1: \documentclass[12pt,preprint]{aastex}
2: \bibliographystyle{apj}
3:
4: \def\gtrsim{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$>$}}}}
5: \def\lesssim{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$<$}}}}
6: \newcommand{\msun}{$M_{\odot}$}
7: \newcommand{\ngc}{NGC~4254}
8: \newcommand{\VV}{VIRGOHI~21}
9: \newcommand{\hi}{H{\sc i}}
10: \newcommand{\etal}{et~al.}
11: \newcommand{\swift}{{\sl Swift}}
12: \newcommand{\ergcms}{ergs~cm$^{-2}$~s$^{-1}$}
13: \newcommand{\fxfo}{$f_{\rm X}/f_{O}$}
14: \newcommand{\ergs}{ergs~s$^{-1}$}
15: \newcommand{\ergl}{ergs~s$^{-1}$}
16: \newcommand{\xrt}{\it XRT\rm}
17: \newcommand{\beq}{\begin{equation}}
18: \newcommand{\eeq}{\end{equation}}
19: \newcommand{\kms}{km~s$^{-1}$}
20: \newcommand{\name}{SWIFT-XRT~J121723+1439.5}
21:
22: \begin{document}
23:
24: \title{Swift XRT Observations of the Possible Dark Galaxy \VV}
25: \author{Massimiliano~Bonamente\altaffilmark{1,2},
26: Douglas~A.~Swartz\altaffilmark{3},
27: Martin~C.~Weisskopf\altaffilmark{2}, and
28: Stephen~S.~Murray\altaffilmark{4}}
29: \altaffiltext{1}{Department of Physics, University of Alabama, Huntsville, AL, USA}
30: \altaffiltext{2}{NASA Marshall Space Flight Center, Huntsville, AL, USA}
31: \altaffiltext{3}{Universities Space Research Association,
32: NASA Marshall Space Flight Center, VP62, Huntsville, AL, USA}
33: \altaffiltext{4}{Harvard-Smithsonian Center for Astrophysics, 60 Garden Street,
34: MS-67, Cambridge, MA 02138}
35:
36: \begin{abstract}
37: \swift\ \xrt\ observations of the \hi\ line source \VV\ were performed
38: on April 22 and April 26, 2008 for a total exposure time of 9.2~ks.
39: This is the first pointed X-ray observation of \VV, a putative
40: dark galaxy in the Virgo cluster, and no photons were detected from this source.
41: The non-detection of extended X-ray emission
42: within the angular extent of the \hi\ source
43: corresponds to a 99\% confidence upper limit of $2.1 \times 10^{-14}$~\ergcms\
44: in the 0.3--2.0~keV band.
45: The equivalent upper limit to the amount of diffuse hot gas
46: associated with \VV\ is in the range $4\times 10^{7} - 2 \times 10^8$ \msun\
47: for a hot gas temperature between 0.1 and 1 keV.
48: The non-detection also corresponds to a 99\%-confidence upper limit on the flux from
49: a point-like source
50: of $8\times 10^{-15}$~\ergcms\ in the 0.3--2.0~keV
51: band.
52: We discuss the constraints on the nature of \VV\ imposed by these
53: observations and the theoretical implications of these results.
54: \end{abstract}
55:
56: \keywords{dark matter --- galaxies: individual (\VV) --- X-rays: galaxies}
57:
58: \section{Introduction}
59: \hi\ surveys of the Virgo Cluster \citep{davies2004,davies2006,
60: minchin2005,minchin2007,haynes2007}
61: have revealed a remarkable %isolated
62: \hi\ cloud, \VV, with apparent evidence of rotation yet unaccompanied by any optical
63: emission.
64: The \hi\ mass ($M_{HI} \sim 3\times 10^7$~\msun),
65: inferred circular velocity ($V\sim 100$~km~s$^{-1}$)
66: and physical size ($R\sim 8$~kpc) are all typical of luminous galaxies of
67: dynamical masses $10^{10}-10^{11}$~\msun\ yet deep optical follow-up observations
68: provide an upper limit of only 31.1 $I$ mag~arcsec$^{-2}$, corresponding to
69: a stellar content of 2$\times$10$^5$~$L_{\odot}$ \citep{minchin2007}.
70: Thus, \VV\ is currently considered the
71: prototypical dark galaxy candidate: a dynamically-massive
72: dark-matter-dominated galaxy
73: with a rotating gaseous disk but devoid of stars.
74:
75: If \VV\ is a massive dark galaxy, then it
76: has important implications for the theory of structure formation.
77: The highly-successful cosmological hierarchical clustering model originated by
78: \citet{white1978} is based on the idea of accretion and merging of cold dark matter
79: (CDM) from small sizes ($\sim$10$^8$~\msun) up to the large galaxy clusters observed
80: at the present epoch. The CDM model predicts a large number of
81: $\gtrsim$10$^8$~\msun\ dark matter (DM) halos should remain unassimilated following
82: hierarchical growth and thus still be present today \citep[e.g.,][]{moore1999,
83: klypin1999,davies2006}.
84: %
85: To date, however,
86: the number of observed low-mass galaxies falls far short of these
87: predictions. This is true locally, where only a handful of satellites
88: surround the massive Local Group spirals
89: \citep[e.g.,][]{mateo1998},
90: and is manifest in
91: the faint end of the luminosity function that shows a slope more shallow than
92: $\Lambda$CDM predictions
93: \citep{blanton2001,benson2003}.
94: This shortcoming of the $\Lambda$CDM model can be overcome either
95: by preventing small DM halos from forming in the first place
96: or by suppressing star formation so that they remain optically faint.
97: %
98: The latter possibility implies a population of
99: low-mass DM halos in
100: which the baryon content remains optically-dark but still
101: visible chiefly through \hi\ gas emission.
102: The rarity of isolated extragalactic
103: \hi\ clouds lacking optical counterparts
104: \citep{doyle2005}
105: suggests that \VV, if truly massive, is a
106: rare case in favor of this scenario.
107:
108: An alternative to the dark galaxy interpretation for \VV\
109: \citep{bekki2005, haynes2007, duc2008} is that
110: it is simply a part of an elongated \hi-dominated tidal
111: tail emanating from the nearby luminous galaxy \ngc\ (M99). The tidal tail was
112: caused either by a fly-by encounter with another (massive) galaxy \citep{vollmer2005}
113: or by ram pressure stripping within the Virgo environment.
114: Recent observations show the \hi\ gas in the tail extends from
115: \ngc, located $\sim$120 kpc to the south of \VV,
116: to about 130~kpc beyond \VV\ to the north \citep{haynes2007}.
117: The peculiar
118: velocity structure in the vicinity of \VV\ is, in this case, ascribed to
119: projection effects on the streaming motions of the debris tail
120: instead of to rotation within a gravitationally bound disk.
121: In this scenario, the estimated \hi\ mass is close to the total mass of the object;
122: there is no massive halo accompanying \VV.
123:
124: In
125: \S~\ref{analysis} we describe the \swift\
126: observations leading to upper limits of the X-ray emission
127: from \VV, and in \S~\ref{discussion} we provide our interpretation
128: of the \swift\ non-detection.
129: In this paper we assume
130: a distance to the source of $D=16.5$ Mpc corresponding to the mean
131: of the Virgo Cluster \citep{mei2007} and a scale of 4.9~kpc per arcminute.
132:
133: \section{SWIFT observations of \VV\ and data analysis}
134: \label{analysis}
135:
136: The \VV\ field was observed by \swift\ on 2008 April 22 and 26 with the
137: \xrt\ operated in photon counting mode.
138: The \xrt\ has a very low detector background, which makes it especially suitable for observations
139: of faint diffuse X-ray sources.
140: These data are the first pointed X-ray observations of \VV.
141:
142: The data were screened for bad pixels and other detector artifacts, and only photon
143: grades 0-12 were included in the analysis. The screening resulted in 9.2 ks
144: of clean data. In this analysis we considered photon energies in the 0.3-7 keV range,
145: where the calibration of the \xrt\ is better understood.
146: The data reduction was performed using \tt xselect \rm and the \tt FTOOLS\rm,
147: and the spectral analysis with \tt XSPEC\rm.
148:
149: \VV\ lies about 1 Mpc to the NW of the center of the Virgo Cluster
150: at R.A.$=$12$^{\rm h}$17$^{\rm m}$53.6$^{\rm s}$
151: decl.$=$+14\arcdeg45\arcmin25\arcsec\ (J2000.0).
152: The angular size of \VV, 3.5\arcmin$\times$1.375\arcmin, lies well within
153: the 24\arcmin$\times$24\arcmin\ field of view of the \xrt\
154: \citep{cusumano2006,romano2005}.
155: X-ray images of the \VV\ field are shown in
156: Figure \ref{image}, in which the gray contour are from the \hi\ observations
157: of \citet{minchin2007}. The upper part of the \hi\ feature was interpreted
158: by \citet{minchin2007} as a rotating disc, and the lower part as an \hi\ bridge connecting
159: \VV\ with NGC~4254, located to the south and outside the
160: \swift\ field of view.
161:
162: As with normal galaxies, \VV\ was expected to be a source of
163: both diffuse and point-source X-ray emission.
164: Even in the absence of stars,
165: gas falling into a galaxy's gravitational potential from the intergalactic medium
166: should be heated to roughly the (X-ray-emitting) virial temperature.
167: In addition,
168: accretion onto a central supermassive black hole from the gas reservoir
169: in \VV, even at a fraction of the Bondi rate, should appear as a localized source of X-radiation though not necessarily optically bright.
170: No X-ray emission associated with \VV\ was detected.
171: %
172: An unrelated weak (41 counts) point-like source was detected near
173: R.A.=12$^{h}$17$^{m}$23$^{s}$, decl.=14$^{\circ}$39\arcmin29\arcsec\
174: (Figure \ref{image}).
175: As no object is known within 1\arcmin\ of this location,
176: we regard this as the discovery of the X-ray source \name.
177: %
178: In the following, we place upper limits to the presence of diffuse
179: and point source emission from \VV.
180:
181: \subsection{Diffuse X-ray emission}
182: \label{diffuse}
183: In order to place upper limits to the diffuse soft X-ray emission
184: from \VV, we accumulated X-ray photons from a 3.5'$\times$1.375'
185: region located at the position of the putative \hi\ disc (black box
186: in Figure \ref{image}). We detected a total of 8 counts in the 0.3-2 keV
187: band, and no counts in the 2-7 keV band.
188: We estimate the number of background counts, first by extracting the data
189: from the entire field of view, excluding a region of enhanced emission
190: to the south (Figure \ref{image}) and regions close to the detector
191: boundaries. The region included encompassed 337.3 square arcminutes.
192: We found $623$ counts in the 0.3-2 keV band and $319$ counts in the 2-7 keV
193: band. Ignoring the possibility that some of these events might be vignetted by
194: the telescope one therefore expects in the region of interest at least
195: $8.9\pm0.4$ counts in the soft band and $4.6\pm0.3$ in the hard band. We
196: note that it is mildly interesting and somewhat improbable that no counts
197: were measured in the hard band from the region containing \VV.
198:
199: We use the soft X-ray band data in order to set upper limits as to source
200: detection, since this is the band in which
201: the X-ray emission from diffuse gas (and point sources) is expected.
202: Since the measured background predicts at least a mean of 8.9$\pm$0.4 counts
203: and we detected 8 counts it is therefore evident that the \VV\ region is not
204: an X-ray emitter.
205: To determine upper limits we increased the expected number of counts in the
206: soft band by 10\% to 9.8 counts to account for the possibility that some of
207: the measured background events arise from faint, unresolved X-ray sources.
208: We determine 99\%-confidence upper limits assuming 18 counts as detection.
209:
210: \subsubsection{Uniform distribution}
211: \label{sec:uniform}
212: We assume that the \hi\ disk is seen edge-on, and thus the emitting volume
213: is $V=\pi R^2 L$, where $R=1.75$' and $L=1.375$'.
214: We use an optically-thin plasma
215: of uniform density and of Solar abundance (\tt apec \rm in \tt XSPEC\rm), a Galactic
216: \hi\ column density of $N_H=2.7\times 10^{20}$ cm$^{-2}$
217: \citep{dickey1990,kalberla2005} with \citet{morrison1983} cross-sections (\tt wabs \rm
218: in \tt XSPEC\rm), and calculate the emission measure
219: required to achieve 8.1 source counts as a function of plasma temperature using \tt XSPEC\rm.
220: The emission measure calculated by \tt XSPEC \rm is given by the model normalization, $K$, as
221: \begin{equation}
222: K= \frac{10^{-14}}{4 \pi D_A^2} \int n_e n_H dV,
223: \label{norm}
224: \end{equation}
225: in which $n_e$ and $n_H$ are, respectively, the electron and hydrogen
226: number density and $D_A$ the angular size distance (for $z\ll 1$).
227: From this, we calculate the gas density, and thus the gas mass.
228: The results are shown in Figure \ref{limits}, showing that these 9.2 ks \swift\
229: \xrt\ observations set upper limits of the order of $10^8 M_{\odot}$.
230: These upper limits correspond to a flux of $2.1 \times 10^{-14}$ \ergcms\ in the 0.3-2 keV band.
231:
232: \subsubsection{$\beta$ model distribution}
233: \label{sec:beta}
234: A putative hot halo may be maintained in hydrostatic equilibrium by the
235: gravitational potential of the galaxy. In this case,
236: the gas density will decrease with radius. We therefore
237: consider a more realistic model of the gas density
238: given by the $\beta$ model,
239: \begin{equation}
240: n_{e}=n_{e0} \left( 1 + \frac{r^2}{r_c^2} \right)^{-\frac{3}{2} \beta}
241: \label{beta}
242: \end{equation}
243: as often observed in galaxy clusters \citep{cavaliere1978}. The parameters
244: $r_c$ and $\beta$ describe, respectively, the radius within which
245: the density is approximately constant, and the decrease
246: of the density at large radii. For this interpretation, we use a spherically
247: symmetrical model in which the galaxy has a radius of $R=1.75$'.
248: In this region, the background in the 0.3-2 keV band has a mean of
249: 16.3 counts, and we thus place the 99\% confidence upper
250: limit at 27 total counts (or 10.7 source counts), following the method described in
251: \S~\ref{sec:uniform}. Using the $\beta$ model (Equation \ref{beta})
252: into Equation \ref{norm}, we obtain upper limits to the central
253: gas density:
254: \begin{equation}
255: n_{e0}=\left(\frac{K 10^{14} 4 \pi D^2}{4 \pi r_c^3 I_1(R/r_c)}\right)
256: ^{\frac{1}{2}}
257: \label{neo}
258: \end{equation}
259: in which the integration of the density profile has led to the integral
260: %\begin{equation}
261: $I_1(R/r_c) = \int_{0}^{R/r_c} x^2(1+x^2)^{-3 \beta} dx$.
262: %\end{equation}
263: We use two reference values for the index $\beta$:
264: a value of $\beta$=2/3, as usually measured in galaxy
265: clusters \citep[e.g.,][]{laroque2006},
266: and a shallower slope of $\beta$=1/3,
267: which may be typical of lower-mass halos (e.g., galaxy groups
268: and galaxies).
269: We find analytical solutions
270: for the integral, $I_1(n)=n/(2+2n^2)+ \arctan(n)/2$ for $\beta$=2/3,
271: and $I_1(n)=n-\arctan(n)$ for $\beta$=1/3.
272:
273: We then estimate the upper limits to the gas mass by integration of the
274: density profile,
275: \begin{eqnarray}
276: M_{gas} &=& \mu_e m_H \int_0^{R} n_{e0} \left( 1 + \frac{r^2}{r_c^2} \right)^{-\frac{3}{2} \beta}
277: 4 \pi r^2 dr \nonumber \\ &=& 4 \pi \mu_e m_H n_{e0} r_c^3 I_2(R/r_c),
278: \label{mgas}
279: \end{eqnarray}
280: in which the $\beta$ model profile has led to the integral
281: %\begin{equation}
282: $I_2(R/r_c) = \int_0^{R/r_c} x^2(1+x^2)^{-3\beta/2} dx$.
283: %\end{equation}
284: This integral is $I_2(n) = n - \arctan(n)$ for $\beta=2/3$, and $I_2(n)=\ln(x+\sqrt{1+x^2})$
285: for $\beta=1/3$.
286:
287: We estimate upper limits to the gas mass
288: as function
289: of plasma temperature using Equations \ref{neo} and \ref{mgas},
290: using a fiducial value of the core radius of 0.5' (Figure \ref{limits}).
291: We determine that the upper limits to the gas mass depend weakly on the choice of
292: the core radius, given that varying the fiducial value by $\pm$100\% results in
293: variation of less than 10\%.
294: This model results in upper limits to the gas mass that are in the range
295: of $M_{gas} \simeq 4\times 10^7 - 2 \times 10^8$ $M_{\odot}$.
296: Upper limits to the gas mass are therefore similar to those obtained from
297: the simple uniform density model of Section \ref{sec:uniform}.
298:
299: \subsection{Point source X-ray emission}
300: \label{ps}
301: We set upper limits to the luminosity of an unobserved point source anywhere
302: within the 3.5' x 1.375' region of interest. The upper limit takes into account
303: that there are 68 statistically-independent SWIFT spatial resolution elements
304: within this region, given that the half-power diameter of the \xrt\ point spread function
305: at 1.5 keV is 18". The measured background, per 18"-diameter spatial
306: resolution element is 0.131, or 0.144 using a +10\% allowance for
307: vignetting. Thus, considering the Poisson distribution with a
308: mean of 0.144 and a total of 4 counts, there is a probability of only $1.6\times 10^{-5}$ of
309: measuring 4 or more counts in just one of these resolution elements. Of
310: course we would have been happy to measure 4 or more counts in any one of the
311: 68 elements so the joint probability is correspondingly larger. Thus, we may
312: state, with approximately 99.9\%-confidence that we have not detected a
313: a source of 4 counts or brighter from this region.
314:
315: We consider a power-law emission model with photon index $\Gamma=2.0$
316: and determine the upper limits to the source luminosity as function
317: of the absorbing \hi\ column density using \tt XSPEC\rm.
318: The upper limits are shown in Figure \ref{limits},
319: and are in the $L=3\times 10^{38}-10^{40}$ \ergl\ range. This upper limit
320: corresponds to a flux of 8$\times 10^{-15}$ \ergcms\ in the 0.3-2 keV
321: band.
322:
323: \subsection{UV and IR emission}
324:
325: For completeness, we also analyzed extant non-X-ray images of \VV.
326:
327: \swift\ UVOT images using the UVM2 filter
328: (effective wavelength 2231 \AA) were obtained
329: concurrently with the XRT data. No sources were detected within
330: the angular extent of \VV. The {\tt FTOOLS} utility {\tt uvotsource}
331: was used to estimate a limiting magnitude for the UV emission from
332: the 3.5\arcmin$\times$1.375\arcmin\ region enclosing \VV.
333: This limit does not improve our estimates of the upper limit to
334: the mass of diffuse hot gas nor to the flux from a point source
335: as estimated above from the X-ray data.
336:
337: We also analyzed archival {\sl Spitzer} data (Program ID 30725)
338: which included all four IRAC and three MIPS spectral bands.
339: Again, no extended emission was detected within the \hi\ extent of
340: \VV. We estimate an upper limit to the mass of dust associated with
341: \VV\ to be 10$^4$--10$^6$~\msun\ depending on the dust temperature
342: assumed.
343:
344: \section{Discussion and conclusions}
345: \label{discussion}
346:
347: A search for X-ray emission from the \hi\ line source \VV\ using the
348: \swift\ \xrt\ obtained only upper limits.
349: The non-detection indicates that any diffuse X-ray emitting gas at $kT=0.1-1$ keV
350: must be less massive than $4\times 10^{7} - 2 \times 10^8$ \msun,
351: depending on the gas temperature and its spatial distribution. Given that the \swift\
352: \xrt\ detector has no effective area approximately below 0.1 keV, it is possible that
353: significant amounts of gas at sub-virial temperature have gone undetected by this
354: X-ray observation.
355: Similarly, any X-ray point source associated with \VV\ will have a
356: luminosity lower than $5 \times 10^{38}-10^{40}$ \ergs, depending on the
357: absorbing column density of gas along the line of sight.
358:
359: How do these limits compare with expectations? There are well-established
360: correlations between optical and X-ray light from galaxies: Typical
361: \fxfo\ ratios for normal galaxies range from $10^{-4}$ to $10^{-2}$
362: and for galaxies hosting AGNs the range extends up to unity
363: \citep[e.g.,][]{kim2006}.
364: %
365: For the dark galaxy \VV, these ratios predict an X-ray
366: flux of $10^{-17}$ to $10^{-15}$ for a normal galaxy up to $10^{-13}$
367: if an AGN is present.
368: %
369: Of course, these \fxfo\ correlations reflect the causal link between
370: stellar processes and X-ray light. They are not expected to hold in the
371: absence of stars.
372:
373: Instead, our
374: expectation that \VV\ should be a source of diffuse X-ray emission
375: stems from the fundamental picture \citep{reesostriker1977} that gas
376: falling into a DM halo is first heated to the virial temperature.
377: %%%before slowly cooling radiatively and settling onto the disk where it
378: %%%is expected to eventually form stars.
379: The accumulation of hot gas over a Hubble time could be substantial for
380: a massive DM halo like \VV\ \citep{whitefrenk1991}.
381: %
382: We would expect a gas virial temperature of
383: approximately 0.4~keV based on the observed kinematics of \VV\ %%%following
384: \citep[from][]{minchin2007} and thus our best guess to the
385: amount of hot gas associated with \VV\ is $\leq 5\times 10^7$~\msun.
386: %%%This is comparable to the amount of \hi\ gas detected from \VV\
387: %%%\citep{minchin2007} and
388: %%%to about 10\% of the total \hi\ mass in the vicinity \citep{haynes2007}.
389: %%%This, in itself, does not constitute a very restrictive limit on the nature of \VV.
390: %
391: Scaling from simulations of gas infall onto stable disk galaxies
392: \citep{benson2000, toft2002},
393: the X-ray luminosity of this gas should be up to
394: 5$\times$10$^{38}$~\ergl\ for rotation speeds similar to that
395: of \VV. %%%predict X-ray luminosities up to
396: %%%5$\times$10$^{38}$~\ergl\ for rotation speeds similar to that
397: %%%of \VV\ (scaling as $L_{\rm X} \propto V_c^5$).
398: %%%This is comparable to our upper limit, and
399: %%%does not place strong limits on \VV.
400:
401: The above estimates are based on the assumption of an isolated `quiescent' massive
402: galaxy accreting slowly (i.e., with a long cooling time)
403: from the surrounding intergalactic medium.
404: In the particular case of \VV, in contrast,
405: a considerable gas reservoir exists within the tidal tail emanating
406: from \ngc\ that should lead to a
407: relatively high gas accretion rate and hence high X-ray luminosity.
408: Furthermore, turbulence and shocks following the interaction between
409: \VV\ and \ngc\ would very likely heat some portion of the disk \hi\ gas component
410: of \VV\ to X-ray emitting temperatures.
411: We therefore consider the above estimates to be minimal expectations.
412: Of course, any prediction of the amount of hot gas or its luminosity
413: will scale with the dynamical mass of
414: the accreting object.
415:
416: The observed upper limits reported here, therefore,
417: show that the extreme object \VV\ is {\sl not} an extreme X-ray object
418: (in the sense of a high \fxfo) nor has it accreted very much gas
419: from its surroundings during its lifetime.
420:
421: \input{ms.bbl}
422:
423: \clearpage
424:
425: \begin{figure}
426: \begin{center}
427: \includegraphics[angle=-90,width=6in]{f1.eps}
428: \end{center}
429: \caption{\swift\ \xrt\ images of \VV \label{image}, smoothed with a Gaussian
430: kernel of $\sigma$=7". Grey contours are reproduced
431: from the \citet{minchin2007} \hi\ observations with the Westerbork telescope.
432: The black box represent the location of the putative \hi\ rotating disc.
433: The circle centered at RA=12$^{h}$17$^{m}$23$^{s}$, $\delta$=14$^{\circ}$39'29" represents
434: the region from which we extracted the counts of the serendipitous source \name.
435: }
436: \end{figure}
437:
438: \begin{figure}
439: \includegraphics[angle=-90,width=5.0in]{f2.eps}
440: \caption{(a): Upper limits to the mass of diffuse gas in \VV, estimated assuming a
441: detection threshold of 17 counts in these \xrt\ observations. The \swift\ \xrt\ bandpass
442: is not sensitive to emission from diffuse gas at temperature $kT \leq 0.1$ keV.
443: Solid line : uniform
444: density model, using a detection threshold of 8.2 counts in a 3.5'$\times$1.375' box.
445: Dashed lines : $\beta$ model with $r_c$=0.5', using a detection threshold
446: of 10.7 counts in a circle of radius 1.75' (bottom line: $\beta$=0.67; top line: $\beta$=0.33).
447: Given that the gas mass is proportional
448: to the square root of the number of detected counts, the estimates are not very sensitive
449: to changes in the detection threshold.
450: (b) Upper limits to the luminosity of point sources in the \VV\ field, estimated
451: assuming a detection threshold of 4 source counts. The luminosity is proportional to the
452: number of detected counts.
453: \label{limits}}
454: \end{figure}
455:
456:
457: \end{document}
458:
459: