1: \documentclass[12pt,preprint]{aastex}
2: %\usepackage{mkfig}
3: %\usepackage{emulateapj5}
4:
5: \begin{document}
6:
7: \title{\bf Mira~B Rejuvenated\altaffilmark{1}}
8:
9: \author{Brian E. Wood\altaffilmark{2} and Margarita Karovska\altaffilmark{3}}
10:
11: \altaffiltext{1}{Based on observations with the NASA/ESA Hubble Space
12: Telescope and the NASA-CNES-CSA Far Ultraviolet Spectroscopic Explorer.
13: The former observations were obtained at the Space Telescope Science
14: Institute, which is operated by the Association of Universities for
15: Research in Astronomy, Inc., under NASA contract NAS5-26555.
16: FUSE is operated for NASA by the Johns Hopkins University under NASA
17: contract NAS5-32985.}
18: \altaffiltext{2}{JILA, University of Colorado, Boulder, CO
19: 80309-0440; woodb@origins.colorado.edu.}
20: \altaffiltext{3}{Smithsonian Astrophysical Observatory, 60 Garden St.,
21: Cambridge, MA 02138; mkarovska@cfa.harvard.edu.}
22:
23: \begin{abstract}
24:
25: Recent ultraviolet spectra from the {\em Hubble Space Telescope}
26: (HST) and {\em Far Ultraviolet Spectroscopic Explorer} (FUSE) satellites
27: demonstrate that UV line and continuum fluxes observed from Mira~B are
28: increasing back towards the levels that the {\em International Ultraviolet
29: Explorer} observed in 1979--1980 and 1990--1995, after having been
30: found in a very low state by HST and FUSE in 1999--2001. The UV emission
31: is associated with accretion of material onto Mira~B from Mira~A's
32: massive wind, so the variability is presumably due to variations
33: in accretion rate. From wind absorption features, we estimate a
34: Mira~B mass loss rate of $2.5\times 10^{-12}$ M$_{\odot}$~yr$^{-1}$,
35: indicating that Mira~B's wind has increased in strength along with the
36: accretion rate. The UV variability may be associated with a previously
37: reported 14-year periodicity in Mira~B's optical emission.
38:
39: \end{abstract}
40:
41: \keywords{accretion, accretion disks --- binaries: close --- stars:
42: individual (o Ceti) --- stars: winds, outflows --- ultraviolet: stars}
43:
44: \section{INTRODUCTION}
45:
46: Mira~AB is the nearest symbiotic system at a distance of only
47: $\sim130$~pc, \citep{macp97}, and the only symbiotic binary in
48: which the components of the system have been resolved at wavelengths
49: ranging from X-rays to radio \citep{mk97,mk05,ldm06}. The primary (Mira,
50: o~Ceti, HD~14386) is a very large pulsating AGB star (photospheric
51: diameter of $\sim500$~R$_{\odot}$), with a massive, cool wind of
52: $\sim 10^{-7}$ M$_{\odot}$~yr$^{-1}$ \citep{pfb88}.
53: The companion (Mira~B, VZ~Ceti) is accreting material from Mira~A, but
54: the exact nature of the accretor is uncertain \citep{mj84};
55: it has usually been assumed to be a white dwarf \citep{ahj54,mk05}.
56: The current angular separation between the components is
57: $\sim$ $0.6^{\prime\prime}$ \citep{mk97}. The orbit of the
58: binary is poorly known, but \citet{jlp02} estimate a long
59: orbital period of $\sim 500$ years, and this orbit suggests that the
60: distance from Mira~A to Mira~B is close to the apparent separation in
61: the plane of the sky, about 70~AU. This interacting binary, being one
62: of the few accretion systems in which the components of the system are
63: resolvable, provides a unique opportunity to study the spectral energy
64: distributions of the components of an interacting binary individually,
65: as well as being a laboratory for testing accretion models.
66:
67: In 1995, the Faint Object Camera (FOC) instrument on the {\em Hubble
68: Space Telescope} (HST) resolved the Mira system for the first time
69: spatially and spectrally at UV and optical wavelengths \citep{mk97}.
70: Since then, the system has been monitored from UV to radio
71: wavelengths. In the past few years, we have been witnessing dramatic
72: changes in Mira~B, especially at UV wavelengths \citep{mk97,bew01}.
73: Spectra from the Space Telescope Imaging
74: Spectrograph (STIS) instrument on HST imply a substantial decrease in
75: accretion luminosity in the late 1990s, compared with fluxes observed at
76: earlier times by the {\em International Ultraviolet Explorer} (IUE)
77: and HST/FOC. Furthermore, a forest of Ly$\alpha$-fluoresced H$_{2}$
78: emission lines dominated the HST spectra of Mira~B in 1999, despite not
79: being seen at all in any previous IUE or HST spectra \citep{bew01,bew02}.
80: Observations from the {\em Far Ultraviolet Spectroscopic
81: Explorer} (FUSE) demonstrated that the low accretion luminosity and
82: prominent H$_{2}$ emission persisted into 2001 \citep{bew04}.
83: On 2003 December 6, {\em Chandra} observations detected an unprecedented
84: soft X-ray outburst from Mira~A \citep{mk05}, further
85: evidence of remarkable dynamicism in the Mira~AB system. The outburst
86: was also detected in contemporaneous H$\alpha$ observations and is
87: likely associated with a magnetic flare followed by a mass ejection or
88: jet-like activity \citep[][Karovska 2006, in prep.]{mk05}.
89:
90: We here present recent UV observations of Mira~B from HST and FUSE,
91: obtained following the X-ray outburst. We compare these data with past
92: observations by these satellites and by IUE. We discuss the long term
93: variations of Mira~B and a possible correlation with long term
94: optical variability.
95:
96: \section{UV SPECTROSCOPY OF MIRA B FROM 1979--2001}
97:
98: Numerous UV spectra of Mira~AB were obtained in 1979--1980 and
99: 1990--1995 by IUE. This data set includes 32 spectra taken in IUE's
100: short-wavelength, low-resolution mode (i.e., SW-LO spectra,
101: covering $\sim 1150-1950$~\AA). These far-UV (FUV) spectra are dominated
102: by broad, hot lines such as C~II $\lambda$1335, Si~IV $\lambda$1400,
103: C~IV $\lambda$1550, Si~III] $\lambda$1892, and C~III] $\lambda$1908,
104: which are produced by material accreting onto Mira~B from
105: Mira~A's massive wind. The few high-resolution spectra of this
106: wavelength region (i.e., SW-HI spectra) are very noisy, but they suggest
107: very broad widths for these lines, consistent with emission from a
108: rapidly rotating accretion disk \citep{dr85}. Unlike
109: IUE, HST can resolve the binary, and UV spectra from HST confirm
110: that the broad lines are from Mira~B, and that Mira~A contributes very
111: little to the IUE spectra below 2000~\AA\ \citep{mk97,bew01}.
112:
113: Only a modest level of UV variability is present within the large IUE
114: SW-LO data set, with line and continuum fluxes varying by about
115: a factor of 2. However, when HST observed Mira~B in 1999 August using
116: STIS, the character of Mira~B's FUV spectrum had changed dramatically.
117: The UV continuum and all of the aforementioned UV emission lines had
118: weakened by at least a factor of 20, and the FUV spectrum was now dominated
119: by many narrow H$_{2}$ lines fluoresced by H~I Ly$\alpha$ \citep{bew01},
120: which are not detectable at all in the IUE spectra.
121: The accretion onto Mira~B results in a warm, fast wind from the star
122: that is most clearly apparent in broad absorption features in the Mg~II
123: h \& k lines at 2800~\AA. Given the presence of accretion, this wind
124: is envisioned to be of a bipolar nature. Analysis of the Mg~II absorption
125: demonstrates that the substantial drop in UV flux from Mira~B seen by HST
126: is accompanied by a corresponding drop in its mass loss rate \citep{bew02}.
127:
128: The low UV fluxes observed by HST in 1999 are most naturally
129: explained by a substantial decrease in the accretion rate onto Mira~B,
130: which also led to a weaker wind from the star. The sudden prominence
131: of the H$_{2}$ lines in the HST spectra suggests that unlike other UV
132: lines, the H~I Ly$\alpha$ line responsible for fluorescing these lines
133: had {\em not} decreased in strength relative to the IUE era. Studying
134: Ly$\alpha$ in IUE spectra is very difficult due to geocoronal
135: contamination, but \citet{bew02} still uncovered evidence to
136: support the contention that Ly$\alpha$ fluxes are similar in the HST
137: and IUE spectra.
138:
139: A 2001 November observation from FUSE found Mira~B to be still in a
140: very low state, with C~III $\lambda$1175 fluxes comparable to those in
141: the 1999 HST/STIS spectrum \citep{bew04}. Like the STIS data,
142: this FUSE spectrum also shows numerous narrow H$_{2}$ lines.
143: Ground-based observations have suggested that Mira~B's optical
144: light curve may have a period of 14 years, with the last published minimum
145: being in 1971 \citep{ahj54,yy77}. \citet{bew04}
146: noted that the the numerous IUE SW-LO observations are all
147: near predicted maxima of this cycle, while the 1999 HST and 2001 FUSE
148: observations are near a predicted minimum. Thus, it is possible that
149: the observed long-term UV variability could be correlated with previously
150: detected optical variations. If the optical periodicity is real and if a
151: connection between the optical and UV light curves truly exists, Mira~B's
152: UV fluxes should increase towards a 2006--2007 maximum predicted by the
153: 14-year optical periodicity. The new 2004 HST and FUSE spectra presented
154: here allow us to test this prediction, and to see
155: if the character of Mira~B's UV spectrum has changed once again.
156:
157: \section{RECENT HST AND FUSE OBSERVATIONS}
158:
159: \subsection{Spectral Line Variability}
160:
161: Mira~B was observed by HST with the STIS instrument on 2004
162: February 16. A near-UV (NUV) spectrum from 2274--3119~\AA\ was taken
163: using the moderate resolution E230M grating, with an exposure time
164: of 600~s, and Mira~B's FUV spectrum from 1140--1729~\AA\ was observed
165: using the E140M grating, with an exposure time of 2100~s. Both spectra were
166: taken through the $0.2^{\prime\prime}\times 0.2^{\prime\prime}$ aperture.
167:
168: Mira was also observed by FUSE on 2004 August 31. This was an
169: 8603~s exposure through the LWRS aperture, covering the
170: 905--1187~\AA\ spectral range. This represents only a fraction of the
171: requested 30~ksec exposure, so the new FUSE spectrum is much noisier
172: than the 2001 FUSE observation \citep{bew04}. A
173: C~III $\lambda$1175 flux can still be measured from the data and some
174: of the strongest H$_{2}$ lines can still be discerned.
175:
176: Figures~1 and 2 compare UV spectral regions observed by HST in 2004
177: with the old 1999 spectra. Five of the Figure~1 panels focus on
178: spectral regions containing FUV emission lines of H~I, O~I, C~II, Si~IV,
179: and C~IV. Some of these lines are blended with H$_{2}$ emission, so the
180: locations of strong, narrow H$_{2}$ lines are also indicated in the figure.
181: The last panel shows the NUV spectrum in the region of the Fe~II UV1
182: multiplet lines. The line locations indicated in Figure~1 are with
183: respect to Mira's rest frame, where the rest wavelengths are taken from
184: \citet{ha93} and \citet{dcm03}, and we assume a radial velocity
185: of 56 km~s$^{-1}$ for the Mira system \citep{pfb88,pp90,ej00}.
186:
187: Figures~1 and 2 show that UV fluxes from Mira~B have increased
188: dramatically since 1999, with the O~I, C~II, Si~IV, C~IV, and Mg~II lines
189: being much stronger than they were. The notable exceptions to this
190: rule, the H~I Ly$\alpha$ line and the H$_{2}$ lines fluoresced by
191: Ly$\alpha$, will be discussed further in \S3.3. The UV continuum is also
192: much stronger in 2004 (see the Fe~II panel in Fig.~1). In Figure~3, we
193: plot measured fluxes as a function of time for many different lines,
194: including measurements from IUE, HST, and FUSE. For
195: C~III $\lambda$1175, O~I $\lambda$1300, C~II $\lambda$1335,
196: Si~IV $\lambda$1400, and C~IV $\lambda$1550, the IUE fluxes are measured
197: from the 32 IUE SW-LO observations mentioned in \S2. Examples
198: of these spectra have been previously presented by \citet{dr85}
199: and \citet{bew01}. Note that the fluxes in Figure~3 are
200: the total flux of these multiplets (e.g., for Si~IV the flux includes
201: both the $\lambda$1393 and $\lambda$1402 lines), given that the
202: individual components of the multiplets are not resolved in the IUE SW-LO
203: spectra.
204:
205: Blended H$_{2}$ lines are removed from
206: the data before measuring fluxes from the HST and FUSE data. This
207: cannot be done for the low resolution IUE spectra, but thanks to the
208: higher line fluxes seen by IUE, H$_{2}$ is not as strong a contaminant
209: in the IUE data, so the correction is not as important \citep{bew01}.
210: The C~III $\lambda$1175 line is the only line observable by
211: all three satellites, though in the HST/STIS spectra and the 2004 FUSE
212: spectrum the C~III line is observed at very low S/N. In the higher quality
213: 2001 FUSE spectrum, the H$_{2}$ lines add
214: $1.6\times 10^{-14}$ ergs~cm$^{-2}$~s$^{-1}$ to the C~III feature
215: \citep{bew04}, so we simply subtract this flux from each C~III
216: flux measurement to correct for the strong H$_{2}$ contamination in C~III.
217:
218: The IUE's inability to resolve the Mira binary is generally not
219: a problem since Mira~A does not contribute significantly to most UV
220: emission lines. However, Mira~A can contribute to Mira~AB's Mg~II lines
221: at certain pulsation phases \citep{bew00}. Fortunately, in
222: long-wavelength, high-resolution (LW-HI) IUE spectra, Mira~A's
223: narrow Mg~II lines are mostly contained within the saturated core of
224: Mira~B's Mg~II wind absorption. Therefore, the Mira~A Mg~II line can
225: be removed before measuring a Mg~II flux for Mira~B. The IUE Mg~II
226: fluxes plotted in Figure~3 are all from LW-HI spectra. We do not use
227: LW-LO spectra because of potential Mira~A contamination. The Mg~II
228: fluxes in Figure~3 include both the h and k lines.
229:
230: Figure~3 shows the dramatic decrease in UV line fluxes
231: observed in 1999--2001. The highest temperature line, C~IV $\lambda$1550,
232: exhibits a particularly large decrease of nearly a factor of 100.
233: The fluxes increase significantly in 2004, but they are still lower
234: than fluxes observed by IUE in 1979--1980 and 1990--1995.
235: As discussed in \S2, a 14-year periodicity
236: has been detected in Mira~B's optical light curve \citep{ahj54,yy77},
237: and dotted lines in Figure~3 indicate the
238: phase and period of this light curve. The increase in UV flux that we
239: see in our 2004 spectra is roughly consistent with the predictions of the
240: optical light curve, assuming that the UV emission follows the same
241: periodic behavior as that inferred for the optical light.
242: It therefore remains a plausible hypothesis that Mira~B's long-term
243: UV variability is associated with the previously detected periodic
244: optical variability. However, truly verifying this connection would
245: require UV observations over at least another cycle.
246:
247: \subsection{Wind Absorption}
248:
249: The blue sides of all the resonance lines in Figures~1 and 2 are
250: suppressed by absorption from Mira~B's wind, with the exception of C~IV
251: (see below). The wind absorption is more readily apparent in the new
252: 2004 data than in the old 1999 spectra. This is in part due to much
253: better S/N in the new data thanks to the higher
254: fluxes, but it is also due to the wind absorption being intrinsically
255: stronger, which is particularly evident for the Fe~II and Mg~II lines.
256: The mass loss rate of Mira~B's wind can be estimated
257: most easily from the Mg~II wind absorption profiles, which have the
258: highest S/N.
259:
260: From analyses of the Mg~II k line at 2796~\AA, we have previously
261: estimated a mass loss rate of
262: $\dot{M}=5\times 10^{-13}$ M$_{\odot}$ yr$^{-1}$ from the 1999 HST spectrum,
263: and $\dot{M}=1\times 10^{-11}$ M$_{\odot}$ yr$^{-1}$ from
264: a typical IUE spectrum \citep{bew02}. Using the exact
265: same methodology, we have fitted a wind
266: absorption profile to the 2004 Mg~II k line in Figure~2. The
267: observed absorption profile is not smooth, suggesting a more complicated
268: and structured wind than we can fit with such a simple model.
269: Nevertheless, the mass loss rate inferred by the fit should still be
270: a reasonable estimate. This fit suggests
271: a mass loss rate of $\dot{M}=2.5\times 10^{-12}$ M$_{\odot}$ yr$^{-1}$
272: and a wind terminal velocity of $V_{\infty}=450$ km~s$^{-1}$.
273: For comparison, the absorption predicted by the wind absorption profile
274: from 1999 is also shown in Figure~2,
275: assuming $\dot{M}=5\times 10^{-13}$ M$_{\odot}$ yr$^{-1}$ and
276: $V_{\infty}=250$ km~s$^{-1}$. The wind absorption is significantly
277: stronger in 2004, meaning that Mira~B's wind has clearly
278: strengthened, but the mass loss rate is still not as high as the
279: $\dot{M}=1\times 10^{-11}$ M$_{\odot}$ yr$^{-1}$ value estimated for
280: the IUE era. The mass loss rate variations are consistent with
281: the accretion rate variations inferred from
282: the UV fluxes (see Fig.~3). Clearly, Mira~B's accretion rate is highly
283: correlated with the strength of its wind.
284:
285: The only atomic lines in Figure~1 that do not definitively possess
286: wind absorption are the C~IV lines. Small decreases in flux blueward of
287: the center of the two C~IV lines may indicate the presence of {\em some}
288: C~IV wind absorption, but this absorption is not nearly as prominent as
289: the saturated absorption seen for the other lines. The presence of
290: strong wind absorption for the Si~IV $\lambda$1392 line, and the
291: relative lack of absorption for C~IV
292: implies a maximum temperature for the wind between $T=6.0\times 10^{4}$~K
293: (Si~IV's line formation temperature) and $T=1.0\times 10^{5}$~K (C~IV's
294: line formation temperature).
295:
296: \subsection{The Anomalous Ly$\alpha$ Line}
297:
298: The one clearly anomalous line in Figure~1 is H~I Ly$\alpha$.
299: Although the 2004 Ly$\alpha$ flux is slightly higher than in 1999, this
300: line does not show the dramatic flux increase that the other emission
301: lines show. The singular lack of variability for this line was also
302: noted by \citet{bew02}, who inferred that the Ly$\alpha$ flux during
303: the IUE era cannot have been much stronger than in 1999, unlike all the
304: other lines. The lack of variability for Ly$\alpha$ means that the H$_{2}$
305: lines that are fluoresced by Ly$\alpha$ are also relatively constant,
306: as shown by the H$_{2}$ lines blended with C~II and Si~IV in Figure~1.
307:
308: It is difficult to understand why the Ly$\alpha$ line behaves
309: so differently from the continuum and all other emission lines, including
310: lines formed at similar temperatures (O~I, C~II, and Mg~II).
311: Does this mean that the Ly$\alpha$ emission is coming from a
312: different place than the other lines? If so, why? Another possibility
313: is that the Ly$\alpha$ flux actually {\em does} vary along with the
314: other lines, but the extremely high opacity of Mira~B's wind
315: to Ly$\alpha$ moderates the variation. Wind absorption explains why
316: the observed Ly$\alpha$ emission is entirely redshifted relative to
317: line center (see Fig.~1). Analogous absorption is also seen for many of
318: the other lines, but wind opacity will be orders of magnitude higher for
319: Ly$\alpha$ since hydrogen is much more abundant than other atomic species.
320:
321: Given that increases in UV flux are clearly accompanied by increases
322: in wind strength, perhaps the enhanced wind obscuration largely cancels the
323: flux increase for Ly$\alpha$. This hypothesis would require that Ly$\alpha$
324: photons are actually destroyed rather than just scattered by the wind.
325: Perhaps this destruction could be accomplished by scattering into Mira~B's
326: accretion disk, or extinction by dust entrained within Mira~B's wind.
327:
328: \section{CAUSES OF LONG-TERM VARIABILITY}
329:
330: The fundamental cause of Mira~B's substantial variability
331: remains uncertain. Variable mass loss from Mira~A could have caused
332: the changes in the UV flux of Mira~B observed in the past 10 years,
333: signaling that the system is undergoing important transformations.
334: These transformations may have set the stage for the remarkable X-ray
335: outburst from Mira~A that was observed by {\em Chandra} on
336: 2003 December 6 \citep{mk05}.
337: %This outburst was not present in an XMM observation from
338: %2003 July 23 \citep{jhk04}, and had faded by the time of a
339: %follow-up {\em Chandra} observation on 2004 January 11.
340:
341: If the variability is associated with the possible 14-year optical
342: periodicity, the reason for the periodicity is a mystery. The UV lines
343: in particular are believed to originate from accretion processes on Mira~B,
344: so a periodicity in these lines would suggest a periodicity for the accretion
345: rate. Accretion instabilities could easily lead to variability, but it
346: is not clear why they would be periodic. Perhaps the gravitational
347: influence of an undetected stellar or planetary companion could lead to
348: a corresponding periodicity in the accretion rate, perhaps through
349: wind focusing effects. Hints of a third
350: body in the Mira~AB system have been previously reported in the literature
351: \citep{pb80,mk92,mk93}. Since Mira~A's wind is the source of the
352: accreting material, a periodicity in Mira~A's mass loss rate could in
353: principle also lead to periodicity in the accretion rate.
354: No such periodicity has yet been found in the light curve of Mira
355: compiled by AAVSO, but searches for very low amplitude, long-term
356: periodicity are ongoing (M.\ Templeton, private communication).
357:
358: We hope to explore these possibilities further in the future,
359: as we reach the expected maximum in 2006--2007 and over the next
360: 14-year cycle, using different observational techniques including
361: optical/IR interferometry. However, future UV high resolution spectra
362: are very much in doubt due to the failure of the STIS instrument in
363: 2004 August, and the reaction wheel and gyro failures on FUSE that
364: currently limit observations to high declinations. Thus, new UV
365: instrumentation may be necessary to continue spectroscopic observations
366: in the ultraviolet.
367:
368: \acknowledgments
369:
370: This work was supported by NASA grant NAS8-39073 to the
371: Smithsonian Astrophysical Observatory and NASA grant NNG05GC14G to
372: the University of Colorado.
373: MK is a member of the Chandra X-ray Center, which is operated
374: by the Smithsonian Astrophysical Observatory under contract to
375: NASA NAS8-39073.
376:
377: \clearpage
378:
379: \begin{thebibliography}{}
380:
381: \bibitem[Abgrall et al.(1993)]{ha93}
382: Abgrall, H. A., Roueff, E., Launay, F., Roncin, J. -Y., \& Subtil, J. -L.
383: 1993, A\&AS, 101, 273
384: \bibitem[Baize(1980)]{pb80}
385: Baize, P. 1980, A\&AS, 39, 86
386: \bibitem[Bowers \& Knapp(1988)]{pfb88}
387: Bowers, P. F., \& Knapp, G. R. 1988, ApJ, 332, 299
388: \bibitem[Josselin et al.(2000)]{ej00}
389: Josselin, E., Mauron, N., Planesas, P., \& Bachiller, R. 2000, A\&A, 362, 255
390: \bibitem[Joy(1954)]{ahj54}
391: Joy, A. H. 1954, ApJS, 1, 39
392: \bibitem[Jura \& Helfand(1984)]{mj84}
393: Jura, M., \& Helfand, D. J. 1984, ApJ, 287, 785
394: \bibitem[Karovska(1992)]{mk92}
395: Karovska, M. 1992, in Complementary Approaches to Double and Multiple Star
396: Research, ed. H. A. McAlister \& W. I. Hartkopf (ASP Conf. Ser., Vol. 32),
397: 558
398: \bibitem[Karovska et al.(1997)]{mk97}
399: Karovska, M., Hack, W., Raymond, J., \& Guinan, E. 1997, ApJ, 482, L175
400: \bibitem[Karovska et al.(1993)]{mk93}
401: Karovska, M., Nisenson, P. \& Beletic, J. 1993, ApJ, 402, 311
402: \bibitem[Karovska et al.(2005)]{mk05}
403: Karovska, M., Schlegel, E., Hack, W., Raymond, J. C., \& Wood, B. E. 2005,
404: ApJ, 623, L137
405: %\bibitem[Kastner \& Soker(2004)]{jhk04}
406: %Kastner, J. H., \& Soker, N. 2004, ApJ, 616, 1188
407: \bibitem[Matthews \& Karovska(2006)]{ldm06}
408: Matthews, L. D., \& Karovska, M. 2006, ApJ, 637, L49
409: \bibitem[Morton(2003)]{dcm03}
410: Morton, D. C. 2003, ApJS, 149, 205
411: \bibitem[Perryman et al.(1997)]{macp97}
412: Perryman, M. A. C., et al. 1997, A\&A, 323, L49
413: \bibitem[Planesas et al.(1990)]{pp90}
414: Planesas, P., Bachiller, R., Martin-Pintado, J., \& Bujarrabal, V. 1990,
415: ApJ, 351, 263
416: \bibitem[Prieur et al.(2002)]{jlp02}
417: Prieur, J. L., Aristidi, E., Lopez, B., Scardia, M., Mignard, F., \&
418: Carbillet, M. 2002, ApJS, 139, 249
419: \bibitem[Reimers \& Cassatella(1985)]{dr85}
420: Reimers, D., \& Cassatella, A. 1985, ApJ, 297, 275
421: \bibitem[Wood \& Karovska(2000)]{bew00}
422: Wood, B. E., \& Karovska, M. 2000, ApJ, 535, 304
423: \bibitem[Wood \& Karovska(2004)]{bew04}
424: Wood, B. E., \& Karovska, M. 2004, ApJ, 601, 502
425: \bibitem[Wood et al.(2001)]{bew01}
426: Wood, B. E., Karovska, M., \& Hack, W. 2001, ApJ, 556, L51
427: \bibitem[Wood et al.(2002)]{bew02}
428: Wood, B. E., Karovska, M., \& Raymond, J. C. 2002, ApJ, 575, 1057
429: \bibitem[Yamashita \& Maehara(1977)]{yy77}
430: Yamashita, Y., \& Maehara, H. 1977, PASJ, 29, 319
431:
432: \end{thebibliography}
433:
434: \clearpage
435:
436: \begin{figure}
437: \plotone{f1.eps}
438: \caption{A comparison of UV spectra of Mira~B taken in 1999 (dotted line)
439: and 2004 (solid line). The spectra are smoothed slightly for the sake
440: of appearance. Dashed lines indicate the rest wavelengths
441: (in Mira's rest frame) of the line or lines indicated by the panel
442: labels (e.g., H~I Ly$\alpha$, O~I, C~II, Si~IV, C~IV, and Fe~II).
443: Dot-dashed lines indicate the locations of strong, narrow H$_{2}$
444: lines.}
445: \end{figure}
446:
447: \clearpage
448:
449: \begin{figure}
450: \plotone{f2.eps}
451: \caption{A comparison of Mira~B's Mg~II k line profile from 1999 (dotted
452: histogram) and 2004 (solid histogram), plotted on a velocity scale
453: in Mira~B's rest frame. A wind absorption profile has
454: been fitted to the 2004 data (thick solid line), which suggests a mass
455: loss rate of $\dot{M}=2.5\times 10^{-12}$ M$_{\odot}$ yr$^{-1}$ and a
456: terminal velocity of $V_{\infty}=450$ km~s$^{-1}$. Also shown is
457: the absorption predicted by a fit to the 1999 wind absorption
458: (dot-dashed line), assuming
459: $\dot{M}=5\times 10^{-13}$ M$_{\odot}$ yr$^{-1}$ and
460: $V_{\infty}=250$ km~s$^{-1}$ (Wood \& Karovska 2004). This greatly
461: underpredicts the amount of absorption observed in 2004, illustrating
462: the increase in wind strength from 1999 to 2004.}
463: \end{figure}
464:
465: \clearpage
466:
467: \begin{figure}
468: \plotone{f3.eps}
469: \caption{Fluxes of various UV emission lines as a function of time, based
470: on spectra from IUE (triangles), HST (boxes), and FUSE (diamonds).
471: The fluxes are normalized to the average value observed by IUE.
472: The lines and normalization factors (in units of
473: ergs~cm$^{-2}$~s$^{-1}$) are
474: C~III $\lambda$1175: $4.1\times 10^{-13}$,
475: O~I $\lambda$1300: $8.1\times 10^{-13}$,
476: C~II $\lambda$1335: $6.5\times 10^{-13}$,
477: Si~IV $\lambda$1400: $5.5\times 10^{-13}$,
478: C~IV $\lambda$1550: $1.7\times 10^{-12}$, and
479: Mg~II $\lambda$2800: $1.3\times 10^{-11}$.
480: The dotted lines are schematic representations of the 14 year
481: periodicity of Mira~B suggested by optical observations (Joy 1954;
482: Yamashita \& Maehara 1977).}
483: \end{figure}
484:
485: \end{document}
486:
487:
488: