0807.0003/ms.tex
1: %\documentclass{article}
2: %\usepackage{emulateapj}
3: 
4: %\documentclass[12pt,preprint]{aastex}
5: \documentclass{emulateapj}
6: 
7: %\documentclass[preprint2]{aastex}
8: %\documentstyle[harvey,epsf]{article}
9: %\input epsf.tex
10: 
11: \begin{document}
12: 
13: \shorttitle{Evolution of Globular Clusters}
14: \shortauthors{Kalirai et~al.}
15: 
16: \title{A Glimpse into the Past: The Recent Evolution of Globular Clusters\altaffilmark{1,2}}
17: 
18: \author{
19: Jasonjot S. Kalirai\altaffilmark{3,5}, 
20: Jay Strader\altaffilmark{4,5}
21: Jay Anderson\altaffilmark{6}, and
22: Harvey B. Richer\altaffilmark{7}
23: }
24: \altaffiltext{1} {Based on observations with the NASA/ESA Hubble Space Telescope, obtained at 
25: the Space Telescope Science Institute, which is operated by the Association of Universities 
26: for Research in Astronomy, Inc., under NASA contract NAS5-26555.  These observations are 
27: associated with proposal GO-10424.}
28: \altaffiltext{2} {Based on observations obtained at the Gemini Observatory, which is operated by the
29: Association of Universities for Research in Astronomy, Inc., under a cooperative agreement
30: with the NSF on behalf of the Gemini partnership: the National Science Foundation (United
31: States), the Particle Physics and Astronomy Research Council (United Kingdom), the
32: National Research Council (Canada), CONICYT (Chile), the Australian Research Council
33: (Australia), CNPq (Brazil) and CONICET (Argentina).}
34: \altaffiltext{3}{University of California Observatories, University of 
35: California at Santa Cruz, Santa Cruz CA, 95060; jkalirai@ucolick.org}
36: \altaffiltext{4}{Harvard-Smithsonian Center for Astrophysics, Cambridge MA, 02138; 
37: jstrader@cfa.harvard.edu}
38: \altaffiltext{5}{Hubble Fellow}
39: \altaffiltext{6}{Space Telescope Science Institute, Baltimore MD, 21218; jayander@stsci.edu}
40: \altaffiltext{7}{Department of Physics and Astronomy, University of British Columbia, 
41: Vancouver, BC, Canada, V6T~1Z1; richer@astro.ubc.ca}
42: 
43: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
44: 
45: \begin{abstract}
46: 
47: We present the serendipitous discovery of 195 extragalactic globular 
48: clusters (GCs) in one of the deepest optical images ever obtained, a 126 orbit 
49: {\it Hubble Space Telescope} ({\it HST}) Advanced Camera for Surveys (ACS) 
50: imaging study of the nearby Galactic GC NGC~6397.  The 
51: distant GCs are all found surrounding a bright elliptical 
52: galaxy in the field, and are among the faintest objects detected in the image, 
53: with magnitudes 26 $\lesssim F814W \lesssim$ 30.  We measure the redshift of the 
54: parent elliptical galaxy, using the Gemini Multi-Object Spectrograph (GMOS) on 
55: Gemini South, to be $z$ = 0.089 (375~Mpc).  This galaxy, and its associated 
56: clusters, therefore ranks as one of the most distant such systems discovered 
57: to date.  The measured light from these clusters was emitted 1.2~Gyr ago 
58: (the lookback time) and therefore the optical properties hold clues for 
59: understanding the evolution of GCs over the past Gyr.  
60: We measure the color function of the bright GCs and find that both a 
61: blue and red population exist, and that the colors of each sub-population 
62: are redder than GCs in local elliptical galaxies of comparable 
63: luminosity.  For the blue clusters, the observed color difference from $z = 0.089$ 
64: to today is only slightly larger than predictions from stellar evolution (e.g., 
65: changes in the luminosity and color of the main-sequence turnoff and the 
66: morphology of the horizontal branch).  A larger color difference is found in the 
67: red clusters, possibly suggesting that they are very metal-rich and/or significantly 
68: younger than 12~Gyr.
69: 
70: \end{abstract}
71: 
72: \keywords{galaxies: elliptical and lenticular, cD -- galaxies: star clusters 
73: -- globular clusters: general -- stars: evolution}
74: 
75: \section{Introduction} \label{introduction}
76: 
77: The formation of globular star clusters (GCs) is closely linked to the 
78: evolutionary histories of their parent galaxies.  Major epochs of star 
79: formation in isolated galaxies as well as interactions with other 
80: galaxies lead to subsequent spikes in the number of GCs 
81: \citep{schweizer87,ashman98,harris01,brodie06}.  Indeed, 
82: it appears that the processes involved in galaxy formation imprint a 
83: signature on the clusters as reflected by the similar properties of 
84: the host galaxy and its associated clusters (e.g., age and metallicity).  
85: The prevalence of a large number of GCs around all 
86: massive galaxies, as well as their optically bright luminosities ($-$10 $ 
87: \lesssim M_{V} \lesssim -$5), makes them excellent tracers of galaxy 
88: formation processes.
89: 
90: %%%%%%%%%%%%%%%%%%%%%%%%%
91: %%%%%%%%%%%%%%%%%%%%%%%%%
92: 
93: \begin{figure*}
94: \epsscale{0.95}
95: \plotone{f1.eps}
96: \figcaption{A 3.4$'$ $\times$ 1.7$'$ section of the 126 orbit 
97: {\it HST} exposure of NGC~6397 (this is approximately one-half of 
98: the total ACS field of view).  The strong gradient across the image 
99: from top-left to bottom-right reflects the decreasing radial density 
100: of NGC~6397 stars (the cluster center is located $\sim$6.5 arcminutes 
101: to the N-W of our field).  The inset shows a closer look (8$''$ $\times$ 
102: 8$''$ -- 15.9 $\times$ 15.9~kpc at the distance of the elliptical galaxy) 
103: of the large elliptical galaxy at 
104: image coordinates ($x$, $y$) = (789.9, 750.9) pixels.  The point source 
105: enhancement surrounding the elliptical galaxy is very high and suggests 
106: that these objects are GCs themselves, orbiting this 
107: galaxy.  Over the area of the inset, only one cluster white dwarf is 
108: expected to contaminate our sample.  Five of the GCs 
109: are highlighted with small circles in the inset.  Higher resolution 
110: version of this figure is available at http://www.ucolick.org/$\sim$jkalirai/DistantGlobs/.
111: Image credit: NASA, ESA, H.\ Richer (UBC), J.\ Kalirai (UCSC).
112: %furthest globular in inset is < 8 kpc from center of galaxy.
113: \label{fig:image}}
114: \end{figure*}
115: 
116: %%%%%%%%%%%%%%%%%%%%%%%%%
117: %%%%%%%%%%%%%%%%%%%%%%%%%
118: 
119: %%%%%%%%%%
120: 
121: To date, GCs have been detected in hundreds of nearby 
122: galaxies (see Table~1 in Brodie \& Strader 2006 for those systems with 
123: accurate photometry).  However, direct age measurements from resolved 
124: main-sequence turnoff fitting are only possible for the nearest clusters 
125: in the Milky Way, LMC, SMC, and M31 (e.g., see the recent study of SKHB~312 
126: by Brown et~al.\ 2004).  Properties of more distant systems are typically 
127: constrained with integrated light photometry and, where possible, 
128: spectroscopy.  Such work has demonstrated that the luminosity and color 
129: distributions of GCs in most galaxies are remarkably similar (e.g., 
130: Harris \& Racine 1979), with most populations exhibiting a bimodal color 
131: distribution reflecting old metal-poor and metal-rich subpopulations.  
132: Whether these populations reflect GCs that formed at two different 
133: times (e.g., Forbes et~al.\ 1997) is still debated.  Unfortunately, 
134: directly probing the properties of these clusters at earlier times is a 
135: difficult task.  For example, at the distance of the nearby Coma cluster, 
136: a typical GC with $M_{V}$ = $-$7.5 has an apparent 
137: magnitude of $V$ = 27.5.
138: 
139: In {\it HST} Cycle 13, our team (PI: Richer - GO-10424) was granted 126 
140: orbits of ACS time in a single pointing to explore the white dwarf cooling 
141: sequence, faint main-sequence, and space motion of the Milky Way 
142: core-collapsed GC NGC~6397 
143: \citep{richer06,hansen07,kalirai07a}.  The resulting data set represents 
144: one of the deepest optical images ever obtained in astronomy, with a 50\% 
145: (25\%) completeness limit of $F814W$ = 28 ($F814W$ = 28.75) and a 
146: photometric error of $\sigma$ = 0.3 at $F814W$ = 28.0 and $F606W$ = 29.25.  
147: In this {\it Letter}, we present the discovery of a distant, background 
148: galaxy in this data set that harbors its own GC population.  
149: We measure the redshift of the galaxy to be $z$ = 0.089, and therefore this 
150: system resides $\sim$375~Mpc from the Milky Way at a lookback time of 
151: 1.2~Gyr.  The color function of the GCs in this galaxy is 
152: found to show differences when compared to local samples suggesting the 
153: possible detection of evolutionary changes in GC properties 
154: over the past Gyr.
155: 
156: \section{Imaging and Spectroscopic Observations} \label{observations}
157: 
158: The {\it HST}/ACS imaging observations of NGC~6397 from GO-10424 have been 
159: described in several papers (e.g., Richer et~al.\ 2006; Hansen et~al.\ 2007; 
160: Kalirai et~al.\ 2007a; Anderson et~al.\ 2008).  To summarize, we 
161: obtained 252 images in $F814W$ and 126 in $F606W$, for a total 
162: integration time of 126 orbits in a single pointing centered at 
163: $\alpha_{\rm J2000}$ = 17:41:03, $\delta_{\rm J2000}$ = $-$53:44:21.  
164: Sources detected at the same position on 90 of the 252 $F814W$ 
165: images were shown to be a 3$\sigma$ detection (from simulations) 
166: and cleaned to remove false detections in the wings of bright stars, those 
167: caused by intersecting diffraction spikes, and objects morphologically 
168: inconsistent with stars (e.g., galaxies and cosmic rays).  PSF 
169: photometry on these sources was performed as described in 
170: \cite{anderson08}.  
171: 
172: Our expectation was that the final cleaned catalog of $\sim$8000 objects 
173: from this analysis represented stars on the image.  However, a 
174: spatial plot of the faint-blue sources indicated a clustering of 
175: 244 objects centered on a bright $F814W$ = 16.7 elliptical galaxy 
176: at pixel location $x$ = 789.9, $y$ = 750.9 (see Figure~\ref{fig:image}).  
177: We reduced the initial catalogue by 6 
178: objects that look extended, another 5 objects that are very bright 
179: and red, and 38 objects that were only detected in one filter and therefore 
180: for which we have no color information.  This final sample of 195 objects 
181: must be GCs themselves, in orbit around the background 
182: galaxy.  The clusters are among the faintest objects in our catalogue, 
183: ranging in brightness from 26 $\lesssim F814W \lesssim$ 30.  Given 
184: the ACS pixel scale and distance to the galaxy (see below), these 
185: clusters are all located within a radial distance of 1.7 -- 20~kpc 
186: (0.85$''$ -- 10$''$) from the center of this galaxy.  The 
187: incompleteness in our sample increases rapidly as we approach the center of 
188: the galaxy and therefore we preferentially detect the bluer clusters.  We can roughly 
189: estimate the masses of the clusters assuming they are old (12~Gyr) and 
190: metal-poor ([Fe/H] = $-$1.5).  For a Kroupa initial-mass function, the 
191: \cite{maraston05} models indicate that the GCs with $F814W = 27.5$ 
192: are $\sim$2 $\times$ 10$^6$~$M_\odot$ whereas the brightest globulars in 
193: this galaxy are $\sim$10$^7$~$M_\odot$.
194: 
195: To measure the distance to the galaxy accurately, we first obtained a ground 
196: based image of the field in excellent seeing conditions and low air-mass with 
197: the Gemini Multi-Object 
198: Spectrograph (GMOS) on Gemini South (Program ID: GS-2006A-DD-16, executed on 
199: 18~Aug.\ 2006).  Despite the confusion from NGC~6397 stars, the elliptical 
200: galaxy was easily delineated at $\alpha_{\rm J2000}$ = 17:41:12.2, 
201: $\delta_{\rm J2000}$ = $-$53:43:16.5 in a 120~s exposure with the $g'$ filter.  
202: Next, we obtained a low resolution spectrum of the galaxy with a 1$''$ longslit 
203: and the B600 grating ($R$ = 1600) centered at 5200~${\rm \AA}$.  The position 
204: angle was set to 325~degrees to avoid contamination in the slit from nearby 
205: stars.  Four 900~s target exposures, as well as observations of a flux 
206: standard, a Cu/Ar arc, and a flat field were obtained.  The data were reduced 
207: using the Gemini IRAF Package, Version 1.8.  
208: 
209: The final reduced spectrum of the elliptical galaxy in the observed wavelength 
210: frame of reference is shown in Figure~\ref{fig:cmd} (top).  The arrows mark 
211: several well-defined spectral features along with their rest-frame 
212: wavelengths.. A cross-correlation with a template spectrum of 
213: an old, metal-rich single stellar population yields a best-fit 
214: redshift of $z = 0.0894$. Assuming standard concordance 
215: cosmology with $\Omega_{m} = 0.3$, $\Omega_{\Lambda} = 0.7$, 
216: and $h = 0.7$, this redshift corresponds to a proper radial 
217: distance of 375~Mpc and a lookback time of 1.2~Gyr. The luminosity 
218: distance of 409~Mpc gives a distance modulus of $m-M = 38.06$.  
219: While this is not the most distant galaxy in which GCs 
220: have been detected (that currently belongs to Abell 1689 at $z=0.183$; 
221: Mieske et~al.\ 2004), the photometry in this galaxy is considerably 
222: deeper than in any previous study of a galaxy at a comparable distance.
223: 
224: \section{Results -- Photometric Properties of \\ Distant Globular Clusters} \label{results}
225: 
226: To probe the properties of these distant GCs, we compare 
227: their color distribution to that of globulars in local elliptical 
228: galaxies.  These galaxies nearly always have a bimodal distribution of 
229: GC colors, with peaks near $V-I \sim 0.95$ and 1.18.  There 
230: is a mild dependence of the peak GC colors with host 
231: galaxy luminosity (Strader, Brodie, \& Forbes 2004; Peng et~al.\ 2006), 
232: and so we correct these values to $V-I \sim 0.93$ and 1.16 for a galaxy 
233: with $M_{\rm V}$ = $-$20.5 (assuming $V-I$ = 1.2).  We note that our 
234: measurement of the integrated brightness of the galaxy is only roughly 
235: correct given the large number of foreground contaminating stars in the 
236: outer parts.  This does not effect our results as a 0.5 magnitude 
237: difference in the galaxy luminosity would translate to a $<$0.01 mean 
238: magnitude offset in the expected peak color of the GCs.  
239: 
240: In Figure~\ref{fig:cmd} (middle) we present the color magnitude diagram of 196 
241: GCs.  This includes the foreground globular NGC~6397 where 
242: we resolve all of the individual stars ({\it small points}) as well as the 
243: 195 extragalactic objects ({\it larger points below the white dwarf cooling 
244: sequence of NGC~6397}).  A large fraction of these distant GCs 
245: ($\sim$60\%) have very faint magnitudes and lie below the 50\% completeness limit 
246: ({\it dashed line}, $F814W$ = 28).  A total of 46 objects, or $\sim$64 
247: after accounting for incompleteness, have $F814W < 27.5$.  The foreground 
248: reddening in the direction of NGC~6397 is E($F606W - F814W$) = 0.18 
249: \citep{hansen07}, giving $A_{F814W}$ = 0.33 for a standard reddening curve. 
250: Using the distance modulus in \S\,2 and this reddening, the expected peak 
251: of the approximately lognormal GC luminosity function (GCLF) 
252: is at $F814W$ $\sim$ 30.0. It follows that we have only observed the 
253: tip of the GCLF in this galaxy. Assuming a standard $\sigma = 1.3$ for the 
254: GCLF (e.g., Kundu \& Whitmore 2001), down to $F814W = 27.5$ only $\sim$2.7\% 
255: of the total GC population is visible (assuming a Gaussian form
256: for the entire GCLF). Even though this estimate is 
257: very rough, it appears that there must be more than a thousand 
258: GCs in this galaxy. This puts the galaxy squarely in 
259: the regime of massive ellipticals, and we can safely use the peak 
260: GC colors of local ellipticals as a reference.
261: 
262: In Figure~\ref{fig:cmd} (bottom) we show a color histogram of GC 
263: candidates with $F814W < 27.5$ and a density estimate overplotted. There 
264: is a broad peak of blue objects with $F606W - F814W$ between $\sim$0.8 
265: and 1.2, and a tail to redder values. Mixture modeling of the color 
266: distribution with Nmix (see discussion in Strader et~al.\ 2006) finds 
267: that the preferred fit is two Gaussians centered at $F606W - F814W$ = 
268: 1.03 $\pm$ 0.02 and 1.29 $\pm$ 0.05. The errors are derived through 
269: bootstrapping.  While increasing photometric errors 
270: tend to broaden the color distribution as one goes to fainter magnitudes, 
271: performing the same fitting to a limit of $F814W = 28$ yields essentially 
272: the same result: a preference for two populations of GCs, 
273: with peaks at $F606W - F814W$ = 1.03 and 1.31. The small number of red 
274: GCs makes that peak value uncertain, but the derived blue 
275: peak appears consistent with a visual estimate from Figure~\ref{fig:cmd} 
276: (bottom).  
277: 
278: %%%%%%%%%%%%%%%%%%%%%%%%%
279: %%%%%%%%%%%%%%%%%%%%%%%%%
280: 
281: \begin{figure}
282: \epsscale{1.2} \plotone{f2.eps} \figcaption{{\it Top} -- A Gemini GMOS 
283: spectrum of the elliptical galaxy shown in Figure~\ref{fig:image} 
284: reveals its redshift to be $z = 0.0894$, indicating a luminosity 
285: distance of 409~Mpc.  Several absorption features and their rest-frame 
286: wavelengths are illustrated. {\it Middle} -- The proper motion selected 
287: color-magnitude diagram of the foreground cluster NGC~6397 shows a 
288: beautiful main-sequence extending from the turnoff down to the lowest 
289: mass stars that burn hydrogen (red points).  The rich white dwarf cooling 
290: sequence of the cluster is also visible.  The larger dots in the 
291: faint-blue part of the CMD represent the 195 extragalactic GCs found 
292: in the vicinity of the large elliptical galaxy in 
293: Figure~\ref{fig:image}.  {\it Bottom} -- Histogram and 
294: density estimate for the colors of 46 GCs with $F814W < 27.5$.  
295: The majority of GCs belong to the blue subpopulation, with a tail 
296: towards red objects. \label{fig:cmd}}
297: \end{figure}
298: 
299: %%%%%%%%%%%%%%%%%%%%%%%%%
300: %%%%%%%%%%%%%%%%%%%%%%%%%
301: 
302: We now compare these observed peak values to those expected from a 
303: scaling of local GC systems in a galaxy with same luminosity 
304: as our host elliptical.  To convert from $V-I$ to $F606W-F814W$, we use the 
305: empirical relation of \cite{harris07}, derived from photometry of NGC~2419. This 
306: relation is $F606W - F814W = 0.707 (V-I) + 0.033 (V-I)^{2}$. To these values we 
307: add Galactic reddening of E($F606W - F814W$) = $0.18 \pm 0.03$. Finally, we add a 
308: $k$-correction derived from Bruzual \& Charlot (2003) stellar 
309: population models. This depends slightly on the (unknown) metallicity 
310: of the stellar population, but assuming values of [Fe/H] $\sim -1.2$ 
311: and $-0.2$ for the two subpopulations gives $k$-corrections of 0.088 
312: and 0.063, respectively. If our assumed metallicities are incorrect, 
313: the effect on the $F606W - F814W$ colors is minor: a 0.1 dex change in 
314: [Fe/H] corresponds to about 0.01 mag, and 0.2 dex is a reasonable 
315: assumption for the uncertainty in the metallicity. Combining these 
316: corrections together gives predicted peak values of $F606W - F814W = 
317: 0.95 \pm 0.03$ and $1.11 \pm 0.04$ \emph{disregarding} evolution in 
318: the stellar population.  We also computed the predicted peak colors 
319: assuming the \cite{sirianni05} transformations between Vega magnitudes 
320: and ACS filters and found the same results, within the small error 
321: bars.
322: 
323: The observed change in the peak colors of blue and red GCs from 
324: $z = 0.089$ to today is 0.08 $\pm$ 0.04 and 0.18 
325: $\pm$ 0.06 magnitudes. What is the expected evolution?  Both the 
326: \cite{maraston05} and \cite{bruzual03} models predict that the red metal-rich 
327: GCs should be about 0.01 mag redder in $F606W - F814W$ at 
328: the present day, and that there should be essentially \emph{no} 
329: ($< 0.01$ mag) evolution in the blue metal-poor GCs.  
330: Therefore, to first order, our results suggest that more evolution has 
331: occurred than is predicted by the models, especially for the red 
332: clusters\footnote{Based on the tightness of NGC 6397's main-sequence, the differential 
333: reddening along this line of sight is negligible.}.  We note 
334: that some GC systems in the local universe 
335: show the peculiar property of a correlation between blue GC 
336: magnitude and color, such that the more luminous GCs are 
337: redder \citep{harris06,strader06,mieske06}. Since we sample 
338: only the brightest GCs in this distant galaxy, the mean color 
339: of the blue GCs derived above may be redder than the typical 
340: object, therefore bringing the predicted and observed values closer to 
341: the expected difference.  We therefore conclude that the behavior of the 
342: blue GCs is essentially consistent with simple stellar population 
343: models, while the red GCs are substantially redder at a lookback 
344: time of $\sim 1.2$ Gyr than expected.
345: 
346: \section{Discussion}
347: 
348: The most obvious candidate for any change in blue GC colors 
349: is the horizontal branch (HB). At old ages, its morphology is highly sensitive 
350: to a number of variables, including age.  At younger ages, GCs 
351: should typically have red HBs, which transition to blue HBs as the stellar 
352: population ages.  For old systems, the quantitative evolution in a relatively 
353: red color such as $F606W - F814W$ is expected to be small.  Our results confirm 
354: this, suggesting that blue GCs have become bluer by just 
355: 0.08 $\pm$ 0.04 magnitudes over the past 1.2~Gyr.  The \cite{maraston05} 
356: models for a 12--13 Gyr metal-poor ([$Z$/H] = $-1.35$) single stellar population 
357: predict a difference of only 0.02 mag between an object with a blue and a red HB, 
358: consistent with our result at 1.5$\sigma$.  Observations of Galactic GCs show 
359: no correlation between residuals 
360: of $V - I$ colors from a fiducial $V - I$ to [Fe/H] relation and two different 
361: parameterizations of HB morphology, suggesting that---at least for the range of HBs 
362: observed in the Galaxy---the HB has little effect on the $V - I$ color 
363: (Smith \& Strader 2007).
364: 
365: For the red GCs, our mixture modeling fit of the color distribution 
366: indicates a peak at $F606W - F814W$ = 1.29 $\pm$ 0.05, 0.18 $\pm$ 0.06 
367: magnitudes redder than the local sample.  Although our data set contains few red 
368: GCs, the observed peak is found to be at the same color irrespective 
369: of the magnitude limit adopted (i.e., pushing the sample to $F814W <$ 28).  
370: Over the past 1.2~Gyr, the expected evolution from models of a single stellar 
371: population is much smaller than this, as noted above.  One possible way to 
372: reconcile {\it some} of these observed differences is to consider whether a fraction 
373: of these clusters are significantly younger than 12~Gyr, and more metal-rich.  
374: In this case, the predicted color difference resulting from standard stellar 
375: evolution between a redshift of $z = 0.089$ and now would be larger than that 
376: calculated above.  
377: 
378: An additional effect may explain the observed color difference if the metallicities 
379: of the clusters are in fact supersolar.  Recently, \cite{kalirai07b} have shown 
380: that the remnant population of white dwarfs in the old (8~Gyr), supersolar metallicity 
381: Galactic star cluster NGC~6791 are undermassive when compared to more metal-poor 
382: systems.  These stars likely formed through a unique channel involving enhanced 
383: mass loss of the progenitor star on the red giant branch.  Such evolution naturally 
384: explains both the absence of stars near the tip of the red giant branch in this 
385: cluster and the population of the extreme horizontal branch of the system (see 
386: also Castellani \& Castellani 1993).  Although the morphology of the horizontal branch does not 
387: heavily affect the color evolution of a population in our filters, the depletion 
388: of stars near the tip of the red giant branch can be more important and lead to 
389: an increasingly bluer population as the system ages.  Of course, this mechanism 
390: would be even more efficient if the clusters are older since the envelope mass 
391: of the evolving stars is lower.  However, this mechanism does require the 
392: clusters to have [Fe/H] $\gtrsim$ $+$0.2 \citep{kilic07}, and we see few systems 
393: nearby that are this metal-rich.
394: 
395: Finally, we note that these objects are far enough away that we cannot rule out 
396: a contaminating population of ultra-compact dwarfs (UCDs).  The properties of 
397: these potentially transition objects (between GCs and dwarf galaxies) 
398: vary widely among galaxies.  For example, in NGC~3311, they are are seen as very 
399: red objects more luminous than the old and red GC population 
400: (Wehner \& Harris 2007).  However, in NGC~1407, the UCDs have larger sizes than 
401: globulars but have colors that fall between the blue and red subpopulations 
402: (Harris et~al.\ 2006).  We note that UCDs are typically found in massive galaxy 
403: clusters or ``groups'', and it is difficult to assess the environment near our 
404: elliptical galaxy given the presence of NGC~6397 stars out to several arcminutes 
405: (1$'$ = 120~kpc).
406: 
407: \section{Conclusions} \label{conclusions}
408: 
409: We have presented a study of the photometric properties of one of the most 
410: distant sample of GCs yet discovered, 195 objects at $d$ = 400~Mpc.  
411: These clusters are all found around a bright elliptical galaxy 
412: that is itself located behind one of the nearest Galactic GCs 
413: to the Sun, NGC~6397.  Given their large distance, the light from these 
414: clusters has taken 1.2~Gyr to reach us, and therefore these objects 
415: offer us a rare chance to probe the evolution of GCs in 
416: the recent past.  By comparing the color function of these clusters to nearby 
417: GCs, we find differences that suggest the mean colors of GCs have become 
418: bluer in the past Gyr.  This evolutionary change can be 
419: explained assuming standard stellar evolution for the blue clusters.  However, 
420: we find differences that suggest the mean colors of the red GCs 
421: were significantly redder 1.2~Gyr ago as compared to samples in local giant 
422: ellipticals, much redder than evolutionary models predict. 
423: 
424: \acknowledgements
425: We would like to thank the entire {\it HST}/ACS team from GO-10424 for 
426: their help in obtaining the imaging observations.  We are also grateful 
427: to the Gemini South observatory for approving our program under the Director's 
428: Discretionary Time allotment (Program ID: GS-2006A-DD-16).  We would also 
429: like to thank an anonymous referee for several useful suggestions that have 
430: improved the quality of this Letter, and for his/her patience through the 
431: refereeing process.  JSK and JS are both supported by NASA through Hubble Fellowship 
432: grants, awarded by the Space Telescope Science Institute, which is operated by the 
433: Association of Universities for Research in Astronomy, Incorporated, under NASA 
434: contract NAS5-26555. Support for this work was also provided by grant 
435: HST-GO-10424 from NASA/STScI.  The research of HBR is supported 
436: by grants from the Natural Sciences and Engineering Research Council of Canada.  
437: He also thanks the Canada-US Fulbright Program for the award of a Fulbright 
438: Fellowship.
439: 
440: \begin{thebibliography}{}
441: 
442: \bibitem[Anderson et~al.(2008)]{anderson08} Anderson, J., et~al.\ 2008, \aj, 135, 2114
443: 
444: \bibitem[Ashman \& Zepf(1998)]{ashman98} Ashman, K.~M, \& Zepf, S.~E.\ 1998, Globular 
445: Cluster Systems, Cambridge, U.~K.; New York : Cambridge University Press, 1998
446: 
447: \bibitem[Brodie \& Strader(2006)]{brodie06} Brodie, J.~P., \& Strader, J.\ 2006, 
448: \araa, 44, 193
449: 
450: \bibitem[Brown et~al.(2004)]{brown04} Brown, T.~M., Ferguson, H.~C., Smith, E., Kimble, R.~A., 
451: Sweigart, A.~V., Renzini, A., Rich, R.~M., \& VandenBerg, D.~A.\ 2004, \apjl, 613, L125
452: 
453: \bibitem[Bruzual \& Charlot(2003)]{bruzual03} Bruzual, G., \& Charlot, S.\ 2003, 
454: \mnras, 344, 1000
455: 
456: \bibitem[Castellani \& Castellani(1993)]{castellani93} Castellani, M., \& Castellani, 
457: V.\ 1993, \apj, 407, 649
458: 
459: \bibitem[Forbes, Brodie, \& Grillmair(1997)]{forbes97} Forbes, D.~A., Brodie, J.~P., \& 
460: Grillmair, C.~J.\ 1997, \aj, 113, 1652
461: 
462: \bibitem[Hansen et~al.(2007)]{hansen07} Hansen, B.~M.~S., et~al.\ 2007, 
463: \apj, 671, 380
464: 
465: \bibitem[Harris \& Racine(1979)]{harris79} Harris, W.~E., \& Racine, R.\ 1979, \araa, 17, 241
466: 
467: \bibitem[Harris(2001)]{harris01} Harris, W.~E.\ 2001, Extragalactic Star Clusters, 
468: IAU Symposium 207, ed. E. Grebel, D. Geisler, D. Minniti, astro-ph/0108355
469: 
470: \bibitem[Harris et~al.(2006)]{harris06} Harris, W.~E., Whitmore, B.~C., Karakla, D., 
471: Okon, W., Baum, W.~A., Hanes, D.~A., \& Kavelaars, J.~J.\ 2006, \apj, 636, 90
472: 
473: \bibitem[Harris et~al.(2007)]{harris07} Harris, W.~E., Harris, G.~L.~H., 
474: Layden, A.~C., \& Stetson, P.~B.\ 2007, \aj, 134, 43
475: 
476: \bibitem[Kalirai et~al.(2007a)]{kalirai07a} Kalirai, J.~S., et~al.\ 2007a, \apjl, 657, L93
477: 
478: \bibitem[Kalirai et~al.(2007b)]{kalirai07b} Kalirai, J.~S., Bergeron, P., Hansen, B.~M.~S., Kelson, 
479: D.~D., Reitzel, D.~B., Rich, R.~M., \& Richer, H.~B.\ 2007b, \apj, 671, 748
480: 
481: \bibitem[Kilic, Stanek, \& Pinsonneault(2007)]{kilic07} Kilic, M., Stanek, K.~Z., \& 
482: Pinsonneault, M.~H.\ 2007, \apj, 671, 761
483: 
484: \bibitem[Kundu \& Whitmore(2001)]{kundu01} Kundu, A., \& Whitmore, B.~C.\ 2001, \aj, 
485: 121, 2950
486: 
487: \bibitem[Maraston(2005)]{maraston05} Maraston, C.\ 2005, \mnras, 362, 799
488: 
489: \bibitem[Mieske et~al.(2004)]{mieske04} Mieske, S., et~al.\ 2004, \aj, 128, 1529
490: 
491: \bibitem[Mieske et~al.(2006)]{mieske06} Mieske, S., Hilker, M., Infante, L., 
492: \& Jordan, A.\ 2006, \aj, 131, 2442
493: 
494: \bibitem[Peng et~al.(2006)]{peng06} Peng, E.~W., et~al.\ 2006, \apj, 639, 95
495: 
496: \bibitem[Richer et~al.(2006)]{richer06} Richer, H.~B., et~al.\ 2006, Science, 313, 936
497: 
498: \bibitem[Schweizer(1987)]{schweizer87} Schweizer, F.\ 1987, Philos.\ Trans.\ R.\ Soc.\ 
499: London Ser.\ A 358, 2063
500: 
501: \bibitem[Sirianni et~al.(2005)]{sirianni05} Sirianni, M., et~al.\ 2005, \pasp, 117, 1049
502: 
503: \bibitem[Smith \& Strader(2007)]{smith07} Smith, G.~H., \& Strader, J.\ 2007, 
504: Astronomische Nachrichten, 328, 107 
505: 
506: %\bibitem[Sohn et~al.(2006)]{sohn06} Sohn, S.~T., O'Connell, R.~W., Kundu, A., 
507: %Landsman, W.~B., Burstein, D., Bohlin, R.~C., Frogel, J.~A., \& Rose, J.~A.\ 
508: %2006, \aj, 131, 866.
509: 
510: \bibitem[Strader et~al.(2004)]{strader04} Strader, J., Brodie, J.~P., Forbes, D.~A., 
511: 2004, \aj, 127, 3431
512: 
513: \bibitem[Strader et~al.(2006)]{strader06} Strader, J., Brodie, J.~P., Spitler, 
514: L., \& Beasley, M.\ 2006, \aj, 132, 2333
515: 
516: \bibitem[Wehner \& Harris(2007)]{wehner07} Wehner, E.~M.~H., \& Harris, W.~E.\ 
517: 2007, \apjl, 668, L35
518: 
519: \end{thebibliography}
520: 
521: \end{document}
522: