1: %\documentclass[manuscript]{aastex}

2: \documentclass[12pt,preprint]{aastex}

3: %\documentclass[11pt,preprint]{aastex}

4: 

5: %\documentclass{emulateapj} 

6: %\usepackage{apjfonts}

7: 

8: \newcommand\mdot{\rm \dot{M}}

9: \newcommand\msun{\rm M_{\odot}}

10: \newcommand\rsun{\rm R_{\odot}}

11: \newcommand\lsun{\rm L_{\odot}}

12: \newcommand\msunyr{\rm M_{\odot}\,yr^{-1}}

13: \newcommand\be{\begin{equation}}

14: \newcommand\en{\end{equation}}

15: \newcommand\cm{\rm cm}

16: 

17: \newcommand\kms{\rm{\, km \, s^{-1}}}

18: \newcommand\K{\rm K}

19: \newcommand\etal{{\rm et al}.\ }

20: \shorttitle{GM Cep}

21: \shortauthors{Sicilia-Aguilar et al.}

22: 

23: \begin{document}

24: 

25: \title{The Rapid Outbursting Star GM Cep: An EX-or in Tr 37?}

26: 

27: \author{Aurora Sicilia-Aguilar\altaffilmark{1}, Bruno Mer\'{\i}n\altaffilmark{2}, Felix Hormuth\altaffilmark{1}, P\'{e}ter \'{A}brah\'{a}m\altaffilmark{3},}

28: \author{Thomas Henning\altaffilmark{1},  M\'{a}ria Kun\altaffilmark{3}, Nimesh Patel\altaffilmark{4}, Attila Juh\'{a}sz\altaffilmark{1}, Wolfgang Brandner\altaffilmark{1},}

29: \author{Lee W. Hartmann\altaffilmark{5}, Szil\'{a}rd Csizmadia\altaffilmark{3}, Attila Mo\'{o}r\altaffilmark{3}}

30: 

31: \altaffiltext{1}{Max-Planck-Institut f\"{u}r Astronomie, K\"{o}nigstuhl 17, 69117 Heidelberg, Germany}

32: \altaffiltext{2}{Research and Scientific Support Department, European Space Agency, ESTEC, Keplerlaan 1, 2200 AG Noordwijk, The Netherlands}

33: \altaffiltext{3}{Konkoly Observatory, H-1525 Budapest, P.O.Box 67, Hungary}

34: \altaffiltext{4}{Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138}

35: \altaffiltext{5}{University of Michigan, 830 Dennison 500 Church St., Ann Arbor, MI 48109}

36: 

37: \email{sicilia@mpia.de}

38: 

39: \begin{abstract}

40: 

41: We present optical, IR and millimeter observations of the solar-type star 13-277, 

42: also known as GM Cep, in the 4 Myr-old cluster Tr 37. GM Cep experiences rapid

43: magnitude variations of more than 2 mag at optical wavelengths. We explore the 

44: causes of the variability, which seem to be dominated by strong increases in the 

45: accretion, being similar to EX-or episodes. The star shows high, variable accretion 

46: rates (up to $\sim$10$^{-6}$ M$_\odot$/yr), signs of powerful winds, 

47: and it is a very fast rotator (V$sini \sim$43 km/s). Its strong mid-IR excesses reveal

48: a very flared disk and/or a remnant envelope, most likely out of hydrostatic 

49: equilibrium. The 1.3 millimeter fluxes suggest a relatively  massive disk

50: (M$_D \sim$0.1 M$_\odot$). Nevertheless, the millimeter mass is not enough to sustain

51: increased accretion episodes over large timescales, unless the mass is underestimated 

52: due to significant grain growth. We finally explore the possibility of GM Cep having

53: a binary companion, which could trigger disk instabilities producing the enhanced 

54: accretion episodes.

55: 

56: \end{abstract}

57: 

58: \keywords{stars: variable --- accretion disks ---  stars: pre-main sequence --- stars:individual:GM Cep (13-277)}

59: 

60: \section{Introduction \label{intro}}

61: 

62: The open cluster Tr 37 is one of the best studied intermediate-aged 

63: young star formation regions. Aged $\sim$ 4 Myr (Sicilia-Aguilar et al.

64: 2004, 2005; hereafter Paper I, Paper II), and located at 900 pc distance 

65: (Contreras et al. 2002), it contains a rich population of T Tauri stars 

66: (TTS; $\sim$180 members with spectral types G to M2) of which about 48\% 

67: still show IR excesses consistent with protoplanetary disks at different 

68: evolutionary stages (Sicilia-Aguilar et al. 2006a, b; hereafter Paper III, 

69: Paper IV). While most of the stars in Tr 37 show important evidence of

70: disk evolution, with lower accretion rates and near-IR excesses than in 

71: younger regions, a few objects still display characteristics of much 

72: younger systems. The most remarkable one is the solar-type star 13-277, 

73: also called GM Cep (Morgenroth 1939). Because of its spatial location within 

74: the cluster, it belongs most likely to the Tr 37 main population, with 

75: average ages $\sim$4 Myr ($\sim$85\% of the population is older than 2 Myr, 

76: and $\sim$95\% is older than 1 Myr), rather than to the young population 

77: associated to the Tr 37 globule, aged $\sim$1 Myr (Paper II). Its continuum 

78: spectrum and the strong and broad H$\alpha$ emission suggested an accretion 

79: rate \.{M}$\sim$3 10$^{-7}$ M$_\odot$/yr, about 2 orders of magnitude over 

80: the median accretion rate of TTS in Tr 37 (Paper IV). Its bolometric 

81: luminosity  (L$\sim$26 L$_\odot$ in 2000) is about 1 order of magnitude higher 

82: than other stars with similar late G-early K spectral type. Finally, we notice 

83: an unusually high mid-IR flux for a solar-type star: its MIPS flux at 70 $\mu$m 

84: is comparable to that of the Tr 37 Herbig Be star MVA-426, being one of the only 

85: four cluster members detected at this wavelength (Paper III). 

86: 

87: All observations suggest that GM Cep is a variable star of the EX-or 

88: type, with an unstable disk and variable accretion rate, which is

89: remarkable in an ``old'' cluster like Tr 37, where disk evolution seems

90: ubiquitous. FU-ors and EX-or objects have been suggested to be either normal stages 

91: within the very early TTS evolution (see Hartmann \& Kenyon 1996 for a review) or 

92: a special type of young object, probably binary (Herbig 2003 and references therein). 

93: Here we present the results of our photometry monitoring campaign during 

94: 2006-2007\footnote{Based on observations collected at the German-Spanish Astronomical 

95: Center, Calar Alto, jointly operated by the Max-Planck-Institut f\"{u}r Astronomie 

96: Heidelberg and the  Instituto de Astrof\'{\i}sica de Andaluc\'{\i}a (CSIC)}, lucky 

97: imaging, millimeter continuum, and $^{12}$CO data, together with optical spectra 

98: (high- and low-resolution), and a compilation of the data available from the 

99: literature. The multiwavelength data are described in Section \ref{observations}. In 

100: Section \ref{analysis} we explore the characteristics of the star: optical and IR 

101: variability, spectral type, accretion, winds, disk mass, and potential companions. 

102: Section \ref{outburst} discusses the possible causes of variability and the comparison 

103: with similar stars, and in Section \ref{conclu} we summarize our results.

104: 

105: \section{Observations and Data Reduction \label{observations}}

106: 

107: 

108: \subsection{Data from the Literature \label{data-literature}}

109: 

110: The photometric data and epochs available for GM Cep are given in Table \ref{obs-table}. 

111: The first reference to GM Cep, Morgenroth (1939), classifies it as a ``long-period 

112: variable'' with visual magnitudes 13.5-15.5, although no epoch nor period are mentioned. 

113: Suyarkova (1975) studied the light variations over several months,

114: detecting rapid variations (days to weeks) between 14.2 and 16.4 mag, mixed 

115: with stability periods up to $\sim$100 days at both the lower and the upper 

116: magnitude. Kun (1986) lists the star among the variables in Cep OB2, providing V 

117: photometry for three epochs. The databases Vizier, Simbad,

118: and SuperCosmos list optical photometric data from the USNO-A2.0, USNO-B1.0, 

119: GSC2.2, and SuperCosmos catalogs, plus IR data from IRAS and MSX6C

120: (Egan et al. 2003). There are two mentions of GM Cep in the amateur database of 

121: the American Association of Variable Star Observers (AAVSO), but they do not add 

122: any extra information to our data. Our previous studies of Tr 37 include 

123: optical photometry taken with the 1.2m telescope at the Fred Lawrence Whipple 

124: Observatory (FLWO) between 2000 and 2004 (Paper I, II), low resolution optical 

125: spectroscopy taken with the FAST spectrograph on the 1.5m telescope at the FLWO 

126: (Paper I), high-resolution optical spectroscopy from Hectochelle/MMT at the FLWO 

127: (Paper IV), and IRAC and MIPS data from our Spitzer study (Paper III). We also 

128: include the near-IR data from 2MASS, presented in Paper I. All these 

129: data confirmed the variability of GM Cep and the presence of strong IR excesses 

130: from a very luminous disk. 

131: 

132: In order to compare the data from the different studies in Table \ref{obs-table}, 

133: we transformed the photographic and CCD data from different filters into the 

134: UBVR$_c$I$_c$ system, which is the closest match to the data prior to $\sim$1986. 

135: The photographic R63F and R59F red data are comparable to the modern R$_c$ filter. 

136: R$_c$-R63F=0.096 and 0.081 mag, for giants and dwarfs, respectively, and 

137: R$_c$-R59F=-0.031 (Bessell 1986). Following Blair \& Gilmore (1982), R$_c$=R63F if 

138: R$_c$-I$_c <$0.9 mag (for GM Cep, R$_c$-I$_c$ $\sim$0.8 mag). The GG395-B$_j$ 

139: photographic passband is similar to the Johnsons B filter, although the zero point 

140: is different (Bessell et al. 1986). The errors involved in these conversions are 

141: less than 0.1 mag, below the typical errors from the photographic plates 

142: and well below the observed variations of GM Cep. For the modern observations, 

143: when contemporary V magnitude is available, R$_J$ and I$_J$ can be transformed 

144: into R$_c$ and I$_c$ for comparison (Bessel 1979; Fernie 1983) with errors under 

145: 0.01 mag for R, and around 0.1-0.2 mag for I. 

146:  

147: 

148: 

149: \subsection{Optical Photometry \label{data-optical}}

150: 

151: The first set of optical observations was taken with the KING 70 cm telescope of 

152: the Max-Planck-Institut f\"{u}r Astronomie, at the K\"{o}nigstuhl in Heidelberg

153: (see Table \ref{obs-table}). The telescope has a field of view of 

154: $\sim$18'x18', and is equipped with a Loral CCD L3-W17 detector and a set of 

155: standard UBVR$_J$I$_J$ filters. The filters have very small color transform 

156: coefficients, and the second order extinction is almost zero, so relative 

157: photometry is feasible. Exposure times ranged from 30s to 3 x 120s depending on 

158: filter and weather conditions. We also obtained VR$_J$I$_J$ photometry  

159: (Table \ref{obs-table})with the CAFOS imager on the 2.2m telescope in Calar 

160: Alto (Director's Discretionary Time, DDT), which covers a field of view of diameter 16'. 

161: Taking 3 x 10 s or 3 x 30s exposures, we reach similar magnitude and area coverage 

162: as the K\"{o}nigstuhl observations. 

163: 

164: The data were reduced following standard routines within the IRAF\footnote{IRAF 

165: is distributed by the National Optical Astronomy Observatories,

166: which are operated by the Association of Universities for Research

167: in Astronomy, Inc., under cooperative agreement with the National

168: Science Foundation.} tasks \textit{noao.imred.ccdred} to do the bias and flat 

169: field corrections. Aperture photometry was performed within IRAF 

170: \textit{noao.digiphot.apphot}, and the data in each band were calibrated via 

171: relative photometry, comparing with our previous observations from 

172: the FLWO (Paper I). Depending on the night and instrument, $\sim$150-800 stars 

173: were used for calibrating VRI, and $\sim$20-200 for B and U. Since the FLWO 

174: observations used the UVR$_c$I$_c$ filters, the R$_c$ and I$_c$ bands were 

175: transformed into the Johnsons system (Fernie 1983; Bessel 1979). Taking

176: into account the large number of stars involved in the calibration, the errors 

177: derived from band transformation are minimal and the final error is dominated by 

178: the signal to noise, being typically less than 0.01 mag. Due to the lack of 

179: calibrated B data in the FLWO survey, we used the B photometry from the literature 

180: for the relative photometry. About 150 stars with B band photometry are listed by

181: SIMBAD, mostly bright objects from the Marshall \& Van Altena (1987) catalog.

182: Only 15 of these stars are not saturated in the K\"{o}nigstuhl frames and could

183: be used for calibration. Whereas this is insufficient for 

184: studying colors, the relative changes between different nights are unaffected. 

185: 

186: \subsection{Optical Spectroscopy: HIRES/Keck and CAFOS/2.2m\label{data-spec}}

187: 

188: GM Cep was observed with the high-resolution spectrograph HIRES mounted on Keck I on 

189: 2001 June 30. The spectra were taken before the HIRES upgrades in 2004, using the 

190: rg610 filter and a 1.148'' slit, which provides a wavelength coverage between 6350 

191: and 8750 \AA\ , with some gaps in between orders, and a resolution R$\sim$ 34,000.

192: The resolution is similar to that of Hectochelle/MMT (Sicilia-Aguilar 

193: et al. 2005a; Paper IV), but the wavelength coverage is much larger, including

194: the H$\alpha$ and Li I 6708\AA\ regions, several forbidden lines ([O I] at 6368 \AA, 

195: [S II] at 6717 and 6731 \AA, [N II] at 6548 and 6583 \AA), the Ca II IR lines at 8498 

196: and 8662 \AA, and numerous photospheric absorption lines. The exposure time was 3 x 900 s. 

197: The data was reduced using the MAKEE Keck Observatory HIRES data reduction software, 

198: developed by T. Barlow. The wavelength calibration was done with a ThAr lamp, being 

199: accurate up to 1-2 km/s. No flux calibration was required.

200: 

201: A low-resolution spectrum was taken with the CAFOS camera in the long-slit spectroscopic 

202: mode on the 2.2m telescope  as a part of our Calar Alto DDT program. We used the 

203: B-100, G-100, and R-100 grisms, resulting in a wavelength coverage from 3200 to 9000 \AA\ 

204: with no gaps and some overlap, and a resolution of 2 \AA\ per pixel (Figure \ref{lowrespec-fig}). 

205: The spectra were reduced using the standard IRAF tasks (\textit{noao.imred.ccdred} 

206: and \textit{specred.twodspec}), and a wavelength solution was derived using a HgHeRb lamp. 

207: A total of 3 x 600 s exposures were taken with each grism, in order to achieve a good 

208: cosmic ray removal. In addition, we obtained two CAFOS acquisition images in 

209: V and R$_J$, with exposure times of 60s. These two images, although lower in quality and 

210: with a smaller field of view than the standard CAFOS imaging, were reduced and calibrated 

211: using the techniques from Section \ref{data-optical}, to provide simultaneous photometry.

212: 

213: 

214: 

215: \subsection{Lucky Imaging with AstraLux \label{data-astra}}

216: 

217: In order to look for wide- and intermediate-distance companions, we observed

218: GM Cep with the high spatial resolution Lucky Imaging camera AstraLux, operating 

219: at the 2.2m telescope in Calar Alto. AstraLux is a high-speed camera

220: with an electron multiplying CCD, which allows full frame rates of up to 

221: 34\,Hz with virtually zero readout noise. It has a field of view  of

222: 24''$\times$24'' at a pixel scale of 47\,mas/px.

223: The AstraLux data consist typically of several thousands single frames with 

224: integration times between 15 and 100\,ms. Software post-processing selects 

225: the frames which are least affected by distortions due to atmospheric turbulence, 

226: by measuring the Strehl ratio in each single frame. The 

227: high-quality frames ($\sim$ 1--5\% of all images) are combined to a final result 

228: with improved angular resolution and Strehl ratio (Lucky Imaging technique; 

229: Tubbs et al. 2002; Law et al. 2006). Under good seeing conditions and with the 

230: SDSS z' filter, AstraLux provides spatial resolution better than 100\,mas and 

231: Strehl ratios of up to 25\%, depending on the magnitude. Observations of GM Cep 

232: were conducted in July and November 2006, and May and June 2007 

233: (see Tables \ref{obs-table} and \ref{astralux-table} for details). 

234: Though the November 2006 data is slightly affected by atmospheric  

235: dispersion, leading to an elongated PSF, the superb seeing of 0.6''

236: still resulted in a Strehl ratio of 14\%. Figure \ref{astralux-fig} shows

237: the November 2006 data together with a simulated theoretical PSF,

238: including atmospheric dispersion effects. We present a detailed analysis of 

239: these data in Section \ref{companion}. 

240: 

241: \subsection{JHK Photometry \label{data-JHK}}

242: 

243: Near-IR photometry was obtained with the IR camera CAIN-2 at the 

244: 1.52m Carlos S\'{a}nchez telescope in the Teide Observatory (Canary Islands)

245: during 5 nights in early June 2007 (see Table \ref{obs-table}). Observations 

246: in the J H and Ks bands consist of 5 dithered exposures around GM Cep to ensure 

247: proper sky image subtraction. For each dither point we took several frames 

248: using short exposure time to avoid saturation, which were later averaged to 

249: obtain the final reduced image. The minimum exposure time was 1~s for individual 

250: exposures and the total integration time per filter was about 6 min. The data were 

251: reduced using the package \textit{caindr} developed within IRAF by 

252: J.A. Acosta-Pulido. The processing included sky subtraction, flatfielding 

253: and combination of the dither positions. Aperture photometry was done, and the 

254: relative photometry calibration was based on 5 to 13 nearby IR sources taken from 

255: the 2MASS catalog. The errors derived from this procedure are dominated by the 

256: background and CCD read noise, ranging from 0.1 to 0.01 mag depending on weather 

257: conditions. The star remained stable during the whole run (see Table \ref{JHK-table}), 

258: taking into account the errors, and its magnitude was similar to the  

259: 2MASS 2001 data. Similar stability over a few days had been observed in the optical data

260: taken in September 2000. Simultaneous optical observations cover only 1 of the nights, 

261: during which the star was near maximum. 

262: 

263: 

264: \subsection{IRAM 1.3 Millimeter Continuum \label{data-mm}}

265:  

266: We observed GM Cep at the IRAM 30m telescope in Pico Veleta using the 37 channel 

267: bolometer, MAMBO-1. The observations were done during night time, at 4:18h UT on 

268: June 2nd 2006, and 2:58h on June 3rd. The opacity at zenith was Tau=0.4.

269: Each observation consisted of a 20 min integration standard \textit{onoff}, 

270: with 10 min integration on source. The wobbler throw for the \textit{off} 

271: position was 32'' and 50'', respectively, so the \textit{off} measurement falls 

272: in different locations, to avoid the potential contamination from other sources. 

273: The data were reduced using the MOPSIC software, developed by R. Zylka for the 30m 

274: bolometer data. The pointing and flux calibration were done with the sources 

275: LK H$\alpha$ 234, Cep A, and NGC 7538 from the IRAM Pool catalog, and the planet 

276: Uranus. The final flux at 1.3mm is 13.9 $\pm$ 1.0 mJy. The 1.3 millimeter continuum 

277: observations are analyzed in Section \ref{diskmodels}. 

278:  

279: \subsection{IRAM $^{12}$CO(1-0) and $^{12}$CO(2-1) Observations \label{data-co}}

280:  

281: We also observed GM Cep with the heterodyne receivers at the 30m IRAM 

282: telescope. The observations were performed under fair weather conditions 

283: on June 3rd (UT 7:00-8:15), June 4th (UT 7:30-9:20), and June 5th 

284: (UT 8:00-10:00). We used the standard \textit{onoff} mode with wobbler shift, 

285: that results in better baselines, with wobbler throws from 70'' to 120''. 

286: We used the highest resolution mode available (2 km/s), 

287: and the backends A100, B100 (for frequencies around 115 GHz)

288: and A230, B230 (for frequencies around 230 GHz). The instrument allows to 

289: observe two frequencies simultaneously, which we tuned to the $^{12}$CO(1-0) 

290: and $^{12}$CO(2-1) transitions. Since the LRS velocity of the source (considering 

291: the average cluster CZ=-15.0$\pm$3.6 km/s) at the time of the observation

292: was around 0 km/s, the receivers were tuned to 115.271-115.275 GHz 

293: and 230.538-230.543 GHz, varying the central frequencies during different 

294: integrations to avoid that the line falls always on the same place 

295: on the detector. The total integration time was 4 h, with 

296: 2h on source. For pointing and calibration, we observed the standard

297: sources J2253+161, NGC 7027, and Cep A, from the IRAM Pool catalog.

298:  

299: The data were reduced using the CLASS software within the GILDAS package, 

300: specially developed for the reduction of IRAM heterodyne data. The single 

301: scans were examined to remove the bad ones, matched in frequency, and added. 

302: The baseline was fitted as a 1-order polynomial and subtracted. The reduced 

303: data is plotted in Figure \ref{lines-fig}. The $^{12}$CO(1-0) is identified

304: as a 5$\sigma$ detection with peak value 0.3mJy, but the weaker $^{12}$CO(2-1) 

305: transition is not detected. No line broadening was observed for the 

306: $^{12}$CO(1-0), which may be an effect of the weakness of the detection, 

307: and/or a consequence of the inclination of the object. In addition to the 

308: 5$\sigma$ emission, we find a 4$\sigma$ absorption around the line. This 

309: absorption does not seem to originate in a single point in the off position, 

310: given the changes in the wobbler throw and its position angles. A possible 

311: explanation would be the presence of substantial amounts of gas in the Tr 37 bubble. 

312: According to the work of Patel et al. (1998) using the 14m FCRAO antenna, the region 

313: around GM Cep is relatively clean (Figure \ref{lines-fig} right). Nevertheless, the

314: sensitivity of the 30m is more than double than the FCRAO, so the absorption could 

315: be due to gas emission in the expanding IC 1396 H II region barely detected in the 

316: FCRAO maps. The $^{12}$CO lines are weaker than expected in typical disks for the

317: same 1.3 mm continuum emission, which could be due to oversubtraction of ambient 

318: $^{12}$CO emission, and is difficult to correct without deep mapping of the area. 

319: 

320: 

321: \section{Analysis \label{analysis}}

322:  

323: \subsection{The Lightcurve: Outburst History \label{lightcurve}}

324: 

325: The lightcurve of GM Cep is depicted in Figure \ref{lightcurve-fig}. The star 

326: suffers rapid and repeated irregular variations in all bands, with amplitudes 

327: 2-2.5 mags. It increased in brightness after $\sim$1986, reached a low magnitude 

328: level in $\sim$2002, and raised between $\sim$2003 and $\sim$July 2006. From 

329: July 2006 on, when our time sampling is better, GM Cep experienced a decrease 

330: in magnitude, followed by a 1 mag increase in January 2007, a further decline by 

331: April 2007, increasing again to near-maximum levels by June 2007. 

332: The changes in magnitude can be very rapid ($>$0.1 mag in consecutive days in 

333: 2000 and 2001; 2 mag from Dec. 15 to  Dec. 29, 2006; 1 mag from 

334: Jan. 2 to Jan. 18, 2007). This suggests that GM Cep 

335: probably suffered similar rapid oscillations during the undersampled periods.

336: 

337: The simultaneous multiband data show that the color changes 

338: are small and relatively random (Figure \ref{colorJ-fig}). The changes in color do 

339: not seem  purely related to extinction variations, considering a typical interstellar 

340: extinction law (Cardelli et al. 1989), although the material surrounding a young star is 

341: likely to be different from typical ISM matter. If the changes in magnitude were due to 

342: extinction alone, we would need to assume significant grain growth and a non-standard 

343: extinction law: To produce the same amplitude of variations in all VRI bands, the obscuring 

344: grains should be very large ($>$10$\mu$m) to ensure grey extinction (Eiroa et al. 2002). 

345: Simultaneous JHK and optical observations would help to constraint the

346: effects of extinction in the lightcurve, but unfortunately only one of our JHK data

347: sets has simultaneous UBVRI photometry. The small color variations suggest as well that 

348: eclipses by a companion of different mass are not responsible for the changes. The amplitude 

349: of the variations (up to 2.5 magnitudes) is too large for eclipses, even if we consider 

350: an equal-mass companion, unless we consider some special binary configuration in addition 

351: to obscuring material, somehow analogous to KH 15D (Hamilton et al. 2005 and references therein). 

352: Detailed studies of the periodicity  in the future should clarify the presence of eclipses. 

353: Nevertheless, both the extinction and the eclipse scenarios fail to explain the high luminosity

354: of GM Cep (up to 30-40 L$_\odot$ at maximum) unless the star had a much earlier

355: spectral type (A) or belonged to another luminosity class.

356: 

357: \subsection{Stellar Properties and Activity \label{spectra}}

358: 

359: The spectra of GM Cep are dominated by very strong and broad H$\alpha$ emission 

360: with strong P-Cygni profile with a deep blueshifted absorption that goes under the 

361: continuum level, characteristic of very strong accretion (Edwards et al. 1994; 

362: Hartigan et al. 1995). Other emission lines associated to winds and accretion 

363: are present as well (Figures \ref{lowrespec-fig}, \ref{lines1-fig}, and \ref{lines2-fig}). 

364: The main lines are listed in Table \ref{lines-table}. Given that none of the spectra 

365: is flux calibrated, it is not possible to determine the variations in the continuum level. 

366: The spectral features in the high-resolution data from 2001 and 2004 are very similar. 

367: The H$\alpha$ equivalent width (EW) was larger in the 2004 (EW=-14 \AA) than in 2001 

368: (EW=-6 \AA), and the 10\% H$\alpha$ velocity width was also higher in 2004 (660 km/s; 

369: Paper IV) than in 2001 (580 km/s), suggesting a higher accretion rate in the first case. 

370: The narrow H$\alpha$ feature in the HIRES spectrum may be affected by nebular emission 

371: of the H II region, or may have been oversubtracted in the Hectochelle spectrum. The 

372: low-resolution spectra give H$\alpha$ EW -10 and -19 \AA\ in June 2001 and April 2007, 

373: respectively, suggesting H$\alpha$ variability caused by variable accretion/winds added to

374: changes in the continuum level. 

375: 

376: In the 2001 low-resolution FAST spectrum, the H$\beta$ line appears in absorption.

377: Typically, accreting stars show the Balmer H Series in emission, with similar 

378: profiles in all the lines, suggesting that they came from the same mass of emitting gas

379: (Muzerolle et al. 1998, 2001). Nevertheless, in the case of very strong

380: accretors, the strong winds can dominate the most optically thick lines

381: (H$\alpha$) overrunning the signs of magnetospheric accretion, which would

382: dominate optically thinner lines (like H$\beta$; Muzerolle et al. 2001). 

383: Intermediate-mass variables of the UX-or class tend to have H$\beta$ in absorption

384: (Grinin et al. 2001; Tambovtseva et al. 2001), but H$\alpha$ P-Cygni profiles

385: are very rare in them, and the Ca II IR lines appear typically in absorption.

386: For GM Cep, the Ca II IR triplet ($\lambda \lambda$ 8498, 8542, 8662 \AA), indicator 

387: of accretion, shows strong emission in both the HIRES and the CAFOS spectra 

388: (Figures \ref{lowrespec-fig} and \ref{lines1-fig}). The lines were stronger in 2001, 

389: and the ratio of the 8498 and the 8662 line varies in the two epochs from $\sim$1.5 

390: to $\sim$2.5. The profiles of these two lines are not similar either, but the 8662 \AA\ 

391: line can be blended the Paschen 13 line. 

392: 

393: The high-resolution spectra show forbidden [N II] emission at 6548 and 6583 \AA, 

394: typically associated to winds and shocks (Hartmann \& Raymond 1989). The [N II] lines 

395: are stronger in the 2004 data, which has the stronger H$\alpha$. In addition, the [N II] 

396: lines appear redshifted by about 18 km/s in 2004 with respect to 2001. The shock origin 

397: of the [N II] emission results typically in small blueshifts of the lines 

398: (Hartigan et al. 1995), so the velocity shift observed could be related to changes 

399: in the velocity and/or location of the shocked matter, or in the radial velocity

400: of the star. Considering the strength of the lines, we believe that GM Cep had stronger 

401: winds during 2004 than in 2001. The [O I] line at 6300 \AA\ is  not covered, and the [O I] 

402: line at 6363\AA\ is not detected, but the [S II] line at 6729 \AA\ is detected in emission.

403: The forbidden emission lines ([O I], [N II], [S II]) can be used to constrain the wind 

404: and shock density and the temperature of the regions where they originate, given their 

405: different critical densities and ionization potentials (Hartigan et al. 1990). If a jet 

406: or wind does not encounter matter in its way, the forbidden line emission may not be 

407: produced (Hartigan et al. 1995). Higher densities tend to reduce the [N II] and [S II] 

408: emission compared to [O I], and higher shock velocities increase ionization and the 

409: [N II]/[O I] ratio. The different velocities observed for the different lines suggest 

410: different formation volumes: [N II] could be produced in a low-density shock far from 

411: the star, whereas the [O I] line could be related to denser, hotter wind maybe

412: associated to the material producing the H$\alpha$ absorption.

413: 

414: GM Cep presents a remarkable, double-peaked O I emission at 8446 \AA, with peaks 

415: centered at 8442 and 8448 \AA, respectively ($\sim$-140 km/s and $\sim$+80 km/s; 

416: Figure \ref{lines1-fig}). This line is characteristic of very strong winds, 

417: being usually present in Herbig Ae/Be stars and in strongly accreting TTS like RW Aur 

418: (Alencar et al. 2005). The double peak observed in GM Cep could in fact indicate two 

419: jet-like wind components. There is a prominent and broad O I absorption at 7774\AA, 

420: maybe related to the hot disk (Grinin et al. 2001). We detect as well some weak 

421: Fe II and Fe I lines in emission  (Table \ref{lines-table}), which are seen in 

422: spectra of other strongly accreting stars like V1647 (Fedele et al. 2007).

423: 

424: The new data allow to revise the spectral type of the star. The low-resolution spectra are  

425: veiled and lack broad band features, giving an uncertain spectral type around G. The 

426: high-resolution HIRES spectrum shows numerous photospheric lines in the whole wavelength 

427: range, despite important blending due to the fast rotation of GM Cep. Comparing to standard 

428: spectral libraries (Montes et al. 1997; Coluzzi et al. 1993), the Fe I and Ca I lines, 

429: together with the Li I absorption (0.16\AA) suggest a late-type in the range G7V-K0V 

430: (1$\sigma$; G5-K3 for 3$\sigma$). The main contributors to the uncertainty in the spectral 

431: type are the moderate veiling ($\sim$0.2-0.4, depending on spectral type) and the fast 

432: rotation, added to the intrinsic lack of strong features in G-type stars. 

433: 

434: Cross-correlation of the Hectochelle spectrum with spectra of slow rotators of a similar 

435: spectral type revealed a rotational velocity V$sini$=51.6$\pm$9.4 km/s (Paper IV). We repeated 

436: the cross-correlation with each one of the HIRES orders, removing the zones affected by 

437: emission lines, bad pixels, and strong atmospheric features (mostly O$_2$ and H$_2$O bands; 

438: Curcio et al. 1994). The results of the cross-correlation are summarized in Table \ref{xcor-table}, 

439: and displayed in Figure \ref{xcor-fig}. The HIRES data confirm the very high rotational velocity

440: over a large range of wavelengths, V$sini$=43.2 $\pm$ 1.6 km/s in average (standard deviation 

441: $\sigma$=4.2 km/s). Compared to the average rotational velocity of the rest of Tr 37 members 

442: (V$sini$=10.2$\pm$3.9 km/s), GM Cep is a very fast rotator: the next fast 

443: rotator in Tr 37 is the M0 weak-lined TTS 21-763, with V$sini$=28.3$\pm$8.1 km/s. We do not 

444: observe any correlation between V$sini$ and the wavelength (Figure \ref{xcor-fig}) nor between 

445: the line broadening and the line transition lower excitation potential, as has been 

446: suggested for FU-ors (Hartmann \& Kenyon 1987; Welty et al. 1990, 1992; Kenyon \& Hartmann 1996). 

447: Therefore, we believe that the absorption lines arise rather from the stellar photosphere than 

448: from a hot disk or shell. The cross-correlation provides as well a measure of the radial 

449: velocity. The average radial velocity in Tr 37 is CZ=-15.0$\pm$3.6 km/s, so GM Cep is slightly

450: off-cluster with CZ=-7.5$\pm$6.7 km/s (Paper IV). The HIRES data produces a different off-cluster

451: value, CZ=-21.0$\pm$2.6 km/s. On the other hand, the radial velocity of the $^{12}$CO (1-0) 

452: peak is consistent with the cluster average -15$\pm$2 km/s (Section \ref{data-co}).

453: Nevertheless, the broad lines lead to large errors, so the evidence of spectroscopic binarity 

454: is only 2$\sigma$ and therefore, not conclusive.

455: 

456: 

457: 

458: \subsection{Accretion \label{accretion}}

459: 

460: The U band observations combined with simultaneous VRI photometry can be used to 

461: determine the accretion rate via the excess U band luminosity (L$_U$; Gullbring et al. 1998; 

462: Paper I, II, IV). The excess in U band luminosity is well correlated to the accretion

463: luminosity L$_{acc}$, which can be transformed into an accretion rate. Following

464: Gullbring et al. (1998):

465: 

466: \begin{equation}

467: \log(L_{acc}/L_\odot) = 1.09 \log(L_U / L_\odot) + 0.98

468: \end{equation}

469: 

470: \begin{equation}

471: L_{acc} \sim G M_* \dot{M} / R_* (1-R_* / R_{in}) 

472: \end{equation}

473: 

474: The mass and radius of the star (M$_*$, R$_*$) can be obtained from the V vs. V-I

475: diagram and evolutionary tracks (Siess et al. 2000). L$_U$ can be estimated as 

476: the difference between the measured U band luminosity and the

477: photospheric luminosity, calculated from the measured I magnitude and using the U-I 

478: color of a non-accreting star with the same spectral type (Kenyon \& Hartmann 1995). 

479: All magnitudes must be corrected from extinction (derived from the VRI colors using a

480: standard galactic extinction law; Cardelli et al. 1989), and the transformation of 

481: magnitudes to luminosity can be done with the zero point flux and the bandwith 

482: for U band (4.19 10$^{-9}$ erg s$^{-1}$ cm$^{-2}$ \AA$^{-1}$ and 680 \AA, respectively). For 

483: GM Cep, and for an spectral type G7-K0, the extinction is A$_V$$\sim$2-3 mag, so the 2006-2007

484: data results in variable L$_U \sim$0.5 - 4 L$_\odot$ and a variable accretion rate in 

485: the range \.{M}$\sim$10$^{-7}$ - 5 10$^{-6}$ M$_\odot$/yr. The mass and radius of the star 

486: are uncertain, given the variations within the color-magnitude diagram (Figure \ref{colorJ-fig}), 

487: being the most probable values M$_*$=2.1 M$_\odot$ and a radius between

488: 3 to 6 R$_\odot$. Added to the uncertainty in the spectral type and

489: extinction, this result is accurate within 1 order of magnitude. In any case, 

490: this accretion rate is 2-3 orders of magnitude higher than the median in Tr 37 

491: (Paper IV), and 1-2 orders of magnitude higher than the rates in typical Taurus stars.  

492: 

493: Using the parametrization in Natta et al. (2004), we can convert the H$\alpha$ velocity 

494: wings at 10\% of the maximum (V$_{H\alpha 10\%}$) into accretion rates:

495: 

496: \begin{equation}

497: \log(\dot{M} / M_\odot yr^{-1}) = -12.89 + 9.7 \times 10^{-3} V_{H\alpha 10\%} / km s^{-1}

498: \end{equation}

499: 

500: For the measured velocities of 580 and 660 km/s, the accretion rates are 5 10$^{-8}$ M$_\odot$/yr 

501: and 3 10$^{-7}$ M$_\odot$/yr, respectively. These values may be underestimated given the wind

502: absorption in the H$\alpha$ profiles and depending on the viewing angle. 

503: As the dispersion associated to this parametrization is around 1-2 orders of magnitude, the 

504: 3 10$^{-7}$ M$_\odot$/yr rate derived from Hectochelle in 2004 is roughly consistent with the 

505: values derived from U band photometry. The moderate veiling estimated from the HIRES spectrum

506:  ($\sim$0.2-0.4) agrees with the accretion rate estimated from H$\alpha$ assuming an intermediate 

507: luminosity for GM Cep.

508: 

509: 

510: \subsection{IR Variability and Disk Models \label{diskmodels}}

511: 

512: In order to study the structure of the disk around GM Cep, we trace the spectral energy 

513: distribution (SED) using the millimeter, Spitzer, 2MASS, and optical data. The main 

514: uncertainty here is that the different observations are not simultaneous. 

515: The 2MASS JHK data are nearly contemporary to the optical photometry (2000), but the Spitzer 

516: data were taken in 2003-2004, and the millimeter continuum in 2006. Short time variations of 

517: $\sim$1 mag in JHK have been observed for stars similar to GM Cep (Eiroa et al. 2002), 

518: suggesting structural changes in the disks. According to the  D'Alessio et al. (2005) disk 

519: models, the flux at wavelengths shorter than $\sim$ 300 $\mu$m scales linearly with the total 

520: (stellar and accretion) luminosity, but the disk flux at longer wavelengths is independent 

521: (Mer\'{\i}n et al. 2004). Juh\'{a}sz et al. (2007) and K\'{o}sp\'{a}l et al. (2007) found 

522: similar evidences of mid-IR excess variations with luminosity. If the causes of the outbursts 

523: were extinction variations, the mid- and far-IR variability would be milder, and the near-IR 

524: would be only moderately variable.  The differences between the fluxes measured by IRAS 12$\mu$m, 

525: IRAC 8$\mu$m, and the MSX6C A band ($\sim$8 $\mu$m) are suggestive of IR variability. On the 

526: other hand, the millimeter emission reflects the total mass of the disk, changing little even 

527: if the star suffers strong variability.The brightness in V varied in the whole range 

528: ($\sim$2-2.5 mags) during the time when the different IR observations were taken 

529: (Figure \ref{lightcurve-fig}), so for modelling the disk we take the optical data from

530: September 2000, which has an intermediate brigthness, and keep in mind that the SED

531: may be affected by variable IR fluxes.

532: 

533: Typical disk models assume that the disk is in hydrostatic equilibrium. Nevertheless, the 

534: strong IR emission suggests that the disk is compact, dense, and, if the variability is triggered 

535: by episodes of enhanced accretion, probably out of hydrostatic equilibrium (Clarke \& Lin 1990; 

536: Eiroa et al. 2002), so we must be cautious interpreting the models. We use the standard 

537: models for irradiated accretion disks around TTS of D'Alessio et al. (2005), which trace the 

538: flared disk self-consistently, with a flaring angle determined by the hydrostatic equilibrium, 

539: and include stellar illumination and accretion as heating sources for the disk. The dust 

540: opacities are calculated for a collisional distribution of 

541: grain sizes, where the amount of grains of a given size ``a" varies as a power law

542: n(a)$\propto$ a$^{-3.5}$ and the maximum grain size is 10$\mu$m. We assume an average inclination 

543: of 60 degrees; the effects of intermediate inclination variations being only important for the 

544: near-IR emission. The stellar parameters, derived from the optical photometry of September 2000

545: using the HR diagram, are M=3.0 M$_\odot$, R=5.14 R$_\odot$, T=5000 K. We also include an accretion 

546: rate of  10$^{-6}$ M$_\odot$/yr. These parameters are one of the main uncertainties in the modeling, 

547: given the variability of the star. The best fit model is displayed in Figure \ref{bestfit-fig}.  

548: The result is, not surprising, that the disk cannot be modeled as a standard disk. If we try to fit 

549: the MIPS data points at 24 and 70 $\mu$m, the 1.3mm continuum flux would be overestimated by more 

550: than one order of magnitude, and even the 24 and 70 $\mu$m contemporary data alone cannot be 

551: accurately reproduced with hydrostatic equilibrium disk models for any observed stellar luminosity. 

552: The mid-IR flux of GM Cep is comparable to that of FU Orionis itself (Quanz et al. 2006), 

553: although its disk (or disk plus envelope) is less massive. Since the flux at millimeter wavelengths 

554: is independent of the rapid optical variability, a standard hydrostatic equilibrium model can 

555: reproduce the flux at 1.3mm only if the grains are either much smaller or much larger than 

556: 10 $\mu$m (maximum grain size $<$1$\mu$m or $>$1cm) or if the disk is very small (up to 30-50 AU). 

557: Even in these cases, an additional source of IR emission should be included to produce the high 

558: 24 and 70 $\mu$m flux (in a small disk) or to fit the near-IR excess (if the grain sizes have 

559: cm sizes). 

560: 

561: A way to increase the IR fluxes without changing the mm flux would be by  varying the 

562: flaring angle at a certain distance from the star. The flaring angle in the 

563: D'Alessio et al. (2005) models is fixed by the hydrostatic equilibrium, but

564: strong turbulence related to the strong accretion can produce an unstable

565: disk. As an experiment, the mid-IR fluxes can be increased by adding a black body 

566: emission with T$\sim$150 K to a compact 30 AU disk model. For the luminosity of 

567: GM Cep, this would mean a distance of $\sim$ 25 AU. The excess mid-IR emission that 

568: needs to be added could be interpreted as an increase in the flaring angle by a factor

569: of $\sim$2 at distances of $\sim$20-30 AU (a ``bump'' in the disk), as an illuminated 

570: wall at $\sim$25 AU, or as a shell or envelope located at a similar distance. If the 

571: innermost disk were very flared or had a high inner wall, some degree of self-shadowing 

572: in the near-IR may occur, leading to a ``bump'' at longer wavelengths (Dullemond et al. 2001).

573: 

574: A physical mechanism to produce a ``bump'' or a local change in the flaring angle could 

575: be to disrupt the disk by a massive object (a stellar companion and/or a giant planet) 

576: triggering thermal instability and producing rims and/or spiral waves (Clarke \& Lin 1990; 

577: Clarke \& Syer 1996; Lodato \& Clarke 2004). This picture would also be useful to explain 

578: the observed outbursts as periods of increased accretion activity. A close-in companion, 

579: a companion embedded in the disk, or a companion out of the disk (typically, at distances

580: 2-3 times the size of the disk; Artymowicz \& Lubow 1994) could produce disk instabilities 

581: and increased accretion episodes, with the advantage that companion-triggered outbursts can 

582: occur at later stages of evolution in Class II objects, more consistent with the ages of Tr 37

583: (Lodato \& Clarke 2004). Detailed studies of the periodicity and radial velocity will be 

584: used to test these scenarios.

585: 

586: The shell/envelope picture presents a main problem: the relatively low extinction of GM Cep. 

587: Assuming the reasonable spectral type range (G5-K3) the extinction can vary between A$_V$=1.5-3 mag, 

588: roughly consistent with the cluster average A$_V$=1.67$\pm$0.45 mag (Paper II), and with the 

589: minimum interstellar extinction for an object located at 900 pc, A$_V \sim$1 mag.

590: Therefore, the column density in the envelope would be of the order of N(H$_2$)=10$^{21}$ cm$^{-2}$ 

591: at most (Frerking et al. 1982; Taylor et al. 1993), resulting in a total mass of 

592: $\sim$4$\times 10^{-5}$M$_\odot$ if the envelope has a $\sim$30 AU radius, and still 

593: only 0.001 M$_\odot$ for a $\sim$500 AU radius, insignificant compared to the total

594: disk mass. The envelope could be more dense and massive if it had polar gaps, which is 

595: reasonable given the strong winds, and if we assume a low viewing angle, but it cannot be 

596: too massive to reproduce the millimeter flux. A low viewing angle would be consistent with 

597: the lack of broadening in the $^{12}$CO(1-0) line, and would suggest that GM Cep is rotating close

598: to the breakup velocity, $\sim$200 km/s. Accretion from an envelope onto the disk could produce 

599: episodes of increased accretion onto the star. Extinction by a non-uniform envelope 

600: can produce magnitude changes as well, but the rapid variation timescales require that, if 

601: extinction plays an important role in the magnitude changes, the extincting matter must be 

602: very close to the star. In order to distinguish between all these scenarios, simultaneous 

603: observations at all wavelengths are required.

604: 

605: 

606: \subsection{Disk Mass Estimates \label{diskmass}}

607: 

608: The millimeter continuum data can be used to estimate the

609: total disk mass, given that the whole disk is optically thin

610: at these wavelengths. The total dust mass is written as:

611: 

612: \begin{equation}

613: M_{dust} =  F_\nu D^2 / \kappa_\nu B_\nu(T_{dust})

614: \end{equation}

615: 

616: where F$_\nu$ is the flux at millimeter wavelengths, D is the distance, 

617: $\kappa_\nu$ is the dust mass absorption coefficient at the same wavelength

618: (2 cm$^{2}$ g$^{-1}$ at 1.3mm; Beckwith et al. 1990; Miyake \& Nakagawa 1993; 

619: Kr\"{u}gel \& Siebenmorgen 1994). B$_\nu$(T$_{dust}$) is the Planck function 

620: for the temperature of the bulk of the dust. The total mass of the disk can be 

621: obtained assuming a gas-to-dust ratio of 100 (Beckwith et al. 1990). The main 

622: uncertainties are the dust absorption coefficient (if substantial grain growth

623: has occurred), the gas-to-dust ratio (if the disk is older than typical disks), 

624: and, in an unstable disk, the temperature of the bulk of the dust. We use the 

625: dust absorption coefficient assumed for Taurus disks (Beckwith et al. 1990; 

626: Osterloh \& Beckwith 1995), which is larger than the $\kappa_\nu$ for the interstellar

627: medium (1 cm$^{2}$ g$^{-1}$) or than the one of disks with grains larger than $\sim$100 $\mu$m 

628: (Kr\"{u}gel \& Siebenmorgen 1994). The typical dust temperature for outer disks around TTS is 

629: 30-40 K (Dutrey et al. 1997; Thi et al. 2001). For GM Cep, the dust temperature could be 

630: higher if the strong accretion is adding extra heating to the disk, and/or if the bulk of 

631: the dust is closer to the star (because of changes in the flaring and/or envelope configuration). 

632: In order to be conservative, we consider dust temperatures in the range 30-150 K, 

633: resulting in total disks masses from 0.07 to 0.01 M$_\odot$, several times the 

634: minimum mass solar nebula.

635: 

636: The emission of optically thin transitions in molecules can be used to estimate the

637: mass in cold gas (Thi et al. 2001). The $^{12}$CO(1-0) line detected at the IRAM 30m 

638: telescope is most likely optically thick, so this would  underestimate the disk mass. 

639: Nevertheless, if we make the approximate calculation, assuming that the line is 

640: thermalized, we have:

641: 

642: \begin{equation}

643: F_{(1-0)} = (h \nu / 4 \pi D^2) N(CO) A_{1-0} \chi_\nu A

644: \end{equation}

645: 

646: where F$_{(1-0)}$ is the total flux in the $^{12}$CO(1-0) line,

647: N(CO) is the $^{12}$CO column density, D is the distance, A$_{1-0}$ is the

648: Einstein coefficient for the transition,  $\chi_\nu$ is the 

649: level occupation at a given temperature, and A is the disk surface area. 

650: The main error in this approach is that the line is not optically thin. 

651: If we assume the ratio of H$_2$ to $^{12}$CO to be 10$^4$

652: (Thi et al. 2001), and we consider the total flux integrated 

653: over the line profile F$_{(1-0)}$, the mass of the disk can be written as:

654: 

655: \begin{equation}

656: M_{disk}= \mu m_{H_2} F_{(1-0)} 4 \pi D^2 (N(H_2)/N(CO))/ h \nu A_{1-0} \chi_\nu

657: \end{equation}

658: 

659: With A$_{1-0}$ and $\chi_\nu$ from the Leiden Atomic and Molecular Database (LAMBDA; 

660: Sch\"{o}ier et al. 2005) and the Cologne Database of Molecular Spectroscopy (CDMS; 

661: M\"{u}ller et al. 2001, 2005), and assuming gas temperatures between 40 and 80 K 

662: (this variation results only in a factor of 2), the total disk mass inferred from the 

663: integrated flux F$_{(1-0)}$=0.27 Jy km/s is of the order of 0.0005 M$_\odot$,

664: about 100 times smaller than the estimate from the dust

665: continuum emission. This is reasonable since the line is optically thick, and 

666: the typical CO depletion values in TTS disks are of the order of 10-1000 

667: (Thi et al. 2001). Moreover, our $^{12}$CO fluxes may be underestimated

668: because of the absorption from ambient gas (Section \ref{data-co}).

669: 

670: The disk mass obtained integrating the 30 AU disk model that gives the best

671: fit to the millimeter point (Section \ref{diskmodels}) is about 3 times larger,

672: 0.19 M$_\odot$. The model suggests a temperature around 30-40 K for the bulk

673: of dust. The main source of disagreement is the dust 

674: mass absorption coefficient, which in the model is variable with the radius of 

675: the disk and depends on the distribution of grains of different sizes, having

676: typical values between 0.4 and 1.8 cm$^2$ g$^{-1}$, smaller than the $\kappa _\nu$

677: from Beckwith et al. (1990). The stronger inconsistency is that for an accretion 

678: rate \.{M}$\sim$10$^{-6}$M$_\odot$/yr, the disk would survive less than 

679: 0.1 Myr, very little compared with the ages of Tr 37 and even with the ages of the Tr 37 

680: globule. This inconsistency is not exclusive to GM Cep, but many of the strong accretors like 

681: DR Tau and RW Aur have similar \.{M} and even smaller millimeter disk masses

682: (0.03 and 0.0003 M$\odot$, respectively; Beckwith et al. 1990; Osterloh \& Beckwith 1995), so 

683: their survival times would be even shorter. In case of RW Aur, the problem can

684: be solved assuming that the current accretion episodes are transient events caused by 

685: disk disturbances by its companion(s) (Petrov et al. 1991; Cabrit 2005). EX-ors are 

686: supposed to have variable accretion, so they could spend most of their lives accreting at 

687: lower rates. Another way to solve the disagreement would be to change the opacities by 

688: increasing the maximum grain sizes to $\sim$1-10 cm or decreasing it under 1$\mu$m. Such dust

689: distributions are compatible with the millimeter fluxes if the disk has a mass 

690: $\sim$0.5-0.6 M$_\odot$ and outer radius $>$100 AU. Mannings \& Emerson (1994) suggested that 

691: including large grains and fractal dust structures could increase the disk mass of DR Tau, 

692: and Rodmann et al. (2006) found evidence of cm-sized grains in RW Aur.

693: Important grain growth seems appropriate for older TTS like GM Cep and should

694: be explored with more observations in the submm range to constrain the SED slope.

695: 

696: \subsection{Does GM Cep have a Wide-Orbit Companion? \label{companion}}

697: 

698: Given the anomalous SED of GM Cep, the repeated outbursts, and the fact that many of 

699: the well-known variable and EX-or/FU-or objects are multiple systems with separations

700: $\sim$100-300 AU (Reipurth \& Aspin 2004; Vittone \& Errico 2005; Ghez et al. 1993), we 

701: investigate the presence of companions. The presence of binaries has been invoked as a 

702: mechanism to produce EX-or and FU-or outbursts in old and more evolved disks without 

703: an envelope (Clarke \& Lin 1990; Bonnell \& Bastien 1992). Some very 

704: variable TTS like RW Aur are such multiple systems (Ghez et al. 1993), and companions at 

705: distances $\sim$100 AU affect the evolution of disks around Herbig AeBe stars 

706: (Chen et al. 2006) and TTS (Bouwman et al. 2006). The high-resolution spectroscopy is not 

707: conclusive, but the detection of spectroscopic binaries may be difficult if the star is 

708: a very fast rotator, and/or if the viewing angle is low. The lucky imaging data from 

709: AstraLux does not reveal a wide companion. The small elongation of the PSF is compatible 

710: with the predicted atmospheric dispersion effects, and cannot be taken as an indicator 

711: of an unresolved companion. The ability to image companions with AstraLux is highly 

712: dependent on the magnitude difference between the two stars, and GM Cep was always imaged 

713: during high magnitude phases, so the identification of companions 

714: at 100-200 AU would not be possible even if the two stars have a similar mass. 

715: 

716: In order to constrain the separation and magnitude difference of a possible companion, we 

717: analyzed the maximum brightness differences of a hypothetical companions at different 

718: angular separations from GM Cep for the November 2006 AstraLux data, which has the best 

719: Strehl ratio. We adopted a 5$\sigma$ peak detection over the background noise as visibility 

720: criterion. The absence of static aberrations and asymmetries in the AstraLux data beyond the 

721: first diffraction ring allow to determine the intensity standard deviation in concentric 

722: annuli around GM Cep, resulting in robust estimates of the detection limits 

723: (Table \ref{binary-table}). Binary simulations were constructed by adding a scaled copy of 

724: the real observations to itself with the given magnitude differences and separations 

725: (Figure \ref{simulations-fig}). Since not only the stellar PSF is added, but also the noise, 

726: this method increases the background noise especially at small magnitude differences, leading 

727: to more pessimistic estimates of the achievable detection limits. The observed PSF is in 

728: good agreement with a theoretical simulation which includes atmospheric dispersion effects, 

729: so there is no evidence for a bright companion at angular separations between 50-100 mas

730: (45-90 AU). We conclude that GM Cep does not have a companion with a magnitude 

731: difference $\le$1 mag at $\approx$100 AU, and none farther than 200 AU with a magnitude

732: difference $\le$2 mag. However, we cannot rule out the presence of a close G- or K-type 

733: companion with normal luminosity. We will obtain further AstraLux observation of GM Cep near

734: its minimum in parallel to our photometric monitoring.

735:  

736: \section{Discussion: Potential Variability Mechanisms in GM Cep\label{outburst}}

737: 

738: There are different mechanisms responsible for the magnitude variations of 

739: pre-main sequence stars. Herbst et al. (1994) classifies them in Type I, 

740: due to rotation of a star with cool spots, Type II, due to rotation of a 

741: star with hot spots (related to accretion), and Type III, 

742: also called UX-or variables. Young, accreting, solar-type stars experience 

743: Type I and II variations of the order of few 0.01-0.5 mag in a few days 

744: (Bertout 1989; Brice\~{n}o et al. 2001; Eiroa et al. 2002). The irregular 

745: UX-or variations,  with more complex timescales of days/weeks combined with 

746: few-years cycles, do not usually exceed $\Delta$mag$\sim$1 mag, and occur 

747: typically in nearly edge-on HAeBe stars (Grinin et al. 2000), although other 

748: authors claim that stars as early as K0 can experience Type III variations

749: (Herbst \& Shevchenko 1999) and that the viewing angle can be as low as 45

750: degrees (Natta \& Whitney 2000). The mechanisms triggering the UX-or variations

751: are the most controversial, and they have been described as obscuration by shells 

752: of matter close to the star (Shevchenko et al. 1993; Natta et al. 1997; 

753: Grinin et al. 2000; Tambovtseva et al. 2001) or as thermal instabilities similar to

754: those triggering FU-or outbursts (Herbst \& Shevchenko 1999; Natta et al. 2000).

755: Strong variations ($\Delta$mag$\sim$2-5 mag) due to enhanced accretion episodes 

756: are seen in FU-or or EX-or outbursts, which are recurrent at least in the case 

757: of EX-or objects (Hartmann \& Kenyon 1996; Lehman et al. 1995; Herbig et al. 2001). 

758: These extreme accretors tend to be very fast rotators (Herbig et al. 2003; although 

759: the fast rotation detected may be actually the rotation of the very hot inner 

760: disk), to have IR fluxes that deviate from typical disk models (Hartmann \& Kenyon 1996), 

761: and to have in some cases companions at 100-300 AU (Reipurth \& Aspin 2004). FU-ors 

762: and probably EX-ors could be either a normal phase within the very early evolution 

763: and disk formation in TTS (Clarke et al. 1990; Hartmann \& Kenyon 1996, among others), 

764: or a special type of systems (certain binary or multiple systems; 

765: Herbig et al. 2003; Quanz et al. 2006). The long-time evolution

766: of these objects is not known, so the study of relatively old variables within

767: these classes could help to understand their nature.

768: 

769: The lightcurve of GM Cep is not purely consistent with any known FU-or, EX-or 

770: nor UX-or lightcurves, although these objects are largely variable within their 

771: classes. The magnitude changes ($\Delta$mag$\sim$2-2.5 mag) and multiple maxima 

772: are consistent with an EX-or or UX-or type variable. The high IR flux is another 

773: characteristic of EX-or/FU-or systems, due to their high luminosity and/or to 

774: the presence of envelopes. The H$\alpha$ and [N II] variability suggest changes 

775: in the accretion rates and in the strength of the winds, and although the 

776: H$\beta$ absorption is typical of UX-ors, the blueshifted absorption in H$\alpha$ 

777: is very rare in this class. The magnitude changes are not consistent with variable 

778: obscuration alone considering a standard galactic extinction law. Nevertheless,

779: if the extinction were produced by material in the circumstellar disk/shell

780: of GM Cep, containing reprocessed dust and large grains, the color changes during

781: high extinction episodes could be different. The type of obscuration suggested for UX-or

782: variables (Shevchenko et al. 1993; Natta et al. 1997) results in bluer colors at low 

783: magnitude, due to obscuration of the red disk and scattering (Grinin et al. 2001), which 

784: have not been seen so far in GM Cep. Moreover, pure extinction episodes cannot explain the 

785: high luminosity of GM Cep unless the spectral type and/or the luminosity class are 

786: changed beyond reasonable limits. Extinction variations alone could account 

787: only for up to $\sim$1.5 mags in V and cannot produce the observed changes in color.

788: If variable extinction were produced by the disk material in the vicinity 

789: of the star, the system should be close to edge-on or should have an anomalous 

790: distribution of circumstellar matter. Given the observed V$sini$, the angle of view must 

791: be larger than $\sim$15 degrees to prevent breakup, and the near-IR fluxes and 

792: the P-Cygni H$\alpha$ profile suggest an intermediate angle. If the excess 

793: mid-IR emission were due to an envelope, the strong winds could contribute to clean the 

794: polar gaps, reducing the extinction. GM Cep could be then one of the oldest EX-or variables 

795: known, since these phenomena seem to occur preferentially at ages $\sim$1 Myr. According 

796: to the picture developed by Clarke \& Syer (1996), the presence of a planetary or 

797: stellar companion in the disk of a Class II object at few AU distances can trigger 

798: thermal instability and produce outbursts with amplitude and frequency depending on 

799: the mass of the companion (Lodato \& Clarke 2004). Given the ages of Tr 37, it is 

800: most likely that GM Cep is a special object, probably a binary or multiple system, 

801: rather than a typical very young TTS.

802: 

803: In order to understand the variations of GM Cep, we compare it to similar cases:

804: 

805: The activity of the EX-ors EX Lupi and DR Tau resembles strongly that of GM Cep. 

806: EX Lupi suffers periods of quiescence followed by strong variability with

807: 2-3 mag amplitude and changes in its emission line spectrum (Lehmann et al. 1995). 

808: Like GM Cep, it is presumably older than other EX-ors (Herbig 2007). The K5-K7 

809: eruptive star DR Tau is known to have irregular variability episodes since 1900, with 

810: amplitudes $\sim$1.5 to $\sim$3-5 mag. Its variable and strong stellar winds,

811: large UV excess and veiling suggest high and variable accretion rates (Chavarria 1979;

812: Mora et al. 2001). DR Tau can remain relatively stable at both the low-magnitude and 

813: high-magnitude phases during years/decades, as was seen by Suyarkova (1975) for GM Cep. 

814: The variability of DR Tau cannot be explained by hot and cold spots and variable obscuration

815: alone (Eiroa et al. 2002). Like for GM Cep, the mass of the disk of DR Tau 

816: ($\sim$0.03 M$_\odot$; Beckwith et al. 1990) seems to be too small for its accretion rate 

817: if we assume the small typical grain sizes.

818: 

819: Suyarkova (1975) classified GM Cep as a RW Aur-type variable. The ``extreme CTTS'' RW Aur 

820: shares with GM Cep the strong and variable P-Cygni H$\alpha$ profile, a powerful disk, 

821: a large accretion rate, and a strong double-peaked O I emission lines at 8446 \AA, with 

822: velocity shifts $\pm$100 km/s (Ghez et al. 1993; Alencar et al. 2005). It is a triple 

823: system and the prototype of the RW Aur-class of irregular variable stars (Hoffmeister 1957),

824: characterized for rapid magnitude variations with $\Delta$mag$\sim$1-4 mag and a typical 

825: spectral type around G5.  The very small millimeter disk mass of RW Aur ($\sim$0.0003 M$_\odot$;  

826: Osterloh \& Beckwith 1995) suggests a timescale for disk removal of the order 

827: of centuries to millenia, too short unless we assume brief and transient accretion episodes 

828: (Cabrit 2005) and/or important grain growth to cm sizes (Rodmann et al. 2006). The accretion 

829: seems to be non-axisymmetric, as expected from interactions with a close companion

830: (Petrov et al. 2001).

831: 

832: The 1 Myr-old star GW Ori is one of the most massive TTS known (M=2.5 M$_\odot$, R=5.6 R$_\odot$; 

833: Mathieu et al. 1991), resembling strongly GM Cep. With an accretion rate $\sim$10$^{-6}$M$_\odot$/yr, 

834: it is a very fast rotator (V$sini$=43 km/s; Bouvier et al. 1986), and shows variability up to 1 mag

835: in JHK (Samus et al. 2004). It is a binary (maybe triple) system, with a $\sim$G5 primary with 

836: luminosity 26 L$_\odot$, similar to GM Cep and abnormally high for its spectral type (Mathieu et al. 1991).  

837: It is also  has a very strong IR excess, and is one of the most luminous TTS at millimeter wavelengths 

838: (Mathieu et al. 1991, 1995), suggesting a disk (or disk+shell) mass $\sim$1.5M$_\odot$, being the only 

839: one in its class able to sustain high accretion rates over several Myr. 

840:  

841: The K3 star CW Tau shares with GM Cep the rapid rotation (V$sini$=28 km/s; 

842: Muzerolle  et al. 1998), the H$\alpha$ P-Cygni profile and a deep, and a broad OI absorption 

843: at 7773\AA. It shows magnitude variations up to 2 magnitudes (General Catalog of Variable Stars,

844: Kukarkin et al. 1971), and has a strong outflow (Hirth et al. 1994; G\'{o}mez de Castro 1993) 

845: and strong [N II] forbidden line emission. As for GM Cep, the mass of the disk is not specially high

846: despite the signs of strong accretion ($<$0.02 M$_\odot$; Beckwith et al. 1990).

847: 

848: The outburst of IRAS 05436-0007/V1647 (Eisl\"{o}ffel \& Mundt 1997) revealed the 

849: bright McNeil Nebula in 2004 (McNeil et al. 2004). This very extincted (A$_V$$\sim$13 mag) 

850: object shows a flat mid-IR spectrum (in contrast to typical EX-or and similar objects, 

851: which have TTS-looking IR SEDs). It has been considered as a Class 0 embedded object 

852: (Lis et al. 1999), but may be some type of EX-or or FU-or Class II object, surrounded by 

853: a massive circumstellar envelope (M$\sim$0.5 M$_\odot$; \'{A}brah\'{a}m et al. 2004), 

854: with a 5.6 L$_\odot$ luminosity, typical of TTS, or an intermediate case between EX-or 

855: and FU-or objects (Fedele et al. 2007). Its emission line spectrum at optical wavelengths 

856: is similar to the spectrum of GM Cep.

857: 

858: KH 15D is a pre-main sequence binary system with a precessing disk or ring, seen

859: edge on, so one of the components is permanently occulted by the disk, whereas

860: the other is visible only during half of the period (Hamilton et al. 2005). 

861: The color changes are affected by the eclipses and scattering. 

862: Although such special configurations allow much large amplitudes and atypical color

863: variations, a similar scenario fails to explain the high luminosity of GM Cep without 

864: changing its spectral type and luminosity class or introducing additional mechanims (i.e., 

865: hot shells) to reproduce the optical absorption lines. 

866: 

867: \section{Conclusions \label{conclu}}

868: 

869: GM Cep is an extremely variable late G star in the 4 Myr-old cluster Tr 37, 

870: resembling younger EX-or objects. Its complex variability in color 

871: and magnitude is not consistent with a single typical variability 

872: mechanism in TTS (hot/cold spots, variable obscuration, 

873: or changes in the accretion rate). The amplitude of the variations, the

874: high accretion rate, the luminous mid-IR disk, and the high stellar 

875: luminosity suggest variable accretion to be the stronger contributor, maybe

876: mixed with variable extinction by circumstellar material (similar to

877: UX-or variables), and some minor influence of cold spots and scattering.

878: Increased accretion episodes are thought to produce the strong, irregular

879: variations in young EX-or and FU-or objects, typically still surrounded  by

880: massive infalling envelopes. Nevertheless, GM Cep has a medium-mass disk and a small

881: or inexistent envelope, and belongs to a cluster where strong disk evolution

882: is ubiquitous. Large changes in the accretion  rates can result

883: in changes in the stellar and disk structure, and conversely,

884: the disruption of the disk structure (for instance, by companions, or

885: via gravitational instability if the disk is very massive and compact),

886: can increase the accretion rate. The presence of companions 

887: (stellar, substellar or planetary) is a plausible mechanism to produce

888: disk instabilities in a relatively old star like GM Cep. Simultaneous 

889: multiwavelength observations, including IRS spectra scheduled for 2007, 

890: providing coverage in the 5-35$\mu$m range, and detailed 

891: monitoring in the coming years will help us to reveal the nature of the 

892: variability in GM Cep. Sub-millimetre observations in the future should 

893: be used to determine the SED slope and constrain the size of the grains.

894: Old but extremely accreting stars are thus a key to understand the processes 

895: of disk accretion and evolution and the consequences for planet formation.

896: 

897: \acknowledgements

898: We want to thank the Calar Alto Observatory for 

899: allocation of Director's Discretionary Time to this programme.

900: We thank L. A. Hillenbrand for kindly providing the HIRES spectrum

901: of GM Cep taken in 2001, J.A. Acosta-Pulido for providing us with his 'caindr'

902: package, and P. D'Alessio, for providing the tools for modelling the disk. 

903: P.\'A. acknowledges the support of the Hungarian Research Fund No. K62304. 

904: This publication makes use of data products from the Two Micron All Sky Survey, which is 

905: a joint project of the University of Massachusetts and the Infrared Processing and Analysis 

906: Center/California Institute of Technology, funded by the National Aeronautics and Space 

907: Administration and the National Science Foundation. This research has made use of the 

908: SIMBAD database, operated at CDS, Strasbourg, France, of the SuperCOSMOS Sky Survey,

909: provided by WFAU, Institute for Astronomy, Royal Observatory, Blackford Hill, 

910: Edinburgh, and of the Digitized Sky Surveys, produced at the Space Telescope Science 

911: Institute under U.S. Government grant NAG W-2166. We thank as well K. Tristram, 

912: S. Birkmann, B. Braunecker, and J. Stegmaier for their help at the 70 cm

913: K\"{o}nigstuhl telescope, and S. Pedraz-Marcos for his help with the Calar Alto

914: observations. We thank O. Fischer, J. Greiner, P. Kroll, D. Fedele,  C. Eiroa, 

915: M. Wiedner, H. Beuther,  V. Roccatagliata, and B. Conn for their comments and suggestions.

916: We finally thank the anonymous referee for his/her comments and suggestions to improve

917: the quality of this paper.

918: 

919: \begin{thebibliography}{}

920: 

921: \bibitem[\'{A}brah\'{a}m et al.(2004)]{abraham04} \'{A}brah\'{a}m, P., K\'{o}sp\'{a}l, A., Csizmadia, S., Mo\'{o}r, A., Kun, M., Stringfellow, G., 2004, A\&A 419, L39

922: 

923: \bibitem[Alencar et al.(2005)]{alencar05} Alencar, S., Basri, G., Hartmann, L., Calvet, N., 2005, A\&A 440, 595

924: 

925: \bibitem[Beckwith et al.(1990)]{beckwith90} Beckwith, S., Sargent, A., Chini, R., G\"{u}sten, R., 1990, AJ, 99, 924 

926: 

927: \bibitem[Bessell(1979)]{bessell79} Bessell, M., 1979, PASP, 91, 589 

928: 

929: \bibitem[Bessell(1986)]{bessell86} Bessell, M., 1986, PASP, 98, 1303

930: 

931: \bibitem[Blair \& Gilmore(1982)]{blair92} Blair, M., \& Gilmore, G., 1982, PASP 94, 742

932: 

933: \bibitem[Bonnell \& Bastien(1992)]{bonnell92} Bonnell, L., \& Bastien, P., 1992, ApJ, 401, L31

934: 

935: \bibitem[Bouwman et al. (2006)]{bouwman06} Bouwman, J., Lawson, W., Dominik, C., Feigelson, E., Henning, Th., Tielens, A., Waters, L., 2006, ApJ, 653, L57

936: 

937: \bibitem[Cardelli et al.(1989)]{cardelli89}  Cardelli, J., Clayton, G., Mathis, J., 1989, ApJ, 345, 245

938: 

939: \bibitem[Chavarria(1979)]{chavarria79} Chavarria-K., C., 1979, A\&A, 79, 18 

940: 

941: \bibitem[Chen et al.(2006)]{chen06}   Chen, X. P.; Henning, T.; van Boekel, R.; Grady, C. A., 2006, A\&A, 445, 331

942: 

943: \bibitem[Clarke et al.(1990)]{clarke90}  Clarke, C.,  Lin, D., Pringle, J., 1990, MNRAS, 242, 439

944: 

945: \bibitem[Clarke \& Syer(1996)]{clarkesyer96} Clarke, C., \& Syer, D., 1996, MNRAS, 278, L23

946: 

947: \bibitem[Clarke et al.(2005)]{clarke05} Clarke, C., Lodato, G., Melnikov, S., Ibrahimov, M., 2005, MNRAS, 361, 942

948: 

949: \bibitem[Coluzzi(1993)]{coluzzi93} Coluzzi, R., 1993, ILLSS Catalogue, Bull. Inf. CDS, 43, 7 

950: 

951: \bibitem[Contreras et al.(2002)]{contreras02}  Contreras, M.E., Sicilia-Aguilar, A., Muzerolle, J., Calvet, N., Berlind, P., Hartmann, L. 2002, AJ, 124, 1585

952: 

953: \bibitem[Couch \& Newell(1980)]{couch80} Couch, W., \& Newell, E., 1980, PASP, 92, 746

954: 

955: \bibitem[Curcio et al.(1964)]{curcio64} Curcio, J., Drummeler, L., Knestrick,G., 1964, Applied Optics Vol. 3, N. 12, 1401 

956: 

957: \bibitem[D'Alessio et al.(1999)]{dale99} D'Alessio, P., Calvet, N., Hartmann, L., Lizano, S., Cant\'{o}, J., 1999, \apj, 527, 893

958: 

959: \bibitem[D'Alessio et al. (2005)]{dale05a} D'Alessio, P., Mer\'{\i}n, B., Calvet, N., Hartmann, L., \& Montesinos, B., 2005, Rev. Mex. A. A., 41, 1

960: 

961: \bibitem[Dullemond et al. 2001]{dullemond01} Dullemond, C., Dominik, C.,  Natta, A., 2001, ApJ 560, 957 

962: 

963: \bibitem[Dutrey et al.(1997)]{dutrey97} Dutrey, A., Guilloteau, A., Guelin, M., 1997, A\&A, 317, L55 

964: 

965: \bibitem[Edwards et al.(1994)]{edwards94} Edwards, S., Hartigan, P., Gandour, L., Andrulis, C., 1994, AJ 108, 1056 

966: 

967: \bibitem[Eisl\"{o}ffel \& Mundt(1997)]{eisloffelmundt97} Eisl\"{o}ffel, J.; Mundt, R. 1997 AJ, 114, 280

968: 

969: \bibitem[Fedele et al.(2007)]{fedele07} Fedele, D., van den Ancker, M., Petr-Gotzens, M., Rafanelli, P., 2007, AA in press 

970: 

971: \bibitem[Fernie(1983)]{fernie83} Fernie, J., 1983, PASP, 95, 782 

972: 

973: \bibitem[Frerking et al.(1982)]{frerking82} Frerking, M., Langer, W., Wilson, R., 1982, ApJ, 262, 590 

974: 

975: \bibitem[Ghez et al.(1993)]{ghez93} Ghez, A., Neugebauer, G., Matthews, K., 1993, AJ 106, 2005 

976: 

977: \bibitem[G\'{o}mez de Castro(1993)]{gomez93} G\'{o}mez de Castro, A., 1993, ApJ, 412, 43

978: 

979: \bibitem[Grinin et al. (2000)]{grinin00} Grinin, V., Rostopchina, A., Shakhovskoy, D., 2000, IAUS, 200, 115

980: 

981: \bibitem[Grinin et al. (2001)]{grinin01} Grinin, V., Kozlova, O., Natta, A., Ilyin, I., Tuominen, I., Rostopchina, A., Shakhovskoy, D., 2001, A\&A, 397, 482

982: 

983: \bibitem[Gullbring et al.(1998)]{gullbring98} Gullbring, E., Hartmann, L., Brice\~{n}o, C., Calvet, N., 1998, ApJ 492, 323

984: 

985: \bibitem[Hamilton et al.(2005)]{hamilton05} Hamilton, C., Herbst, W., Vrba, F., \& 14 more coauthors, 2005, AJ, 130, 1896

986: 

987: \bibitem[Hartmann et al.(1986)]{hartmann86} Hartmann, L., Hewett, R., Stahler, S., Mathieu, R., 1986, \apj, 309, 275

988: 

989: \bibitem[Hartmann \& Kenyon(1987)]{hartmannkenyon87} Hartmann, L., \& Kenyon, S., 1987, ApJ 322, 393

990: 

991: \bibitem[Hartmann \& Raymond(1989)]{harmtannraymond89} Hartmann, L., \& Raymond, J., 1989,  ApJ , 337, 903

992: 

993: \bibitem[Hartmann \& Kenyon(1996)]{hartmannkenyon96} Hartmann, L., \& Kenyon, S., 1996, ARAA, 34, 207

994: 

995: \bibitem[Herbig et al.(2001)]{herbig01} Herbig, G., Aspin, C., Gilmore, A., Imhoff, C., Jones, A., 2001, PASP, 113, 1547

996: 

997: \bibitem[Herbig et al.(2003)]{herbig03} Herbig, G., Petrov, P., Duemmler, R., 2003, ApJ, 595, 384

998: 

999: \bibitem[Herbig(2007)]{herbig07} Herbig, G., 2007, AJ, 133, 2679 

1000: 

1001: \bibitem[Herbst et al.(1994)]{herbst94}  Herbst, W., Herbst, D., Grossman, E., Weinstein, D., 1994, AJ 108, 1906

1002: 

1003: \bibitem[Herbst \& Shevchenko(1999)]{herbst99} Herbst, W., Shevchenko, V., 1999, AJ 118, 1043

1004: 

1005: \bibitem[Hirth et al.(1994)]{hirth94} Hirth, G. A.; Mundt, R.; Solf, J., 1994, A\&A, 285, 929

1006: 

1007: \bibitem[Hoffmeister(1957)]{hoffmeister57} Hoffmeister, C., 1957, IAUS, 3, 22

1008: 

1009: \bibitem[Juh\'{a}sz et al.(2007)]{juhasz07} Juh\'{a}sz, A., Prusti, T., \'{A}brah\'{a}m, P., Dullemond, C., 2007, MNRAS, 374, 1242

1010: 

1011: \bibitem[K\'{o}sp\'{a}l et al.(2007)]{kospal07}  K\'{o}sp\'{a}l, A.; \'{A}brah\'{a}m, P.; Prusti, T.; Acosta-Pulido, J.; Hony, S.; Mo\'{o}r, A.; Siebenmorgen, R., 2007, A\&A in press

1012: 

1013: \bibitem[Kr\"{u}gel \& Siebenmorgen(1994)]{kruegel94} Kr\"{u}gel, E., \& Siebenmorgen, R., 1994, A\&A, 288, 929

1014: 

1015: \bibitem[Kun(1986)]{kun86} Kun, M., 1986, Commission 27 of the I.A.U.,  Information Bulleting on Variable Stars, 2961

1016: 

1017: \bibitem[Kukarkin et al.(1971)]{kukarkin71}Kukarkin B.V., Kholopov P.N., Pskovski Y.P., Efremov Y.N., Kukarkina N.P., Kurochkin N.E. and Medvedeva G.I., 1971, General Catalogue of Variable Stars, 3rd ed.

1018: 

1019: \bibitem[Kurtz et al.(1992)]{kurtz92} Kurtz, M.J., Mink, D.J., Wyatt, W.F., Fabricant, D.G., Torres, G., 

1020: Kriss, G., and Tonry, J.L. 1992, in Astronomical Data Analysis Software and Systems I, 

1021: ASP Conf. Ser., Vol. 25, eds. D.M. Worral, C. Biemesderfer, and J. Barnes, 432

1022: 

1023: \bibitem[Law et al.(2006)]{law06} Law, N. M., Mackay, C. D., and Baldwin, J. E., 2006, A\&A, 446, 739

1024: 

1025: \bibitem[Lehman et al.(1994)]{lehman94}Lehman, T., Reipurth, B., Brandner, W., 1994, AA, 300, L9

1026: 

1027: \bibitem[Lin and Papaloizou(1986)]{lin86} Lin, D.N.C., \& Papaloizou, J., 1986, \apj, 309, 846 

1028:  

1029: \bibitem[Lodato \& Clarke(2004)]{lodatoclarke04} Lodato, G., Clarke, C., 2004, MNRAS, 353, 841  

1030: 

1031: \bibitem[Mannings \& Emerson(1994)]{mannings94} Mannings, V., Emerson, J., 1994, MNRAS, 267, 361

1032:  

1033: \bibitem[Marschall \& van Altena(1987)]{marschall87} Marschall, L.A. \& van Altena, W.F., 1987, \aj, 94,71

1034: 

1035: \bibitem[Mathieu et al.(1991)]{mathieu91} Mathieu, R., Adams, F., Latham,D., 1991, AJ, 101, 2184;

1036: 

1037: \bibitem[Mathieu et al.(1995)]{mathieu95} Mathieu, R., Adams, F., Fuller, G., Jensen, E.,  Koerner, D., Sargent, A., 1995,  AJ, 109, 2655

1038: 

1039: \bibitem[McNeil et al.(2004)]{mcneil04} McNeil, J., Reipurth, B., Meech, K., 2004, IAUC 8284, 1 

1040: 

1041: \bibitem[Mer\'{\i}n (2004)]{merin04} Mer\'{i}n, B., 2004 PhD Thesis, Universidad Aut\'{o}noma de Madrid 

1042: 

1043: \bibitem[Miyake \& Nakagawa(1993)]{miyakenakagawa93} Miyake, K.,  \& Nakagawa, Y., 1993, ICARUS, 106, 20

1044: 

1045: \bibitem[Montes et al.(1997)]{montes97} Montes, D., Mart\'{\i}n, E., Fern\'{a}ndez-Figueroa, M., Cornide, M, De Castro, E., 1997, A\&AS, 123, 473 

1046: 

1047: \bibitem[Mora et al.(2001)]{mora01} Mora, A.; Mer\'{\i}n, B.; Solano, E.; Montesinos, B.; de Winter, D.; Eiroa, C.; Ferlet, R.; Grady, C. A.; Davies, J. K.; Miranda, L. F.; and 12 coauthors, A\&A 378, 116

1048:  

1049: \bibitem[Morgenroth(1939)]{morgenroth39} Morgenroth, O., 1939, Astron. Nachr. 268, 273

1050: 

1051: \bibitem[M\"{u}ller et al.(2001)]{mueller01} M\"{u}ller, M., Thornwirth, S., Roth, D., \& Winnewisser, G., 2001, A\&A 370, L49

1052: 

1053: \bibitem[M\"{u}ller et al.(2005)]{mueller05} M\"{u}ller, M., Schl\"{o}der, F., Stutski, J., \& Winnewisser, G., 2005, J Mol. Struct. 742, 215 

1054: 

1055: \bibitem[Muzerolle et al.(1998)]{muzerolle98} Muzerolle, J., Hartmann, L., Calvet, N., 1998, AJ 116, 455

1056: 

1057: \bibitem[Muzerolle et al.(2000)]{muzerolle00} Muzerolle, J., Calvet, N., Brice\~{n}o, C., Hartmann, L. \& Hillenbrand, L., 2000, \apj, 535, L47

1058: 

1059: \bibitem[Natta et al.(1997)]{natta97} Natta, A., Grinin, V., Manings, V., \& Ungerechts, H., 1997, ApJ 491, 885 

1060: 

1061: \bibitem[Natta \& Whitney(2000)]{nattawhitney00} Natta, A., Whitney, B., 2000, A\&A, 364, 633 

1062: 

1063: \bibitem[Natta et al.(2000)]{natta00} Natta, A., Grinin, V., Tambovtseva, L., 2000, ApJ, 542, 421 

1064: 

1065: \bibitem[Natta et al.(2004)]{natta04} Natta, A., Testi, L., Muzerolle, J., Randich, S., Comer\'{o}n, F., Persi, P., 2004, A\&A 424, 603

1066: 

1067: \bibitem[Osterloh \& Beckwith(1995)]{osterloh95} Osterloh, M., \& Beckwith, S., 1995, ApJ 439, 288

1068: 

1069: \bibitem[Patel et al.(1998)]{patel98} Patel, N.A.,  Goldsmith, P.F., Heyer, M.H. \& Snell, R.L., 1998, ApJ , 507, 241

1070: 

1071: \bibitem[Petrov et al.(2001)]{petrov01} Petrov, P., Gahm, G., Gameiro, G., Duemmler, R., \& 4 more coauthors, 2001, A\&A, 369, 993 

1072: 

1073: \bibitem[Quanz et al.(2006)]{quanz06} Quanz, S., Henning, Th., Bouwman, J., Ratzka, T., \& Leinert, C., 2006, ApJ, 648, 472

1074: 

1075: \bibitem[Ralchenko et al.(2007)]{ralchenko07} Ralchenko, Yu., Jou, F.-C., Kelleher, D.E., Kramida, A.E., Musgrove, A., Reader, J., Wiese, W.L., and Olsen, K. (2007). NIST Atomic Spectra Database (version 3.1.2); http://physics.nist.gov/asd3. National Institute of Standards and Technology, Gaithersburg, MD

1076: 

1077: \bibitem[Reipurth \& Aspin(2004)]{reipurth04} Reipurth, B., Aspin, C., 2004, ApJ 608, L65 

1078: 

1079: \bibitem[Rodmann et al.(2006)]{rodmann06} Rodmann, J.; Henning, Th.; Chandler, C. J.; Mundy, L. G.; Wilner, D. J., 2006, A\&A, 446, 211

1080: 

1081: \bibitem[Samus et al.(2004)]{samus04} Samus et al., 2004, Combined General Catalogue of Variable Stars,  The GCVS4.2 Catalog (Vol. I-III) 

1082: 

1083: \bibitem[Sch\"{o}ier et al.(2005)]{schoier05} Sch\"{o}ier, F., van der Tak, F., van Dishoeck, E., Black, J., 2005, A\&A432, 369 

1084: 

1085: \bibitem[Shevchenko et al.(1993)]{shevchenko93} Schevchenko, V., Grankin, K., Ibraghimov, M., Melnikov, S., Yakubov, S., 1993, Ap\&SS, 202, 137 

1086: 

1087: \bibitem[Sicilia-Aguilar et al. Paper I(2004)]{sic04} Sicilia-Aguilar, A., Hartmann, L., Brice\~{n}o, C., Muzerolle, J.,Calvet, N., 2004, \aj 128, 805, Paper I

1088: 

1089: \bibitem[Sicilia-Aguilar et al. (2005)]{sic05onc} Sicilia-Aguilar, A., Hartmann, L., Szentgyorgyi, A., Roll, J., Conroy, M., Calvet, N., Fabricant, D., \& Hern\'{a}ndez, J., 2005, AJ, 129, 363 

1090: 

1091: \bibitem[Sicilia-Aguilar et al. Paper II(2005))]{sic05opt} Sicilia-Aguilar, A., Hartmann, L., Hern\'{a}ndez, J., Brice\~{n}o, C., Calvet, N., 2005, \aj 130, 188, Paper II

1092: 

1093: \bibitem[Sicilia-Aguilar et al. Paper III(2006))]{sic05ir} Sicilia-Aguilar, A., Hartmann, L., Calvet, N., Megeath, S.T., Muzerolle, J., Allen, L., D'Alessio, P., Mer\'{\i}n, B., Stauffer, J., Young, E., Lada, C., 2006,ApJ  638, 897, Paper III

1094: 

1095: \bibitem[Sicilia-Aguilar et al. Paper IV(2006))]{sic06echelle} Sicilia-Aguilar, A., Hartmann, L., F\"{u}r\'{e}sz, G., Henning, Th., Dullemond, C., Brandner, W., 2006, AJ 132, 2135. Paper IV

1096: 

1097: \bibitem[Siess et al.(2000)]{sies00} Siess, L., Dufour, E. \&  Forestini, M. 2000 A\&A , 358, 593

1098: 

1099: \bibitem[Suyarkova(1975)]{suyarkova75} Suyarkova, O., 1975, Variable stars 20, 167 

1100: 

1101: \bibitem[Tambovtseva et al.(2001)]{tambovtseva01} Tambovtseva, L., Grinin, V., Rodgers, B., Kozlova, O., 2001, Astronomy Reports, 45, 6, 442

1102: 

1103: \bibitem[Taylor et al. (1993)]{taylor93} Taylor, S., Hartquist, T., Williams, D., 1993, MNRAS, 264, 929

1104: 

1105: \bibitem[Thi et al.(2001)]{thi01} Thi, W., van Dishoeck, E., Blake, G., van Zadelhoff, G., and 7 more authors, 2001, ApJ, 561, 1074;

1106: 

1107: \bibitem[Tonry \& Davis(1979)]{tonry79} Tonry, J., \& Davis, M., 1979, AJ 84, 1511

1108: 

1109: \bibitem[Tubbs et al.(2002)]{tubbs02} Tubbs, R. N., Baldwin , J. E., Mackay, C. D., and Cox, G. C., 2002, A\&A, 387. L21

1110: 

1111: \bibitem[Vittone \& Errico(2005)]{vittoneerrico05} Vittone, A. \& Errico, L., 2005, Mem. S.A.It., 76, 320

1112: 

1113: \bibitem[Welty et al.(1990)]{welty90} Welty, A., Strom, S., Strom, K., Hartmann, L., Kenyon, S., Grasdalen, G., Stauffer, J.,1990, ApJ 349, 328

1114: 

1115: \bibitem[Welty et al.(1992)]{welty92} Welty, A., Strom, S., Edwards, S., Kenyon, S., Hartmann, L., 1992, ApJ 397, 250

1116: 

1117: \end{thebibliography}

1118: 

1119: 

1120: 

1121: \clearpage

1122: 

1123: \input tab1.tex

1124: 

1125: \clearpage

1126: 

1127: \input tab2.tex

1128: 

1129: \input tab3.tex

1130: 

1131: \clearpage

1132: \pagestyle{empty}

1133: \input tab4.tex

1134: \clearpage

1135: \pagestyle{plaintop}

1136: \input tab5.tex

1137: 

1138: \input tab6.tex

1139: 

1140: \input tab7.tex

1141: 

1142: \input tab8.tex

1143: 

1144: \clearpage

1145: 

1146: 

1147: 

1148: \begin{figure}

1149: \plotone{f1.eps}

1150: \caption{Low-resolution spectra of GM Cep. The main emission and absorption

1151: features, as well as the atmospheric H$_2$O and O$_2$ bands, are labeled.\label{lowrespec-fig}}

1152: \end{figure} 

1153: 

1154: \begin{figure}

1155: \plotone{f2.eps}

1156: \caption{AstraLux observations of GM Cep in the SDSS z' filter. The left image has a 

1157: field of view of 2\arcsec$\times$2\arcsec\ and was generated from the best 

1158: 2\% of 15,000 single frames. The effective integration time is 7.5\,s at a 

1159: single frame exposure time of 25\,ms. The middle image shows the inner 

1160: 0\farcs9$\times$0\farcs9. The right image is the theoretical PSF with same 

1161: pixel scale, taking atmospheric dispersion effects into account. 

1162: All images are square root scaled up to saturation. \label{astralux-fig}}

1163: \end{figure} 

1164: 

1165: \begin{figure}

1166: \plottwo{f3a.eps}{f3b.eps}

1167: \caption{Left: $^{12}$CO(1-0) and $^{12}$CO(2-1) spectra. The zero frequencies are

1168: 115.2712 and 230.538 GHz, respectively. A narrow emission is

1169: seen at approximately the cluster velocity only in$^{12}$CO(1-0),

1170: suggeting that the radial velocity of GM Cep is consistent with the radial 

1171: velocity of the cluster. Right: FCRAO CO J=1-0 peak intensity map of the IC1396 region 

1172: near GM Cep (Patel et al. 1998). The position of GM Cep is marked as a ``+''. The  

1173: halftone greyscale levels are linearly scaled from 0.5 to 8 K, with  

1174: contour levels: -3, 3, 5, 10, 15, 20... times the rms of the map (0.5  

1175: K). \label{lines-fig}}

1176: \end{figure} 

1177: 

1178: 

1179: \begin{figure}

1180: \plotone{f4.eps}

1181: \caption{Lightcurve for GM Cep, including the data in Table \ref{opt-table}.

1182: The left pannel shows the complete lightcurve.

1183: The right pannel shows a zoom on the most recent data.

1184: The datapoints have been connected for better visualization, although due to the undersampling,

1185: further magnitude oscillations probably occurred in between the measurements. \label{lightcurve-fig}}

1186: \end{figure} 

1187: 

1188: 

1189: \begin{figure}

1190: \plotone{f5.eps}

1191: \caption{Color evolution of GM Cep, compared to the 1, 10, and 100 Myr isochrones

1192: (from right to left, short-dashed, long-dashed and thin lines; 

1193: Siess et al. 2000) and to the reddening vector (dashed-dotted line marked at intervals

1194: of A$_V$=0.5 mag). The data from September 2000

1195: has been converted from the Cousins into the Johnson system (Fernie 1983).

1196: The individual data points have been connected to show the direction of the variations.

1197: Typical errors are smaller than the dots except for the September 2000 converted data,

1198: which has errors up to $\sim$0.07 and $\sim$0.2 mags in R$_J$ and in I$_J$, respectively.

1199: The isochrones have been reddened by the cluster average (A$_V$=1.67 mag) following

1200: a standard reddening law (Cardelli et al. 1989).

1201:  \label{colorJ-fig}}

1202: \end{figure} 

1203: 

1204: 

1205: \begin{figure}

1206: \plotone{f6.eps}

1207: \caption{Emission lines in GM Cep, from HIRES (solid line)

1208: and from Hectochelle (dotted line) spectra. The zero velocity position has been marked

1209: according to the HIRES radial velocity CZ=-21.0$\pm$2.6 km/s. \label{lines1-fig}}

1210: \end{figure} 

1211: 

1212: 

1213: \begin{figure}

1214: \plotone{f7.eps}

1215: \caption{Emission lines in GM Cep observed with HIRES. The zero velocity position has been marked

1216: according to the HIRES radial velocity CZ=-21.0$\pm$2.6 km/s. The Li I absorption at 6708\AA\ and 

1217: the Paschen 13 H line at 8663\AA\ have been marked by dotted-dashed lines.  \label{lines2-fig}}

1218: \end{figure} 

1219: 

1220: 

1221: 

1222: \begin{figure}

1223: \plotone{f8.eps}

1224: \caption{Rotational velocity (V$sini$) and radial velocity (CZ) of GM Cep versus

1225: wavelength, derived from the HIRES spectra (dots) and from Hectochelle data (crosses).

1226: The average V$sini$ and CZ of the Tr 37 members are displayed as a solid line,

1227: together with the 1$\sigma$ deviations seen in the cluster (dashed lines).

1228: Whereas the radial velocity of GM Cep is consistent with the Tr 37 members within

1229: the errors, the V$sini$ is much higher than any typical CTTS or WTTS. \label{xcor-fig}}

1230: \end{figure} 

1231: 

1232: 

1233: \begin{figure}

1234: \plotone{f9.eps}

1235: \caption{SED data and disk models (D'Alessio et al. 2005) for GM Cep. The fitted data 

1236: points (see text) are represented by filled circles. Open circles mark the rest of data 

1237: points not included in the fitting, and inverted triangles are upper limits. The dotted 

1238: lines trace three different hydrostatic equilibrium disk models for a star with luminosity 

1239: 26 L$_\odot$, accretion rates 10$^{-7}$-10$^{-6}$ M$_\odot$/yr, and disk radii ranging from 

1240: 50 to 600 AU. The short dashed line represent the model that best fits the millimetre point,

1241: consisting of a disk with \.{M}$\sim$10$^{-6}$ M$_\odot$/yr and 30 AU outer radius, and the

1242: bold line is the 30 AU model plus a black body with T=150 K to simulate an excess flaring

1243: or envelope in order to reproduce the high mid-IR fluxes. The long dashed line is the photospheric 

1244: emission for a K0 star (Kenyon \& Hartmann 1995). The data is de-reddened for A$_V$=2.5 mag.

1245: The unability to fit simultaneously the mid-IR (24 and 70$\mu$m) and the millimeter point

1246: with the standard equilibrium models suggests that the disk is unstable and has probably

1247: suffered considerable grain growth. \label{bestfit-fig}}

1248: \end{figure} 

1249: 

1250: \begin{figure}

1251: \plotone{f10.eps}

1252: \caption{Simulations of different binary cases as they would be imaged with AstraLux, 

1253: for different separations and magnitude differences. From top to bottom, and 

1254: left to right: separation 100 mas (90 AU), $\Delta$(mag)=1.20 mag; 

1255: separation 150 mas (135 AU),$\Delta$(mag)=1.92 mag; separation 200 mas

1256: (180 AU), $\Delta$mag)=2.20 mag; separation 250 mas (225 AU), 

1257: $\Delta$(mag)=2.50 mag; separation 300 mas (270 AU), $\Delta$(mag)=3.00 mag; 

1258: separation 400 mas (360 AU),$\Delta$(mag)=3.50 mag. 

1259: All images are displayed on a linear scale up to saturation. The field

1260: of view is 3''$\times$3''. 

1261: \label{simulations-fig}}

1262: \end{figure} 

1263: 

1264: 

1265: \end{document}

1266: 

1267: