hii\ region,
130: cataloged as P98 \citep{Petit98},   %% PA
131:  and within 
132:  30~pc of the center of a bright blue emission region. 
133: The optical spectrum of this emission region is presented 
134:  in \S~\ref{s:SPEC}.
135: We model its spectrum as emission from a young star cluster combined with a
136: nebula photoionized by a cluster hosting an X-ray source (\S~\ref{s:RESULTS}).
137: This model reproduces the observed $4000 - 7300$~\AAA\ continuum spectrum but 
138:  under-predicts several strong observed emission lines. 
139: We argue (\S~\ref{s:shocks}) these lines are due, at least in part, to shock excitation
140:  associated with the ULX and that radiative and mechanical feedback plays
141:  an important part in the overall energetics of the emission region.
142: We briefly summarize our results and compare the 
143:  properties of the environment of the ULX in \ngc\ to other ULX nebulae
144:  in \S~\ref{s:discussion}.
145: 
146: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
147: \section{X-ray and Optical Observations} \label{s:OBS}
148: 
149: \ngc\ is a highly-inclined, i = 77$^\circ$ \citep{ggal},
150: SA(s)b spiral galaxy subtending 10.\arcmin 5
151: on the sky. We adopt a distance of 14.6~Mpc \citep{ferrarese};
152: 1\arcsec\ corresponds to 71~pc.
153: 
154: Swartz \etal\ (2004) identified three ULX candidates within the $D_{25}$
155:  isophote of the \cxo\ field of \ngc. 
156: All three lie in the northeast portion of the spiral arm structure
157:  beyond the large ring seen in the Spitzer Infrared Nearby Galaxies Survey 
158:  images  
159:  \citep{sings}.
160: An \ha\ image of this region is shown in Figure~\ref{f:ha_image} with the
161:  approximate positions of the three ULX candidates superposed.
162: Only the brightest ULX candidate has an obvious association with optical
163:  emission features in this or any other images we have studied. 
164: For the remainder of this work, we will refer to this ULX, which is formally
165:  designated CXOU~J223706.6+342620, as \cxou.
166: 
167: 
168: 
169: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
170: \subsection{\cxo\ observations}
171: 
172: \cxo\ obtained a 29.5~ks observation of a portion of \ngc\ on 12 Aug 2001
173: (ObsID 2198).
174: The observation was taken in sub-array mode using the Advanced CCD Imaging 
175:  Spectrometer (ACIS) with the galaxy center imaged near the aimpoint on the 
176:  back-illuminated S3 CCD.
177: \cxou\ is approximately 1.\arcmin 85 off-axis in this observation.
178: Using the known position of the nucleus\footnote{$\alpha = 22^h37^m04^s\!.096, \ \delta = +34^\circ24^\prime56^{\prime\prime}\!.29$ (J2000); \citet{opos}.}  of \ngc, 
179: we estimate the absolute positional
180:  uncertainty of the ULX is $\sim0\farcs4$.
181: Within a 3$\sigma = 1\farcs86$ circle about the position of the source,
182:  175 X-ray events were detected corresponding to a S/N of 10.6 
183:  (in the 0.5--8.0 keV band).
184: Acceptable fits to the spectrum can be made using a variety of models. The
185:  best-fitting model attempted is an absorbed power law 
186:  ($N_{H} = 6.9_{-2.4}^{+3.1} \times 10^{21}$~cm$^{-2}$, $\Gamma = 2.26_{-0.52}^{+0.63}$,
187: $\chi^2=5.1$ for 5 dof after grouping to ensure at least 20~counts per spectral bin).
188: The corresponding intrinsic source luminosity is 
189:  $L_{\rm X} = (2.8\pm0.6)\times 10^{39}$~\ergl\ 
190:  according to our previous analysis \citep{swartz}.  
191: (Here and elsewhere, errors are extremes on the single interesting parameter 1$\sigma$ confidence intervals.)
192: The luminosity of \cxou\ lies at the median value of the 97 ULX candidates in our sample 
193:  of ULXs in spiral galaxies. 
194: The average luminosity for that subsample is $5.9\times 10^{39}$~\ergl\
195:  and the highest luminosities exceeded several $10^{40}$~\ergl.
196: If we assume the X-ray luminosity approximates the bolometric luminosity and 
197:  that the source radiates isotropically at the Eddington limit
198:  during the \cxo\ observation, then the 
199:  mass of the compact accretor is 22$\pm$5~\msun.
200: 
201: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
202: \subsection{\hst\ observations}\label{s:hst}
203: 
204: \hst/WFPC2 observations\footnote{Available from http://archive.stsci.edu/} of NGC7331 were carried out on 
205:  1997 August 13 and 14 using F450W and F814W filters.
206: Astrometric corrections were made using USNO stars
207: \citep{usno}. 
208: The 1$\sigma$ nominal
209: absolute positional uncertainities in the \hst\ images are 
210: $0\farcs3$ giving a combined uncertainty of 
211: about $0\farcs5$ between the \hst\ and \cxo\ registration.
212: 
213: Figure~\ref{f:FOV} displays an 8\arcsec $\times$8\arcsec\ 
214:  portion of the F450W \hst\ image around \cxou.  
215: The \hst\ \ image is dominated by two bright extended sources; presumably young
216:  stellar clusters. The X-ray source
217:  is located within the NW cluster, within the formal error.
218: The optical source has a FWHM of $\sim0\farcs1$ or 7~pc at the distance of \ngc\
219: (the source is not completely symmetric, \hst\ images reveal 
220: additional emission in the wing).
221: The total background-corrected brightness
222: within a $1\farcs5$ diameter circle (equivalent to the seeing conditions in the 
223:  ground-based data, see below) centered on the X-ray source position
224:  is $m_{\rm F450W}=20.44$~mag.
225: This corresponds to an absolute magnitude $M_{450W} = -10.4$~mag,
226:  lying at about the maximal luminosity of the Galactic open 
227:  clusters \citep{larsen}. 
228: However, taking into account interstellar absorption
229:  (\S~\ref{s:RESULTS}), the object intrinsic brightness may be significantly higher.
230: 
231: 
232: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
233: \subsection{Isaac Newton Telescope observations} \label{s:INT}
234: 
235: Archival optical images of \ngc, obtained with the Wide Field Camera (WFC) on the
236: Isaac Newton Telescope (INT) at La~Palma\footnote{Available from http://casu.ast.cam.ac.uk/casuadc/archives/ingarch}, 
237: were examined for structure local to \cxou.
238: We used images taken in narrow filters \oiii\ (central wavelength 5008~\AA, FWHM=100~\AA),
239: H$\alpha$ (6568~\AA, 95~\AA) and in the Harris broad-band filters B (4298~\AA, 1065~\AA),
240: V (5425~\AA, 975~\AA) and R (6380~\AA, 1520~\AA). All the imaging observations were
241: carried out in 2000 and 2001. We have estimated seeing from the FWHM of unsaturated
242:  point-like sources in the images to be 
243: $0\farcs7$ in H$\alpha$, $0\farcs8$ in \oiii\ and from $0\farcs9$ to $1\farcs2$ in the broad bands. 
244: 
245: Astrometric corrections were made using USNO-B1 and USNO-B2 standards 
246: \citep{usno}. The astrometrical error obtained in all these images 
247: (using 22 to 27 astrometrical standard 
248:  stars) is less than $0\farcs2$.
249: We made photometric calibration of the B- and V-band images using 
250: 19 USNO standard stars that range in brightness from 13.0 to 20.5 mag.
251: A systematic error of the USNO photometric standards as high as 0.2-0.3 
252: magnitude is possible for the fainter stars \citep{usno}. 
253: 
254: We created new \ha\ and \oiii\ images free of continuum emission by dividing the narrow band 
255: images over the R and V images, respectively. In the resulting images the majority 
256: of single stars (presumably non-emission-line stars) disappear.
257: From these images one may estimate the spatial structure in the emission lines
258: and measure the corresponding emission line equivalent widths.
259: Isophotes at the 20, 50 and 85\% levels above the background in these \ha\ and \oiii\ 
260: images are shown in Figure~\ref{f:FOV} (100\% corresponds to the maximum value in the 
261: 8\arcsec$\times$8\arcsec\ field). 
262: Taking the 50\% isophote as representative of the extent 
263: of the emission in these lines, one obtains the size of the emission-line
264: region to be $\sim$2\arcsec\ in \ha\ and \oiii\ 
265: corresponding to 140 pc (FWHM).
266: The peak of the \ha\ emission is coincident with the NW cluster resolved in the \hst\ image.
267: The peak \oiii\ emission lies between the two clusters suggesting each cluster contributes 
268:  equally to the \oiii\ emission.  
269: 
270: The estimated B-band magnitude within a circle of $1\farcs5$ diameter around the NW
271: cluster is $19\mag{.}97\pm0.06$ (USNO-B2) and $20\mag{.}32\pm0.06$ (USNO-B1).
272: For modelling the optical spectrum (see below),
273: we adopt the value B=$20\mag{.}0$.
274: 
275: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
276: \section{Optical Spectroscopy} \label{s:SPEC}
277: 
278: Optical spectral data were obtained using the 6m Russian telescope 
279:  with the SCORPIO 
280: focal reducer \citep{scorpio} in the long-slit (LS) mode  
281: with a $0\farcs75$-wide slit, providing
282: spectral resolution of about 10 \AA.
283: The data cover the spectral range $4000-7300$ \AA.
284: The object was observed on August 29, 2005
285: at $1\farcs4$ seeing conditions.
286: The reduction process includes all the standard procedures.
287: 
288: The spectrum is shown in Figure~\ref{fig:ospect_mod}. 
289: Many strong emission lines are present above a well-defined blue continuum.
290: Table~1 lists the observed and dereddened integral fluxes in all the emission
291: lines detected.
292: 
293: 
294: We extracted slit images in the strongest lines (H$\beta$, \oiii, \ha, and \sii).
295: The extent of the \oiii\ and \ha\ emission in the slit images is consistent 
296:  with the structure seen in the INT images.
297: The emission in the slit images can best be described as the sum of a point-like
298: structure (FWHM$\sim1\farcs5$ consistent with the seeing conditions)
299: and more extended (FWHM$\sim5\farcs0$) emission. 
300: That is, the two clusters are not resolved.
301: The centroids of the point-like emission 
302: in all the lines are coincident within their errors;
303: but in the extended component the \oiii\ centroid 
304: is shifted about $0\farcs4$ to the SSE relative to \ha\ along the slit 
305: (cf. Figure~\ref{f:FOV}). The \sii\ centroid is shifted in the opposite direction,
306: about $0\farcs2$ to the NNW relative to the \ha\ peak.
307: 
308: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
309: \section{Spectrum modelling} \label{s:RESULTS}
310: 
311: The size and luminosity of the bright sources in the \hst\ blue (F450W) image 
312:  suggest they are compact young star clusters. 
313: The INT narrow-band images and SAO slit images suggest the star clusters are 
314: embedded in a larger emission line nebula
315:  (with the slit sampling most of the cluster emission but not all of the 
316:  larger emission line region).
317: We constructed synthetic spectra based on a model for the emission 
318:  consisting of an underlying young stellar population 
319:  and a photoionized nebula. 
320: Starburst99 \citep{sb2005} was used for the stellar contribution which 
321:  provides model spectra for the continuum emission as functions of age
322:  and metallicity of the cluster and normalized by the cluster mass.
323: CLOUDY96.01 \citep{cloudy} was used to simulate the emission line spectrum
324:  from the photoionized nebula as a function of the input photoionizing source
325:  spectrum and ionization parameter, the nebular abundances, and the density.
326: The spectral modelling began by finding the best-fitting 
327:  cluster model. This model was then subtracted from the observed spectrum
328:  and the best-fitting model 
329:  for the residual spectrum was found using the photoionization code.
330: 
331: \subsection{The contribution from the Star Cluster}
332: 
333: The oxygen abundance and abundance gradient in \ngc\ is given by \citet{pilyugin}. 
334: At the location of \cxou, about 90\arcsec\ from the center
335:  of the galaxy, 
336: $12+\log ({\rm O/H}) = 8.32$ or about 2--3 times less than the solar value.
337: Using this as a guide,
338: we constructed Starburst99 models with [Fe/H]$=$0.0, $-0.4$, $-0.7$.
339: %% The results are not sensitive to metallicities in this range.
340: We computed a grid of models for cluster ages, $t_c$, ranging from 1 to 30~Myr.
341: The model spectra were fit to the observed spectra allowing for a 
342: variable interstellar absorption component (using the reddening curve of 
343: \citet{CCM}).
344: Since the Starburst99 stellar cluster model does not predict 
345: a nebular emission line spectrum it was necessary to exclude 
346: all the bright observed emission lines in the model fitting. 
347: We find the best-fit parameters of the cluster model are
348:  $t_c=(4.25\pm0.25)$~Myr, $A_V = 1\mag{.}43 \pm 0\mag{.}05$,
349:  $M = (1.1\pm 0.2)\times10^5$~\msun\ where $M$ is the cluster mass 
350:  which scales linearly with the overall model normalization. 
351: The $\chi_r^2$ statistic for this model is 1.30 for $\sim$200 degrees of freedom
352: (the spectral range divided by the spectral resolution element with emission 
353: lines omitted and reduced by the number of model fit parameters). 
354: Solar metallicity provides a significantly better fit.
355: The best-fit model spectrum (scaled by a factor of $1/3$ for clarity) is shown
356: as the lower curve in Figure~\ref{fig:ospect_mod}.
357: 
358: The age and metallicity derived here are consistent with the presence of
359: the broad but weak emission features observed at 4650~\AA\ and 5806~\AA.
360: We interpret these features as the C~{\sc iii}$+$N~{\sc iii} and
361:  C~{\sc iv} line blends, respectively, due to the presence of Wolf-Rayet 
362: stars \citep{wr_conti,SV98} in the cluster. We estimate the equivalent
363: widths to be  1.7$\pm$0.2~\AA\ for the 
364: blue blend and 4.2$\pm$0.2~\AA\ for the red blend of lines. 
365: (The quoted errors are formal errors using a piece-wise linear approximation to
366: the local continua; inspection of Figure~\ref{fig:ospect_mod} suggests 
367: the uncertainties are likely higher.)
368: The theoretical blue-to-red ratio \citep{SV98} 
369: most closely approaches the observed ratio at 3--6 Myr where the 
370: theoretical ratio is $\approx$1 for a wide range of metallicities.
371: 
372: \subsection{The contribution from photoionization}
373: 
374: The best-fit stellar population model spectrum was subtracted 
375:  from the dereddened observed spectrum to give an estimate 
376:  of the purely nebular spectrum. 
377: We used the $3\le t_c \le 6$~Myr
378:  UV spectrum from the best-fit unabsorbed cluster model plus an additional
379:  X-ray component based on the \cxo\ spectrum of \cxou\
380:  ($\Gamma=2.2$, 3.2$\times$10$^{48}$~photons s$^{-1}$ blueward of 
381:  the Ly~edge) for the photoionization source for the nebula model.
382: We calculated a set of spherically-symmetric photoionization
383:  models with fixed inner (1~pc) and outer (30~pc) radii using CLOUDY96.01. 
384: The inner radius is typical for a massive young star cluster 
385: (e.g., \citet{OCnilakshi}).
386: The outer radius corresponds to the
387:  half-width of the slit in projection (at the galaxy distance of 14.6~Mpc).
388: We have found that the choice of
389:  inner and outer radii do not strongly affect the results.
390: The nebula gas density and the low-energy cutoff for the ionizing 
391:  X-ray spectrum were allowed to vary in the modelling.
392: The best-fitting gas density is about 30~cm$^{-3}$
393:  (consistent with the observed \sii\ $\lambda6717/\lambda6731$ ratio).
394: The models proved insensitive to 
395:  the additional X-ray and corresponding EUV radiation from the ULX.
396: This was anticipated because the number flux from the cluster model
397:  in the Ly~continuum 
398:  ($\sim$10$^{51}$~s$^{-1}$) far exceeds the flux of X-ray photons
399:  even after extrapolating this rather steep X-ray spectrum to the Ly~edge.
400: 
401: The results of the photoionization model predictions and the observed 
402:  emission line parameters are listed in Table~1.
403: For each emission line observed, the ratio of the observed
404:  (Column~2) and dereddened (Column~3; $A_V = 1\mag{.}4$) line fluxes to the 
405:  H$\beta$ flux are tabulated along with the integral dereddened
406:  line luminosity (Column~4), the photoionization model predicted line
407:  luminosity (Column~5), and the residual line luminosity after subtracting
408:  the photoionization model prediction (Column~6).
409: The best-fit stellar continuum model combined with the best-fit photoionization
410:  model (convolved with the instrumental response function) is shown  
411:  as the middle (dotted) curve in Figure~\ref{fig:ospect_mod}
412: (displaced by $1/2$ for clarity).
413: 
414: \subsection{The contribution from shocks} \label{s:shocks}
415: 
416: These pure photoionization models predict strong \oiii, He~{\sc i}, 
417:  and H Balmer line emission excited by the UV emission from the underlying
418:  star cluster
419:  but are unable to reproduce the observed strength of several key emission lines
420:  (see Table~1
421: and Figure~\ref{fig:ospect_mod}).
422: A significant excess is observed in 
423:  \sii\ $\lambda\lambda$6716, 6731,
424:  [O{\sc i}] $\lambda\lambda$6300, 6364 
425: and [N{\sc ii}] $\lambda\lambda$6548, 6583~\AA\ doublets.
426: We find the additional dereddened luminosity in the \sii\ doublet, for 
427:  example, is $L_{\rm{[SII]}}=1.72\times 10^{38}$~\ergl.
428: Further, the residual \oiii\ $\lambda(4959+5007)/\lambda 4363$ ratio is
429:  low, $\sim$20, compared to typical photoionized \hii\ regions.
430: 
431: The observed excesses in these low-excitation lines and the low 
432:  \oiii\ ratio are likely caused by  
433:  shock excitation in the nebula (cf. \citet{mappings}). 
434: We can identify three plausible sources for shock excitation: 
435: ({\it i}) individual SNe in the stellar cluster, 
436: ({\it ii}) the cumulative stellar wind in the cluster, and 
437: ({\it iii}) the ULX source. 
438: Here we make only rough estimates of the possible contributions from these three 
439: sources; we lack sufficient knowledge of the physical conditions in the nebula
440:  (e.g., shock velocity, pressure, density, temperature, and pre-shock ionization)
441:  to apply a full shock-excitation model for the emission lines. 
442: 
443: The Starburst99 models predict the total supernova production rate
444:  in a cluster of age $t_c \sim 4.25$~Myr and mass $M \sim 10^5$~\msun\
445:  and a Salpeter IMF is about $6\times 10^{-5}$~yr$^{-1}$.  %% PA
446: Since SNRs are visible in emission lines for
447:  up to $10^5$ years \citep{Cioffi}, the effective
448:  number of SNRs contributing to the spectrum is at most $\sim 6$. 
449: Typical H$\alpha$ and \sii\
450:  luminosities from an individual SNR are 
451:  $L_{{\rm H}\alpha} \sim L_{\rm [SII]} \sim (3-6) \times 10^{36}$~\ergl\ 
452:  \citep{braunm31,MF_cat}.
453: The expected luminosity in these lines from 6 SNRs is then 
454:  $\sim (1.8-3.6) \times 10^{37}$~\ergl, or 5 to 10 times less than the 
455:  residual luminosities listed in Table~1. 
456: 
457: At the estimated age of 4.25~Myr, the Starburst99 model predicts equal contributions
458: from SNe and stellar winds to the mechanical luminosity of the cluster. If the 
459: conversion efficiencies from mechanical luminosity input to emission line radiation
460: are the same for SNe and stellar winds, winds will contribute a comparable 
461: amount to the shock-excited lines. The combination of SNe and stellar winds still leaves 
462: the \sii\ $\lambda\lambda$6717, 6731 
463: production about 2.5 to 5 times below that needed to explain the observed luminosity. 
464: 
465: The third plausible source for shock excitation  is the ULX.
466: The presence of such an object in the cluster may contribute to
467: the residual line emissions if mechanical energy from the X-ray source can be
468:  imparted to its surroundings in the form of an outflow. It has been postulated 
469: that some ULXs are microquasars \citep{Kingetal01}, whose radiation is
470: collimated along the jet axis, or supercritical accretion disks with outflows
471: like that in SS433 \citep{FaMe01}. 
472: 
473: In the case of microquasars, outflow in the 
474:  form of a leptonic jet could occur during the low state and during outbursts
475:  associated with transitions within the very-high/intermediate states
476:  (Fender, Pooley, \& Gallo 2004; Corbel \etal\ 2003).
477: However, utilizing the known correlation between X-ray and radio luminosities for
478:  black holes in the low-hard X-ray state (Corbel \etal\ 2003;
479:  Gallo, Fender, \& Pooley 2003) to estimate the mechanical power in the jet and 
480: assuming the \cxo\ observation of \cxou\ was obtained when the source 
481:  was near Eddington, then the power 
482:  in the jet is $L_J \lesssim  10^{38}$~\ergl\ even allowing for additional
483:  transient ejection events (Fender, Belloni, \& Gallo 2004).
484: The mechanical power imparted to the surrounding nebula in the form
485:  of outflow from microquasar-like jets from the ULX is then  
486:  far less than needed to explain the residual line emission.
487: [Note that the Galactic microquasars like 1E~1740.7-2942 and GRS~1758-258
488:  (Mirabel \etal\ 1993), %% PA 
489: for instance, have presistent lobe-like emission
490:  but are old stellar systems lacking associations with young energetic nebulae.]
491: 
492: In the case of SS433-like objects, outflow is in the form of persisent
493:  heavy jets that impart a large fraction of their accretion energy into their 
494:  surroundings in the form of mechanical energy.
495: The observed kinetic luminosity in the two jets of SS433
496: is $\sim 2 \times 10^{39}$~\ergl\ \citep{Fab04}. 
497: This is comparable to the predicted mechanical luminosity from stellar winds and SNe
498:  in the star cluster associated with \cxou, $3.5 \times 10^{39}$~\ergl,
499:  at its age of 4.25~Myr.
500: 
501: Likely all three sources of shock excitation contribute to the observed 
502:  line emission. The ULX can be an important, and even dominant, source of 
503:  radiative and mechanical feedback in the overall energetics of the emission region.
504: 
505: \section{Discussion} \label{s:discussion}
506: 
507: We have presented optical photometric and spectroscopic data that 
508:  help to describe the physical environment surrounding the ULX \cxou\ in \ngc.
509: \cxou\ is spatially-coincident (within astrometrical errors) with a bright,
510: young, $t_c=4.25\pm0.25$~Myr, star cluster of mass
511: $M = (1.1\pm0.2)\times 10^5$~\msun.
512: If the ULX is a part of this bright stellar cluster, 
513:  then the progenitor of the compact accretor must have been at least as massive as 40-50~\msun\
514:  in order to have already evolved through the supernova stage. 
515: The companion star is also likely massive and evolved in order to provide the 
516:  high mass accretion rate needed to power the ULX.
517: 
518: The emission line spectrum of the surrounding nebula can be 
519:  interpreted as a result of photoinization produced
520:  by the cluster UV spectrum, but an additional source of shock excitation
521:  is needed to understand the observed \sii, [O{\sc i}] and [N{\sc ii}] 
522:  line intensities. 
523: The mechanical input from SNe and stellar winds from the star cluster 
524:  can account for only
525: $\sim$20--40\% of the required shock excitation so some additional energy 
526: source is needed. The obvious candidate is the ULX itself.
527: 
528: Emission nebulae associated with nearby ULXs are 
529: apparently a common phenomenon.
530: Although no formal census has been undertaken, we find potential associations
531:  with bright, extended optical structures for 67 of the 154 ULX
532:  candidates tabulated in the ULX survey of Swartz \etal\ (2004). 
533: Although these associations are based mainly upon Digitized Sky Survey images
534:  and must await more conclusive analysis, many detailed studies 
535:  of ULXs with associated nebulae have appeared in
536:  the recent literature \citep{Pakul_Mir03,Leh_ea05,
537: kuntzm101, Pakul06, GPM1, GPM2,Abol_ea07}.
538: 
539: In particular, the nebula associated with 
540:  \cxou\ in \ngc\ can be compared with  the compact nebula MF\,16 associated with the ULX
541: NGC\,6946~X-1 \citep{MF16_BFS}. 
542: The MF\,16 nebula is an isolated object: It does not appear to be 
543:  contaminated by associated star
544:  clusters or nearby H\,II regions. 
545: The line luminosities of MF\,16 \citep{MF16_BFS,Abol_ea06} are 
546:  $L_{{\rm [SII]}} \approx L_{{\rm H\alpha}} \approx 2 \times 10^{38}$~\ergl, 
547:  $L_{{\rm H\beta}} \approx 6
548: \times 10^{37}$~\ergl; equal to  
549: the residual line luminosities in the \ngc\ nebula (Table~1). 
550: On the other hand, \oiii\ $\lambda\lambda 4959, 5007$ and \oi\ $\lambda 6300$ \AA\
551: lines in MF\,16 are 3.5 times brighter than those in the \ngc\ residuals. These two lines
552: are excited in different ways: by UV photoionization in the case of
553: \oiii\ $\lambda\lambda 4959, 5007$ and by shock excitation in the case of 
554: \oi\ $\lambda 6300$ \AA. 
555: 
556: Additional sources of mechanical energy in the form of more or less collimated winds
557: or jets is not unprecedented in ULX nebulae.
558: In the nebulae of ULXs NGC\,6946~X-1 and
559: Holmberg\,II~X-1 (another ``isolated'' ULX nebula), radial velocity gradients of $50-100$~km~s$^{-1}$ have been found 
560: \citep{Leh_ea05,Fab_ea06}. It was concluded that the nebulae have to be
561: powered by the central ULXs and that, therefore, the ULXs must also provide a source of gas collisional
562: excitation in the ULX-associated nebulae. 
563: %%SF
564: It was found recently through optical spectroscopy of 8 ULX-associated nebulae
565: \citep{Abol_ea07} that in at least half of them shock excitation
566: dominates. We cannot say definitely about the kinematical properties of the gas in 
567: the \cxou\ counterpart -- we lack both spectral and spatial 
568: resolution for this -- but it is likely the ULX plays an important role here as well. 
569: 
570: \acknowledgements
571: The authors thank N.\,Borisov for help with the observations.
572: This work has been supported in part by Russian
573: RFBR grants NN\, 06-02-16865 and 07-02-00909
574: and by a joint RFBR/JSPS grant N\,05-02-12710.
575: Additional support was provided by NASA under grant NNG04GC86G issued 
576: through the Office of Space Science and by the 
577: Space Telescope Science Institute under the grant HST/AR-10954.
578: 
579: 
580: \begin{thebibliography}{}
581: 
582: \bibitem[Abolmasov et al.(2006)]{Abol_ea06}
583: Abolmasov, P., Fabrika, S., Sholukhova, O. \& Afanasiev, V. 2005,
584: in Science Perspectives for 3D Spectroscopy, ed. 
585: M. Kissler-Patig, M. M. Roth. \& J. R. Walsh
586: (Springer Berlin / Heidelberg); astro-ph/0602369
587: 
588: \bibitem[Abolmasov et al.(2007)]{Abol_ea07}
589: Abolmasov, P., Fabrika, S., Sholukhova, O. \& Afanasiev, V. 2007,
590: Astrophysical Bulletin, 62, N\,1, 36; astro-ph/0602369
591: 
592: \bibitem[Afanasiev \& Moiseev(2005)]{scorpio}
593: Afanasiev, V. \& Moiseev, A. 2005, Astronomy Letters, 31, 194
594: 
595: \bibitem[Argyle \& Clements(1990)]{opos}
596: Argyle, R. W. \& Clements, E. D. 1990, The Observatory, 110, 93
597: 
598: \bibitem[Blair et al.(2001)]{MF16_BFS}
599: Blair, W. P., Fesen, R. A. \& Schlegel, E. M.  2001, AJ, 121, 1497
600: 
601: \bibitem[Braun \& Walterbos(1993)]{braunm31}
602: Braun, R. \& Walterbos, R. A. M. 1993, 
603: A\&ASS, 98, 327
604: 
605: \bibitem[Cardelli et al.(1998)]{CCM}
606: Cardelli, J. A., Clayton, G. C. \& Mathis, J. S. 1998
607: ApJ, 345, 245
608: 
609: \bibitem[Cioffi et al.(1988)]{Cioffi}
610: Cioffi, D. F., McKee, C. F. \& Bertschinger, E. 1988
611: ApJ, 334, 252
612: 
613: \bibitem[Conti et al.(1983)]{wr_conti}
614: Conti, P. S., Leep, M. E. \& Perry, D. N. 1983
615: ApJ, 268, 228
616: 
617: \bibitem[Dopita et al.(2000)]{mappings}
618: Dopita, M. A., Kewley, L. J., Heisler, C. A. \& Sutherland, R. S.
619: 2000 ApJ, 542, 224
620: 
621: \bibitem[Dopita \& Sutherland(1996)]{DoSutI}
622: Dopita, M. A. \& Sutherland, R. S. 1996, ApJSS, 102, 161
623: 
624: \bibitem[Fabbiano(1989)]{fabbiano89} 
625: Fabbiano, G. 1989, ARA \& A, 27, 87
626: 
627: \bibitem[Fabbiano(2006)]{fabbiano2006}
628: Fabbiano, G. 2006, ARA\&A, 44, 323
629: 
630: \bibitem[Fabbiano \& White(2006)]{fw2006}
631: Fabbiano, G. \& White, N. E. 2006, in Compact stellar X-ray sources, ed. W. Lewin \& M. van der Klis, Cambridge Astrophysics Series No. 39, (Cambridge:Cambridge University Press), 475
632: 
633: \bibitem[Fabrika \& Mescheryakov(2001)]{FaMe01}
634: Fabrika, S., Mescheryakov A. 2001.
635:       In: \textit{Galaxies and their Constituents at the Highest Angular
636:       Resolution}, IAU Symp. N205, ed. R.T. Schilizzi, p.\ 268
637: 
638: \bibitem[Fabrika(2004)]{Fab04}
639:       Fabrika, S., 2004, Astrophys. and Space Phys. Rev., 12, 1
640: 
641: \bibitem[Fabrika et al.(2006)]{Fab_ea06}
642: Fabrika, S., Karpov, S., Abolmasov, P. \& Sholukhova, O. 2006.
643: In Populations of High Energy Sources in Galaxies, IAU Symposium 230,
644: ed. E. J. A. Meurs \& G. Fabbiano (Cambridge : Cambridge University Press);
645: astro-ph/0510491
646: 
647: \bibitem[Fender et al.(2004)]{FeBeGa04}
648: Fender, R. P., Belloni, T. P. \& Gallo, E. 2004, MNRAS, 355, 1105
649: 
650: \bibitem[Ferland et al.(1998)]{cloudy}
651: Ferland, G. J., Korista, K. T., Verner, D. ., Ferguson, J. W., Kingdon,
652: J. B. \& Verner, E. M. 1998, PASP, 110, 761
653: 
654: \bibitem[Ferrarese et al.(2000)]{ferrarese}
655: Ferrarese, L., et al.
656: 2000, ApJ, 529, 745
657: 
658: \bibitem[Gallo et al.(2003)]{GaFePo03}
659: Gallo, E., Fender, R. P. \& Pooley, G. G. 2003, MNRAS, 344, 60
660: 
661: \bibitem[Garc\'{\i}a-G\'{o}mez et al.(2002)]{ggal}
662: Garc\'{\i}a-G\'{o}mez, C., Athanassoula, E. \& Barber, C. 2002,
663: A \& A, 389, 68
664: 
665: \bibitem[Goad et al.(2002)]{goad5402}
666: Goad, M. R., Roberts, T. P., Knigge, C., \& Lira, P.
667: 2002, MNRAS, 335, 67
668: 
669: \bibitem[Gris\'{e} et al.(2006a)]{GPM1}
670: Gris\'{e}, F., Pakull, M. W. \& Motch, C. 2006,
671: in Populations of High Energy Sources in Galaxies, IAU Symposium 230, 
672: ed. E. J. A. Meurs \& G. Fabbiano (Cambridge : Cambridge University Press);
673: astro-ph/0603768
674: 
675: \bibitem[Gris\'{e} et al.(2006b)]{GPM2} 
676: Gris\'{e}, F., Pakull, M. W. \& Motch, C. 2006,  %% preprint, astro-ph/0603769 
677: in Populations of High Energy Sources in Galaxies, IAU Symposium 230, 
678: ed. E. J. A. Meurs \& G. Fabbiano (Cambridge : Cambridge University Press) 
679:  
680: \bibitem[Helfand(1984)]{helfand84}
681: Helfand, D. J. 1984, PASP, 96, 913
682: 
683: \bibitem[Kaaret et al.(2004)]{hoii_kaaret}
684: Kaaret, P., Ward, M. J., \& Zezas, A. 2004, MNRAS, 351, 83
685: 
686: \bibitem[King et al.(2001)]{Kingetal01}
687:       King, A. R., Davies, M. B., Ward, M. J., Fabbiano, G., Elvis, M. 2001,
688:       ApJ (Letters),  552, L109 (2001)
689: 
690: \bibitem[Kuntz et al.(2005)]{kuntzm101}
691: Kuntz, K. D., Gruendl, R. A. \& Chu, Y. -H. et al. 2005
692: ApJ, 620, 31
693: 
694: \bibitem[Larsen(2001)]{larsen}
695: Larsen, S. S. 2001,
696: in Extragalactic Star Clusters, IAU Symposium 207,
697: ed. D. Geisler, E. K. Grebel \& D. Minniti 
698: (Astronomical Society of the Pacific, San Francisco), 421
699: 
700: \bibitem[Lehmann et al.(2005)]{Leh_ea05}
701:       Lehmann, I., et al. 2005,
702:       A\& A, 431, 847
703: 
704: \bibitem[Leitherer et al.(1999)]{starburst1999}
705: Leitherer, C., et al. 1999, ApJSS, 123, 3
706: 
707: \bibitem[Long \& van Speybroeck(1983)]{lvs83}
708: Long, K. S. \& van Speybroeck, L. P. 1983,
709: Accretion Driven Stellar X-ray Sources,
710: ed. W. H.G. Lewin, \& E. P.J. van den Heuvel
711: (Cambridge: Cambridge University Press), 141
712: 
713: \bibitem[Matonick \& Fesen(1997)]{MF_cat}
714: Matonick, D. M., Fesen, R. A.,~1997 ApJSS,~112, 49
715: 
716: 
717: \bibitem[Mirabel et al.(1993)]{mir1993}
718: Mirabel, I. F., Rodriguez, L. F., Cordier, B., Paul, J. \& Lebrun, F.
719: 1993, A\& A SS, 97, 193
720: 
721: \bibitem[Monet et al.(2003)]{usno}
722: Monet. D. G., et al. 2003,
723: AJ 125, 984
724: 
725: \bibitem[Nilakshi et al.(2002)]{OCnilakshi}
726: Nilakshi, S. R., Pandey, A. K. \& Mohan, V. 2002, A\&A, 383, 153
727: 
728: \bibitem[Pakull \& Mirioni(2003)]{Pakul_Mir03} 
729: Pakull, M. W. \& Mirioni, L. 2003, in Winds, Bubbles, and Explosions: 
730: a conference to honor John Dyson,
731: RevMexAA (Serie de Conferencias) 15, 197, ed. J. Arthur \& W. J. Henney
732: 
733: \bibitem[Pakufll et al.(2006)]{Pakul06} 
734: Pakull, M. W., Gris\'{e}, F., Motch, C.  2006,
735: in Populations of High Energy Sources in Galaxies, IAU Symposium 230, 
736: ed. E. J. A. Meurs \& G. Fabbiano (Cambridge : Cambridge University Press);
737: astro-ph/0603771
738: 
739: \bibitem[Pilyugin et al.(2003)]{pilyugin}
740: Pilyugin, L. S., Thuan, L. X. \& V\'{i}lchez, J. M. 2003,
741: A\&A, 397, 487
742: 
743: \bibitem[Petit(1998)]{Petit98}
744: Petit, H. 1998, A\&AS, 131, 317
745: 
746: \bibitem[Regan et al.(2005)]{sings}
747: Regan, M. W., et al. 2004
748: ApJSS, 154, 204
749: 
750: \bibitem[Schaerer \& Vacca(1998)]{SV98}
751: Schaerer, D. \& Vacca, W. D. 1998, ApJ, 497, 618
752: 
753: \bibitem[Soria et al.(2005)]{soria5408}
754: Soria, R., Fender, R. P., Hannikainen, D. C., Read, A. M. \&
755: Stevens, I. R. 2006, MNRAS, 368, 1527 
756: 
757: \bibitem[Swartz et al.(2004)]{swartz}
758: Swartz, D. A., Ghosh, K. K., Tennant, A. F. \& Wu, K. 2004
759: ApJS, 154, 519
760: 
761: \bibitem[V\'{a}zquez \& Leitherer(2005)]{sb2005}
762: V\'{a}zquez, G. A. \& Leitherer, C. 2005, ApJ, 621, 695
763: 
764: %%additional ref%%%%%%%%
765: 
766: \end{thebibliography}
767: 
768: \clearpage
769: 
770: \begin{table} \label{tab:lines}
771: \begin{center}
772: \caption{Integrated Emission Line Fluxes and Luminosities}
773: \begin{tabular}{lccccc} 
774: \hline \hline 
775: line & $F / F(H_\beta)$ & $F / F(H_\beta)$ & $L$ & $L_{mod}$& $L_{res}$\\ \cline{4-6}
776:      &                  & (dereddened)     & \multicolumn{3}{c}{10$^{37}$ \ergl} \\ \hline
777:  H$\delta$                 &0.05$\pm$0.07    &0.211$\pm$0.012  &4.5$\pm$0.3    &3.8  & 0.7\\
778:  H$\gamma$                 &0.34$\pm$0.02    &0.444$\pm$0.016  &9.5$\pm$0.5    &6.8  & 2.7\\
779:  $[$O{\sc iii}$]$ $\lambda$4363 &0.03$\pm$0.01    &0.04$\pm$0.02    &0.87$\pm$0.6   &0.0007&0.87\\
780:  H$\beta$                  &1.000$\pm$0.05   &1.000$\pm$0.006  &21.45$\pm$0.19 &14.59 &6.86 \\
781:  $[$O{\sc iii}$]$ $\lambda4959$ &0.35$\pm$0.04    &0.292$\pm$0.008  &6.28$\pm$0.17  &2.32  & 3.96\\
782:  $[$O{\sc iii}$]$ $\lambda5007$ &1.15$\pm$0.05    &0.919$\pm$0.006  &19.72$\pm$0.19 &6.98  & 12.74\\
783:  $[$N{\sc i}$]$ $\lambda5200$   &0.08$\pm$0.04    &0.050$\pm$0.006  &1.08$\pm$0.17  &0.004 &1.08\\
784:  $[$N{\sc ii}$]$ $\lambda5755$  &0.02$\pm$0.03    &0.011$\pm$0.005  &0.24$\pm$0.11  &0.006 &0.24\\
785:  HeI $\lambda5876$   &0.13$\pm$0.04    &0.085$\pm$0.005  &1.81$\pm$0.14  &2.01  & -0.19\\
786:  $[$O{\sc i}$]$ $\lambda$6300$+$
787:  $[$S{\sc iii}$]$ $\lambda6310$ &0.15$\pm$0.04    &0.085$\pm$0.007  &1.83$\pm$0.15  &0.01  &1.81\\
788:  $[$O{\sc i}$]$ $\lambda6363$   &0.02$\pm$0.02    &0.008$\pm$0.003  &0.18$\pm$0.09  &0.004  &0.18\\
789:  $[$N{\sc ii}$]$ $\lambda6548$  &0.668$\pm$0.003  &0.349$\pm$0.002  &7.48$\pm$0.05  & 0.33 & 7.15\\
790:  H$\alpha$                 &6.375$\pm$0.009  &3.360$\pm$0.005  &72.09$\pm$0.16 &44.77  &27.32 \\
791:  $[$N{\sc ii}$]$ $\lambda6583$  &2.00$\pm$0.009   &1.046$\pm$0.005  &22.43$\pm$0.15 & 0.97  & 21.46\\
792:  HeI $\lambda6678$   &0.04$\pm$0.03    &0.023$\pm$0.004  &0.50$\pm$0.11  & 0.57  &-0.07\\
793:  $[$S{\sc ii}$]$ $\lambda6717$  &0.97$\pm$0.02    &0.493$\pm$0.011  &10.5$\pm$0.3   & 0.6  &10.0\\
794:  $[$S{\sc ii}$]$ $\lambda6731$  &0.71$\pm$0.02    &0.356$\pm$0.011  &7.6$\pm$0.3    & 0.4 & 7.2\\
795: \hline
796: \end{tabular}
797: \end{center}
798: \end{table}
799: 
800: \clearpage
801: 
802: \begin{figure}
803: \includegraphics[angle=0,width=0.97\columnwidth]{f1.eps}
804: \figcaption{Isaac Newton Telescope 
805:  \ha\ image of a portion of the \ngc\ field showing the location of 
806:  three ULXs in relation to the central regions of the galaxy. 
807: At the distance of 14.6~Mpc to \ngc, 1\arcmin\ corresponds to 4.25~kpc.
808: Circles denote  2\arcsec\ radius locations
809:  of the three ULX candidates detected. \cxou\ is the middle source.
810:  \label{f:ha_image} } 
811: \end{figure}
812: 
813: \clearpage
814: 
815: \begin{figure}
816: \includegraphics[width=1.0\columnwidth]{f2.eps}
817: \figcaption{
818: 8\arcsec $\times$8\arcsec\ \hst\ F450W image of the field around the ULX source \cxou. 
819: The X-ray source position is marked by the "X" of $\pm 0\farcs5$ radius 
820:  corresponding to the combined \cxo\ and \hst\ positional uncertainty. 
821: Ground-based \ha\  (black) and \oiii\ $\lambda 5007$ (white)
822: isophotes at the 20, 50 and 85\% levels above their backgrounds 
823: are superimposed (see \S~\ref{s:INT}). The slit position of the optical 
824: spectrum (\S~\ref{s:SPEC}) is also shown.
825:     \label{f:FOV} }
826: \end{figure}
827: 
828: \clearpage
829: 
830: \begin{figure}
831: \includegraphics[width=\textwidth]{f3.eps}
832: \figcaption{Integral spectrum of the cluster near \cxou\ (top). 
833: The lower thick solid curve 
834: is the best-fitting stellar spectrum model estimate,
835: divided by a factor of 3 for clarity, and the dotted curve is the combination of this model 
836: and the best-fitting photoionization model (see \S~\ref{s:RESULTS}),
837: divided by a factor of 2. The brightest emission lines including
838: Wolf-Rayet blends are superscribed.
839:   \label{fig:ospect_mod} }
840: \end{figure}
841: 
842: \end{document}
843: 
844: