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

2: \documentclass[preprint2]{aastex}

3: 

4: 

5: %%Macros

6: \newcommand{\geff}{$g_{\rm eff}$}

7: \newcommand{\vsini}{\mbox{$v\sin i$}}

8: \newcommand{\B}{$\vert {\bf B}\vert$}

9: \newcommand{\kms}{km s$^{-1}$}

10: \newcommand{\Teff}{$T_{\rm eff}$}

11: \newcommand{\logg}{$\log\thinspace g$}

12: \newcommand\vmac{\mbox{$v_{\rm mac}$}}

13: \newcommand\vmic{\mbox{$v_{\rm mic}$}}

14: 

15: 

16: %\slugcomment{Submitted to The Astrophysical Journal}

17: 

18: \def\ltsima{$\; \buildrel < \over \sim \;$}

19: \def\simlt{\lower.5ex\hbox{\ltsima}}

20: \def\gtsima{$\; \buildrel > \over \sim \;$}

21: \def\simgt{\lower.5ex\hbox{\gtsima}}

22: \def\Msun{M$_\odot$}

23: 

24: %\received{2001 October 16}

25: 

26: \begin{document}

27: 

28: \title{The Magnetic Fields of Classical T Tauri Stars}

29: 

30: \author{Christopher M. Johns--Krull\altaffilmark{1}}

31: \affil{Department of Physics \& Astronomy, Rice University, 6100 Main St.

32:        MS-108, Houston, TX 77005}

33: \email{cmj@rice.edu}

34: 

35: \altaffiltext{1}{Visiting Astronomer, Infrared Telescope Facility, operated

36: for NASA by the University of Hawaii}

37: 

38: \begin{abstract} 

39: 

40: We report new magnetic field measurements for 14 classical T Tauri stars 

41: (CTTSs).  We combine these data with one previous field determination in

42: order to compare our observed field strengths with the field strengths

43: predicted by magnetospheric accretion models.

44: We use literature data on the stellar mass, radius, rotation period, and

45: disk accretion rate to predict the field strength that should be present on

46: each of our stars according to these magnetospheric accretion models.  We

47: show that our measured field values do not correlate with the field 

48: strengths predicted by simple magnetospheric accretion theory.  We also use

49: our field strength measurements and literature X-ray luminosity data to test

50: a recent relationship expressing X-ray luminosity as a function of surface 

51: magnetic flux derived from various solar feature and main sequence star 

52: measurements.  We find that the T Tauri stars we have observed have

53: weaker than expected X-ray emission by over an order of magnitude on 

54: average using this relationship.  We suggest the cause for this is actually

55: a result of the very strong fields on these stars which decreases the 

56: efficiency with which gas motions in the photosphere can tangle magnetic

57: flux tubes in the corona.

58: 

59: \end{abstract}

60: 

61: \keywords{accretion, accretion disks ---

62: line: profiles ---

63: stars: atmospheres ---

64: stars: formation ---

65: stars: magnetic fields ---

66: stars: pre--main-sequence ---}

67: 

68: 

69: 

70: \section{Introduction} 

71: 

72: It is now generally accepted that accretion of circumstellar disk material

73: onto the surface of a classical T Tauri star (CTTS) is controlled by a

74: strong stellar magnetic field (e.g. see review by Bouvier et al. 2006).

75: It is within the disks around these low mass pre-main sequence stars that 

76: solar systems similar to our own form.  Understanding the processes through 

77: which young stars interact with and eventually disperse their disks is 

78: critical for understanding the rotational evolution of stars and the formation

79: of planets.  A key question is to understand how young stars can accrete

80: large amounts of disk material with high specific angular momentum, yet

81: maintain rotation rates that are observed to be relatively slow

82: (e.g. Hartmann \& Stauffer 1989, Edwards et al. 1994).  This problem is solved

83: in current magnetospheric accretion models by having the stellar magnetic 

84: field truncate the inner disk, typically near the corotation radius, and 

85: channel the disk material onto the stellar surface, most often at high stellar

86: latitude.  

87: 

88: Support for magnetospheric accretion in CTTSs is significant.  Current models

89: can account for the relatively slow rotation of most CTTSs (Camenzind 1990;

90: K\"onigl 1991; Collier Cameron \& Campbell 1993; Shu et al. 1994; Paatz \& 

91: Camenzind 1996; Long, Romanova, \&

92: Lovelace 2005).  Studies of the spectroscopic and photometric variability of 

93: CTTSs have been interpreted in terms of magnetically controlled accretion 

94: (e.g. Bertout, Basri, \& Bouvier 1988; Bertout et al.  1996; Herbst, 

95: Bailer--Jones, \& Mundt 2001; Alencar, Johns--Krull, \& Basri 2001) with the 

96: magnetic axis inclined to the rotation axis in some cases (e.g. Kenyon et al.

97: 1994, Johns \& Basri 1995, Bouvier et al. 1999, Romanova et al. 2004).  

98: Models of high 

99: resolution Balmer line profiles, computed in the context of magnetospheric 

100: accretion, reproduce many aspects of observed line profiles (Muzerolle, 

101: Calvet, \& Hartmann 1998, Muzerolle et al. 2000).  Such models can also

102: reproduce aspects of the observed \ion{Ca}{2} infrared triplet lines

103: (Azevedo et al. 2006).  T Tauri stars (TTSs) are

104: observed to be strong X-ray sources indicating the presence of strong magnetic

105: fields on their surfaces (see Feigelson \& Montmerle 1999 for a review), and 

106: several TTSs have been observed to have strong surface magnetic fields 

107: (Basri, Marcy, \& Valenti 1992; Guenther et al. 1999; Johns--Krull, Valenti, 

108: \& Koresko 1999b, hereafter Paper I; Johns--Krull, 

109: Valenti, \& Saar 2004, hereafter Paper II; Yang, Johns--Krull, \& Valenti 

110: 2005, hereafter Paper III), and strong magnetic fields have been observed in 

111: the formation region of the \ion{He}{1} emission line at 5876 \AA\ 

112: (Johns--Krull et al.  1999a; Valenti \& Johns--Krull 2004; Symington et al. 

113: 2005), which is believed to be produced in a shock near the stellar surface 

114: as the disk material impacts the star (Beristain, Edwards, \& Kwan 2001).

115: 

116: Despite these successes, open issues remain.  Most current theoretical models

117: assume the stellar field is a magnetic dipole with the magnetic axis aligned

118: with the rotation axis.  However, recent spectropolarimetric measurements

119: show that the fields on TTSs are probably not dipolar, either aligned with the

120: rotation axis or not (Johns--Krull et al. 1999a; Valenti \& Johns--Krull 

121: 2004; Daou, Johns--Krull, \& Valenti 2005; Yang, Johns--Krull, \& Valenti

122: 2006).  On the other hand, it is expected

123: that even for complex magnetic geometries, the dipole component of the field 

124: should dominate at distance from the star where the interaction with the disk 

125: is taking place, so this may not contradict current theory.  In the case of

126: the Sun, the dipole component appears to become dominant at $2.5 R_\odot$ or 

127: closer (e.g. Luhmann et al. 1998).  For expected disk truncation radii of 

128: 3 -- 10 $R_*$ in CTTSs (see below and Table \ref{sample}), this suggests the 

129: dipole component will govern the stellar interaction with the disk.  

130: Additionally, Gregory et al. (2006) show that accretion can occur from a 

131: truncated disk even when the stellar field geometry is quite complex; however,

132: no study has considered the torque balance between a star and its disk in the

133: case of a complex stellar field geometry.  

134: 

135: Magnetospheric accretion models for CTTSs have been developed by a number of

136: investigators.  K\"onigl (1991) first applied the work of Ghosh and Lamb 

137: (1979) to CTTSs, showing that an equilibrium state could exist if the stellar

138: field was strong enough.  The field would truncate the inner disk and the

139: star would spin at the same angular velocity as the disk where it was

140: truncated just inside the corotation radius.  Additional authors have

141: have studied this equilibrium state, utilizing different assumptions regarding

142: the details of how the stellar field couples to the disk.  Paper I

143: examined the models of K\"onigl (1991), Collier Cameron and

144: Campbell (1993), and Shu et al. (1994) and presented equations from each

145: work which predict the stellar magnetic field strength given the stellar 

146: mass ($M_*$), radius ($R_*$), rotation period ($P_{rot}$), and mass accretion

147: rate ($\dot M$).  These three studies each assume the stellar field is a dipole

148: aligned with the rotation axis.  In this simple case, there should be no 

149: photometric or spectroscopic variability associated with accretion assuming

150: the disk is uniform.  As mentioned above, the observations clearly show 

151: substantial accretion variability, and they also show that the fields are

152: probably not aligned dipoles, at least at the stellar surface.  Nevertheless,

153: it may well be that equilibrium models mentioned above give an indication of

154: the mean, or time averaged behavior of a CTTS.  

155: 

156: Several authors have performed

157: numerical simulations of magnetospheric accretion.  Typically, aligned stellar

158: dipole fields are assumed and the simulations examine the instabilities

159: that can result between the star, its magnetosphere, and the disk.  Many

160: studies find that winds and/or jets are produced (e.g. Hayashi, Shibata, \& 

161: Matsumoto 1996; Goodson, Winglee, \& B\"ohm 1997; Miller \& Stone 1997; 

162: Goodson, B\"ohm, \& Winglee 1999; Ferreira, Pelletier \& Appl 2000; Romanova

163: et al. 2005).  These studies usually 

164: examine only a few select cases and generally do not address how the rotation

165: of the star may evolve: it is a fixed parameter of the model.  These 

166: simulations may be testable through variability studies (see Goodson et al. 

167: 1999).   Recently, Long et al. (2005) present numerical simulations, again

168: assuming an aligned dipole field geometry, with the aim of determining if

169: a time averaged equilibrium rotation rate is established.  Despite variability

170: in the accretion and wind flows, these authors find that the star attains

171: an equilibrium rotation rate with a period equal to that of the disk where it

172: is truncated inside, but close to, the corotation radius.  Long et al. 

173: (2005) use the results of their simulations to derive a relationship for

174: the equilibrium rotation period of CTTSs (their equation 12).  This can

175: be solved for the stellar magnetic field strength and results in an equation

176: with the same dependence on the stellar ($M_*$, $R_*$, $P_{rot}$) and

177: accretion parameters ($\dot M$) as found for K\"onigl (1991) and Shu et al.

178: (1994) as given in Paper I.  It should be noted however,

179: that Matt and Pudritz (2004, 2005a) have recently asserted that the spin

180: down torque produced on the star in magnetospheric accretion models may be

181: up to an order of magnitude lower than most previous investigators have

182: calculated.  Matt and Pudritz (2005) suggest that an accretion powered

183: {\it stellar} wind is responsible for spinning down CTTSs.  

184: 

185: Given the theoretical uncertainty involved with magnetospheric accretion, we 

186: can try to use observations to test whether the disk locking scenario 

187: proposed by the equilibrium models actually holds 

188: in CTTSs.  One of the first attempts at this is the work of Stassun et al. 

189: (1999) who find no correlation between rotation period and the presence

190: of an infrared (IR) excess indicative of a circumstellar disk in a sample of

191: 254 stars in Orion.  However, IR excess alone is not a good measure of the 

192: accretion rate.  Muzerolle, Calvet, and Hartmann (2001) note that current 

193: theory predicts a correlation between rotation period and mass accretion rate

194: which they do not observe.  Muzerolle et al. (2001) suggest that variations in 

195: the stellar magnetic field strength from star to star may account for the lack

196: of correlation.  Indeed, Paper I emphasizes that there are

197: several stellar and accretion parameters that enter into the equilibrium

198: relationship, and the stellar magnetic field remains the quantity measured 

199: for the fewest number of CTTSs.  Due primarily to this lack of data,

200: Johns--Krull and Gafford (2002) made the assumption that the field in fact

201: does not vary significantly from star to star and they looked for correlations

202: among the remaining stellar and accretion parameters.  The predictions based

203: on aligned dipole field geometries are not present in the data; however, if

204: the dipole assumption is dropped, Johns--Krull and Gafford showed that the

205: predicted correlations are present in the samples they examined from CTTSs

206: in Taurus.

207: 

208: The success of the Johns--Krull and Gafford (2002) study depends on the

209: constancy of the magnetic field from one TTS to the next.  From a dynamo

210: perspective, this may be a good assumption.  For cool main sequence stars

211: there is a well defined positive correlation between magnetic activity indices

212: (e.g.  \ion{Ca}{2} H\&K emission, X-ray emission) and inverse Rossby number

213: (the convective turnover time divided by the rotation period) with saturation

214: setting at large inverse Rossby number (e.g. Vilhu 1984, Noyes et al. 1984).

215: Johns--Krull, Valenti, and Linsky (2000) showed that due to their long 

216: convective turnover times and relatively short rotation periods, TTSs generally

217: all lie in the saturated portion of this relationship.  On the other hand, most

218: TTSs are fully convective or nearly so, so a solar-like interface dynamo

219: (Durney \& Latour 1978) is likely very inefficient or non-existent in these

220: stars.  Another possibility is a convective/turbulent dynamo such as 

221: originally proposed by Durney, DeYoung, \& Roxburgh (1993).  Recent studies

222: by Chabrier and K\"uker (2006) and Dobler, Stix, and Brandenburg (2006) do 

223: predict that dynamo action in fully convective stars does correlate with 

224: rotation rate.  A final potential contributor to

225: the fields of TTSs is primordial fields entrained in the star during the

226: star formation process (e.g. Tayler 1987, Moss 2003).  However, little

227: is known about how these fields might vary from star to star in a given star

228: formation region.  As a result, additional measurements of the magnetic

229: field strength on a sizeable sample of CTTSs are required to test current

230: magnetospheric accretion models.

231: 

232: To that end, we have been conducting observations of TTSs at the NASA Infrared

233: Telescope Facility (IRTF) to measure the Zeeman broadening of several K

234: band \ion{Ti}{1} lines which are prominent in the spectra of cool stars.

235: The first use of this line to measure a field on a TTS is in Paper I, where

236: we lay out most of the basics of our observation and analysis technique.

237: Details have been expanded upon somewhat in Papers II and III.  In particular,

238: in Paper III, it was shown that errors in effective temperature of $\sim 200$

239: K as well as errors of 0.5 dex in log$g$ result in errors of less than

240: 10\% in the derived mean magnetic field on TTSs.  This is because we are

241: measuring actual Zeeman broadening of the photospheric absorption lines

242: instead of some secondary effect such as a change in line equivalent width.

243: As a result, it is no longer necessary to perform the detailed analysis of

244: optical high resolution spectra to obtain the best estimates of \Teff\ and

245: \logg\ in order to obtain a good magnetic field measurement.  What is needed

246: is a good estimate of \vsini, and all the stars in our sample have reliable

247: measurements from the literature.  

248: Here we focus on CTTSs for which we have the ancillary data

249: required to predict the stellar magnetic field strength from magnetospheric

250: accretion models.  To do that, we use the equations (1-3) given in Paper I 

251: for the studies of K\"onigl (1991), Collier Cameron and Campbell (1993),

252: and Shu et al. (1994), and we use equation (12) of Long et al. (2005).  In

253: order to make the field predictions, we need stellar and accretion parameters.

254: Rotation periods come from Bouvier et al. (1993, 1995), with the exception

255: of TW Hya which comes from Mekkaden (1998).  Spectral type, stellar 

256: radius, and the mass accretion rate for most stars come from Gullbring et al. 

257: (1998).  For DG Tau, DH Tau, and T Tau we take these values from White and 

258: Ghez (2001).  In some cases (e.g. DH Tau), the spectral type is listed as 

259: spanning two subtypes (e.g. M1-2) and here we average the subtypes (e.g. M1.5).

260: For all stars, we use pre-main sequence evolutionary tracks of Siess, Dufour,

261: and Forestini (2000) to derive the stellar mass.

262: Our sample of stars is given in Table \ref{sample}.  Included in Table 

263: \ref{sample} is the observed X-ray luminosity for each star.  These come

264: from Neuhauser et al. (1995) except for DF Tau, DH Tau, and GM Aur which 

265: come from Damiani et al. (1995); and for TW Hya which comes from Kastner et

266: al. (2002).  The goal of this paper is to measure the mean magnetic field 

267: strength for all the stars in Table \ref{sample} so that the observations 

268: can be compared with the theoretical predictions.  Some of the stars in our

269: sample are known close binaries; however, in these cases one member of the

270: binary dominates the optical and IR light, and it is this member to which 

271: the various stellar, accretion, and magnetic parameters apply.

272: Included in our sample

273: is TW Hya which we do not analyze here, but was analyzed in Paper III.  In 

274: \S 2 we describe our observations and data reduction.  The magnetic field 

275: analysis is described in \S3.  In \S 4 we give our results and we compare 

276: our measured and the predicted field strengths.  We also compare our measured

277: field strengths and the observed X-ray luminosity of our targets to 

278: theoretical predictions based on main sequence stars.  A discussion of our 

279: results is given in \S 5.

280: 

281: \section{Observations and Data Reduction}

282: 

283: All spectra were obtained with the CSHELL spectrometer (Tokunaga et

284: al.  1990, Greene et al.\ 1993) on the 3.0 m telescope at the NASA IRTF

285: on Mauna Kea in Hawaii. Observations occurred during four observing runs: 

286: I) 10--12 December 1996, II) 14--20 December 1997, III) 7--12 January 2000,

287: and IV) 23--28 November 2004 (see Table \ref{obs}).  On the first run a

288: 1\farcs0 slit gave a FWHM of $\sim 4.7$ pixels on the $256 \times 256$ InSb

289: array detector, corresponding to a spectral resolving power of 

290: $R \equiv \lambda /\Delta\lambda \sim 21,500$.  For the second

291: and fourth runs, a 0\farcs5 slit yielded a FWHM of $\sim 2.8$ pixels on

292: the detector giving $R \sim 36,000$. For the third observing run, the same 

293: 0\farcs5 slit yielded a spectral resolving

294: power of only $R \sim 27,000$. [Soon after run III, CSHELL was serviced

295: to restore the spectral resolution.] CSHELL uses a continuously variable

296: filter (CVF) to isolate individual orders of the echelle grating. Each

297: star was observed in three settings, each covering $\sim 0.0057$

298: \micron. The first setting (1) contains two strong \ion{Ti}{1} lines at

299: 2.2211 and 2.2233 \micron. The second (2) contains two more strong

300: \ion{Ti}{1} lines at 2.2274 and 2.2311 \micron. The third setting (3)

301: centered at 2.3120 \micron\ contains $\sim 9$ CO lines from the $v =

302: 2-0$ first overtone band. 

303: 

304: Each star was observed at two positions along the slit separated by

305: 10\arcsec.  Multiple pairs of exposures were made of each CTTS.  Total

306: exposure time varied some from object to object and wavelength setting to

307: wavelength setting.  The minimum exposure time was 20 minutes and the maximum

308: was 1.5 hours.  A typical total exposure time was 1 hour.  Taking the 

309: difference of each image pair with the star moved along the slit removed 

310: detector bias, dark current, and the average background due to night sky

311: emission. Difference images were flat fielded using a normalized

312: spectrum of an internal continuum lamp that fully illuminated the

313: slit.  Stellar spectra were extracted following the procedure described

314: in Paper I, which includes the determination of an

315: oversampled slit function and optimal extraction of the spectrum.

316: Wavelengths were determined by fitting $n\lambda$ as a function of line

317: position for 7--10 lamp emission lines and then dividing by the order

318: number, $n$, of the desired spectral region. Calibration lines from several

319: orders were obtained by changing the order-sorting CVF while keeping

320: the grating fixed.  In addition to CTTSs, each night we observed hot, rapidly

321: rotating stars spanning a range of airmasses.  Using these data, all spectra 

322: shown in this paper have been corrected for telluric absorption.

323: For each CTTS in our sample, Table \ref{obs} gives a log of when

324: the star was observed and which wavelength settings were covered.

325: 

326: \section{Analysis}

327: 

328:      The most successful approach for measuring fields on late--type 

329: stars in general has been to measure Zeeman broadening of spectral lines in

330: unpolarized light (e.g., Robinson 1980; Saar 1988; Valenti, Marcy, \& Basri

331: 1995; Johns--Krull \& Valenti 1996; Papers I, II, \& III).  In the presence

332: of a magnetic field, a given spectral line will split up into a number

333: of components depending the atomic structure of the levels contributing to

334: the line.  Taking into account rotational and turbulent broadening, a

335: strong magnetic field typically produces a change in the shape of magnetically

336: sensitive line profiles.

337: There are several \ion{Ti}{1} lines in the K band which are excellent probes

338: of magnetic fields in late-type stars (e.g. Saar \& Linsky 1985), and here 

339: we observe 4 of them (see \S 2).  In addition to the 4 strongly Zeeman

340: sensitive \ion{Ti}{1} lines, our wavelength settings also cover a relatively

341: weak, but detectable, \ion{Sc}{1} line at 2.2265 $\mu$m which has 

342: reduced Zeeman sensitivity due to lower Land\'e-$g$ values.  We also

343: observe several magnetically insensitive CO lines near 2.313 $\mu$m which

344: serve as a check on our values of \vsini\ and turbulent broadening terms.

345: Since we are observing CTTSs, there is some concern that these lines could

346: form in the disk and trace disk kinematics instead of stellar properties.

347: This issue was studied for CO by Casali and Eiroa (1996) and Johns--Krull

348: and Valenti (2001).  Both studies find that the K band CO lines of the

349: vast majority of CTTSs arise in the stellar photosphere, and the CO lines

350: observed below are fully consistent with formation on the star.  Based on

351: their behavior as a function of effective temperature, the \ion{Ti}{1} lines

352: in the K band appear to trace similar temperature material as that

353: traced by the CO (Wallace \& Hinkle 1997).  In addition, the limited 

354: studies of TTS which show no K band excess (Johns--Krull et al. 2004,

355: Yang et al. 2005) reveal \ion{Ti}{1} and CO line profiles with the same 

356: qualitative behavior as that found below: CO lines well matched by 

357: rotationally broadened stellar photosphere models and excess

358: broadening in the magnetically sensitive \ion{Ti}{1} lines.  We therefore

359: assume that the line profiles are dominated by the stellar photosphere.

360: 

361: The goal in this paper is to measure the magnetic field on a sample of CTTSs

362: by modeling the profiles of their K band photospheric absorption 

363: lines.  Our spectrum synthesis code and detailed analysis technique for 

364: measuring magnetic fields on TTSs is described in Papers I -- III.  Here,

365: we simply review some of the specific details relevant to the results shown

366: here.  In order to synthesize the stellar spectrum, we must first specify

367: the atmospheric parameters: effective temperature (\Teff), gravity (\logg), 

368: metallicity ([M/H]), microturbulence (\vmic), macroturbulence (\vmac), and

369: rotational velocity (\vsini).  Following Papers I --- III,

370: a fixed value of 2 \kms\ is adopted for \vmac.  Valenti, Piskunov,

371: \& Johns--Krull 1998 found that microturbulence and macroturbulence 

372: were degenerate in M dwarfs, even with very high quality spectra.  Therefore,

373: microturbulence is neglected here,

374: allowing \vmac\ to be a proxy for all turbulent broadening.

375: Solar metallicity is assumed for all stars.

376: We use \vsini\ values measured using

377: the CO spectra presented here by Johns--Krull and Valenti (2001) for all

378: the target stars analyzed in that paper.  We adopt

379: \vsini\ values from Hartmann et al. (1986) for CY Tau and GM Aur, and we

380: adopt the \vsini\ value of Basri and Batalha (1990) for DG Tau.

381: The \Teff\ is based on spectral type using the calibration of Johnson

382: (1966) and \logg\ is estimated by placing objects in the HR diagram.

383: This typically produces uncertainties of $< 200$ K in \Teff\ and $< 0.5$

384: dex in \logg\ which translate into uncertaities of 10\% or less in the

385: derived magnetic field values (Paper III).

386: Once we have estimates of \Teff\ and \logg, we use the

387: ``next generation" (NextGen) model atmosphere (Allard \& Hauschildt 1995)

388: from the published grid which is closest to these values.  In the range of 

389: interest here, the NextGen models are on a grid every 100 K (e.g. 3900, 4000,

390: 4100) in \Teff\ and every 0.5 dex (e.g. 3.5, 4.0, 4.5) in \logg.  For BP Tau, 

391: we use the NextGen model with \Teff$=4100$ K which is closest in temperature 

392: to the value we derived in Paper I (its spectral type of K7 suggests 

393: \Teff$=4000$ K).

394: 

395: Lastly, we solve for the K band veiling in each star 

396: along with the stellar magnetic field properties.

397: We initially decided to fit for a single veiling parameter which would apply to

398: all three spectral regions we observe in each star.  This turned out to not

399: be practical, as a few stars could not be well fit by a single veiling 

400: value.  The different wavelength regions for our stars

401: were observed with a time interval between settings of several hours up to

402: a couple of days.  Classical TTSs regularly show significant variations in 

403: their K band flux on timescales as short as a day (and occasionally shorter),

404: likely as a result of accretion variability (Skrutskie et al. 1996, Eiroa et

405: al. 2002).  We therefore decided to let the veiling be a free parameter for

406: each star in each of the three wavelength settings analyzed here.

407: 

408: The free parameters of our fit to the observed spectra are then the value

409: of the K band veiling in each wavelength region (this is the only free

410: parameters in the CO wavelength region) and the magnetic field properties

411: of the star.  It was found in Papers I -- III that the spectra of the TTSs

412: studied there could not be fit with a single value of the magnetic field

413: strength.  Instead, a distribution of magnetic field strengths is required.

414: It was also found that fits to the spectra are degenerate in the derived

415: field values unless we limit the fit to specific values of the magnetic

416: field strength, separated by about 2 kG, which is the approximate ``magnetic

417: resolution" of our data.  Therefore, we use the same limitations when

418: fitting the data presented here.  We assume the star is composed of regions

419: of 0, 2, 4, and 6 kG magnetic field, and we solve for the filling factor

420: of each of these components.  The different regions are assumed to be 

421: well mixed over the surface of the star -- different components are not

422: divided up into well defined spots or other surface features.  Another key

423: assumption (which will be discussed further in \S 5) is that the temperature

424: structure in all the field regions is assumed to be identical for the

425: purpose of spectrum synthesis:  again, the fields are not confined to cool

426: spots or hot plage--like regions.  In total then, we fit for 6 free parameters:

427: the 3 veiling values and the filling factor of the 3 non-zero field

428: regions (the sum of the filling factors must be 1.0).  In Figures \ref{highb}

429: and \ref{lowb} we show our spectral fits to two of our CTTSs, the first

430: (DK Tau) with a high average surface field, and the second (DE Tau) with a

431: low average surface field.

432: We characterize the field on all our CTTSs by computing the mean

433: magnetic field from our fits: $\bar B = \Sigma B_i f_i$ where $B_i$ is the

434: value of the field (0, 2, 4, 6 kG) in each fitted region and $f_i$ is the

435: filling factor of these field components.

436: The formal uncertainty in the fit to all our stars is quite small.  The

437: true uncertainty is dominated by uncertainties in the field resulting from

438: errors in our adopted temperature and gravity, which as discussed above

439: are $\sim 10$\%.  We therefore adopt this value for the uncertainty in our

440: mean magnetic field for each star.  

441: 

442: While DK Tau's \ion{Ti}{1} lines show obvious broadening relative to the

443: nonmagnetic model in Figure \ref{highb}, the situation is a little less clear

444: for DE Tau in Figure \ref{lowb}.  Blowing up the plot, it is apparent that

445: all the \ion{Ti}{1} lines are broader than the null field model which is a

446: good match to the width of the CO lines.  This extra broadening in the

447: \ion{Ti}{1} lines can be quantified by a cross correlation analysis as well.

448: We construct a kernel which is the zero field model but with a \vsini\ $=1.0$

449: \kms\ instead of 7 \kms\ which is the measured value for DE Tau.  We then

450: cross-correlate this kernel with the observed DE Tau spectrum and fit the

451: result with a Gaussian profile to determine the FWHM of the cross correlation

452: function.  Doing this on the CO portion of the spectrum (lower panel of

453: Figure \ref{lowb}), we find a FWHM$ = 6.35$ pixels.  Cross correlating the

454: 1.0 \kms\ kernel with the actual fit to the CO lines (dashed line of Figure

455: \ref{lowb} where \vsini$=7.0$ \kms), we find the FWHM$ = 6.54$ pixels (which

456: is slightly greater - by 0.19 pixels - than for the observed spectrum).  We

457: can estimate the range of FWHM we should measure by taking the fit to the

458: spectrum shown in the dashed line and adding random noise to it to simulate

459: the signal-to-noise ratio (S/N) of our observed spectrum (115 for DE Tau).

460: We then cross correlate this synthetic data with the same kernel above and

461: repeat this experiment 100 times.  Doing so, we recover the a mean FWHM of

462: 6.54 pixels and find a standard deviation in our recovered FWHM values of

463: 0.09 pixels.  Thus, the FWHM we recover from the observations of the 

464: magnetically insensitive CO lines is only 2.1$\sigma$ {\it less} than that 

465: recovered from the CO lines of the null field model for DE Tau.  When we 

466: perform the same analysis on the magnetically sensitive \ion{Ti}{1} lines,

467: we find that the FWHM of the cross correlation function using the observations

468: is 9.97 pixels, whereas the null field model gives a FWHM$ = 8.66$ pixels

469: with a standard deviation of 0.13 pixels.  Thus the observed magnetically

470: sensitive \ion{Ti}{1} lines give a width 10.1$\sigma$ {\it broader} than does

471: the null field model.  Using the actual magnetic fit to DE Tau and again

472: adding 100 different realizations of random noise to it to create synthetic

473: data for the cross correlation analysis yields a FWHM$ = 10.12 \pm 0.17$

474: pixels, which is within 0.9$\sigma$ of that recovered from the observed

475: spectrum.  We are therefore confident in the excess width of the \ion{Ti}{1}

476: lines in DE Tau ($\bar B = 1.12$ kG) and in the other ``low" field measurement

477: stars (see below).  The analysis of Paper III shows that even for

478: these low mean field values, the uncertainty due to temperature and gravity

479: errors again lead to a mean field uncertainty of $\sim 10$\%, which is the

480: field uncertainty adopted here.

481: 

482: \section{Results}

483: 

484: The results of our spectral fitting are given in Table \ref{results}.  

485: Reported here are the derived K band veiling, $r_K$, the surface averaged

486: magnetic field, $\bar B$, for each star, and a prediction for the X-ray

487: luminosity based on the measured field values (discussed below).  One

488: star in our sample, GM Aur, was observed in only one wavelength setting:

489: the \ion{Ti}{1} setting (2) described above.  Results for this star are

490: somewhat less firm that for the others.

491: 

492: \subsection{K Band Veiling}

493: 

494: We determine 3 estimates of the K band veiling ($r_K$) for each star (one 

495: from each wavelength region) which we can compare against each other and 

496: against previous veiling estimates.

497: In Table \ref{results} we tabulate the mean value of $r_K$ determined

498: from the 3 settings and give as an uncertainty estimate the standard deviation

499: of the 3 $r_K$ values.  These veiling estimates generally agree with previous

500: studies (Folha \& Emerson 1999, Johns--Krull \& Valenti 2001).

501: We note though that by solving for the veiling in each

502: wavelength region separately, our magnetic field results are free of systematic

503: effects caused by differences in line strength between the actual star and

504: our spectrum synthesis.  Such differences are adjusted for by changes in the

505: derived $r_K$ values.  It is the {\it shape} of the \ion{Ti}{1} lines that

506: constrains the magnetic field vales we derive.  In support of this idea, we

507: looked for a correlation between the veiling values found here and the mean

508: fields measured.  We also looked for a correlation between the difference in

509: the veiling found here compared to in Johns--Krull and Valenti (2001) and

510: the mean field values.  No correlation is found in either case.

511: 

512: For one star, T Tau, our derived $r_K$ is substantially different than found

513: in previous studies.  At K0, T Tau is the earliest

514: spectral type star in our study, and one of the earliest spectral type

515: CTTS known.  The veiling of Johns--Krull and Valenti (2001) is derived

516: based on a later spectral type template star.  Similarly,

517: Folha and Emerson (1999) report a veiling of $r_K = 2.5$ based on comparison

518: with a later type template.  The \ion{Ti}{1} and CO lines used here get much

519: weaker in the observed spectra of dwarfs and giants as you go from K7

520: to K0 (e.g. Wallace \& Hinkle 1997).  This difference in intrinsic line

521: strength will artificially enhance the derived veiling for a K0 star when

522: using a later template.  The veiling we derive here is quite low, which 

523: might be the result of the line strength in our synthetic spectra being

524: too low as a result of some error in the line data or the model 

525: atmosphere.  To test this, we compared our model spectra for T Tau

526: to observed spectra of the K0 dwarf, HR 166.  This star was observed in

527: setting (2) during the 4th observing run (see \S 2).  Valenti and Fischer

528: (2005) report \Teff$ = 5221$ K and [M/H] $ = 0.16$ for this K0V star.

529: We observe the lines in HR 166 to be quite weak, and the 2.2311 $\mu$m 

530: \ion{Ti}{1} line is strongly affected by telluric absorption, so we ignore 

531: it here.  We measure an equivalent width of $W_{eq} = 49.4 \pm 4.8$ m\AA\ for

532: the 2.2274 $\mu$m based on the 4 spectra observed of HR 166.  Our 0 kG

533: synthetic spectrum for T Tau gives an equivalent width of $W_{eq} = 58.0$

534: m\AA, in good agreement with our observed value.  The fact that the 

535: synthetic spectrum has a larger equivalent width means that we may

536: actually be over estimating the veiling, not underestimating it; however,

537: the correction is small ($\sim 0.2$).  Therefore, we suggest our veiling

538: measurement for T Tau is reliable, and that previous estimates are too

539: large as a result of the later type template stars used for comparison.

540: 

541: \subsection{Magnetic Field Properties}

542: 

543: \subsubsection{Comparison with Magnetospheric Accretion Theory}

544: 

545: Table \ref{results} gives the mean magnetic field values, $\bar B$, for each

546: star analyzed here.  These can be compared directly with the predicted field 

547: strengths given in Table \ref{sample}.  Figure \ref{magcomp} compares the 

548: values of $\bar B$ we measure here with the predictions of magnetospheric 

549: accretion theory.  As mentioned above, all 4 field predictions given in Table 

550: \ref{sample} are very well correlated with each other.  From the standpoint

551: of looking for a correlation between the observed and predicted fields,

552: it does not matter which set of theoretical predictions we use.  Figure

553: \ref{magcomp} plots the predictions based on the work of Shu et al. (1994).

554: As is quite evident from a visual inspection of the figure, there does not

555: appear to be a correlation between the observed and predicted field values.

556: We can quantify this by computing the linear correlation coefficient (also

557: called Pearson's $r$ value) and its significance (e.g. Press et al. 1986).

558: Computing this for the data in Figure \ref{magcomp}, we find a correlation

559: coefficient of $r = 0.08$ which has an associated false alarm probability

560: of 79\%, indicating no correlation in the data at all.  While there is no

561: apparent correlation in Figure \ref{magcomp}, the good news is that the

562: observed fields generally lie on or above the line indicating equality.

563: Thus, the measured fields on these CTTSs are approximately the right magnitude

564: for magnetospheric accretion to work.  We will return to this point further

565: in \S 5 below.

566: 

567: When examining Tables \ref{sample} and \ref{results} and Figure \ref{magcomp},

568: worry may be aroused by the specific choice of mass accretion rates used.  It

569: is well known that mass accretion rates for CTTSs are difficult to measure,

570: and estimates for specific stars can vary by an order of magnitude or

571: more (see discussion in Gullbring et al. 1998).  Fortunately, the predicted

572: field strengths vary only as the square root of the mass accretion rate

573: (Johns--Krull et al. 1999b),

574: making them somewhat less sensitive to the problems associated with measuring

575: the accretion rate.  In addition, different studies generally do find good

576: correlation of their measured accretion rates from star to star, even if

577: they differ on the overall magnitude (again, see Gullbring et al. 1998).

578: Therefore, we have some confidence that the variation of the predicted

579: fields from star to star can be reasonably well estimated.  In an effort to

580: see how the results presented here depend on the stellar and accretion

581: parameters used for the stars studied, Table \ref{results2} gives the

582: mass accretion rates and predicted fields for 5 relatively large studies

583: which contain many of the stars observed here.  Since the derived mass

584: accretion rates depend to some extent on the mass and radius assigned to

585: the star, in Table \ref{results2} the predicted fields are calculated

586: using the mass and radius adopted for each star in the relevant study.

587: The studies presented in the table are Valenti et al. (1993, labelled VBJ),

588: Hartigan et al. (1995, labelled HEG), Gullbring et al. (1998, labelled

589: GHBC), Calvet and Gullbring (1998, labelled CG), and White and Ghez

590: (2001, labelled WG).  While accretion rates can vary by over an order of

591: magnitude bewteen these different studies, there is generally good correlation

592: in the predicted fields based on the different studies, particularly when

593: allowing removal of a single obviously discrepant star.  For example, 

594: comparing field predictions from HEG with those of GHBC produces a 

595: correlation ($r = 0.73$) with a false alarm probability of 1\% even though

596: the accretion rates differ by an order of magnitude in many cases.  Comparing

597: field predictions from VBJ with GHBC gives a correlation ($r = 0.74$) with

598: a false alarm probability of 2\% (due to the smaller number of stars in

599: common), and when removing the obviously discrepant GI Tau, the correlation

600: improves to $r = 0.99$ with a false alarm probability of $2.7 \times 10^{-6}$.

601: 

602: On the other hand, none

603: of these studies produced a good correlation with the observed mean field

604: strengths.  The bottom line of the table gives the correlation

605: coefficient and associated false alarm probability when comparing the

606: predicted fields from each study with the observed fields.

607: Plots of the observed versus predicted fields appear similar to

608: Figure \ref{magcomp}, showing a flat relation with no apparent correlation

609: even when the most discrepant point is ignored.

610: While, mass accretion rate is the parameter that varies the

611: most for each star from one study to the next, the predicted fields turn

612: out to be very sensitive to the adopted value of the stellar radius.  The

613: predicted field strengths vary as $R_*^3$ (Johns--Krull et al. 1999b),

614: which is usually determined by measuring the stellar luminosity and

615: effective temperature.  Effective temperatures usually come from spectral

616: type and an adopted spectral type -- effective temperature calibration;

617: however, this has a number of its own problems (e.g. Padgett 1996,

618: Huerta 2007) and is deserving of more attention.

619: 

620: \subsubsection{Relationship to X-ray Emission}

621: 

622: Stellar magnetic fields are also believed to give rise to ``activity" in

623: late-type stars.  Activity is typically traced by line emission or broad

624: band emission at high energy wavelength such as X-rays.  Pevtsov et al.

625: (2003) study the relationship between magnetic flux and X-ray luminosity

626: on the Sun and cool stars, generally finding excellent correlation between

627: these two quantities ranging over almost 11 orders of magnitude in each

628: quantity.  Pevtsov et al. (2003) include preliminary measurements for

629: 6 of the CTTSs shown here taken from Johns--Krull and Valenti (2000) and

630: Johns--Krull et al. (2001).  We use the X-ray luminosity--magnetic flux

631: relationship of Pevtsov et al. (2003) with our magnetic field measurements

632: and the stellar radii given in Table \ref{sample} to calculate the expected

633: X-ray luminosity for each of our stars.  These predicted X-ray values are

634: given in the last column of Table \ref{results}.  Figure \ref{xcomp} plots

635: the measured X-ray luminosities from Table \ref{sample} versus these

636: predicted X-ray values.  All but one of the CTTSs in our sample show lower

637: X-ray emission than would be predicted by their magnetic field properties,

638: and many of them are low by more than an order of magnitude.  Only TW Hya

639: shows a higher level of measured X-ray emission relative to the prediction,

640: and perhaps significantly, this CTTSs is the star closest to the main

641: sequence in our sample due to its age of $\sim 10$ Myr.  We discuss this 

642: issue further in the next section.

643: 

644: \subsubsection{Field Measurements and Pressure Equipartition}

645: 

646: From the perspective of cool star research, the high apparent field

647: strengths on these CTTSs is a little surprising.  Spruit \& Zweibel (1979)

648: computed flux tube equilibrium models, showing that magnetic field

649: strength is expected to scale with gas pressure in the surrounding

650: non-magnetic photosphere.  Similar results were found by Rajaguru, Kurucz,

651: and Hasan (2002) who find that convective collapse of thin magnetic flux

652: tubes will set a field strength limit of $\sim 750$ G in the photosphere of

653: a star with \Teff$ = 4000$ K and \logg$ = 3.5$.  Convective collapse is a 

654: mechanism for forming partially evacuated flux tubes, and Rajaguru et al. 

655: (2002) find evidence that this mechanism may not operate effectively 

656: at low \Teff; however, they find a maximum field strength of $\sim 1300$ G if

657: the field is to remain in pressure equilibrium with the surrounding 

658: unmagnetized photosphere.   Field strengths set by such pressure equipartition 

659: considerations appear to be observed in G and K dwarfs (e.g. Saar 1990, 1994, 

660: 1996) and possibly in M dwarfs (e.g. Johns--Krull \& Valenti 1996).  

661: 

662: TTSs have relatively low surface gravities and hence low photospheric gas 

663: pressures, so that equipartition flux tubes would have relatively low 

664: magnetic field strengths compared to cool dwarfs.  Indeed, Safier (1999) 

665: examined in some detail the ability of TTS photospheres to confine magnetic

666: flux tubes via pressure balance with the surrounding quiet photosphere, 

667: concluding that the maximum field strength allowable on TTSs is substantially

668: below the few detections reported at that time.  The maximum field strength

669: allowed for a confined magnetic flux tube is $B_{eq} = (8 \pi P_g)^{1/2}$

670: where $P_g$ is the gas pressure at the observed level in the stellar

671: atmosphere.  Here, we take as a lower limit (upper limit in pressure) the

672: level in the atmosphere where the local temperature is equal to the effective

673: temperature.  This is the approximate level at which the continuum forms,

674: with the \ion{Ti}{1} lines forming over a range of atmospheric layers

675: above this level.  Our computed values of $B_{eq}$ are given in Table 

676: \ref{results}.  In Figure \ref{eqcomp} we plot our observed mean magnetic

677: field strengths versus these $B_{eq}$ values for our entire sample.  

678: In all cases, the observed mean magnetic field is greater than the field

679: predicted by pressure equipartition arguments.  In many cases the observed

680: mean field exceeds $B_{eq}$ by a factor of more than 2.

681: 

682: \subsubsection{Origins of the Strong Fields}

683: 

684: Lastly, we compare our mean field measurements with various stellar 

685: parameters in an effort to uncover the origin of the strong magnetic 

686: fields observed on these stars.  We compare both the mean magnetic

687: field strength, $\bar B$, and the total magnetic flux, $\Phi$, to a number

688: of stellar parameters, including those thought to be important in the

689: dynamo generation of magnetic fields.  The results of these comparisons are

690: summarized in Table \ref{stelpar}.  Given in this table is the Pearson linear

691: correlation coefficient, $r$, and its false alarm probability, $f_p$,

692: which measures the degree to which the quantities being compared are

693: correlated with one another.  As Table \ref{stelpar} shows, there is no

694: significant correlation of $\bar B$ or $\Phi$ with any stellar parameter

695: or common dynamo variable, including the convective turnover time, $\tau_c$

696: [values of $\tau_c$ were kindly provided by Y.-C. Kim and were computed

697: using models described in Kim and Demarque (1996) with input physics 

698: described in Preibisch et al. (2005)].

699: The most significant correlation found is $\Phi$ versus $L_*$; however,

700: we must be careful with such a correlation due to the common role the

701: stellar radius plays in both variables.  The magnetic flux is $\Phi = \bar B 

702: R_*^2$, and the stellar radius is calculated from the measured $L_*$ and

703: $T_{eff}$.  Thus, we expect to find a correlation, and when comparing

704: $R_*$ versus $L_*$ we find $r = 0.73$ and $f_p = 0.002$.  This correlation

705: is stronger than that found using $\Phi$, suggesting the correlation of

706: $\Phi$ with $L_*$ is driven entirely by the stellar radius, with the

707: magnetic field playing no role.  Interestingly, the next most significant

708: correlation is $\Phi$ versus $\tau_c$; however, this correlation is again 

709: likely driven by the correlation of $R_*$ with $\tau_c$ ($r = -0.63, f_p = 

710: 0.01$), with the magnetic field playing no role.  Within this sample of

711: 15 CTTSs, we then find no correlation of the measured magnetic field 

712: values with any stellar or dynamo parameters.

713: 

714: \section{Discussion}

715: 

716: We provide new measurements of the mean magnetic field strength on 14

717: CTTSs.  The detected fields range from 1.12 to 2.90 kG.  These measurements

718: combined with previous reports of magnetic fields on TTSs (Basri et al.

719: 1992; Guenther et al. 1999; Johns--Krull et al. 1999b, 2004; Yang et al.

720: 2005) suggest that the strong majority of TTSs are covered by kilogauss

721: magnetic fields.  As found in Papers I - III, we use a distribution of magnetic

722: field strengths to fit the observed K band \ion{Ti}{1} profiles.  We

723: model this distribution as if it were a spatial distribution on the stellar

724: surface: the model assumes distinct regions with different field 

725: strengths.  A concern is that the distribution is in fact a distribution

726: with depth in the magnetic regions.  If this were the case, we would be

727: overestimating the total filling factor of magnetic regions as well as the

728: mean field strength.  Observations of sunspots in the K band show that the

729: \ion{Ti}{1} lines in this spectral region display profiles characteristic of

730: a single magnetic field value (e.g. Wallace \& Livingston 1992), despite the 

731: likely variation of the field with depth in a sunspot (e.g. Mathew et al.

732: 2003).  Therefore, we believe the field distributions found in Papers I - III

733: and used here represent actual spatial distributions of field strengths on 

734: the stellar surface.

735: 

736: The observed mean field values for the CTTSs are all larger than the field

737: strength for these stars predicted by pressure equilibrium arguments (Figure

738: \ref{eqcomp}).  This suggests that the magnetic pressure dominates the gas

739: pressure, in some cases by a significant amount, in the photospheres of these

740: stars.  If this is the case, there should be no field free regions on these

741: stars, though we do use a 0 kG field component when fitting their spectra.

742: This issue is discussed in \S 4.2.3 and in Papers I and II.

743: In these papers, it was found that using fits with 0, 2, 4, and 6 kG

744: components gave essentially identical mean field measurements as those

745: from fits using 1, 3, 5, and 7 kG components.  Thus, the data do not demand

746: a field free region, and such a region may not exist on these stars.

747: With strong magnetic field essentially covering the entire surface of

748: TTSs, we expect to see some effect of the fields on the atmosphere, perhaps

749: analogous to the fields of sunspots inhibiting convection (e.g. Foukal 1990).

750: Such an effect might be diagnosed through the study of line bisectors

751: (e.g. Gray 1982); however, little work in this area has been done for

752: late K spectral types, so the intrinsic bisector shape for stars of this

753: spectral type is unknown.  

754: 

755: On the other hand, we may have evidence for the

756: feedback of these strong fields on the star through the X-ray activity

757: displayed by TTSs.  Feigelson et al. (2003),  find that X-ray emission in

758: TTSs saturates on average at a level approximately an order of magnitude

759: below that displayed by main sequence stars.  In addition to this apparent

760: global reduction in X-ray emission of TTSs, observations also reveal that

761: accreting young stars as a group show X-ray emission levels that are reduced

762: by a factor of 2--3 relative to non-accreting young stars, even after 

763: dependencies on mass and age are taken into account (Flaccomio et al.

764: 2003, Stassun et al. 2004, Preibisch et al. 2005, Feigelson et al. 2007).  

765: At least two explanations have been advanced that can explain this difference

766: between stars with and without accretion signatures.  Preibisch et al.

767: (2005) suggest that the higher density of mass loaded accreting field

768: lines prevents these portions of CTTSs magnetospheres from heating up to

769: temperatures required for X-ray emission.  Jardine et al. (2006) present a

770: model of TTS cornonae with realistic levels of complexity and show that 

771: this can produce a similar range of X-ray emission levels as that which

772: is observed in for example, the Chandra Orion Ultradeep Project (COUP). 

773: Jardine et al. (2006) suggest that the presence of a disk,

774: particularly in lower mass CTTSs, acts to strip the outer parts of the stellar

775: corona, thereby reducing the total X-ray emission.  This naturally predicts

776: an anticorrelation between X-ray emission and the presence of a disk,

777: at least in the lower mass young stars such as those studied here.  While 

778: there is no observed correlation of X-ray emission with the presence or 

779: absence of an inner disk as diagnosed by K-band emission for the COUP

780: dataset (e.g. Feigelson et

781: al.  2007), this coronal stripping may be related to the action of accretion.

782: While these ideas may explain the factor of 2--3 difference in accreting 

783: versus non-accreting young stars, they do not explain the order of magnitude

784: reduction in X-ray emission for TTSs as a whole noted by Feigelson et al. 

785: (2003).  Here, we find that the observed

786: X-ray emission from our sample of CTTSs is substantially below what is 

787: expected given their strong magnetic field values (Figure \ref{xcomp}),

788: often by more than an order of magnitude.

789: While the origin of coronal heating in the Sun and solar like stars is still

790: debated, most theories assume it is due in part to convective motions 

791: moving and buffeting photospheric magnetic flux tubes (e.g. Fisher et al. 

792: 1998, Foukal 1990).  We suggest that the strong magnetic fields which appear

793: to cover the surface of TTSs reduce the efficiency with which convective 

794: motions can stress these stellar magnetic fields, which in turn results in 

795: reduced coronal heating relative to what would otherwise be expected.

796: 

797: The original motivation for this study was to test magnetospheric accretion

798: models for CTTSs.  Figure \ref{magcomp} shows the relationship between 

799: magnetic field strengths predicted from these models compared with the measured

800: field strengths.  Clearly, there is no correlation.

801: Magnetospheric accretion models typically assume an aligned dipolar magnetic

802: field at the surface of a CTTS.  It is now fairly clear that the fields

803: of TTSs are not dipolar (Johns--Krull et al. 1999a, Valenti \& Johns--Krull

804: 2004, Daou et al. 2006, Yang et al. 2006).  The magnetic geometry must be

805: complicated at the surface of TTSs, as on the Sun and all other cool stars. 

806: However, the dipolar component may dominate at the inner edge of the 

807: accretion disk, several stellar radii from the surface.  This notion is

808: supported by observations of circular polarization in the narrow component

809: of the \ion{He}{1} emission line which forms near the base of the accretion

810: flows where the material strikes the stellar surface (Johns--Krull et al.

811: 1999a).  Time series polarization in this line shows smooth variations 

812: characterized by few if any polarity reversals, and are consistent with the

813: idea that the disk is interacting with a dipole-like field geometry

814: (Valenti \& Johns--Krull 2004, Symington et al. 2005, Yang et al. 2007).

815: 

816: Magnetospheric accretion models use an assumed dipolar field geometry

817: to predict which stellar magnetic field lines couple the stellar

818: surface to the disk. Since the actual magnetic geometry is more complicated,

819: the fraction of the stellar surface with field lines coupled to

820: the disk (including those along which accretion will occur) can differ 

821: significantly from predicted values (see Gregory et al. 2006) for a 

822: discussion of this point.  This means the relationships given in

823: Paper I and used to calculate the predicted fields given in Table \ref{sample}

824: are actually underconstrained, even though we have now measured $\bar B$

825: at the surface.

826: However, Johns--Krull and Gafford (2002) point out that a more general

827: relationship for the Shu et al. (1994) model can be

828: determined which does not rely on the assumption of a dipolar field

829: geometry.  This formulation characterizes the magnetic coupling in 

830: terms of the surface

831: flux that threads the disk outside corotation, rather than the

832: equatorial field strength of an assumed dipole.  Valenti and Johns--Krull

833: (2004) use the results of Johns--Krull and Gafford (2002) to derive an 

834: equation for the filling factor of accretion zones, $f_a$, on CTTSs [their

835: equation (1)].  We use our magnetic field results with this equation to

836: predict $f_a$ for each of our stars.  These accretion filling factors range

837: from 0.57\% to 3.09\%, in good agreement with the range of observed $f_a$

838: values (Valenti, Johns, \& Basri 1993; Calvet \& Gullbring 1998).

839: The predicted $f_a$ values range over a factor of 5.4 while the observed

840: mean field strengths range over a factor of 2.6.  In the original 

841: magnetospheric accretion models, $f_a$ is essentially assumed constant and 

842: the predicted field strengths range over a factor of 8 (Table \ref{sample}).

843: We suggest that $f_a$ is more likely than $\bar B$ to vary

844: by a large factor.  Our measurements of $\bar B$ support this conclusion.  

845: Additionally, one could look for a correlation between the fields predicted

846: in Table \ref{sample} and the fields anchored in the accretion flows and

847: diagnosed by the \ion{He}{1} line.  Johns--Krull et al. (2001) do this,

848: finding a good correlation; however, we caution that they only had a 

849: sample of 4 stars.

850: 

851: Given that the magnetic fields on the surface of TTSs may not be

852: dipolar, one question that comes to mind is whether the observed fields on

853: CTTSs are indeed strong enough to enforce disk locking in these young stars.

854: The magnetospheric accretion models require some (though admittedly uncertain)

855: field strength in the disk, several stellar radii above the star, to enforce

856: disk locking.  The dipole assumption can then be used to predict what field

857: strength is implied at the stellar surface.  If the field geometry is much

858: more complex than this, a larger field strength will be required since the

859: higher order field components have a stronger dependence on radius than 

860: a dipole does.  The contribution of the various field components at the

861: inner disk edge is uncertain, but as mentioned above the \ion{He}{1}

862: observations suggest the dipole component may dominate.  This field

863: component is uncertain observationally, but it can be constrained through

864: spectropolarimetry.   Smirnov et al. (2003) report a marginal detection of

865: a net field of 150 G on T Tau which was not confirmed by Smirnov et al.

866: (2004) or Daou et al. (2006).  Yang et al. (2007) detect a net field of 

867: 149 G on TW Hya on one night of their 6 night monitoring campaign on this

868: star, finding only upper limits of $\sim 100$ G on the other nights.  Other

869: studies also only find upper limits (3$\sigma$) of 100--200 G 

870: (Johns--Krull et al. 1999a, 

871: Valenti \& Johns--Krull 2004).  The relationship between the net field

872: observed spectropolarimetrically and the dipole component of the star depends

873: on knowing the angle between the line of sight and the dipole axis; however,

874: the general lack of detection suggests the dipole component on TTSs is small.

875: For example, a dipole with a 3 kG polar field strength (as suggested for

876: BP Tau by Symington et al. 2005) observed at an

877: angle of 45$^\circ$ would produce a net field of 690 G, well above current

878: detection limits.  Accretion hot spots are expected to have essentially

879: no effect on net field measurements from spectropolarimetry (Smirnov et 

880: al. 2005) since the acretion filling factors are typically $\sim 1$\%

881: (Valenti et al. 1993, Calvet \& Gullbring 1998). 

882: 

883: Compared with at least some of the predicted values in Table \ref{sample}, the

884: general lack of strong polarimetric field detections in TTSs might suggest the

885: fields are actually too weak to enforce disk locking in these stars.  However,

886:  the general agreement 

887: in magnitude between the predicted and observed accretion filling factors 

888: described above and in Valenti and Johns--Krull (2004) suggests that the 

889: trapped flux model of Shu et al. (1994, see also Ostriker \& Shu 1995) can 

890: still enforce disk locking with the field strengths observed here.  Indeed,

891: Johns--Krull and Gafford (2002) find that this trapped flux model predicts

892: correlations among stellar and accretion parameters that are actually

893: observed in the data, whereas correlations based on assuming a dipole field

894: geometry are not observed in the data.  As sectropolarimetric field

895: measurements get more precise, and the dipole component gets more constrained,

896: comparisons with all the field predictions in Table \ref{sample} will

897: become more critical.  In addition, more theoretical work is 

898: called for to explore whether realistic magnetic field geometries and 

899: strengths can produce disk locking in magnetospheric accretion models.

900: First steps in exploring the role of realistic field geometries in 

901: CTTSs accretion have been made by Gregory et al. (2006); however,

902: torque balance calculations still need to be made for these types of

903: models.

904: 

905: Lastly, we consider the origin of the strong fields observed in TTSs.  

906: Chabrier and K\"uker (2006) assert that primordial, or fossil, magnetic 

907: fields in fully convective stars can only survive for a timescale of

908: $\tau_d \sim R_*^2/\eta$, where $R_*$ is the stellar radius and $\eta$ is

909: the turbulent magnetic diffusivity.  The value of $\tau_d$ works out to

910: be on the order of 10 -- 100 years.  Thus, Charbier and K\"uker (2006)

911: assert that the fields of fully convective stars, such as most TTSs, 

912: must be produced by a turbulent dynamo.  On the other hand, Tayler (1987)

913: and Moss (2003) suggest that primordial fields can survive much longer

914: in pre-main sequence stars.  Chabrier and K\"uker (2006) and Dobler et al.

915: (2006) both find that dynamo action in fully convective stars should

916: correlate with rotation rate.  Motivated by this, we looked for correlations

917: between the observed magnetic fields and various stellar parameters 

918: including rotation (see \S 4.2.4).  No correlations were found, and 

919: indeed stars like CY Tau and DH Tau show very similar stellar properties

920: (\Teff, mass, radius, rotation rate, convective turnover time) but have

921: mean fields that differ by over a factor of 2.3.  In addition, the

922: work of Chabrier and K\"uker (2006) and Dobler et al. (2006) generally

923: find magnetic field strengths somewhat lower than the equipartition

924: values, while the observed fields are significantly stronger than this.

925: Additional constraints on fully convective dynamo models can also be

926: made by comparing predicted field geometries with those derived from

927: Zeeman Doppler imaging of fully convective stars.  Donati et al.

928: (2006) do this for an M dwarf, finding inconsistencies in the observed

929: field geometry relative to the predictions.  Therefore, 

930: it appears difficult for dynamo theories to produce the fields observed

931: on TTSs.  As a result, we suggest that the fields on these young stars

932: may indeed be primordial in origin, as suggested by Tayler (1987) and

933: Moss (2003).

934: 

935: 

936: \acknowledgements

937: I am pleased to acknowledge numerous, stimulating discussions with J. Valenti

938: on all aspects of the work reported here.  I wish to thank J. Valenti, S.

939: Saar, and

940: H. Yang for their assistance in obtaining some of the spectra used in this 

941: study.  I also wish to thank staff of the NASA IRTF for their hospitality and

942: help during my observing runs there.  Finally, I wish to acknowledge partial 

943: support from the NASA Origins of Solar Systems program through grant numbers

944: NAG5-13103 and NNG06GD85G made to Rice University.  Finally, I wish to

945: acknowledge many useful comments from an anonymous referee.

946: 

947: 

948: \begin{references}

949: \reference{} Alencar, S.\ H.\ P., Johns--Krull, C.\ M., \& Basri, G., 2001,

950:    AJ, 122, 3335

951: \reference{} Allard, F. \& Hauschildt, P.H. 1995, in Bottom of the Main

952:    Sequence and Beyond, ed. C.G. Tinney (Berlin: Springer), 32

953: \reference{} Azevedo, R., Calvet, N., Hartmann, L., Folha, D.~F.~M., Gameiro, 

954:    F., \& Muzerolle, J.\ 2006, \aap, 456, 225 

955: \reference{} Basri, G., \& Batalha, C.\ 1990, \apj, 363, 654 

956: \reference{} Basri, G., Marcy, G.\ W., \& Valenti, J.\ A. 1992, ApJ, 390, 622

957: \reference{} Beristain, G., Edwards, S., \& Kwan, J.\ 2001, \apj, 551, 1037

958: \reference{} Bertout, C., Basri, G., \& Bouvier, J. 1988, ApJ, 330, 350

959: \reference{} Bertout, C., Harder, S., Malbet, F., Mennessier, C., \&

960:    Regev, O. 1996, \aj, 112, 2159

961: \reference{} Bouvier, J., et al. 1999, A{\&}A, 349, 619

962: \reference{} Bouvier, J., Cabrit, S., Fenandez, M., Martin, E.L., \& Matthews,

963:    J.M. 1993, A{\&}A, 272, 176

964: \reference{} Bouvier, J., Alencar, S.H.P., Harries, T.J., Johns--Krull, C.M.,

965:    \& Romanova, M.M. 2006, in Protostars and Planets V, B. Reipurth, D. Jewitt,

966:    \& K. Keil (eds.), (Tucson: Univ. of Arizona Press), in press

967: \reference{} Bouvier, J., Covino, E., Kovo, O., Mart\'in, E.L., Matthews, J.M.,

968:    Terranegra, L., \& Beck, S.C. 1995, A{\&}A, 299, 89

969: \reference{} Calvet, N., \& Gullbring, E.\ 1998, \apj, 509, 802 

970: \reference{} Camenzind, M. 1990, Rev. Modern Astron., (Berlin: Springer-Verlag),

971:    3, 234

972: \reference{} Casali, M.~M., \& Eiroa, C. 1996, \aap, 306, 427

973: \reference{} Chabrier, G. \& K\"uker, M. 2006, \aap, 446, 1027

974: \reference{} Collier Cameron, A. \& Campbell, C.\ G. 1993, \aap, 274, 309

975: \reference{} Daou, A.\ G., Johns--Krull, C.\ M., \& Valenti, J.\ A. 2006, \aj,

976:    131, 520

977: \reference{} Damiani, F., Micela, G., Sciortino, S., \& Harnden, F.\ R.\ Jr.

978:    1995, \apj, 446, 331

979: \reference{} Dobler, W., Stix, M., \& Brandenburg, A. 2006, \apj, 638, 336

980: \reference{} Donati, J.-F., Forveille, T., Cameron, A.~C., Barnes, J.~R., 

981:    Delfosse, X., Jardine, M.~M., \& Valenti, J.~A.\ 2006, Science, 311, 633 

982: \reference{} Durney, B.\ R., De Young, D.\ S., \& Roxburgh, I.\ W. 1993, 

983:    Sol. Phys., 145, 207

984: \reference{} Durney, B.\ R. \& Latour, J. 1978, Geophys. Astrophys. Fluid

985:    Dynamics, 9, 241

986: \reference{} Edwards, S., Hartigan, P., Ghandour, L., \& Andrulis, C. 1994,

987:    \aj, 108, 1056

988: \reference{} Eiroa, C., et al.\ 2002, \aap, 384, 1038

989: \reference{} Feigelson, E.~D., Gaffney, J.~A., III, Garmire, G., Hillenbrand,

990:    L.~A., \& Townsley, L.\ 2003, \apj, 584, 911 

991: \reference{} Feigelson, E.\ D. \& Montmerle, T. 1999, ARAA, 37, 363

992: \reference{} Feigelson, E., Townsley, L., G{\"u}del, M., \& Stassun, K.\ 2007,

993:    Protostars and Planets V, 313 

994: \reference{} Ferreira, J., Pelletier, G., \& Appl, S. 2000, MNRAS, 312, 387

995: \reference{} Fisher, G.~H., Longcope, D.~W., Metcalf, T.~R., \& Pevtsov, 

996:    A.~A.\ 1998, \apj, 508, 885 

997: \reference{} Flaccomio, E., Micela, G., \& Sciortino, S. 2003, \aap, 402, 277

998: \reference{} Folha, D.~F.~M., \& Emerson, J.~P.\ 1999, \aap, 352, 517 

999: \reference{} Foukal, P.\ 1990, Solar Astrophysics, (New York: 

1000:    Wiley-Interscience)

1001: \reference{} Ghosh, P. \& Lamb, F.\ K. 1979, ApJ, 232, 259

1002: \reference{} Goodson, A.~P., B\"ohm, K.-H., \& Winglee, R.~M. 1999, \apj, 

1003:    524, 142

1004: \reference{} Goodson, A.~P., Winglee, R.~M., \& B\"ohm, K.-H. 1997, \apj,

1005:    489, 199

1006: \reference{} Gray, D.~F.\ 1982, \apj, 255, 200

1007: \reference{} Greene, T.~P., Tokunaga, A.~T., Toomey, D.~W., \& Carr, J.~B.\

1008:    1993, \procspie, 1946, 313

1009: \reference{} Gregory, S.~G., Jardine, M., Simpson, I., \& Donati, J.-F. 2006,

1010:    \mnras, 371, 999

1011: \reference{} Guenther, E.\ W., Lehmann, H., Emerson, J.\ P., \& Staude, J.

1012:    1999, A{\&}A, 341, 768

1013: \reference{} Gullbring, E., Hartmann, L., Brice\~no, C., \& Calvet, N. 1998,

1014:    ApJ, 492, 323 

1015: \reference{} Hartigan, P., Edwards, S., \& Ghandour, L.\ 1995, \apj, 452, 736 

1016: \reference{} Hartmann, L., Hewett, R., Stahler, S., \& Mathieu, R.~D.\ 1986,

1017:    \apj, 309, 275

1018: \reference{} Hartmann, L. \& Stauffer, J.\ R. 1989, AJ, 97, 873

1019: \reference{} Herbst, W., Bailer--Jones, C.\ A.\ L., \& Mundt, R. 2001, ApJ,

1020:    554, L197

1021: \reference{} Hayashi, M.~R., Shibata, K., \& Matsumoto, R. 1996, \apj, 468, 

1022:    L37

1023: \reference{} Huerta, M. 2007, Ph.D. Thesis, Rice University

1024: \reference{} Jardine, M., Cameron, A.~C., Donati, J.-F., Gregory, S.~G., \& 

1025:   Wood, K.\ 2006, \mnras, 367, 917

1026: \reference{} Johns, C.\ M. \& Basri, G. 1995, ApJ, 449, 341

1027: \reference{} Johns--Krull, C.\ M. \& Gafford, A.\ D. 2002, \apj, 573, 685

1028: \reference{} Johns--Krull, C.\ M. \& Valenti, J.\ A. 1996, \apjl, 459, L95

1029: \reference{} Johns--Krull, C.\ M. \& Valenti, J.\ A. 2000, in ASP Conf. Ser.

1030:    198, Stellar Clusters and Associations, ed. R. Pallavicini, G. Micela, \&

1031:    S. Sciortino (San Francisco: ASP), 371

1032: \reference{} Johns--Krull, C.~M.~\& Valenti, J.~A.\ 2001, \apj, 561, 1060

1033: \reference{} Johns--Krull, C.\ M., Valenti, J.\ A., Hatzes, A.\ P., \&

1034:    Kanaan, A. 1999a, 510, L41

1035: \reference{} Johns--Krull, C.\ M., Valenti, J.\ A., \& Koresko, C. 1999b, ApJ,

1036:    516, 900 (Paper I)

1037: \reference{} Johns--Krull, C.\ M., Valenti, J.\ A., \& Linsky, J.\ L. 2000,

1038:    \apj, 539, 815

1039: \reference{} Johns--Krull, C.\ M., Valenti, J.\ A., \& Saar, S.\ H. 2004,

1040:    \apj, 617, 1204 (Paper II)

1041: \reference{} Johns--Krull, C.\ M., Valenti, J.\ A., Saar, S.\ H., \& Hatzes,

1042:    A.\ P. 2001, Cool Stars, Stellar Systems, and the Sun: 11th Cambridge

1043:    Workshop, R.\ J.\ Garc\'ia L\'opez, R. Rebolo, \& M.\ R.\ Zapatero Osorio

1044:    (eds.), ASP Conf. Series, 223, 521

1045: \reference{} Johnson, H.\ L. 1966, ARAA, 4, 193

1046: \reference{} Kastner, J.\ H., Huenemoerder, D.\ P., Schulz, N.\ S., Canizares, 

1047:    C.\ R., Weintraub, D.\ A. 2002, \apj, 567, 434

1048: \reference{} K\"onigl, A. 1991, ApJ, 370, L39

1049: \reference{} Kenyon, S.\ J., Hartmann, L., Hewett, R., Carrasco, L., 

1050:    Cruz--Gonzalez, I., Recillas, E., Salas, L., Serrano, A., Strom, K.\ M.,

1051:    Strom, S.\ E., \& Newton, G. 1994, AJ, 107, 2153

1052: \reference{} Kim, Y.-C. \& Demarque, P. 1996, \apj, 457, 340

1053: \reference{} Long, M., Romanova, M.\ M., \& Lovelace, R.\ V.\ E. 2005, \apj,

1054:    634, 1214

1055: \reference{} Luhmann, J.~G., Gosling, J.~T., Hoeksema, J.~T., \& Zhao, X.\ 

1056:    1998, \jgr, 103, 6585

1057: \reference{} Mathew, S.~K., et al.\ 2003, \aap, 410, 695 

1058: \reference{} Matt, S. \& Pudritz, R.\ E. 2004, \apj, 607, L43

1059: \reference{} Matt, S. \& Pudritz, R.\ E. 2005, MNRAS, 356, 167

1060: \reference{} Matt, S., \& Pudritz, R.~E.\ 2005, \apjl, 632, L135 

1061: \reference{} Mekkaden, M.\ V. 1998, \aap ,340, 135

1062: \reference{} Miller, K.~A. \& Stone, J.~M. 1997, \apj, 489, 890

1063: \reference{} Moss, D. 2003, A\&A, 403, 693

1064: \reference{} Muzerolle, J., Calvet, N., Brice\~eno, C., Hartmann, L., \&

1065:    Hillenbrand, L. 2000, ApJ, 535, L47

1066: \reference{} Muzerolle, J., Calvet, N., \& Hartmann, L. 1998, ApJ, 492, 743

1067: \reference{} Muzerolle, J., Calvet, N., \& Hartmann, L. 2001, ApJ, 550, 944

1068: \reference{} Neuhauser, R., Sterzik, M.\ F., Schmitt, J.\ H.\ M.\ M., 

1069:    Wichmann, R., \& Krautter, J. 1995, \aap, 297, 391

1070: \reference{} Noyes, R.~W., Hartmann, L.~W., Baliunas, S.~L., Duncan, D.~K., 

1071:    \& Vaughan, A.~H.\ 1984, \apj, 279, 763 

1072: \reference{} Ostriker, E.~C., \& Shu, F.~H.\ 1995, \apj, 447, 813 

1073: \reference{} Paatz, G. \& Camenzind, M. 1996, A{\&}A, 308, 77

1074: \reference{} Padgett, D.~L.\ 1996, \apj, 471, 847 

1075: \reference{} Preibisch, T., Kim, Y. -C., Favata, F., Feigelson, E. D., 

1076:    Flaccomio, E., Getman, K., Micela, G., Sciortino, S., Stassun, K., Stelzer,

1077:    B., \& Zinnecker, H. 2005, \apjs, 160, 401

1078: \reference{} Press, W.~H., Flannery, B.~P., Teukolsky, S.~A., \& Vetterling,

1079:    W.~T. 1986, Numerical Recipes, (New York: Cambridge Univ. Press)

1080: \reference{} Pevtsov, A.~A., Fisher, G.~H., Acton, L.~W., Longcope, D.~W.,

1081:    Johns--Krull, C.~M., Kankelborg, C.~C., \& Metcalf, T.~R.\ 2003, \apj, 598,

1082:    1387 

1083: \reference{} Rajaguru, S.~P., Kurucz, R.~L., \& Hasan, S.~S.\ 2002, \apjl, 

1084:    565, L101

1085: \reference{} Romanova, M.~M., Ustyugova, G.~V., Koldoba, A.~V., \& Lovelace,

1086:    R.~V.~E.\ 2004, \apj, 610, 920 

1087: \reference{} Romanova, M.~M., Ustyugova, G.~V., Koldoba, A.~V., \& Lovelace,

1088:    R.~V.~E.\ 2005, \apjl, 635, L165 

1089: \reference{} Robinson, R.\ D. 1980, \apj, 239, 961

1090: \reference{} Saar, S.\ H. 1988, \apj, 324, 441

1091: \reference{} Saar, S.\ H. 1990, IAU Symp. 138, Solar Photosphere: Structure,

1092:    Convection, and Magnetic Fields, ed.  J.\ O. Stenflo (Dordrecht: Kluwer),

1093:    427

1094: \reference{} Saar, S.\ H. 1994, IAU Symp. 154, Infrared Solar Physics, eds. 

1095:    D.\ M. Rabin et al. (Dordrecht: Kluwer), 493

1096: \reference{} Saar, S.\ H. 1996, IAU Symp. 176, Stellar Surface Structure, eds. 

1097:    K.\ G. Strassmeier \& J.\ L. Linsky (Dordrecht: Kluwer), 237

1098: \reference{} Saar, S.\ H. \& Linsky, J.\ L. 1985, \apj, 299, L47

1099: \reference{} Safier, P.~N.\ 1999, \apjl, 510, L127 

1100: \reference{} Shu, F.\ H., Najita, J., Ostriker, E., Wilkin, F., Ruden, S., \&

1101:    Lizano, S. 1994, ApJ, 429, 781

1102: \reference{} Siess, L., Dufour, E., \& Forestini, M. 2000, \aap, 358, 593

1103: \reference{} Skrutskie, M.~F., Meyer, M.~R., Whalen, D., \& Hamilton, C.\ 1996,

1104:    \aj, 112, 2168

1105: \reference{} Smirnov, D.~A., Fabrika, S.~N., Lamzin, S.~A., \& Valyavin, 

1106:    G.~G.\ 2003, \aap, 401, 1057

1107: \reference{} Smirnov, D.~A., Lamzin, S.~A., Fabrika, S.~N., \& Chuntonov, 

1108:    G.~A.\ 2004, Astronomy Letters, 30, 456 

1109: \reference{} Smirnov, D.~A., Romanova, M.~M., \& Lamzin, S.~A.\ 2005, 

1110:    Astronomy Letters, 31, 335 

1111: \reference{} Spruit, H.~C., \& Zweibel, E.~G.\ 1979, \solphys, 62, 15 

1112: \reference{} Stassun, K.~G., Ardila, D.~R., Barsony, M., Basri, G., \&

1113:    Mathieu, R.~D. 2004, \aj, 127, 3537

1114: \reference{} Stassun, K.~G., Mathieu, R.~D., Mazeh, T., \& Vrba, F.~J.\ 1999,

1115:    \aj, 117, 2941

1116: \reference{} Symington, N.~H., Harries, T.~J., Kurosawa, R., \& Naylor, T.\ 

1117:    2005, \mnras, 358, 977

1118: \reference{} Tayler, R.J. 1987, MNRAS, 227, 553

1119: \reference{} Tokunaga, A.~T., Toomey, D.~W., Carr, J., Hall, D.~N.~B., \& 

1120:    Epps, H.~W.\ 1990, \procspie, 1235, 131

1121: \reference{} Valenti, J.~A., Basri, G., \& Johns, C.~M.\ 1993, \aj, 106, 2024 

1122: \reference{} Valenti, J.~A., \& Fischer, D.~A.\ 2005, \apjs, 159, 141 

1123: \reference{} Valenti, J.\ A. \& Johns--Krull, C.\ M. 2004, Ap\&SS, 292, 619

1124: \reference{} Valenti, J.\ A., Marcy, G.\ W., \& Basri, G. 1995, \apj, 439, 939

1125: \reference{} Valenti, J.~A., Piskunov, N., \& Johns-Krull, C.~M.\ 1998, \apj,

1126:    498, 851

1127: \reference{} Vilhu, O.\ 1984, \aap, 133, 117

1128: \reference{} Wallace, L., \& Hinkle, K.\ 1997, \apjs, 111, 445 

1129: \reference{} Wallace, L., \& Livingston, W.~C.\ 1992, NSO Technical Report 

1130:    \#92-001; Tucson: National Solar Observatory, 1992, 

1131: \reference{} White, R.~J.~\& Ghez, A.~M.\ 2001, \apj, 556, 265 

1132: \reference{} Yang, H., Johns--Krull, C.\ M., \& Valenti, J.\ A. 2005, \apj,

1133:    635, 466 (Paper III)

1134: \reference{} Yang, H., Johns--Krull, C.\ M., \& Valenti, J.\ A. 2007, \aj,

1135:    133, 73

1136: \end{references}

1137: 

1138: 

1139: \clearpage

1140:  

1141: \begin{deluxetable}{lcccccccccc}

1142: \tablewidth{17.5truecm}   %8.8truecm maximum

1143: \tablecaption{Sample of Stars\label{sample}}

1144: \tablehead{

1145:    \colhead{}&

1146:    \colhead{Spec.}&

1147:    \colhead{$M_*$}&

1148:    \colhead{$R_*$}&

1149:    \colhead{$\dot{M} \times 10^8$}&

1150:    \colhead{$P_{rot}$}&

1151:    \colhead{$B_{eq}$}&

1152:    \colhead{$B_{eq}$}&

1153:    \colhead{$B_{eq}$}&

1154:    \colhead{$B_{eq}$}&

1155:    \colhead{$L_X$}\\[0.2ex]

1156:    \colhead{Star}&

1157:    \colhead{Type}&

1158:    \colhead{(M$_\odot$)}&

1159:    \colhead{(R$_\odot$)}&

1160:    \colhead{$(M_\odot {\rm yr}^{-1})$}&

1161:    \colhead{(days)}&

1162:    \colhead{(G)}&

1163:    \colhead{(G)}&

1164:    \colhead{(G)}&

1165:    \colhead{(G)}&

1166:    \colhead{($10^{30}$ erg s$^{-1}$)}

1167: }

1168: \startdata

1169: AA Tau  &K7& 0.70 & 1.74 & 0.33 & 8.20 & 1020 & 290 & 1210 & 380 &  0.439 \\

1170: BP Tau  &K7& 0.70 & 1.99 & 2.88 & 7.60 & 1850 & 620 & 2180 & 880 &  0.707 \\

1171: CY Tau  &M1& 0.40 & 1.63 & 0.75 & 7.90 & 1130 & 380 & 1330 & 530 & $<0.627$ \\

1172: DE Tau  &M2& 0.32 & 2.45 & 2.64 & 7.60 &  490 & 190 &  580 & 230 & $<0.287$ \\

1173: DF Tau  &M1& 0.40 & 3.37 & 17.7 & 8.50 &  680 & 290 &  790 & 320 &  0.646 \\

1174: DG Tau  &K7.5& 0.65 & 2.05 & 4.57 & 6.30 & 1610 & 560 & 1900 & 760 & $<0.223$ \\

1175: DH Tau  &M1.5& 0.36 & 1.39 & 0.10 & 7.20 &  550 & 160 &  640 & 260 &  1.318 \\

1176: DK Tau  &K7& 0.68 & 2.49 & 3.79 & 8.40 & 1190 & 410 & 1400 & 560 & $<0.223$ \\

1177: DN Tau  &M0& 0.51 & 2.09 & 0.35 & 6.00 &  320 & 100 &  380 & 150 &  0.564 \\

1178: GG Tau A&K7& 0.68 & 2.31 & 1.75 & 10.3 & 1280 & 420 & 1510 & 610 & $<0.107$ \\

1179: GI Tau  &K6& 0.93 & 1.74 & 0.96 & 7.20 & 1900 & 560 & 2240 & 900 &  0.241 \\

1180: GK Tau  &K7& 0.69 & 2.15 & 0.64 & 4.65 &  390 & 110 &  450 & 180 &  0.241 \\

1181: GM Aur  &K7& 0.70 & 1.78 & 0.96 & 12.0 & 2540 & 800 & 2990 &1200 &  0.562 \\

1182:  T Tau  &K0& 2.30 & 3.31 & 4.40 & 2.80 &  420 & 110 &  490 & 200 &  0.568 \\

1183: TW Hya  &K7& 0.74 & 1.00 & 0.20 & 2.20 &  950 & 250 & 1120 & 450 &  1.380 \\

1184: \enddata

1185: \tablecomments{All predicted fields are the equatorial magnetic field

1186: strength for an assumed dipolar magnetic field.  In order, the predicted

1187: fields come from K\"onigl (1991), Collier Cameron and Campbell (1993),

1188: Shu et al. (1994), and Long et al. (2005).}

1189: \end{deluxetable}

1190: 

1191: \clearpage

1192: 

1193: \begin{deluxetable}{lcc}

1194: \tablewidth{5.5truecm}   %8.8truecm maximum

1195: \tablecaption{Observing Runs\label{obs}}

1196: \tablehead{

1197:    \colhead{}&

1198:    \colhead{Obs.}&

1199:    \colhead{Wavl.}\\[0.2ex]

1200:    \colhead{Star}&

1201:    \colhead{Run}&

1202:    \colhead{Set}

1203: }

1204: \startdata

1205: AA Tau   &  III & 1,2,3 \\

1206: BP Tau   &  II  & 1,2,3 \\

1207: CY Tau   &  IV  & 1,2,3 \\

1208: DE Tau   &  III & 1,2,3 \\

1209: DF Tau   &  II  & 1,2,3 \\

1210: DG Tau   &  III & 1,2,3 \\

1211: DH Tau   &  III & 1,2,3 \\

1212: DK Tau   &  II  & 1,2,3 \\

1213: DN Tau   &  III & 1,2,3 \\

1214: GG Tau A &  III & 1,2,3 \\

1215: GI Tau   &  III & 1,2,3 \\

1216: GK Tau   &  III & 1,2,3 \\

1217: GM Aur   &  I   & 2 \\

1218: T Tau    &  II  & 1,2,3 \\

1219: \enddata

1220: \end{deluxetable}

1221: 

1222: \clearpage

1223: 

1224: \begin{deluxetable}{lcccc}

1225: \tablewidth{10.0truecm}   %8.8truecm maximum

1226: \tablecaption{Results\label{results}}

1227: \tablehead{

1228:    \colhead{}&

1229:    \colhead{}&

1230:    \colhead{$\bar B$}&

1231:    \colhead{$B_{eq}$}&

1232:    \colhead{$L_X$}\\[0.2ex]

1233:    \colhead{Star}&

1234:    \colhead{$r_K$}&

1235:    \colhead{(kG)}&

1236:    \colhead{(kG)}&

1237:    \colhead{($10^{30}$ erg s$^{-1}$)}

1238: }

1239: \startdata

1240: AA Tau  & $0.52 \pm 0.09$ & 2.78 & 1.02 & 4.32 \\

1241: BP Tau  & $1.08 \pm 0.25$ & 2.17 & 1.03 & 4.42 \\

1242: CY Tau  & $0.65 \pm 0.21$ & 1.16 & 1.04 & 1.36 \\

1243: DE Tau  & $1.16 \pm 0.17$ & 1.12 & 1.04 & 3.34 \\

1244: DF Tau  & $1.37 \pm 0.15$ & 2.90 & 1.04 & 20.7 \\

1245: DG Tau  & $2.67 \pm 0.41$ & 2.55 & 1.00 & 8.19 \\

1246: DH Tau  & $1.77 \pm 0.49$ & 2.68 & 1.42 & 2.20 \\

1247: DK Tau  & $1.46 \pm 0.23$ & 2.64 & 1.02 & 9.27 \\

1248: DN Tau  & $0.54 \pm 0.09$ & 2.00 & 1.01 & 4.51 \\

1249: GG Tau A& $0.35 \pm 0.19$ & 1.24 & 1.02 & 3.28 \\

1250: GI Tau  & $1.00 \pm 0.11$ & 2.73 & 1.44 & 4.23 \\

1251: GK Tau  & $1.47 \pm 0.31$ & 2.28 & 1.02 & 5.59 \\

1252: GM Aur  & $1.25$          & 2.22 & 1.02 & 3.51 \\

1253:  T Tau  & $0.45 \pm 0.77$ & 2.37 & 0.84 & 15.8 \\

1254: TW Hya\tablenotemark{a}  & $0.07 \pm 0.04$ & 2.61 & 1.76 & 1.13 \\

1255: \enddata

1256: \tablenotetext{a}{Veiling and field measurements taken from Yang et al.

1257: (2005).}

1258: \end{deluxetable}

1259: 

1260: \clearpage

1261: 

1262: \begin{deluxetable}{lcccccccccccccc}

1263: \tabletypesize{\scriptsize}

1264: \tablewidth{18.0truecm}   %8.8truecm maximum

1265: \tablecaption{Range of Predicted Fields\label{results2}}

1266: \tablehead{

1267:    \colhead{}&

1268:    \multicolumn{2}{c}{VBJ}&

1269:    \colhead{}&

1270:    \multicolumn{2}{c}{HEG}&

1271:    \colhead{}&

1272:    \multicolumn{2}{c}{GHBC}&

1273:    \colhead{}&

1274:    \multicolumn{2}{c}{CG}&

1275:    \colhead{}&

1276:    \multicolumn{2}{c}{WG}\\[0.2ex]

1277:    \cline{2-3}

1278:    \cline{5-6}

1279:    \cline{8-9}

1280:    \cline{11-12}

1281:    \cline{14-15}

1282:    \colhead{}&

1283:    \colhead{}&

1284:    \colhead{}&

1285:    \colhead{}&

1286:    \colhead{}&

1287:    \colhead{}&

1288:    \colhead{}&

1289:    \colhead{}&

1290:    \colhead{}&

1291:    \colhead{}&

1292:    \colhead{}&

1293:    \colhead{}&

1294:    \colhead{}&

1295:    \colhead{}&

1296:    \colhead{}\\[0.2ex]

1297:    \colhead{}&

1298:    \colhead{$\dot{M} \times 10^8$}&

1299:    \colhead{$B_{eq}$\tablenotemark{a}}&

1300:    \colhead{}&

1301:    \colhead{$\dot{M} \times 10^8$}&

1302:    \colhead{$B_{eq}$\tablenotemark{a}}&

1303:    \colhead{}&

1304:    \colhead{$\dot{M} \times 10^8$}&

1305:    \colhead{$B_{eq}$\tablenotemark{a}}&

1306:    \colhead{}&

1307:    \colhead{$\dot{M} \times 10^8$}&

1308:    \colhead{$B_{eq}$\tablenotemark{a}}&

1309:    \colhead{}&

1310:    \colhead{$\dot{M} \times 10^8$}&

1311:    \colhead{$B_{eq}$\tablenotemark{a}}\\[0.2ex]

1312:    \colhead{Star}&

1313:    \colhead{$(M_\odot {\rm yr}^{-1})$}&

1314:    \colhead{(kG)}&

1315:    \colhead{}&

1316:    \colhead{$(M_\odot {\rm yr}^{-1})$}&

1317:    \colhead{(kG)}&

1318:    \colhead{}&

1319:    \colhead{$(M_\odot {\rm yr}^{-1})$}&

1320:    \colhead{(kG)}&

1321:    \colhead{}&

1322:    \colhead{$(M_\odot {\rm yr}^{-1})$}&

1323:    \colhead{(kG)}&

1324:    \colhead{}&

1325:    \colhead{$(M_\odot {\rm yr}^{-1})$}&

1326:    \colhead{(kG)}

1327: }

1328: \startdata

1329: AA Tau  & 0.71 & 1.56 & & 12.6 & 4.04 & & 0.33 & 0.96 & & 0.40 & 1.05 & & 0.65 & 2.34 \\

1330: BP Tau  & 2.43 & 2.46 & & 15.8 & 4.06 & & 2.88 & 1.62 & & 2.30 & 1.45 & & 1.32 & 1.81 \\

1331: CY Tau  &\nodata&\nodata& & 0.63 & 2.62 & & 0.75 & 1.38 & & 0.80 & 1.43 & & 0.14 & 0.75 \\

1332: DE Tau  & 18.0 & 0.07 & & 31.6 & 1.18 & & 2.64 & 0.49 & & 4.60 & 0.65 & & 4.07 & 0.62 \\

1333: DF Tau  & 21.9 & 0.37 & &125.9 & 0.67 & & 17.7 & 0.57 & & 20.8 & 0.62 & & 1.00 & 2.03 \\

1334: DG Tau  &\nodata&\nodata& &199.5 & 9.09 & &\nodata&\nodata& & 91.9 & 7.36 & & 4.57 & 2.44 \\

1335: DH Tau  & 2.83 & 0.25 & &\nodata&\nodata& &\nodata&\nodata& &\nodata&\nodata& & 0.11 & 0.93 \\

1336: DK Tau  & 0.61 & 1.93 & & 39.8 & 2.19 & & 3.79 & 0.95 & & 0.50 & 0.35 & &\nodata&\nodata\\

1337: DN Tau  & 0.13 & 0.15 & & 3.20 & 0.84 & & 0.35 & 0.30 & & 0.20 & 0.23 & & 0.19 & 0.38 \\

1338: GG Tau A& 3.01 & 1.48 & & 20.0 & 1.41 & & 1.75 & 1.05 & & 2.50 & 1.26 & & 1.26 & 1.79 \\

1339: GI Tau  & 2.02 & 9.16 & & 12.6 & 1.06 & & 0.96 & 1.70 & & 0.60 & 1.35 & & 0.83 & 1.19 \\

1340: GK Tau  & 0.05 & 0.23 & & 6.30 & 0.86 & & 0.64 & 0.32 & & 1.40 & 0.48 & & 0.65 & 0.57 \\

1341: GM Aur  & 0.74 & 4.30 & & 2.50 & 5.20 & & 0.96 & 2.34 & & 0.70 & 2.00 & & 0.66 & 6.86 \\

1342:  T Tau  &\nodata&\nodata& &\nodata&\nodata& &\nodata&\nodata& &\nodata&\nodata& & 3.16 & 0.39 \\

1343: {\it Cor}\tablenotemark{b}& 0.27& 0.43 & & 0.21 & 0.51 & & 0.07 & 0.85 & & 0.21 & 0.52 & & 0.18 & 0.55 \\

1344: \enddata

1345: \tablenotetext{a}{The predicted field strength is the equatorial field

1346: strength of an assumed dipole field using the Shu et al. (1994) model.}

1347: \tablenotetext{b}{Given on this line in each $\dot M$ column is the value

1348: of the correlation coefficient, $r$, between the predicted field strenghts 

1349: and the observed mean field strengths.  Given on this line in each $B_{eq}$

1350: column is the associated false alarm probability.}

1351: \end{deluxetable}

1352: 

1353: \clearpage

1354: 

1355: \begin{deluxetable}{lcccc}

1356: \tablewidth{6.0truecm}   %8.8truecm maximum

1357: \tablecaption{Correlations of Derived Magnetic Parameters with Stellar Parameters\label{stelpar}}

1358: \tablehead{

1359:    \colhead{Quantities}&

1360:    \colhead{}&

1361:    \colhead{}\\[0.2ex]

1362:    \colhead{Compared}&

1363:    \colhead{$r$}&

1364:    \colhead{$f_p$}

1365: }

1366: \startdata

1367: $\bar B$ vs $P_{rot}^{-1}$ & 0.22 & 0.44 \\

1368: $\Phi $ vs $P_{rot}^{-1}$ & 0.05 & 0.85 \\

1369: $\bar B$ vs $\tau_c P_{rot}^{-1}$ & 0.23 & 0.41 \\

1370: $\Phi $ vs $\tau_c P_{rot}^{-1}$ & $-0.24$ & 0.40 \\

1371: $\bar B$ vs $\tau_c$ & 0.03 & 0.92 \\

1372: $\Phi $ vs $\tau_c$ & $-0.57$ & 0.03 \\

1373: $\bar B$ vs $T_{eff}$ & 0.26 & 0.36 \\

1374: $\Phi $ vs $T_{eff}$ & 0.37 & 0.18 \\

1375: $\bar B$ vs $M_*$ & 0.42 & 0.12 \\

1376: $\Phi $ vs $M_*$ & 0.20 & 0.48 \\

1377: $\bar B$ vs $L_*$ & 0.10 & 0.73 \\

1378: $\Phi $ vs $L_*$ & 0.66 & $<0.01$ \\

1379: $\bar B$ vs Age & 0.23 & 0.41 \\

1380: $\Phi $ vs Age & $-0.31$ & 0.26 \\

1381: \enddata

1382: \tablecomments{Convective turnover times, $\tau_c$, were kindly computed 

1383: by Y.-C. Kim, based on models described in Kim \& Demarque (1996) and 

1384: Preibisch et al. (2005).}

1385: \end{deluxetable}

1386: 

1387: \clearpage

1388: 

1389: \begin{figure} 

1390: \epsscale{.75}

1391: \plotone{f1.ps}

1392: \caption{K band spectra of DK Tau are shown in the black histogram.  The 

1393: top 2 panels show Zeeman sensitive Ti I lines and a Sc I line.  The bottom 

1394: panel shows Zeeman insensitive CO lines.  The dashed curve shows a model with

1395: no magnetic field.  The smooth solid curve shows the model fit for a star 

1396: covered by a mixture of regions which include 0, 2, 4, and 6 kG fields.  Here,

1397: the mean field strength averaged over the stellar surface is 2.64 kG.

1398: \label{highb}}

1399: \end{figure}

1400: 

1401: \clearpage

1402: 

1403: \begin{figure} 

1404: \plotone{f2.ps}

1405: \caption{K band spectra of DE Tau are shown in the black histogram.  The 

1406: top 2 panels show Zeeman sensitive Ti I lines and a Sc I line.  The bottom 

1407: panel shows Zeeman insensitive CO lines.  The dashed curve shows a model with

1408: no magnetic field.  The smooth solid curve shows the model fit for a star 

1409: covered by a mixture of regions which include 0, 2, 4, and 6 kG fields.  Here,

1410: the mean field strength averaged over the stellar surface is 1.12 kG.

1411: \label{lowb}}

1412: \end{figure}

1413: 

1414: \clearpage

1415: 

1416: \begin{figure} 

1417: \plotone{f3.ps}

1418: \caption{The observed mean magnetic field strength, $\bar B$, plotted

1419: versus the predicted equatorial magnetic field strength from the Shu et al.

1420: (1994) treatment of the magnetospheric accretion model.

1421: \label{magcomp}}

1422: \end{figure}

1423: 

1424: \begin{figure} 

1425: \plotone{f4.ps}

1426: \caption{The measured X-ray luminosity plotted against the predicted

1427: X-ray luminosity for our sample of stars.  The line of equality is shown

1428: by the dashed line.  The majority of these CTTSs appear underluminous in

1429: X-rays given their magnetic field properties.

1430: \label{xcomp}}

1431: \end{figure}

1432: 

1433: 

1434: \begin{figure} 

1435: \plotone{f5.ps}

1436: \caption{The observed mean magnetic field on our sample of CTTSs plotted

1437: versus the maximum allowed magnetic field strength predicted assuming 

1438: pressure equilibrium with a surrounding non-magnetic atmosphere.  The observed

1439: fields tend to be much stronger than the pressure equilibrium predictions,

1440: implying magnetic filling factors near unity.

1441: \label{eqcomp}}

1442: \end{figure}

1443: 

1444: \end{document}

1445: 

1446: