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

2: %%\documentstyle[12pt,aasms4]{article}

3: % 

4: \newbox\grsign \setbox\grsign=\hbox{$>$} \newdimen\grdimen \grdimen=\ht\grsign

5: \newbox\laxbox \newbox\gaxbox

6: \setbox\gaxbox=\hbox{\raise.5ex\hbox{$>$}\llap

7:      {\lower.5ex\hbox{$\sim$}}}\ht1=\grdimen\dp1=0pt

8: \setbox\laxbox=\hbox{\raise.5ex\hbox{$<$}\llap

9:      {\lower.5ex\hbox{$\sim$}}}\ht2=\grdimen\dp2=0pt

10: \def\gax{\mathrel{\copy\gaxbox}}

11: \def\lax{\mathrel{\copy\laxbox}}

12: %

13: %Other definitions

14: 

15: \def\as     {\ifmmode {\rlap.}$\,$''$\,$\! \else ${\rlap.}$\,$''$\,$\!$\fi}

16: \def\degg   {\ifmmode {\rlap.}$\,$^{\circ}$\,$\! \else ${\rlap.}$\,$^{\circ}$\,$\!$\fi}

17: \def\s      {\ifmmode {\rlap.}$\,$^{s}$\,$\! \else ${\rlap.}$\,$^{s}$\,$\!$\fi}

18: \def\kms    {\ifmmode{{\rm ~km~s}^{-1}}\else{~km~s$^{-1}$}\fi}

19: \def\eg     {e.g.,~}

20: \def\ie     {i.e.,~}

21: \def\hho    {H$_2$O}

22: \def\Msun   {$M_{\odot}$}   

23: \def\Lsun   {$L_{\odot}$}

24: %  These provide a simple means of producing references in ApJ style.

25: %

26: %\def\ref#1  {\noindent \hangindent=24.0pt \hangafter=1 {#1} \par}

27: %\def\vol#1  {{\bf {#1}{\rm,}\ }}

28: %\def\etal   {{\it et~al.}}

29: \def\apj    {{ApJ}}

30: %\def\apjl   {{Ap.~J.~(Letters) }}

31: \def\apjs   {{ApJS}}

32: \def\aj     {{AJ}}

33: \def\chaap  {{\it Ch.~Astr.~Ap. }}

34: \def\aap    {{A\&A}}

35: \def\aas    {{A\&AS }}

36: \def\anrev  {{ARAA}}

37: \def\baas   {{BAAS }}

38: \def\mnras  {{MNRAS}}

39: \def\nat    {{Nature}}

40: \def\rmex   {{Rev.\ Mexicana Astr.~Ap. }}

41: \def\astmas  {{in Astrophysical Masers, ed. A. W. Clegg \& G. E. Nedoluha

42: (Berlin: Springer), 137}}

43: \def\austjp {{\it Aust.~J.~Phys. }}

44: %

45: 

46: %\slugcomment{2003 January 15}

47: 

48: %% If you wish, you may supply running head information, although

49: %% this information may be modified by the editorial offices.

50: %% The left head contains a list of authors,

51: %% usually a maximum of three (otherwise use et al.).  The right

52: %% head is a modified title of up to roughly 44 characters.  Running heads

53: %% will not print in the manuscript style.

54: 

55: \shorttitle{New Class of OH Masers}

56: \shortauthors{Argon, Reid \& Menten}

57: 

58: %% This is the end of the preamble.  Indicate the beginning of the

59: %% paper itself with \begin{document}.

60: 

61: \begin{document}

62:  

63: \title{A Class of Interstellar OH Masers Associated with Protostellar Outflows}

64: \author{Alice L. Argon and Mark J. Reid}

65: \affil{Harvard-Smithsonian Center for Astrophysics,

66: 60 Garden Street/MS42, Cambridge MA 02138}

67: 

68: \author{Karl M. Menten}

69: \affil{Max-Planck-Institut f\"ur Radioastronomie,

70: Auf dem H\"ugel, D-53121 Bonn, Germany}

71: 

72: 

73: \begin{abstract}

74:         Using the Very Large Array, we have detected weak OH maser emission 

75: near the Turner--Welch protostellar source in the W3~OH region.  

76: Unlike typical interstellar OH masers, which are associated with

77: ultra-compact HII regions, our measured positions and 

78: proper motions (from Very Long Baseline Interferometry)

79: indicate that these OH masers are associated 

80: with a bipolar outflow traced by strong \hho\ masers.  These OH

81: masers may be part of a class of interstellar OH masers that are

82: associated with very young stars which have yet to, or may

83: never, create ultra-compact HII regions.

84: This class of OH masers appears to form near the edges of very dense 

85: material (within which \hho\ masers form), where 

86: total densities drop precipitously and interstellar UV radiation is 

87: sufficient to dissociate the \hho\ molecules.

88: Observations of this class of OH masers may be an important

89: way to probe the distribution of this important molecule

90: in interstellar shocks at arcsecond resolution or better.      

91: \end{abstract}

92: 

93: \keywords{ISM: individual (W3)---ISM: jets and outflows---masers---radiation

94: mechanisms: nonthermal---radio lines: ISM---stars: formation}

95: 

96: 

97: \section{Introduction}

98: 

99:         In the Milky Way, hydroxyl masers (OH) are found associated 

100: with either evolved stars (stellar masers), regions of massive star 

101: formation (interstellar masers), or the interface between SNRs and molecular 

102: material.  Interstellar OH masers are often observed 

103: from molecular material surrounding ultra-compact HII (UCHII) regions.

104: The OH masers in the source W3~OH are some of the strongest and best studied 

105: of their kind.  In this source, the OH maser spots

106: have been mapped with VLBI observations and found projected toward 

107: an UCHII region \citep{RHB80}.  Proper motions of these

108: masers \citep{BRM92} suggest a slow, general expansion 

109: with a velocity of a 

110: few km~s$^{-1}$.

111: 

112:         Interstellar \hho\ masers are often found in the same star 

113: forming regions as OH masers.  However, frequently the \hho\ masers

114: do not appear to be {\it directly} associated with UCHII regions.

115: In fact, rarely can one detect a stellar source at the center of the 

116: \hho\ maser flow, not even at IR wavelengths, suggesting a deeply embedded 

117: and probably very young object.  Moreover, \hho\ masers are also found 

118: in regions of low-mass star formation, which are devoid of OH masers. 

119: \hho\ masers often exhibit strong outflow motions, with expansion speeds

120: $\sim30$~km~s$^{-1}$, and sometimes high velocity features moving

121: at hundreds of km~s$^{-1}$.   

122: 

123:         In this paper we present a detailed analysis of weak OH masers  

124: in the W3~OH region that do {\it not} appear to be associated with an UCHII 

125: region.  Instead they seem to form at the edges of the 

126: outflow from the Turner--Welch object \citep{TW84}, 

127: a protostellar source or cluster of sources \citep{WSW99} 

128: that drives the \hho\ masers near W3~OH

129: \citep{AMM93} and a synchrotron jet \citep{RAM95,WRM99}.  Our observations 

130: suggest that protostellar outflows may be an important mechanism for the 

131: production of weak OH masers when no ionizing stars are present.

132: 

133: 

134: \section{Observations and Results}

135: 

136: \subsection{VLA Observations and Results}

137: 

138: W3~OH was observed in the 1612.231, 1665.4018, 1667.359, and 1720.530 MHz 

139: OH transitions with the NRAO\footnote{The National Radio Astronomy Observatory 

140: (NRAO) is operated by Associated Universities, Inc., under a cooperative 

141: agreement with the National Science Foundation.}  

142: Very Large Array (VLA) on 1992 December 27 using all 27

143: antennas in the A-configuration.  Three 1.5 minute scans were obtained in

144: the 1665 transition, three 2.5 minute scans in the 1667 transition,

145: and one 2.5 minute scan in each of the remaining two transitions.  When a

146: transition was observed more than once, scans were spread over about 3 hours.

147: Weather conditions were excellent.

148: 

149: Observations were made in both left-circular (LCP) and right-circular (RCP)

150: polarizations.  These observations were part of a project to map a large 

151: sample of interstellar OH masers \citep{ARM00}.  The 

152: observing bandwidth of 0.1953 MHz was centered at an 

153: LSR velocity of $-$44.0 km s$^{-1}$.  For each band the inner 128 spectral 

154: channels (out of a total of 256) were correlated, yielding a channel 

155: separation of 0.14 km s$^{-1}$.

156: 

157: Calibration and imaging were done in the Astronomical Image Processing System 

158: (AIPS).  Channel ``0'' observations (an average of the central 75$\%$ of 

159: each bandpass) were used to calibrate the complex instrumental 

160: gain for each antenna.  We employed the standard method of boot-strapping the 

161: flux density of a (variable) secondary calibrator, B0212$+$735, from

162: the flux density of the primary calibrator, 3C~286.  We then determined 

163: complex gains for the secondary calibrator using the AIPS task CALIB, 

164: interpolated those gains to the W3~OH observation times with the task CLCAL, 

165: and applied them to the full spectral-line database with the tasks TACOP and 

166: SPLIT.  Observations of the secondary calibrator were made immediately before 

167: and after each W3~OH scan to ensure accurate interpolation.

168: 

169: To remove the effects of any residual atmospheric phase corruption, we

170: ``self-calibrated'' the data.  This was accomplished by selecting a strong

171: unresolved feature in one of the polarizations for each transition (called the

172: ``reference feature''), shifting the interferometer phase center to the

173: position of that feature, determining residual phases using a point source, 

174: and subtracting those residual phases from the

175: data in all spectral channels in both polarizations.

176: The reference features chosen were the $v_{LSR}=-48.9$~\kms\ and 

177: $-44.4$~\kms\ channels at 1665 and 1667 MHz, respectively, both in 

178: LCP.  

179: 

180: A comparison of the positions of the strong OH masers in our study 

181: \citep{ARM00} and another study (\citet{BRM92} 

182: and references therein), taking into account the different map center 

183: locations, suggests that our absolute position error for the 1665 MHz 

184: transition is $\approx0\as1$ in each direction.  Self-calibration was followed 

185: by imaging, whereby each spectral channel was mapped and CLEANed.  

186: We searched an area of $128'' \times 128''$ ($512 \times 512$ pixels). 

187: Note that because our sources can be 100\% circularly polarized, we divided 

188: the maps produced by MX by a factor of two so that Stokes I is obtained by 

189: summing (rather than averaging) the RCP and LCP maps.

190: 

191: After imaging, selected peaks were fit with a two-dimensional Gaussian 

192: brightness distribution to determine their flux density, size, and 

193: offset from the reference feature.

194: While strong OH maser emission for the UCHII region has been mapped before,  

195: we searched for emission beyond the innermost $5'' \times 5''$.  

196: Spurious maser features were rejected by demanding that the

197: following four constraints be met:

198: 

199: \begin{itemize}

200: \item[] 1) {Peaks had to have a flux density of at least 4.5 times the channel 

201: rms noise.  Typical channel rms noise levels were 28 mJy for the 

202: 1665.4018 MHz transition and 26 mJy for the 1667.359 MHz transition.}

203: \item[] 2) {Peaks had to have a flux density of at least 1.3 times the absolute 

204: value of the largest negative peak in the channel map.}

205: \item[] 3) {Peaks could not be any weaker than 1/150 of the peak in the map

206: owing to dynamic range limitations, which are particularly severe at 

207: 1665 MHz because of the strong OH emission from the UCHII region

208: between $-43<v_{LSR}<-47$ ~km s$^{-1}$.  Two ``peaks'' appear in the spectrum 

209: shown in Figure 1 that do not satisfy this criterion.  They are artifacts of 

210: a 132 Jy ($-$46.3 km s$^{-1}$, LCP) and a 201 Jy ($-$45.0 km s$^{-1}$, RCP) 

211: feature.}

212: \item[] 4) {Peaks had to persist over three or more adjacent channels 

213: to within a synthesized beam width.}

214: \end{itemize}

215: 

216: \begin{figure}

217: \plotone{f1.eps}

218: \figcaption[]{

219: Spectra of two ``outlying'' OH maser features ($-42.4$~\kms, LCP at 1665 MHz 

220: and $-49.3$~\kms, LCP at 1667 MHz) at the positions of their maximum 

221: emission.  Heavy lines indicate RCP and light lines LCP emission.  The 

222: ``features'' at v$_{LSR}$ $-46.3$~\kms\ (LCP) and $-45.0$~\kms\ (RCP) in the 

223: 1665 MHz transition are artifacts generated by spatial sidelobes of the 

224: strongest 1665 LCP (132 Jy) and RCP (201 Jy) features, respectively, which are 

225: located toward the UCHII region.  The outlying OH features are {\it not} 

226: artifacts since there is no strong LCP emission at v$_{LSR}=-42.4$~\kms\ in 

227: 1665 or at $-49.3$~\kms\ in 1667 elsewhere in the map.  A $4\sigma$ RCP 

228: detection at $-41.3$~\kms\ in the 1665 MHz transition may be part of a Zeeman 

229: pair, the other member being the $-42.4$~\kms\ feature. 

230: \label{fig1} 

231: }

232: \end{figure}

233: 

234: \begin{deluxetable}{lccccc}

235: \footnotesize

236: \tablecaption{VLA Observations of OH Maser Emission$^a$ \label{}}

237: \tablenum{1}

238: \tablewidth{0pt}

239: \tablehead{

240: \colhead{Transition (Pol.)}& \colhead{ $v_{\rm LSR}$$^b$ }& 

241: \colhead{Peak Flux Density$^b$}& \colhead{$\Delta v^b$}& 

242: \colhead{$\Delta\Theta_x^c$}& \colhead{$\Delta\Theta_y^c$}\nl

243: & \colhead{ (km s$^{-1}$) }& \colhead{(Jy)}& 

244: \colhead{ (km s$^{-1}$) }& \colhead{$('')$}& \colhead{$('')$}

245: }

246: 

247: \startdata

248: 

249: 1665.4018            (LCP)& $-42.4$& $0.25\pm0.03$& $0.29$& $4.17\pm0.06$& 

250: $-0.09\pm0.06$\nl

251: 1667.359\phantom{8}  (LCP)& $-49.3$& $0.20\pm0.03$& $0.37$& $6.81\pm0.07$& 

252: $-0.06\pm0.07$\nl

253: 

254: \tablecomments{

255: $^a$~~Maser features $>2\as5$ from center of the UCHII region in W3~OH.

256: \noindent

257: $^b$~~The quantities $v_{\rm LSR}$, $\Delta v$ (FWHM), and peak flux density 

258: were determined by fitting a Gaussian profile to the three highest spectral 

259: points.

260: \noindent  

261: $^c$~~Positions are offsets eastward ($\Delta\Theta_x$) and northward 

262: ($\Delta\Theta_y$) from the map center of Fig.~2, located at 

263: $\alpha_{B1950} = 02^h 23^m 16\s499, \delta_{B1950} = 61^{\circ} 38' 

264: 57\as21$, and are weighted averages of the contributing channels.  

265: Specifically, the position offsets are relative to the 

266: 1665 MHz LCP reference feature at $-46.3$~\kms\ 

267: and then shifted by adding ($-0\as706,-0\as210$)  

268: to the relative maser positions to register them with the map in Fig.~2.

269: }

270: 

271: \enddata

272: \end{deluxetable}

273: 

274: \begin{deluxetable}{lccccc}

275: \footnotesize

276: \tablecaption{VLBI Observations of 1665 MHz LCP OH Maser Emission \label{}}

277: \tablenum{2}

278: \tablewidth{0pt}

279: \tablehead{

280: \colhead{Epoch}& \colhead{ $v_{\rm LSR}$ }& \colhead{Peak Flux Density}& 

281: \colhead{$\Delta v$}& \colhead{$\Delta\Theta_x$}& \colhead{$\Delta\Theta_y$}\nl

282: & \colhead{ (km s$^{-1}$) }& \colhead{(Jy)}& 

283: \colhead{ (km s$^{-1}$) }& \colhead{$('')$}& \colhead{$('')$}

284: }	

285: 

286: \startdata

287: 

288: 1978& $-42.9$& $0.75\pm0.10$& $0.26$& $4.222\pm0.001$& $-0.052\pm0.001$\nl

289: 1986& $-42.4$& $0.17\pm0.04$& $0.29$& $4.199\pm0.001$& $-0.094\pm0.001$\nl

290: 

291: \tablecomments{The positional errors are {\it formal} uncertainties.  See 

292: text for details and the Table 1 caption for other information.}

293: 

294: \enddata

295: \end{deluxetable}

296: 

297: We found two spots of OH maser emission satisfying all four criteria, one 

298: at 1665 and the other at 1667 MHz (see Table 1).  Since the expected 

299: uncertainty for a point source is approximately half the synthesized 

300: beam ($1''$) divided by the signal-to-noise ratio, the relative position 

301: error in each coordinate is $\approx0\as06$ and $\approx0\as07$ for the 

302: 1665 and 1667 MHz transitions, respectively.  Spectra of the two features 

303: at the pixel of their maximum flux density are shown in Figure 1.  Both

304: features are $\gax8\sigma$ detections and were detected by Gaume \& Mutel

305: (1987), who, while noting that they were in the vicinity of the water masers

306: and HCN emission, did not discuss their significance.  Our $-42.4$ and $-49.3$ 

307: km s$^{-1}$ features were reported at $-42.8$ and $-48.3$ km s$^{-1}$, 

308: respectively, by Gaume \& Mutel, who list velocities of the peak channel 

309: emission in 1.1 km s$^{-1}$ wide channels.  Thus, the agreement in both 

310: position and velocity for these two measurements is reasonable.  Gaume \&  

311: Mutel also observed one additional spot of weak emission (0.1 Jy, a $3\sigma$ 

312: detection) {\it outside of} the central $5''$ x $5''$, that we fail to confirm.

313: Our $1\sigma$ noise level was 0.026 Jy in this transition, so if the maser's

314: flux density was constant (or increasing) in time between their observations

315: in 1985 and ours in 1992, we would have a $4\sigma$ detection.  Either Gaume

316: \& Mutel's detection is spurious, which is not highly unlikely in a

317: $\sim 10''$ x $10''$ field (containing about 100 independent beams), or the

318: feature weakened over the years.

319: 

320: \subsection{VLBI Data and Results}

321: 

322: In an attempt to confirm the 1665 MHz VLA detection and obtain a proper 

323: motion, we re-mapped the data from two epochs of VLBI observations: one 

324: in 1978 by \citet{GBR88} and the other in 1986 by \citet{BRM92}.  

325: (No archival VLBI data for the 1667.359 transition were available).  

326: At both epochs, RCP and LCP were observed and there 

327: were 96 spectral channels per IF, giving channel separations of 0.12 km 

328: s$^{-1}$.  Observations were centered at an LSR velocity of $-$46.0 km 

329: s$^{-1}$.  At both epochs, we searched for emission within

330: ($\pm 0\as1$) of the positions in Table 1.  Since the features in W3~OH are weak

331: and heavily resolved on long baselines,

332: we used a (u--v)-taper of 20~M$\lambda$.  The resulting 

333: rms noise levels in the channel maps were about 100 mJy and 35 mJy for 1978 

334: and 1986 observations, respectively.  

335: 

336: The VLBI results are given in Table 2 and displayed in Figure 2.  The 1665 MHz 

337: LCP feature near 

338: $v_{LSR}=-42.4$~\kms\ in our VLA data was found to be present in the 1978 

339: and 1986 VLBI maps, albeit at a slightly different velocity ($-42.9$~\kms) in 

340: 1978.  Note that \citet{GM87} report a velocity of $-42.8$~\kms\ for 

341: this feature.  We registered the maps at the three epochs (two VLBI, one VLA) 

342: by using the strongest 1665 MHz LCP feature (132 Jy at $-46.3$~\kms\ in the 

343: VLA data), which is located at $\alpha_{B1950} = 02^h 23^m 16\s322~\pm~0\s01, 

344: \delta_{B1950} = 61^{\circ} 38' 57\as66~\pm~0\as1$.  

345: The VLBI relative position errors 

346: in Table 2 are $\approx$~1 mas.  Assuming these features trace a single cloud 

347: of gas, we find a proper motion of about $-32$~\kms\ eastward and $-58$~\kms\ 

348: northward, with a {\it formal uncertainty} of about $2$~\kms\ in each 

349: direction.  A distance of 2.2 kpc was assumed.  Because the line center 

350: velocity of the feature changes by $\approx0.5$~\kms\ between the two epochs, 

351: a realistic proper motion uncertainty should include the possibility of 

352: structural changes in the feature.  The span of position offsets over the 

353: three or more adjacent channels that comprise a feature is one way to estimate 

354: the magnitude of possible structural changes.  A typical span was found to be 

355: $\approx2$~mas, which gives rise to a more {\it realistic proper motion 

356: uncertainty} of $\approx4$~\kms\ in each direction.  Only the two VLBI epochs 

357: were used to estimate the motion since their positional accuracies are far 

358: better than the positional accuracy of the VLA epoch.  The VLA position, 

359: however, is consistent with the derived VLBI motion and our OH motion is 

360: consistent with the bipolar outflow seen in the \hho\ masers \citep{AMM93}.  

361: 

362: Why does the Doppler velocity change over the eight years between our two

363: VLBI epochs?  The possibility that structural changes have occurred in the

364: feature was mentioned above.  A second possibility is that these masers, unlike

365: typical HII region masers, are likely to occur in a strongly accelerated

366: region, the region between the terminus of the H$_2$O outflow and the dense

367: ambient molecular cloud.  If this is the case, it would not be unreasonable 

368: to find a change of 0.5 km s$^{-1}$ in Doppler velocity between the two epochs.

369: 

370: %\citet{BRM92} note a flux density calibration uncertainty of about 

371: Bloemhof, Reid \& Moran (1992) note a flux density calibration uncertainty of 

372: about 20$\%$ in the VLBI data.   Since this can not account for the 

373: difference among the three epochs, we conclude that 

374: this feature is variable.  

375: 

376: \section{Discussion}

377: 

378: \subsection {OH Masers Associated with the TW protostar}

379: 

380: The OH maser features described in this paper are not associated with the 

381: UCHII region in W3~OH, where strong OH and class II CH$_3$OH 

382: maser emission is found \citep{RHB80,BRM92,M91,MMW99}.  Instead,

383: as seen in Fig.~2 their positions, and the 1665 MHz feature proper motion, 

384: suggest an association with the TW object and its intense H$_2$O masers 

385: \citep{AMM93}.  The association of some OH masers in W3~OH with the TW object 

386: and its H$_2$O masers suggests that some interstellar OH 

387: masers that are offset from HII regions may be associated with protostellar 

388: outflows and very young stars which have yet to, or may never, ionize their 

389: surroundings.

390: %Thus,the masers discussed in this paper may be part of a newly recognized

391: %class of interstellar OH masers.

392: 

393: The TW object is thought to be a newly formed or forming massive star 

394: \citep{TW84,WRM99} or stars \citep{WSW99},

395: which have yet to ionize their placental material.  

396: The ``hot core'' nature of the TW object was established by 

397: observations of high excitation NH$_3$~(5,5) inversion lines, which yield

398: an  NH$_3$ rotation temperature of 180~K \citep{MWW86}. 

399: While the TW object shows no detectable thermal bremsstrahlung, 

400: surprisingly, it is a source of synchrotron emission \citep{RAM95}. 

401: When imaged at high resolution ($0\as2$) and sensitivity (10~$\mu$Jy rms) 

402: with the VLA, the synchrotron source appears as an elongated, wiggling, 

403: double-sided jet \citep{WRM99}.  

404: The predominantly east--west orientation of the synchrotron jet 

405: matches both the spatial distribution and proper motions of the H$_2$O masers.

406: Also, the center of the synchrotron jet lies, within measurement 

407: uncertainties of $0\as1$, at the center of expansion of the

408: H$_2$O masers \citep{AMM93}.

409: 

410: 

411: Comparison of the positions and velocities of the TW source 

412: OH masers with the dust continuum and molecular line emission mapped by 

413: \citet{WHS97} and \citet{WSW99} also indicates a very close correspondence.

414: Both the 1665 and the 1667 MHz masers project on the edges of

415: the dust continuum/molecular emission.  The thermally excited molecular 

416: gas displays an apparent shift in velocity across

417: the source, with $v_{LSR}$ changing from about $-52$~km s$^{-1}$

418: at the eastern edge to about $-45$~km s$^{-1}$ at the western edge.

419: This trend of velocity increasing from east to west 

420: is consistent with our OH maser velocities.

421: 

422: The arrangement of the OH masers at, or just beyond, the ends of the 

423: H$_2$O maser outflow is noteworthy.   One possible explanation for this 

424: arrangement is that the H$_2$O molecules in the outflow are dissociated at the 

425: extremities of the source, producing OH molecules with sufficient density to 

426: yield detectable maser emission.  H$_2$O masers are estimated to require H$_2$ 

427: densities between $10^8$ and $10^{10}$ cm$^{-3}$, whereas OH masers require 

428: densities between $10^5$ and $10^{7}$ cm$^{-3}$ (\citet{E92} and references 

429: therein).  The upper limits to the H$_2$ densities are established by 

430: collisional thermalization of the level populations and quenching of the 

431: population inversions.  The lower limits arise from estimates of the minimum 

432: column density necessary to achieve strong maser amplification over reasonable 

433: path lengths.  Therefore, the transition from H$_2$O to OH masing in a source

434: like the TW object suggests a rapid decrease in density with

435: distance and a sharp boundary for the elongated region of H$_2$O maser

436: activity.  

437: 

438: Confirmation of a sharp molecular boundary comes from mm-wavelength

439: observations of dust continuum emission.

440: The molecular column density inside of the OH masers in the TW object, 

441: inferred from the dust continuum, of $N_{H_2} \sim 3\times 10^{24}$~cm$^{-2}$

442: indicates a volume density of $n_{H_2} \sim 10^8$~cm$^{-3}$, assuming

443: a line of sight path length of 0.01~pc (P. Schilke, pers. comm.).

444: The relatively sharp boundary of the dust continuum emission suggests

445: that the molecular density must drop by a factor of 10 or more

446: within a projected distance of about 0.005 pc at the edge of the source.  

447: Possibly, when molecular densities fall below $\sim10^7$~cm$^{-3}$, 

448: interstellar ultraviolet photons can dissociate H$_2$O molecules and 

449: create OH radicals.

450: 

451: \subsection {Comparison to Other Regions: Shock Chemistry}

452: 

453: While most interstellar OH masers do not show evidence of strong

454: outflows, there is conclusive evidence that \hho\

455: masers in regions of massive star formation partake in strong outflows 

456: from both 1) the large radial velocities observed, exceeding 

457: $\pm100$ \kms\ in Orion-KL \citep{G81a}, and 2) the proper

458: motions measured there and in W51 Main \citep{G81b}, 

459: Sgr B2 \citep{RSM88}, W49 N \citep{GMR92}.  Of course 

460: the \hho\ masers in W3 OH-TW also show a clear bipolar flow 

461: \citep{AMM93}, and the $\sim50$~\kms\ proper motion of one of the

462: OH masers from the TW object suggests that at least some of the weak OH 

463: masers in the new class introduced by \citet{GM87} 

464: are also associated with strong outflows.

465: 

466: The shock chemistry calculations of \citet{ND89} show that behind 

467: the front of a dissociative (``$J$'') shock of velocity 80 \kms, OH 

468: re-forms at a high abundance approaching $[{\rm OH/H}] \approx 10^{-5}$,

469: while the \hho\ abundance is about an order of magnitude lower.

470: In contrast, behind a non-dissociative (``$C$'') shock almost all the oxygen

471: becomes tied up in \hho\ (e.g., Draine, Roberge, \&\ Dalgarno 1983).

472: The velocities in most \hho-emitting outflows are lower than 50 \kms,

473: indicating non-dissociative shocks (in whose aftermath conditions

474: are very conducive for pumping \hho\ masers; \citet{E92}). This

475: might explain why the \hho\ masers are always much more powerful 

476: (by up to more than 4 orders of magnitude) than the OH masers in 

477: W3~OH-TW (and Orion-KL as discussed below).

478: Alternatively, as mentioned above, the densities necessary

479: for 22 GHz \hho\ maser action are two orders of magnitude higher

480: than the values usually invoked for OH masers, and the

481: OH maser emission should be quenched at these densities.

482: 

483: The masers associated with the TW object and those associated with

484: source-I in the Orion-KL region (Menten \&\ Reid 1995) have definite similarities.

485: Both are associated with strong outflows.  They

486: have weak (or undetectable) thermal continuum emission

487: at radio wavelengths and no clear, direct IR detections. 

488: Also both have strong \hho\ and weaker OH masers.

489: If the OH masers in Orion-KL and W3~OH-TW have a similar

490: origin, then why is it that toward Orion-KL \citet{JMN89} find 

491: many dozen features in the 1665 MHz (and about 2 dozen in the 1612 MHz) 

492: OH line, while we detect only one each in the 1665 and 1667 MHz lines      

493: (and none in the 1612 MHz line) toward W3~OH-TW?

494: Given our $(4\sigma)$ detection limit of roughly 0.1~Jy, we would

495: detect a 2.5 Jy strong feature in Orion put at the roughly

496: 5 times larger distance of W3OH. Thus, we would detect 25 of the Johnston

497: et al.~1665 MHz features, many of them barely, and one

498: of their 1612 MHz features. Although no strong correlation between

499: \hho\ and OH luminosity seems to have been established yet,

500: we note that the maser luminosity in both species is about an order of 

501: magnitude greater in Orion-KL than W3 OH-TW, as is the case for the 

502: infrared luminosity in both regions. This makes the paucity of OH 

503: features observed toward W3 OH-TW compared to Orion-KL understandable.

504: 

505: \subsection {OH Maser/UCHII Region Association}

506: 

507: In regions of massive star formation, one often finds interstellar OH 

508: masers.  The traditional view is that interstellar hydroxyl masers

509: (and class II methanol masers) form in the slowly 

510: expanding (3 to 10 \kms), dense ($\lax 10^7$~cm$^{-3}$), 

511: and warm ($\approx150$K) molecular layers driven by the

512: expanding UCHII region (\eg \citet{EJ78,CW91,M97}), where the OH is likely

513: formed by photo-dissociation of \hho.  

514: Since the strongest OH sources are generally associated with UCHII 

515: regions, most researchers have concluded that OH masers are probably 

516: associated with massive OB-type (ionizing) stars.   

517: However, it may well be that the ``traditional'' OH--UCHII region association

518: picture emerged due to a selection effect, caused by the fact that

519: earlier searches for OH masers were conducted predominantly 

520: toward HII regions. 

521: At least three surveys suggest that this is indeed the case.

522: 

523: In \citet{ARM00}, we selected for observation interstellar OH 

524: masers whose declinations were above $-45^{\circ}$ and peak flux densities 

525: were stronger than 1 Jy in both circular polarizations in at least one OH 

526: main-line transition.  We observed 91 OH masers and noted that our sample was  

527: at least $80\%$ complete.  Towards each of the observed OH maser regions, we  

528: also looked for 8.4 GHz continuum emission detectable above a noise level 

529: of $\approx 0.15$ mJy~beam$^{-1}$.    The Maser--UCHII region association was 

530: judged by analogy to W3~OH.   In W3~OH, the OH masers closer than $3''$ 

531: to the center of the brightest 8.4 GHz continuum emission ($\approx 2''$ 

532: from the edge) appear to be

533: associated with the UCHII region, while the OH masers farther away 

534: (i.e., those we have associated with the TW source) are not.   At a 

535: distance of 2.2~kpc (Humphreys 1978) 

536: to W3~OH, $2''$ corresponds to 0.02 pc.   We then examined all sources

537: in the \citet{ARM00} survey and found that in only 49 of the 91 maser fields 

538: %($54\%$) were {\it all} of the OH masers closer than 0.03 pc to an 

539: ($54\%$) were {\it any} of the OH masers closer than 0.02 pc to the edge of an 

540: UCHII region. 

541: 

542: In another recent survey, \citet{FC00} presented a search for 8.5 GHz radio 

543: continuum emission toward 45 sites of OH and \hho\ maser emission with 

544: Galactic longitudes in the range $352\degg52$ to $24\degg79$.  Only toward

545: 17 of them did they find continuum emission above a detection threshold of 

546: approximately 0.5 mJy~beam$^{-1}$.  (At this frequency, according to 

547: \citet{FC00}, an ionization-bounded, optically thin, HII region produced by a 

548: ZAMS star of spectral type B1 at a distance of 5 kpc is 1 mJy and would have 

549: been detected.) 

550: %{\bf (I haven't yet 

551: %checked this number) Do!! It's taken from FC.})

552: If one considers only the OH maser regions and uses the maser-continuum 

553: association criterion described above, one finds that only 16 of the 26 

554: ($62\%$) regions that contain OH masers  have associated continuum emission.  

555: 

556: Finally, \citet{GM87} studied the ground state OH maser and 15 GHz 

557: continuum emission associated with 11 regions of star formation, demanding only

558: that the 1720 MHz OH transition be present when selecting regions for study.   

559: %They noted that $20-25\%$ of the maser clusters within their 11 regions 

560: %did not appear to be closely associated with HII emission.  

561: They noted that $75-80\%$ of the maser clusters within their 11 regions 

562: appeared to be closely associated with HII emission.  Detection limits 

563: (rms noise levels) varied widely from source to source but typically were in

564: the range of a 0.3 mJy~beam$^{-1}$ to 3  mJy~beam$^{-1}$.  

565: Although this percentage is not directly comparable to the percentages 

566: from the two previous surveys, since we did not attempt to subdivide maser 

567: regions into maser clumps when discussing those surveys, it does strengthen 

568: the argument that many OH masers have no associated continuum.  

569: Gaume \& Mutel (1987) grouped these unassociated masers into a separate 

570: ``class'' and noted that they tended to be much weaker than other OH masers.  

571: 

572: How does one explain the presence of OH masers with no associated continuum?  

573: At least two possibilities come to mind.  First, the masers may be associated 

574: with a very young HII region that is currently below the detection limit

575: because of its small size.  While one might expect this phase in the

576: development of an UCHII region to be very brief, \citet{Keto2002}

577: show that the strong gravitational attraction of material very close to

578: a massive star can significantly extend the lifetime of this phase.  

579: In the future, these HII regions can be expected to grow larger and become 

580: detectable.  An alternative possibility for the absence of an UCHII region 

581: near some OH masers is that these masers may be associated with a star that 

582: will never give rise to an HII region.  The TW source and other outflow 

583: regions, for example, may contain a central star less massive than a 

584: B3 star and not produce sufficient ionizing photons to create a detectable

585: UCHII region.

586: 

587: It is uncertain how many of the OH masers with no detected continuum emission 

588: are similar to the TW object masers.  In general, OH masers associated 

589: with outflows can be expected to have weak features spanning tens of 

590: \kms\ in radial velocity.  In the survey of \citet{ARM00}

591: $\approx 20$\% of interstellar OH maser spectra have velocity spreads 

592: $\gax10$~\kms\ and $\approx 5$\% have velocity spreads $\approx30$~\kms, 

593: the latter being inconsistent with the expansion velocity of a compact 

594: HII region.  Unfortunately, it is possible that 

595: most surveys could have missed weak, high velocity, maser features

596: owing to bandwidth and sensitivity limitations.  

597: 

598: While the spread in radial velocity of the OH masers toward W3~OH is 

599: not large ($\approx9$~\kms), the observed proper motion of the 1665 MHz maser 

600: toward the TW source is very large, implying that the outflow in this 

601: source lies close to the plane of the sky.  This is corroborated by the 

602: proper motions measured for the TW \hho\ masers by \citet{AMM93}.  

603: This may not be a coincidence, since there may be a preferential 

604: orientation for observing detectable masers in a bipolar outflow.

605: For outflows, the longest gain paths are likely to be perpendicular 

606: to the outflow axis.   Thus, these masers may be strong only when the 

607: outflow axis is nearly perpendicular to our line of sight, and they

608: would show small Doppler shifts from the systemic velocity.

609: If this effect is important, then a significant fraction of the OH 

610: masers with no detected continuum emission could be associated with outflows. 

611: 

612: 

613: %Since the youngest (hyper-compact) HII regions

614: %may not yet have expanded enough to produce easily observable radio emission,

615: %the lack of detected continuum emission might only indicate extreme youth.

616: %\citet{K02} shows that strong gravitation forces can retard the growth of

617: %a hyper-compact HII region and greatly extend such a youthful phase.

618: 

619: 

620: \begin{figure}

621: \epsscale{0.89}

622: \plotone{f2.eps}

623: \figcaption[]{ 

624: {\it (Top panel)} 1665 and 1667 MHz OH masers from \citet{ARM00}

625: and two ``outlying'' OH masers (this paper) superimposed on the 8.4 GHz VLA 

626: contour plot of Wilner et al. (1999).  The ``stars'' represent 1665 MHz OH 

627: maser emission and the ``squares'' 1667 MHz OH maser emission.  Open symbols 

628: are RCP and closed symbols LCP.  `TW' stands for the Turner-Welch object 

629: (Turner \& Welch 1984). {\it (Lower panel)} H$_2$O and outlying OH maser 

630: emission near the Turner Welch object.  Positions and proper motions of the 

631: H$_2$O masers are indicated by filled circles and light arrows, 

632: respectively (Alcolea {\it 

633: et al.} 1993) and the large cross marks the \hho\ maser center of 

634: expansion determined by these authors. Positions of the OH masers are

635: indicated by either a ``star'' or a 

636: ``square'' (this paper) and the single proper motion by a heavy 

637: arrow. The origin of coordinates corresponds to 

638: $\alpha_{B1950} = 02^h 23^m 16\s499, \delta_{B1950} = 61^{\circ} 38' 57\as21$. 

639: \label{fig2} }

640: \end{figure}

641: 

642: 

643: 

644: \section{Conclusions}

645: During an interferometric survey of OH masers in star-forming regions, we 

646: detected weak OH maser emission associated with the Turner-Welch object,

647: a very deeply embedded protostellar object. 

648: Using archival VLBI data, we were able to determine the proper motion of 

649: one of the maser features, which is of comparable magnitude and 

650: direction as the proper motions of the \hho\ masers,

651: which arise from a bipolar outflow centered on the TW object.

652: While conventional wisdom associates OH masers with the slowly expanding 

653: molecular envelopes of ultracompact HII regions, we argue that the W3 OH/TW OH

654: masers are part of a new class of OH masers that trace protostellar outflows.

655: Another example of this class of OH masers might be the 

656: Orion-KL/IRc~2/Source-I masers and, possibly, some of the sources mapped 

657: by \citet{FC00} in which OH and \hho\ masers are coexistent and are 

658: offset from continuum emission.

659:  

660: The near spatial coincidence of OH and \hho\ maser emission in outflow

661: sources may be explained by OH production by photo-dissociation of \hho\

662: by the interstellar radiation field.

663: The densities required for \hho\

664: maser action are roughly two orders of magnitude higher than those suitable 

665: for OH masers.  This is consistent with the observation that the OH maser

666: spots are located at the outer edges of the \hho\ emission region.

667: One important avenue of future research suggested by our results

668: is high spatial resolution observations of this class of OH masers, 

669: since OH is an important chemical constituent of post-shock gas not 

670: otherwise observable with comparable resolution.

671: 

672: 

673: \begin{thebibliography}{}

674: %\bibitem[Alcolea, Menten, Moran, \& Reid(1993)]{AMM93} 

675: %   Alcolea, J., Menten, K.~M., Moran, J.~M., \& Reid, M.~J.\ 1993, 

676: %   Astrophysical masers; Proceedings of the Conference, Arlington, VA, 

677: %   Mar.~9-11, 1992 (A93-52776 23-90), p.~225-228., 225 

678: \bibitem[Alcolea, Menten, Moran, \& Reid(1993)]{AMM93} 

679:    Alcolea, J., Menten, K.~M., Moran, J.~M., \& Reid, M.~J.\ 1993, 

680:    in Astrophysical Masers, ed. A.~W. Clegg \& G.~E. Nedoluha 

681:    (Springer Verlag: Berlin), 225

682: \bibitem[Argon, Reid, \& Menten(2000)]{ARM00} Argon, A.~L., 

683:    Reid, M.~J., \& Menten, K.~M.\ 2000, \apjs, 129, 159 

684: %\bibitem[Becklin(1997)]{B97} Becklin, E. E. 1997, in Proc. ESA Symp. 

685: %   "The Far Infrared And Submillimetre Universe", ed. A. Wilson, (Nordwijk: 

686: %    ESA SP-401) p. 201

687: %\bibitem[Beuther et al.(2002)]{BSM02} Beuther, H., Schilke, 

688: %   P., Menten, K.~M., Motte, F., Sridharan, T.~K., \& Wyrowski, F.\ 2002, 

689: %   \apj, 566, 945 

690: \bibitem[Bloemhof, Reid, \& Moran(1992)]{BRM92} Bloemhof, 

691:    E.~E., Reid, M.~J., \& Moran, J.~M.\ 1992, \apj, 397, 500 

692: \bibitem[Cesaroni \& Walmsley(1991)]{CW91} Cesaroni, R.~\& 

693:    Walmsley, C.~M.\ 1991, \aap, 241, 537 

694: %\bibitem[Cesaroni, Felli, \& Walmsley(1999)]{CFW99} Cesaroni, 

695: %   R., Felli, M., \& Walmsley, C.~M.\ 1999, \aaps, 136, 333 

696: \bibitem[Draine, Roberge, \& Dalgarno(1983)]{DRD83} Draine, 

697:    B.~T., Roberge, W.~G., \& Dalgarno, A.\ 1983, \apj, 264, 485 

698: %\bibitem[Dreher \& Welch(1981)]{DW81} Dreher, J.~W.~\& 

699: %   Welch, W.~J.\ 1981, \apj, 245, 857 

700: \bibitem[Elitzur(1992)]{E92} Elitzur, M.\ 1992, \araa, 30, 75 

701: \bibitem[Elitzur \& de Jong(1978)]{EJ78} Elitzur, M.~\& de 

702:    Jong, T.\ 1978, \aap, 67, 323 

703: %\bibitem[Elitzur, Hollenbach, \& McKee(1989)]{EHM89} Elitzur, 

704: %   M., Hollenbach, D.~J., \& McKee, C.~F.\ 1989, \apj, 346, 983 

705: %\bibitem[Fiebig \& Guesten(1989)]{FG89} Fiebig, D.~\& 

706: %   Guesten, R.\ 1989, \aap, 214, 333 

707: \bibitem[Forster \& Caswell(2000)]{FC00} Forster, J.~R.~\& 

708:    Caswell, J.~L.\ 2000, \apj, 530, 371 

709: %\bibitem[Garay, Moran, \& Reid(1987)]{GMR87} Garay, G., 

710: %   Moran, J.~M., \& Reid, M.~J.\ 1987, \apj, 314, 535 

711: \bibitem[Garcia-Barreto et al.(1988)]{GBR88} Garcia-Barreto, 

712:    J.~A., Burke, B.~F., Reid, M.~J., Moran, J.~M., Haschick, A.~D., \& 

713:    Schilizzi, R.~T.\ 1988, \apj, 326, 954 

714: \bibitem[Gaume \& Mutel(1987)]{GM87} Gaume, R.~A.~\& Mutel, 

715:    R.~L.\ 1987, \apjs, 65, 193 

716: \bibitem[Genzel, Reid, Moran, \& Downes(1981a)]{G81a} Genzel, 

717:    R., Reid, M.~J., Moran, J.~M., \& Downes, D.\ 1981, \apj, 244, 884 

718: \bibitem[Genzel et al.(1981b)]{G81b} Genzel, R.~et al.\ 1981,

719:    \apj, 247, 1039

720: %\bibitem[Greenhill et al.(1998)]{GGS98} Greenhill, L.~J., 

721: %   Gwinn, C.~R., Schwartz, C., Moran, J.~M., \& Diamond, P.~J.\ 1998, \nat, 

722: %   396, 650 

723: \bibitem[Gwinn, Moran, \& Reid(1992)]{GMR92} Gwinn, C.~R., 

724:    Moran, J.~M., \& Reid, M.~J.\ 1992, \apj, 393, 149 

725: %\bibitem[Hansen et al.(1977)]{HMR77} Hansen, S.~S., Moran, 

726: %   J.~M., Reid, M.~J., Johnston, K.~J., Spencer, J.~H., \& Walker, R.~C.\ 

727: %   1977, \apjl, 218, L65 

728: %\bibitem[Hartquist, Menten, Lepp, \& Dalgarno(1995)]{HML95} 

729: %   Hartquist, T.~W., Menten, K.~M., Lepp, S., \& Dalgarno, A.\ 1995, \mnras, 

730: %   272, 184

731: \bibitem[Humphreys(1978)]{H78} Humphreys, R.~M.\ 1978, \apjs, 38, 309 

732: \bibitem[Johnston, Migenes, \& Norris(1989)]{JMN89} Johnston, 

733:    K.~J., Migenes, V., \& Norris, R.~P.\ 1989, \apj, 341, 847 

734: %\bibitem[Keto(2002)]{K02} Keto, E.\ 2002, \apj, 

735: %   580, 980 

736: \bibitem[Keto(2002)]{Keto2002} Keto, E.\ 2002, \apj, 580, 980 

737: \bibitem[Mauersberger, Wilson, \& Walmsley(1986)]{MWW86} 

738:    Mauersberger, R., Wilson, T.~L., \& Walmsley, C.~M.\ 1986, \aap, 166, L26 

739: \bibitem[Menten(1991)]{M91} Menten, K.~M.\ 1991, in ASP Conf. Proc. 16, 

740:    Atoms, Ions and Molecules: New Results in Spectral Line Astrophysics, 

741:    ed. A.~D. Haschick \& P.~T.~P. Ho (ASP: San Francisco), 119 

742: \bibitem[Menten(1997)]{M97} Menten, K.~M.\ 1997, 

743:    in IAU Symp. 178, Molecules in Astrophysics: Probes \&\ Processes, 

744:    ed. E. van Dishoeck (Kluwer: Dordrecht), 163 

745: %\bibitem[Menten et al.(1992)]{MRP92} Menten, K.~M., Reid, 

746: %   M.~J., Pratap, P., Moran, J.~M., \& Wilson, T.~L.\ 1992, \apjl, 401, L39 

747: %\bibitem[Menten(1996)]{M96} Menten, K.~M.\ 1996, 

748: %   in IAU Symp. 178: Molecules in Astrophysics -- Probes \&\ Processes,

749: %   ed. E. van Dishoeck (Dordrecht: Kluwer), p. 163

750: \bibitem[Menten \& Reid (1995)]{MR95} Menten, K.~M.~\& Reid, M.~J.\ 1995, 

751:     \apj, 445, L157 

752: \bibitem[Moscadelli, Menten, Walmsley, \& Reid(1999)]{MMW99} Moscadelli, L., 

753:    Menten, K.~M., Walmsley, C.~M., \& Reid, M.~J.\ 1999, \apj, 

754:    519, 244 

755: %\bibitem[Mueller, Shirley, Evans, \& Jacobson(2002)]{MSE02} 

756: %   Mueller, K.~E., Shirley, Y.~L., Evans, N.~J., \& Jacobson, H.~R.\ 2002, 

757: %   \apjs, 143, 469 

758: \bibitem[Neufeld \& Dalgarno(1989)]{ND89} Neufeld, D.~A.~\&

759:    Dalgarno, A.\ 1989, \apj, 340, 869

760: %\bibitem[Neufeld \& Melnick(1991)]{NM91} Neufeld, D.~A.~\& 

761: %   Melnick, G.~J.\ 1991, \apj, 368, 215 

762: %\bibitem[Norris(1984)]{N84} Norris, R.~P.\ 1984, \mnras, 

763: %   207, 127 

764: \bibitem[Reid et al.(1995)]{RAM95} Reid, M.~J., Argon, A.~L.,

765:    Masson, C.~R., Menten, K.~M., \& Moran, J.~M.\ 1995, \apj, 443, 238

766: \bibitem[Reid et al.(1980)]{RHB80} Reid, M.~J., Haschick, 

767:    A.~D., Burke, B.~F., Moran, J.~M., Johnston, K.~J., \& Swenson, G.~W.\ 

768:    1980, \apj, 239, 89 

769: \bibitem[Reid et al.(1988)]{RSM88} Reid, M.~J., Schneps,

770:    M.~H., Moran, J.~M., Gwinn, C.~R., Genzel, R., Downes, D., \& Roennaeng,

771:    B.\ 1988, \apj, 330, 809

772: %\bibitem[Schneps et al.(1981)]{SMG81} Schneps, M.~H., Moran, 

773: %   J.~M., Genzel, R., Reid, M.~J., Lane, A.~P., \& Downes, D.\ 1981, \apj, 

774: %   249, 124 

775: %\bibitem[Sridharan et al.(2002)]{SBS02} Sridharan, T.~K., 

776: %   Beuther, H., Schilke, P., Menten, K.~M., \& Wyrowski, F.\ 2002, \apj, 566, 

777: %   931 

778: %\bibitem[Stecklum et al.(2002)]{SBH02} Stecklum, B., Brandl, 

779: %   B., Henning, T., Pascucci, I., Hayward, T.~L., \& Wilson, J.~C.\ 2002, 

780: %   \aap, 392, 1025 

781: %\bibitem[Testi, Felli, Persi, \& Roth(1994)]{TFP94} Testi, 

782: %   L., Felli, M., Persi, P., \& Roth, M.\ 1994, \aap, 288, 634 

783: \bibitem[Turner \& Welch(1984)]{TW84} Turner, J.~L.~\& 

784:    Welch, W.~J.\ 1984, \apjl, 287, L81 

785: \bibitem[Wilner, Reid, \& Menten(1999)]{WRM99} Wilner, D.~J., 

786:    Reid, M.~J., \& Menten, K.~M.\ 1999, \apj, 513, 775 

787: %\bibitem[Wootten(2001)]{W01} Wootten, A. 2001, Science with the 

788: %   Atacama Large Millimeter Array, ASP Conference Proceeding Vol. 235, 

789: %   ed. A. Wootten (San Francisco:ASP)

790: \bibitem[Wyrowski et al.(1997)]{WHS97} Wyrowski, F., Hofner, 

791:    P., Schilke, P., Walmsley, C.~M., Wilner, D.~J., \& Wink, J.~E.\ 1997, 

792:    \aap, 320, L17 

793: \bibitem[Wyrowski, Schilke, Walmsley, \& Menten(1999)]{WSW99} 

794:    Wyrowski, F., Schilke, P., Walmsley, C.~M., \& Menten, K.~M.\ 1999, \apjl, 

795:    514, L43 

796: \end{thebibliography}

797: 

798: 

799: \end{document}

800: