1: %% Using AASTeX v5.0 LaTeX 2e macros.
2:
3: %\documentclass{aastex}
4: \documentclass[12pt,preprint]{aastex}
5: %\documentclass[preprint2]{aastex}
6: %\documentclass[apj]{emulateapj}
7:
8: \def \degpoint {${.}\!\!^{\circ}\!$}
9: \def \secpoint {${.}\!\!^{\prime \prime}\!$}
10: \def \minpoint {${.}\!\!^{\prime}$}
11:
12: \def\ltsim{ \,{}^<_\sim\, }
13: \def\gtsim{ \,{}^>_\sim\, }
14:
15: \shorttitle{Stuctural Parameters of GCs in NGC 5128}
16: \shortauthors{G\'{o}mez and Woodley}
17:
18: \begin{document}
19:
20: \title{Sizes of Confirmed Globular Clusters in NGC 5128: A Wide-Field High-Resolution Study\footnote{This paper includes data gathered with the 6.5 meter Magellan Telescopes located at Las Campanas Observatory, Chile}}
21:
22: \author{Mat\'{\i}as G\'{o}mez}
23: \affil{Grupo de Astronom\'{\i}a, Depto. de F\'{\i}sica, Universidad de
24: Concepci\'{o}n, Casilla 160-C, Concepci\'{o}n, Chile}
25: \email{matias@astro-udec.cl}
26:
27: \author{Kristin A.~Woodley}
28: \affil{Department of Physics \& Astronomy, McMaster University,
29: Hamilton ON L8S 4M1, Canada}
30: \email{woodleka@physics.mcmaster.ca}
31:
32:
33: \begin{abstract}
34: Using Magellan/IMACS images covering a 1.2 x 1.2 sq. degree
35: FOV with seeing of 0.4"-0.6", we have applied convolution techniques
36: to analyse the light distribution of 364 confirmed globular cluster
37: in the field of NGC 5128 and to obtain their structural parameters. Combining
38: these parameters with existing Washington photometry from
39: Harris et al. (2004),
40: we are able to examine the size difference between metal-poor (blue)
41: and metal-rich (red) globular clusters.
42: For the first time, this can be addressed on a sample of confirmed
43: clusters that extends to galactocentric distances about 8 times
44: the effective radius, R$_{eff}$, of the galaxy.
45: Within 1 R$_{eff}$, red clusters are about $30\%$
46: smaller on average than blue clusters, in agreement
47: with the vast majority of extragalactic globular cluster systems
48: studied. As the galactocentric distance increases,
49: however, this difference becomes negligible. Thus, our results
50: indicate that the difference in the clusters' effective radii, r$_e$, could be
51: explained purely by projection
52: effects, with red clusters being more centrally concentrated than blue
53: ones and an intrinsic r$_e$--R$_{gc}$ dependence, like the one
54: observed for the Galaxy.
55: \end{abstract}
56:
57: \keywords{galaxies: elliptical and lenticular, cD --- galaxies:
58: individual (NGC 5128) ---
59: galaxies: star clusters --- globular clusters: general}
60:
61: \section{Introduction}
62: \label{sec:intro}
63: Since sizes and structural parameters of globular clusters (GCs)
64: in different GC systems (GCSs) have first been obtained,
65: it has become clear that some of these properties correlate with
66: global properties of
67: their host galaxies \citep[see for example][]{jordan05,brodie06}.
68: The existence of the so called fundamental plane relation for
69: an increasing number of studied GCSs seems to
70: confirm that GCs populate
71: a narrow region in this parameter space
72: \citep{djorgovski95,mclaughlin00,mclaughlin05,barmby07}.
73: However, there are puzzling trends that are still awaiting confirmation
74: and need to be addressed using larger samples of GCs.
75:
76: It is necessary to study structural parameters of
77: GCs and GC-like objects in different
78: environments before definitive statements can be made
79: regarding their formation.
80: Among the structural parameters that can be studied, the effective
81: (or half-light) radius is of particular importance.
82: Models have shown that this
83: quantity remains fairly constant throughout the entire GC
84: lifetime \citep{spitzer72,aarseth98}, making it a
85: good indicator of proto-GC sizes that are still observable today.
86: A decade ago, HST observations unveiled a systematic size difference
87: between red and blue GCs \citep{kundu98}. Since then, multiple
88: studies have found that the
89: blue GCs are between $17\%-30\%$ larger than their
90: metal-rich counterparts in both spirals and early--type galaxies
91: \citep{kundu99,puzia99,larsen01,larsen_fb01,kundu01,barmby02,jordan05}.
92: However, most of these studies have made use of HST observations
93: and examine only the innermost regions of the galaxy or small
94: fields in regions at galactocentric distances
95: greater than the galaxy's effective radius.
96:
97: According to \cite{larsen03}, the systematic size difference between
98: red and blue GCs is caused merely by a projection effect.
99: Since red (metal-rich) GCs are found to be more
100: centrally concentrated than blue
101: (metal-poor ones) in early type galaxies
102: \citep[][among others]{cote01,dirsch03,woodley05}, the red
103: GCs will appear to lie, on average,
104: at a smaller galactocentric distance. The red clusters will on
105: average be smaller than the blue clusters assuming that both types
106: shares the same relation between the GC size and galactocentric
107: distance. The relation r $\sim \sqrt{\rm{R}_{gc}}$ was first found in the
108: Milky Way by \cite{vandenbergh91}. In this scenario, the difference
109: between the cluster sizes should be most apparent at small
110: galactocentric distance and should decrease strongly beyond 1 galaxy
111: effective radius \citep{larsen03}.
112:
113: Alternatively, \cite{jordan04} suggests that this effect could be
114: explained by an intrinsic difference between metal-rich and metal-poor
115: GCs. Assuming
116: half-mass radii that are independent of metallicity, effects of
117: mass segregation combined with
118: a metallicity-dependent stellar lifetime should lead to different
119: sizes between the blue and red clusters. The brightest stars would be
120: more massive and more centrally concentrated for the metal-rich
121: GCs. This scenario should have little to no dependence
122: on a cluster's distance from the center of its parent galaxy.
123:
124: In a recent study, \cite{spitler06} analysed the GCS
125: of NGC 4594 (Sombrero, at a distance of $ \sim 9$ Mpc) using a
126: six-image mosaic from HST/ACS.
127: They confirm that within the inner 2 arcmin (2.2 R$_{eff}$),
128: the metal-rich GCs are, on average, $17\%$
129: smaller than the metal-poor clusters. However, the
130: size difference becomes negligible at $\sim 3$ arcmin, corresponding
131: to $\sim 3.4$ R$_{eff}$, where R$_{eff} =0.89$
132: arcmin \citep{baggett98}.
133:
134: To further understand the sizes of red and blue clusters, we need a
135: homogeneous survey of a GCS with the ability to eliminate
136: contaminating sources, high resolution to measure structural
137: parameters, and over a large range in galactocentric distance.
138:
139: NGC 5128 is the nearest giant elliptical galaxy, at a distance of 3.8 Mpc
140: \citep{mclaughlin07}. Its GCs are thus
141: easily resolvable with sub-arcsecond seeing \citep{harris06}.
142: In this paper we present effective radius
143: results for 337 GCs from the \cite{woodley07} catalog that are
144: confirmed GCs by either radial velocity measurement from various
145: studies \citep[see the references in ][]{woodley07} or are resolved by
146: HST/ACS images \citep{harris06}. We also present the effective radii
147: of 27 GCs newly confirmed through radial velocity
148: measurements using the Baade 6.5-m telescope with the instrument LDSS-2
149: (data in preparation for publication). This list represents a clean
150: sample of confirmed clusters. All of these also have ellipticities less than
151: 0.4 and effective radii less than 8 pc, both of which are consistent with
152: {\it normal} GC properties in NGC 5128. We find that only an additional $2.4\%$
153: of GCs from the \cite{woodley07} catalog have effective radii greater
154: than the 8 pc boundary we have imposed here (to be discussed in detail
155: in G\'omez \& Woodley, 2008, in preparation).
156: Those few GCs are not considered here as our purpose is to
157: establish the effective radius trends within the bulk of the GC population.
158:
159: \section{Observations}
160: \label{sec:obs}
161:
162: On the night of April 9, 2006, 25 fields were imaged with the
163: Magellan 6.5 m telescope using the Inamori Magellan Areal Camera
164: and Spectrograph (IMACS).
165: In the highest imaging resolution, IMACS offers a FOV of 15.4
166: arcmin on a side, composed of a mosaic of 8 2Kx4KCDs
167: with a scale of 0.111 arcsec/pixel.
168: Our observational material will be fully discussed in
169: Harris et al. 2008 (in preparation). The
170: total field of view of our images is roughly 1.2 x 1.2
171: square degrees and the average seeing is about 0.5" across the
172: entire field with individual
173: frames ranging from 0.35" to 0.7".
174: Images were acquired through B (on 16 of the 25 fields) and R
175: (on all 25 fields) filters with both 10 second and
176: 300 second exposures to avoid saturation of the brightest clusters.
177:
178: We have identified all GCs in the catalog of
179: \cite{woodley07} on our IMACS frames\footnote{Note that the positions of two
180: GCs have been corrected: GC0001 with a corrected right ascension
181: of 13$^h$ 25$^m$ 1.16$^s$ (J2000) and GC0002 with corrected
182: declination of -43$^{o}$ 02$^{\arcmin}$ 42.9$^{\arcsec}$ (J2000).}.
183: We have run the code ISHAPE \citep{larsen99,larsen_s01} individually on each
184: GC in our R filtered IMACS frames, using a
185: stellar point spread function (PSF) modelled from the chip in which
186: the cluster is located. For this, typically 20-30 stars were
187: chosen in each frame and measured with standard tools in IRAF\footnote{IRAF is
188: distributed by the National Optical Astronomy Observatory, which is
189: operated by the Association of Universities for Research in Astronomy
190: Inc., under cooperative agreement with the National Science Foundation.}.
191: ISHAPE convolves the PSF with analytical profiles and
192: compares the result with the input image until a best match is achieved.
193:
194: As the analytical model, we chose King (1962) profiles, given their simplicity
195: and because they are known to provide a good fit to a large family of
196: GCs in different environments. Moffat functions were also
197: tried, but they do not improve the fits except for a handful of large and
198: very elliptical sources that we are not considering for the present study.
199: They will be discussed as special cases in a forthcoming paper.
200: Possible systematic effects in the sizes, arising from the choice of a
201: particular model are discussed in \cite{larsen99}. However, the effective
202: radius seems to be independent of the model for sources that have a similar
203: extension to the stellar PSF, as in our case.
204: For a recent comparison between different models, the reader is referred
205: to \cite{barmby07} and \cite{mclaughlin07}.
206:
207: \cite{king62} profiles are defined by a core-radius r$_c$ and a
208: concentration index, which we define here as $c =$ r$_t/$r$_c$, where
209: r$_t$ is the tidal radius of
210: the cluster. Usually, the concentration parameter is the most uncertain
211: one to constrain \citep{larsen_s01}, but given the high
212: spatial resolution of our IMACS images, we were able to
213: fit this along with the ellipticity, position angle (PA) and r$_c$.
214:
215: The sizes quoted by ISHAPE were transformed into effective radii using
216: the approximation r$_e/$r$_c \approx 0.547 c^{0.486}$, good to
217: $\pm 2\%$ for $c>4$ \citep{larsen_s01}. The median value for the
218: concentration parameter for GCs in NGC 5128 was $c=39.4 \pm 10.2.$
219: Uncertainties in the effective radius were estimated by the
220: standard deviation of the determined value using King profiles with
221: fixed concentration indices of 15, 30, and 100.
222: These concentration parameters were chosen
223: based on typical values observed in our Galaxy as well as the
224: concentration parameters fit freely with ISHAPE for the NGC 5128 data.
225: The r$_e$ determined for any given cluster with varying concentration
226: parameters does not vary more than $\sim 10\%$ for the average
227: GC. The concentration parameter, c, is the most uncertain
228: of the fitted parameters. The extension of the GC is a secondary uncertainty.
229: A GC at the distance of NGC 5128, with an effective radius of 6 pc,
230: would span a diameter of 0.6", marginally larger than the typical stellar FWHM. For
231: smaller or more compact objects, the instrisic size can be as small as
232: 0.1", i.e. completely blurred even with sub-arcsecond imaging.
233:
234: \section{Results}
235: \label{sec:results}
236:
237: We have 69 GCs in common with the HST/ACS structural parameter
238: study of \cite{harris06}. Their r$_e$ values, derived
239: through an isophotal analysis of the resolved clusters, and discussed
240: fully in \cite{mclaughlin07},
241: serve as an external comparison and quality test for our
242: measurements. Figure~\ref{fig:acs_imacs} shows clearly that
243: there is good agreement in the $r_e$ values determined by these independent techniques.
244: We have also examined our measured r$_e$ as a function
245: of ellipticity and luminosity and also found no notable correlation.
246:
247: In an upcoming paper, (G\'omez \& Woodley, 2008, in preparation), we
248: will discuss the structural parameters in detail as well as the new GCs
249: discovered with the Baade 6.5m telescope (mentioned above) that have
250: been used in this study. Here, we focus on the dependence of the
251: GC sizes as a function of galactocentric radius, R$_{gc}$ for the GC
252: subpopulations. A metallicity break has been
253: chosen to represent the red or metal-rich ([Fe/H]$ > -1$) and blue or
254: metal-poor ([Fe/H]$ < -1$) subpopulations of clusters, following the
255: studies of \cite{larsen03,harris04,woodley05,gomez06,woodley07}. The [Fe/H]
256: values were obtained from a $C-T_1$ transformation \citep{harrisharris02}
257: assuming E(B-V) = 0.11 \citep{schlegel98}.
258:
259: Figures~\ref{fig:re_feh} and \ref{fig:N_re} show that, within one effective radius of the
260: galaxy, the red clusters are significantly smaller than the blue
261: ones by $\sim 30\%$.
262: As galactocentric distance increases, however, this
263: difference tends to disappear and beyond a distance of 12 kpc (corresponding to
264: $\sim 2.3$ R$_{eff}$), no difference remains. We have performed a Spearman
265: test to study the trend between metallicity and r$_{e}$. The Spearman
266: non-parametric rank-correlation coefficient ranges from -1 to 1, with -1
267: for a complete anti-correlation, 0 for no correlation and 1 for a complete
268: correlation. We found the coefficients to be -0.4 for R$_{gc} < 1$ R$_{eff}$,
269: -0.1 for 1 R$_{eff} <$ R$_{gc} <$ 2 R$_{eff}$, 0.1 for 2$ R_{eff} <$ R$_{gc} < 3 $R$_{eff}$, and
270: 0.1 for R$_{gc} > 3$ R$_{eff}$. In addition, the Spearman test gives the probability that
271: the two datasets are uncorrelated as well. The values read 0.01, 0.19, 0.43
272: and 0.34, respectively, for the four radial bins presented above. Thus,
273: there is a small, but virtually confirmed anti-correlation within 1 R$_{eff}$.
274:
275: This size trend is even more clear in Figure~\ref{fig:N_re}.
276: Within 1 R$_{eff}$ of the galaxy, the metal-rich and metal-poor clusters have median effective radii of
277: 1.94$\pm0.19$ and 2.98$\pm0.20$ pc, respectively.
278: However, outside of 1 R$_{eff}$, the
279: median r$_e$ is identical within uncertainties.
280:
281: Figure~\ref{fig:re_Rgc} shows r$_{e}$ as a function of projected
282: R$_{gc}$ for the GCs in NGC 5128 and in the Milky Way.
283: The metal-poor GCs in NGC 5128 do not follow the
284: r$_e$--R$_{gc}$ relationship that is evident in the metal-rich
285: GCs. The Milky Way data, on the other hand, has been shown to host
286: the
287: r$_e$--R$_{gc}$ relationship for both metallicity populations,
288: first noted by \cite{vandenbergh91}, using a 3-dimensional R$_{gc}$.
289: However, in projection, the metal-poor GCs in the Milky Way
290: do not appear vastly different from those in NGC 5128.
291:
292: \section{Discussion}
293: \label{sec:discussion}
294: The existence of a systematic difference in the effective radii of
295: blue and red clusters has been extensively studied in other galaxies,
296: with blue clusters typically
297: found to be $17-30\%$ larger than red ones (see
298: Section~\ref{sec:intro}). However, in NGC 5128, \cite{harris02} did
299: not find any correlation between color and size for a sample of 27 GCs using
300: HST/WFPC2. In a subsequent study, \cite{gomez06} found the red GCs
301: to have {\em larger\/} median sizes compared to the blue clusters for a
302: sample of 38 objects, with Magellan/MagIC.
303: Both studies were based on small field images, involved small
304: sample sizes, and were centered at large R$_{gc}$. (At the
305: {\it smallest} R$_{gc}$, this was more than 2 times farther
306: than the R$_{eff}$ of the galaxy light.)
307: According to \cite{larsen03}, the average sizes
308: of red and blue clusters should be similar at about 1 R$_{eff}$ and
309: beyond, if projection
310: effects are to account for the size difference.
311: The results from \cite{harris02} and \cite{gomez06} are
312: consistent with this scenario, bearing in mind the
313: low number statistics of their studies.
314:
315: The only two studies thus far with a large enough sample
316: of GCs that also extends beyond 1 R$_{eff}$, are this work and
317: \cite{spitler06}. Both show that the red clusters are smaller than the blue
318: within 1 R$_{eff}$, {\it and that they are identical in size beyond this
319: distance}. In the ACS Virgo Cluster Survey, \cite{jordan05} have studied the
320: sizes of GCs in 67 early-type galaxies.
321: Their analysis reaches about 3 times the effective
322: radius of the massive ellipticals studied and these GCSs dominate their sample.
323: However, their results, favoring red clusters being consistently smaller than
324: blue clusters are dominated by the inner GCs in these galaxies.
325:
326: Our sample, consists of only confirmed GCs which have been analyzed
327: homogeneously, span a projected galactocentric
328: distance of up to 50 kpc, i.e., 8 galaxy R$_{eff}$.
329: Thus, we are able to draw conclusions about the origin of the size difference
330: with a sample that is both uncontaminated and much more spatially
331: extended than in previous studies.
332:
333: As is evident from Figs.~\ref{fig:re_feh}, \ref{fig:N_re} and \ref{fig:re_Rgc}
334: metal-poor clusters do not show an r$_e$--R$_{gc}$
335: relationship. \cite{jordan05} analyse this trend for metal-poor clusters
336: in their samples and conclude that they are too shallow compared to the
337: Galaxy for the projection effects to account for the size difference. Our
338: results for NGC 5128 agree with this, but at the same time make it clear that
339: the metal-poor subpopulation does not represent the global r$_e$--R$_{gc}$ trend.
340: In fact, only {\em metal-rich} clusters show it.
341: Therefore, projection effects can account for the observed size
342: differences without the need of intrinsic formation and destruction mechanisms
343: between red and blue clusters.
344:
345: \section{Conclusions}
346: \label{sec:conclusions}
347:
348: Using a contaminant-free sample of 364 GCs in NGC 5128,
349: confirmed with radial velocity measurements or by resolved HST images,
350: we have measured effective radii using ISHAPE. Our results indicate
351: that the blue or metal-poor clusters do not show any significant
352: r$_{e}$--R$_{gc}$ relation. However, the red or metal-rich GCs
353: do show a steep relation in which red clusters within 1
354: R$_{eff}$ of the galaxy's light are $30\%$ smaller than the blue
355: clusters. Beyond this distance there is no indication for a size
356: difference between the two metallicity populations. This finding in
357: NGC 5128, not previously seen in any other early-type galaxy, supports
358: the more tentative findings of the Sombrero galaxy's GCS
359: \citep{spitler06}. Both
360: studies support the idea that the size differences are most likely caused by
361: projection effects \citep{larsen03}
362: and not by intrinsic physical differences between the two subgroups.
363:
364:
365: Acknowledgements: M.G. and K.A.W. thank Dean McLaughlin for use of HST
366: structural parameters in advance of publication. M.G. thanks the Dept. of
367: Physics and Astronomy at McMaster University and especially Bill and
368: Gretchen Harris for their hospitality. K.A.W. thanks NSERC and
369: Bill Harris for financial support, and also the Depto. de
370: F{\'i}sica at the Universidad de Concepci{\'o}n, especially Doug
371: Geisler, for their hospitality. We thank the anonymous referee for
372: her/his valuable suggestions and comments.
373:
374:
375: %*************************REFERENCES*************************************
376: %***************************************************************************
377: \clearpage
378:
379: \begin{thebibliography}{}
380:
381:
382: \bibitem[Aarseth \& Heggie(1998)]{aarseth98} Aarseth, S.~J. \& Heggie,
383: D.~C. 1998, \mnras, 297, 794
384:
385: \bibitem[Barmby et al.(2002)]{barmby02} Barmby, P., Holland, S., \&
386: Huchra, J.~P. 2002, \aj, 123, 1937
387:
388: \bibitem[Barmby et al.(2007)]{barmby07} Barmby, P., McLaughlin, D.~E.,
389: Harris, W.~E., Harris, G.~L.~H., \& Forbes, D.~A. 2007, \aj, 133, 2764
390:
391: \bibitem[Baggett et al.(1998)]{baggett98} Baggett, W.~E., Baggett, S.~M., \& Anderson, K.~S. 1998, \aj, 116, 1626
392:
393: \bibitem[Brodie \& Strader(2006)]{brodie06} Brodie, J.~P. \&
394: Strader, J. 2006, \araa, 44, 193
395:
396: \bibitem[C{\^o}t{\'e} et al.(2001)]{cote01} C{\^o}t{\'e}, P.,
397: McLaughlin, D.~E., Hanes, D.~A., Bridges, T.~J., Geisler, D.,
398: Merritt, D., Hesser, J.~E., Harris, G.~L.~H., \& Lee, M.~G. 2001,
399: \apj, 559, 828
400:
401: \bibitem[de Vaucouleurs \& Pence(1978)]{devaucouleurs78} de
402: Vaucouleurs, G. \& Pence, W.~D. 1978, \aj, 83, 1163
403:
404: \bibitem[Dirsch et al.(2003)]{dirsch03} Dirsch, B., Richtler, T.,
405: Geisler, D., Forte, J.~C., Bassino, L.~P., \& Gieren, W.~P. 2003, \aj,
406: 125, 1908
407:
408: \bibitem[Djorgovski(1995)]{djorgovski95} Djorgovski, S. 1995, \apj,
409: 438, 29
410:
411: \bibitem[Drinkwater et al.(2000)]{drinkwater00} Drinkwater, M.~J., Jones, J.~B., Gregg, M.~D., \& Phillipps, S. 2000, PASA, 17, 227
412:
413: \bibitem[G{\'o}mez et al.(2006)]{gomez06}G{\'o}mez, M., Geisler, D., Harris, W.~E., Richtler, T.,
414: Harris, G.~L.~H., \& Woodley, K.~A. 2006, \aap, 447, 877
415:
416: \bibitem[Harris et al.(2004)]{harris04} Harris, G.~L.~H., Harris,
417: W.~E., \& Geisler, D. 2004, \aj, 128, 723
418:
419: \bibitem[Harris(1996)]{harris96} Harris, W.~E. 1996, \aj, 112, 1487
420:
421: \bibitem[Harris \& Harris(2002)]{harrisharris02} Harris, W.~E., \& Harris, G.~L.~H. 2002, \aj, 123, 3108
422:
423: \bibitem[Harris et al.(2002)]{harris02} Harris, W.~E., Harris,
424: G.~L.~H., Holland, S.~T., \& McLaughlin, D.~E. 2002, \aj, 124, 1435
425:
426: \bibitem[Harris et al.(2006)]{harris06} Harris, W.~E., Harris,
427: G.~L.~H., Barmby, P., McLaughlin, D.~E., \& Forbes, D.~A. 2006, \aj,
428: 132, 2187
429:
430:
431: \bibitem[Jord{\'a}n (2004)]{jordan04} Jord{\'a}n, A. 2004, \apj, 613L, 117
432:
433: \bibitem[Jord{\'a}n et al.(2005)]{jordan05} Jord{\'a}n, A.,
434: C{\^o}t{\'e}, P., Blakeslee, J.~P., Ferrarese, L., McLaughlin, D.~E.,
435: Mei, S., Peng, E.~W., Tonry, J.~L., Merritt, D., Milosavljevi{\'c},
436: M., Sarazin, C.~L., Sivakoff, G.~R., \& West, M.~J. 2005, \apj, 634, 1002
437:
438: \bibitem[King (1962)]{king62} King, I.~R. 1962, \aj, 67, 471
439:
440: \bibitem[Kundu \& Whitmore(1998)]{kundu98} Kundu, A. \& Whitmore,
441: B.~C. 1998, \aj, 116, 2841
442:
443: \bibitem[Kundu et al.(1999)]{kundu99} Kundu, A., Whitmore, B.~C.,
444: Sparks, W.~B., Macchetto, F.~D., Zepf, S.~E., \& Ashman, K.~M. 1999,
445: \apj, 513, 733
446:
447: \bibitem[Kundu \& Whitmore(2001)]{kundu01} Kundu, A. \& Whitmore,
448: B.~C. 2001, \aj, 121, 2950
449:
450: \bibitem[Larsen (1999)]{larsen99} Larsen, S.~S. 1999, \aaps, 139, 393
451:
452: \bibitem[Larsen (2001)]{larsen_s01} Larsen, S.~S. 2001, \aj, 122, 1782
453:
454: \bibitem[Larsen et al.(2001a)]{larsen01} Larsen, S.~S., Brodie, J.~P., Huchra, J.~P., Forbes, D.~A., \& Grillmair, C.~J. 2001a, \aj, 121, 2974
455:
456: \bibitem[Larsen et al.(2001b)]{larsen_fb01} Larsen, S.~S., Forbes, D.~A., \&
457: Brodie, J.~P. 2001b, \mnras, 327, 1116
458:
459: \bibitem[Larsen \& Brodie(2003)]{larsen03} Larsen, S.~S. \& Brodie,
460: J.~P. 2003, \apj, 593, 340
461:
462: \bibitem[McLaughlin(2000)]{mclaughlin00} McLaughlin, D.~E. 2000, \apj,
463: 539, 618
464:
465: \bibitem[McLaughlin \& van der Marel(2005)]{mclaughlin05} McLaughlin,
466: D.~E. \& van der Marel, R.~P. 2005, \apjs, 161, 304
467:
468: \bibitem[McLaughlin et al.(2007)]{mclaughlin07} McLaughlin, D.~E.,
469: Barmby, P., Harris, W.~E., Harris, G.~L.~H., \& Forbes, D.~A. 2007, \mnras, submitted
470:
471: \bibitem[Puzia et al.(1999)]{puzia99} Puzia, T.~H., Kissler-Patig, M.,
472: Brodie, J.~P., \& Huchra, J.~P. 1999, \aj, 118, 2734
473:
474: \bibitem[Schlegel et al.(1998)]{schlegel98} Schlegel, D.~J., Finkbeiner, D.~P., \& Davis, M. 1998, \apj, 500, 525
475:
476: \bibitem[Spitler et al.(2006)]{spitler06} Spitler, L.~R., Larsen,
477: S.~S., Strader, J., Brodie, J.~P., Forbes, D.~A., \& Beasley, M.~A,
478: 2006, \aj, 132, 1593
479:
480: \bibitem[Spitzer \& Thuan(1972)]{spitzer72} Spitzer, L.~Jr., Thuan,
481: T.~X. 1972, \apj, 175, 31
482:
483: \bibitem[van den Bergh et al.(1991)]{vandenbergh91} van den Bergh, S.,
484: Morbey, C., \& Pazder, J. 1991, \apj, 375, 594
485:
486: \bibitem[Woodley et al.(2005)]{woodley05} Woodley, K.~A., Harris, W.~E.,
487: \& Harris, G.~L.~H, 2005, \aj, 129, 2654
488:
489: \bibitem[Woodley et al.(2007)]{woodley07} Woodley, K.~A., Harris,
490: W.~E., Beasley, M.~A., Peng, E.~W., Bridges, T.~J., Forbes, D.~A.,
491: \& Harris, G.~L.~H 2007, \aj, 134, 494
492:
493: \end{thebibliography}
494:
495:
496:
497:
498: %\clearpage
499:
500: % *************** FIGURES **********************
501: \clearpage
502:
503: \begin{figure}
504: \plotone{f1.eps}
505: \caption{Comparison of effective radius r$_e$, in parsecs, for 69 GCs in NGC
506: 5128 measured with both our IMACS study and HST/ACS
507: \citep{mclaughlin07}. The solid line is a 1:1 relationship.
508: A least squares fit (dashed line) gives
509: r$_e$~(IMACS)~$ = (1.02 \pm 0.04)$~r$_e$~(ACS)~$ - (0.22 \pm 0.18)$.
510: The IMACS uncertainties
511: correspond to the standard deviation of the effective radii in
512: the ISHAPE fitting, using three different concentration
513: parameters $c=15,30,100$. The agreement between these independent
514: techniques and datasets is evident.}
515: \label{fig:acs_imacs}
516: \end{figure}
517:
518: \clearpage
519:
520:
521: \begin{figure}
522: \plotone{f2.eps}
523: \caption{The effective radius, r$_e$, as a function of [Fe/H] for (from top to
524: bottom) R$_{gc} < 1$ R$_{eff}$ of NGC 5128,
525: 1 R$_{eff} <$ R$_{gc} <$ 2 R$_{eff}$, 2$ R_{eff} <$ R$_{gc} < 3 $R$_{eff}$, and
526: R$_{gc} > 3$ R$_{eff}$, where 1 R$_{eff} = 5.1'$. Least squares best fit
527: and the number of GCs are indicated in each of the four
528: panels. Only GCs with ellipticity less than 0.4 and
529: r$_e < 8 $ pc were considered in the fit. The change in the slope
530: of the fitted lines is due primarily to an increase in the median size of red
531: clusters.}
532: \label{fig:re_feh}
533: \end{figure}
534:
535: \begin{figure}
536: \plotone{f3.eps}
537: \caption{The distributions in
538: r$_e$ of the GCs in the same galactic effective radial
539: bins as in Fig.~\ref{fig:re_feh}, for metal-poor ([Fe/H]$ < -1$) on the left and
540: metal-rich ([Fe/H]$ > -1$) clusters on the right. Associated number of
541: GCs and the mean r$_e$ (open circle) are labelled in
542: each bin. Outside 3 R$_{eff}$ of the galaxy, the data on the GC
543: population suffers from incompleteness and spatial bias. The open circle
544: in each distribution gives the median value for the $r_e$ and its
545: formal error for each subsample. Metal-poor clusters show no significant
546: change of $r_e$ with R$_{gc}$, while, on average, metal-rich clusters
547: follow a clear trend
548: for larger $r_e$ with R$_{gc}$.}
549: \label{fig:N_re}
550: \end{figure}
551:
552: \begin{figure}
553: \plotone{f4.eps}
554: \caption{The effective radius, r$_e$ in pc, as a function of projected
555: galactocentric radius, R$_{gc}$ in kpc, for the GCs in NGC 5128
556: (open circles) and in the Milky Way galaxy (crosses) for both the
557: metal-poor GCs ([Fe/H]$ < -1$) on the top and metal-rich GCs
558: ([Fe/H]$ > -1$) on the bottom. The Milky Way GC data is taken
559: from \cite{harris96} with the projected galactocentric radius defined as
560: R$_{gc} = \sqrt{y^2 + z^2}$ and R$_{eff}=2.7$ kpc
561: \citep{devaucouleurs78}.
562: Best fit curves of the form
563: r$_e =$c(R$_{gc}$/R$_{eff}$)$^\alpha$ yield (c$=3.22\pm0.12$, $\alpha=0.05\pm0.05$)
564: and (c$=2.76\pm0.14$, $\alpha=0.26\pm0.06$) for the metal-poor
565: and metal-rich GCs in NGC 5128 (dashed curves)
566: along with (c$=3.02\pm0.17$, $\alpha=0.17\pm0.04$) and
567: (c$=3.16\pm0.21$, $\alpha=0.36\pm0.07$) for the
568: metal-poor and metal-rich GCs in the Milky Way (solid curves).}
569: \label{fig:re_Rgc}
570: \end{figure}
571:
572: \end{document}
573:
574: