1: \documentclass[apjpt4]{aastex}
2: \shorttitle{Discovery of Blue Hook Stars in M54}
3: \shortauthors{Rosenberg, Recio-Blanco, \& Garc\'\i a-Mar\'\i n.}
4:
5: \begin{document}
6:
7: \title{Discovery of Blue Hook Stars in the Massive Globular Cluster M54}
8:
9: \author{Alfred Rosenberg}
10: \affil{Instituto de Astrof\'\i sica de Canarias, Via L\'actea s/n,
11: E-38205 La Laguna, Tenerife, Spain}
12: \email{alf@ll.iac.es}
13:
14: \author{Alejandra Recio-Blanco}
15: \affil{Dip. di Astronomia, Universit\`a di
16: Padova, Vicolo dell'Osservatorio 2, I-35122 Padova, Italy}
17: \email{recio@pd.astro.it}
18:
19: \and
20:
21: \author{Macarena Garc\'\i a-Mar\'\i n}
22: \affil{Dpto. Astrof\'\i sica Molecular e Infrarroja, Instituto de
23: Estructura de la Materia (DAMIR/IEM-CSIC), Serrano 113bis, E-28006
24: Madrid, Spain}
25: \email{maca@damir.iem.csic.es}
26:
27:
28: \begin{abstract}
29: We present $BV$ photometry centered on the globular cluster M54
30: (NGC~6715). The color--magnitude diagram clearly shows a blue
31: horizontal branch extending anomalously beyond the zero age horizontal
32: branch theoretical models. These kinds of horizontal branch stars
33: (also called ``blue hook'' stars), which go beyond the lower limit of
34: the envelope mass of canonical horizontal branch hot stars, have so
35: far been known to exist in only a few globular clusters: NGC~2808,
36: $\omega$~Centauri (NGC~5139), NGC~6273, and NGC~6388. Those clusters,
37: like M54, are among the most luminous in our Galaxy, indicating a
38: possible correlation between the existence of these types of
39: horizontal branch stars and the total mass of the cluster. A gap in
40: the observed horizontal branch of M54 around $T_{\rm eff} = 27\,000$~K could
41: be interpreted within the late helium flash theoretical scenario, a
42: possible explanation for the origin of those stars.
43:
44: \end{abstract}
45:
46: \keywords{Stars: horizontal branch -- Galaxy: globular
47: clusters: individual: NGC~6715 -- H-R diagram.}
48:
49: \section{Introduction}
50:
51: The horizontal branch (HB) hosts stars with a helium-burning core of
52: about 0.5 $M_{\odot}$, and a hydrogen-burning shell. The masses of the
53: hydrogen envelopes vary from more than 0.2 $M_{\odot}$ to less than
54: 0.02 $M_{\odot}$. Furthermore, the less massive the hydrogen envelope
55: is, the hotter is the corresponding HB star. In the case of a star
56: cluster, we find a color spread of the HB stars which is called the HB
57: morphology. To a first approximation, the different color extensions
58: of observed cluster HBs are described in terms of the variation of
59: metal abundance, the {\it first parameter} (metal-rich clusters tend
60: to have short red HBs, while metal-poor ones exhibit predominantly
61: blue HBs). However, some other parameter (or set of parameters) has
62: also to be at work, as clusters with nearly identical metallicities
63: can show very different HB color distributions
64: \citep{vandenbergh67,sandage67}, leading to the so called
65: {\it second parameter} debate.
66:
67: Horizontal branch stars with very low envelope masses ($\leq$ 0.02
68: $M_{\odot}$, $T_{\rm eff} > 20\,000$~K), known as extended or ``extreme HB''
69: (EHB) stars, are probably the most extreme expression of the second
70: parameter problem. They have lost up to twice the mass during the red
71: giant branch (RGB) ascent than other HB stars in the same cluster
72: \citep{dcruz96}. As a result, in contrast to the more massive blue HB
73: stars, EHB stars do not ascend the asymptotic giant branch (AGB) but
74: evolve directly onto the white dwarf domain \citep{sweigart74}.
75: Recently, \citet{whitney98} and \citet{dcruz00} revealed the existence
76: of a particular kind of EHB star: a population of hot subluminous HB
77: stars, lying up to 0.7 mag below the ZAHB and forming a hook-like
78: feature in the far-UV color--magnitude diagram (CMD) of
79: $\omega$~Centauri. These ``blue hook'' stars have effective
80: temperatures up to $40\,000$~K and cannot be produced by canonical HB
81: evolution \citep{brown01}. In the optical, for effective temperatures
82: higher than 10\,000~K, ultraviolet radiation constitutes the main part
83: of the energy flux coming from the stellar surface, making the HB, in
84: practice, vertical in the classical $V$ {\it vs.} ($B-V$) plane become
85: of bolometric correction. Hence, in optical CMDs blue hook stars are
86: located at the faintest extreme of the HB.
87:
88: In this letter, we present $BV$ photometry centered on the globular
89: cluster (GC) M54. The CMD clearly shows a blue HB anomalously
90: extending beyond ZAHB models. Previous photometric studies of this
91: cluster, in ($V-I$), were not suitable for properly revealing this
92: extremely hot stellar population. Initially, blue hook stars were
93: detected in the clusters NGC~2808 and $\omega$~Centauri. More
94: recently, \citet{busso03} have reported their presence also in the
95: blue HB tail of NGC~6388. In addition, as noted by \citet{brown01},
96: the CMD of NGC~6273 shown by \citet{piotto99} shows a blue HB
97: extending to $M_V$ $>$ 5, and therefore beyond theoretical ZAHB
98: models. All these clusters are among the most massive GCs in the
99: Galaxy, as well as M54, which is the second most massive GC known. In
100: Section 2, we describe the observations and the photometric reduction
101: techniques. In Section 3, we analyze the extended HB of M54 and its
102: interpretation inside the late helium flashers scenario. Finally, in
103: Section 4, we summarize the results and consider their wider
104: implications.
105:
106: \section{Observations and Reduction}
107:
108: The observational data base consist of four images, two in $B$ and two
109: in $V$, of 30 and 900 s each, centered on M54. Images were observed in
110: service mode on 2002 July $8$ at the ESO 3.5 m New Technology
111: Telescope (NTT), with a seeing of $\sim0.5$ arcsec. The detector, the
112: Superb Seeing Imager (SUSI-2), is a mosaic of two 2k $\times$ 4k EEV
113: CCDs, with a size of $0.08$ arcsec per pixel and a total field of view
114: of $5.5$ $\times$ $5.5$ arcmin that were binned $2\times2$.
115:
116: The images were corrected for bias and spatial sensitivity variations
117: using the respective master flats, computed as the median of all
118: available sky flats of the specific run. Afterward, photometry was
119: performed using the DAOPHOT\-/\-ALLSTAR\-/\-ALLFRAME software
120: \citep{stetson87,stetson94}.
121:
122: The absolute calibration of the observations, which include the $BVRI$
123: filters, will be published in a forthcoming paper. It is based on six
124: fields of standard stars from the catalog of \citep{landolt92}, and
125: the absolute zero-point uncertainties of our calibration are
126: $\leq0.02$ mag for each of the four bands.
127:
128: \section{The extended HB of M54 and its interpretation}
129:
130: Figure 1 shows the CMD of M54 in an area away from the crowded region
131: with $R$ $<$ 90 arcsec, where $R$ is the projected distance from the
132: cluster center. We can identify at least three stellar systems: the GC
133: M54, the Sagittarius Dwarf Galaxy, and the Milky Way bulge
134: \citep{layden00}. All stars plotted were selected using the sharp
135: parameter ($-$0.25 $<$ sh $<$ 0.25) and the $V$ error ($\leq0.1$). In
136: this paper we are interested in M54, for which the main sequence (MS),
137: RGB, and HB are clearly visible. The whole CMD has been shifted in
138: color and magnitude by $E(B-V)=0.16$ and $(m-M)_0=17.25$ in order to
139: fit the ZAHB model. The CMD shows an extended HB, which spans almost
140: 4.75 mag in $V$ and extends down to $M_V\simeq$ 5.25, i.e., $\sim$ 1.5
141: mag below the turn-off. The thick line shows the ZAHB model by
142: \citet{cassisi97} for a metallicity of [Fe/H] $=$ $-1.31$.
143:
144: Crowding and completeness experiments where also performed along the
145: HB sequence. The completeness was found to be higher than $50\%$ for
146: the entire CMD shown ($60\%$ and $70\%$ zones are labeled in Fig. 1.)
147:
148: From the crowding experiments, we have obtained the photometric
149: dispersion for synthetic stars located on the HB theoretical model
150: line, to which we have add three times the typical dispersion in color
151: (0.025~mag) measured from the region between $1.0\leq{M_V}\leq4.5$
152: of the HB. These limits are represented by the dashed lines and all 78
153: stars within these two lines are considered as the HB stars in the
154: following discussion.
155:
156: Although, because of bolometric correction, the ($B-V$) color is not a
157: good temperature indicator for the hot HB stars, the CMD clearly shows
158: a horizontal branch $\sim$ 0.5 mag deeper than the ZAHB model. At
159: least nine stars in Fig. 1 seem to have effective temperatures higher
160: than $33\,600$~K, which is the end of the ZAHB model by
161: \citet{cassisi97}. As mentioned in the introduction, a different
162: evolutionary path for populating the very hot end of the HB is
163: therefore needed. One possible explanation is that these stars
164: experience a delayed helium core flash. \citet{castellani93} and
165: \citet{dcruz96,dcruz00}, showed that even if a star suffers
166: a very high mass loss during its RGB phase, it can still ignite He
167: burning after evolving off the RGB.
168:
169: If the helium flash occurs somewhere between the tip of the RGB and
170: the top of the helium dwarf cooling curve ({\it early} hot flashers),
171: the star will settle onto the EHB without any mixing between its
172: helium core and hydrogen envelope, following a canonical evolutionary
173: path to the EHB. The reason for this can be found in the entropy
174: barrier caused by the hydrogen-burning shell, which prevents the
175: convection zone produced by the helium flash from penetrating into the
176: hydrogen envelope \citep{iben76}. Early hot flasher models by
177: \citet{brown01} predict that these stars reach a maximum temperature
178: of $31\,500$~K on the EHB, and that their luminosities are almost
179: indistinguishable from the luminosities of canonical EHB
180: stars. \citet{brown01} claimed that if the helium flash occurs while
181: the star is descending the white dwarf cooling curve ({\it late} hot
182: flashers), the entropy barrier carried by the hydrogen-burning shell
183: becomes a too weak to prevent mixing between the core and the star's
184: envelope. The resulting star will have a temperature of about
185: $37\,000$~K and will lie as much as 0.7 mag below the ZAHB in the {\it
186: UV} CMD, further supporting the argument that blue hook stars are the
187: progeny of late hot flashers. The fact that blue hook stars are both
188: hotter and more helium-rich than classical EHB stars has been
189: observationally confirmed by the spectroscopic analysis of
190: \citet{moehler02} in $\omega$~Centauri. In addition, full evolutionary
191: computation of helium flash induced-mixing in Population II stars has
192: recently been developed by \citet{cassisi03}. They modeled the
193: incursion of the He flash-driven convective zone into the H-rich
194: external layers, with a subsequent surface enrichment in He and C. In
195: agreement with the observations, models experiencing this dredge-up
196: event are significantly hotter than their counterparts with H-rich
197: envelopes. They also compare their abundance predictions of He with
198: measurements by \citet{moehler02}.
199:
200: On the other hand, the sharp transition between the early and the hot
201: flashers, corresponding to a difference in mass loss of only 10$^{-4}$
202: $M_{\odot}$ along the RGB, would produce a gap in the observed stellar
203: distribution. The CMDs of $\omega$~Centauri by \citet{kaluzny97} and
204: \citet{lee99}, and NGC~6273 by \citet{piotto99} do not seem to show an
205: EHB gap. \citet{brown01} suggest that the gap in $\omega$~Centauri is
206: perhaps blurred by the metallicity distribution of the cluster. The
207: differential reddening and magnitude limit could be the reason in the
208: case of NGC~6273. They test their blue hook models by trying to
209: reproduce the luminosity function of NGC~2808 by \citet{bedin00}. They
210: first assume that the EHB and the blue hook zone are uniformly
211: populated and then include the HB evolution of the models to brighter
212: $V$ magnitudes. They find that the gap in the observed ZAHB
213: distribution corresponds to the gap between the canonical ZAHB and
214: their blue hook models with mixed envelopes, further supporting the
215: scenario of the late hot flashers.
216:
217: In Figure 2, we compare the stellar distribution in $M_V$ of the M54
218: horizontal branch from our photometry (shown as filled triangles) with
219: that of NGC~2808 by \citet{bedin00} (open circles). The total number
220: of stars in the NGC~2808 distribution (160) was normalized to that of
221: M54 (78) for a better comparison. For the latter cluster, a distance
222: modulus of ($M-m$)$_V=$ 15.74 \citep{bedin00} was used. Error bars for
223: M54 correspond to the square root of the number of the stars per bin,
224: and the bin size is 0.30 mag. Interestingly, M54 presents a gap in the
225: stellar distribution at the same absolute $V$ magnitude ($M_V$ $\sim$
226: 4.3) as NGC~2808 (marked in Figure 2 with a vertical arrow). As
227: pointed out before, in the late helium flasher scenario, this gap
228: would separate the canonical EHB from the blue hook stars and would be
229: only apparent because of the mixed envelopes of blue hook stars
230: \citep{brown01}. If this is true, a quite abundant blue hook
231: population is clearly seen in M54 at the same absolute magnitudes
232: (effective temperatures) as in NGC~2808. From this histogram, we found
233: that the number of star in the blue hook region, compared to the whole
234: HB star sample (for $1.0 \leq M_V \leq 5.5$) is $35\%$ in the case of
235: M54 and less than $20\%$ in the case of NGC~2808. We do not know if
236: this difference is caused by the metallicities, which are very similar
237: for both clusters ([Fe/H] $= -1.25$ and [Fe/H] $= -1.11$ for M54 and
238: NGC~2808, respectively, in the Carretta \& Gratton 1997 scale
239: \citep{rutledge97}), which helps the direct comparison of their
240: HBs. On the other hand, it would again be necessary to take into
241: account the HB evolution of the models to brighter V luminosities in
242: order to explain the location of the gap at $T_{\rm eff}\sim 27\,000$~K, as
243: already claimed by \citet{brown01} for the case of NGC~2808. On the
244: other hand, we note that the first gap in the HB of NGC~2808 pointed
245: out by \citet{bedin00} at $V\sim18.5$ mag ($M_V\sim2.8$ mag) is not
246: visible in M54. On the other hand, there is an underpopulated zone at
247: lower magnitudes around $M_V\sim3.2$ mag.
248:
249: Finally, the peculiar bimodality of the NGC~2808 HB morphology, which
250: includes both a blue HB and a red HB clump, may also be present in
251: M54. This would increase the similarities in the peculiar features
252: observed in the HB of these two clusters. However, although there is
253: evidence for a red HB both in our diagram and already published $VI$
254: photometry (see, for example, \citet{sarajedini95}), this feature
255: could belong to the Sagittarius (Sgr) dwarf galaxy field.
256: \citet{layden00}, in their extensive $VI$ photometry of M54 and
257: the Sgr galaxy, conclude that their subtractions of the Sgr field from
258: the M54 CMD cannot ascertain beyond doubt whether the anonymous red HB
259: belongs to Sgr or to M54.
260:
261: On the other hand, in the recent work by \citet{monaco03}, it is clear
262: that the Sgr dwarf galaxy also possesses an old and relatively
263: metal-poor stellar component that populates the blue HB
264: tail. Interestingly, \citet{monaco03} find that the the population of
265: faintest ($V>18.6$) blue HB stars is relatively lower in the Sgr field
266: than in M54. Moreover, the spread of the stars in the hottest part of
267: the Sgr field HB is rather high. In our Figure 1 we can see that the
268: stars that we present as blue hook stars are clearly grouped in a
269: relatively small region of the CMD. They are clearly separate from
270: the MS, the rest of our CMD being extremely clean. For all these
271: reasons we believe that the stars in our CMD, located under the GAP,
272: are really cluster members.
273:
274: \section{Discussion}
275:
276: As described in the previous section, we can conclude that the M54
277: horizontal branch hosts a blue hook stellar population that extends
278: the HB to fainter magnitudes than ZAHB models. Understanding the
279: origin of hot HB stars is important not only for our fine-tuning of
280: stellar evolution theory, but has wider applications in
281: astrophysics. Indeed, hot HB are now considered to be the prime
282: contributors to the ultraviolet emission in elliptical galaxies
283: \citep{greggio90,brown01}. Blue hook stars are not, however, a common
284: feature of all GCs with an extended blue HB morphology (see for
285: instance \citep{moehler00}, for the cluster NGC~6752).
286:
287: At this point, it might be useful to examine the similarities in the
288: physical properties of this cluster with the other GCs with already
289: confirmed or suspected blue hook stars in their HBs---that
290: is NGC~2808, NGC~6388, NGC~6273, and $\omega$~Centauri---in order to shed
291: more light on the origin of this peculiar kind of star. The most
292: striking analogy between these five clusters is that they are among
293: the most massive clusters in our Galaxy. Such characteristics would
294: explain the presence of a larger EHB population, but would not, in
295: principle, be directly considered as a justification for the bluer HB
296: morphology of these clusters. The total absolute $V$ magnitude ($M_V$)
297: can give a good estimate of these clusters' total luminosities (as all
298: GCs have similar color indices and hence similar bolometric
299: corrections) and therefore a good measurement of the baryonic mass of
300: these old stellar systems. From (\citet{harris96}, updated to the new
301: catalog version of 2003 February) we find $M_V$ $=$ $-9.18$ for
302: NGC~6273, $-9.39$ for NGC~2808, $-9.42$ for NGC~6388, $-10.01$ for M54,
303: and $-10.29$ for $\omega$~Centauri.
304:
305: Nevertheless, we could think that if there is a distribution in mass
306: loss along the RGB, then the high mass loss tail of this distribution
307: would be more likely to be occupied in a more massive cluster. If this
308: is true, a correlation between the number of hot HB stars per stellar
309: mass in a cluster and its HB extension and/or total mass at constant
310: metallicity should also exist. The first results of a multivariate
311: analysis (Recio-Blanco et al., in preparation), based on a photometric
312: database of 74 GCs \citep{piotto02} seem to exclude such correlation.
313:
314: On the other hand, both M54 and $\omega$~Centauri are suspected of
315: being the nuclei of a current and former dwarf galaxy,
316: respectively. In fact, M54 must play an important role in the star
317: formation history of the Sagittarius galaxy, as it lies in the high
318: density region of Sgr \citep{ibata97}, and, as pointed out by
319: \citet{layden00}, it marks one of the earliest epochs of star
320: formation in Sgr. M54 may be similar to the nuclear star clusters seen
321: in nucleated dwarf elliptical galaxies \citep{sarajedini95}. On the
322: other hand, because of the unusual properties of $\omega$~Centauri
323: (mainly abundance variations and metallicity spread), the scenario
324: that this cluster may be also the core of a disrupted dwarf galaxy
325: (e.g., \citet{freeman93}) had a continuous infall of gas to its
326: center, leading to a variable star formation history, is becoming
327: popular. Moreover, the NGC$~$2808 HB bimodality could be interpreted
328: within a similar scenario of cluster self enrichment if we consider
329: the \citet{dantona02} suggestion of the influence of a possible second
330: generation of He-rich stars in the final cluster HB morphology. In
331: this sense, the correlation between the high mass of these clusters
332: and the existence of blue hook stars (and so of their progenitors as
333: the proposed late hot helium flashers) could also be linked to the
334: second parameter debate regarding the more general problem of GC HB
335: morphology.
336:
337: It is interesting to point out, in addition, that age differences from
338: cluster to cluster would not be enough to explain the second parameter
339: problem. This is the case for NGC$~$2808, coeval with other clusters
340: of similar metallicity but much shorter HBs such as NGC~362, NGC~1261,
341: or NGC~1851, all of them still $20\%$ younger than NGC~288, NGC~5904,
342: or NGC~6218 \citep{rosenberg99,rosenberg00a,rosenberg00b}. In fact,
343: many other massive GCs of different metallicities show particularly
344: extended HB morphologies: NGC$~$6266 ($M_V = -9.19$, [Fe/H] $\sim
345: -1.3$), NGC$~$2419 ($M_V = -9.58$, [Fe/H] $\sim -2.1$=) and NGC$~$6441
346: ($M_V = -9.64$, [Fe/H] $\sim -0.5$). Deeper photometry in the blue or
347: in the ultraviolet could reveal the presence of a blue hook population
348: for those clusters as the one already detected in NGC$~$2808,
349: $\omega$~Centauri, NGC~6388, NGC~6273, and now in M54.
350:
351: \acknowledgments
352:
353: We would like to thank our anonymous referee for many valuable
354: comments and suggestions. This report is based on observations with
355: the ESO {NTT + SUSI2}, located at the La Silla Observatory, Chile (ESO
356: proposal 69.D-0655(A)). We thank L. R.\ (``Rolly'') Bedin for
357: providing the NGC~2808 HB photometry and Santino Cassisi for his ZAHB
358: models. We are grateful to Antonio Aparicio, Carme Gallart, Giampaolo
359: Piotto and Ivo Saviane for allowing us to used these data, from a
360: larger project entitled ``Relative Ages of Outer Halo Globular
361: Clusters'', in advance of publication. ARB recognizes the sustenance
362: from {MIUR} and from {ASI}.
363:
364: \begin{thebibliography}{}
365:
366: \bibitem[Bedin et al.(2000)]{bedin00} Bedin, L.R., Piotto, G., Zoccali, M., Stetson, P.B., Saviane, I., Cassisi, S., \& Bono, G. 2000, \aap, 363, 159
367: \bibitem[Brown et al.(2001)]{brown01} Brown, T.M., Sweigart, A.V., Lanz, T., Landsman, W.B. \& Hubeny, I. 2001, \apj, 562, 368
368: \bibitem[Busso et al.(2003)]{busso03} Busso, G., Piotto, G. \& Cassisi, S. 2003, MemSAI, astro-ph/0308341
369: \bibitem[Carreta and Gratton (1997)]{carreta97} Carreta, E., \& Gratton, R.G. 1997, \aaps, 121, 95
370: \bibitem[Cassisi et al.(2003)]{cassisi03} Cassisi, S., Schlattl, H., Salaris, M. \& Weiss A., 2003,\apjl, 582, L43
371: \bibitem[Cassisi and Salaris (1997)]{cassisi97} Cassisi, S., \& Salaris M. 1997, \mnras, 285, 593
372: \bibitem[Castellani and Castellani (1993)]{castellani93} Castellani, M., \& Castellani, V. 1993, \apj, 407, 649
373: \bibitem[D'Antona et al.(2002)]{dantona02} D'Antona, F., Caloi, V., Montalb\'an, J., Ventura, P. \& Gratton, R. 2002, \aap, 395, 69
374: \bibitem[D'Cruz et al.(1996)]{dcruz96} D'Cruz, N.L., Dorman, B., Rood, R.T. \& O'Connell, R.W. 1996, \apj, 466, 359
375: \bibitem[D'Cruz et al.(2000)]{dcruz00} D'Cruz, N.L., et al. 2000, \apj, 530, 352
376: \bibitem[Freeman (1993)]{freeman93} Freeman, K.C. 1993, ASP Conf. Ser., 48, 27
377: \bibitem[Greggio and Renzini (1990)]{greggio90} Greggio, L., \& Renzini, A. 1990, \apj, 364, 35
378: \bibitem[Harris (1996)]{harris96} Harris, W. 1996, \aj, 112, 1487
379: \bibitem[Ibata et al.(1997)]{ibata97} Ibata, R.A., Wyse, R.F., Gilmore, G., Irwin, M.J. \& Suntzeff, N.B. 1997, \aj, 113, 634
380: \bibitem[Iben (1976)]{iben76} Iben, I. 1976, \apj, 208, 165
381: \bibitem[Kaluzny et al.(1997)]{kaluzny97} Kaluzny, J., Kubiak, M., Szyma\'nski, M., Udalski, A., Krsemi\'nski, W. \& Mateo, M. 1997, \aaps, 125, 343
382: \bibitem[Layden and Sarajedini (2000)]{layden00} Layden, A.C., \& Sarajedini, A. 2000, \aj, 119, 1760
383: \bibitem[Landolt (1992)]{landolt92} Landolt, A.U. 1992, \aj, 104, 340
384: \bibitem[Lee et al.(1999)]{lee99} Lee, Y.W., Joo, J.M., Sohn, Y.J., Rey, S.C., Lee, H.C. \& Walker, A.R. 1999, \nat, 402, 55
385: \bibitem[Moehler et al(2000)]{moehler00} Moehler, S., Sweigart, A.V., Landsman, W.B. \& Heber, U. 2000, \aap, 360, 120
386: \bibitem[Moehler et al(2002)]{moehler02} Moehler, S., Sweigart, A.V., Landsman, W.B. \& Driezler, S. 2002, \aap, 395, 37
387: \bibitem[Monaco et al. (2003)]{monaco03} Monaco, L., Bellazzini, M., Ferraro, F.R., \& Pancino, E. 2003, \apjl, 597, L25
388: \bibitem[Piotto et al.(1999)]{piotto99} Piotto, G., Zoccali, M., King, I.R., Djorgovski, S.G., Sosin, C., Rich, R.M. \& Meylan, G. 1999, \aj, 118, 1727
389: \bibitem[Piotto et al.(2002)]{piotto02} Piotto, G., et al. 2002, \aap, 391, 945
390: \bibitem[Rosenberg et al.(1999)]{rosenberg99} Rosenberg, A., Saviane, I., Piotto, G. \& Aparicio, A. 1999, \aj, 118, 2306
391: \bibitem[Rosenberg et al.(2000a)]{rosenberg00a} Rosenberg, A., Piotto, G., Saviane, I. \& Aparicio, A. 2000a, \aaps, 144, 5
392: \bibitem[Rosenberg et al.(2000b)]{rosenberg00b} Rosenberg, A., Aparicio, A., Saviane, I. \& Piotto, G. 2000b, \aaps, 145, 451
393: \bibitem[Rutledge et al.(1997)]{rutledge97} Rutledge, G.A., Hesser, J.E., \& Stetson, P.B. 1997, \pasp, 109, 907
394: \bibitem[Sandage and Wildey (1967)]{sandage67} Sandage, A., \& Wildey, R. 1967, \apj, 150, 469
395: \bibitem[Sarajedini and Layden (1995)]{sarajedini95} Sarajedini, A., \& Layden, A.C. 1995, \aj, 109, 1086
396: \bibitem[Stetson (1987)]{stetson87} Stetson, P.B. 1987, \pasp, 99, 191
397: \bibitem[Stetson (1994)]{stetson94} Stetson, P.B. 1994, \pasp, 106, 250
398: \bibitem[Sweigart et al.(1974)]{sweigart74} Sweigart, A.V., Mengel, J.G., \& Demarque, P. 1974, \aap, 30, 13
399: \bibitem[Van den Bergh (1967)]{vandenbergh67} Van den Bergh, S. 1967, \aj, 72, 70
400: \bibitem[Walker (1999)]{walker99} Walker, A.R. 1999, \aj, 118, 432
401: \bibitem[Whitney et al.(1998)]{whitney98} Whitney, J.H., et al. 1998, \aj, 495, 284
402:
403: \end{thebibliography}
404:
405: \clearpage
406:
407: \begin{figure}
408: \epsscale{0.7}
409: \plotone{f1.eps}
410: \caption{The $M_V$ {\it vs.} $(B-V)_0$ CMD for 15\,202 stars from the
411: observed field centered on M54. All stars plotted were selected by the
412: sharp parameter ($-$0.25 $<$ sh $<$ 0.25) and the $V$ error ($\leq$ 0.1),
413: and are all farther than 90 arcsec from the cluster center. The
414: overplotted thick line is the ZAHB model by Cassisi \& Salaris (1997),
415: for a metallicity of [Fe/H] $= -1.31$. On both sides, as dashed lines,
416: are shown the color limits of the region within which we have adopted
417: stars as HB stars. Several temperatures, obtained from the ZAHB model,
418: are also shown. Note the GAP between 25\,000~K and 29\,000~K. The 27 stars
419: below the GAP and between the dashed lines are expected to be the blue
420: hook stars. The 70\% and 60\% completeness levels, obtained from a
421: crowding experiment along the HB region, are also
422: labeled. \label{fig1}}
423: \end{figure}
424:
425: \clearpage
426:
427: \begin{figure}
428: \epsscale{0.8}
429: \plotone{f2.eps}
430: \caption{Stellar distribution of the M54 HB (filled triangles)
431: from our photometry and NGC~2808 (open circles) from Bedin et
432: al. (2000). The same gap at $M_V$ $\sim$ 4.3 is observed, possibly
433: separating the canonical HB from blue hook stars.
434: \label{fig2}}
435: \end{figure}
436:
437: \end{document}
438:
439:
440: