1: \documentclass[12pt, preprint]{aastex}
2: %\documentclass{emulateapj}
3: \begin{document}
4:
5: \title{Interferometric $890\micron$ Images of High Redshift Submillimeter Galaxies}
6: \author {D. Iono\altaffilmark{1,2}, A. B. Peck\altaffilmark{1},
7: A. Pope\altaffilmark{3}, C. Borys\altaffilmark{4},
8: D. Scott\altaffilmark{3}, D. J. Wilner\altaffilmark{1},
9: M. Gurwell\altaffilmark{1}, P. T. P. Ho\altaffilmark{1},
10: M. S. Yun\altaffilmark{5},
11: S. Matsushita\altaffilmark{6},
12: G. R. Petitpas\altaffilmark{1},
13: J. S. Dunlop\altaffilmark{7},
14: M. Elvis\altaffilmark{1},
15: A. Blain\altaffilmark{4},
16: E. Le Floc'h\altaffilmark{8}}
17: \altaffiltext{1}{Harvard-Smithsonian CfA, 60 Garden St., Cambridge, MA 02138}
18: \altaffiltext{2}{National Astronomical Observatory of Japan, 2-21-1 Osawa, Mitaka, Tokyo 181-8588; d.iono@nao.ac.jp}
19: \altaffiltext{3}{Dept. of Physics \& Astro., UBC, Vancouver, BC, V6T 1Z1}
20: \altaffiltext{4}{California Institute of Technology, Pasadena, CA 91125}
21: \altaffiltext{5}{Dept. of Astronomy, Univ. of Mass., Amherst, MA 01003}
22: \altaffiltext{6}{Academia Sinica Institute of Astronomy and Astrophysics, P.O. Box 23-141, Taipei 106, Taiwan, R.O.C.}
23: \altaffiltext{7}{Institute for Astronomy, University of Edinburgh, Royal Observatory, Blackford Hill, Edinburgh EH9 3HJ}
24: \altaffiltext{8}{Steward Observatory, 933 N. Cherry Ave, Tucson, AZ 85721}
25: \begin{abstract}
26:
27: We present high resolution $890\micron$ images of two 20~mJy
28: submillimeter galaxies, SMMJ123711+622212
29: and MIPS~J142824.0+352619, obtained using the
30: Submillimeter Array (SMA).
31: Using submillimeter interferometric observations
32: with an angular resolution of $2\farcs5$,
33: the coordinates of these high
34: redshift sources are determined with an accuracy of $0\farcs2$.
35: The new SMA data on SMMJ123711+622212
36: reveal an unresolved submm source
37: offset to the east by $0\farcs8$ from an optical galaxy
38: found in deep $HST$ images, suggesting
39: either a large galaxy with a dusty central region, or an interacting galaxy system.
40: The SMA image of hyper-luminous
41: (L$_{\rm FIR} = 3.2 \times 10^{13}$~L$_{\odot}$)
42: source MIPS~J142824.0+352619 provides
43: a firm upper limit to the source size of
44: $\lesssim 1\farcs2$. This constraint
45: provides evidence that the foreground lens is only weakly
46: affecting the observed high FIR luminosity.
47:
48: \end{abstract}
49:
50: \keywords{galaxies: formation, galaxies: starburst, cosmology: observations, galaxies: high redshift, submillimeter }
51:
52: \section{Introduction}
53:
54: The discovery of high redshift submillimeter sources has significantly
55: improved our understanding of the star formation history in the early universe.
56: Negative k-correction allows observation of the thermal dust emission
57: at $850\micron$
58: almost independent of the redshift up to $z\sim 10$ \citep{blain02}.
59: Deep observations using the SCUBA bolometer on the
60: %James Clerk Maxwell Telescope (JCMT) have played a major
61: %role in unveiling the presence of the distant submillimeter sources
62: James Clerk Maxwell Telescope (JCMT) have
63: unveiled the presence of distant submm sources
64: \citep{smail97,barger98,hughes98,eales00,cowie02, scott02,borys03, webb03, wang04}.
65: The primary origin of the submm emission is believed to be
66: the reprocessed dust emission from newborn stars in young galaxies.
67: %most of which were beginning to form when the universe was less than
68: %a few billion years old.
69: While these discoveries are attended by a great number of high resolution
70: follow-up optical/NIR imaging studies, the $14''$
71: %%angular
72: resolution of the
73: JCMT at 850$\micron$ yields a large error circle which
74: is too coarse for a precise determination of the
75: optical/NIR counterparts to these sources. Deep optical imaging
76: typically shows several optical/NIR
77: sources within the SCUBA beam.
78: To date, the most successful ways
79: to obtain precise astrometry on the target are to obtain high
80: resolution, deep 1.4~GHz radio images
81: \citep{ivison98,barger00,chapman01,ivison02,dunlop04},
82: or to obtain interferometric 1.3~mm continuum images \citep[e.g.][]{downes99}.
83: The former, however, does not identify
84: robust optical/NIR counterparts for all of the sources, revealing counterparts for
85: $\sim75\%$ of the $S_{850\micron} > 5$~mJy sources with
86: $S_{\rm 1.4GHz} > 30$~$\mu$Jy
87: \citep[e.g.][]{ivison02,chapman03,borys04,greve04,wang04}.
88: Precise astrometry obtained using mm or submm interferometers
89: can unambiguously identify the correct counterpart
90: for the remaining radio-faint sources,
91: and sources with multiple radio counterparts.
92: High angular resolution submm observations also allow us
93: to understand the true nature of the submm
94: sources with established optical counterparts
95: in which gravitational lensing is a possibility.
96:
97: We present recent Submillimeter Array
98: \citep[SMA;][]{ho04} detections of two $S_{850\micron} \sim 20$~mJy
99: sources, SMMJ123711+622212 \citep[hereafter GN~20;][]{pope05}
100: and MIPS~J142824.0+352619
101: \citep[hereafter MIPS-J1428;][]{borys05}.
102: The 20.3~mJy source GN~20 was discovered in the recent SCUBA
103: observations of the GOODS North Field \citep{giavalisco04}.
104: %with S/N close to 10 \citep{pope05}.
105: GN~20, with its $10\sigma$
106: %%SCUBA
107: $850\micron$ detection \citep{pope05} and
108: $5\sigma$ 1.3~mm detection at IRAM PdBI
109: (Pope et al. in preparation) make this one of the
110: strongest submm sources discovered to date. % \citep{pope05}.
111: It has very weak radio and undetectable $450\micron$ emission,
112: suggesting it lies at high redshift.
113: %\textit{Spitzer}/IRAC galaxies in
114: %the vicinity are possible counterparts, and high resolution submillimeter
115: %imaging can discriminate between them.
116: %GN~20 is not detected in 450$\micron$, X-ray \citep{alexander03}, or in the
117: %existing 1.4~GHz radio map \citep{richards00}, but its robust
118: %SCUBA detection and a $5\sigma$ 1.3~mm detection at IRAM PdBI
119: %(Pope et al. in preparation) make this one of the strongest submillimeter
120: %sources discovered to date \citep{pope05}.
121: %\textit{Spitzer}/IRAC galaxies in the vicinity are possible counterparts,
122: %and high resolution submillimeter imaging can discriminate between them.
123: MIPS-J1428 was discovered in the \textit{Spitzer} MIPS images of the
124: NDWFS Bo\"{o}tes field \citep{jannuzi99,soifer04}.
125: It was detected \citep{borys05} at $350\micron$
126: using SHARC-II \citep{dowell03} on the CSO, 1.4~GHz radio
127: continuum at the VLA, and
128: subsequent followup observations \citep{borys05} revealed that it is an
129: extremely luminous ${\rm (3.2\pm0.7)\times10^{13}\,L_\odot}$
130: starburst dominated galaxy at $z=1.325$.
131: Recent Keck-DEIMOS spectroscopy has revealed a $z=1.034$ galaxy
132: directly aligned with MIPS-J1428 \citep{borys05}, possibly lensing it.
133: High angular resolution submm observations allow us to determine
134: whether or not the submm emission is coincident with the
135: optical emission, and thereby search for any evidence for amplification
136: of the FIR luminosity by gravitational lensing.
137: We adopt $H_0 = 70$~km~s$^{-1}$~Mpc$^{-1}$,
138: $\Omega_m$ = 0.3, $\Omega_{\Lambda}$ = 0.7.
139:
140: \section{Observation and Data Reduction}
141:
142: GN~20 was observed on February 20 (track~1) and
143: March 5, 2005 (track~2), and MIPS-J1428 was observed on March 8 (track~3)
144: and April 4 (track~4), 2005 using 5 -- 7 antennas in the compact
145: configuration of the SMA.
146: The SIS receivers were tuned to a center frequency of 342.883~GHz in the
147: upper sideband (USB), yielding 332.639~GHz in the lower sideband (LSB).
148: This tuning frequency was chosen to facilitate the
149: receiver tuning and to achieve the optimal receiver performance.
150: The target coordinates were obtained from the
151: 1.3~mm IRAM PdBI detection of GN~20 ($\alpha$~(J2000)~$= 12^h 37^m 11.88^s$,
152: $\delta$~(J2000)~$=62^{\circ}22' 12\farcs00$; Pope et al. in preparation),
153: and the position of
154: the proposed MIR galaxy of MIPS-J1428
155: \citep[$\alpha$~(J2000)~$= 14^h 28^m 24.10^s$,
156: $\delta$~(J2000)~$=35^{\circ}26' 19\farcs00$;][]{borys05}.
157: All tracks were taken under good atmospheric opacity
158: (i.e. $\tau_{225} = 0.04$ -- $0.08$).
159:
160:
161: The SMA data were calibrated using the Caltech software package MIR,
162: modified for the SMA.
163: Antenna based passband calibration was done using
164: all of the planets and bright QSOs observed in a given track.
165: For GN~20, antenna based time-dependent
166: phase calibration was done using 1153+495, a 0.7~Jy QSO 14$^\circ$
167: away from the target. In addition, a 0.6~Jy QSO near GN~20, 1048+717
168: (14$^\circ$ from GN~20, 23$^\circ$ from 1153+495) was observed
169: for a total of 12 minutes during each track. The detection of
170: 1048+717 at the phase center empirically verifies and constrains
171: the accuracy of the
172: phase calibration referenced to 1153+495 (see \S3).
173: Similarly for the MIPS-J1428 tracks, two QSOs 1310+323
174: (0.6~Jy; 16$^\circ$ away from MIPS-J1428)
175: and 1635+381 (1.0~Jy; 25$^\circ$ away from MIPS-J1428) were used together
176: to calibrate the time dependent phase,
177: and 1419+543 (0.4~Jy; 3$^\circ$ away from MIPS-J1428)
178: was used to check the astrometry.
179: Finally, absolute flux calibration was performed using Callisto and Mars.
180: Imaging was carried out in MIRIAD \citep{miriad}.
181: Maximum sensitivity was achieved by
182: adopting natural weighting, which gave a
183: synthesized beam size of $2\farcs9 \times 2\farcs2$ (P.A.=
184: $12.2^\circ$) for GN~20
185: and $2\farcs6 \times 2\farcs4$ (P.A. = $-47.2^\circ$)
186: for MIPS-J1428.
187: The rms noise after combining the two sidebands in two tracks
188: was 2.1~mJy (GN~20) and 2.2~mJy (MIPS-J1428).
189:
190: \section{Results and Discussion}
191:
192: \subsection{SMMJ123711+622212 (GN~20)}
193: The SMA image of GN~20 is shown in Figure~\ref{fig1}~(a).
194: The real part of the visibility amplitudes do not decline as a function of
195: projected baseline length,
196: indicating that the emission is not spatially resolved (Figure~\ref{fig2}).
197: The upper limit on the source size is $1\farcs2$.
198: The derived total flux from a point source model
199: is $22.9 \pm 2.8$~mJy, consistent
200: with the $890\micron$ flux of $18.1$~mJy
201: extrapolated from the SCUBA $850\micron$ flux of
202: $20.3 \pm 2.1$ mJy \citep{pope05}.
203: The derived coordinates for GN~20 are
204: $\alpha$~(J2000)~$= 12^h 37^m 11.92^s$,
205: $\delta$~(J2000)~$=62^{\circ}22' 12\farcs10$, with
206: statistical uncertainties in the fit of
207: $0\farcs1$ for both $\alpha$ and $\delta$. These are consistent with
208: estimated errors of $\sim 0\farcs11$ in $\alpha$ and
209: $\sim 0\farcs15$ in $\delta$
210: from a $10 \sigma$ detection and a beam of $2\farcs9 \times 2\farcs2$.
211:
212:
213: In order to check the robustness of our phase calibration and to estimate
214: the systematic uncertainties in the SMA astrometry,
215: we have imaged the test QSO (1048+717)
216: using the same phase calibration we used to map GN~20.
217: The resultant QSO map after adding the two sidebands and two tracks
218: is shown in Figure~\ref{fig1} (a)(inset).
219: A point source fit to the visibilities of 1048+717 gave a
220: positional offset from the phase center of
221: $\Delta \alpha = 0\farcs01 \pm 0\farcs02$ and $\Delta \delta = 0\farcs06 \pm 0\farcs02$.
222: The coordinates of this and all of the QSOs used in these observations
223: were adopted from the Radio Reference Frame \citep{johnston95},
224: which are accurate to better than 3 mas.
225: Hence the precise detection of 1048+717 at the phase center
226: ensures that our phase calibration referenced to 1154+379 is robust.
227:
228: As an additional check, we fit a point source model
229: to the visibilities of 1048+717 in each sideband of each track separately.
230: The results show that the offsets from the phase center are
231: consistent among the different sidebands and tracks in R.A.
232: ($\Delta \alpha = 0\farcs01$ -- $0\farcs04$),
233: and slightly larger in Dec. for track~2 ($\Delta \delta = 0\farcs02$ -- $0\farcs05$),
234: and a factor 10 larger in Dec. for track~1 ($\Delta \delta \sim 0\farcs30$).
235: The overall average offset from
236: the phase center is $\Delta \alpha = 0\farcs02 \pm 0.02$ and
237: $\Delta \delta = 0\farcs16 \pm 0.14$.
238: The uncertainties in Dec. in track~2 are larger than the
239: uncertainties in R.A. due to
240: beam elongation in the north-south direction, while the
241: source of the factor 10 larger error in track~1
242: is not obvious from the data.
243: These errors are consistent with the uncertainties of
244: $0\farcs1$ -- $0\farcs15$ expected from
245: a maximum baseline error of $0.1\lambda$
246: at 230~GHz using a phase calibrator
247: that is $15^\circ$ away from the target.
248: These systematic tests prove the robustness of our
249: phase calibration in each track
250: and lend high confidence to the resulting positional accuracy
251: of $\lesssim 0\farcs1$ in R.A. and $0\farcs1$ -- $0\farcs2$ in Dec. for GN~20.
252:
253:
254: The new astrometric coordinates allow us to compare the submm
255: source with high resolution images in the publicly available
256: deep $Spitzer$ IRAC \citep{fazio04} (Figure~\ref{fig1}~(b)) and
257: $HST$ ACS \citep{ford98} (Figure~\ref{fig1}~(c)) images
258: of the GOODS North Field.
259: The absolute astrometric accuracy of both of these images is $\sim 0\farcs1$,
260: and they are both tied to the coordinate frame defined by the VLA positions in
261: \citet{richards00}\footnote{After applying the known positional offset in Dec.
262: of $0\farcs38$ between the GOODS-N images and the VLA catalog positions. See
263: http://data.spitzer.caltech.edu/popular/goods/Documents/ for details.}.
264: The IRAC 3.6$\micron$ image reveals a source centered $< 0\farcs5$ west of
265: the submm coordinates, while the higher resolution
266: $HST$ $V$-band image reveals a faint optical
267: source $0\farcs8$ to the west.
268: From the analysis of available ACS images,
269: it is found that this optical source is a B-dropout galaxy ($B = 27.2 \pm 0.4$,
270: $V = 25.2 \pm 0.1$, $i = 24.4 \pm 0.1$) which
271: gives constraints on the probable redshift to be $z\sim 3$--4.
272: We believe that the $0\farcs8$ offset between the SMA position and the
273: \textit{HST} position is significant, and there are several possible
274: astrophysical explanations for this difference.
275: The submm emission may arise
276: from part of a large galaxy where $V$-band emission is completely obscured.
277: Alternatively, GN~20 might be an interacting
278: system where the optical galaxy is a companion to the dusty, more actively
279: star forming galaxy. Although the observed FIR luminosities are
280: significantly different, the apparent separation of
281: these galaxies ($\sim 6$~kpc at $z \sim 3 $)
282: suggest a close similarity to the Antennae system (NGC~4038/39), where
283: the optical galaxies are separated by 7.5~kpc and
284: most of the starburst activity is occurring in the medium between the
285: two galaxies \citep[e.g.][]{wang04b}.
286:
287:
288: GN~20 was suggested to be a two component source, GN~20.1 (20.3~mJy)
289: and GN~20.2 (11.7~mJy), separated by $18''$~(140~kpc)
290: in the low resolution SCUBA image \citep{pope05}.
291: The $890\micron$ flux agreement between our SMA observation and
292: the SCUBA measurement of GN~20.1 implies that GN~20 may be
293: a two component source. However, the suggested position of GN~20.2 is
294: beyond the half power point of the SMA primary beam,
295: where the sensitivity is reduced by more than a factor of two.
296: Detailed discussion in the context of multi-wavelength
297: observations of this source
298: will be provided in a forthcoming paper (Pope et al. in preparation).
299:
300:
301: \subsection{MIPS~J142824.0+352619 (MIPS-J1428)}
302:
303: The SMA map of MIPS-J1428 is shown in Figure~\ref{fig1}~(d).
304: The derived total flux is $18.4 \pm 2.5$ mJy and shows
305: excellent agreement with the $890\micron$ flux of $19.5$~mJy
306: extrapolated from the SCUBA $850\micron$ flux of
307: $21.9 \pm 1.3$~mJy.
308: MIPS-J1428 is not spatially resolved with the $2\farcs5$ beam
309: (see Figure~\ref{fig2}),
310: and the derived coordinates from a point source fit are
311: $\alpha~(2000)= 14^h 28^m 24.06^s$, $\delta~(2000)=35^{\circ} 26' 19\farcs79$,
312: with uncertainties in the fit of $0\farcs1$ for both $\alpha$ and $\delta$.
313: The strong $890\micron$ detection allows us to make a higher
314: resolution image using the visibilities of the longest baselines.
315: %at the expense of signal.
316: The resulting $3\sigma$ unresolved image provides
317: a firm upper limit to the source size of
318: $\lesssim 1\farcs2$ (10~kpc at $z = 1.325$), which is smaller than
319: the size constraint given by the $\sim 1\farcs5$ VLA 1.4~GHz resolution.
320:
321:
322: The map of the test QSO 1419+543
323: after adding the two sidebands and two tracks
324: is shown in Figure~\ref{fig1}~(d)~(inset).
325: A point source fit to the visibilities gave a
326: positional offset from the phase center of
327: $\Delta \alpha = 0\farcs11 \pm 0.05$ and $\Delta \delta = 0\farcs23 \pm 0.05$
328: for 1419+543.
329: As with 1048+717 for GN~20, we fit a point source model
330: to the visibilities of 1419+543 in each sideband of each track separately.
331: The results show that the offsets from the phase center
332: had a wide range of values
333: ($\Delta \alpha = 0\farcs04 -0\farcs28$ and $\Delta \delta = 0\farcs13 - 0\farcs25$)
334: with the overall average offset from
335: the phase center of $\Delta \alpha = 0\farcs14 \pm 0.11$ and
336: $\Delta \delta = 0\farcs21 \pm 0.05$.
337: These uncertainties are larger than those found in 1048+717 for
338: GN~20, and are slightly larger than the uncertainties expected
339: from baseline errors. It is possible that other factors such as image smearing
340: due to large phase noise may have introduced a small error in
341: the source positions.
342: Thus, we assess an uncertainty
343: of $\sim 0\farcs15$ in R.A. and $\sim 0\farcs2$ in Dec. for MIPS-J1428.
344:
345:
346:
347: Figure~\ref{fig1}~(e) and (f) show the SMA $890\micron$ contours
348: of MIPS-J1428 overlaid on the NOAO Deep Wide Field Survey K-band and I-band
349: images. The astrometry of both of these images is tied
350: to the reference frame defined by the USNO-A2.0 catalog,
351: and the typical rms of the residuals is $0\farcs35$ \citep{jannuzi05}.
352: The optical/NIR galaxy seen here is unambiguously
353: the galaxy aligned with the strong submm emission.
354: The accurate astrometry of the submm emission provided by the SMA allows
355: tight constraints on the separation between the bright optical/NIR position,
356: and rules out fainter IRAC sources detected nearby \citep{borys05}. Hence
357: MIPS-J1428 could lie directly behind the foreground $z=1.034$ galaxy, which
358: would potentially result in a large amplification of the submm source.
359: \citet{borys05} use size/luminosity relationships to argue that despite
360: the alignment, the amplification is likely modest since the Einstein
361: ring is of comparable size to the known physical scales of local ULIRGs.
362: Assuming that the lensing is modest, the size
363: is comparable to that of other high-redshift submm sources \citep{chapman04}.
364: Using the star-formation rate estimated
365: in \citet{borys05}, our limit on the angular size, $\theta$, of the object,
366: we derive a lower limit on the SFR density of
367: $>180(\theta/1.2)\mu^{-1}$M$_\odot$yr$^{-1}$kpc$^{-2}$ where $\mu$ is the
368: lensing amplification. This is larger than
369: local ULIRGs \citep{meurer97} and comparable to other high-redshift
370: submm galaxies \citep{chapman04}.
371:
372:
373: \section{Summary}
374:
375: %We present SMA observations of two recently discovered
376: %20~mJy submillimeter sources,
377: We present SMA observations of two 20~mJy submm sources,
378: GN~20 and MIPS-J1428. The positions of the submm sources are
379: determined with $0\farcs1 - 0\farcs2$ accuracy with these data,
380: allowing precise identification of the correct
381: optical galaxy counterpart to the bright submm emission in GN~20, and
382: providing evidence that the foreground lens is only weakly affecting
383: the observed high FIR luminosity in MIPS-J1428.
384: If many of the bright submm galaxies are slightly lensed objects
385: similar to MIPS-J1428, then the implied source counts of
386: inherently bright submm galaxies are over-predicted.
387: Detailed studies, however, exist for only a few sources, and
388: future surveys such as SHADES \citep{mortier05}
389: will provide important information about
390: the star formation properties at high-redshifts.
391:
392: The Submillimeter
393: Array is a joint project between the Smithsonian Astrophysical Observatory
394: and the Academia Sinica Institute of Astronomy and Astrophysics, and is
395: funded by the Smithsonian Institution and the Academia Sinica.
396: This work made use of observations made with the Spitzer
397: Space Telescope, which is operated by JPL,
398: California Institute of Technology under NASA contract 1407.
399: This work made use of images provided by the
400: NOAO Deep Wide-Field Survey \citep{jannuzi99}
401: which is supported by the National Optical Astronomy Observatory (NOAO).
402: NOAO is operated by AURA, Inc., under a cooperative agreement with the
403: National Science Foundation.
404:
405:
406: \begin{thebibliography}{fun}
407: \bibitem[Alexander et al.(2003)]{alexander03} Alexander, D. M., et al. 2003, AJ, 126, 539
408: \bibitem[Barger et al.(1998)]{barger98} Barger, A. J., et al. 1998, Nature, 394, 248
409: \bibitem[Barger, Cowie \& Richards(2000)]{barger00} Barger, A. J., Cowie, L. L. \& Richards, E. A. 2000, AJ, 119, 2092
410: \bibitem[Borys et al.(2003)]{borys03} Borys, C., Chapman, S., Halpern, M. \& Scott, D. 2003, MNRAS, 344, 385
411: \bibitem[Borys et al.(2004)]{borys04} Borys, C., et al. 2004, MNRAS, 352, 759
412: \bibitem[Borys et al.(2005)]{borys05} Borys, C., et al., 2005, \apj, accepted (astro-ph/0509361)
413: \bibitem[Blain et al.(2002)]{blain02} Blain, A. W., Smail, I., Ivison, R. J., Kneib, J. -P. \& Frayer, D. T. 2002, PhR, 369, 111
414: \bibitem[Chapman et al.(2001)]{chapman01} Chapman, S. C., Richards, E. A., Lewis, G. F., Wilson, G. \& Barger, A. J. 2001, ApJ, 548, 147
415: \bibitem[Chapman et al.(2003)]{chapman03} Chapman, S. C., et al. 2003, ApJ, 585, 57
416: \bibitem[Chapman et al.(2004)]{chapman04} Chapman, S.~C., Smail, I., Windhorst
417: , R., Muxlow, T., \& Ivison, R.~J.\ 2004, \apj, 611, 732
418: \bibitem[Cowie, Barger \& Kneib(2002)]{cowie02} Cowie, L. L., Barger, A. J. \& Kneib, J. -P. 2002, AJ, 123, 2197
419: \bibitem[Dowell et al.(2003)]{dowell03} Dowell et al. 2003, SPIE, 4855, 73
420: \bibitem[Downes et al.(1999)]{downes99} Downes et al. 1999, A\&A, 347, 809
421: \bibitem[Dunlop et al.(2004)]{dunlop04} Dunlop et al. 2004, MNRAS, 350, 769
422: \bibitem[Eales et al.(2000)]{eales00} Eales, S., Lilly, S., Webb, T., Dunne, L., Gear, W., Clements, D. \& Yun, M. S. 2000, AJ, 120, 2244
423: \bibitem[Fazio et al.(2004)]{fazio04} Fazio, G. G. et al. 2004, ApJS, 154, 10
424: \bibitem[Ford et al.(1998)]{ford98} Ford, H. C. et al. 1998, SPIE, 3356, 234
425: \bibitem[Greve et al.(2004)]{greve04} Greve, T. R., Ivison, R. J., Bertoldi, F., Stevens, J. A., Dunlop, J. S., Lutz, D. \& Carilli, C. L. 2004, MNRAS, 354, 779
426: \bibitem[Giavalisco et al.(2004)]{giavalisco04} Giavalisco M. et al. 2004, ApJ, 600, L93
427: \bibitem[Ho, Moran \& Lo(2004)]{ho04} Ho, P. T. P., Moran, J. M. \& Lo, K. Y. 2004, ApJ, 616, 1
428: \bibitem[Hughes et al.(1998)]{hughes98} Hughes, D. H., et al. 1998, Nature, 394, 241
429: \bibitem[Ivison et al.(1998)]{ivison98} Ivison, R. J., et al. 1998, MNRAS, 298, 583
430: \bibitem[Ivison et al.(2002)]{ivison02} Ivison, R. J., et al. 2002, MNRAS, 337, 1
431: \bibitem[Jannuzi \& Dey(1999)]{jannuzi99} Jannuzi, B. T. \& Dey, A., 1999, in "Photometric Redshifts and the Detection of High Redshift Galaxies", ASP Conference Series, Vol. 191, Edited by R. Weymann, L. Storrie-Lombardi, M. Sawicki, and R. Brunner. ISBN: 158381-017-X, p. 111
432: \bibitem[Jannuzi et al.(2005)]{jannuzi05} Jannuzi, B. T. et al. 2005, in prep.
433: \bibitem[Johnston et al.(1995)]{johnston95} Johnston, K. J. et al. 1995, AJ, 110, 880
434: \bibitem[Meurer et al.(1997)]{meurer97} Meurer, G. R., Heckman, T. M., Lehnert
435: , M. D., Leitherer, C., \& Lowenthal, J., 1997, AJ, 114, 54
436: \bibitem[Mortier et al.(2005)]{mortier05} Mortier, A. M. J., et al. 2005, MNRAS in press
437: \bibitem[Pope et al.(2005)]{pope05} Pope, A., Borys, C., Scott, D., Conselice, C., Dickinson, M. \& Mobasher, B. 2005, MNRAS, 358, 149
438: \bibitem[Richards(2000)]{richards00} Richards, E. A. 2000, ApJ, 533, 611
439: \bibitem[Sault, Teuben \& Wright(1995)]{miriad} Sault, R. J., Teuben, P. J., \& Wright, M. C. H. 1995, in ASP Conf. Ser. 77, Astronomical Data Analysis Software and Systems IV, ed. R. A. Shaw, H.E. Payne, \& J. J. E. Haynes (San Francisco:ASP), 433
440: \bibitem[Scott et al.(2002)]{scott02} Scott et al. 2002, MNRAS, 331, 817
441: \bibitem[Soifer et al.(2004)]{soifer04} Soifer, B. T., Spitzer/NOAO Team, 2004, AAS, 204, 4805
442: \bibitem[Smail, Ivison \& Blain(1997)]{smail97} Smail, I., Ivison, R. J. \& Blain, A. W. 1997, ApJ, 490, 5
443: \bibitem[Wang, Cowie \& Barger(2004)]{wang04} Wang, W.-H., Cowie, L. L. \& Barger, A. J. 2004, ApJ, 613, 655
444: \bibitem[Wang et al.(2004)]{wang04b} Wang, Z., et al. 2004, ApJS, 154, 193
445: \bibitem[Webb et al.(2003)]{webb03} Webb, T. M., et al. 2003, ApJ, 587, 41
446: \end{thebibliography}
447:
448: \clearpage
449:
450: \begin{figure}
451: \plotone{f1.eps}
452: \caption{(\textit{a}) The SMA 890$\micron$ map of GN~20 and
453: (\textit{inset}) 1048+717.
454: The lowest positive contours represent $2\sigma$, and the contours increase
455: by $1 \sigma$ for GN~20, and by $2 \sigma$ for 1048+717.
456: The lowest negative contour is 2$\sigma$ and increases by $1 \sigma$.
457: (\textit{b}) The SMA $890\micron$ contours overlaid on the $Spitzer$
458: IRAC $3.6\micron$ image and over the (\textit{c}) $HST$ ACS $V$-band image,
459: both obtained from the GOODS archive.
460: The IRAC and $HST$ images are corrected for the known $0\farcs38$ offset
461: in declination.
462: The 4, 6, and 8 $\sigma$ contours from (\textit{a}) are shown, and
463: errorbars near the center show the astrometric accuracy of the SMA image.
464: (\textit{d}) The SMA map of MIPS-J1428 and (\textit{inset}) 1419+543.
465: The contours are the same as in GN~20.
466: The smallness of the astrometric errors ($0\farcs1$ -- $0\farcs2$)
467: from the phase center in the 1048+717 and 1419+543 maps
468: prove the robustness of the astrometry of GN~20 and MIPS-J1428.
469: The SMA $890\micron$ map of MIPS-J1428 is shown overlaid on the NDWFS
470: (\textit{e}) K-band image and the (\textit{f}) I-band image.
471: The 3, 5, and 7 $\sigma$ contours from (\textit{d}) are shown, and
472: errorbars near the center show the astrometric accuracy of the SMA image.
473: }
474: \label{fig1}
475: \end{figure}
476:
477: \clearpage
478:
479: \begin{figure}
480: \plotone{f2.eps}
481: \caption{The real part of the visibility amplitudes vs. projected
482: baseline length for (\textit{top}) GN~20 and (\textit{bottom})
483: MIPS-J1428. The solid horizontal lines represent a point source
484: model with continuum fluxes of 22.9~mJy~(GN~20) and
485: 18.4~mJy~(MIPS-J1428). The upper limit to the source sizes of
486: both of these sources are comparable to the scale constrained
487: by the longest baseline length
488: which is $\sim 1\farcs2$.
489: }
490: \label{fig2}
491: \end{figure}
492:
493: \end{document}
494: