1: \documentclass[apj]{emulateapj}

2: \bibliographystyle{apj}

3: \usepackage{lscape}

4: \usepackage{apjfonts}

5: \usepackage{graphicx}

6: 

7: \slugcomment{Astrophys.J. 671 (2007) 1640}

8: \shorttitle{Galactic Cepheid {\it P-L} Relations}

9: \shortauthors{An et~al.}

10: 

11: \begin{document}

12: 

13: \title{The Distances to Open Clusters from Main-Sequence Fitting. IV.\\

14: Galactic Cepheids, the LMC, and the Local Distance Scale}

15: 

16: \author{Deokkeun An, Donald M.\ Terndrup, and Marc H.\ Pinsonneault}

17: 

18: \affil{Department of Astronomy, Ohio State University,

19: 140 West 18th Avenue, Columbus, OH 43210;\\

20: deokkeun,terndrup,pinsono@astronomy.ohio-state.edu}

21: 

22: \begin{abstract}

23: We derive the basic properties of seven Galactic open clusters

24: containing Cepheids and construct their period-luminosity ({\it P-L})

25: relations.  For our cluster main-sequence fitting we extend previous

26: Hyades-based empirical color-temperature corrections to hotter stars

27: using the Pleiades as a template.  We use $BVI_CJHK_s$ data to test

28: the reddening law and include metallicity effects to perform a more

29: comprehensive study for our clusters than prior efforts.  The ratio of

30: total to selective extinction $R_V$ that we derive is consistent with

31: expectations.  Assuming the LMC {\it P-L} slopes, we find

32: $\langle M_V\rangle=-3.93\pm0.07\ {\rm (statistical)}\pm0.14\ {\rm

33: (systematic)}$ for 10-day period Cepheids, which is generally fainter

34: than those in previous studies.  Our results are consistent with

35: recent {\it HST} and {\it Hipparcos} parallax studies when using the

36: Wesenheit magnitudes $W(VI)$.  Uncertainties in reddening and

37: metallicity are the major remaining sources of error in the $V$-band

38: {\it P-L} relation, but a higher precision could be obtained with deeper

39: optical and near-infrared cluster photometry.  We derive

40: distances to NGC~4258, the LMC, and M33 of $(m - M)_0=29.28\pm0.10$,

41: $18.34\pm0.06$, and $24.55\pm0.28$, respectively, with an additional

42: systematic error of $0.16$~mag in the {\it P-L} relations.  The distance

43: to NGC~4258 is in good agreement with the geometric distance derived

44: from water masers $[\Delta(m - M)_0=0.01\pm0.24]$, our value for M33 is

45: less consistent with the distance from an eclipsing binary

46: [$\Delta(m - M)_0=0.37\pm0.34$], and our LMC distance is moderately shorter

47: than the adopted distance in the {\it HST} Key Project, which formally

48: implies an increase in the Hubble constant of $7\%\pm8\%$.

49: \end{abstract}

50: 

51: \keywords{

52: Cepheids ---

53: distance scale ---

54: galaxies: individual (M33, NGC~4258) ---

55: Magellanic Clouds ---

56: open clusters and associations: general ---

57: stars: distances}

58: 

59: \section{Introduction}

60: 

61: The Cepheid period-luminosity ({\it P-L}) relation has played a key

62: role in the determination of distances within the Local Group and to

63: nearby galaxies.  By extension, it is also crucial for the calibration

64: of secondary distance indicators used to determine the Hubble constant:

65: see, for example, the {\it Hubble Space Telescope (HST)} Key Project on the

66: extragalactic distance scale \citep[][hereafter {\it HST} Key Project]

67: {freedman01} and the {\it HST} program for the luminosity calibration

68: of Type Ia supernovae (SNe~Ia) by means of Cepheids \citep{sandage06b}.

69: The cosmic distance scale was usually established by defining

70: the {\it P-L} relations for Cepheids in the Large Magellanic Cloud (LMC)

71: because of its numerous Cepheids, many of which have been discovered

72: as a result of microlensing campaigns \citep[e.g.,][]{udalski99a}.

73: Despite decades of effort, however, there have been persistent

74: differences in the inferred LMC distance from different methods

75: \citep[see][]{benedict02a}, including ones involving the same basic

76: calibrators such as RR Lyrae or Cepheid variables.

77: 

78: The Galactic {\it P-L} relationship, on the other hand, was traditionally

79: established using open clusters and associations containing Cepheids

80: \citep[see][]{feast87,feast03,sandage06a}.  There is also a steadily

81: growing body of parallax work

82: from the {\it Hipparcos} mission \citep{feast97,madore98,lanoix99,

83: groenewegen00,vanleeuwen07} and the {\it HST} \citep{benedict02b,benedict07},

84: as well as distances derived from the Baade-Becker-Wesselink moving

85: atmospheres method \citep{gieren97,gieren98,gieren05} and interferometric

86: angular diameter measurements \citep{lane00,lane02,kervella04a,kervella04b}.

87: In the current paper we focus on the Cepheid distance scale as inferred

88: from Galactic open clusters and its applications for the extragalactic

89: distance scale.

90: 

91: There are strengths and weaknesses in all of the methods used to establish

92: the Galactic Cepheid distance scale.  The absolute calibration of the

93: {\it P-L} relation requires not only accurate distance measurements but

94: also appropriate accounting for the effects of interstellar extinction and

95: reddening because Galactic Cepheids are heavily obscured with an average

96: $E(B - V)$ of order 0.5~mag \citep{fernie90}.  Reddening can be inferred

97: more precisely for clusters, while field Cepheids are more

98: numerous.  This difficulty can be partially overcome by the usage of

99: the ``reddening-free'' or Wesenheit index \citep{freedman91}, but there are

100: embedded assumptions about the extinction law even in such a system.

101: 

102: We believe that the time is right for a systematic reappraisal of

103: the Cepheid distance scale as inferred from open clusters.  Recent parallax

104: work employing the Wesenheit index \citep{benedict07,vanleeuwen07}

105: has yielded consistent claims of a somewhat smaller LMC distance modulus

106: ($\approx18.40$~mag) than the ones adopted by the {\it HST} Key Project

107: ($18.50$~mag) or \citeauthor{sandage06b} program ($18.54$~mag).  In addition,

108: there are now independent geometric tests of distances to other galaxies,

109: which have large numbers of Cepheids.  These include massive eclipsing

110: binaries in systems such as M31 \citep{ribas05} and M33 \citep{bonanos06}

111: and astrophysical water masers in NGC 4258 \citep{herrnstein99}.  As

112: described below, the availability of both more complete data in various

113: photometric passbands and better theoretical templates permits a more

114: comprehensive look at the open cluster Cepheid distance scale, which can

115: in turn be compared

116: with the alternate methods described above.  We can also use the open

117: clusters with Cepheids to test our ability to derive precise distances

118: and reddenings to heavily obscured and poorly studied systems.  We

119: believe that the net result is a more secure and robust determination

120: of the extragalactic distance scale.

121: 

122: We have been engaged in a long-term effort to create stellar evolution

123: models with the latest input physics and to generate isochrones that

124: are calibrated against photometry in local star clusters with accurate

125: distances.  After verifying that the models are in agreement with

126: the physical parameters of binaries and single stars in the Hyades

127: \citep[][hereafter Papers I and II, respectively]{pinsono03,pinsono04},

128: we developed a procedure for empirically correcting the color-effective

129: temperature ($T_{\rm eff}$)

130: relations to match photometry in the Hyades (Paper II).  Isochrones

131: generated in this way accurately reproduce the shapes of the main-sequence

132: (MS) in several colors, allowing the determination of distances to nearby

133: clusters to an accuracy of $\sim0.04$~mag in distance modulus

134: (or $2\%$ in distance) \citep[][hereafter Paper III]{an07}.

135: 

136: We extend the Hyades-based empirical corrections on theoretical

137: isochrones using the Pleiades and apply our techniques to a set of

138: Cepheid-bearing open clusters that have good multi-color photometry.

139: Selection of cluster sample and compilation of cluster photometry

140: are presented in \S~2.  Procedures on the isochrone calibration and

141: MS-fitting method are described in \S~3.  New cluster distances and

142: extinctions are presented in \S~4.  In \S~5 we construct multi-wavelength

143: Galactic {\it P-L} relations and discuss various systematic errors.

144: In \S~6 we estimate distances and reddenings for the maser galaxy

145: NGC~4258, the LMC, and M33.  We summarize our results in \S~7 and discuss

146: the importance of extinction corrections in the Cepheid distance scale.

147: 

148: \section{Cluster and Cepheid Data}\label{sec:data}

149: 

150: \subsection{Cluster Selection, Metallicity, and Age}\label{sec:mehage}

151: 

152: \input{tab1.tex}

153: 

154: About 30 open clusters and associations in the Galaxy are known to

155: harbor Cepheid variables \citep{feast87,feast99}, but not all of these

156: are useful for MS fitting because many are extremely sparse or poorly

157: studied.  We first excluded systems with overtone or double-mode

158: Cepheids and then examined the available photoelectric and/or CCD

159: photometry for each cluster.  From visual inspection of color-magnitude

160: diagrams (CMDs) we narrowed down the list to 10 promising candidates

161: for MS fitting.  The populous twin clusters $h\&\chi$~Per were

162: excluded from the current analysis because their young age introduces

163: significant uncertainties in

164: the isochrone calibration, and the membership of UY~Per in the clusters is

165: also doubtful \citep[see][and references therein]{walker87}.  The eight

166: remaining clusters are listed in Table~\ref{tab:list} with their

167: 10 Cepheid variables.  We denoted anonymous van den Bergh

168: \citep[C0634+031;][]{vdbergh57} as VDB~1 for brevity.  Most of

169: them are close enough that Two Micron All Sky Survey (2MASS) \citep{skrutskie06}

170: photometry reaches well down the MS.

171: 

172: It is likely that all Cepheids in Table~\ref{tab:list} are

173: members of their associated clusters.  For the three Cepheids in

174: NGC~7790 there exist no radial velocity or proper-motion measurements,

175: although all of them lie near the optical center of the cluster

176: \citep{sandage58}.  V~Cen is located about $25\arcmin$ away from

177: the center of NGC~5662, but \citet{baumgardt00} derived a high

178: membership probability from {\it Hipparcos} proper motions.

179: The rest of the Cepheids are generally considered to be

180: members of their clusters since they are located within the optical

181: radius defined by bright early-type MS stars and red giants.  In

182: many cases, membership is also supported by radial velocities

183: \citep[][and references therein]{mermilliod87} and proper

184: motions \citep[][and references therein]{baumgardt00}.  \citet{orsatti01}

185: suggested that TW~Nor may not be a member of Lyng{\aa}~6 from

186: polarization measurements, but we nevertheless included it here.

187: 

188: Previous studies have neglected the effects of cluster metallicity and

189: age on MS fitting \citep[e.g.,][]{feast87,hoyle03}.  However, the

190: luminosity of the MS at a fixed color (or $T_{\rm eff}$) is sensitive

191: to the metal abundance by $\Delta (m - M)_0 / \Delta {\rm [M/H]}

192: \sim 1$ (Paper~III).  In addition, the cluster age changes the mean

193: color of the upper MS by $\Delta (B - V) \sim 0.05$ between zero-age

194: MS (ZAMS) and $\sim100$~Myr isochrone, which is more appropriate for

195: Cepheid-bearing clusters (see below).  This, in turn, could affect

196: the reddening estimates from the upper MS by $\Delta E(B - V) \sim 0.05$

197: and an MS-fitting distance by $\Delta (m - M) \sim 0.1$.

198: 

199: We adopted Cepheid metal abundances from the high-resolution spectroscopy

200: of \citet{fry97} and assumed the same metallicities for the clusters.

201: In support of this assumption, \citeauthor{fry97} derived metallicities

202: for two dwarfs in M25 that are in agreement with the metallicity of U~Sgr.

203: In addition, \citeauthor{fry97} measured ${\rm \langle [Fe/H] \rangle} =

204: +0.01\pm0.02$ for two Pleiades dwarfs, which is in good agreement with our

205: adopted value for the Pleiades, ${\rm [Fe/H]} = +0.04\pm0.02$.

206: 

207: \citeauthor{fry97} also derived $\alpha$-element abundances (Si, Ca, and Ti)

208: for their sample.  For the stars in this paper, the average enhancement is

209: $\langle{\rm [\alpha/Fe]}\rangle = 0.14\pm0.02$.  We determined

210: an effective metallicity [M/H] from the measured [Fe/H] and [$\alpha$/Fe]

211: using the \citet{kim02} procedure.  These abundances are shown in

212: the fourth column of Table~\ref{tab:list}.  The mean metal abundance

213: of our sample is $\langle {\rm [M/H]}\rangle = +0.03\pm0.03$ and

214: the standard deviation is $0.08$~dex, indicating an intrinsic spread in

215: metallicity for our clusters.  On average, the rescaled [M/H] are larger

216: than [Fe/H] by $0.06$ dex.  We computed errors in [M/H] by propagation

217: of errors in [Fe/H] and [$\alpha$/Fe].

218: 

219: The metallicity scale for the Cepheids is probably robust to $\sim0.1$~dex.

220: As part of a larger study, \citet{yong06} reanalyzed spectra of 11 Cepheids

221: in \citeauthor{fry97} and derived lower [Fe/H] by $\approx0.1$~dex and lower

222: $\alpha$-element enhancement by $\approx0.05$~dex on average.  The mean difference

223: in the total metallicity is $\Delta \langle {\rm [M/H]} \rangle = -0.15

224: \pm0.03$, in the sense that the new analysis indicates lower values.

225: Independently, \citet{andrievsky02} measured abundances of 19 Cepheids in

226: common with \citeauthor{fry97}.  The average difference between these two

227: studies is $\approx0.1$~dex in [Fe/H], the \citeauthor{andrievsky02} values

228: being larger.  They also showed that the alpha abundance ratio ([Si, Ca, Ti/Fe])

229: is nearly solar, which results in a small difference in the total metal

230: abundance between \citeauthor{fry97} and \citeauthor{andrievsky02},

231: $\Delta \langle {\rm [M/H]} \rangle = -0.01\pm0.02$.

232: 

233: We adopted a single representative age (80 Myr) for all clusters, based

234: on comparing our isochrones (\S~\ref{sec:iso}) to the cluster CMDs

235: (\S~\ref{sec:phot}), and took its uncertainty as $\Delta \log{t {\rm

236: (Myr)}} = 0.2$ ($t = 80^{+50}_{-30}$~Myr).  The uncertainty represents

237: both a range of ages among the clusters and the accuracy of the fitting.

238: The average age in the Lyng{\aa} Open Clusters Catalog \citep{lynga87}

239: is 72~Myr for our sample clusters with a standard deviation of 38~Myr.

240: Previous age estimates based on isochrone fitting are generally within

241: our adopted age range.  \citet{mermilliod81} found that NGC~7790 and

242: NGC~6067 are about as old as the Pleiades (78~Myr) from models with

243: core overshooting.  \citet{meynet93} estimated $\approx70$~Myr for

244: NGC~5662 and NGC~6087 when the Pleiades is 100~Myr old from core-overshooting

245: models, while the age of NGC~6067 was estimated to be $\approx170$~Myr.

246: Cluster ages can also be inferred from the masses of Cepheids.  There are

247: several classical Cepheids in binary systems, and these have dynamical

248: masses of about $5 - 6 M_\odot$ \citep{bohm98,evans90,evans97,evans98,evans06}

249: over the period range similar to those of our sample [$0.65 \leq \log P

250: ({\rm days}) \leq 1.05$].  For our models, stars of this mass range have

251: an MS lifetime of about $50 - 80$~Myr.

252: 

253: \subsection{Photometry}\label{sec:phot}

254: 

255: For the photometry of the Cepheids, we used intensity-mean apparent

256: magnitudes in $BVI_C$ from \citet{berdnikov00}, which are on

257: the Cape (Cousins) system as realized by \citet{landolt83,landolt92}.

258: For CEa Cas and CEb Cas, the $I_C$-band measurements were taken from \citet{tammann03}.

259: We adopted $JHK$ photometry from \citet{laney92} on the \citet{carter90}

260: system and transformed it to the LCO system using the \citet{persson04}

261: transformation equations.  We assumed $0.025$~mag zero-point error in

262: the Cepheid photometry and treated this error independently from the cluster

263: photometry (see below).  Table~\ref{tab:list} lists the Cepheid photometry.

264: 

265: \input{tab2.tex}

266: 

267: For the cluster photometry, we compiled photoelectric and CCD $BVI_C$ data

268: on the Johnson-Cousins system from the literature, as well as from WEBDA

269: \citep{mermilliod03}.\footnote{See http://obswww.unige.ch/webda/webda.html.}

270: The references for the photometry are shown in Table~\ref{tab:phot},

271: where we used WEBDA's cross-identification of optical sources among

272: different references except for a few cases where we found missing

273: entries and misidentifications.  For each cluster we picked one or two

274: references to define a local standard in $V$ or $B - V$ by weighing factors

275: such as the number of observations, number of stars with photometry,

276: magnitude range in $V$, the photometric calibration procedure, and

277: whether the photometry was generally in agreement with data in other

278: studies.  The remaining columns of Table~\ref{tab:phot} show the mean

279: differences in $V$ and $B - V$ with respect to the local standard, and

280: the number of stars in common.  The errors shown are standard

281: errors of the mean difference.  For VDB~1 we combined the photoelectric

282: photometry by \citet{arp60} and CCD photometry by \citet[][their Table~2]

283: {turner98}, whose photometry was tied to the former study.

284: 

285: In most cases the differences between one study and another showed no

286: trends with magnitude.  In M25, however, the differences in $B - V$

287: between \citet{johnson60} and \citet{sandage60} were statistically

288: significant ($>2.5\sigma$):

289: \[

290: \Delta (B - V) ({\rm Sandage - Johnson}) = +0.009 (V - 10) - 0.010.\nonumber

291: \]

292: Similarly, in NGC~6087 there were significant trends with magnitude for

293: \citet{fernie61} and \citet{breger66} with respect to \citet{turner86}:

294: \begin{eqnarray}

295: \Delta V ({\rm Fernie - Turner}) = -0.033 (V - 10) - 0.052,\nonumber\\

296: \Delta V ({\rm Breger - Turner}) = -0.014 (V - 10) + 0.002.\nonumber

297: \end{eqnarray}

298: We applied these corrections to put them on

299: the same scale as the reference photometry.  For NGC~7790 we only used

300: the photometry by \citet{stetson00}, although we computed the differences

301: with respect to other studies and included these in Table~\ref{tab:phot}.

302: We applied zero-point corrections to individual values if the average

303: difference was significant at the $2.5\sigma$ level.  After shifting to

304: a common scale, magnitudes and colors in multiple studies were averaged together.

305: The weighted rms differences between the studies in Table~\ref{tab:phot}

306: are 0.026~mag in $V$ and 0.024~mag in $B - V$.  We therefore adopted

307: 0.025~mag as the characteristic size of systematic errors in the photometry.

308: 

309: Useful $V - I_C$ photometry for MS fitting is limited only to four

310: clusters: VDB~1 \citep{turner98}, Lyng{\aa}~6 \citep{walker85a},

311: NGC~6067 \citep{piatti98}, and NGC~7790 \citep{romeo89,gupta00,stetson00}.

312: As with $B - V$ we adopted the \citet{stetson00} photometry for NGC~7790.

313: Compared to Stetson's data, the photometry in \citet{romeo89}

314: is bluer by $0.026\pm0.003$~mag, and the photometry in \citet{gupta00}

315: is redder by $0.023\pm0.001$.  Assuming these values as a characteristic

316: size of error, we adopted $0.025$~mag for the photometric zero-point

317: error in $V - I_C$.

318: 

319: We combined optical photometry with $JHK_s$ measurements from the All

320: Sky Data Release of the 2MASS Point Source

321: Catalog (PSC)\footnote{See http://www.ipac.caltech.edu/2mass/.}.

322: WEBDA provides celestial coordinates for only a small fraction of stars in

323: this study.  For the others we identified each source on the images

324: from the Digitized Sky Survey,\footnote{The Digitized Sky Surveys were

325: produced at the Space Telescope Science Institute under US Government

326: grant NAG W-2166.  The images of these surveys are based on photographic

327: data obtained using the Oschin Schmidt Telescope on Palomar Mountain and

328: the UK Schmidt Telescope. The plates were processed into the present

329: compressed digital form with the permission of these institutions.} or

330: we computed the celestial coordinates from the plate position information

331: in WEBDA.  The rms difference between the retrieved and the 2MASS

332: coordinates was typically $0.5\arcsec$.  For NGC~6067 we matched 2MASS

333: sources with optical photometry only for those stars with good positional

334: accuracy.  The validation of the 2MASS source matches was confirmed from the

335: resulting tight optical and near-infrared color-color relations.  Based

336: on PSC flag parameters, we ignored the infrared data if sources were

337: undetected, blended, or contaminated.  Calibration errors in $JHK_s$

338: were taken as the uncertainty specified in the explanatory supplement

339: to the 2MASS All Sky Data Release: $0.011$ mag in $J$, $0.007$ mag in

340: $H$, and $0.007$ mag in $K_s$.\footnote{See

341: http://www.ipac.caltech.edu/2mass/releases/allsky/doc/explsup.html.}

342: 

343: Most studies do not report individual errors in optical magnitude or color,

344: so we computed the median of $V$ and $B - V$ differences from various studies

345: (after shifting to a common scale using the $>2.5\sigma$ criterion) and

346: assigned this value to represent the random photometric errors for all stars

347: in each cluster.  For $V - I_C$ we assigned 0.02~mag as the random error.

348: Errors for individual stars in  \citet{stetson00} were adopted without change.

349: The $V - K_s$ errors were computed as the quadrature sum of $V$ errors and

350: the catalog's ``total'' photometric uncertainties in $K_s$.

351: 

352: \begin{figure*}

353: \epsscale{1.0}

354: \plotone{f1.eps}

355: \caption{Cluster CMDs in $B - V$.  The bull's-eyes are stars remaining

356: after the photometric filtering, while open circles are stars rejected.  The

357: solid lines are empirically calibrated isochrones at the best-fitting

358: distance (Tables~\ref{tab:dist}), reddening (Tables~\ref{tab:ebv}), and

359: $R_{V,0}$ (Tables~\ref{tab:cluster}).  The boxes on the upper MS and

360: lower MS are regions where the color excess and distance were determined,

361: respectively.\label{fig:bvcmd}}

362: \end{figure*}

363: 

364: \begin{figure*}

365: \epsscale{1.0}

366: \plotone{f2.eps}

367: \caption{Same as Fig.~\ref{fig:bvcmd}, but CMDs in $V - I_C$.

368: \label{fig:vicmd}}

369: \end{figure*}

370: 

371: \begin{figure*}

372: \epsscale{1.0}

373: \plotone{f3.eps}

374: \caption{Same as Fig.~\ref{fig:bvcmd}, but CMDs in $V - K_s$.  The

375: dotted line represents the 2MASS completeness limit.\label{fig:vkcmd}}

376: \end{figure*}

377: 

378: \begin{figure*}

379: \epsscale{1.0}

380: \plotone{f4.eps}

381: \caption{Same as Fig.~\ref{fig:bvcmd}, but CMDs in $J - K_s$.  The

382: dotted line represents the 2MASS completeness limit.\label{fig:jkcmd}}

383: \end{figure*}

384: 

385: The cluster CMDs in various color combinations are shown in

386: Figures~\ref{fig:bvcmd} through \ref{fig:jkcmd}.  Stars rejected as being

387: far from the MS are shown as open circles, while those used in the MS fitting

388: are shown as bull's-eyes.  The curved line in each panel is the best-fitting

389: isochrone on each CMD.  Details of the MS fitting and outlier rejection

390: are given in the next section.  Throughout this paper we denote CMDs by

391: their color and luminosity indices as $(B - V, V)$, $(V - I_C, V)$,

392: $(V - K_s, V)$, $(J - K_s, K_s)$, and $(H - K_s, K_s)$.

393: 

394: \section{Main-Sequence Fitting}

395: 

396: \subsection{Extension of Isochrone Calibration}\label{sec:iso}

397: 

398: In our first two papers of this series (Paper~I and Paper~II),

399: we assessed the accuracy

400: of distances from MS fitting and examined systematic errors in the

401: transformation of theoretical to observational quantities.  In Paper~I

402: we demonstrated that stellar models from the Yale Rotating Evolutionary

403: Code \citep{sills00} are in  agreement with the masses and luminosities

404: of the well-studied Hyades eclipsing binary vB~22 \citep{torres02}.

405: These models also satisfy stringent tests from helioseismology and

406: predict solar neutrino fluxes in line with observations \citep{basu00,

407: bahcall01,bahcall04}.  In Paper~II we showed that the models provide

408: a good match to the spectroscopically determined temperatures

409: \citep{paulson03} of individual Hyades members with good parallaxes

410: \citep{debruijne01}.  However, we found that any of the widely-used

411: color-$T_{\rm eff}$ relations \citep[e.g.,][hereafter LCB]{alonso95,alonso96,

412: lejeune97,lejeune98} fail to reproduce the observed shapes of the MS

413: in the Hyades, having differences in broadband colors as large as

414: $0.1$~mag.  The existence of these systematic errors in the colors, in

415: the presence of agreement between the spectroscopic and theoretical

416: luminosity-$T_{\rm eff}$ scales, led us to argue that the problems lie

417: with the adopted color-$T_{\rm eff}$ relations instead of errors in

418: the theoretical $T_{\rm eff}$ scale.  We then adopted the LCB

419: color-$T_{\rm eff}$ relations and computed empirical corrections to the

420: LCB relations to match photometry in the Hyades.  In Paper~III we showed

421: that isochrones with the empirical color-$T_{\rm eff}$ corrections

422: accurately match the MS shapes of other nearby clusters in three color

423: indices ($B - V$, $V - I_C$, and $V - K_s$), yielding distance estimates

424: with errors as small as 0.04~mag in distance modulus (or 2\% in distance).

425: 

426: However, the Hyades-based calibration is of limited use for this study.

427: The calibration is reliable only at $M_V \ga 3$ because the number of

428: stars in the upper MS of the Hyades is small in this relatively old open cluster

429: \citep[550~Myr from models excluding convective overshoot;][]{perryman98}.

430: On the other hand, photometry of our sample clusters is quite sparse or

431: absent in this lower MS part.  It is therefore desirable to define

432: an extension of the empirical color-$T_{\rm eff}$ corrections for hotter

433: and brighter stars.  A good choice for a calibrating cluster is the Pleiades

434: for several reasons.  First of all, its geometric distance is now known to

435: a precision comparable to that of the Hyades, $(m - M)_0 = 5.63\pm0.02$

436: (see Paper~III and references therein for a detailed discussion of

437: the cluster parameters).  Second, it has low reddening, $E(B - V) = 0.032\pm0.003$.

438: Third, a precise metal abundance from high-resolution spectra is available,

439: ${\rm [Fe/H]} = +0.04\pm0.02$.

440: Furthermore, Cepheid clusters tend to be of ages comparable to that of

441: the Pleiades, making it a good template for distance and reddening

442: estimates.

443: 

444: \input{tab3.tex}

445: 

446: \begin{figure*}

447: \epsscale{0.9}

448: \plotone{f5.eps}

449: \caption{Empirically calibrated isochrones for the Pleiades ({\it solid

450: lines}).  The dot-dashed lines show the isochrones without the color

451: corrections.  The open circles are stars retained for use in the

452: calibration process, and the small plus signs show stars rejected

453: (see text).\label{fig:plcmd}}

454: \end{figure*}

455: 

456: The sources of our Pleiades photometry are described in Paper~III.  We

457: began by removing a few known binaries and then rejected stars that are

458: more than 0.1 mag in color away from a 15 point median of the sample

459: sorted in $V$; this step was independent of the theoretical isochrones.

460: Table~\ref{tab:plphot} lists photometry of the hot Pleiades stars that

461: remained after this selection.  The CMDs of these stars are shown in

462: Figure~\ref{fig:plcmd}.  Small plus signs denote stars rejected as far

463: from the MS, while open circles show stars retained for use in the

464: following calibration process.

465: 

466: \begin{figure}

467: \epsscale{1.1}

468: \plotone{f6.eps}

469: \caption{Differences in the bolometric correction and colors between

470: ``corrected'' and ``original'' LCB tables for MS stars ([M/H] = 0,

471: $\log{g}=4.5$).  Differences are in the sense of the former minus the

472: latter values.\label{fig:lj}}

473: \end{figure}

474: 

475: In the course of this work we became aware of some problematic features

476: in the LCB color tables.  They computed synthetic colors from both the

477: ``original'' and the ``corrected'' model flux distributions (hereafter

478: original and corrected LCB tables, respectively).  In Figure~\ref{fig:lj}

479: we compare the corrected LCB table with the original LCB table for MS stars

480: ($\log{g}=4.5$).  The LCB empirical color-$T_{\rm eff}$ relations

481: (corrected LCB table) are

482: defined only for $T_{\rm eff} < 11500~K$, which results in an artificial

483: jump in CMDs, particularly in the $V$-band bolometric correction.  In

484: addition, there are small scale structures in the color corrections

485: (especially in $V - I_C$) that cause interpolation noise in our isochrones

486: at $T_{\rm eff} \ga 8000$~K.  We therefore used the original LCB table above

487: 8000~K, the corrected LCB table below 6500~K, and a linear ramp between

488: the two tables in order to produce smoother base isochrones.

489: Isochrones at an age of 100~Myr generated with these ``merged''

490: color-$T_{\rm eff}$ relations are shown as dot-dashed lines in

491: Figure~\ref{fig:plcmd}.  However, these still fail to match the

492: observed MS of the Pleiades and therefore require further corrections.

493: 

494: \begin{figure}

495: \epsscale{1.1}

496: \plotone{f7.eps}

497: \caption{Empirical color-$T_{\rm eff}$ corrections ({\it solid line}).

498: The filled circles show the differences in colors between the Pleiades

499: data and the uncorrected isochrone at a constant $M_V$.  The sense of

500: the difference is the former minus the latter values.  The open circles

501: are those for the Hyades stars.\label{fig:corr}}

502: \end{figure}

503: 

504: We derived empirical corrections to the isochrone in $B - V$, $V - I_C$,

505: $V - K_s$, and $J - K_s$ by forcing a match to the Pleiades photometry

506: as done for the Hyades (Paper II).  Figure~\ref{fig:corr} plots the

507: difference in color between individual stars and an isochrone generated

508: with the merged LCB color-$T_{\rm eff}$ relations.  The $T_{\rm eff}$

509: was calculated from $M_V$.  Filled circles are for the Pleiades stars,

510: and the open circles are for the Hyades stars used in the calibration of

511: the lower MS.  The Pleiades stars have smaller error bars than the

512: Hyades points because we assumed a negligible distance spread for the

513: Pleiades, while the Hyades ones reflect parallax errors for individual

514: stars \citep{debruijne01}.  The solid lines in Figure~\ref{fig:corr}

515: represent our empirical corrections, which were computed at a constant

516: $M_V$ in a seven-point moving window.  Given the modest sample size and

517: intrinsic scatter in the upper MS, we did not attempt to resolve out

518: smaller scale structures.  We defined the empirical color-$T_{\rm eff}$

519: corrections as those for the Pleiades at $T_{\rm eff} \geq 6838$~K and

520: for the Hyades at $T_{\rm eff} \leq 6500$~K.  A spline function was used

521: to bridge over the narrow $T_{\rm eff}$ gap between these two corrections.

522: We also used $H - K_s$ in MS fitting but without color corrections

523: because photometric errors were too large in this color to provide

524: a reliable calibration.

525: 

526: \input{tab4.tex}

527: 

528: Applying these corrections to the colors at each $T_{\rm eff}$ defines

529: the empirically calibrated isochrone for the Pleiades, which is tabulated

530: along with the color corrections in Table~\ref{tab:pliso}.  Isochrones

531: incorporating the corrections for the Pleiades are shown as solid lines

532: in Figure~\ref{fig:plcmd}.  Systematic errors in the color corrections

533: are $\sim0.01$~mag from the errors in the adopted distance, metallicity,

534: and reddening values for the Pleiades.  We assumed that our empirical

535: corrections in the upper MS are independent of metallicity and age

536: and applied them to all isochrones generated in Paper~III.  The isochrones

537: constructed in this way are available online

538: at\\{\tt http://www.astronomy.ohio-state.edu/iso/}.

539: 

540: There is one important functional difference between the calibration

541: for luminous stars and that for the lower MS stars.  The MS for FGK-type

542: stars is intrinsically very narrow.  As a result, for the low-mass stars

543: we could plausibly claim that the empirical cluster sequence could be

544: used to redefine the color-$T_{\rm eff}$ relationship.  On the other

545: hand, the dispersion about the mean trend for upper MS stars is

546: significantly larger.  For the Pleiades, we found a dispersion of 0.02~mag

547: in $B - V$ for $0.0 \la B - V \la 0.40$ (after excluding outliers),

548: which is significantly above the photometric precision.  A comparable

549: increase was seen in $V - I_C$ and $V - K_s$ (0.02 and 0.04~mag, respectively).

550: Although this could be due to differential reddening across the cluster

551: field \citep{breger86}, the persistence of rapid rotation in early-type

552: stars can impact both their evolution and the mapping of $T_{\rm eff}$

553: onto colors \citep[e.g.,][]{collins77} as a consequence of the

554: \citet{kraft65} break in rotational properties.  Furthermore, these

555: effects depend on both the rotation rate and the inclination to the line

556: of sight and can be as large as

557: 0.1 mag in $B - V$.  Nonetheless, the mean cluster locus can still be

558: used to define a template for distance studies but should not be

559: interpreted as a direct change in the color-$T_{\rm eff}$ relationship.

560: In other words, our empirical corrections remove systematic trends from

561: the rotation-induced color anomalies, as well as from any intrinsic

562: problems in the adopted color-$T_{\rm eff}$ relationship.

563: 

564: \subsection{Main-Sequence Fitting}\label{sec:msfit}

565: 

566: We determined the distance, reddening, and the ratio of total to

567: selective extinction in $V$ [$R_V \equiv A_V / E(B - V)$] for each cluster

568: via an iterative approach, which fits an isochrone to individual CMDs of

569: various color indices.  All of these parameters are essential to

570: obtain self-consistent solutions for the absolute magnitude of a Cepheid

571: (\S~\ref{sec:pl}).  Our MS-fitting process also includes rejection of

572: stars that are too far away from an MS ridge-line, defined by the locus of

573: points with the highest density on each CMD.  Individual color excesses

574: in each cluster were used to derive $R_V$ according to the extinction law

575: described by \citet[][hereafter CCM89]{cardelli89}.  The rest of this

576: section describes the MS-fitting process in detail.

577: 

578: Our sample clusters do not have detailed radial velocity or proper-motion

579: membership studies.  Before isochrone fitting we therefore

580: applied the photometric filtering procedure described in Paper~III to

581: identify and reject stars that are likely foreground/background objects

582: or cluster binaries.  The procedure iteratively identifies a cluster MS

583: ridge-line independently of the isochrones and rejects stars if they are

584: too far away from the ridge-line.  This was done by computing $\chi^2$

585: in each CMD as

586: \begin{equation}

587: \chi^2 = \sum_{i=1}^N \chi_i^2

588: = \sum_{i=1}^N \frac {(\Delta X_i)^2 }

589:  {\sigma_{X,i}^2 + (\gamma_i^{-1} \sigma_{V,i})^2 + \sigma_0^2},

590: \label{eq:chifilter}

591: \end{equation}

592: where $\Delta X_i$ is the color difference between the $i^{th}$ data

593: point and MS at the same $V$; $\sigma_{X,i}$ and $\sigma_{V,i}$ are

594: photometric errors in color and $V$, respectively.  The error in

595: $V$ contributes to the error in $\Delta X_i$ by the inverse slope of

596: the MS, $\gamma_i^{-1}$.  We added $\sigma_0$ in quadrature to the

597: propagated photometric errors in the denominator to take into account

598: the presence of differential reddening, cluster binaries, non-cluster

599: members, and other effects that would increase the MS width above the

600: photometric precision.  We adjusted the value of $\sigma_0$ so that

601: the total $\chi^2$ is equal to the total number of stars $N$ used in

602: the filtering.

603: 

604: Initially we rejected all data points as outliers if $\chi_i^2$ (the

605: individual contribution to $\chi^2$) was greater than 9 (corresponding

606: to a $3\sigma$ outlier). We repeated adjusting $\sigma_0$ and rejecting

607: outliers with the reduced set of data points until there remained no

608: point with $\chi_i^2 > 9$.  We combined the results from all CMDs, and rejected

609: stars if they were tagged as an outlier at least in one of the CMDs.

610: Missing data in any colors was not a criterion for rejection.  For this

611: study we included $(J - K_s, K_s)$ and $(H - K_s, K_s)$ in the filtering

612: procedure.  We reduced the rejection threshold at each iteration until

613: it was limited to $2.0$--$2.4\sigma$, but our result is insensitive to

614: the final rejection threshold within the fitting errors.

615: 

616: Our filtering results are shown in Figures~\ref{fig:bvcmd}--\ref{fig:jkcmd}.

617: The bull's-eyes represent stars retained and used in the following MS fits,

618: and open circles are the ones rejected by the filtering.  At the end of the

619: iteration, $\sigma_0$ values in $(B - V, V)$ and $(V - K_s, V)$ were

620: (0.055~mag, 0.101~mag) in NGC~129, (0.050~mag, 0.102~mag) in VDB~1,

621: (0.023~mag, 0.064~mag) in NGC~5662, (0.048~mag, 0.102~mag) in Lyng{\aa}~6,

622: (0.031~mag, 0.078~mag) in NGC~6067, (0.033~mag, 0.058~mag) in NGC~6087,

623: (0.043~mag, 0.111~mag) in M25, and (0.011~mag, 0.026~mag) in NGC~7790.

624: These excess dispersions are most likely due to differential reddening

625: of heavily reddened, young open clusters.  Note that the ratio of $\sigma_0$

626: values from $(B - V, V)$ and $(V - K_s, V)$ is also approximately equal to

627: the color-excess ratio $E(V - K_s)/E(B - V) \approx 2.8$ (see below).

628: We discuss in detail the effects of differential reddening on {\it P-L}

629: relations in \S~\ref{sec:pl}.

630: In $(J - K_s, K_s)$ and $(H - K_s, K_s)$, photometric errors were usually

631: large enough that $\sigma_0$ values were automatically set to zero.

632: 

633: The fitting process took advantage of the changing slope of the MS with

634: absolute magnitude.  Near the MS turnoff ($M_V \leq 1.5$), the MS is

635: nearly vertical, allowing a precise determination of the reddening.

636: Further down ($1.5 \leq M_V \leq 3.5$), the slope of the MS is still

637: relatively steep [e.g., $\Delta M_V / \Delta (B - V) \approx 5$], but

638: it nevertheless provides sufficient leverage for determining the distance.

639: We divided each CMD into two zones, shown as boxes in

640: Figures~\ref{fig:bvcmd}--\ref{fig:jkcmd}, which define the regions used

641: for determining the color excess and distance.  The color excess was determined

642: for stars with  $-1.0 \leq M_V \leq 2.0$ in $(B - V, V)$, $(V - I_C, V)$,

643: and $(V - K_s, V)$, and at $-0.8 \leq M_K \leq 1.7$ in $(J - K_s, K_s)$

644: and $(H - K_s, K_s)$.  Distance determinations came from stars with

645: $-0.1 \leq (B - V)_0 \leq 0.5$ and the corresponding color ranges in other

646: colors with the same $M_V$.  For Lyng{\aa}~6 and NGC~7790, $(V - K_s, V)$

647: and $(J - K_s, K_s)$ were not used in the distance estimation because

648: the lower parts of the MS are below the 2MASS completeness limit

649: ($K_s \approx 14.3$).

650: 

651: \input{tab5.tex}

652: 

653: Since the color excess and $R_V$ depend on the intrinsic color of

654: stars, we adopted the color-dependent relations from \citet[][

655: hereafter BCP98]{bessell98} and M. S. Bessell (2007, private communication).

656: Their formulae were

657: based on the theoretical stellar spectral energy distributions with

658: the extinction law from \citet{mathis90}.  Specifically,

659: their formulation yields

660: \begin{equation}

661: R_V = R_{V,0} + 0.22 (B - V)_0,

662: \label{eq:rv}

663: \end{equation}

664: \begin{equation}

665: E(B - V) = E(B - V)_0 \{1 - 0.083 (B - V)_0 \},

666: \label{eq:ebv}

667: \end{equation}

668: where $(B - V)_0$ is an intrinsic color of each star.  The $R_{V,0}$

669: and $E(B - V)_0$ are the values for zero-color stars, where $R_{V,0}$

670: is $3.26$ in their model.  For other colors we used the following

671: color-excess ratios:

672: \begin{eqnarray}

673: \frac{E(V - I_C)}{E(B - V)} &=& 1.30 + 0.06 (V - I_C)_0 \\

674: \frac{E(V - K_s)}{E(B - V)} &=& 2.88 + 0.07 (V - K_s)_0 \\

675: \frac{E(J - K_s)}{E(B - V)} &=& 0.56 + 0.06 (J - K_s)_0 \\

676: \frac{E(H - K_s)}{E(B - V)} &=& 0.20 + 0.15 (H - K_s)_0.

677: \label{eq:color}

678: \end{eqnarray}

679: The first terms represent the color-excess ratios from CCM89 at

680: $R_{V,0} = 3.26$, and the second terms are color-dependent relations

681: from BCP98.  Differences

682: between the original coefficients in BCP98 and those for the 2MASS

683: $JHK_s$ system using transformation equations in \citet{carpenter01}\footnote{

684: Updated color transformations for 2MASS All-sky Data Release can be found at\\

685: http://www.ipac.caltech.edu/2mass/releases/allsky/doc/sec6\_4b.html.}

686: were found to be negligible.  Table~\ref{tab:ebvlaw} lists color-excess

687: ratios for zero-color stars from CCM89 and BCP98, as well as observationally

688: derived values from \citet{schultz75}, \citet{dean78}, \citet{sneden78},

689: and \citet{rieke85}.  Color-excess ratios determined from our cluster

690: sample are also shown in the fourth column, which is discussed in

691: the next section.

692: 

693: The fitting process begins with a guess at the average distance.  At

694: this distance, individual color excesses for zero-color stars [$E(B - V)^{\rm obs}_0$,

695: $E(V - I_C)^{\rm obs}_0$, etc.] were determined using stars on the upper MS.

696: These were used to determine the best-fit $R_{V,0}$ and $E(B - V)_0$

697: from the CCM89 extinction law by minimizing

698: \begin{equation}

699: \chi^2 = \sum_{i=1}^5 \biggr[\frac {E(\lambda_i - V)^{\rm obs}_0

700: - E(\lambda_i - V)^{\rm CCM89}_0}{\sigma_{E(\lambda - V)}} \biggr]^2,

701: \label{eq:rvchi}

702: \end{equation}

703: where $\lambda_i$ represents effective wavelengths in $BI_C$

704: \citep{bessell98} and isophotal wavelengths in $JHK_s$ \citep{cohen03}.

705: The index in the summation runs over $B$, $I_C$, $J$, $H$, and $K_s$.

706: The $\sigma_{E(\lambda - V)}$ is the fitting error in the upper MS,

707: which was determined from the scatter of the points around the isochrone

708: and the individual errors in color.  The CCM89 color excess,

709: $E(\lambda - V)^{\rm CCM89}_0$, can be expressed as

710: \begin{equation}

711: E(\lambda - V)^{\rm CCM89}_0

712: = R_{V,0} E(B - V)_0 \left( \frac {A_{\lambda}}{A_V} - 1 \right),

713: \label{eq:ccm}

714: \end{equation}

715: where ${A_{\lambda}} / {A_V}$ is a function of $R_{V,0}$

716: from polynomial relations in CCM89.  Note that $R_{V,0}$ and $E(B - V)_0$

717: in equation~(\ref{eq:rvchi}) are two parameters to be determined and

718: that $E(B - V)_0$ in equation~(\ref{eq:ccm}) is the normalization factor,

719: which is different from the input color excess in $(B - V, V)$.

720: 

721: The fitting procedure then found weighted median distances on each CMD by

722: using the stars in the lower part of the MS.  The weighted average distance

723: from all CMDs was computed, and the process was repeated until convergence,

724: which occurred when the difference in the average distance from the previous

725: iteration became smaller than its propagated error.

726: 

727: \begin{figure}

728: \epsscale{1.1}

729: \plotone{f8.eps}

730: \caption{Example of determining $R_{V,0}$ and $E(B - V)_0$ for NGC~6067.

731: {\it Top:} Solid line represents the best-fitting CCM89 extinction

732: curve to color excesses ({\it open circles}).

733: {\it Bottom:} Likelihood contours in $R_{V,0}$ and $E(B - V)_0$ shown at

734: $\Delta \chi^2$ = 2.30, 6.17, and 11.8 (68.3\%, 95.4\%, and 99.73\%

735: confidence levels for 2 degrees of freedom).\label{fig:6067.ebv}}

736: \end{figure}

737: 

738: An example of the fit is shown in Figure~\ref{fig:6067.ebv} for NGC~6067.

739: The solid line in the top panel is the best-fitting CCM89 extinction

740: curve to the data ({\it open circles}).  The bottom panel shows likelihood

741: contours in $R_{V,0}$ and $E(B - V)_0$ drawn at $\Delta \chi^2$ = 2.30, 6.17, and 11.8

742: from the minimum $\chi^2 = 4.5$ (68.3\%, 95.4\%, and 99.73\% confidence

743: levels for 2 degrees of freedom).

744: 

745: \section{Cluster Properties}

746: 

747: In this section we present our results on the distance, reddening, and

748: $R_V$ of the sample clusters and evaluate various systematic errors in

749: these parameters.  In particular, we assess the accuracy of MS-fitting

750: based on internal consistency of distances from multi-color CMDs.  We

751: show that our distance estimates are generally shorter than

752: previous ZAMS-fitting results and that this is mainly due

753: to our lower reddening estimates.

754: 

755: \subsection{Distance, Reddening, and $R_V$}\label{sec:par}

756: 

757: \input{tab6.tex}

758: \input{tab7.tex}

759: 

760: Our MS-fitting results on the distance and color excess from individual

761: CMDs are shown in Tables~\ref{tab:dist} and \ref{tab:ebv}, respectively,

762: along with a number of stars used in the fit.  The last two columns in

763: Table~\ref{tab:dist} show the weighted average distance modulus from

764: all available CMDs with its propagated error and the standard deviation

765: ($\sigma$) of individual distances.  The last column in Table~\ref{tab:ebv}

766: lists a standard deviation of $E(B - V)$ from all CMDs, after transforming

767: individual color-excess values in $V - I_C$, $V - K_s$, and $J - K_s$ to

768: $E(B - V)$ using the CCM89 reddening law (Table~\ref{tab:ebvlaw}).  We

769: excluded $E(H - K_s)$ from this computation since we found a large

770: systematic deviation of our color-excess estimates from the CCM89 law

771: as discussed below.

772: 

773: \begin{figure}

774: \epsscale{1.1}

775: \plotone{f9.eps}

776: \caption{Distribution of individual distance moduli for NGC~7790 with

777: respect to the isochrones shown in Fig.~\ref{fig:bvcmd} and \ref{fig:vicmd}.

778: The open histogram is the distribution of all stars in the isochrone

779: fitting range where the distance was determined.  The hatched histogram

780: shows only for those that passed the photometric filtering.  The vertical

781: line indicates the weighted median distance modulus.\label{fig:7790res}}

782: \end{figure}

783: 

784: In Figures~\ref{fig:bvcmd}--\ref{fig:jkcmd} the best-fitting isochrones

785: are overlaid on the cluster CMDs.  While there is some variation in

786: the quality of the data, we could divide the sample clusters into three

787: groups based on how tight the MS is on optical CMDs.  The first group

788: contains NGC~6067 and NGC~7790, which have well-populated CMDs and

789: small fitting errors on each CMD.  They also have a well-defined peak on

790: the distribution of an individual star's distance modulus as shown in

791: Figure~\ref{fig:7790res} for NGC~7790.  The second group is NGC~129,

792: NGC~5662, Lyng{\aa}~6, NGC~6087, and M25.  These clusters have less

793: well-defined peaks, either because of strong differential reddening

794: (\S~\ref{sec:pl}) or because of a bias in the sampling from photoelectric

795: observations at faint magnitudes.  The last group is VDB~1, which

796: has a sparse MS and correspondingly large distance uncertainties.

797: 

798: We can also divide the clusters into groups by examining the difference

799: in derived cluster properties from different colors.  Clusters where

800: the results from multi-color CMDs have a high internal consistency are

801: strong cases, while clusters with substantial differences indicate

802: problems.  For the above first and second groups of clusters except

803: Lyng{\aa}~6, the standard deviations of individual distances

804: [$\sigma_{(m - M)}$ in Table~\ref{tab:dist}] are 0.03--0.08~mag, which

805: are generally consistent with the fitting error on individual CMDs.

806: On the other hand, both Lyng{\aa}~6 and VDB~1 exhibit a larger

807: dispersion in distance [$\sigma_{(m - M)} = 0.21$ and $0.13$~mag,

808: respectively], and they also have a larger standard deviation of $E(B - V)$

809: from different colors ($\sigma$ in Table~\ref{tab:ebv}).

810: 

811: However, the dispersion in Lyng{\aa}~6 is consistent with the statistical

812: scatter expected from the small sample size.  By contrast, VDB~1 has a

813: statistically unreliable result that demands its exclusion from our data

814: set.  For this cluster, the best distance estimates inferred from

815: different colors disagree by much more than their individual error

816: estimates would imply.  One possible explanation for the internal

817: inconsistency is that the cluster MS is contaminated by less reddened

818: field stars as shown by spectral-type study \citep{preston64}.

819: Because these stars happen to lie on the cluster MS, our photometric

820: filtering could have misidentified cluster members.  We therefore

821: exclude VDB~1 from the following analysis and leave it for future studies.

822: 

823: \input{tab8.tex}

824: \input{tab9.tex}

825: 

826: Table~\ref{tab:error} shows the error budget for NGC~6067 as

827: an example.  The first column displays individual sources of error,

828: and the second column shows the size of the errors adopted for each

829: quantity.  The third through fifth columns list error contributions

830: to $R_{V,0}$, $E(B - V)_0$, and $(m - M)_0$, respectively.  The size of

831: systematic errors in cluster metallicity, age, and photometry was

832: discussed in \S~\ref{sec:data}.  The quantity $\Delta Y$ represents

833: the uncertainty in the helium abundance at fixed metallicity and comes

834: from the consideration of the initial solar abundance and those of the Hyades

835: and the Pleiades (Paper~I and Paper~III).  The fitting errors represent

836: the internal precision of the fit and were taken as the larger value

837: of the propagated error or $\sigma_{(m-M)}/\sqrt{N}$ in Table~\ref{tab:dist}.

838: We estimated the fitting errors in $R_{V,0}$ and $E(B - V)_0$ from

839: the size of the $1\sigma$ contours on the $\Delta \chi^2$ distribution

840: (e.g., Fig.~\ref{fig:6067.ebv}).  The total systematic error is the

841: quadrature sum of these errors.  The $R_{V,0}$, $E(B - V)_0$, and

842: $(m - M)_0$ of our cluster sample are listed in Table~\ref{tab:cluster}

843: with total systematic errors in these quantities.

844: 

845: \begin{figure}

846: \epsscale{1.2}

847: \plotone{f10.eps}

848: \caption{Mean color-excess ratios from the cluster sample.  The error

849: bars are from the isochrone fitting only.  The solid line is a fit to

850: the data, constrained to pass through the origin.  The dashed line is

851: the color-excess ratio from the CCM89 reddening law.\label{fig:ebv}}

852: \end{figure}

853: 

854: Independently of the CCM89 reddening law, we also derived color-excess ratios

855: based on our cluster sample.  Figure~\ref{fig:ebv} shows a correlation

856: between $E(B - V)$ and color-excess values in other color indices.

857: Individual clusters are represented as a point, and fitting errors are

858: shown only.  The solid line is a linear fit to the data, constrained to

859: pass through the origin.  Table~\ref{tab:ebvlaw} summarizes our best-fit

860: slopes of these lines or color-excess ratios.  Our values are generally

861: in good agreement with the CCM89 law and other observational estimates.

862: However, our method yields a lower $E(H - K_s) / E(B - V)$ than these

863: studies, which may reflect a zero-point offset in $H - K_s$ of our

864: isochrones.  Note that we did not derive color corrections for $H - K_s$.

865: 

866: \subsection{Comparison with Previous Studies}\label{sec:comp}

867: 

868: \input{tab10.tex}

869: 

870: \begin{figure}

871: \epsscale{1.1}

872: \plotone{f11.eps}

873: \caption{Comparison of cluster distances with those in \citet{feast87},

874: \citet{tammann03}, \citet{hoyle03}, and \citet{gieren05}.  The

875: differences are in the sense of their distances minus those in this

876: paper.  The error bars are quadrature sum from the two studies in the

877: lower two panels.  Individual distances in the top two panels

878: were revised assuming our Pleiades distance scale.\label{fig:comp.dist}}

879: \end{figure}

880: 

881: The most up-to-date compilation of Cepheid parameters in open clusters

882: and associations can be found in \citet{tammann03}.  Their distances are

883: from the earlier compilation by \citet{feast99}.  These ZAMS-fitting

884: distances are the same as in \citet{feast87}, but include more recent

885: distance estimates for NGC~129 \citep{turner92} and NGC~7790 \citep{romeo89}.

886: All of our Cepheids are listed in these tables, and the comparisons with

887: distances in \citeauthor{feast87} and \citeauthor{tammann03} are shown in the upper

888: two panels of Figure~\ref{fig:comp.dist} and listed in

889: Table~\ref{tab:comp.dist}.  Both of these studies estimated distances

890: relative to the Pleiades but assumed different Pleiades distance of

891: $(m - M)_0 = 5.57$ and $5.61$~mag, respectively.  We revised their

892: distances assuming our Pleiades distance, $(m - M)_0 = 5.63$.

893: Since they did not explicitly present errors, we plotted our errors only.

894: The average differences in distance are $\Delta \langle (m - M)_0

895: \rangle = 0.09\pm0.05$ and $0.12\pm0.05$ with \citeauthor{feast87} and

896: \citeauthor{tammann03}, respectively.  The sense of the differences

897: is that our new distance estimates are shorter on average.  The standard

898: deviations of the differences are $0.14$~mag for both

899: comparisons, while the expected size from our error estimates is

900: $0.07$~mag.  If our error estimates are correct, the dispersion

901: indicates a random error of $0.12$~mag in distance estimates from

902: \citeauthor{feast87} and \citeauthor{tammann03}.

903: 

904: In addition to these studies, \citet{hoyle03} derived ZAMS-fitting

905: distances to 11 open clusters with new photometry in $UBVK$.  Our

906: comparison with this study is shown in Figure~\ref{fig:comp.dist}

907: and tabulated in Table~\ref{tab:comp.dist}.  The error bars represent

908: quadrature sums of our and their reported error estimates.  Their

909: distances are marginally consistent with our estimates, and the average

910: difference is $\Delta \langle (m - M)_0 \rangle = 0.04\pm0.07$, our

911: distances being shorter on average.  However, we did not attempt to revise

912: their distances assuming our Pleiades distance scale.  They used a ZAMS from

913: \citet{allen73} for the optical data, but they also claimed that the

914: same results were obtained from the ZAMS in \citet{turner79} and

915: \citet{mermilliod81}.  Since the Pleiades distance modulus of 5.56~mag

916: was adopted in \citeauthor{turner79}, the distance moduli in \citeauthor{hoyle03}

917: would then become longer by $\approx0.07$~mag on average than ours when

918: they are on the same Pleiades distance scale as in this paper.

919: 

920: The bottom panel in Figure~\ref{fig:comp.dist} shows the comparison

921: with distances from the surface brightness technique by \citet{gieren05}.

922: Their distances are also larger on average than our estimates by

923: $0.15\pm0.09$~mag as in the previous MS-fitting studies.  However,

924: they recalibrated the projection factor, which was used to convert

925: observed velocities to pulsational velocities, using MS-fitting

926: distances to cluster Cepheids.  Therefore, their estimates are not

927: completely independent from the above MS-fitting studies.

928: 

929: \input{tab11.tex}

930: 

931: \begin{figure}

932: \epsscale{1.1}

933: \plotone{f12.eps}

934: \caption{Comparison of Cepheid $E(B - V)$ with those in \citet{feast87},

935: \citet{fernie95}, \citet{tammann03}, and \citet{hoyle03}.  The differences

936: are in the sense of their values minus those in this paper.  Error bars

937: represent our error estimates in the top three panels, but they show

938: a quadrature sum from two studies in the bottom panel (Hoyle et~al.).

939: \label{fig:comp.ebv}}

940: \end{figure}

941: 

942: Our short distances are mainly due to our lower $E(B - V)$ values than

943: those in the previous work.  This is because the MS-fitting distances

944: are correlated with reddening by $\Delta (m - M)_0 / E(B - V) \sim 2$

945: (Paper~III).  The top panel in Figure~\ref{fig:comp.ebv} compares

946: the Cepheid reddening with those in \citet{feast87}, and individual

947: estimates are listed in Table~\ref{tab:comp.ebv}.  They estimated the OB

948: star reddening at the location of the Cepheids and then transformed it to

949: the value appropriate for the Cepheid using a color-dependent reddening

950: law (e.g., eqs.~[\ref{eq:rv}] and [\ref{eq:ebv}]).  The mean difference from

951: our study is $\langle E(B - V) \rangle = 0.04\pm0.02$~mag, our values

952: being smaller on average.  The difference in $E(B - V)$ then

953: approximately matches the size expected to produce the difference in

954: distance modulus ($\sim0.1$~mag).

955: 

956: Individual reddening estimates from \citet{fernie95}, \citet{tammann03},

957: and \citet{hoyle03} are also listed in Table~\ref{tab:comp.ebv}, and

958: comparisons with our estimates are shown in Figure~\ref{fig:comp.ebv}.

959: Many studies used an extensive compilation of Cepheid reddening

960: in the David Dunlap Observatory (DDO) database of Galactic classical

961: Cepheids,\footnote{http://www.astro.utoronto.ca/DDO/research/cepheids/}

962: which is from heterogeneous methods \citep{fernie90,fernie94,fernie95}.

963: The mean difference with their data is $\Delta \langle E(B - V) \rangle

964: = 0.05$ in the sense that our reddening is smaller.  \citeauthor{tammann03}

965: adopted these reddening estimates for their cluster Cepheids but made

966: slight corrections to remove a mild correlation of the color excess

967: with residuals from the period-color relation.  This results in a reduced

968: difference with our estimates by $\Delta \langle E(B - V) \rangle = 0.02$.

969: However, the results from \citeauthor{tammann03} are internally less

970: consistent, in the sense

971: that the extinction corrections on the Cepheid magnitudes and the

972: reddening values for the MS-fitting distance estimates were not derived

973: from the same methods.  Finally, we compared our reddening values with those

974: in \citeauthor{hoyle03}, which were based on the cluster $UBV$

975: color-color diagram.  Their values are also systematically

976: larger than ours by $\Delta \langle E(B - V) \rangle = 0.04\pm0.01$.

977: 

978: There are a limited number of reddening estimates from the spectral

979: types in the literature.  \citet{kraft58} estimated $E(B - V) = 0.45$

980: from one B-type star in NGC~129, while \citet{arp59} found $\langle

981: E(B - V) \rangle = 0.53\pm0.01$ from seven stars in the vicinity of its

982: Cepheid (DL~Cas).  In M25, \citet{sandage60} estimated $\langle E(B - V)

983: \rangle = 0.43\pm0.05$ from the spectral types given by \citet{feast57}

984: and $0.43\pm0.06$ by \citet{wallerstein57}, but found a slightly larger

985: reddening, $0.46\pm0.04$, in the vicinity of its Cepheid (U~Sgr).  Finally,

986: \citet{kraft58} estimated $\langle E(B - V) \rangle = 0.49\pm0.02$ from

987: six stars in NGC~7790.  These estimates are generally found between

988: the previous estimates and our new values.  Because of low Galactic

989: latitudes of the clusters ($|b| < 5\arcdeg$), the dust map of

990: \citet{schlegel98} could not provide meaningful upper limits on

991: the color excess.

992: 

993: We contend that our reddening estimates are more reliable than the previous

994: values, which mostly relied on the solar metallicity ZAMS-fitting on

995: $UBV$ color-color diagrams.  In contrast to these studies, we took into

996: account metallicity and age effects on MS fitting, and our estimates are

997: based on simultaneous fits over a wider range of spectral bands,

998: including near-infrared $JHK_s$.  We also note that photometric zero-point

999: errors in $U$-band \citep[e.g.,][]{bessell05} could introduce significant

1000: uncertainties in color excesses derived from $UBV$ color-color diagrams.

1001: 

1002: \section{Galactic Period-Luminosity Relations}\label{sec:pl}

1003: 

1004: With cluster parameters derived in the previous section, we are now

1005: in a position to estimate absolute magnitudes of Cepheids and to

1006: redetermine the Galactic {\it P-L} relations.  We show that our new

1007: {\it P-L} relation in $V$ is generally fainter than in previous studies.

1008: However, it is in good agreement with recent parallax studies

1009: \citep{benedict07,vanleeuwen07} when the Wesenheit magnitudes are employed.

1010: 

1011: \subsection{Absolute Magnitudes of Cepheids}\label{sec:mv}

1012: 

1013: \input{tab12.tex}

1014: 

1015: The absolute $V$ magnitude was computed for each Cepheid from

1016: \begin{eqnarray}

1017: \langle M_V \rangle

1018: = \langle m_V \rangle - (m - M)_0 - R_V E(B - V),\label{eq:mv}

1019: \label{eq:mv}

1020: \end{eqnarray}

1021: where $\langle m_V \rangle$ is the observed intensity-mean magnitude and

1022: $(m - M)_0$ is the extinction-corrected distance modulus.  We used the

1023: same color-dependent prescriptions for $R_V$ and $E(B - V)$ as those for

1024: MS dwarfs (eqs.~[\ref{eq:rv}] and [\ref{eq:ebv}]).  The absolute

1025: magnitudes in other filter passbands were computed from the CCM89

1026: reddening law at each cluster's $R_{V,0}$.  Table~\ref{tab:cepheid}

1027: lists absolute magnitudes in $BVI_CJHK_s$ for all Cepheids in this

1028: study.  In the last column we also show the reddening-insensitive

1029: Wesenheit magnitude, $\langle W_{VI} \rangle \equiv \langle V \rangle -

1030: 2.45 (\langle V \rangle - \langle I_C \rangle)$, which was adopted in the

1031: {\it HST} Key Project.\footnote{We simply adopted $A_V / E(V - I_C) =

1032: 2.45$ from \citet{freedman01} to directly compare our results with those

1033: of other studies, which also employed the same relation.   We would

1034: derive $A_V / E(V - I_C) = 2.56$ for the Cepheids with $\langle B_0

1035: \rangle - \langle V_0 \rangle = 0.7$ at $R_V = 3.35$, computed

1036: from the average $R_{V,0}$ of our cluster sample.  This difference in

1037: $A_V / E(V - I_C)$ has a negligible effect on $W_{VI}$ of Cepheids.

1038: It is noted that \citet{macri01b} tested the CCM89 reddening law using

1039: $VIH$ photometry for 70 extragalactic Cepheids in 13 galaxies.  They

1040: showed that a mean color-excess ratio of $E(V - H) / E(V - I)$ is in

1041: good agreement with the CCM89 predicted value.}

1042: 

1043: \input{tab13.tex}

1044: 

1045: Table~\ref{tab:error2} shows systematic errors in $\langle M_V \rangle$

1046: of the Cepheids.  As in Table~\ref{tab:error}, we estimated

1047: the size of errors by performing MS fits for alternate values of

1048: systematic error quantities.  These are shown in the first and

1049: second columns, which are the same as those in Table~\ref{tab:error}

1050: except the last two rows:  $\Delta m_V$ and $\sigma_{E(B-V)}$.

1051: The $\Delta m_V$ represents a systematic error in the Cepheid photometry,

1052: which was treated as independent of the cluster photometry and any

1053: of the MS-fitting process.  Note that $(m - M)_0$, $R_V$, and

1054: $E(B - V)$ in equation~(\ref{eq:mv}) are correlated with each other

1055: via MS fitting.

1056: 

1057: The $\sigma_{E(B-V)}$ in Table~\ref{tab:error2} represents the error

1058: in the reddening for individual Cepheids.  In equation~(\ref{eq:mv})

1059: we adopted the Cepheid reddening as the mean cluster reddening value.

1060: However, reddening for individual stars becomes more uncertain as

1061: we deal with more differentially reddened clusters.  For these

1062: clusters, the colors for individual stars scatter along a reddening

1063: vector, and this can be detected by a large size of the MS width above

1064: photometric precision.

1065: 

1066: \input{tab14.tex}

1067: 

1068: To estimate the size of differential reddening for each cluster, we

1069: first computed the standard deviation of $B - V$ colors in the upper MS

1070: from Table~\ref{tab:ebv}.  Since the photometric filtering likely

1071: reduces the size of the dispersion, we corrected for it from artificial

1072: cluster CMD tests (Paper~III).  We generated solar metallicity, 80~Myr

1073: isochrones with photometric errors of $0.02$~mag in colors and magnitudes.

1074: We used the \citet{salpeter55} mass function for the primaries and

1075: a flat mass function for the secondaries for a 40\% binary

1076: fraction.\footnote{The binary fraction is defined as the number of binaries

1077: divided by the total number of systems in the considered fitting range.}

1078: Single stars and binaries were then randomly displaced from the MS

1079: assuming a normal distribution of individual reddening with a standard deviation

1080: $\sigma_{E(B-V)}$.  After applying the photometric filtering, we then estimated

1081: the size of the reduction in dispersion as a function of $\sigma_{E(B-V)}$.

1082: In this way, we inferred $\sigma_{E(B-V)}$ for each cluster

1083: from the observed dispersion in $B - V$, which are shown in the second

1084: column of Table~\ref{tab:art}.  We made no corrections to those for NGC~5662

1085: and NGC~7790 because the dispersions in $B - V$ were equal to or smaller

1086: than our assumed photometric precision ($\sigma_{B-V} = 0.02$).

1087: 

1088: In addition to previously known clusters with differential reddening

1089: (NGC~129, NGC~5662, and M25), we also found that the remaining clusters

1090: except NGC~7790 have larger dispersions than photometric errors.  While

1091: this could be due to remaining binaries, background stars, rapid

1092: rotators, or an underestimation of the photometric errors, we took

1093: the excess dispersion in Table~\ref{tab:art} as a characteristic

1094: measure of differential reddening as a conservative error estimate.

1095: The $\sigma_{E(B-V)}$ in Table~\ref{tab:art} was then multiplied by

1096: $R_V$ to estimate the error in $\langle M_V \rangle$, which is shown in

1097: Table~\ref{tab:error2}.  For all clusters except NGC~7790, differential

1098: reddening is the largest error source in $\langle M_V \rangle$.

1099: 

1100: In previous studies, local Cepheid reddening was often determined

1101: using photometry of neighboring stars.  We experimented with this

1102: approach for M25.  We estimated individual Cepheid's reddening

1103: on $(B - V, V)$ and assumed that the dispersion around the best-fitting

1104: isochrone is solely due to differential reddening.  However, we could

1105: not place a strong constraint on the local color excess for U~Sgr

1106: because there were not a sufficient number of stars with good $E(B - V)$.

1107: In addition, cluster binaries cannot be easily distinguished from

1108: highly reddened single stars on CMDs, which could lead to an overestimation

1109: of reddening.

1110: 

1111: \subsection{The Galactic {\it P-L} Relations}\label{sec:galpl}

1112: 

1113: Our Cepheid sample does not span a wide range of periods, which leads to

1114: a large error in {\it P-L} slopes.  Instead, we adopted the {\it P-L} slopes

1115: derived from LMC Cepheids, assuming that the {\it P-L} relations for Cepheids

1116: in the Galaxy and in the LMC have the same slopes.  Some investigators have

1117: claimed that the Galactic {\it P-L} relations are steeper than those of

1118: the LMC Cepheids, and that the LMC {\it P-L} relations have a break at

1119: $P\approx10$~days, possibly due to different metal contents of Cepheids

1120: \citep{kanbur04,sandage04,ngeow05}.  However, \citet{macri06} recently

1121: showed that {\it P-L} relations from the LMC Cepheids in the OGLE-II

1122: catalog \citep{udalski99a} are a good fit to Cepheids in the two

1123: fields of NGC~4258, which have approximately the LMC and the Galactic

1124: metal abundances (see \S~\ref{sec:4258}).

1125: 

1126: We derived LMC {\it P-L} slopes using fundamental mode Cepheids in the

1127: OGLE-II catalog \citep{udalski99a} and those in \citet{persson04}.

1128: The OGLE-II Cepheids have an average period of $\langle \log{P} {\rm (days)}

1129: \rangle \approx 0.6$ with about half of them being in the same

1130: period range as our Galactic Cepheid sample.  The {\it P-L} relations in

1131: the LMC from the OGLE-II database were originally derived by \citet{udalski99b}

1132: and then revised according to the new photometric calibration \citep{udalski00}.

1133: We rederived the period-{\it apparent} magnitude relations in $BVI_C$

1134: using the revised data from the OGLE-II Internet

1135: archive\footnote{http://www.astrouw.edu.pl/\~{}ogle/ogle2/cep\_lmc.html}

1136: after iterative $2.5\sigma$ rejection from 680 Cepheids with $\log{P} >

1137: 0.4$ \citep{udalski99b}.  We also derived similar relations in $JHK_s$

1138: on the LCO system from the photometry in \citet{persson04}.  As in their

1139: analysis, we excluded four stars (HV~883, HV~2447, and HV~2883 have

1140: periods longer than 100 days, and HV~12765 is about 0.2~mag brighter

1141: than the mean {\it P-L} relation), leaving 88 Cepheids with $\log{P} > 0.4$.

1142: 

1143: We then derived the Galactic {\it P-L} relations in $BVI_CJHK_s$ after

1144: correcting for interstellar extinction:

1145: \begin{eqnarray}

1146: \langle M_B     \rangle = (-3.169\pm0.085) - (2.503\pm0.048) (\log{P} - 1),\nonumber\\

1147: \langle M_V     \rangle = (-3.932\pm0.066) - (2.819\pm0.032) (\log{P} - 1),\nonumber\\

1148: \langle M_{I_C} \rangle = (-4.712\pm0.057) - (3.004\pm0.021) (\log{P} - 1),\nonumber\\

1149: \langle M_J     \rangle = (-5.271\pm0.076) - (3.148\pm0.053) (\log{P} - 1),\nonumber\\

1150: \langle M_H     \rangle = (-5.593\pm0.069) - (3.233\pm0.044) (\log{P} - 1),\nonumber\\

1151: \langle M_{K_s} \rangle = (-5.718\pm0.064) - (3.282\pm0.040) (\log{P} - 1).\nonumber

1152: \label{eq:galpl}

1153: \end{eqnarray}

1154: The slope errors are those estimated from the LMC Cepheids.  The zero-point

1155: errors were estimated from the magnitude dispersion among the Cepheids,

1156: assuming that the magnitude errors are uncorrelated with each

1157: other.  We took the average magnitudes and periods of the three Cepheids

1158: in NGC~7790 as one data point.  Cepheids in the two northern clusters

1159: (NGC~129 and NGC~7790) have no photometry in \citet{laney92}, so the

1160: {\it P-L} relations in $JHK_s$ were derived from five clusters.

1161: 

1162: \begin{figure}

1163: \epsscale{1.1}

1164: \plotone{f13.eps}

1165: \caption{Galactic {\it P-L} relation in $V$.  The bull's-eyes are our

1166: Galactic calibrators with their $1\sigma$ errors, and the solid line is

1167: a fit to these points assuming the LMC {\it P-L} slope.  The small boxes

1168: are LMC fundamental mode Cepheids from the OGLE-II survey, assuming

1169: $(m - M)_0 = 18.50$ and the OGLE-II reddening map.  The dashed line is

1170: a fit to these points.\label{fig:pl}}

1171: \end{figure}

1172: 

1173: \begin{figure*}

1174: \epsscale{0.8}

1175: \plotone{f14.eps}

1176: \caption{Same as Fig.~\ref{fig:pl}, but in all $BVI_CJHK_s$.

1177: \label{fig:pl2}}

1178: \end{figure*}

1179: 

1180: Figures~\ref{fig:pl} and \ref{fig:pl2} display {\it P-L} relations for

1181: our Galactic Cepheids ({\it bull's-eyes}) in all filter passbands.  In

1182: $V$-band the $\chi^2$ of the fit is $8.0$ for 6 degrees of freedom,

1183: which indicates our reasonable error estimation.  The standard deviation

1184: of the Galactic Cepheids around the {\it P-L} relation is $0.18$~mag, which

1185: is comparable to that of the LMC counterparts \citep{udalski99b}.  This

1186: value is smaller than previous estimates for the Galactic Cepheids

1187: from MS fitting: \citet{tammann03} found 0.26~mag for 25 Cepheids

1188: \citep[see also][]{sandage04}, and \citet{fouque03} estimated 0.27~mag

1189: from 24 Cepheids.

1190: 

1191: In Figure~\ref{fig:pl} the OGLE-II Cepheids are also shown at

1192: a distance modulus of 18.50 mag, which was adopted in the {\it HST} Key

1193: Project ({\it small boxes}).  Magnitudes of these Cepheids were corrected

1194: for extinction based on the OGLE-II reddening map, which was derived

1195: from the $I_C$-band magnitudes of red clump stars \citep{udalski99a}.

1196: As seen in the figure, the LMC {\it P-L} relation ({\it dashed line}) is

1197: brighter than our Galactic {\it P-L} relation ({\it solid line}) by

1198: $\sim0.3$~mag.  This is due to metallicity and reddening effects, as well

1199: as the LMC distance as discussed in \S~\ref{sec:lmc}.

1200: 

1201: \subsection{Comparison with Previous Studies}\label{sec:plcomp}

1202: 

1203: \input{tab15.tex}

1204: 

1205: \begin{figure}

1206: \epsscale{1.1}

1207: \plotone{f15.eps}

1208: \caption{Comparison of Cepheid absolute magnitudes with those in

1209: \citet{feast87}, \citet{tammann03}, \citet{hoyle03}, and \citet{gieren05}.

1210: The differences are in the sense of their magnitudes minus those in this

1211: paper.  The error bars represent our error estimates only.  Individual

1212: $\langle M_V \rangle$ estimates in the top two panels were revised

1213: assuming our Pleiades distance scale.\label{fig:comp.mv}}

1214: \end{figure}

1215: 

1216: Figure~\ref{fig:comp.mv} compares our $\langle M_V \rangle$ with those of

1217: \citet{feast87}, \citet{tammann03}, \citet{hoyle03}, and \citet{gieren05}.

1218: Error bars represent our error

1219: estimates only.  We revised absolute magnitudes in \citeauthor{feast87}

1220: and \citeauthor{tammann03} assuming our Pleiades distance ($5.63$~mag in

1221: distance modulus).  Individual $\langle M_V \rangle$ estimates are shown

1222: in Table~\ref{tab:comp.mv}.  We restricted our comparison to $\langle M_V

1223: \rangle$ since absolute magnitudes in other passbands are correlated with

1224: each other.  The weighted average differences are $\Delta \langle M_V

1225: \rangle = -0.17\pm0.07$, $-0.09\pm0.09$, $-0.14\pm0.10$, and $-0.22\pm0.11$

1226: from \citeauthor{feast87}, \citeauthor{tammann03}, \citeauthor{hoyle03},

1227: and \citeauthor{gieren05}, respectively, in the sense that our estimates

1228: are fainter on average.  This is mainly due to the fact that our distance

1229: and reddening estimates are smaller on average than those in the previous

1230: studies (\S~\ref{sec:comp}).

1231: 

1232: \begin{figure*}

1233: \epsscale{1.0}

1234: \plottwo{f16a.eps}{f16b.eps}

1235: \caption{Comparison of the $V$-band Galactic {\it P-L} relation with

1236: those in \citet[][{\it left}]{tammann03} and \citet[][{\it right}]{benedict07}.

1237: The open circles are Cepheids in these studies,

1238: and the solid lines are their best-fit {\it P-L} relations.  Our

1239: Galactic calibrators are shown as filled circles, and the dotted lines

1240: represent our {\it P-L} relation.  Magnitudes in \citeauthor{tammann03}

1241: were revised assuming the same Pleiades distance scale as in this paper.

1242: \label{fig:plcomp}}

1243: \end{figure*}

1244: 

1245: In addition, we also made a comparison with all cluster Cepheids in

1246: \citet{tammann03} as shown in the left panel of Figure~\ref{fig:plcomp}.

1247: As found in the above Cepheid-to-Cepheid comparison, the difference with

1248: their $\langle M_V \rangle$ is statistically insignificant for Cepheids

1249: with $\log{P} \la 1.1$.  However, their {\it P-L} relation is steeper

1250: than ours because their long-period Cepheids at $\log{P} \ga 1.1$ are

1251: brighter than expected from our {\it P-L} relation.  If we assume their

1252: {\it P-L} slope for our sample, we would derive $\langle M_V \rangle =

1253: -4.01$ for 10-day period Cepheids, which is $0.25$~mag fainter than

1254: their {\it P-L} zero point.

1255: 

1256: In the right panel of Figure~\ref{fig:plcomp}, we compare our Galactic

1257: {\it P-L} relation in $V$ with that of the recent {\it HST} parallax study

1258: \citep{benedict07}, which provided accurate parallaxes (average $8\%$)

1259: for nine Galactic Cepheids using the {\it HST} Fine Guidance Sensor

1260: \citep[see also][]{benedict02b}.  Their Cepheid parallaxes were estimated

1261: with respect to photometric parallaxes of reference stars, which set the

1262: absolute distance scale of each reference frame.  The magnitudes were

1263: then corrected for interstellar extinction from either color-color relations

1264: or the DDO Galactic Cepheid database.  We have no Cepheids in common with

1265: their study since their distance measurements were focused on nearby

1266: field Cepheids.  If we assume their {\it P-L} slope, our 10-day period

1267: Cepheids would be $0.20$~mag fainter than their zero point.  If the

1268: difference is solely due to the error in our reddening estimates,

1269: our estimates would have been underestimated by $\Delta \langle E(B - V)

1270: \rangle \sim 0.04$.  However, it is noted that reddening can be

1271: determined far more accurately for cluster Cepheids than it can for

1272: field stars.  Our error bars in Figure~\ref{fig:plcomp} are slightly

1273: larger than those of the {\it HST} study because of our conservative

1274: error estimates for the extinction correction.

1275: 

1276: \begin{figure}

1277: \epsscale{1.1}

1278: \plotone{f17.eps}

1279: \caption{Same as Fig.~\ref{fig:plcomp}, but a comparison of the {\it

1280: P-L} relation in $M_{W(VI)}$ with that of \citet[][{\it solid line}]

1281: {benedict07}.  Residuals are shown with respect to their {\it P-L}

1282: relation in the bottom panel.\label{fig:plmw}}

1283: \end{figure}

1284: 

1285: On the other hand, the {\it HST}-based {\it P-L} relation is in better

1286: agreement with ours when the Wesenheit magnitude $W(VI)$ is used as

1287: shown in Figure~\ref{fig:plmw}.  The weighted mean difference of our

1288: Galactic Cepheids from their best-fitting {\it P-L} relation ({\it solid

1289: line}) is $\Delta \langle M_{W(VI)} \rangle = 0.05\pm0.06$ in the sense

1290: that our {\it P-L} relation is fainter.  Since the Wesenheit magnitude

1291: was designed to minimize the effect of errors in the reddening, the

1292: agreement may indicate that the large difference in $\langle M_V \rangle$

1293: is due to the difference in the distance estimation.

1294: 

1295: Previous studies based on {\it Hipparcos} data typically yielded

1296: $\langle M_V \rangle \approx -4.2$ for 10-day period Cepheids

1297: \citep{feast97,feast98,lanoix99,groenewegen00}, while \citet{luri98}

1298: obtained $\langle M_V \rangle \approx -3.9$ from statistical parallaxes.

1299: Recently, \citet{vanleeuwen07} reexamined the Galactic {\it P-L} relations

1300: based on the new reduction of the {\it Hipparcos} astrometric data

1301: \citep{vanleeuwen05b}, where improvements of up to a factor of 2 in

1302: parallax accuracy have been achieved by reconstructing the satellite's

1303: attitude \citep{vanleeuwen05a}.  They analyzed their data together with

1304: the {\it HST} parallaxes to derive a {\it P-L} relation in $W(VI)$.

1305: If we adopt their {\it P-L} slope in $W(VI)$, our {\it P-L} relation

1306: would be $\Delta \langle M_{W(VI)} \rangle = 0.07\pm0.06$ fainter than

1307: their best-fitting {\it P-L} relation.

1308: 

1309: \subsection{Systematic Errors in {\it P-L} relations}\label{sec:plsys}

1310: 

1311: \input{tab16.tex}

1312: 

1313: When determining the {\it P-L} relations in the previous section, we

1314: assumed that the errors in $\langle M_V \rangle$ of individual Cepheids

1315: (Table~\ref{tab:error2}) are uncorrelated with each other.  However,

1316: there are a number of error sources that could systematically change

1317: $\langle M_V \rangle$.  These are shown in Table~\ref{tab:plerr} with

1318: their estimated size of error contributions to the {\it P-L} zero point,

1319: which is described below in detail.  In short, the calibration error is

1320: from the uncertainty in the adopted distance of the Pleiades (Paper~III).

1321: Errors in the metallicity scale of $\pm0.1$~dex and the age scale of $\Delta

1322: \log{t {\rm (Myr)}} = 0.2$ were adopted.  The error from the reddening

1323: laws represents the case of using different color-dependent prescriptions

1324: for $E(B - V)$ and $R_V$ from \citet{feast87} and \citet{laney93}.

1325: We performed artificial cluster CMD tests to estimate a probable size

1326: of a bias in the MS-fitting technique from differential reddening and

1327: unresolved cluster binaries.

1328: 

1329: From the comparison with \citet{yong06} and \citet{andrievsky02},

1330: we inferred that our adopted metallicity scale \citep{fry97} has

1331: a $0.1$~dex systematic error (\S~\ref{sec:mehage}).  Metallicity changes

1332: of $\pm0.1$~dex would vary the $\langle M_V \rangle$ of 10-day period

1333: Cepheids from $-3.84$ to $-4.04$~mag.

1334: 

1335: In this paper, only the relative age scale is relevant with respect

1336: to the Pleiades.  Our assumed age range in \S~\ref{sec:mehage} reflects

1337: the cluster-to-cluster variations with respect to the adopted

1338: age of $80$~Myr.  However, it is possible that the Cepheid clusters

1339: are all systematically younger or older than the Pleiades.  This would

1340: change the reddening values since the mean color of the upper MS

1341: varies with the cluster age.  In fact, some of the CMDs have stars

1342: that extend far above the top end of the $80$~Myr isochrone.  If

1343: these bright stars are cluster members, they would suggest younger

1344: ages than $80$~Myr.  Nevertheless, the {\it P-L} zero point is relatively

1345: insensitive to the age.  If the overall age scale is uncertain at

1346: the $1\sigma$ level of individual cluster ages ($50$--$120$~Myr), we

1347: would have $\langle M_V \rangle = -3.90$ and $-3.96$~mag for 10-day

1348: period Cepheids at $50$ and $120$~Myr, respectively.

1349: 

1350: We estimated the error from the reddening laws using different

1351: prescriptions for the color-dependent $E(B - V)$ and $R_V$ in

1352: \citet{feast87} and \citet{laney93}.  For the same $E(B - V)_0$ and

1353: $R_{V,0}$ values derived from the cluster MS fitting, we found that

1354: these prescriptions result in a fainter {\it P-L} relation by $\Delta

1355: \langle M_V \rangle \approx 0.05$ than our default case

1356: (eqs.~[\ref{eq:rv}] and [\ref{eq:ebv}]).

1357: 

1358: The average $R_{V,0}$ that we inferred from the clusters in this study is

1359: $3.19$ (Table~\ref{tab:cluster}), which is in good agreement with the

1360: average Galactic value for the diffuse interstellar medium, $R_V \sim

1361: 3.1$.  The formal standard deviation of $R_{V,0}$ is $0.17$, but

1362: individual errors are too large to detect a cluster-to-cluster

1363: variation.  Even if we adopt a different $R_{V,0}$, it would have

1364: a reduced impact on $\langle M_V \rangle$.  This is because a larger

1365: $R_V$ makes an MS-fitting distance shorter (Paper~III), but it makes the

1366: extinction correction larger at the same time (eq.~[\ref{eq:mv}]).

1367: 

1368: In \S~\ref{sec:mv} we considered the case where the presence of

1369: differential reddening makes it difficult to estimate individual

1370: extinctions for Cepheids.  This is independent of the MS-fitting

1371: technique in the sense that it does not affect any of the cluster

1372: parameters.  However, differential reddening could also modify

1373: cluster parameters because of cluster binaries.  Unresolved binaries

1374: are typically brighter and redder than a single-star MS, which would make

1375: a distance apparently shorter.  If there exists differential reddening

1376: across the cluster field, the distribution of these binaries on CMDs

1377: would be modified, and the MS-fitting parameters would subsequently change.

1378: 

1379: \begin{figure}

1380: \epsscale{1.1}

1381: \plotone{f18.eps}

1382: \caption{Effects of differential reddening from artificial cluster CMD

1383: tests (see text for details).  A bias in each quantity is shown assuming

1384: 40\% binary fraction.  The thick solid line is the median of these values

1385: from 200 artificial cluster CMDs as a function of differential reddening,

1386: and thin lines on either side are the first and third quartiles.  The dashed

1387: line indicates a zero bias.  The differences are in the sense that Cepheids

1388: look brighter at $\sigma_{E(B-V)} \approx 0.05$ with an overestimation of

1389: $E(B - V)$.\label{fig:diff}}

1390: \end{figure}

1391: 

1392: To estimate the effect of differential reddening, we performed artificial

1393: cluster CMD tests as described earlier in \S~\ref{sec:mv}.  We assumed

1394: that $E(B - V)$ has a normal distribution with a standard deviation

1395: $\sigma_{E(B-V)}$.  For each set of artificial cluster CMDs we applied

1396: the photometric filtering and estimated $R_{V,0}$, $E(B - V)_0$, $(m - M)_0$,

1397: and $\langle M_V \rangle$ as for the actual cluster data, assuming $\langle

1398: B_0 \rangle - \langle V_0 \rangle = 0.7$ for Cepheids. Each panel in

1399: Figure~\ref{fig:diff} shows the bias in each of these quantities as

1400: a function of $\sigma_{E(B-V)}$.  The thick solid line shows the median

1401: of the bias from 200 artificial cluster CMDs, and each set of CMDs contained

1402: 60 single stars and 40 unresolved binaries in our MS-fitting range for

1403: the distance determination.  Thin lines on either side are the first and

1404: third quartiles of these estimates.  The dashed line indicates a zero bias.

1405: 

1406: Even in the presence of a small differential reddening, low mass-ratio

1407: binaries cannot be easily detected from the photometric filtering.

1408: Because these unresolved binaries are redder than a single-star MS at

1409: a given $M_V$, the reddening would be overestimated as shown in

1410: Figure~\ref{fig:diff} for $\sigma_{E(B-V)} \la 0.05$.  A higher $E(B - V)$

1411: then results in a longer distance, and a Cepheid would look brighter

1412: by several hundredths of a magnitude in $\langle M_V \rangle$.  As

1413: differential reddening becomes stronger, binaries would be more scattered

1414: around the MS, and it would become further difficult to distinguish

1415: them from highly reddened single stars.  As a result, both cluster

1416: reddening and distance estimates decrease at $\sigma_{E(B-V)} \ga 0.05$,

1417: while semi-interquartile ranges continuously increase.  We also

1418: experimented with 30\% and 50\% binary fractions but found insignificant

1419: changes within the semi-interquartile ranges.

1420: 

1421: The last column in Table~\ref{tab:art} shows the size of the bias in

1422: $\langle M_V \rangle$ for each Cepheid, which was determined at

1423: $\sigma_{E(B-V)}$ in the second column.  Applying these corrections to

1424: $\langle M_V \rangle$ in Table~\ref{tab:cepheid} leads to $0.05$~mag

1425: reduction in $\langle M_V \rangle$ for 10-day period Cepheids, again

1426: making worse the agreement with the brighter calibration in the previous

1427: studies.  The total $\chi^2$ of the fit becomes smaller ($6.9$ for 6

1428: degrees of freedom) if we add in quadrature the semi-interquartile range

1429: to the error in $\langle M_V \rangle$.

1430: 

1431: In summary, our Galactic {\it P-L} relation in $V$-band has a statistical

1432: error of 0.07~mag and a systematic error of 0.14~mag in the zero point

1433: (Table~\ref{tab:plerr}).  If we add these quantities in quadrature, the

1434: combined error would be 0.16~mag.  Since we determined the Galactic

1435: {\it P-L} relations in other filter passbands with the same cluster

1436: distance, reddening, and $R_V$ estimates as in $V$, their zero-point

1437: errors are correlated with each other.  Therefore, we took $0.16$~mag

1438: as a systematic error in the Cepheid distance scale in the following section.

1439: 

1440: \section{Extragalactic Distance Scales}

1441: 

1442: Cepheids have long served as our most commonly used standard candles

1443: for extragalactic distance studies.  The conventional way of estimating

1444: extragalactic Cepheid distances is to use a fiducial {\it P-L} relation

1445: derived from the LMC Cepheids with an adopted distance to the LMC.

1446: However, geometric distances to a few galaxies have become available

1447: in recent years, providing an opportunity to compare with the Cepheid

1448: distance scale.  In this section we therefore apply our Galactic

1449: {\it P-L} relations to estimate a distance to NGC~4258 (M106) that has

1450: an accurate geometric distance measurement from its water maser sources

1451: \citep{herrnstein99}.  We then infer a distance to the LMC and compare

1452: our result with those adopted by the {\it HST} Key Project and the

1453: SN~Ia calibration program \citep{sandage06b}, from which we consider

1454: a possible increase in their Hubble constant $H_0$ measurements.  We

1455: finish with a discussion of the potentially interesting case of M33

1456: (NGC~598).  Our most fundamental conclusion is that the distance to

1457: NGC~4258 is in good agreement with the geometric measurement when

1458: using the {\it P-L} relations inferred from open clusters.

1459: We also compare our LMC distance with those from the {\it HST}

1460: \citep{benedict07} and the {\it Hipparcos} parallax studies

1461: \citep{vanleeuwen07}.

1462: 

1463: \subsection{The Maser-host Galaxy NGC~4258}\label{sec:4258}

1464: 

1465: \citet{herrnstein99} inferred a geometric distance to NGC~4258 from

1466: the orbital motions of water maser sources on its nucleus

1467: and found $(m - M)_0 = 29.29\pm0.09{\rm\ (statistical)}\pm0.12

1468: {\rm\ (systematic)}$ \citep[see also][]{argon07}.  Many Cepheids

1469: were also observed in this galaxy, providing an opportunity to check

1470: the Cepheid distance scale \citep{maoz99,newman01}.  Recently,

1471: \citet{macri06} discovered 281 Cepheids and provided accurate $BVI$

1472: photometry using the Advanced Camera for Surveys/Wide Field Camera

1473: (ACS/WFC) onboard the {\it HST}.  In particular, they observed two

1474: fields in the galaxy, located at two different galactocentric distances

1475: with significantly different gas-phase metal abundances, and derived

1476: the metallicity sensitivity of the Cepheid luminosity.

1477: 

1478: We reexamined the \citeauthor{macri06} data set in light of our revised

1479: {\it P-L} relationship.  Following their selection procedure, we used

1480: the ``restricted'' sample of 69 Cepheids and estimated distance moduli

1481: for the two groups of Cepheids in the inner (with a period cut of

1482: $P > 12$~days) and outer fields ($P > 6$~days).  For each group of this

1483: sample we corrected for extinction using our {\it P-L} relations in

1484: $BVI_C$.  The absolute distance to NGC~4258 was then estimated by

1485: anchoring our Galactic sample to a reference gas-phase metal abundance

1486: within the galaxy.

1487: 

1488: \input{tab17.tex}

1489: 

1490: \begin{figure*}

1491: \epsscale{0.75}

1492: \plotone{f19.eps}

1493: \caption{{\it Top left}: Apparent distance moduli of inner field Cepheids

1494: in NGC~4258 with the best-fitting CCM89 extinction curve ({\it solid line}).

1495: {\it Top right}: Likelihood contours in the average reddening and distance

1496: modulus of inner field Cepheids shown at $\Delta \chi^2$ = 2.30, 6.17, and

1497: 11.8 (68.3\%, 95.4\%, and 99.73\% confidence levels for 2 degrees of

1498: freedom).  The horizontal lines represent a geometric distance from the

1499: water maser sources and its $1\sigma$ error \citep{herrnstein99}.  {\it

1500: Bottom left}: Same as the top left panel, but for outer field Cepheids.

1501: {\it Bottom right}: Same as the top right panel, but for outer field

1502: Cepheids.\label{fig:4258}}

1503: \end{figure*}

1504: 

1505: Figure~\ref{fig:4258} shows how we determined the average distance

1506: and reddening.  In the left two panels {\it apparent} distance moduli in

1507: $BVI_C$ are shown for Cepheids in the inner and outer fields ({\it top} and

1508: {\it bottom}, respectively).  Distances from shorter wavelength filters are

1509: systematically longer, and the data are well fitted by a CCM89 extinction

1510: curve ({\it solid line}).  Here we assumed $R_V = 3.35$ for the Cepheids,

1511: which was computed from the average $R_{V,0}$ of our cluster sample with

1512: equation~(\ref{eq:rv}) at $\langle B \rangle_0 - \langle V \rangle_0

1513: = 0.72$.  At this value, the CCM89 extinction law yields absorption ratios

1514: $A_{\lambda}/A_V$ of 1.30 and 0.61 for $B$ and $I_C$, respectively,

1515: where we took effective wavelengths of these filters from BCP98.

1516: We added 0.02~mag in quadrature to the fitting error to take into

1517: account the photometric zero-point error.  At the limit of $\lambda^{-1}

1518: \rightarrow 0$, an intercept on the ordinate is the true or unreddened

1519: distance modulus \citep[e.g.,][]{madore91,madore98}.  The right two panels

1520: in Figure~\ref{fig:4258} show likelihood contours in the average reddening

1521: and distance modulus from these fits.  Contours are shown at $\Delta

1522: \chi^2$ = 2.30, 6.17, and 11.8 (68.3\%, 95.4\%, and 99.73\% confidence

1523: levels) for 2 degrees of freedom.  Horizontal lines represent the

1524: distance estimate from the water maser sources and its combined $1\sigma$

1525: error \citep{herrnstein99}.  Distance moduli at the best-fit reddening

1526: are listed in Table~\ref{tab:4258} for each of the two galactic fields.

1527: 

1528: When the reddening is constrained in this way, distances in both fields

1529: are in good agreement with the maser distance.  However, the distance

1530: from the inner field is shorter than the distance from the outer field

1531: by $\Delta (m - M)_0 = 0.07$.  The most natural interpretation of this

1532: difference is that it measures the metallicity dependence of

1533: the Cepheid luminosity.  This effect has been extensively discussed

1534: in the literature and tested empirically \citep{freedman90,gould94,

1535: sasselov97,kochanek97,kennicutt98,udalski00,kennicutt03,groenewegen04,

1536: sakai04,storm04,romaniello05,macri06}.

1537: 

1538: To estimate the true distance modulus of the galaxy, we first estimated

1539: metal abundances in the two NGC~4258 fields and then inferred a distance

1540: at the metallicity of the Galactic Cepheids.  Stellar abundances cannot

1541: be measured directly for most of the Cepheid galaxies.  Instead, the

1542: {\it HST} Key Project adopted ``empirical'' gas-phase [O/H] abundances

1543: from \citet{zaritsky94}.  \citeauthor{macri06} computed oxygen

1544: abundances at each Cepheid's location from the \ion{H}{2} region

1545: abundance gradient in \citeauthor{zaritsky94}.  The average abundances

1546: from this table are $\langle 12 + \log{\rm (O/H)} \rangle = 8.94$ and

1547: $8.54$~dex for the inner and outer field Cepheids, respectively, when

1548: the LMC Cepheids have $12 + \log{\rm (O/H)} = 8.50$ on this scale

1549: \citep{sakai04}.  To relate these abundances to those of

1550: our Galactic Cepheid sample, we adopted a stellar abundance difference

1551: between the LMC Cepheids and Galactic supergiants from \citet{luck98} of

1552: $\Delta{\rm \langle [Fe/H] \rangle (Galaxy - LMC)} \approx 0.30$.  They

1553: also found a similar $\alpha$-element enhancement between the LMC and

1554: the Galactic samples.  In the following discussion we adopted $\Delta

1555: \langle \log{\rm (O/H)} \rangle {\rm (Galaxy - LMC)} = 0.30\pm0.06$,

1556: assuming that the stellar abundance difference is the same as the

1557: difference in the gas-phase oxygen abundance.

1558: 

1559: Since the metallicity of the Galactic Cepheids is found between the metal

1560: abundances for the inner and outer fields, the true distance would lie

1561: between distance estimates from the two galactic fields.  From a linear interpolation,

1562: we obtained $(m - M)_0 = 29.28\pm0.10\pm0.16$ ({\it P-L} zero point) for

1563: a distance to NGC~4258, which is in good agreement with the current maser

1564: distance.  The first error is from the fit to the {\it HST} ACS data

1565: ($0.10$~mag), an error in $R_V$ ($\mp0.03$~mag for $\pm0.5$ change in

1566: $R_V$), and an error in the metallicity scale ($\mp0.02$~mag for

1567: $\pm0.10$~dex).  We adopted an error in the metallicity scale of

1568: $\pm0.10$~dex to account for a possible difference in the stellar and

1569: gas-phase abundances, but it has only a moderate effect on the distance

1570: determination.

1571: 

1572: Previous studies found somewhat longer Cepheid distances to NGC~4258

1573: although these values are consistent within the errors with the maser

1574: solution.  As a part of the {\it HST} Key Project, \citet{maoz99}

1575: estimated $(m - M)_0 = 29.54\pm0.12\pm0.18$ using the Wide-Field

1576: Planetary Camera 2 (WFPC2), and \citet{newman01} revised it

1577: to be $(m - M)_0 = 29.47\pm0.09\pm0.15$, where the first errors represent

1578: those unique to each determination and the second errors are systematic

1579: uncertainties in the {\it HST} Key Project distances.

1580: 

1581: However, the reddening in the {\it HST} Key Project is significantly larger

1582: than our estimates.  The final result in the {\it HST} Key Project lists

1583: $\langle E(V - I)_C \rangle = 0.22\pm0.04$ or $\langle E(B - V) \rangle

1584: = 0.17\pm0.03$ for Cepheids

1585: located in the middle of galactocentric distances to the two ACS fields.

1586: On the other hand, we found $\langle E(B - V) \rangle = 0.08\pm0.04$

1587: and $0.06\pm0.05$ for the inner- and outer-field Cepheids, respectively,

1588: where the errors are the $1\sigma$ contour sizes in Figure~\ref{fig:4258}.

1589: \citeauthor{macri06} also estimated individual Cepheid reddenings using

1590: the OGLE-II LMC {\it P-L} relations, and their average values are also

1591: larger than ours, being $\langle E(B - V) \rangle = 0.18$ and $0.13$~mag

1592: for the inner and outer fields, respectively.

1593: 

1594: In Table~\ref{tab:4258} we also listed the metallicity sensitivity

1595: ($\gamma$) in each band, which was computed from the distances and the average

1596: abundance difference in the two galactic fields.  The magnitude of this

1597: effect is smaller than the original estimate in \citeauthor{macri06},

1598: $\gamma=-0.29\pm0.09\ {\rm (random)}\pm0.05\ {\rm (systematic)}$ mag dex$^{-1}$.

1599: Since our estimate neglects the change in color with metallicity,

1600: we employed the metallicity corrections from \citet{kochanek97} in

1601: the next two sections.  \citeauthor{kochanek97} derived corrections in

1602: $UBVRIJHK$ based on a simultaneous analysis of Cepheids in 17 galaxies

1603: with a proper treatment of correlations between extinction, temperature,

1604: and metallicity.  The corrections are $+0.20\pm0.18$, $-0.08\pm0.14$,

1605: $-0.21\pm0.14$, $-0.26\pm0.13$, $-0.34\pm0.14$, and $-0.40\pm0.13$

1606: mag dex$^{-1}$ in $BVI_CJHK$, respectively, in the sense that

1607: metal-rich Cepheids have a decreased flux in $B$ but increased fluxes

1608: in $VI_CJHK$, possibly due to line-blanketing and back-warming among

1609: other effects. These estimates are based on $\langle E(B - V) \rangle

1610: = 0.15$ for LMC Cepheids, but their dependence on the LMC reddening

1611: is small.  If we adopt these corrections for NGC~4258, we would derive

1612: $(m - M)_0 = 29.34\pm0.13$ for the inner-field and $29.17\pm0.19$~mag

1613: for the outer-field Cepheids, where the errors represent the $1\sigma$

1614: statistical uncertainty.

1615: 

1616: \subsection{LMC}\label{sec:lmc}

1617: 

1618: Our Galactic {\it P-L} relations can also be applied to estimate the

1619: distance to the LMC and its mean reddening.  We included $JHK_s$

1620: photometry of the LMC Cepheids from \citet{persson04}, as well as $BVI_C$

1621: photometry from the OGLE-II catalog in the multi-wavelength fitting

1622: technique.  This approach assumes that Cepheids in these two surveys

1623: were drawn from the same population.\footnote{The OGLE-II Cepheids were

1624: observed along the LMC bar, while the Cepheids observed in $JHK_s$

1625: \citep{persson04} are more dispersed over the LMC.}

1626: 

1627: \begin{figure*}

1628: \epsscale{1.15}

1629: \plotone{f20.eps}

1630: \caption{{\it Left}: Apparent LMC distance moduli in $BVI_CJHK_s$ with

1631: the best-fitting CCM89 extinction curve ({\it solid line}).  {\it Middle}:

1632: Likelihood contours in the average LMC reddening and its distance modulus

1633: in $BVI_CJHK_s$ shown at $\Delta \chi^2$ = 2.30, 6.17, and 11.8 (68.3\%,

1634: 95.4\%, and 99.73\% confidence levels for 2 degrees of freedom).  The

1635: error bar represents the adopted distance and its error by {\it HST} Key

1636: Project at the average reddening from the OGLE-II reddening map.  {\it

1637: Right}: Same as the left panel, but likelihood contours when $BVI_C$

1638: photometry is used only.\label{fig:lmc}}

1639: \end{figure*}

1640: 

1641: Figure~\ref{fig:lmc} shows likelihood contours in the solution for

1642: the average LMC reddening and its distance modulus ({\it middle panel})

1643: from the fits on the apparent distance moduli in $BVI_CJHK_s$ ({\it left

1644: panel}). We applied metallicity corrections for $\Delta \langle \log

1645: {\rm (O/H) \rangle (Galaxy - LMC)} = 0.30$ and assumed that the

1646: metallicity sensitivities on the Glass-Carter $JHK$ system in \citet{kochanek97}

1647: are the same as in the LCO $JHK_s$ system.  As in the NGC~4258 case, we

1648: assumed $R_V = 3.35$ for the Cepheids, which yields absorption ratios

1649: $A_{\lambda}/A_V$ of 0.29, 0.18, and 0.12 for $JHK_s$, respectively,

1650: from the CCM89 extinction law.  In this calculation we took the effective

1651: filter wavelengths from \citet{persson04} for $JHK_s$ on the LCO system.

1652: We added $0.02$~mag in $BVI_C$ \citep{udalski99a} and $0.03$~mag in

1653: $JHK_s$ \citep{persson04} in quadrature to the fitting errors to account

1654: for photometric zero-point errors.

1655: 

1656: \input{tab18.tex}

1657: 

1658: As shown in Figure~\ref{fig:lmc}, our best-fit solution yields

1659: $(m - M)_0 = 18.34\pm0.06\pm0.16$ ({\it P-L} zero point).

1660: Extinction-corrected distance moduli at the best-fit reddening are

1661: listed in Table~\ref{tab:lmc}.  The first error in distance is a

1662: quadrature sum of a fitting error ($\pm0.05$~mag) and an error from

1663: $R_V$ ($\mp0.01$~mag for $\pm0.5$ in $R_V$).  We also added in

1664: quadrature an error from the metallicity ($\mp0.03$~mag for

1665: $\pm0.06$~dex).  As in the NGC~4258 case the zero-point error

1666: of the Galactic {\it P-L} relations dominates the combined error

1667: in the LMC distance modulus.  Without metallicity corrections, we would

1668: derive $(m - M)_0 = 18.48$.

1669: 

1670: We also derived $\langle E(B - V) \rangle = 0.11\pm0.02$, where the

1671: error is from the $1\sigma$ contour in Figure~\ref{fig:lmc}.  Our

1672: average reddening is 0.03~mag smaller than the average reddening from

1673: the OGLE-II, $\langle E(B - V) \rangle = 0.143$, which was adopted in

1674: the {\it HST} Key Project.\footnote{The {\it HST} Key Project adopted

1675: the OGLE-II LMC {\it P-L} relations and scaled them to $(m - M)_0 = 18.50$.

1676: However, the {\it P-L} relations were not scaled to $\langle E(B - V)

1677: \rangle = 0.10$ as they claimed \citep[see also][]{fouque03}.}  However,

1678: our reddening estimate is closer to the value of $E(B - V) \sim 0.10$

1679: \citep{walker99} from several different reddening estimates for LMC

1680: stars \citep[see Table~II in][]{mcnamara80}.  It is also in agreement

1681: with the reddening value determined for the LMC Cepheids, $\langle

1682: E(B - V) \rangle = 0.076$ \citep[]{caldwell85,laney07}.  The zero point

1683: of the OGLE-II reddening map was set by two LMC clusters (NGC~1850 and

1684: NGC~1835) and one eclipsing binary (HV~2274), but the reddening for

1685: HV~2274 has different values from various studies:  OGLE-II adopted

1686: $E(B - V) = 0.149\pm0.015$ \citep{udalski98}, which is larger than

1687: $E(B - V) = 0.120\pm0.009$ \citep{guinan98}, $0.088\pm0.023$

1688: \citep{nelson00}, and $0.103\pm0.007$ \citep{groenewegen01}.

1689: Furthermore, as shown in Table~\ref{tab:lmc}, extinction corrections

1690: from the OGLE-II reddening map result in systematically shorter

1691: distances at shorter wavelengths, which may indicate an overcorrection

1692: for extinction \citep[see also][]{fouque03}.

1693: 

1694: Our distance modulus is $0.16$~mag smaller (or $7\%$ in distance) than

1695: the adopted distance modulus by the {\it HST} Key Project ($18.50$~mag).

1696: Formally, this implies an increase in their $H_0$ by $7\%\pm8\%$.  Our

1697: LMC distance is also shorter by $0.20$~mag (or $9\%$ in distance) than

1698: the distance modulus adopted by the \citet{sandage06b} program ($18.54$~mag).

1699: In addition to our study, recent parallax studies from the {\it HST}

1700: \citep{benedict07} and the revised {\it Hipparcos} catalog \citep{vanleeuwen07}

1701: used the Wesenheit index $W_{VI}$ to derive $(m - M)_0 = 18.40\pm0.05$

1702: and $18.39\pm0.05$~mag, respectively, also suggesting a downward revision

1703: of the LMC distance.  Finally, \citet{macri06} estimated

1704: a distance modulus of NGC~4258 relative to the LMC of $10.88\pm0.04

1705: {\rm\ (statistical)}\pm0.05{\rm\ (systematic)}$.  Given the maser

1706: distance, this was then translated into the LMC distance modulus of

1707: $(m - M)_0 = 18.41\pm0.10{\rm\ (statistical)} \pm0.13{\rm\ (systematic)}$,

1708: which is within $1\sigma$ of our estimate.

1709: 

1710: We note that stronger constraints on distance can be achieved when several

1711: photometric bands are used including near-infrared data.  As illustrated

1712: in Figure~\ref{fig:lmc}, the statistical error in distance becomes

1713: larger by a factor of 3 when only $BVI_C$ information is used

1714: ({\it right panel}) rather than from $BVI_CJHK_s$ ({\it middle panel}).

1715: The error would be even larger when the distance and reddening values

1716: are estimated from only two bands, such as the Wesenheit index.

1717: 

1718: \subsection{M33}

1719: 

1720: As our final case, we estimated the average distance and reddening

1721: for M33 Cepheids using the same method as in the previous section.

1722: The distance to this nearby spiral galaxy has recently been measured

1723: from a detached eclipsing binary, $(m - M)_0 = 24.92\pm0.12$

1724: \citep{bonanos06}, and from two water maser sources in \ion{H}{2}

1725: regions, $(m - M)_0 = 24.32^{+0.45}_{-0.57}$ \citep{brunthaler05}.

1726: 

1727: \begin{figure*}

1728: \epsscale{0.75}

1729: \plotone{f21.eps}

1730: \caption{{\it Left:} Apparent distance moduli in $BVI_C$ with the

1731: best-fitting CCM89 extinction curve ({\it solid line}).  {\it Right:}

1732: Likelihood contours in the average reddening and distance modulus of M33

1733: Cepheids shown at $\Delta \chi^2$ = 2.30, 6.17, and 11.8 (68.3\%, 95.4\%,

1734: and 99.73\% confidence levels for 2 degrees of freedom).  Two open

1735: circles represent the distance and reddening estimates from an eclipsing

1736: binary \citep{bonanos06} and those from the {\it HST} Key Project.

1737: The horizontal lines represent a distance from the water maser sources

1738: and its $1\sigma$ error \citep{brunthaler05}.\label{fig:m33}}

1739: \end{figure*}

1740: 

1741: Figure~\ref{fig:m33} shows likelihood contours in the solution for

1742: the average reddening and distance modulus.  As in the {\it HST} Key

1743: Project, we used $BVI_C$ photometry in \citet{freedman91} for their

1744: 10 best observed Cepheids, which was obtained at the 3.6 m

1745: Canada-France-Hawaii Telescope (CFHT), the Kitt Peak National

1746: Observatory (KPNO) 4 m telescope, and the Palomar 1.5 m telescope.

1747: We assumed $0.03$~mag for the photometric zero-point error in each

1748: band.  On the \citet{zaritsky94} scale, M33 Cepheids have $\langle 12 +

1749: \log{\rm (O/H)} \rangle = 8.82$, which was adopted in the {\it HST}

1750: Key Project.  We therefore applied metallicity corrections from

1751: \citet{kochanek97} for $\Delta {\rm \langle \log (O/H) \rangle

1752: (Galaxy - M33}) = -0.02$.

1753: 

1754: At the best-fit reddening, we obtained an extinction-corrected distance

1755: modulus of $24.51\pm0.11$, $24.60\pm0.09$, and $24.53\pm0.08$ in

1756: $BVI_C$, respectively, where the uncertainties are fitting errors only.

1757: The average distance is $(m - M)_0 = 24.55\pm0.28\pm0.16$

1758: ({\it P-L} zero point).  The first error was determined by adding in

1759: quadrature the fitting error (0.28~mag), the error from $R_V$

1760: ($\mp0.03$~mag for $\Delta R_V = \pm0.5$), and the error from the

1761: metallicity ($\mp0.08$~mag for $\pm0.15$~dex).  We adopted the error in

1762: the metallicity from the {\it HST} Key Project.  If we adopt metal abundances

1763: derived from electron temperatures \citep{sakai04}, the abundance

1764: difference between the LMC and M33 becomes $\Delta \langle {\rm \log

1765: (O/H) \rangle} = +0.21$, which leads to $\Delta {\rm \langle \log (O/H)

1766: \rangle (Galaxy - M33)} = +0.09$.  We would derive $(m - M)_0 = 24.53$

1767: without metallicity corrections.

1768: 

1769: The {\rm HST} Key Project used the same ground-based photometry for

1770: this galaxy and derived a metallicity-corrected distance modulus of

1771: $(m - M)_0 = 24.62\pm0.15$ based only on $VI$ as for the other

1772: {\it HST} sample galaxies.  While this distance is in good

1773: agreement with our estimate, our average reddening of $\langle E(B - V)

1774: \rangle = 0.08\pm0.09$ is lower than their estimate,

1775: $\langle E(V - I) \rangle = 0.27\pm0.05$ or $\langle E(B - V) \rangle

1776: = 0.21\pm0.04$.\footnote{We note that \citet{freedman91} originally

1777: derived $\langle E(B - V) \rangle = 0.10\pm0.09$ and $(m - M)_0 =

1778: 24.64\pm0.09$ using $BVRI$ photometry, assuming the LMC distance modulus

1779: of 18.5~mag and its mean reddening of 0.10~mag.}  A part of the

1780: difference is due to different zero points of {\it P-L} relations.

1781: However, our reddening is in good agreement with the value of the

1782: eclipsing binary, $E(B - V) = 0.09\pm0.01$, which was obtained from

1783: $BVRJHK_s$ photometry \citep{bonanos06} and with previous estimates

1784: based on several different methods (see Table~7 of \citeauthor{bonanos06}).

1785: 

1786: A similar result was obtained from a larger Cepheid sample in

1787: \citet{macri01a}.  They provided $BVI_C$ photometry for 251 Cepheids

1788: obtained at the Fred Lawrence Whipple Observatory (FLWO) 1.2 m telescope

1789: and the MDM 1.3 m telescope \citep[see also][]{mochejska01a,mochejska01b}.

1790: From the restricted sample of 90 Cepheids in $1.1 \leq \log{P} \leq

1791: 1.7$ similar to those of the \citeauthor{freedman91} sample, we derived

1792: $(m - M)_0 = 24.67\pm0.18$ and $\langle E(B - V) \rangle = 0.04\pm0.06$,

1793: where the uncertainties are fitting errors only.

1794: 

1795: Distances from these two independent sets of photometry are about

1796: $0.3$~mag shorter than the distance from the eclipsing binary, although

1797: our distance has a large error.  One potential explanation for the

1798: short Cepheid-based distance is the unresolved blends by a strong

1799: star-to-star correlation function.  \citet{mochejska01c} investigated

1800: the influence of blending by comparing {\it HST} WFPC2 and

1801: \citeauthor{macri01a} ground-based images for a sample of 102 Cepheids

1802: \citep[see also][]{mochejska00}.  They found that the average flux

1803: contribution from unresolved luminous companions on the ground-based

1804: images could be on average $\sim20\%$ in $BVI$ for $\log{P}>1$, which

1805: would systematically underestimate the Cepheid distance by $\sim10\%$.

1806: In addition, \citet{lee02} obtained single-epoch $I$-band photometry

1807: for a subset of the \citeauthor{macri01a} sample using {\it HST} WFPC2

1808: and found that the {\it HST} photometry is on average $\approx0.2$~mag

1809: fainter than \citeauthor{macri01a} photometry.

1810: 

1811: Follow-up observations with a higher spatial resolution

1812: \citep[e.g.,][]{macri04} would be helpful for a definitive test of the

1813: distance to M33, although there are indications that our estimated reddening

1814: is more accurate than the value in the {\it HST} Key Project.  Note

1815: that the effect of stellar blending is less significant in our distance

1816: estimates for NGC~4258 and the LMC.  The ground-based photometry with

1817: the full width at half-maximum (FWHM) of $\sim1.5\arcsec$ would

1818: correspond to $\sim0.3$~pc at the distance of the LMC, compared to

1819: $\sim7$~pc at the distance of M33.  The ACS/WFC FWHM of 0.09\arcsec

1820: \citep{ford03} corresponds to $\sim3$~pc at the distance of NGC~4258.

1821: 

1822: \section{SUMMARY AND DISCUSSION}

1823: 

1824: We have continued our effort to improve the accuracy of isochrones and

1825: distances derived from MS fitting.  We extended the previous Hyades-based

1826: calibration to the upper MS by constructing empirical color-$T_{\rm eff}$

1827: corrections to match the observed Pleiades MS at the cluster's accurately

1828: known distance.  We applied these empirically calibrated sets of isochrones

1829: to Cepheid-bearing Galactic open clusters to derive distances, reddenings,

1830: and $R_V$ at the spectroscopic metal abundances and obtained Galactic

1831: {\it P-L} relations based on nine Cepheids in seven clusters.  Our distance

1832: modulus of individual Cepheids has an accuracy of $\sim0.08$~mag (or

1833: $4\%$ in distance), which is compatible to those of recent parallaxes

1834: from the {\it HST} \citep{benedict07} and the revised {\it Hipparcos}

1835: catalog \citep{vanleeuwen07}.

1836: 

1837: Using these {\it P-L} relations, we derived a distance to NGC~4258 and

1838: found that our Cepheid distance is in excellent agreement with the maser

1839: distance, supporting our distance scale from the MS-fitting technique.

1840: From $BVI_CJHK_s$ photometry we derived the LMC

1841: distance modulus of $(m - M)_0 = 18.34\pm0.06\pm0.16$ ({\it P-L} zero

1842: point) after applying metallicity corrections.  This is also in close

1843: agreement with distance estimates from the recent {\it HST} and {\it

1844: Hipparcos} parallax studies.  However, our revised LMC distance is lower

1845: than the distance adopted by the {\it HST} Key Project, which formally

1846: implies an increase in their $H_0$ by $7\%\pm8\%$.  A similar size of an

1847: increase in $H_0$ is expected from the SN~Ia calibration program

1848: by \citet{sandage06b}.

1849: 

1850: Our reddening estimates are systematically lower than those in the

1851: {\it HST} Key Project.  Part of the reason is the {\it P-L} relations

1852: with different assumptions about the reddening.  They adopted the color

1853: excess for LMC Cepheids from the OGLE-II reddening map, but its average

1854: value is $\Delta \langle E(B - V) \rangle \approx 0.03$ larger

1855: than our LMC reddening estimate.  Nonetheless, relative distances

1856: between the LMC and target galaxies in the {\it HST} Key Project would

1857: remain less affected because the reddening-free or Wesenheit index

1858: adopted in the {\it HST} Key Project is designed to avoid

1859: the problem of knowing the absolute zero point of the reddening scale.

1860: In other words, the absolute reddening value would not be important

1861: as long as the difference in reddening between target Cepheids and

1862: the calibrating system is well-defined.  Therefore, the Wesenheit index

1863: can be effectively used to build the cosmic distance scale once highly

1864: accurate distances to calibrating Cepheids are available.

1865: 

1866: However, our $E(B - V)$ estimates for NGC~4258 and M33 are larger by

1867: $\langle E(B - V) \rangle \sim0.1$ than those in the {\it HST} Key

1868: Project, more than expected from the difference in the LMC reddening.

1869: This is due to two reasons.  First, the Cepheid's color is not only

1870: affected by reddening, but it can be also affected by the metal content.

1871: The {\it HST} Key Project corrected a distance modulus for a metallicity

1872: effect after deriving a color excess.  On the other hand, we applied

1873: metallicity corrections to apparent distance moduli from

1874: \citet{kochanek97}, and then derived a reddening value.

1875: Because the sign and size of the metallicity corrections could not be

1876: the same in different broadband filters, the difference in the Cepheid

1877: metallicity could lead to different reddening estimates from those in

1878: the {\it HST} Key Project.  Second, the Wesenheit magnitude is

1879: based on only two broadband filters.  However, reddening and distance

1880: estimates are naturally correlated in the standard procedure to estimate

1881: these parameters using {\it P-L} relations, so a small zero-point error

1882: in the photometry, for example, can be translated into a large error in

1883: reddening.  Therefore, obtaining Cepheid photometry at least in three

1884: or more passbands including near-infrared filters is of great importance

1885: in the distance and reddening estimation \citep[e.g.,][]{gieren06,soszynski06}.

1886: 

1887: In the determination of individual Cepheids' absolute magnitudes, the extinction

1888: is one of the largest sources of error because of strong differential

1889: reddening in Cepheid-bearing young open clusters.  Although we found

1890: an error of $\sim0.03$~mag in the mean $E(B - V)$ for a cluster, most

1891: Cepheid-bearing clusters have patchy reddening that makes the

1892: individual Cepheid reddening uncertain by up to $\sigma_{E(B-V)} \sim

1893: 0.12$.  An average reddening of cluster members located near a Cepheid

1894: was often used to determine the Cepheid's reddening, but differential

1895: reddening over a small scale could not be resolved unless we have a sufficient

1896: number of nearby stars with well-determined reddening.  However, it is

1897: noted that the most secure reddening estimates for Cepheids are considered

1898: to be given by the cluster Cepheids.  In addition, our LMC reddening is

1899: in close agreement with the previous standard value, $\langle E(B - V) \rangle

1900: \sim0.10$ \citep{walker99}, and our average reddening for M33 Cepheids

1901: is in good agreement with the value inferred from $BVRJHK_s$ photometry

1902: of an eclipsing binary \citep{bonanos06}.

1903: 

1904: In terms of the zero point of the Galactic {\it P-L} relations, the

1905: uncertainty in the metallicity scale of the Galactic open clusters is

1906: the largest error source, with a change of $\Delta {\rm [M/H]}\sim0.1$~dex

1907: producing a $\sim0.1$~mag shift in distance modulus.  No other source

1908: is unusually large.  In Paper~III we have shown that several color

1909: indices can be effectively used in MS fitting to determine a cluster

1910: metallicity because of differential sensitivities of these colors on

1911: metal abundance.  Although the noise in the current photometry prevented

1912: the estimation of the photometric metal abundance, deep multi-color

1913: photometry combined with near-infrared data will be useful to determine

1914: $R_V$, $E(B - V)$, $(m - M)_0$, and [M/H], simultaneously.  A more

1915: accurate metallicity scale could then be used to reduce the size of the systematic

1916: error in the {\it P-L} zero point.  In addition, the zero-point error

1917: can be further reduced by deeper photometry of other Cepheid-bearing clusters,

1918: which were not included in this paper because of poor photometric quality.

1919: These are C1814-191, Collinder~394, Mon~OB2, NGC~1647, NGC~6649,

1920: NGC~6664, Platais~1, Trumpler~35, Vel~OB5, and Vul~OB1.

1921: 

1922: \acknowledgements

1923: We are very grateful to Krzysztof Stanek for his valuable comments

1924: and suggestions.  We gratefully acknowledge Michael Bessell for

1925: answering a number of questions about his computations of

1926: color-excess ratios, Fiona Hoyle for providing her photometry,

1927: and Imants Platais for pointing out some problematic features of

1928: the isochrones, which has led us to investigate the LCB color correction

1929: tables more carefully.  We also would like to thank Andrew Gould,

1930: Lucas Macri, Fritz Benedict, Christopher Kochanek, and Jennifer Johnson

1931: for useful discussions.  We thank Richard Pogge, Subo Dong, and

1932: Kristen Sellgren for their help.

1933: The work reported here was partially supported by grants AST-0206008

1934: and AST-0205789 from the National Science Foundation to the Ohio

1935: State Research Foundation, and with funds provided by the Ohio State

1936: University Department of Astronomy.

1937: This publication makes use of data products from the Two Micron All Sky

1938: Survey, which is a joint project of the University of Massachusetts

1939: and the Infrared Processing and Analysis Center/California Institute of

1940: Technology, funded by the National Aeronautics and Space Administration

1941: and the National Science Foundation.

1942: 

1943: \begin{thebibliography}{}

1944: 

1945: \bibitem[Allen(1973)]{allen73}

1946: Allen, C. W. 1973, Astrophysical Quantities (London: Athlone)

1947: 

1948: \bibitem[Alonso et~al.(1995)Alonso, Arribas, \& Mart\'{\i}nez-Roger]{alonso95}

1949: Alonso, A., Arribas, S., \& Mart\'{\i}nez-Roger, C.  1995, \aap, 297, 197

1950: 

1951: \bibitem[Alonso et~al.(1996)Alonso, Arribas, \& Mart\'{\i}nez-Roger]{alonso96}

1952: Alonso, A., Arribas, S., \& Mart\'{\i}nez-Roger, C.  1996, \aaps, 117, 227

1953: 

1954: \bibitem[An et al.(2007)]{an07}

1955: An, D., Terndrup, D.~M., Pinsonneault, M.~H., Paulson, D.~B.,

1956: Hanson, R.~B., \& Stauffer, J.~R.\ 2007, \apj, 655, 233 (Paper~III)

1957: 

1958: \bibitem[Andrievsky et al.(2002)]{andrievsky02}

1959: Andrievsky, S.~M., et al.\ 2002, \aap, 381, 32 

1960: 

1961: \bibitem[Argon et al.(2007)]{argon07}

1962: Argon, A.~L., Greenhill, L.~J., Reid, M.~J., Moran, J.~M., \& Humphreys,

1963: E.~M.~L.\ 2007, \apj, 659, 1040 

1964: 

1965: \bibitem[Arp(1960)]{arp60}

1966: Arp, H.\ 1960, \apj, 131, 322

1967: 

1968: \bibitem[Arp et al.(1959)Arp, Sandage, \& Stephens]{arp59}

1969: Arp, H., Sandage, A., \& Stephens, C.\ 1959, \apj, 130, 80

1970: 

1971: \bibitem[Bahcall \& Pinsonneault(2004)]{bahcall04}

1972: Bahcall, J.  N., \& Pinsonenault, M.  H.  2004, Phys.  Rev.  Lett., 92, 121301

1973: 

1974: \bibitem[Bahcall et~al.(2001)Bahcall, Pinsonneault, \& Basu]{bahcall01}

1975: Bahcall, J.  N., Pinsonneault, M.  H., \& Basu, S.  2001, \apj, 555, 990

1976: 

1977: \bibitem[Basu et~al.(2000)Basu, Pinsonneault, \& Bahcall]{basu00}

1978: Basu, S., Pinsonneault, M.  H., \& Bahcall, J.  N.  2000, ApJ, 529, 1084

1979: 

1980: \bibitem[Baumgardt et al.(2000)Baumgardt, Dettbarn, \& Wielen]{baumgardt00}

1981: Baumgardt, H., Dettbarn, C., \& Wielen, R.\ 2000, \aaps, 146, 251

1982: 

1983: \bibitem[Benedict et al.(2002a)]{benedict02a}

1984: Benedict, G.~F., et al.\ 2002a, \aj, 123, 473

1985: 

1986: \bibitem[Benedict et al.(2002b)]{benedict02b}

1987: Benedict, G.~F., et al.\ 2002b, \aj, 124, 1695

1988: 

1989: \bibitem[Benedict et al.(2007)]{benedict07}

1990: Benedict, G.~F., et al.\ 2007, \aj, 133, 1810

1991: 

1992: \bibitem[Berdnikov et al.(2000)Berdnikov, Dambis, and Vozyakova]{berdnikov00}

1993: Berdnikov, L.~N., Dambis, A.~K., \& Vozyakova, O.~V.\ 2000, \aaps, 143, 211

1994: 

1995: \bibitem[Bessell(2005)]{bessell05}

1996: Bessell, M.~S.\ 2005, \araa, 43, 293 

1997: 

1998: \bibitem[Bessell et~al.(1998)Bessell, Castelli, \& Plez]{bessell98}

1999: Bessell, M.  S., Castelli, F., \& Plez, B.  1998, \aap, 333, 231 (BCP98)

2000: 

2001: \bibitem[B{\"o}hm-Vitense et al.(1998)]{bohm98} 

2002: B{\"o}hm-Vitense, E., Evans, N.~R., Carpenter, K., Albrow, M.~D., Cottrell, 

2003: P.~L., Robinson, R., \& Beck-Winchatz, B.\ 1998, \apj, 505, 903

2004: 

2005: \bibitem[Bonanos et al.(2006)]{bonanos06}

2006: Bonanos, A.~Z., et al.\ 2006, \apj, 652, 313

2007: 

2008: \bibitem[Breger(1966)]{breger66}

2009: Breger, M.\ 1966, \pasp, 78, 293

2010: 

2011: \bibitem[Breger(1986)]{breger86}

2012: Breger, M.\ 1986, \apj, 309, 311

2013: 

2014: \bibitem[de Bruijne et~al.(2001)de Bruijne, Hoogerwerf, \& de Zeeuw]{debruijne01}

2015: de Bruijne, J.  H.  J., Hoogerwerf, R., \& de Zeeuw, P.  T.

2016: 2001, \aap, 367, 111

2017: 

2018: \bibitem[Brunthaler et al.(2005)]{brunthaler05}

2019: Brunthaler, A., Reid, M.~J., Falcke, H., Greenhill, L.~J.,

2020: \& Henkel, C.\ 2005, Science, 307, 1440

2021: 

2022: \bibitem[Caldwell \& Coulson(1985)]{caldwell85}

2023: Caldwell, J.~A.~R., \& Coulson, I.~M.\ 1985, \mnras, 212, 879

2024: 

2025: \bibitem[Cardelli et al.(1989)Cardelli, Clayton, \& Mathis]{cardelli89}

2026: Cardelli, J.~A., Clayton, G.~C., \& Mathis, J.~S.\ 1989,

2027: \apj, 345, 245 (CCM89)

2028: 

2029: \bibitem[Carpenter(2001)]{carpenter01}

2030: Carpenter, J.~M.\ 2001, \aj, 121, 2851

2031: 

2032: \bibitem[Carter(1990)]{carter90}

2033: Carter, B.~S.\ 1990, \mnras, 242, 1

2034: 

2035: \bibitem[Clari{\'a} et al.(1991)Clari{\'a}, Lapasset, \& Bosio]{claria91}

2036: Clari{\'a}, J.~J., Lapasset, E., \& Bosio, M.~A.\ 1991, \mnras, 249, 193

2037: 

2038: \bibitem[Cohen et al.(2003)Cohen, Wheaton, \& Megeath]{cohen03}

2039: Cohen, M., Wheaton, W.~A., \& Megeath, S.~T.\ 2003, \aj, 126, 1090 

2040: 

2041: \bibitem[Collins \& Sonneborn(1977)]{collins77}

2042: Collins, G.~W., II, \& Sonneborn, G.~H.\ 1977, \apjs, 34, 41

2043: 

2044: \bibitem[Dean et al.(1978)Dean, Warren, \& Cousins]{dean78}

2045: Dean, J.~F., Warren, P.~R., \& Cousins, A.~W.~J.\ 1978, \mnras, 183, 569 

2046: 

2047: \bibitem[Evans et al.(1997)]{evans97}

2048: Evans, N.~R., B{\"o}hm-Vitense, E., Carpenter, K., Beck-Winchatz, B.,

2049: \& Robinson, R.\ 1997, \pasp, 109, 789

2050: 

2051: \bibitem[Evans et al.(1998)]{evans98}

2052: Evans, N.~R., B{\"o}hm-Vitense, E., Carpenter, K., Beck-Winchatz, B.,

2053: \& Robinson, R.\ 1998, \apj, 494, 768 

2054: 

2055: \bibitem[Evans \& Bolton(1990)]{evans90}

2056: Evans, N.~R., \& Bolton, C.~T.\ 1990, \apj, 356, 630 

2057: 

2058: \bibitem[Evans et al.(2006)]{evans06}

2059: Evans, N.~R., Massa, D., Fullerton, A., Sonneborn, G.,

2060: \& Iping, R.\ 2006, \apj, 647, 1387

2061: 

2062: \bibitem[Feast(1957)]{feast57}

2063: Feast, M.~W.\ 1957, \mnras, 117, 193 

2064: 

2065: \bibitem[Feast(1999)]{feast99}

2066: Feast, M.\ 1999, \pasp, 111, 775

2067: 

2068: \bibitem[Feast(2003)]{feast03}

2069: Feast, M.\ 2003, in Stellar Candles for the Extragalactic Distance Scale,

2070: ed. D. Alloin \& W. Gieren (Berlin: Springer), 45

2071: 

2072: \bibitem[Feast \& Catchpole(1997)]{feast97}

2073: Feast, M.~W., \& Catchpole, R.~M.\ 1997, \mnras, 286, L1

2074: 

2075: \bibitem[Feast et al.(1998)]{feast98}

2076: Feast, M., Pont, F., \& Whitelock, P.\ 1998, \mnras, 298, L43

2077: 

2078: \bibitem[Feast \& Walker(1987)]{feast87}

2079: Feast, M.~W., \& Walker, A.~R.\ 1987, \araa, 25, 345

2080: 

2081: \bibitem[Fernie(1961)]{fernie61}

2082: Fernie, J.~D.\ 1961, \apj, 133, 64

2083: 

2084: \bibitem[Fernie(1990)]{fernie90}

2085: Fernie, J.~D.\ 1990, \apjs, 72, 153

2086: 

2087: \bibitem[Fernie(1994)]{fernie94}

2088: Fernie, J.~D.\ 1994, \apj, 429, 844

2089: 

2090: \bibitem[Fernie et al.(1995)]{fernie95}

2091: Fernie, J.~D., Beattie, B., Evans, N.~R., \& Seager, S.\ 1995, IBVS No. 4148

2092: 

2093: \bibitem[Ford et al.(2003)]{ford03}

2094: Ford, H.~C., et al.\ 2003, \procspie, 4854, 81

2095: 

2096: \bibitem[Fouqu{\'e} et al.(2003)Fouqu{\'e}, Storm, \& Gieren]{fouque03}

2097: Fouqu{\'e}, P., Storm, J., \& Gieren, W.\ 2003,

2098: in Stellar Candles for the Extragalactic Distance Scale,

2099: ed. D. Alloin \& W. Gieren (Berlin: Springer), 21

2100: 

2101: \bibitem[Freedman \& Madore(1990)]{freedman90}

2102: Freedman, W.~L., \& Madore, B.~F.\ 1990, \apj, 365, 186

2103: 

2104: \bibitem[Freedman et al.(1991)Freedman, Wilson, \& Madore]{freedman91}

2105: Freedman, W.~L., Wilson, C.~D., \& Madore, B.~F.\ 1991, \apj, 372, 455

2106: 

2107: \bibitem[Freedman et al.(2001)]{freedman01}

2108: Freedman, W.~L., et al.\ 2001, \apj, 553, 47

2109: 

2110: \bibitem[Fry \& Carney(1997)]{fry97}

2111: Fry, A.~M., \& Carney, B.~W.\ 1997, \aj, 113, 1073

2112: 

2113: \bibitem[Gieren et al.(1997)Gieren, Fouqu{\'e}, \& G{\'o}mez]{gieren97}

2114: Gieren, W.~P., Fouqu{\'e}, P., \& G{\'o}mez, M.\ 1997, \apj, 488, 74

2115: 

2116: \bibitem[Gieren et al.(1998)Gieren, Fouqu{\'e}, \& G{\'o}mez]{gieren98}

2117: Gieren, W.~P., Fouqu{\'e}, P., \& G{\'o}mez, M.\ 1998, \apj, 496, 17

2118: 

2119: \bibitem[Gieren et al.(2006)]{gieren06}

2120: Gieren, W., Pietrzy{\'n}ski, G., Nalewajko, K., Soszy{\'n}ski, I.,

2121: Bresolin, F., Kudritzki, R.-P., Minniti, D., \& Romanowsky, A.\ 2006,

2122: \apj, 647, 1056

2123: 

2124: \bibitem[Gieren et al.(2005)]{gieren05}

2125: Gieren, W., Storm, J., Barnes, T.~G., III, Fouqu{\'e}, P.,

2126: Pietrzy{\'n}ski, G., \& Kienzle, F.\ 2005, \apj, 627, 224

2127: 

2128: \bibitem[Groenewegen \& Oudmaijer(2000)]{groenewegen00}

2129: Groenewegen, M.~A.~T., \& Oudmaijer, R.~D.\ 2000, \aap, 356, 849 

2130: 

2131: \bibitem[Groenewegen \& Salaris(2001)]{groenewegen01}

2132: Groenewegen, M.~A.~T., \& Salaris, M.\ 2001, \aap, 366, 752

2133: 

2134: \bibitem[Groenewegen et al.(2004)]{groenewegen04}

2135: Groenewegen, M.~A.~T., Romaniello, M., Primas, F.,

2136: \& Mottini, M.\ 2004, \aap, 420, 655

2137: 

2138: \bibitem[Gould(1994)]{gould94}

2139: Gould, A.\ 1994, \apj, 426, 542 

2140: 

2141: \bibitem[Guinan et al.(1998)]{guinan98}

2142: Guinan, E.~F., et al.\ 1998, \apjl, 509, L21

2143: 

2144: \bibitem[Gupta et al.(2000)]{gupta00}

2145: Gupta, A.~C., Subramaniam, A., Sagar, R.,

2146: \& Griffiths, W.~K.\ 2000, \aaps, 145, 365

2147: 

2148: \bibitem[Haug(1978)]{haug78}

2149: Haug, U.\ 1978, \aaps, 34, 417

2150: 

2151: \bibitem[Herrnstein et al.(1999)]{herrnstein99}

2152: Herrnstein, J.~R., et al.\ 1999, \nat, 400, 539

2153: 

2154: \bibitem[Hertzsprung(1947)]{hertzsprung47}

2155: Hertzsprung, E.\ 1947, Annalen van de Sterrewacht te Leiden, 19, 1

2156: 

2157: \bibitem[Hoag et al.(1961)]{hoag61}

2158: Hoag, A.~A., Johnson, H.~L., Iriarte, B., Mitchell, R.~I., Hallam, K.~L.,

2159: \& Sharpless, S.\ 1961, Publications of the U.S.~Naval Observatory Second Series, 17, 345

2160: 

2161: \bibitem[Hoyle et al.(2003)Hoyle, Shanks, \& Tanvir]{hoyle03}

2162: Hoyle, F., Shanks, T., \& Tanvir, N.~R.\ 2003, \mnras, 345, 269

2163: 

2164: \bibitem[Johnson(1960)]{johnson60}

2165: Johnson, H.~L.\ 1960, \apj, 131, 620

2166: 

2167: \bibitem[Kanbur \& Ngeow(2004)]{kanbur04}

2168: Kanbur, S.~M., \& Ngeow, C.-C.\ 2004, \mnras, 350, 962

2169: 

2170: \bibitem[Kennicutt et al.(2003)Kennicutt, Bresolin, \& Garnett]{kennicutt03}

2171: Kennicutt, R.~C., Jr., Bresolin, F., \& Garnett, D.~R.\ 2003,

2172: \apj, 591, 801

2173: 

2174: \bibitem[Kennicutt et al.(1998)]{kennicutt98}

2175: Kennicutt, R.~C., Jr., et al.\ 1998, \apj, 498, 181

2176: 

2177: \bibitem[Kervella et al.(2004a)]{kervella04a}

2178: Kervella, P., Bersier, D., Mourard, D., Nardetto, N.,

2179: \& Coud{\'e} du Foresto, V.\ 2004a, \aap, 423, 327

2180: 

2181: \bibitem[Kervella et al.(2004b)]{kervella04b}

2182: Kervella, P., Nardetto, N., Bersier, D., Mourard, D.,

2183: \& Coud{\'e} du Foresto, V.\ 2004b, \aap, 416, 941

2184: 

2185: \bibitem[Kim et al.(2002)]{kim02}

2186: Kim, Y.-C., Demarque, P., Yi, S.~K.,

2187: \& Alexander, D.~R.\ 2002, \apjs, 143, 499

2188: 

2189: \bibitem[Kochanek(1997)]{kochanek97}

2190: Kochanek, C.~S.\ 1997, \apj, 491, 13

2191: 

2192: \bibitem[Kraft(1958)]{kraft58}

2193: Kraft, R.~P.\ 1958, \apj, 128, 161

2194: 

2195: \bibitem[Kraft(1965)]{kraft65}

2196: Kraft, R.~P.\ 1965, \apj, 142, 681

2197: 

2198: \bibitem[Landolt(1983)]{landolt83}

2199: Landolt, A.~U.\ 1983, \aj, 88, 439

2200: 

2201: \bibitem[Landolt(1992)]{landolt92}

2202: Landolt, A.~U.\ 1992, \aj, 104, 340 

2203: 

2204: \bibitem[Lane et al.(2000)]{lane00}

2205: Lane, B.~F., Kuchner, M.~J., Boden, A.~F., Creech-Eakman, M.,

2206: \& Kulkarni, S.~R.\ 2000, \nat, 407, 485

2207: 

2208: \bibitem[Lane et al.(2002)Lane, Creech-Eakman, \& Nordgren]{lane02}

2209: Lane, B.~F., Creech-Eakman, M.~J., \& Nordgren, T.~E.\ 2002,

2210: \apj, 573, 330

2211: 

2212: \bibitem[Laney \& Stobie(1992)]{laney92}

2213: Laney, C.~D., \& Stobie, R.~S.\ 1992, \aaps, 93, 93

2214: 

2215: \bibitem[Laney \& Stobie(1993)]{laney93}

2216: Laney, C.~D., \& Stobie, R.~S.\ 1993, \mnras, 263, 921

2217: 

2218: \bibitem[Laney \& Caldwell(2007)]{laney07}

2219: Laney, C.~D., \& Caldwell, J.~A.~R.\ 2007, \mnras, 377, 147

2220: 

2221: \bibitem[Lanoix et al.(1999)Lanoix, Paturel, \& Garnier]{lanoix99}

2222: Lanoix, P., Paturel, G., \& Garnier, R.\ 1999, \mnras, 308, 969 

2223: 

2224: \bibitem[Lee et al.(2002)]{lee02}

2225: Lee, M.~G., Kim, M., Sarajedini, A., Geisler, D.,

2226: \& Gieren, W.\ 2002, \apj, 565, 959

2227: 

2228: \bibitem[Lee \& Lee(1999)]{lee99}

2229: Lee, J.-D., \& Lee, S.-G.\ 1999,

2230: Journal of Korean Astronomical Society, 32, 91 

2231: 

2232: \bibitem[Lejeune et al.(1997)Lejeune, Cuisinier, \& Buser]{lejeune97}

2233: Lejeune, Th., Cuisinier, F., \& Buser, R.\ 1997, \aaps, 125, 229 (LCB)

2234: 

2235: \bibitem[Lejeune et al.(1998)Lejeune, Cuisinier, \& Buser]{lejeune98}

2236: Lejeune, Th., Cuisinier, F., \& Buser, R.\ 1998, \aaps, 130, 65 (LCB)

2237: 

2238: \bibitem[Luck et al.(1998)]{luck98}

2239: Luck, R.~E., Moffett, T.~J., Barnes, T.~G., III,

2240: \& Gieren, W.~P.\ 1998, \aj, 115, 605 

2241: 

2242: \bibitem[Luri et al.(1998)]{luri98}

2243: Luri, X., G{\'o}mez, A.~E., Torra, J., Figueras, F., \& Mennessier, M.~O.\ 1998,

2244: \aap, 335, L81

2245: 

2246: \bibitem[Lyng{\aa}(1987)]{lynga87}

2247: Lyng{\aa}, G.\ 1987, Catalogue of Open Cluster Data, 5th ed.

2248: (Lund: Lund Observatory)

2249: 

2250: \bibitem[Macri(2004)]{macri04}

2251: Macri, L.~M.\ 2004, IAU Colloq.~193: Variable Stars in the Local Group, 310, 33

2252: 

2253: \bibitem[Macri et~al.(2006)]{macri06}

2254: Macri, L.~M., Stanek, K.~Z., Bersier, D., Greenhill, L.~J.,

2255: \& Reid, M.~J.\ 2006, \apj, 652, 1133

2256: 

2257: \bibitem[Macri et al.(2001a)]{macri01a}

2258: Macri, L.~M., Stanek, K.~Z., Sasselov, D.~D., Krockenberger, M.,

2259: \& Kaluzny, J.\ 2001a, \aj, 121, 870

2260: 

2261: \bibitem[Macri et al.(2001b)]{macri01b}

2262: Macri, L.~M., et al.\ 2001b, \apj, 549, 721

2263: 

2264: \bibitem[Madore(1975)]{madore75}

2265: Madore, B.~F.\ 1975, \aap, 38, 471

2266: 

2267: \bibitem[Madore \& Freedman(1991)]{madore91}

2268: Madore, B.~F., \& Freedman, W.~L.\ 1991, \pasp, 103, 933

2269: 

2270: \bibitem[Madore \& Freedman(1998)]{madore98}

2271: Madore, B.~F., \& Freedman, W.~L.\ 1998, \apj, 492, 110 

2272: 

2273: \bibitem[Maoz et al.(1999)]{maoz99}

2274: Maoz, E., Newman, J.~A., Ferrarese, L., Stetson, P.~B.,

2275: Zepf, S.~E., Davis, M., Freedman, W.~L., \& Madore, B.~F.\ 1999,

2276: \nat, 401, 351

2277: 

2278: \bibitem[Mathis(1990)]{mathis90}

2279: Mathis, J.~S.\ 1990, \araa, 28, 37

2280: 

2281: \bibitem[McNamara \& Feltz(1980)]{mcnamara80}

2282: McNamara, D.~H., \& Feltz, K.~A., Jr.\ 1980, \pasp, 92, 587

2283: 

2284: \bibitem[Mermilliod(1981)]{mermilliod81}

2285: Mermilliod, J.~C.\ 1981, \aap, 97, 235

2286: 

2287: \bibitem[Mermilliod et al.(1987)Mermilliod, Mayor, \& Burki]{mermilliod87}

2288: Mermilliod, J.~C., Mayor, M., \& Burki, G.\ 1987, \aaps, 70, 389 

2289: 

2290: \bibitem[Mermilliod \& Paunzen(2003)]{mermilliod03}

2291: Mermilliod, J.-C., \& Paunzen, E.\ 2003, \aap, 410, 511

2292: 

2293: \bibitem[Meynet et al.(1993)Meynet, Mermilliod, \& Maeder]{meynet93}

2294: Meynet, G., Mermilliod, J.-C., \& Maeder, A.\ 1993, \aaps, 98, 477

2295: 

2296: \bibitem[Mochejska et al.(2001a)]{mochejska01a}

2297: Mochejska, B.~J., Kaluzny, J., Stanek, K.~Z., Sasselov, D.~D.,

2298: \& Szentgyorgyi, A.~H.\ 2001a, \aj, 121, 2032

2299: 

2300: \bibitem[Mochejska et al.(2001b)]{mochejska01b}

2301: Mochejska, B.~J., Kaluzny, J., Stanek, K.~Z., Sasselov, D.~D.,

2302: \& Szentgyorgyi, A.~H.\ 2001b, \aj, 122, 2477

2303: 

2304: \bibitem[Mochejska et al.(2000)]{mochejska00}

2305: Mochejska, B.~J., Macri, L.~M., Sasselov, D.~D.,

2306: \& Stanek, K.~Z.\ 2000, \aj, 120, 810

2307: 

2308: \bibitem[Mochejska et al.(2001c)]{mochejska01c}

2309: Mochejska, B.~J., Macri, L.~M., Sasselov, D.~D.,

2310: \& Stanek, K.~Z.\ 2001c, preprint (astro-ph/0103440)

2311: 

2312: \bibitem[Moffat \& Vogt(1973)]{moffat73}

2313: Moffat, A.~F.~J., \& Vogt, N.\ 1973, \aaps, 10, 135

2314: 

2315: \bibitem[Moffat \& Vogt(1975)]{moffat75}

2316: Moffat, A.~F.~J., \& Vogt, N.\ 1975, \aaps, 20, 155 

2317: 

2318: \bibitem[Nelson et al.(2000)]{nelson00}

2319: Nelson, C.~A., Cook, K.~H., Popowski, P., \& Alves, D.~R.\ 2000,

2320: \aj, 119, 1205

2321: 

2322: \bibitem[Newman et al.(2001)]{newman01}

2323: Newman, J.~A., Ferrarese, L., Stetson, P.~B., Maoz, E., Zepf, S.~E.,

2324: Davis, M., Freedman, W.~L., \& Madore, B.~F.\ 2001, \apj, 553, 562

2325: 

2326: \bibitem[Ngeow et al.(2005)]{ngeow05}

2327: Ngeow, C.-C., Kanbur, S.~M., Nikolaev, S., Buonaccorsi, J.,

2328: Cook, K.~H., \& Welch, D.~L.\ 2005, \mnras, 363, 831

2329: 

2330: \bibitem[Orsatti et al.(2001)Orsatti, Vega, \& Marraco]{orsatti01}

2331: Orsatti, A.~M., Vega, E.~I., \& Marraco, H.~G.\ 2001, \aap, 380, 130

2332: 

2333: \bibitem[Paulson et~al.(2003)Paulson, Sneden, \& Cochran]{paulson03}

2334: Paulson, D.  B., Sneden, C., \& Cochran, W.  D.  2003, \aj, 125, 3185

2335: 

2336: \bibitem[Perryman et~al.(1998)]{perryman98}

2337: Perryman, M.~A.~C., et~al.\ 1998, \aap, 331, 81

2338: 

2339: \bibitem[Persson et al.(2004)]{persson04}

2340: Persson, S.~E., Madore, B.~F., Krzemi{\'n}ski, W., Freedman, W.~L.,

2341: Roth, M., \& Murphy, D.~C.\ 2004, \aj, 128, 2239

2342: 

2343: \bibitem[Phelps \& Janes(1994)]{phelps94}

2344: Phelps, R.~L., \& Janes, K.~A.\ 1994, \apjs, 90, 31

2345: 

2346: \bibitem[Piatti et al.(1998)Piatti, Clari{\'a}, \& Bica]{piatti98}

2347: Piatti, A.~E., Clari{\'a}, J.~J., \& Bica, E.\ 1998, \apjs, 116, 263

2348: 

2349: \bibitem[Pinsonneault et~al.(2003)]{pinsono03}

2350: Pinsonneault, M.  H., Terndrup, D.  M., Hanson, R.  B., \& Stauffer, J.  R.

2351: 2003, \apj, 598, 588 (Paper~I)

2352: 

2353: \bibitem[Pinsonneault et~al.(2004)]{pinsono04}

2354: Pinsonneault, M.  H., Terndrup, D.  M., Hanson, R.  B., \& Stauffer, J.  R.

2355: 2004, \apj, 600, 946 (Paper~II)

2356: 

2357: \bibitem[Preston(1964)]{preston64}

2358: Preston, G.~W.\ 1964, \pasp, 76, 165 

2359: 

2360: \bibitem[Ribas et al.(2005)]{ribas05}

2361: Ribas, I., Jordi, C., Vilardell, F., Fitzpatrick, E.~L.,

2362: Hilditch, R.~W., \& Guinan, E.~F.\ 2005, \apjl, 635, L37

2363: 

2364: \bibitem[Rieke \& Lebofsky(1985)]{rieke85}

2365: Rieke, G.~H., \& Lebofsky, M.~J.\ 1985, \apj, 288, 618 

2366: 

2367: \bibitem[Romaniello et al.(2005)]{romaniello05}

2368: Romaniello, M., Primas, F., Mottini, M., Groenewegen, M.,

2369: Bono, G., \& Fran{\c c}ois, P.\ 2005, \aap, 429, L37

2370: 

2371: \bibitem[Romeo et al.(1989)]{romeo89}

2372: Romeo, G., Bonifazi, A., Fusi Pecci, F., \& Tosi, M.\ 1989, \mnras, 240, 459

2373: 

2374: \bibitem[Sagar \& Cannon(1997)]{sagar97}

2375: Sagar, R., \& Cannon, R.~D.\ 1997, \aaps, 122, 9

2376: 

2377: \bibitem[Sakai et al.(2004)]{sakai04}

2378: Sakai, S., Ferrarese, L., Kennicutt, R.~C., Jr.,

2379: \& Saha, A.\ 2004, \apj, 608, 42

2380: 

2381: \bibitem[Salpeter(1955)]{salpeter55}

2382: Salpeter, E.~E.\ 1955, \apj, 121, 161

2383: 

2384: \bibitem[Sandage(1958)]{sandage58}

2385: Sandage, A.\ 1958, \apj, 128, 150

2386: 

2387: \bibitem[Sandage(1960)]{sandage60}

2388: Sandage, A.\ 1960, \apj, 131, 610

2389: 

2390: \bibitem[Sandage \& Tammann(2006)]{sandage06a}

2391: Sandage, A., \& Tammann, G.~A.\ 2006, \araa, 44, 93

2392: 

2393: \bibitem[Sandage et al.(2004)Sandage, Tammann, \& Reindl]{sandage04}

2394: Sandage, A., Tammann, G.~A., \& Reindl, B.\ 2004, \aap, 424, 43

2395: 

2396: \bibitem[Sandage et al.(2006)]{sandage06b}

2397: Sandage, A., Tammann, G.~A., Saha, A., Reindl, B., Macchetto, F.~D.,

2398: \& Panagia, N.\ 2006, \apj, 653, 843

2399: 

2400: \bibitem[Sasselov et al.(1997)]{sasselov97}

2401: Sasselov, D.~D., et al.\ 1997, \aap, 324, 471 

2402: 

2403: \bibitem[Schlegel et al.(1998)Schlegel, Finkbeiner, \& Davis]{schlegel98}

2404: Schlegel, D.~J., Finkbeiner, D.~P., \& Davis, M.\ 1998, \apj, 500, 525

2405: 

2406: \bibitem[Schultz \& Wiemer(1975)]{schultz75}

2407: Schultz, G.~V., \& Wiemer, W.\ 1975, \aap, 43, 133

2408: 

2409: \bibitem[Sills et~al.(2000)Sills, Pinsonneault, \& Terndrup]{sills00}

2410: Sills, A., Pinsonneault, M.  H., \& Terndrup, D.  M.  2000, \apj, 534, 335

2411: 

2412: \bibitem[Skrutskie et al.(2006)]{skrutskie06}

2413: Skrutskie, M.~F., et al.\ 2006, \aj, 131, 1163

2414: 

2415: \bibitem[Sneden et al.(1978)]{sneden78}

2416: Sneden, C., Gehrz, R.~D., Hackwell, J.~A., York, D.~G.,

2417: \& Snow, T.~P.\ 1978, \apj, 223, 168 

2418: 

2419: \bibitem[Soszy{\'n}ski et al.(2006)]{soszynski06}

2420: Soszy{\'n}ski, I., Gieren, W., Pietrzy{\'n}ski, G., Bresolin, F.,

2421: Kudritzki, R.-P., \& Storm, J.\ 2006, \apj, 648, 375

2422: 

2423: \bibitem[Stetson(2000)]{stetson00}

2424: Stetson, P.~B.\ 2000, \pasp, 112, 925

2425: 

2426: \bibitem[Storm et al.(2004)]{storm04}

2427: Storm, J., Carney, B.~W., Gieren, W.~P., Fouqu{\'e}, P.,

2428: Latham, D.~W., \& Fry, A.~M.\ 2004, \aap, 415, 531

2429: 

2430: \bibitem[Tammann et al.(2003)Tammann, Sandage, \& Reindl]{tammann03}

2431: Tammann, G.~A., Sandage, A., \& Reindl, B.\ 2003, \aap, 404, 423

2432: 

2433: \bibitem[Thackeray et al.(1962)Thackeray, Wesselink, \& Harding]{thackeray62}

2434: Thackeray, A.~D., Wesselink, A.~J., \& Harding, G.~A.\ 1962, \mnras, 124, 445

2435: 

2436: \bibitem[Torres \& Ribas(2002)]{torres02}

2437: Torres, G., \& Ribas, I.  2002, \apj, 567, 1140

2438: 

2439: \bibitem[Trumpler(1921)]{trumpler21}

2440: Trumpler, R.~J.\ 1921, Lick Observatory Bulletin, 10, 110

2441: 

2442: \bibitem[Turner(1979)]{turner79}

2443: Turner, D.~G.\ 1979, \pasp, 91, 642 

2444: 

2445: \bibitem[Turner(1986)]{turner86}

2446: Turner, D.~G.\ 1986, \aj, 92, 111

2447: 

2448: \bibitem[Turner et al.(1992)Turner, Forbes, \& Pedreros]{turner92}

2449: Turner, D.~G., Forbes, D., \& Pedreros, M.\ 1992, \aj, 104, 1132

2450: 

2451: \bibitem[Turner et al.(1998)Turner, Pedreros, \& Walker]{turner98}

2452: Turner, D.~G., Pedreros, M.~H., \& Walker, A.~R.\ 1998, \aj, 115, 1958

2453: 

2454: \bibitem[Udalski(2000)]{udalski00}

2455: Udalski, A.\ 2000, Acta Astronomica, 50, 279

2456: 

2457: \bibitem[Udalski et al.(1998)]{udalski98}

2458: Udalski, A., Pietrzy{\'n}ski, G., Wo{\'z}niak, P., Szyma{\'n}ski, M.,

2459: Kubiak, M., \& {\.Z}ebru{\'n}, K.\ 1998, \apjl, 509, L25

2460: 

2461: \bibitem[Udalski et~al.(1999a)]{udalski99a}

2462: Udalski, A., Soszy{\'n}ski, I., Szyma{\'n}ski, M., Kubiak, M., Pietrzy{\'n}ski, G.,

2463: Wo{\'z}niak, P., \& {\.Z}ebru{\'n}, K.\ 1999a, Acta Astronomica, 49, 223 

2464: 

2465: \bibitem[Udalski et al.(1999b)]{udalski99b}

2466: Udalski, A., Szyma{\'n}ski, M., Kubiak, M., Pietrzy{\'n}ski, G., Soszy{\'n}ski, I.,

2467: Wo{\'z}niak, P., \& {\.Z}ebru{\'n}, K.\ 1999b, Acta Astronomica, 49, 201

2468: 

2469: \bibitem[van den Bergh(1957)]{vdbergh57}

2470: van den Bergh, S.\ 1957, \apj, 126, 323

2471: 

2472: \bibitem[van den Bergh \& Harris(1976)]{vdbergh76}

2473: van den Bergh, S., \& Harris, G.~L.~H.\ 1976, \apj, 208, 765

2474: 

2475: \bibitem[van Leeuwen(2005)]{vanleeuwen05a}

2476: van Leeuwen, F.\ 2005, \aap, 439, 805

2477: 

2478: \bibitem[van Leeuwen \& Fantino(2005)]{vanleeuwen05b}

2479: van Leeuwen, F., \& Fantino, E.\ 2005, \aap, 439, 791

2480: 

2481: \bibitem[van Leeuwen et al.(2007)]{vanleeuwen07}

2482: van Leeuwen, F., Feast, M.~W., Whitelock, P.~A.,

2483: \& Laney, C.~D.\ 2007, \mnras, 379, 723

2484: 

2485: \bibitem[Walker(1985)]{walker85a}

2486: Walker, A.~R.\ 1985, \mnras, 213, 889

2487: 

2488: \bibitem[Walker(1987)]{walker87}

2489: Walker, A.~R.\ 1987, South African Astronomical Observatory Circular, 11, 131

2490: 

2491: \bibitem[Walker(1999)]{walker99}

2492: Walker, A.~R.\ 1999, in Post-Hipparcos Cosmic Candles,

2493: ed. A. Heck \& F. Caputo (Dordrecht: Kluwer), 125

2494: 

2495: \bibitem[Walker \& Coulson(1985)]{walker85b}

2496: Walker, A.~R, \& Coulson, I.~M.\ 1985,

2497: South African Astronomical Observatory Circular, 9, 97

2498: 

2499: \bibitem[Wallerstein(1957)]{wallerstein57}

2500: Wallerstein, G.\ 1957, \pasp, 69, 172 

2501: 

2502: \bibitem[Wampler et al.(1961)]{wampler61}

2503: Wampler, J., Pesch, P., Hiltner, W.~A., \& Kraft, R.~P.\ 1961,

2504: \apj, 133, 895

2505: 

2506: \bibitem[Yong et al.(2006)]{yong06}

2507: Yong, D., Carney, B.~W., Teixera de Almeida, M.~L.,

2508: \& Pohl, B.~L.\ 2006, \aj, 131, 2256 

2509: 

2510: \bibitem[Zaritsky et al.(1994)Zaritsky, Kennicutt, \& Huchra]{zaritsky94}

2511: Zaritsky, D., Kennicutt, R.~C., Jr., \& Huchra, J.~P.\ 1994,

2512: \apj, 420, 87

2513: 

2514: \end{thebibliography}

2515: 

2516: \end{document}

2517: