1: %
2:
3: %% This produces a one-column, single-spaced document.
4: %\documentclass[12pt,preprint]{aastex}
5:
6: %% This produces a one-column, single-spaced document.
7: \documentclass[12pt,preprint]{aastex}
8:
9: %% preprint2 produces a double-column, single-spaced document:
10: % \documentclass[preprint2]{aastex}
11:
12: %% manuscript produces a one-column, double-spaced document:
13: % \documentclass[manuscript]{aastex}
14:
15: %% This emulates the ApJ
16: % \documentstyle[emulateapj5]{aastex}
17:
18: % Next 14 lines from Haldan to embed ps figures
19: %\documentclass[12pt,preprint]{aastex}
20: %\usepackage{emulateapj5}
21: %\usepackage{epsfig}
22: %\usepackage{apjfonts}
23: %
24: %\makeatletter
25: %\newenvironment{tablehere}
26: % {\def\@captype{table}}
27: % {}
28: %
29: %\newenvironment{figurehere}
30: % {\def\@captype{figure}}
31: % {}
32: %\makeatother
33:
34:
35: \newcommand{\nd}{\nodata}
36: \newcommand{\mdot}{M$_{\odot}$}
37: \newcommand{\mv}{$M_V$}
38: \newcommand{\hst}{{\it HST}}
39: \newcommand{\cha}{{\it Chandra}}
40: \newcommand{\teff}{$\mathrm{T_{eff}}$}
41: \newcommand{\lx}{$L_{x}$}
42: \newcommand{\fx}{$F_{x}$}
43: \newcommand{\fxfopt}{$F_X/F_{opt}$}
44: \newcommand{\fopt}{$F_{opt}$}
45: \newcommand{\ergs}{erg~s$^{-1}$}
46:
47: %\slugcomment{Draft version \#6}
48: %\slugcomment{Accepted for publication by {\it The Astrophysical Journal}}
49:
50: \shorttitle{Optical Counterpart to Supersoft Source in M3}
51: \shortauthors{Edmonds et al.}
52:
53: \begin{document}
54:
55: \title{\hst\ Discovery of an Optical Counterpart to the Supersoft X-ray
56: Source in the Globular Cluster M3\altaffilmark{1}}
57:
58: \altaffiltext{1}{Based on observations with the NASA/ESA {\it Hubble Space
59: Telescope} obtained at STScI, which is operated by AURA, Inc. under NASA
60: contract NAS 5-26555.}
61:
62: \author{Peter D. Edmonds\altaffilmark{2}, Peter
63: Kahabka\altaffilmark{3}, \& Craig O. Heinke\altaffilmark{2}}
64: \altaffiltext{2}{Harvard-Smithsonian Center
65: for Astrophysics, 60 Garden St, Cambridge, MA 02138;
66: pedmonds@cfa.harvard.edu, cheinke@cfa.harvard.edu}
67: \altaffiltext{3}{Sternwarte Universit\"at Bonn, Auf dem H\"ugel 71, D-53121 Bonn,
68: Germany; pkahabka@astro.uni-bonn.de}
69:
70: \begin{abstract}
71:
72: We report the detection, with \hst, of an optical counterpart to the
73: transient supersoft X-ray source 1E~1339.8+2837, in the globular cluster
74: M3. The counterpart is found near the faint end of the subgiant branch in
75: the $V$ vs $V-I$ color magnitude diagram, but is extremely bright in
76: $U$. Variability is detected over a range of timescales suggesting the
77: presence of an accretion disk and perhaps also ellipsoidal variations of
78: the subgiant secondary. The optical colors of the binary are similar to
79: those of cataclysmic variables recently discovered in 47~Tucanae and
80: NGC~6397. We suggest that magnetically channeled accretion may explain the
81: relatively low X-ray luminosity of this source's supersoft state.
82:
83: \end{abstract}
84:
85: \keywords{binaries: general -- globular clusters: individual (M3)
86: -- techniques: photometric -- X-rays: binaries}
87:
88: \section{Introduction}
89: \label{sect.intro}
90:
91: Supersoft X-ray sources (SSSs) are characterized by high bolometric
92: luminosities ($10^{36}-10^{38}$ \ergs) and extremely soft X-ray spectra
93: (blackbody temperatures of 15--80 eV). These characteristics are best
94: explained by white dwarfs (WDs) that are burning hydrogen-rich matter
95: accreted onto their surface from a companion star at the rate of about
96: $10^{-7}$ \mdot\ per year (Kahabka \& van den Heuvel 1997). These accretion
97: rates are about a hundred times higher than those typically found in
98: cataclysmic variables (CVs), the more common form of mass accreting binary
99: containing a WD. Such high accretion rates can be supplied by mass transfer
100: on a thermal time scale from a slightly evolved companion star (mass
101: 1.3--2.5 \mdot) that is more massive than the WD (Kahabka \& van den Heuvel
102: 1997).
103:
104: Of the nine known SSSs in the Milky Way (Kahabka \& van den Heuvel 2003),
105: only one is found in a globular cluster. A low luminosity X-ray source
106: (\lx\ $\sim 7 \times 10^{33}$ \ergs) named 1E~1339.8+2837 was discovered
107: just inside the 33$''$ core radius of M3 (NGC~5272) by {\it Einstein}
108: (Hertz \& Grindlay 1983; we assume a distance of 10.1 kpc to M3, using
109: Djorgovski 1993). This source had a much higher luminosity of $2
110: \times 10^{35}$ \ergs\ when observed by ROSAT in January 1991 and January
111: 1992 (Hertz, Grindlay, \& Bailyn 1993) and a very soft spectrum ($kT
112: \approx 20$ eV, Hertz et al. 1993; $kT = 36 \pm 20$ eV, Dotani et
113: al. 1999). This supersoft outburst is 10-100 times less luminous than most
114: other SSSs (see Kahabka \& van den Heuvel 2003 and Kahabka \& van den Heuvel
115: 1997). The ROSAT luminosity and temperature imply a WD radius of only $9
116: \times 10^{7}$ cm, suggesting that only part of the WD surface is
117: undergoing burning. The source has been seen in quiescence (\lx\ $\sim
118: 10^{33}$ \ergs) with a hard X-ray spectrum several times since the
119: 1991/1992 outburst (6/1992, 7/1993, 7/1994, 7/1995 with ROSAT, and 1/1997
120: with ASCA; Dotani et al. 1999).
121:
122: Because 1E~1339 is located in a globular cluster its distance and reddening
123: are well known, helping in modeling of the system. Also, excellent
124: constraints on the binary parameters are possible if the secondary in the
125: system is found, combined with good optical photometry. No successful
126: identification of an optical companion has yet been made, but very few
127: searches have been reported in the literature. Hertz et al. (1993) report a
128: star in the ROSAT error circle with UV magnitudes consistent with those of
129: a horizontal branch star, and Geffert (1998) reports a blue star and a
130: variable star within a 5$''$ error circle, but provides no further details.
131:
132: Here, we report the first detection of an optical companion to 1E~1339 in
133: M3 using the ROSAT X-ray position determined by Verbunt (2001). \hst\
134: imaging is used to combat crowding near the center of the cluster. We
135: briefly discuss the implications of this detection, addressing the origin
136: of the supersoft X-ray emission, and the parameters of the binary system.
137: More details about the implications of these results will be given in
138: Kahabka et al. (2004; in preparation).
139:
140: \section{Observations, Data Analysis and Results}
141: \label{sect.obs}
142:
143: The two archival WFPC2 datasets analyzed for this paper are program GO-5496
144: obtained in April 1995 (PI: F. Pecci) and program GO-6805 obtained in May
145: 1998 (PI: C. Bailyn). These observations were obtained when the SSS was
146: likely in quiescence (see above). Only a subset of the WF2 images in
147: GO-5496 were analyzed (those with intermediate length exposures), while all
148: of the WF2 images for GO-6805 were analyzed (see Table~\ref{tab.data}). In
149: both of these datasets the candidate for the optical counterpart to 1E~1339
150: was found on the WF2 chip, and only the images for this chip are analyzed
151: here.
152:
153: \subsection{Photometry}
154: \label{sect.phot}
155:
156: For both the GO-5496 and GO-6805 datasets, the F336W, F555W and F814W
157: images were combined into deep, doubly oversampled images using the drizzle
158: routines of Fruchter \& Hook (2002). This procedure allowed for the removal
159: of cosmic rays. Figure~\ref{fig.fchart} shows a small region of the WF2
160: images in each filter for both datasets. The dashed circle (radius =
161: 2\farcs2, or 2-$\sigma$) is centered on the ROSAT position of 1E~1339
162: (Verbunt 2001). Note the shift in the position of this circle between the
163: two \hst\ observing epochs because of $\sim 1''$ uncertainties in the \hst\
164: astrometry. One star in the X-ray error circle appears from inspection to
165: be relatively bright in $U$, and is labeled as Star A. The offsets between
166: the nominal X-ray position of the SSS and Star A are 0\farcs97
167: (0.88-$\sigma$) and 0\farcs36 (0.33-$\sigma$) for the GO-5496 and GO-6805
168: datasets respectively.
169:
170: To create color magnitude diagrams (CMDs) for the WF2 chip, star lists were
171: created in each of the three filters and PSF fitting was performed on the
172: deep images. The instrumental F336W, F555W and F814W magnitudes were then
173: converted into $U$, $V$ and $I$ magnitudes using the zeropoints and color
174: corrections of Holtzman et al. (1995). Stars were included in a CMD if they
175: were detected in either $U$ and $V$ or $V$ and $I$ (the maximum matching
176: distance between filters was 1.5 oversampled pixels).
177:
178: The resulting $U$ vs $U-V$ and $V$ vs $V-I$ CMDs are shown in
179: Fig.~\ref{fig.cmds}. Stars within 1\farcs5 of the nominal position of
180: 1E~1339 are circled. The photometric scatter for these objects is slightly
181: larger than average because the SSS is located in the most crowded region
182: of the WF2 chip, the area closest to the center of M3. Only two of the
183: circled objects clearly have unusual colors, and these stars (denoted Stars
184: A and B) are labeled in each CMD. Star A is bright and has extremely blue
185: colors in the $U$ vs $U-V$ CMD, but falls close to the subgiant branch
186: ridgeline in the $V$ vs $V-I$ CMD (the photometry is given in
187: Table~\ref{tab.phot}). A comparison of the CMD position of Star A between
188: the 1995 and 1998 epochs shows evidence for variability, since the star
189: appears to be slightly fainter and redder in the 1998 dataset in both the
190: $U$ vs $U-V$ and $V$ vs $V-I$ CMDs (the difference is less obvious with the
191: redder color). This evidence for variability is confirmed by time series
192: analysis given in \S~\ref{sect.tseries}.
193:
194: Star A is not a blue straggler because its $U-V$ colors in both epochs
195: ($-0.6$ and $-0.45$) are much bluer than the mean $U-V$ color of the blue
196: straggler sequence ($\sim 0.25$), and it lies on the subgiant branch in the
197: $V$ vs $V-I$ CMD. This `blue excess' in $U$ and secondary (subgiant)
198: dominated luminosity in $V$ and $I$ is similar to that found for the CVs in
199: NGC~6397 (Cool et al. 1998) and 47~Tuc (Edmonds et al. 2003), suggesting
200: that 1E~1339 is also a CV. Extra support for this hypothesis is given in
201: \S~\ref{sect.tseries}. A key difference between this system and nearly all
202: of the CVs found in 47~Tuc and NGC~6397 is that Star A includes a subgiant
203: secondary, with \mv\ ranging between 3.01 and 3.19. The one exception is
204: AKO~9, the 1.16 day period CV in 47~Tuc (Edmonds et al. 2003).
205:
206: Star B is located near the outer edge of the 2-$\sigma$ error circle in the
207: GO-5496 data, and just outside the 2-$\sigma$ error circle in GO-6805. Its
208: positions in the $U$ vs $U-V$ and $V$ vs $V-I$ CMDs suggest that the star
209: is a luminous red giant that falls well behind the cluster. Therefore Star
210: B is unlikely to be responsible for the supersoft X-ray emission.
211:
212: \subsection{Time Series}
213: \label{sect.tseries}
214:
215: Figure~\ref{fig.tseries} shows time series in F336W, F555W and F814W for
216: Star A, with example time series for two stars with similar
217: magnitudes, offset by $\sim$0.5 mag (for clarity). To derive errors, we
218: analyzed the time series of 25 stars with magnitudes that are within
219: several tenths of a magnitude of Star A. The 1-$\sigma$ error bars
220: plotted in Fig.~\ref{fig.tseries} show the mean standard deviation for
221: these 25 stars, with 3-$\sigma$ deviations removed.
222:
223: Star A is clearly variable. The standard deviations of the time series for
224: Star A exceed the mean standard deviations of the 25 reference stars by
225: factors of 9.8 (F336W), 4.3 (F555W) and 2.6 (F814W) for GO-5496, and by
226: factors of 5.6 (F336W), 3.2 (F555W) and 3.0 (F814W) for GO-6805. We fit
227: straight lines ($y=ax+b$) to the data (where $y$ is in units of magnitudes
228: and $x$ is in units of time measured in days) using least squares fitting
229: (see Fig.~\ref{fig.tseries}). For the GO-6805 data we derived
230: $a=-(0.38\pm0.14)$ mag/day (2.8-$\sigma$; F336W), $a=-0.23\pm0.05$ mag/day
231: (4.7-$\sigma$; F555W), and $a=-0.20\pm0.04$ mag/day (5.1-$\sigma$; F814W).
232: None of the reference time series in any of the 3 filters had $a > 0.088$
233: mag. Similar fits are shown for the GO-5496 data, although the short time
234: series mean that the derived parameters are less useful.
235:
236: After subtracting the fitted straight lines from the GO-6805 data, we found
237: that the standard deviation of the F336W time series is still significantly
238: larger than expected (4.9-$\sigma$), while the excesses for the F555W and
239: F814W filters are only marginally significant (2.5-$\sigma$ and 2.0-$\sigma$
240: respectively). Clearly, the GO-6805 variability shows a drift over
241: the 0.5 day observation added to variations at higher frequencies.
242:
243: The slow variations shown in Fig.~\ref{fig.tseries} may, in part, be
244: ellipsoidal variations. These are expected for systems where the secondary
245: is filling its Roche lobe, and should be detectable, given adequate
246: signal-to-noise, in systems where the secondary dominates the light in
247: particular filters (as the SSS does in $V$ and $I$). If the variations are
248: ellipsoidal in origin, then their approximately linear nature shows we are
249: observing no more than a quarter of the SSS's orbital period. This implies
250: that the orbital period of the SSS must be at least $\sim$ 2 days. Using
251: the stellar evolution models of Bergbusch \& Vandenberg (1992), Kepler's
252: third law and the Roche lobe equation of Paczy{\'n}ski (1971) we estimate
253: that the period of the SSS should be just over 2 days (assuming that the
254: star's radius does not change significantly from its equilibrium
255: value). Longer monitoring with \hst\ could easily test for the expected
256: periodic variations. Ground-based observations will require excellent
257: seeing or adaptive optics because the SSS is only 0\farcs9 away from a much
258: brighter neighboring star\footnote{This saturated star is either a red
259: giant or a horizontal branch star and is identified as the variable star
260: V224, when using the finding chart of Bakos, Benko, \& Jurcsik
261: (2000). However, Corwin and Carney (2001) found no evidence for it being
262: variable. A possible solution to this mystery is that the nearby SSS went
263: through an outburst and its variability was then so extreme that it
264: affected measurements of the nearby bright star.}.
265:
266: Changes in the brightness of the accretion disk may also contribute to the
267: slow variations. For example, some contribution from an accretion disk may
268: be visible in the $V$ vs $V-I$ CMD for GO-5496, where the SSS is bluewards
269: of the subgiant branch by 0.05--0.1 mag. Also, the magnitude differences
270: between the GO-5496 and GO-6805 observations are probably caused by some
271: difference in accretion luminosity. The shorter term variations are almost
272: certainly flickering from the accretion disk.
273:
274: \section{Discussion}
275: \label{sect.disc}
276:
277: %
278: % Absolute mag: 18.13-15.12=3.01; 18.31-15.12=3.19
279: %
280:
281: To summarize, the properties of Star A (blue $U-V$ color, variability, and
282: the close astrometric match with the ROSAT position of the SSS) mean that
283: it is almost certainly the optical counterpart of 1E~1339. The optical
284: colors of the SSS are similar to those of the CVs in 47~Tuc and NGC~6397,
285: suggesting that 1E~1339 may be a type of CV, consistent with the scenarios
286: predicted by Hertz et al. (1993) and Dotani et al. (1999).
287:
288: One popular scenario to allow the high mass transfer rates required for
289: steady nuclear burning on the surface of the WD is for mass transfer to
290: occur on a thermal timescale. This is believed to occur in binaries with a
291: slightly evolved companion star that is more massive than the WD primary
292: (Kahabka \& van den Heuvel 1997). With a secondary mass for the SSS of no
293: more than 0.9 \mdot\ (less depending on how much mass has been lost), it
294: should be easy to fulfill this requirement by normal stellar evolution in a
295: primordial binary or an exchange collision between a primordial binary and
296: a WD. WDs with masses of $\sim$0.55 \mdot\ are currently being produced in
297: the cluster (applying the initial-to-final mass relation of Weidemann
298: 2000), and WDs with similar masses will be common. Therefore, our results
299: are broadly consistent with expectations of high mass transfer rates to
300: explain nuclear burning on the surface of the WD primary. However, a
301: slightly evolved companion star with a mass between 1.3 and 2.5
302: \mdot\ is ruled out by these observations, since the only globular cluster
303: stars in this mass range (besides neutron stars) are blue stragglers.
304:
305: A useful diagnostic for this system is the \fxfopt\ ratio. We used the
306: 0.5-2.5 keV X-ray flux values of Dotani et al. (1999) from April and June
307: 1995, and January 1997, and \fopt\ $=10^{-0.4V-5.43}$ to derive
308: \fxfopt=0.9--2.8. These large \fxfopt\ values are consistent with those
309: found for magnetic (DQ~Her) systems by Verbunt et al. (1997). We speculate
310: that magnetic behavior might explain why only part of the WD surface
311: appeared to undergo hydrogen burning in 1991/1992. It has already been
312: argued that large numbers of magnetic CVs may be found in globular clusters
313: (Grindlay et al. 1995 \& Edmonds et al. 1999). We note that studies of
314: field CVs have shown that some CVs can turn on as SSSs (Greiner et al. 1999
315: \& Patterson et al. 1998), with luminosities ($10^{35}-10^{36}$ \ergs) that
316: are lower than those of the original set of Milky Way and Magellanic Cloud
317: SSSs. One of these, T~Pyx, might be a magnetic CV (Patterson et al. 1998).
318:
319: \cha\ observations, led by C.O.H., will provide new constraints on the
320: X-ray spectrum and variability of 1E~1339, and, with a larger sample of
321: optically identified X-ray sources, should give us excellent (0.1-0.2")
322: relative astrometry between \hst\ and \cha. One dataset was obtained in
323: November 2003 and two will follow in 2004. The second of these \cha\
324: observations will be coordinated with optical and UV observations with
325: \hst/ACS in 2004. Spectroscopic studies of this relatively bright binary
326: should be possible with \hst/STIS, but will be a challenge for ground-based
327: telescopes because of the nearby giant star.
328:
329: \acknowledgments We acknowledge Ron Elsner for useful discussions.
330:
331: %\clearpage
332:
333: \begin{thebibliography}{}
334: \bibitem[Bakos, Benko, \& Jurcsik(2000)]{bak00} Bakos, G.~A., Benko, J.~M.,
335: \& Jurcsik, J.\ 2000, Acta Astronomica, 50, 221
336: \bibitem[Bergbusch \& Vandenberg(1992)]{ber92} Bergbusch, P.~A.~\&
337: Vandenberg, D.~A.\ 1992, \apjs, 81, 163
338: \bibitem[Corwin \& Carney(2001)]{car01} Corwin, T.~M.~\& Carney, B.~W.\
339: 2001, \aj, 122, 3183
340: \bibitem[Cool et al.(1998)]{coo98} Cool, A.~M., Grindlay, J.~E., Cohn,
341: H.~N., Lugger, P.~M., \& Bailyn, C.~D.\ 1998, \apjl, 508, L75
342: \bibitem[Djorgovski(1993)]{djo93} Djorgovski, S.\ 1993, ASP
343: Conf.~Ser.~ 50: Structure and Dynamics of Globular Clusters, 373
344: \bibitem[Dotani, Asai, \& Greiner(1999)]{dag99} Dotani, T., Asai, K., \&
345: Greiner, J.\ 1999, \pasj, 51, 519
346: \bibitem[Edmonds et al.(1999)]{edm99} Edmonds, P.~D.,
347: Grindlay, J.~E., Cool, A., Cohn, H., Lugger, P., \& Bailyn, C.\ 1999,
348: \apj, 516, 250
349: \bibitem[Edmonds, Gilliland, Heinke, \& Grindlay(2003)]{edm03} Edmonds,
350: P.~D., Gilliland, R.~L., Heinke, C.~O., \& Grindlay, J.~E.\ 2003, \apj,
351: 596, 1177
352: \bibitem[Fruchter \& Hook(2002)]{fru02} Fruchter, A.~S.~\& Hook, R.~N.\
353: 2002, \pasp, 114, 144
354: \bibitem[Geffert(1998)]{gef98} Geffert, M.\ 1998, \aap, 340, 305
355: \bibitem[Greiner et al.(1999)]{gre99} Greiner, J., Tovmassian, G.~H., Di
356: Stefano, R., Prestwich, A., Gonz{\' a}lez-Riestra, R., Szentasko, L., \&
357: Chavarr{\'{\i}}a, C.\ 1999, \aap, 343, 183
358: \bibitem[Grindlay et al.(1995)]{gri95} Grindlay, J.~E., Cool, A.~M.,
359: Callanan, P.~J., Bailyn, C.~D., Cohn, H.~N., \& Lugger, P.~M.\ 1995,
360: \apjl, 455, L47
361: \bibitem[Hertz \& Grindlay(1983)]{her83} Hertz, P.~\& Grindlay, J.~E.\
362: 1983, \apj, 275, 105
363: \bibitem[Hertz, Grindlay, \& Bailyn(1993)]{hgb93} Hertz, P., Grindlay,
364: J.~E., \& Bailyn, C.~D.\ 1993, \apjl, 410, L87
365: \bibitem[Holtzman et al.(1995)]{hol95} Holtzman, J.~A., Burrows, C.~J.,
366: Casertano, S., Hester, J.~J., Trauger, J.~T., Watson, A.~M., \&
367: Worthey, G.\ 1995, \pasp, 107, 1065
368: \bibitem[Kahabka \& van den Heuvel(1997)]{kv97} Kahabka, P.~\& van den
369: Heuvel, E.~P.~J.\ 1997, \araa, 35, 69
370: \bibitem[Kahabka \& van den Heuvel(2003)]{kv03} Kahabka, P.~\& van den
371: Heuvel, E.~P.~J.\ 2003, in Compact Stellar X-Ray Sources,
372: eds. W.~H.~G. Lewin and M. van der Klis, Cambridge University Series
373: (submitted)
374: \bibitem[Paczy{\'n}ski(1971)]{pac71} Paczy{\'n}ski, B.\ 1971,
375: \araa, 9, 183
376: \bibitem[Patterson et al.(1998)]{pat98} Patterson, J.~et al.\ 1998, \pasp,
377: 110, 380
378: \bibitem[Verbunt, Bunk, Ritter, \& Pfeffermann(1997)]{ver97} Verbunt, F.,
379: Bunk, W.~H., Ritter, H., \& Pfeffermann, E.\ 1997, \aap, 327, 602
380: \bibitem[Verbunt(2001)]{ver01} Verbunt, F.\ 2001, \aap, 368, 137
381: \bibitem[Weidemann(2000)]{wei00} Weidemann, V.\ 2000, \aap, 363, 647
382:
383:
384: \end{thebibliography}
385:
386: \pagebreak
387:
388: \input{tab1.tex}
389:
390: \input{tab2.tex}
391:
392: \pagebreak
393:
394: %%--------------------------FIGURE----------------------------------
395: \begin{figure*}
396: %\epsfig{file=sf1.eps,width=13.5cm}
397: \epsscale{0.85}
398: \plotone{sf1.eps}
399: \caption{Finding charts for the optical counterpart to the SSS (Star
400: A). F336W ($\sim U$), F555W ($\sim V$) and F814W ($\sim I$) images are
401: shown for both the GO-5496 (1995) and GO-6805 (1998) datasets. The
402: dashed-line shows a 2-$\sigma$ (2\farcs2) error circle for the SSS from
403: Verbunt (2001). Star A, the likely optical counterpart to the SSS is
404: labeled with two arrows, and Star B, a possible background red giant, is
405: labeled with a single arrow.}
406: \label{fig.fchart}
407: \end{figure*}
408: %%-------------------------------------------------------------------
409:
410: %%--------------------------FIGURE----------------------------------
411: %\begin{figurehere}
412: \begin{figure}
413: %\epsfig{file=sf2.eps,width=11.5cm}
414: \epsscale{0.7}
415: \plotone{sf2.eps}
416: \caption{$U$ vs $U-V$ and $V$ vs $V-I$ CMDs for the WF2 chip from the
417: GO-5496 and GO-6805 datasets. Star A, the SSS ID, and Star B, a likely
418: non-cluster member are labeled with `X's, and stars within 1\farcs5 of the
419: nominal position of 1E~1339 are circled. Note the different positions of
420: Star A in the GO-5496 and GO-6805 CMDs.}
421: \label{fig.cmds}
422: %\end{figurehere}
423: \end{figure}
424: %%-------------------------------------------------------------------
425:
426: %%--------------------------FIGURE----------------------------------
427: %\begin{figurehere}
428: \begin{figure}
429: %\epsfig{file=f3.eps,width=11.5cm}
430: \epsscale{0.7}
431: \plotone{f3.eps}
432: \caption{Time series for Star A in the GO-5496 and GO-6805 datasets. Time
433: series for two reference stars with similar magnitudes to those of Star A
434: are also shown (with appropriate zeropoint shifts to avoid
435: confusion). Straight-line fits to the time series of Star A are shown, plus
436: 1-$\sigma$ error bars (with approximately the same size as the plotted
437: symbols in some cases).}
438: \label{fig.tseries}
439: %\end{figurehere}
440: \end{figure}
441: %%-------------------------------------------------------------------
442:
443: %---------------------------------------------------------------------------
444:
445: \end{document}
446:
447:
448: