1: \documentclass{emulateapj}
2: \usepackage{psfig}
3: \usepackage{apjfonts,times}
4:
5: %\usepackage[figuresright]{rotating}
6:
7: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
8: % AUTHOR'S MACROS
9: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
10: \newcommand{\beq}{\begin{equation}}
11: \newcommand{\eeq}{\end{equation}}
12: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
13: % AUTHOR'S SYMBOLS
14: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
15: \newcommand{\gapprox}{$\stackrel {>}{_{\sim}}$}
16: \newcommand{\lapprox}{$\stackrel {<}{_{\sim}}$}
17: \def\alp{\mbox{$\alpha$}}
18: \def\farcd{\hbox{$.\mkern-4mu^\circ$}}
19: \def\farcm{\hbox{$.\mkern-4mu^\prime$}}
20: \def\farcs{\hbox{$.\!\!^{\prime\prime}$}}
21: \def\earth{\hbox{$\oplus$}}
22: \def\arcmin{\hbox{$^\prime$}}
23: \def\arcsec{\hbox{$^{\prime\prime}$}}
24: \def\solar{\mbox{$_{\normalsize\odot}$}}
25: \def\fs{\hbox{$.\!\!^{\rm s}$}}
26: \def\deg{\hbox{$^\circ$}}
27: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28: \newcommand{\ttbs}{\char'134}
29: \newcommand{\AmS}{{\protect\the\textfont2
30: A\kern-.1667em\lower.5ex\hbox{M}\kern-.125emS}}
31: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
32: \newcommand{\lsim}{\ \raise
33: -2.truept\hbox{\rlap{\hbox{$\sim$}}\raise5.truept\hbox{$<$}\ }}
34: \newcommand{\gsim}{\ \raise
35: -2.truept\hbox{\rlap{\hbox{$\sim$}}\raise5.truept\hbox{$>$}\ }}
36: \newcommand{\simsim}{\ \raise
37: -2.truept\hbox{\rlap{\hbox{$\sim$}}\raise5.truept\hbox{$\sim$}\ }}
38: \newcommand{\FeH}{\mbox{[Fe/H]}\,}
39: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
40:
41: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
42: % add words to TeX's hyphenation exception list
43: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
44: \hyphenation{author another re-commend-ed Post-Script}
45:
46: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
47: %% You can insert a short comment on the title page using the command
48: %% below.
49: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
50: \slugcomment{Accepted for Publication in ApJ October 2, 2007}
51: %\slugcomment{Draft Version \today}
52:
53:
54: \shorttitle{Star Formation and Clustering in NGC~346/N~66 in the SMC}
55: \shortauthors{E. Hennekemper et al.}
56:
57: \begin{document}
58:
59: \title{NGC~346 in the Small Magellanic Cloud. III. Recent Star Formation
60: and Stellar Clustering Properties in the Bright HII Region
61: N~66\altaffilmark{1}}
62:
63: %\footnotetext[]{Research supported by the Deutsche Forschungsgemeinschaft
64: %(German Research Foundation)}
65:
66: \author{Eva Hennekemper, Dimitrios A. Gouliermis, Thomas Henning,
67: Wolfgang Brandner}
68:
69: \affil{Max-Planck-Institute for Astronomy, K\"onigstuhl 17, 69117
70: Heidelberg, Germany}
71:
72: \and
73:
74: \author{Andrew E. Dolphin}
75:
76: \affil{Raytheon Corporation, USA}
77:
78:
79: %\author{Eva Hennekemper\altaffilmark{1},
80: % Dimitrios A. Gouliermis\altaffilmark{1},
81: % Thomas K. Henning\altaffilmark{1},\\
82: % Wolfgang Brandner\altaffilmark{1,2}, and
83: % Andrew E. Dolphin\altaffilmark{3}
84: %}
85: %\altaffiltext{1}{Max Planck Institute for Astronomy, K\"onigstuhl 17, 69117
86: %Heidelberg, Germany}
87: %\altaffiltext{2}{UCLA, Div. of Astronomy, 475 Portola Plaza, Los Angeles, CA
88: %90095-1547, USA}
89: %\altaffiltext{3}{Raytheon Corporation, USA}
90:
91: \altaffiltext{1}{Research supported by the Deutsche Forschungsgemeinschaft (German
92: Research Foundation)}
93:
94: %\email{dgoulier@mpia.de}
95:
96: \begin{abstract}
97:
98: In the third part of our photometric study of the star-forming region
99: NGC~346/N~66 and its surrounding field in the Small Magellanic Cloud
100: (SMC), we focus on the large number of low-mass pre-main sequence (PMS)
101: stars revealed by the Hubble Space Telescope Observations with the
102: Advanced Camera for Surveys. We investigate the origin of the observed
103: broadening of the pre-main sequence population in the $V-I$, $V$ CMD.
104: The most likely explanations are either the presence of differential
105: reddening or an age spread among the young stars. Assuming the latter,
106: simulations indicate that we cannot exclude the possibility that stars
107: in NGC 346 might have formed in two distinct events occurring about 10
108: and 5 Myr ago, respectively. We find that the PMS stars are not
109: homogeneously distributed across NGC 346, but instead are grouped in at
110: least five different clusters. On spatial scales from 0.8$''$ to 8$''$
111: (0.24 to 2.4\,pc at the distance of the SMC) the clustering of the PMS
112: stars as computed by a two-point angular correlation function is
113: self-similar with a power law slope $\gamma \approx -0.3$. The
114: clustering properties are quite similar to Milky Way star forming
115: regions like Orion OB or $\rho$\,Oph. Thus molecular cloud
116: fragmentation in the SMC seems to proceed on the same spatial scales as
117: in the Milky Way. This is remarkable given the differences in
118: metallicity and hence dust content between SMC and Milky Way star
119: forming regions.
120:
121: % The positions of these stars in the $V-I$, $V$
122: %CMD exhibit a broadening, which can be easily misinterpreted as an age
123: %spread. We check if other factors can produce such a widening. We find
124: %that reddening, as well as binarity and variability may cause the observed
125: %broadening, suggesting that the PMS population of the association NGC~346
126: %could be considered as the product of a single star formation event.
127: %Specifically, if reddening is indeed an important factor for this
128: %widening, then these PMS stars may appear much younger than what they
129: %really are. Under these circumstances, the previously estimated age of
130: %these stars (3 - 5 Myr) is too small. However, our tests show that we
131: %cannot exclude the possibility that these stars are the product of
132: %multi-epoch star formation.
133: %, assuming that
134: %
135: %%the interstellar reddening is somewhat lower, comparable to the one
136: %estimated from the OB stars of the association ($E(B-V)\simeq0.04$~mag).
137:
138: \end{abstract}
139:
140: \keywords{Magellanic Clouds --- Color-Magnitude diagram --- stars:
141: evolution --- galaxies: star clusters --- clusters: individual (NGC~346)}
142:
143:
144: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
145: \begin{figure*}[t!]
146: \epsscale{1.15}
147: \plottwo{f1a.eps}{f1b.eps}
148: \caption{Left: ACS/WFC observations of NGC 346,
149: composed of three overlapping pointings in the $I$-band.
150: The association NGC~346 covers the
151: center of the field, located almost in the middle of the ``bar" of N~66,
152: as it is outlined by the nebulosity. The intermediate-age cluster BS~90
153: (Rochau et al. 2007) is seen as the circular stellar concentration to
154: the north of the association.
155: {\em Right:} On the ACS/WFC H\alp\ image of NGC~346/N~66
156: stars with intrinsic H\alp\ excess are marked by circles.
157: The spatial distribution of the excess stars outlines the
158: H{\sc ii} region, and coincides with the overall distribution
159: of the PMS stars.}
160: \label{fig-ima}
161: \end{figure*}
162: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
163: %\footnote{MultiDrizzle Home Page is
164: %http://stsdas.stsci.edu/multidrizzle/}
165:
166: \section{Introduction}
167:
168: It is well known that Galactic OB associations also host large numbers
169: of fainter, low-mass pre-main sequence (PMS) stars (e.g.\ Preibisch \&
170: Zinnecker 1999; Sherry et al.\ 2004; Brice\~{n}o et al.\ 2005).
171: Low-mass PMS stars in stellar associations provide a longer-lasting
172: record of the most recent star formation events than the short-lived
173: high-mass stars. Large-scale surveys have identified hundreds of PMS
174: stars in nearby OB associations (Brice\~{n}o et al. 2007). In order to
175: understand the star formation triggering and history, or the Initial
176: Mass Function (IMF), one has to study both high- and low-mass stars in
177: star forming regions.
178:
179: In many cases the low-mass populations of galactic OB associations
180: cannot be easily distinguished from foreground main-sequence stars (e.g.
181: Brice\~{n}o et al. 2001). Recently PMS stars have been discovered for
182: the first time in an extragalactic stellar association, which suffers
183: much less from contamination by the galactic disk. {\em Hubble Space
184: Telescope} (HST) observations of the association LH\,52 in the Large
185: Magellanic Cloud (LMC) revealed $\approx$500 PMS stars with masses down
186: to 0.3\,M$_\odot$ (Gouliermis et al. 2006a). The investigation of
187: individual extragalactic pre-main sequence stars opens a new field of
188: study. As both high angular resolution and wide field-of-view are
189: required, HST is the ideal observatory to carry out these studies. While
190: HST observations of extragalactic associations cannot reach the
191: detection limits achieved for local associations, they provide a unique
192: opportunity for studying low-mass star formation in other galaxies.
193:
194: The OB association NGC~346 in the Small Magellanic Cloud (SMC) is
195: located in the central part of the brightest {\sc H ii} region in the
196: SMC, named LHA~115-N~66 or in short N~66 (Henize 1956). With 33
197: spectroscopically confirmed O and B stars, NGC~346 hosts the largest
198: sample of young, massive stars in the SMC (Walborn 1978; Walborn \&
199: Blades 1986; Niemela et al. 1986; Massey et al. 1989; Walborn et al.
200: 2000; Evans et al. 2006). Photoionization models by Rela\~{n}o et al.
201: (2002) imply that these stars are a major source of ionizing flux for
202: the surrounding diffuse interstellar medium (ISM). Recent imaging from
203: the Wide-Field Channel (WFC) of the {\em Advanced Camera for Surveys}
204: (ACS) on-board HST (GO Program 10248; PI: A. Nota) revealed the PMS
205: stellar content of the general region of NGC~346/N~66 down to the
206: sub-solar mass regime (Nota et al.\ 2006; Gouliermis et al.\ 2006b --
207: hereafter Paper I -- ; Sabbi et al.\ 2007). Nota et al.\ (2006) suggest
208: that all PMS stars in the association are the product of a single star
209: formation event, taking place 3 to 5 Myr ago. The PMS distribution in
210: the $V-I$, $V$ color-magnitude diagram (CMD), however, shows a prominent
211: widening, which may be explained by an age spread of $\approx$10 Myr.
212: This raises the question {\em if the PMS stars in NGC\,346 are indeed
213: the result of a single star formation event.}
214:
215: Sabbi et al.\ suggest that the PMS population is mainly concentrated in
216: a number of subclusters (three of them at the central part, where the
217: association is located), which formed at the same time from the
218: turbulence-driven density variations, and not following a sequential
219: process. Star counts, however, reveal only a few compact PMS clusters
220: with a significant number of stars in the area around the association.
221: The main body of the association cannot be divided into separate
222: subclusters (Paper I). {\em Could the remote clusters be the product of
223: a star formation event occurring before or after the event which
224: triggered the formation of the association?}
225:
226:
227: In the first part of our study of this extraordinary star-forming region
228: (Paper I), we compiled our ACS photometry of almost 100,000 stars, and
229: presented preliminary results on stellar types and their distinct
230: spatial distributions. We confirmed the co-existence of massive OB stars
231: and low-mass PMS stars in NGC 346. In the present paper, we explore in
232: greater detail the properties of the low-mass PMS stars with a focus on
233: clustered star formation in NGC~346/N~66.
234:
235: The paper is organized as follows: In \S~2 we describe the datasets from
236: HST/ACS, present the color-magnitude diagram, carry out a comparison of
237: our photometry of the brightest stars with previous photometric studies,
238: and discuss the reddening in the region. The broadening of the PMS
239: distribution in the CMD, a thorough discussion of possible causes and
240: explanations of this phenomenon as well as the spatial variations in the
241: CMD are presented in \S~3. In \S~4 we present the clustering properties
242: of the PMS stars and discuss them in terms of hierarchical star
243: formation. Our analysis of H\alp\ observations with HST/ACS as well as
244: previous {\em Spitzer} Space Telescope results on Young Stellar Objects
245: (YSOs) are presented in \S~5 and \S~6, respectively. In \S~7 we
246: summarize the results.
247:
248: \section{Description of the Data\label{sec-dat}}
249:
250: The observations used in this study, taken with HST/ACS (GO Program
251: 10248), and the photometry with the package {\tt DOLPHOT}\footnote{The
252: ACS mode of {\tt DOLPHOT} is an adaptation of the photometry package
253: {\tt HSTphot} (Dolphin 2000). It can be downloaded from {\tt
254: http://purcell.as.arizona.edu/dolphot/}.} is thoroughly presented in
255: Paper I. These observations cover an area of about 5\arcmin\ $\times$
256: 5\arcmin\ wide from imaging of three overlapping ACS fields observed
257: with the Wide-Field Channel (WFC) of the camera centered on NGC~346
258: (Table 1 of Paper I) in the filters $F555W$ and $F814W$ ($\equiv V, I$).
259: The $F814W$ image of the whole field is shown in Fig.~\ref{fig-ima}
260: (left). This area covers the intermediate-age cluster BS 90 (Bica \&
261: Schmidt 1995), the association NGC~346, and all known components of the
262: bright nebula N~66, also known as DEM L 103 (Davies, Elliott \& Meaburn
263: 1976). The field covering the central area, at RA (J2000) = 00$^{\rm
264: h}$59$^{\rm m}$07$^{\rm s}$, DEC (J2000) =
265: $-$72{\deg}10{\arcmin}31{\arcsec}, was also observed in the narrow-band
266: filter $F658N$, equivalent to H\alp\ (``Advanced Camera for Surveys
267: Instrument Handbook'' Ver. 7, Oct 2006) with three 512s exposures
268: (Fig.~\ref{fig-ima}, {\em right}). In the present study we used {\tt
269: DOLPHOT} to obtain photometry from the H\alp\ images, and to combine the
270: results with our $V$ and $I$ photometry (see \S~5).
271:
272: %\clearpage
273: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
274: \begin{figure*}[]
275: \epsscale{1.15}
276: \plottwo{f2a.eps}{f2b.eps}
277: \caption{Isodensity contour maps constructed from star counts of Upper
278: Main Sequence ({\em left}), and Pre-Main Sequence ({\em right}) stars, as
279: they have been selected in Paper I according to their distinct positions
280: in the CMD. The lowest isopleth corresponds to the background surface
281: density and isopleths of higher surface density are plotted in steps of
282: 1$\sigma$, where $\sigma$ is the standard deviation of the background
283: density. On the right, the positions of candidate YSOs found in the
284: S$^3$MC Spitzer Survey in the observed area around NGC~346 (Simon et al.
285: 2007) are also plotted with crosses and dots, while the brightest 31 OB
286: stars, spectroscopically verified by Massey et al. (1989), included in our
287: photometry are plotted with open boxes. North is up and east to the left
288: on the map.\label{fig-cms}}
289: \end{figure*}
290: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
291:
292: \subsection{The Color-Magnitude Diagram of NGC~346}
293:
294: Based on star counts performed in Paper I, we identified the areas where
295: stars of different types are concentrated. The isodensity contour maps
296: from counting the bright upper main sequence (UMS) and PMS stars are
297: shown in Fig.~\ref{fig-cms}. In Paper I these maps allowed the
298: identification of the region which is covered by the association alone.
299: This region is confined within a radius of about 0\farcm6 around the
300: geometric center of the system. The $V-I$, $V$ Color-Magnitude Diagram
301: (CMD) of the stars detected in this region, as the most representative
302: of the association NGC~346, is shown in Fig. \ref{fig-cmdsys}. This CMD
303: is characterized by a very bright upper main sequence and a prominent
304: sequence of PMS stars, both contaminated by the stellar content of the
305: general field of the SMC, but also of the second dominant stellar
306: concentration, the cluster BS 90. We also identified the area covered by
307: the latter, which is the subject of the second part of our study (Rochau
308: et al. 2007). The part we present here deals with the recent star
309: formation of this active region, and thus it focuses on the young
310: populations as they are revealed from the bright UMS and the faint PMS
311: stars found by our photometry.
312:
313: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
314: \begin{deluxetable*}{rcccccclcccc}
315: \tablewidth{0pt}
316: %\rotate
317: \tabletypesize{\scriptsize}
318: \tablecaption{Photometry of OB stars in NGC 346 by Massey et al. (1989),
319: and from our ACS analysis. Colors and spectral types are also taken from
320: Massey et al. (1989). The corresponding reddening free parameters $Q$
321: and the derived individual un-reddened colors, color excesses and
322: extinctions are also given. Star ID numbers and celestial coordinates
323: (in J2000) are from our catalog published in Paper I. Units of
324: declination are degrees, arcminutes, and arcseconds; units of right
325: ascension are hours, minutes and seconds.\label{tab1}}
326: \tablehead{
327: \colhead{No} &
328: \colhead{RA} &
329: \colhead{DEC} &
330: \colhead{$V_{\rm ACS}$} &
331: \colhead{$V_{\rm CTIO}$} &
332: \colhead{($U-B$)$_{\rm CTIO}$} &
333: \colhead{($B-V$)$_{\rm CTIO}$} &
334: \colhead{Sp. Type} &
335: \colhead{$Q$} &
336: \colhead{($B-V$)$_{\rm 0}$} &
337: \colhead{$E(B-V)$} &
338: \colhead{$A_{\rm V}$}
339: }
340: \startdata
341: 2&00:59:04.46 &$-$72:10:24.49&12.44&12.57&$-$1.09&$-$0.20&O5.5If &$-$0.95&$-$0.31&0.11&0.37\\
342: 5&00:59:00.72 &$-$72:10:27.91&13.52&13.50&$-$1.12&$-$0.23&O3V((f*)) &$-$0.95&$-$0.32&0.09&0.28\\
343: 6&00:59:00.01 &$-$72:10:37.67&13.69&13.66&$-$1.05&$-$0.23&O5.5V &$-$0.88&$-$0.29&0.06&0.20\\
344: 9&00:58:57.36 &$-$72:10:33.38&13.98&14.02&$-$1.05&$-$0.24&O4V((f)) &$-$0.88&$-$0.29&0.05&0.16\\
345: 10&00:59:01.77 &$-$72:10:30.94&14.38&14.18&$-$1.08&$-$0.23&O5.5V((f$+$))&$-$0.91&$-$0.30&0.07&0.24\\
346: 13&00:59:06.71 &$-$72:10:41.02&14.65&14.53&$-$1.01&$-$0.22&O6.5V &$-$0.85&$-$0.28&0.06&0.20\\
347: 14&00:59:06.31 &$-$72:09:55.84&14.37&14.26&$-$0.91&$-$0.10&B0.5V &$-$0.84&$-$0.28&0.18&0.57\\
348: 15&00:59:02.88 &$-$72:10:34.68&14.45&14.39&$-$1.08&$-$0.22&O7V &$-$0.92&$-$0.31&0.09&0.28\\
349: 18&00:59:14.50 &$-$72:11:59.53&14.97&14.98&$-$0.90&$-$0.16&B0V &$-$0.78&$-$0.26&0.10&0.32\\
350: 19&00:59:28.75 &$-$72:10:16.28&15.09&15.14&$-$0.95&$-$0.08&B0V &$-$0.89&$-$0.30&0.22&0.69\\
351: 20&00:59:12.28 &$-$72:11:07.66&14.79&14.68&$-$0.96&$-$0.19&O7V &$-$0.82&$-$0.27&0.08&0.27\\
352: 21&00:59:12.66 &$-$72:11:08.84&14.85&14.52&$-$0.96&$-$0.17&O8V &$-$0.84&$-$0.28&0.11&0.35\\
353: 24&00:59:18.58 &$-$72:11:09.64&14.96&14.87&$-$1.02&$-$0.24&O8V &$-$0.85&$-$0.28&0.04&0.13\\
354: 26&00:59:07.29 &$-$72:10:25.10&15.09&15.13&$-$0.95&$-$0.30&O8V &$-$0.73&$-$0.24&0.00&0.00\\
355: 27&00:59:05.86 &$-$72:10:50.09&15.01&15.10&$-$1.00&$-$0.31&O8V &$-$0.78&$-$0.26&0.00&0.00\\
356: 29&00:59:01.86 &$-$72:10:43.07&15.03&15.03&$-$1.00&$-$0.28&O9.5V &$-$0.80&$-$0.27&0.00&0.00\\
357: 30&00:59:12.78 &$-$72:10:52.14&15.18&15.00&$-$0.93&$-$0.21&O9.5V &$-$0.78&$-$0.26&0.05&0.16\\
358: 31&00:59:15.48 &$-$72:11:11.44&14.98&14.82&$-$0.98&$-$0.15&O6V &$-$0.87&$-$0.29&0.14&0.45\\
359: 32&00:59:09.80 &$-$72:10:58.76&15.25&15.26&$-$0.96&$-$0.22&O8V &$-$0.80&$-$0.27&0.05&0.15\\
360: 33&00:59:05.97 &$-$72:10:44.69&15.20&15.04&$-$0.96&$-$0.21&B0V &$-$0.81&$-$0.27&0.06&0.19\\
361: 37&00:59:11.61 &$-$72:09:57.31&15.23&14.96&$-$1.01&$-$0.16&O5.5V &$-$0.89&$-$0.30&0.14&0.44\\
362: 38&00:58:58.74 &$-$72:10:51.10&15.21&15.20&$-$0.97&$-$0.27&O7.5V &$-$0.78&$-$0.26&0.00&0.00\\
363: 39&00:58:58.84 &$-$72:10:38.57&15.20&15.15&$-$0.97&$-$0.17&B0V &$-$0.85&$-$0.28&0.11&0.36\cr
364: 42&00:59:05.17 &$-$72:10:38.24&15.20&15.20&$-$1.04&$-$0.18&B0V &$-$0.91&$-$0.30&0.12&0.39\cr
365: 44&00:59:07.59 &$-$72:10:48.07&15.41&15.33&$-$0.95&$-$0.24&O6V &$-$0.78&$-$0.26&0.02&0.06\cr
366: 48&00:59:04.76 &$-$72:11:02.69&15.39&15.33&$-$0.87&$-$0.18&O7.5V &$-$0.74&$-$0.25&0.07&0.21\cr
367: 53&00:59:11.88 &$-$72:10:55.56&15.68&15.62&$-$0.99&$-$0.20&B0V &$-$0.85&$-$0.28&0.08&0.26\cr
368: 54&00:59:00.92 &$-$72:11:09.02&15.60&15.50&$-$1.03&$-$0.26&O6.5V &$-$0.84&$-$0.28&0.02&0.06\cr
369: 57&00:59:08.65 &$-$72:10:13.87&15.62&15.50&$-$1.02&$-$0.28&O7V &$-$0.82&$-$0.27&0.00&0.00\cr
370: 65&00:59:02.82 &$-$72:10:37.20&15.63&15.53&$-$1.02&$-$0.22&B0V &$-$0.86&$-$0.29&0.07&0.21\cr
371: 74&00:58:48.90 &$-$72:09:51.84&15.86&15.85&$-$0.91&$-$0.16&B0 &$-$0.79&$-$0.26&0.10&0.33
372: \enddata
373: \end{deluxetable*}
374: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
375:
376: \subsection{Completeness}
377:
378: The photometric completeness is defined as a function of magnitude for
379: both $V$ and $I$ band by artificial star experiments described in Paper
380: I. In Fig. 2 of Paper I the ``overall'' completeness functions for the
381: whole observed stellar sample are shown. Here, we show the completeness
382: factors for the region of the association (Fig.~\ref{fig-cmp}), which
383: corresponds to the CMD of Fig.~\ref{fig-cmdsys}. A comparison of the
384: completeness in the area of the association with the overall
385: completeness of the entire region (Paper I) and the one in the area of
386: the cluster BS 90 (Rochau et al. 2007) shows that the association NGC
387: 346 has the lowest completeness. This is not due to higher crowding,
388: since the area of BS 90 is much more crowded, but most probably due to
389: the higher extinction related to this stellar system.
390:
391: \subsection{Comparison with previous catalogs \label{massey-compare}}
392:
393: Our photometry also includes the brightest stars in NGC~346. Since the
394: bright main-sequence stars in NGC~346 have been photometrically and
395: spectroscopically studied in detail by Massey et al. (1989), it is worth
396: comparing our results with the ones from these authors. We found, thus,
397: for each star in the Massey et al. (1989) catalog its corresponding record
398: in our catalog. We identified 31 out of a total of 33 bright OB stars for
399: which Massey et al. performed spectroscopy. Our photometry is in excellent
400: agreement with the photometry of these authors, in the sense that their
401: differences are negligible. This is demonstrated in Fig.~\ref{fig-phtcmp},
402: where the $V$ magnitudes from CTIO 0.9-m Telescope photometry by Massey et
403: al. and our $V$ magnitudes from ACS for the 31 common stars are compared
404: (see also Table \ref{tab1}).
405:
406: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
407: \begin{figure}[]
408: \epsscale{1.15}
409: \plotone{f3.eps}
410: \caption{The $V-I$, $V$ Color-Magnitude Diagram (CMD) of the association
411: NGC~346 from HST/ACS photometry. A region within a radius of
412: 0\farcm6 is selected as the best representation of the stellar
413: population, which is characterized by two features in the CMD:
414: the upper main sequence, with 33 spectroscopically confirmed OB stars,
415: and the lower mass Pre-Main Sequence. An isochrone model for an age
416: of $\approx$4\,Myr Girardi et al.\ (2002) for $E(B-V)=0.08$ mag and a
417: distance modulus $m-M=$ 18.9 mag is overplotted, and confirms
418: the young age of the association.}
419: \label{fig-cmdsys}
420: \end{figure}
421: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
422:
423: %\clearpage
424: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
425: \begin{figure}[t!]
426: \epsscale{1.15}
427: \plotone{f4.eps}
428: \caption{Completeness factors as derived by {\tt DOLPHOT} for all V
429: and I band datasets in NGC~346 within a radius of about 0\farcm6 around
430: its center. Note that due to crowding and the presence of diffuse emission,
431: the completeness is almost one magnitude worse than when considering the
432: entire stellar sample as is shown in Fig.\ 2 in Paper I.}
433: \label{fig-cmp}
434: \end{figure}
435: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
436:
437: \subsection{Reddening in the observed area\label{sec-red}}
438:
439: In order to evaluate the reddening toward different regions of the
440: observed area, we estimated the reddening of individual stars by
441: comparing to ``un-reddened'' isochrone models. Specifically, the
442: ``youngest'' available Padova isochrone in the ACS filter system
443: (Girardi et al. 2002) was compared with main sequence stars brighter
444: than $V\simeq 22$~mag (``upper main sequence'' population in Paper I).
445: The observed position in the CMD of each main sequence star with
446: $V$~\lsim~22~mag was compared to the expected position the star would
447: have according to the Padova isochrone model of $\log{\tau}=6.6$,
448: assuming that this is the age of the star. Metallicity of $Z=0.004$ and
449: a distance modulus of $m-M=18.9$~mag, both typical values for the SMC,
450: were considered for the model. The difference between the observed
451: magnitude and color of each star and the values expected by the model
452: provide a measurement of the color excess, $E(V-I)$, and extinction,
453: $A_{\rm V}$, for the star. The produced reddening map and the
454: corresponding reddening distribution are shown in Fig.~\ref{fig-genred}.
455: This distribution peaks at $E(B-V) \simeq$~0.08~mag. The reddening seems
456: to vary from one region to the next and within the boundaries of the
457: association NGC~346 itself, as we show in the next section.
458:
459: %Considering that only OB stars can be used for an accurate estimation of
460: %the reddening in a region (Massey et al. 1989), this selection gives only
461: %a first-order approximation of how reddening behaves from one region to
462: %the next.
463:
464: \subsection{Extinction around the OB stars of NGC~346\label{sec-difred}}
465:
466: Differential reddening is important for the studies of young stellar
467: populations in bright massive stellar systems like NGC~346. In order to
468: quantify the interstellar reddening in different regions within the
469: association, we make use of previous spectroscopic studies of the
470: brightest main-sequence stars observed with our photometry. We have
471: already matched these stars with the OB stars of NGC~346 studied by
472: Massey et al. (1989) (see \S \ref{massey-compare}). Therefore, we know
473: their exact spectral types. In addition, the $UBV$ photometry of these
474: authors can be used for estimating the reddening-free parameter $Q$:
475: $$Q=(U-B)-0.72 \times (B-V)$$ Based on the spectral classification of OB
476: stars in our Galaxy by Schmidt-Kaler (1982) and Peletier (1989) and the
477: photometry of standards of such spectral types by Johnson \& Morgan
478: (1953), it is known that the ``un-reddened'' color index $(B-V)_{\rm 0}$
479: is related to $Q$ by a simple function (Binney \& Merrifield 1998): $$
480: (B-V)_{\rm 0} \simeq 0.332 Q$$ We used this relation to estimate
481: $(B-V)_{\rm 0}$ for each of our OB stars with known spectral types from
482: Massey et al. (1989). Therefore, we are able to estimate their
483: individual color excess $E(B-V)$ by comparing the observed color indices
484: $(B-V)$ to their un-reddened ones $(B-V)_{\rm 0}$, as calculated from
485: their $Q$ parameters, and consequently their individual extinction
486: $A_{\rm V}$. The extinction and reddening values found for each OB star
487: are given in Table \ref{tab1}, where it is shown that individual
488: reddening for the OB stars varies from $E(B-V) \simeq$ 0.0 to 0.2 mag
489: (0.0 \lsim\ $A_{\rm V}$ \lsim\ 0.7 mag).
490:
491: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
492: \begin{figure}[t!]
493: \epsscale{1.15}
494: \plotone{f5.ps}
495: \caption{Comparison between the $V$ magnitudes from our photometry with
496: ACS and the photometry with the 0.9-m CTIO Telescope for the 31 brightest
497: OB stars in NGC~346, for which Massey et al. determined accurate spectral
498: types. The drawn line corresponds to the line of excellent coincidence.}
499: \label{fig-phtcmp}
500: \end{figure}
501: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
502:
503: The OB stars are centrally concentrated in the association NGC~346
504: itself, as shown from their spatial distribution in Fig.~\ref{fig-cms}
505: (right; their positions are indicated by the open boxes), and therefore
506: a comparison between the spatial distribution of their reddening and the
507: one derived from all upper main sequence stars for the whole observed
508: region is not possible. However, typical reddening values derived for
509: the OB stars in the area of the association are found to be
510: systematically lower than those derived in the previous section from
511: upper main sequence stars in the same location. Moreover, the reddening
512: found for the OB stars follow a normal number distribution peaked at
513: $E(B-V) \simeq$~0.04~mag, which is half of the one found from the upper
514: main sequence stars of the whole region (shown in
515: Fig.~\ref{fig-genred}). From the spatial distribution of the reddening
516: derived from the OB stars, we find that $A_{\rm V}$ is higher in the
517: inner parts of the association with values between $A_{\rm V} \simeq$
518: 0.2 and 0.4 mag ($E(B-V) \simeq$ 0.06 - 0.12 mag). A possible
519: explanation for the lower reddening toward the OB stars is that probably
520: strong winds from these stars have blown away the gas around them, and
521: therefore reddening appears lower in their vicinities. Considering that
522: the reddening estimation in the previous section is based on a larger
523: statistical sample of stars than the one presented here only for the OB
524: stars, in the subsequent analysis we assume a mean interstellar
525: reddening of $E(B-V) \simeq$~0.08~mag, a result which also agrees with
526: the one derived for the same region by Sabbi et al. (2007).
527:
528: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
529: \begin{figure*}[]
530: \epsscale{1.15}
531: \plottwo{f6a.ps}{f6b.ps}
532: \caption{{\em Left}: Reddening Map around NGC~346. This map is based on the
533: upper main sequence stars
534: down to $V = 22.15$~mag, as they were selected in Paper I. The reddening
535: for each star was estimated by comparison to a
536: Padova isochrone for an age of $\approx$4\,Myr (Girardi et al. 2002). Darker
537: regions, indicating high reddening, are apparent in the general region of
538: the association and along the related molecular filaments of the N~66
539: ``bar''. The white area to the north corresponds to the cluster BS~90,
540: where no upper main sequence stars are found (Rochau et al. 2007), and
541: therefore there is no reddening measurement for this area. This map also
542: demonstrates the patchy behavior of the reddening further away from the
543: association. {\em Right}: The reddening distribution from the upper main
544: sequence stars, as it is estimated from comparison with isochrone
545: models. The peak corresponds to color excess $E(B-V)\simeq0.08$ mag.}
546: \label{fig-genred}
547: \end{figure*}
548: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
549:
550: \section{Pre-Main Sequence Stars in NGC~346 \label{sec-pms}}
551:
552: The ages of low-mass PMS stars are usually determined from their
553: location in the CMD by comparison to theoretical isochrones (e.g.
554: Brice\~{n}o et al. 2007). Quite often, a widening of the observed PMS
555: in the CMD gives evidence for an age spread (Palla \& Stahler, 2000).
556: However, this spread could also occur because of several other factors
557: which can cause considerable deviations of an individual star's position
558: in the CMD from the location predicted by theoretical models. Such
559: factors include variability, binarity, and the distribution of distances
560: along the line of sight. According to Shu's et al. (1987) picture of
561: star formation a typical galactic molecular cloud could sustain star
562: formation for a timescale on the order of the diffusion time,
563: $\sim10^{7}$~yrs, and an observable consequence would be an age spread
564: in star forming regions, with age of PMS stars comparable to this
565: timescale. Still, as Brice\~{n}o et al. (2007) reports, no evidence for
566: a wide age spread has been found in the PMS populations of $\sigma$~Ori,
567: Ori OB1, $\lambda$~Ori and Upper Sco in our Galaxy. These authors
568: conclude that the data from these different star forming regions support
569: the hypothesis that star formation occurs fast and synchronized in
570: molecular clouds. {\em Do we observe the same behavior in star-forming
571: regions of the MCs like NGC~346/N~66 based on our ACS photometry of its
572: PMS population?} In the following paragraphs we explore this question.
573:
574: \subsection{Location of PMS Stars in the CMD \label{sec-pmsloc}}
575:
576: In the CMD a well populated faint red sequence of PMS stars is present
577: (Fig.\ \ref{fig-cmdsys}). The location of the PMS stars in NGC~346 is in
578: good agreement with the PMS location identified in LH~52 in the LMC
579: (Gouliermis et al.\ 2006a). The faintest red part of the CMD of NGC~346,
580: where these stars are located, is shown in Fig.~\ref{fig-multcmd}. The
581: whole stellar sample is plotted in this CMD, with the points corresponding
582: to the stars confined within the region of the association ($\sim$
583: 0\farcm6 around the center of NGC~346) being plotted with thicker symbols.
584: In the left panel the tentative limits that were selected in Paper I to
585: distinguish the PMS stars from other stellar types are drawn, while in the
586: right panel PMS isochrone models for ages between 0.5 and 15 Myr from
587: Siess et al.\ (2000) are overplotted. We assume a distance modulus of $m-M
588: =18.9$ mag and, although typical value for the metallicity of the SMC is
589: $Z=0.004$, we use PMS models for metallicity of $Z=0.01$, because it is
590: the lowest available in the Siess et al. grid of models. A mean reddening
591: of $E(B-V)=0.08$~mag (\S~\ref{sec-red}) was assumed.
592:
593: %\clearpage
594: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
595: \begin{figure*}[]
596: \epsscale{1.}
597: \plotone{f7.ps}
598: \caption{$V-I$, $V$ Color-Magnitude Diagram (CMD) of all
599: stars detected in the area of NGC~346 with ACS imaging, centered on the
600: region of the observed PMS stars. The plot in the left panel shows the
601: area selected in Paper I as the best representative of these
602: stars. In the right panel PMS isochrones by Siess et al. (2000) for ages
603: of 0.5, 0.75, 1.0, 1.5, 2.5, 5, 10 and 15 Myr are overplotted. In both plots the
604: stars confined in an area with radius of about 0\farcm6 around the
605: center of the association are shown with thicker symbols. The ZAMS is
606: drawn as a solid line. {\tt DOLPHOT} fitting errors in color and
607: magnitude are also shown. The PMS models predict PMS stars well above
608: the limit of $V\simeq$ 22 mag, but the contamination by the background
609: field and mostly by the close-by intermediate-age cluster BS~90 does not
610: allow us to distinguish the PMS stars from the evolved ones for this
611: magnitude range. The wide range of the isochrones which fit the loci of
612: the PMS stars suggests multi-epoch star
613: formation. However the observed broadening of the PMS stars in the CMD
614: might also be the result of several other factors, most prominently among them
615: differential reddening (see text in \S 3).}
616: \label{fig-multcmd}
617: \end{figure*}
618: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
619:
620:
621: \subsection{Broadening of the PMS stellar sequence \label{sc-pmsbroad}}
622:
623: It has previously been suggested that the PMS stars of NGC~346 represent
624: a single star formation event that took place 3 - 5 Myr ago (Nota et al.
625: 2006). However, a widening of the sequence of PMS stars is apparent in
626: the CMD of Fig.~\ref{fig-multcmd}. Comparison with theoretical
627: isochrones indicates that it could correspond to an age spread of
628: $\approx$15 Myr (Fig.~\ref{fig-multcmd}, right panel). This suggests
629: that the stars may not be the product of a single star formation event,
630: but instead star formation lasted much longer in NGC 346. In the
631: following, we test this hypothesis and discuss the observational
632: limitations and physical effects, which apart from an age spread can
633: produce the observed broadening of the PMS distribution. Specifically,
634: we analyze how photometric uncertainties, stellar variability,
635: unresolved binarity and variable extinction affect an observed sequence
636: of PMS stars. The individual effect of each of these factors, but also
637: their cumulative effect on the CMD is discussed.
638:
639: \subsubsection{Photometry errors}
640:
641: Typical {\tt DOLPHOT} photometry errors, derived from the PSF fitting,
642: are shown on the right panel of Fig.~\ref{fig-multcmd} to illuminate the
643: dependency of the color uncertainties on the brightness of the detected
644: PMS stars. The errors derived from the incompleteness are significantly
645: lower than the photometry errors, and therefore are not considered here.
646: A reasonable explanation for the observed widening of the distribution
647: of the PMS stars could be their photometric errors. However, the
648: observed broadening is wider than the mean photometric errors per
649: magnitude range. For stars with $V$ \lapprox\ 23.5 mag the sequence of
650: PMS stars is found to cover a color range of at least 5$\sigma$ or more
651: ($\sigma$ being the typical color uncertainty), while for fainter stars
652: with 23.5 \lapprox\ $V$ \lapprox\ 25.5 mag, this range becomes at least
653: 3$\sigma$ wide. This implies that the photometric errors alone cannot
654: account for the observed spread in the whole magnitude range of the PMS
655: stars.
656:
657: \subsubsection{Binarity}
658:
659: Another factor expected to move PMS stars away from the colors and
660: magnitudes predicted by models is their binarity. Using the simulations
661: of the PMS population with $24\leq V \leq 27$~mag and $1.0 \leq V-I \leq
662: 2.2$~mag of $\sigma$ Ori by Sherry et al. (2004), we expect at the
663: distance of NGC~346 ($m-M \simeq 18.9$~mag) a brightness spread of
664: $\Delta V\simeq$ 0.75 mag. The difference in magnitudes between the PMS
665: models for 1.5 and 15 Myr, for the same magnitude and color limits, is
666: $\Delta V\simeq$ 2.5 - 1.6 mag, at least twice as large as the spread
667: produced by binarity.
668:
669: %Distance modulus for sigma Ori is 8.2 mag
670:
671: \subsubsection{Variability}
672:
673: It is almost certain that the PMS stars of NGC~346 are T~Tauri stars, as
674: can be seen from a comparison of their loci in the CMD with the location
675: of T~Tauri stars observed, for example, in the Orion OB1 association in
676: the Milky Way (Brice\~{n}o et al. 2005). These stars are known to
677: exhibit optical variability, that could cause the PMS spread observed in
678: our CMD. The variability in the $V$-band for Classical T~Tauri stars
679: (CTTSs) shows an amplitude up to 3 mag, while for Weak-line T~Tauri
680: stars (WTTSs) the variability is between 0.05 and 0.6 mag (Herbst et al.
681: 1994). In young (3 - 5 Myr) star forming regions the ordinary fraction
682: for CTTSs is 30\% - 50\% (Preibisch \& Zinnecker, 1999), but this
683: fraction decreases with age so that by an age of a few Myr most low-mass
684: PMS stars are WTTSs (Sherry et al. 2004). If we assume that this region
685: is indeed very young, then we should consider that no more than 50\% of
686: the T~Tauri stars in our sample are CTTSs. These PMS stars are
687: characterized by irregular photometric variability (e.g. Sherry 2003)
688: and therefore it is quite difficult to estimate the probability of
689: observing them near the extreme limit of their variability. On the other
690: hand, the photometric variability of these objects is extremely rapid
691: (e.g. Sherry 2003) and consequently the probability of observing its
692: extreme limit is very low. Under these circumstances, a percentage of
693: about 50\% of CTTSs in our sample can only partly account for the
694: observed spread. If we consider that the region is older, as suggested
695: by Massey et al. (1989; $\sim$~12~Myr), then the probability that the
696: observed widening due to variability is even smaller, since these stars
697: would be mostly WTTSs.
698:
699: \subsubsection{Reddening}
700:
701: Differential reddening is probably the most important factor that might
702: produce the observed broadening. In \S~2, we derive the individual
703: interstellar reddening for each of the OB stars within the association.
704: This leads to an estimation of their color excess $E(B-V)$ and
705: consequently their individual extinction $A_{\rm V}$ from their
706: un-reddened color indices $(B-V)_{\rm 0}$, and reddening-free $Q$
707: parameters (\S~\ref{sec-difred}; Table \ref{tab1}). We find that
708: extinction varies from $A_{\rm V} \simeq$ 0.06 to 0.39 mag (with a peak
709: at $A_{\rm V} \sim$ 0.13 mag). Such an extinction-spread corresponds to
710: a spread in color excess between $E(V-I) \simeq$ 0.03 and 0.20 mag,
711: which is much smaller than the observed widening in color of the
712: sequence of PMS stars. On the other hand, if we consider the reddening
713: distribution based on the upper MS stars of the entire observed region
714: (Fig.~\ref{fig-genred}; $\langle E(B-V)\rangle \sim 0.08$~mag), then
715: reddening becomes a very important factor for the observed PMS
716: broadening. A further support to the crucial role of reddening is given
717: by the comparison of the spatial variations in the locus of the PMS
718: stars with the reddening map of the observed region derived in
719: \S~\ref{sec-red} (shown in Fig.\ref{fig-genred} left) and the spatial
720: variations of the reddening of the central 31 OB stars
721: (\S~\ref{sec-difred}). This comparison shows that indeed the PMS stars
722: located in the more reddened areas seem to occupy a preferentially
723: redder locus in the CMD. The importance of reddening is demonstrated
724: below with the use of simple simulations of the observed loci of PMS
725: stars in the CMD.
726:
727: %$E(B-V) \simeq$ 0.02 and 0.12 mag and therefore 0.03 \lapprox\ $E(V-I)$
728: %\lapprox\ 0.20 mag (assuming $R_{\rm V} = 3.2$ and since $E(V-I)=1.6
729: %E(B-V)$; Mihalas \& Binney 1981)
730:
731:
732: \subsection{Simulation of the Observed PMS Broadening
733: \label{sec-simbroad}}
734:
735: We discussed above several factors that may explain the observed PMS
736: spread in NGC~346, and we find that none of them can actually produce
737: this spread in the observed degree {\em if it acts alone}. However, one
738: should consider that these factors are not isolated phenomena, but, on
739: the contrary, they are expected to act on the observed stellar
740: parameters in a cumulative manner. Taking this into account, we simulate
741: the observed widening of the PMS assuming that the ``suspected'' factors
742: act in an additive way on a sequence of single-age PMS stars. We
743: constructed a toy model of the observed PMS stars, assuming that all
744: stars are the product of a single star formation event. Then, we applied
745: each of the aforementioned biases. We conclude from this test that a
746: single-aged PMS population can indeed be misinterpreted as the product
747: of star-formation events that took place during different periods. The
748: apparent age spread can be the product of the cumulative action of
749: several factors.
750:
751: An example of the results of our toy model is given in
752: Fig.~\ref{fig-sim}, where we assume that the observed PMS population of
753: NGC 346 is 4 Myr old. First, we assume a concentration of PMS stars, as
754: they are selected in Paper I according to their location in the CMD
755: (Fig.~\ref{fig-sim}{\em a}), which populates a single 4~Myr PMS
756: isochrone (Fig.~\ref{fig-sim}{\em b}). Then we apply the effects of the
757: suspected factors one on top of the other. Here we describe the
758: procedure:
759:
760: \begin{itemize}
761:
762: \item[(i)] {\em Binarity}: As far as the existence of unresolved
763: binaries in our sample concerns, there are no studies on the expected
764: fraction of binary systems in associations of the Magellanic Clouds.
765: Therefore, we based our simulations on the information we have on
766: galactic OB associations. These systems show a binary fraction that does
767: not differ significantly from the field (Mathieu 1994), which is found
768: for G dwarfs to be around $\sim$ 58 \% (Duquennoy \& Mayor 1991). The
769: well-studied Orion Nebular Cluster (ONC) is also found to be consistent
770: with the field binary fraction (e.g. Prosser et al. 1994; Petr et al.
771: 1998; K{\"o}hler et al. 2006). In many clusters, a clear binary sequence
772: is observed to lie 0.75 magnitudes above the single star main sequence
773: (Tout 1991), and this is almost equal to the brightness shift that a
774: single star will suffer toward brighter magnitudes, if it is an
775: unresolved binary system with equal mass components (de Bruijne et al.
776: 2001).
777:
778: Simulations of unresolved low-mass PMS stars in the Orion OB1b association
779: centered on $\sigma$ Ori by Sherry (2003) showed that for a 3 Myr old
780: binary system with a 0.5 M{\solar} primary the position of the system on
781: the $V-I$, $V$ CMD can be shifted up to 0.13 mag on the $V-I$ axis. The
782: actual shift depends on the mass of the lower-mass companion, which ranges
783: from 0.1 to 0.5 M{\solar}. This system also brightens by slightly less
784: than 0.75 magnitudes. It should be noted that recent observational studies
785: on the mass ratios of T~Tauri binaries are consistent with a uniform
786: distribution of mass ratios between 0.2 and 1.0 (Woitas et al. 2001). In
787: order to simulate the effect of unresolved binaries to the modeled
788: sequence of 4 Myr old PMS stars we randomly applied a brightening of no
789: more than $V \simeq$ 1 mag and a reddening \lapprox\ 0.13 mag, assuming
790: different binary fractions. It is worth noting that Elson et al. (1998)
791: measured a binary fraction in the young rich LMC cluster NGC~1818, and
792: found $35 \pm 5$\% in the core. This is roughly comparable to the Galactic
793: values which are considered here. In Fig.~\ref{fig-sim}{\em c} (top panel)
794: the result for a binary fraction of 50\% is shown. This result clearly
795: suggests that binarity can play an important role to the broadening of a
796: single-age PMS population in the $V-I$, $V$ CMD.
797:
798: \item[(ii)] {\em Variability:} In order to simulate the effect of
799: variability on our modeled sequence of PMS stars, we assigned to each
800: star a random amplitude drawn from a Gaussian probability distribution,
801: following Sherry (2003), who assumed normally distributed $V$-band
802: variations in WTTSs with a mean of $M_{V} = 0.15$ mag and a standard
803: deviation of $\sigma =$ 0.1 mag, based on data for 42 WTTSs with known
804: $V$-band variations. The $V-I$ colors of these stars should appear
805: redder when they are fainter, due to rotational modulation. According to
806: Herbst et al. (1994) the ratio of color variations, $\Delta I/\Delta V$,
807: ranges from 0.39 to 0.88 mag. We assumed several values for this ratio
808: within these limits. In Fig.~\ref{fig-sim}{\em d} (top panel) we show
809: the results assuming a mean value of $\Delta I = 0.6 \Delta V$ and
810: consequently $\Delta(V-I)=0.4 \Delta V$. It should be noted that for
811: these results we assumed that only WTTSs are present at the assumed age
812: of NGC~346. This is certainly the reason for a rather small effect on
813: the variability in our simulations. In the case that some fraction of
814: CTTSs had also been included, the spread implied by the model could be
815: as large as 5 times more, due to the higher amplitude these stars
816: exhibit.
817:
818: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
819: \begin{figure*}[]
820: \epsscale{.75}
821: \plotone{f8a.ps}
822: \plotone{f8b.ps}
823: \caption{{\em Top}: An example of the results from our toy model used to
824: simulate the cumulative effect of several factors on a single-age sequence
825: of PMS stars. (a) The observed PMS stars in the region of NGC~346 are
826: selected according their positions in the $V-I$, $V$ CMD. (b) Their
827: positions are modeled based on the 4~Myr PMS isochrone from Siess et al.
828: (2000). Factors, such as (c) binarity, (d) variability, (e) reddening and
829: (f) photometric uncertainties are applied in order to explore the
830: broadening they cause to a single-age PMS population, making it appearing
831: as the result of multi-epoch star formation (see \S~\ref{sec-simbroad}).
832: {\em Bottom}: Results from a toy model similar to the one on the top, but
833: assuming that the PMS stars are formed during two star formation events,
834: one 10~Myr and the other 4~Myr ago. In this example 50\% of the population
835: is assumed to be formed during the first and 50\% during the second
836: event. These results imply that the possibility of more than one star
837: formation events in NGC~346/N~66 cannot be ruled out.}
838: \label{fig-sim}
839: \end{figure*}
840: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
841:
842: \item[(iii)] {\em Reddening}: We explore the displacement of the PMS stars
843: from their theoretical loci on the CMD due to reddening. We assume that
844: these stars are subject to a) the reddening of the whole area as estimated
845: from the upper main sequence stars (\S~\ref{sec-red}), by randomly
846: applying the observed reddening distribution shown in Fig.
847: \ref{fig-genred} ({\em right}), and b) the much lower reddening of the
848: area of NGC~346 alone found from the OB stars (\S~\ref{sec-difred}). For
849: Fig. \ref{fig-sim}{\em e} (top panel), the assumed reddening distribution
850: was based on stars in the whole region with $V$ \lsim\ 21 mag, and is
851: chosen to show the maximum effect. The corresponding distribution for even
852: brighter stars is thinner with smaller mean values and standard
853: deviations. This would also be the result for the (more accurately
854: measured) reddening found for OB stars. As shown in Fig. \ref{fig-sim}{\em
855: e}, high extinction is indeed the most important factor, in addition to
856: binarity, that can cause the observed broadening of single-age PMS stars.
857:
858: \item[(iv)] {\em Photometry}: Photometric accuracy is a function of
859: magnitude, being higher for the brighter stars. In order to explore the
860: effect of photometric uncertainties to a ``perfect'' sequence of PMS
861: stars, we randomly applied the observed mean photometric errors per
862: magnitude (shown in Fig.~\ref{fig-multcmd} {\em right}) to the modeled PMS
863: population. Naturally, the derived broadening is wider for the fainter
864: stars. This is shown in Fig.~\ref{fig-sim}{\em f} (top panel).
865:
866: \end{itemize}
867:
868: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
869: \begin{figure*}[]
870: \epsscale{.75}
871: \plotone{f9.eps}
872: \caption{{\em Top}: A 1\farcm5 $\times$ 1\farcm2 region centered on the
873: association NGC~346, with the borders of 20 selected sub-regions
874: 18\arcsec$\times$18\arcsec\ wide each overplotted. {\em Bottom}: The
875: CMDs of the sub-regions, with PMS isochrones overplotted for sub-regions
876: 11 to 15. PMS stars are present in almost the entire region,
877: and they seem to be aligned across the nebula. This plot shows that the
878: CMD varies in number of PMS stars and width of their broadening from one
879: region to the other.}
880: \label{fig-cntr346}
881: \end{figure*}
882: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
883:
884: From the results of our toy model for the influence of several factors
885: on a ``perfect'' single-age sequence of PMS stars, it can be seen that
886: indeed a single-age PMS population may appear broad enough in the $V-I$,
887: $V$ CMD so that it could be misinterpreted as a group of stars, produced
888: by multi-epoch star formation. We specifically note the importance of
889: reddening and binarity of such stars for this broadening. Variability
890: may also play an important role depending on the nature of these stars.
891: Although each of these factors acting alone cannot cause the observed
892: widening of PMS loci in the CMD, it is their cumulative action that
893: leads to the broadening of the sequence of PMS stars in NGC 346.
894: However, it should be noted that if indeed reddening is the most
895: important cause for this widening, then the PMS population of NGC~346
896: should be expected to be older than 3 - 5 Myr, since reddening
897: ``pushes'' PMS stars to redder colors making them appear {\em younger}.
898: Specifically, PMS models for ages on the order of 10 Myr proved to fit
899: the observed sequence of PMS stars and their broadening better. This
900: result is in line with Massey et al. (1989), who claim that although the
901: location of the bright stars of NGC~346 in the CMD implies an age of
902: less than 5~Myr for the association, there is a spatial distinct
903: subgroup of evolved $\sim$~15~M{\solar} stars {\em in the association}
904: with ages $\simeq$~12~Myr.
905:
906: Furthermore, another very interesting result of our simulations is that
907: the possibility of multi-epoch star formation cannot be excluded.
908: Specifically, our toy model showed that if we consider that the
909: interstellar reddening may be lower than what was previously assumed
910: (using e.g. $E(B-V)\simeq0.04$~mag found from the OB stars in the region
911: of the association), then the observed spread of the PMS stars can only be
912: reproduced by (at least) two separate PMS isochrones. This is demonstrated
913: in Fig.~\ref{fig-sim} (bottom panel), where we show the results of the toy
914: model if two star formation events around 4 and 10 Myr ago are assumed. In
915: this figure it is shown that the final broadening of the PMS loci in the
916: CMD does not differ significantly (although it is a bit wider) from the
917: one resulted from a single star formation event. Typical ages derived from
918: the simulations of a double star formation event (half of the stars formed
919: in each epoch) are 3 - 4 and 10 - 12 Myr in agreement with both
920: hypotheses. Conclusively, we {\em cannot} completely exclude the
921: possibility of a true age spread of the sequence of PMS stars in NGC~346
922: due to separate star formation events.
923:
924: \subsection{CMD Variations within the Region of NGC~346\label{sec-cmdvar}}
925:
926: In order to study variations in the CMD we selected a region of size
927: 1\farcm5 $\times$ 1\farcm2, centered on the association NGC~346. This
928: region was divided into 20 18\arcsec$\times$18\arcsec\ sub-regions.
929: Indeed, we see that the CMDs differ from one region to the next in the
930: nature of the stars, their numbers and the apparent PMS broadening. The
931: individual CMDs of all sub-regions are shown in Fig.~\ref{fig-cntr346}.
932: In order to avoid confusion, PMS isochrones are overplotted only for
933: sub-regions 11 to 15, so that a rough estimation of the age differences
934: can be extracted. Specifically, isochrones for 1.5 and 10 Myr are
935: displayed in the CMD of sub-region 11, the 0.75 and 15 Myr isochrones
936: are shown in the CMDs of sub-regions 12 and 13, whereas for the CMD of
937: sub-region 14 the older age limit seems to be better fit by a 10 Myr
938: isochrone, but for a very low number of faint PMS stars.
939:
940: From Fig.~\ref{fig-cntr346}, it can be seen that PMS stars are present
941: in almost the entire selected region, except for the southwestern
942: (sub-regions 5 and 10) and northeastern (sub-region 16) areas, which are
943: located away from the association (and the nebula). From the CMDs of the
944: remaining sub-regions we can conclude that PMS stars seem to follow the
945: southeast (bottom left corner) to northwest (top right corner) direction
946: of the \emph{nebulosity}, with a higher concentration in the main body
947: of the association (sub-region 13) and around it (e.g. sub-regions 12,
948: 14 and 18)\footnote{The concentration of PMS stars in sub-regions 3, 4,
949: 5, 10, 11, 15, 16 and 17 is lower than the one along the
950: southeast-northwest axis.}. Reddening in the selected region is expected
951: to be higher in the part centered on the main body of NGC~346 due to the
952: nebula, and indeed the corresponding CMDs exhibit the wider broadening
953: in their PMS populations. This suggests that reddening {\em is} indeed
954: (at least partly) responsible for the observed PMS broadening. This is
955: more clearly seen in the CMD of sub-region 13, the best representative
956: of the population of the association (similar to the CMD used in our
957: toy-model).
958:
959: From the discussion so far on the PMS stars of association NGC~346 alone
960: we {\em cannot} safely conclude that these stars represent a coherent
961: population, formed in a single star formation event. Nota et al. (2006)
962: and Sabbi et al. (2007) argue that such an event happened 3 - 5 Myr ago,
963: while Massey et al. (1989) suggest a timescale three times longer,
964: closer to our results presented in \S 3.3. However, taking into account
965: the low reddening found from the OB stars in the association, we tend to
966: conclude that a true age spread may also contribute to the observed PMS
967: broadening.
968:
969: NGC~346 does not seem to represent the only product of recent star
970: formation in the general area. As shown from the isodensity contour
971: maps of Fig. \ref{fig-cms}, density peaks of PMS stars can be seen away
972: from the association, forming small compact concentrations. In the
973: following sections we study these concentrations, which we refer to as
974: ``PMS clusters''. We also present our results from H\alp\ observations
975: and previous results from observations with the {\em Spitzer} Space
976: Telescope in the mid- and far-infrared, which reveal loci of on-going
977: star formation in the general region of NGC~346/N~66.
978:
979: \section{PMS Clusters in the region of NGC~346/N~66 \label{sec-pmsclus}}
980:
981: \subsection{The most significant PMS clusters in the Region}
982:
983: The isodensity contour map of the region of NGC~346/N66 constructed from
984: star counts of the PMS stars is shown in Fig.~\ref{fig-cms} (right
985: panel). In this map several distinct concentrations of PMS stars can be
986: identified. We selected the statistically most significant
987: concentrations with a density \gsim\ 3$\sigma$ above the background
988: (where $\sigma$ is the standard deviation of the background density).
989: There are five such concentrations, which fit the description of ``PMS
990: clusters''. The contour map of the whole region, as well as the ones of
991: the local regions around each of these clusters, numbered from north to
992: south, are shown in Fig.~\ref{fig-pmsclus} (top).
993:
994: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
995: \begin{figure*}[t!]
996: \centerline{
997: \vbox{
998: \psfig{figure=f10a.eps,width=12cm,angle=270}
999: \psfig{figure=f10b.ps,width=13cm}
1000: }
1001: }
1002: \caption{Top: Isodensity contour map from star counts of the PMS stars of the
1003: area of NGC~346/N~66, as they have been selected in Paper I. Five PMS
1004: clusters in the area around the association NGC~346 are selected based on
1005: their high density of PMS stars and the corresponding local contour maps
1006: are shown enlarged. Bottom:
1007: $V-I$, $V$ CMDs of the stars included within the area of each
1008: of the selected PMS clusters.
1009: An isochrone for an age of $\approx$4\,Myr by Girardi et
1010: al.\ (2002) is overplotted for the main sequence populations. For the PMS
1011: stars, isochrone models by Siess et al. (2000) are overplotted to
1012: indicate the time range within which these stars have presumably formed.
1013: Cluster 1 seems to be rather a more evolved open cluster, while cluster
1014: 5 meets the characteristics of the association itself. Clusters 2a, 2b
1015: and 3 probably represent the youngest products of clustered star
1016: formation away from the association NGC~346.
1017: \label{fig-pmsclus}}
1018: \end{figure*}
1019: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1020:
1021: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1022: \begin{figure}[]
1023: \epsscale{1.15}
1024: \plotone{f11.ps}
1025: \caption{Average surface density of companions as a function of angular
1026: separation from each star for the PMS clusters identified in the region
1027: of NGC~346 (Fig.~\ref{fig-pmsclus}). From our observations we cannot
1028: distinguish separations in the binary regime. Therefore, the power-law
1029: index is given only for larger separations, typically in the range
1030: 0.8$''$ to 8$''$, corresponding to physical scales of 0.24\,pc to 2.4\,pc at
1031: the distance of the SMC. The fitting functions, presented
1032: by solid lines, and the corresponding indices, $\gamma$, are given for
1033: each cluster. Their values indicate that all PMS clusters should be
1034: considered as results of a hierarchical clustering process (see section 4.2).}
1035: \label{fig-clsd}
1036: \end{figure}
1037: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1038:
1039: The size of each PMS cluster was defined by a closer inspection of their
1040: contour map, as shown in Fig.~\ref{fig-pmsclus} (top). The center was
1041: chosen to be the density peak. A circular area was selected to represent
1042: the extent of each cluster with a radius determined by the isopleth,
1043: corresponding to the mean density of the surrounding region (first
1044: isopleth in the individual maps of the clusters). This isopleth defines
1045: the size of each cluster. It should be noted that cluster 2 consists of
1046: two distinct regions, which we will treat in the following study
1047: separately. Clusters 1 and 3 are clearly detected in the contour maps of
1048: both PMS and Upper Main Sequence stars (see Fig.~\ref{fig-cms}).
1049: Therefore, their size was determined by the contour maps constructed
1050: with both PMS and UMS stars. In order to define an age range within
1051: which each PMS cluster was presumably formed we construct the CMD of the
1052: stars included in the area covered by each cluster.
1053:
1054: \subsubsection{Color-Magnitude Diagrams of the PMS Clusters}
1055:
1056: Although it is quite difficult to disentangle the actual age of the
1057: discovered PMS clusters, because of the artificial broadening of the PMS
1058: loci in the CMD discussed in \S~\ref{sec-pms}, it is worthwhile to
1059: specify the limits within which their ages may range, in order to
1060: identify any age-differences between them. Therefore, the stellar
1061: populations of the PMS clusters are analyzed in terms of their CMDs
1062: shown in Fig \ref{fig-pmsclus} (bottom). Isochrone models for both MS
1063: (Girardi et al. 2002) and PMS (Siess et al. 2000) populations are
1064: overplotted. These models suggest that cluster 1 is in a rather evolved
1065: state with the best fitting PMS model indicating an age range of 5 to 15
1066: Myr. The stellar content of this cluster, which also includes a few
1067: bright MS stars, suggests that it should be treated as a small open
1068: cluster rather than a recently formed PMS cluster.
1069:
1070: The models indicate that all the remaining PMS clusters seem to be much
1071: younger than cluster 1. Specifically, clusters 2 to 4 do not exceed an
1072: age of 5 Myr, with clusters 2a, 2b and 3 being the youngest in the sample
1073: with ages for their PMS stars younger than 2.5 Myr. All isochrone models
1074: have been plotted for a mean reddening of $E(B-V)\simeq$~0.08~mag.
1075: Differential reddening is found to play an important role in the
1076: broadening of the PMS stars in the CMD. However, a constant mean reddening
1077: should be considered for the isochrone models, and we found that the
1078: selection of this reddening according to the measurements of
1079: \S~\ref{sec-red} and \S~\ref{sec-difred} does not seem to be very
1080: important for the derived age-span of the clusters. This is demonstrated
1081: in Fig \ref{fig-pmsclus} for the CMD of cluster 2a, where we plot the PMS
1082: isochrones assuming three cases: No reddening (dash-dotted lines) and
1083: reddening values of $E(B-V)\simeq$~0.08~mag (found from the upper main
1084: sequence for the whole region; dashed lines) and $E(B-V)\simeq$~0.04~mag
1085: (found from the OB stars in the association; long-dashed lines).
1086:
1087: The spread of the PMS stars in the CMD of cluster 5 covers a wider range
1088: of ages, comparable to the one we observe in the main body of the
1089: association NGC~346 (discussed in \S 3). This is not a surprising result
1090: considering that cluster 5 is the closest PMS cluster to the association,
1091: and probably it shares the same star formation history with it. In
1092: general, from the individual CMDs of the PMS clusters (except for cluster
1093: 1, which is a rather evolved open cluster with a PMS population) we see
1094: that the northern part of the observed field (north of the association)
1095: covers younger clusters of PMS stars than in the southern part and the
1096: association itself. It will be discussed below whether this is an
1097: indication that these clusters do not share the same star formation
1098: history with NGC~346, but rather that they are the products of more recent
1099: star formation.
1100:
1101: \subsection{Cluster Properties of PMS Stars}
1102:
1103: The degree of clustering of PMS stars can be evaluated through their
1104: surface density as a function of angular separation from each star.
1105: Gomez et al. (1993) applied this technique to the study of the spatial
1106: distribution of PMS stars in the Taurus-Auriga molecular cloud, and they
1107: found that a power law form (with index $\gamma \simeq -1.2$) reproduces
1108: the overall shape of the actual PMS distribution in Taurus, indicating
1109: the presence of real clustering. Larson (1995) extended the analysis of
1110: Gomez et al. to smaller separations and he measured a power-law relation
1111: $\Sigma \propto \theta^{\gamma}$ for young star clusters, where $\Sigma$
1112: is the surface density of PMS stars, and $\theta$ the angular separation
1113: from each star. He found that the Taurus young stars exhibit
1114: self-similar or fractal clustering on the largest scales, but there is a
1115: clear break from self-similarity at a scale of about 0.04 pc which
1116: divides the regime of binary and multiple systems on smaller scales from
1117: that of true clustering on larger scales.
1118:
1119: A survey of young clusters in the Orion, Ophiuchus, Chamaeleon, Vela,
1120: and Lupus star-forming regions by Nakajima et al. (1998) showed an index
1121: for $\Sigma(\theta)$ of $\gamma \simeq -2$ at stellar separations in the
1122: binary regime, smaller than 0.1 pc, whereas at large separations (0.1 to
1123: 1.0 pc.), the power index seems to vary from region to region with $-0.8
1124: < \gamma < -0.1$. The break in the power law was found to occur on
1125: physical scales of 0.01 to 0.1 pc. Pudritz (2002) argues that this
1126: spatial range probably does not reflect the initial conditions for
1127: isolated star formation in clusters, because a star moving at $\simeq
1128: 1.0~{\rm km} {\rm s}^{-1}$ in such a cluster will move a distance of 1.0
1129: pc in a million years, thereby erasing the initial conditions. According
1130: to Larson, the power law break provides clear evidence for an intrinsic
1131: scale in the star formation process, which is found to be essentially
1132: equal to the Jeans length in typical molecular cloud cores. He also
1133: notes that there is evidence that nearly all stars are formed in binary
1134: or multiple systems, and that some of these systems are subsequently
1135: disrupted by interactions in the denser star-forming environments to
1136: produce the observed mixture of single, binary and multiple stars
1137: (Larson 1995). Goodwin et al. (2007) review the observed properties of
1138: multiple systems and find that the multiplicity of stars changes,
1139: probably due to the dynamical decay of small-N systems and/or the
1140: destruction of multiple systems within dense clusters. They note that
1141: models of the fragmentation of rotating and turbulent molecular cores
1142: almost always produce multiple systems, but the properties of those
1143: systems do not match the observations.
1144:
1145: OB associations are known to contain subgroups of smaller clusters
1146: (Blaauw 1983), which often contain smaller groups of stars themselves,
1147: made of multiple-star systems. Efremov \& Elmegreen (1998) suggest that
1148: star formation is generally clumped into a hierarchy of structures, from
1149: small multiple systems to giant star complexes and beyond and that OB
1150: associations are just one level in this hierarchy. Even the ISM has been
1151: observed to be arranged in a hierarchical structure (Scalo 1985;
1152: V{\'a}zquez-Semadeni 2004). Turbulent motions have been proposed as a
1153: cause for the hierarchical structure (Elmegreen 1998). The velocity
1154: structure in turbulent regions is suggested to be scale-free, meaning
1155: that {\em the same physical processes and the same velocity-separation
1156: relations occur over a wide range of absolute scales} (Elmegreen 2006).
1157: According to this hypothesis, large clumps originate from the large
1158: velocities which compress the gas on large scales and small clumps
1159: result from the compression of the gas characterized by small velocities
1160: on small scales. Considering that stars are formed by the collapse of
1161: such hierarchical structured clumps, they are themselves hierarchically
1162: structured. The break in the power-law relation $\Sigma \propto
1163: \theta^{\gamma}$ defines the two different regimes of star formation:
1164: The binary regime at smaller scales, and the regime of hierarchical
1165: clustering at larger ones.
1166:
1167: In order to determine the clustering properties of the PMS clusters
1168: found in the region of NGC~346, we apply the same technique and
1169: construct the average surface density of companions $\Sigma(\theta)$ for
1170: every PMS member of each of the PMS clusters as a function of the
1171: angular projected separation $\theta$ (see Fig.~\ref{fig-clsd}). We
1172: apply a linear fit for the estimation of the power law index, $\gamma$.
1173: The function $\Sigma(\theta)$ drops sharply near the largest separations
1174: owing to the size of the region selected for each cluster. Consequently,
1175: our analysis is based only on the smaller observed separations. In
1176: addition our observations do not allow us to separate stars in the
1177: binary regime and, therefore, our study does not include them. The
1178: errors shown in Fig.~\ref{fig-clsd} reflect the low number statistics in
1179: our samples of PMS stars per cluster.
1180:
1181: The derived values of the power-law index for all clusters are found to
1182: be in the narrow range $-0.4 < \gamma < -0.2$, comparable to the results
1183: of Nakajima et al. (1998) for separations in the regime of hierarchical
1184: clustering. The power-law index relates to the nearest-neighbor
1185: distance, in the sense that if the latter can be fitted by the Poisson
1186: distribution, then the power index is close to 0. The distribution of
1187: the nearest-neighbor distance is much broader than the Poisson one, when
1188: the surface density varies appreciably within the region. Then
1189: $\Sigma(\theta)$ has a steep $\theta$-dependence, and the power-law
1190: index is negative (Nakajima et al. 1998), comparable to the values we
1191: find for our clusters. Consequently, these values can be interpreted as
1192: the result of the variation of the surface density within the region,
1193: and clearly suggest that the PMS clusters identified in the region of
1194: NGC~346/N~66 are the products of hierarchical clustering.
1195:
1196: %Both the self-similar clustering of the Taurus stars on the larger scales
1197: %and the power-law form of the upper stellar IMF may have their origin in
1198: %hierarchical, and perhaps fractal-like, cloud structure.
1199:
1200: %\clearpage
1201: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1202: \begin{figure*}[t!]
1203: \epsscale{1.15}
1204: \plottwo{f12a.ps}{f12b.ps}
1205: %\hbox{
1206: %\psfig{figure=f11a.ps,width=8cm}
1207: %\psfig{figure=f11b.ps,width=8cm}
1208: %}
1209: \caption{{\em Left}: In a $R-{\rm H}\alpha$, $V-I$ color-color diagram
1210: stars without H\alp\ emission form an almost horizontal band around $R-{\rm
1211: H}\alpha \simeq 0.0$ mag, with most of the low main-sequence stars
1212: concentrated around $V-I \simeq 0.5$ mag. The spread of normal MS
1213: stars in $R-{\rm H}\alpha$ is mostly due
1214: to the photometric errors in the narrow-band H\alp\ magnitudes. The
1215: group of cooler stars with $R-{\rm H}\alpha \geq 0.4$ mag are mostly
1216: faint PMS stars, which exhibit H\alp\ excess. Few bright stars with
1217: strong H\alp\ emission, which may be Herbig Ae/Be stars, or Be type stars,
1218: are also found. {\em Right}: The $V-I$, $V$ CMD of all
1219: emission line sources (thick symbols) overlayed on the CMD of the whole
1220: sample of stars. Both bright and faint PMS
1221: stars with H\alp\ excess can be readily identified, but the
1222: nature of the stars located on the main sequence just below the turn-off
1223: is still unclear. The rather bright detection limit for the emission
1224: stars is due to the low sensitivity of the H\alp\ filter.\label{fig-haemit}}
1225: \end{figure*}
1226: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1227:
1228: \section{H\alp\ Observations of the NGC~346/N~66 Region\label{sec-ha}}
1229:
1230: \subsection{Stars with H\alp\ excess}
1231:
1232: In total, more than 23,000 stars are detected with our photometry in all
1233: three $V$, $I$ and H\alp\ bands with photometric uncertainties less than
1234: 0.1 mag. By comparing the H\alp\ and $R$ magnitudes of the detected
1235: sources, one can identify stars with H\alp\ emission (e.g. Grebel et al.
1236: 1992, Keller et al. 1999). As our dataset is lacking observations in
1237: $R$, we determined a ``synthetic''$R$ magnitude for every star derived
1238: by accurate interpolation between $V$-, $R$- and $I$-band data for the
1239: filter set of ACS, as ${\rm m}_{675}=0.38{\rm m}_{555}+0.62{\rm
1240: m}_{814}$ (De Marchi, {\em private communication}). We identified the
1241: H\alp\ emission stars in our sample by plotting the $R-{\rm H}\alpha$
1242: color index against case $V-I$ (see Fig.~\ref{fig-haemit}, {\em left}),
1243: and then applying the selection criterion requiring all H\alp\ emission
1244: stars to have $R-{\rm H}\alpha \geq 0.4$ mag above the sequence of
1245: non-emission stars (Keller et al. 1999). This selection revealed 309
1246: stars, which are marked by fat dots in the $V-I$, $V$ CMD in Fig.\
1247: \ref{fig-haemit} ({\em right}). The bright sources of this sample are
1248: candidate Herbig Ae/Be stars or evolved Be stars, while the fainter are
1249: T~Tauri stars still in their pre-main sequence phase.
1250:
1251: \subsection{Spatial Distribution of Stars with H\alp\ Excess}
1252:
1253: The spatial distribution of the stars with H\alp\ excess coincides with
1254: the outline of the H{\sc ii} region (Fig.~\ref{fig-ima} ({\em right}) as
1255: well as the general spatial distribution of the PMS stars (Fig.
1256: \ref{fig-cms}, {\em right}) and the infrared emission from {\em Spitzer}
1257: observations (Simon et al. 2007; see also next section). Two major
1258: concentrations of emission stars can be readily identified, one located
1259: in the nebula N~66, following its general southeast-northwest trend, and
1260: the other to the north-northeast part of the association, coinciding
1261: with clusters 2a, 2b and 3.
1262:
1263: \section{{\em Spitzer} Observations of NGC~346\label{sec-spitzer}}
1264:
1265: As part of the {\em Spitzer} Survey of the Small Magellanic Cloud
1266: (S$^3$MC), Bolatto et al.\ (2007) imaged the SMC in all seven MIPS and
1267: IRAC wavebands, and compiled a photometric catalog of 400,000 mid- and
1268: far-infrared point sources. A color-composite image from {\em
1269: Spitzer}/IRAC imaging of the general area of NGC~346/N~66, retrieved
1270: from the {\em Spitzer} Data Archive is shown in Fig.~\ref{fig-mirima}.
1271: Bolatto et al. (2007) identified 282 bright candidate Young Stellar
1272: Objects (YSOs) as bright 5.8 \micron\ sources with very faint optical
1273: counterparts and very red mid-infrared colors ([5.8]$-$[8.0] $>$ 1.2
1274: mag, where [5.8] and [8.0] are the magnitudes of the sources in the 5.8
1275: \micron\ and 8.0 \micron\ IRAC bands, respectively. 16 of these
1276: candidate YSOs fall within the NGC 346 region observed with HST/ACS. By
1277: comparison with our ACS photometric catalog (Paper I) we obtained
1278: optical photometry for the candidate YSOs. Because of the difference in
1279: angular resolution of at least $\approx$20 between {\em Spitzer} and
1280: ACS, a unique identification of the optical counterparts of the YSO
1281: candidates is not always possible. In Table \ref{tab-yso} we summarize
1282: the IRAC and MIPS photometry of the YSOs identified by Bolatto et al.\
1283: (2007) as well as the ACS photometry of the brightest star located
1284: within 1\farcs5 of the {\em Spitzer} position. We also quote the number
1285: of stars detected on ACS within this radius, and find evidence for
1286: significant clustering. Up to $\approx$20 brighter main sequence and
1287: fainter pre-main sequence stars can be found within 1\farcs5 of the
1288: YSOs, i.e.\ we might be seeing compact stellar clusters. The mean
1289: density of these stars for the whole region ($\sim$ 0.15
1290: stars/$\Box$\arcsec) is much lower than the average stellar density of
1291: the stars found within 1\farcs5 around candidate YSOs. A study of the
1292: clustering of PMS stars at the loci of YSOs in the SMC association
1293: NGC~602 is presented by Gouliermis et al. (2007). These authors note
1294: that high-resolution near-infrared imaging and spectroscopy are required
1295: to better understand the stellar populations and clarify if they are
1296: signatures of on-going clustered star formation.
1297:
1298: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1299: \begin{deluxetable*}{ccccccccrcc}
1300: \tablewidth{0pt}
1301: \tabletypesize{\scriptsize}
1302: \tablecaption{Candidate YSOs found with {\em Spitzer} in the area centered
1303: on NGC~346 observed with ACS. ID numbers shown in column 1 are from
1304: Bolatto et al. 2007. Celestial coordinates are given in J2000. Magnitudes
1305: in all four IRAC channels (3.6 \micron\, 4.5 \micron\, 5.8 \micron\, 8.0
1306: \micron) and the 24 \micron\ channel of MIPS are given in columns 4, 5, 6, 7
1307: and 8 respectively. None of these objects was detected in the longer
1308: wavebands of MIPS (70 \micron\ and 160 \micron). The Numbers of the stars
1309: found to coincide from our ACS photometry (Paper I) with the loci of
1310: these YSOs (within 1\farcs5) and the $F555W$ and $F814W$ ($V$, $I$) VEGA
1311: magnitudes of the brightest ones are given in the last three
1312: columns.\label{tab-yso}}
1313: \tablehead{
1314: \colhead{ID} &
1315: \colhead{RA (deg)} &
1316: \colhead{DEC (deg)} &
1317: \colhead{[3.6]} &
1318: \colhead{[4.5]} &
1319: \colhead{[5.8]} &
1320: \colhead{[8.0]} &
1321: \colhead{[24]} &
1322: \colhead{No} &
1323: \colhead{$m_{\rm F555W}$} &
1324: \colhead{$m_{\rm F814W}$}
1325: }
1326: \startdata
1327: 185 & 14.7150 & $-$72.1710 & - & 14.91 & 12.62 & 10.33 &-& 4&22.55&22.04\\
1328: 186 & 14.7167 & $-$72.1673 & - & 14.56 & 12.58 & 11.21 &-& 5&18.39&17.41\\
1329: 187 & 14.7190 & $-$72.1602 & 15.17 & 15.86 & 12.11 & 10.31 &-& 7&23.14&22.50\\
1330: 188 & 14.7202 & $-$72.1783 & 15.75 & 14.97 & 12.33 & 10.64 &-& 4&23.82&21.64\\
1331: 189 & 14.7231 & $-$72.1624 & 14.57 & 15.32 & 12.52 & 10.97 &-& 1&18.86&17.76\\
1332: 190 & 14.7275 & $-$72.1642 & 14.57 & 15.28 & 12.16 & 10.55 &-& 8&20.04&19.46\\
1333: 191 & 14.7514 & $-$72.1681 & - & 12.02 & 10.64 & ~~9.19 &5.43& 6&16.27&15.96\\
1334: 192 & 14.7582 & $-$72.1762 & 14.76 & 14.35 & 12.59 & 11.25 &-&13&18.99&19.13\\
1335: 194 & 14.7605 & $-$72.1687 & 12.38 & 11.48 & 10.43 & ~~9.22 &4.41&10&18.67&17.68\\
1336: 195 & 14.7622 & $-$72.1765 & 13.43 & 13.26 & 11.49 & 10.02 &-& 5&14.45&14.70\\
1337: 196 & 14.7631 & $-$72.1774 & 13.93 & 13.73 & 11.78 & 10.24 &-& 9&18.39&18.26\\
1338: 197 & 14.7672 & $-$72.1797 & - & 12.74 & 11.67 & 10.31 &-& 8&16.41&15.11\\
1339: 198 & 14.7973 & $-$72.1670 & - & 13.91 & 11.90 & 10.02 &-& 7&15.91&16.09\\
1340: 200 & 14.8008 & $-$72.1663 & 11.78 & 11.28 & ~~9.96 & ~~8.43 &5.10& 2&16.77&16.04\\
1341: 203 & 14.8208 & $-$72.1544 & 13.90 & 13.58 & 11.87 & 10.12 &-&19&19.95&19.51\\
1342: 204 & 14.8293 & $-$72.1585 & - & 14.83 & 12.67 & 11.12 &-& 9&19.39&18.41
1343: \enddata
1344: \end{deluxetable*}
1345: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1346: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1347:
1348: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1349: \begin{figure}[t!]
1350: \epsscale{1.15}
1351: \plotone{f13.eps}
1352: \caption{Color-composite image of NGC~346/N~66 from {\em Spitzer}/IRAC
1353: with 3.6 \micron\ shown in blue, 4.5 \micron\ in green, and 8 \micron\ in red.}
1354: \label{fig-mirima}
1355: \end{figure}
1356: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1357:
1358:
1359: More recently, Simon et al. (2007) used the data of the S$^{3}$MC survey
1360: and detected a significant population of bright, red infrared sources.
1361: They expanded the catalog of candidate YSOs in the general area of N~66,
1362: and carried out spectral energy distribution (SED) fits based on models
1363: by Whitney et al. (2003a,b, 2004) to select $\approx$110 YSOs and YSO
1364: candidates. Of these, 65 fall within the field of NGC~346 observed with
1365: HST/ACS. The positions of these candidate YSOs, kindly provided by J.D.\
1366: Simon, are overplotted on the isodensity contour map of the PMS stars in
1367: the region of NGC~346 in Fig.~\ref{fig-cms} (right). The majority of the
1368: candidate YSOs from the Simon et al.\ catalog are located where we find
1369: the highest densities of PMS stars. Specifically, almost all of the
1370: ``definite'' YSOs (solid circle symbols) coincide with the density peaks
1371: of the map of Fig.~\ref{fig-cms} (right), and most of them are located
1372: in the association and its surrounding area. In addition, four of the
1373: PMS clusters we identified earlier (\S 4) coincide with ``definite''
1374: YSOs (clusters 2, 3, 4, and 5). According to Simon et al. (2007) the
1375: spatial distribution of the YSOs reveal that star formation in N~66 did
1376: not proceed from the southwest toward the center of the {\sc H ii}
1377: region as proposed by Massey et al (1989). Simon et al. also note that
1378: the most-embedded YSOs are more centrally concentrated than the more
1379: evolved ones, and that the YSOs are mass segregated with the most
1380: massive ones being located closer to the center of N~66. Also, the
1381: clumps of molecular gas found by Rubio et al. (2000) from CO(2-1)
1382: emission, and the peaks of dust emission at 7\micron\ analyzed by
1383: Contursi et al. (2000) coincide with many of these YSOs.
1384:
1385: \section{Summary}
1386:
1387: In this paper we present the results from our photometric study on the
1388: recent star formation history of N~66 the brightest {\sc H~ii} region in
1389: the SMC, related to the OB association NGC~346, as it is recorded in the
1390: observed PMS population of the region. Our deep photometry revealed an
1391: extremely large number of low-mass PMS stars in the association and the
1392: surrounding region, easily distinguishable in the $V-I$, $V$ CMD. We
1393: show that factors such as reddening, binarity and variability can cause
1394: a broadening of the positions of these stars in the CMD of the
1395: association, wide enough so that they can be misinterpreted as the
1396: result of multi-epoch star formation, and not as the product of a single
1397: star formation event. We found that a modest reddening like the one we
1398: found for the general area of NGC~346 ($E(B-V)\simeq 0.08$ mag) can make
1399: the PMS stars appear younger than what they actually are, and therefore
1400: an age of 10 Myr better fits the observed sequence of PMS stars.
1401: However, our results do not exclude the possibility of multi-epoch star
1402: formation in the area of the association if the reddening is even lower,
1403: as suggested from the OB stars of this region ($E(B-V)\simeq 0.04$ mag).
1404: In this case, two star formation events 10 and 5 Myr ago can explain the
1405: observed broadening of the PMS stars due to age spread and factors such
1406: as reddening and binarity. No specific dependence of the estimated ages
1407: of the PMS stars to their loci within the association, as a signature of
1408: sequential star formation, was found.
1409:
1410: It has been previously suggested that three different generations of
1411: stars occurred through sequential star formation in the region of
1412: NGC~346/N~66 (Rubio et al. 2000) within the last 3 Myr (based on the age
1413: of the youngest OB stars) and that the PMS stars of NGC~346 represent a
1414: star formation event that took place 3 - 5 Myr ago (Nota et al. 2006).
1415: However, the study by Massey et al. (1989) based on the presence of
1416: evolved 15~M{\solar} stars suggests that there might have been earlier
1417: star formation events in the region making NGC~346 as old as $\sim$~12
1418: Myr. In addition, there are recent indications that NGC~346 might host
1419: classical Be-type stars (Evans et al. 2006), and if so, the age of the
1420: association should be at least 10~Myr, a threshold given by Fabregat \&
1421: Torrej{\'o}n (2000) for classical Be-type stars to form. These results
1422: fit very well to the hypothesis that star formation in NGC~346 did
1423: already occur about 10~Myr ago, as our observations and simulations of
1424: the PMS stars suggest.
1425:
1426: From star counts based on our ACS photometry we identify at least five
1427: PMS clusters across the region, covering a range of ages. On spatial
1428: scales from 0.8$''$ to 8$''$ (0.24 to 2.4\,pc at the distance of the
1429: SMC) the clustering of the PMS stars as computed by a two-point angular
1430: correlation function is self-similar with a power law slope $\gamma
1431: \approx -0.3$. The clustering properties are quite similar to Milky Way
1432: star-forming regions like the Orion OB association or $\rho$\,Oph. Thus
1433: molecular cloud fragmentation in the SMC seems to proceed on the same
1434: spatial scales as in the Milky Way. This is remarkable given the
1435: differences in metallicity and hence dust content between SMC and Milky
1436: Way star-forming regions.
1437:
1438: The youngest PMS stars are located mostly to the north of the bar of
1439: N~66, where three PMS clusters are identified. This area is also
1440: characterized by a high concentration of candidate YSOs (Simon et al.
1441: 2007), H\alp-excess stars (found with our photometry), and IR-emission
1442: peaks (Rubio et al. 2000). This indicates that star formation probably
1443: still takes place in an arc-like feature, as it is outlined by the
1444: spatial distribution of these sources. In an accompanying letter, we
1445: combine these results with previous multi-wavelength studies of the
1446: region to investigate the star formation history, which helped to shape
1447: NGC~346/N~66 (Gouliermis et al. 2007b).
1448:
1449: \acknowledgments
1450:
1451: D. A. Gouliermis kindly acknowledges the support of the German Research
1452: Foundation (Deutsche Forschungsgemeinschaft - DFG) through the
1453: individual grant GO 1659/1-1. This paper is based on observations made
1454: with the NASA/ESA Hubble Space Telescope, obtained from the data archive
1455: at the Space Telescope Science Institute. STScI is operated by the
1456: Association of Universities for Research in Astronomy, Inc. under NASA
1457: contract NAS 5-26555. It is also based on observations made with the
1458: Spitzer Space Telescope, which is operated by the Jet Propulsion
1459: Laboratory, California Institute of Technology under a contract with
1460: NASA.
1461:
1462: %%%%%%%%%%%%%%%%%%%%%%%%%%% BIBLIOGRAPHY %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1463: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1464: %\newpage
1465: \begin{references}
1466: \reference{} Blaauw, A. 1983, IrAJ, 16, 141
1467: \reference{} Brice\~{n}o, C., et al. 2001, Sci, 291, 93
1468: \reference{} Brice{\~n}o, C., et al. 2005, AJ, 129, 907
1469: \reference{} Brice\~{n}o, C., et al. 2007, in ``Protostars and Planets V'',
1470: Eds. B. Reipurth, D. Jewitt, and K. Keil, University of Arizona Press,
1471: Tucson, p. 345
1472: \reference{} Bolatto, A., et al. 2007,ApJ, 655, 212
1473: \reference{} Chu, Y.-H.\ 1997, \aj, 113, 1815
1474: \reference{} Chu, Y.-H., \& Kennicutt, R.~C., Jr.\ 1988, \aj, 95, 1111
1475: \reference{} Contursi, A., et al. 2000, A\&A, 362, 310
1476: \reference{} Danforth, C.~W., Sankrit, R., Blair, W.~P., Howk, J.~C., \&
1477: Chu, Y.-H.\ 2003, \apj, 586, 1179
1478: \reference{} Davies, R.~D., Elliott, K.~H., \& Meaburn, J.\ 1976, MmRAS,
1479: 81, 89
1480: \reference{} de Bruijne, J.~H.~J., Hoogerwerf, R., \& de Zeeuw, P.~T.\
1481: 2001, A\&A, 367, 111
1482: \reference{} Dolphin, A.~E.\ 2000, \pasp, 112, 1383
1483: \reference{} Duquennoy, A., \& Mayor, M.\ 1991, A\&A, 248, 485
1484: \reference{} Efremov, Y.~N., \& Elmegreen, B.~G.\ 1998, \mnras, 299, 588
1485: \reference{} Elmegreen, B. G.\ 1998, PASA, 15, 74
1486: \reference{} Elmegreen,B. G.\ 2006, in ``Globular Clusters, Guide to
1487: Galaxies'' ed. T. Richtler, et al., ESO/Springer
1488: ({\tt arXiv:astro-ph/0605519})
1489: \reference{} Evans, C.~J., Lennon, D.~J., Smartt, S.~J., \& Trundle, C.\
1490: 2006, \aap, 456, 623
1491: \reference{} Elson, R.~A.~W., Sigurdsson, S., Davies, M., Hurley, J., \&
1492: Gilmore, G.\ 1998, MNRAS, 300, 857
1493: \reference{} Fabregat, J., \& Torrej{\'o}n, J.~M.\ 2000, \aap, 357, 451
1494: \reference{} Fukuda N. \& Hanawa T. 2000, ApJ, 533, 911
1495: \reference{} Girardi, L.,et al. 2002, A\&A, 391, 195
1496: \reference{} Gomez, M., Hartmann, L., Kenyon, S.~J., \& Hewett, R.\
1497: 1993, \aj, 105, 1927
1498: \reference{} Goodwin, S. P., Kroupa, P., Goodman, A., Burkert, A. 2007,
1499: in ``Protostars and Planets V'', Eds. B. Reipurth, D. Jewitt, and K.
1500: Keil, University of Arizona Press, Tucson, p. 951
1501: \reference{} Gouliermis, D., Brandner, W., \& Henning, T.\ 2006a, ApJL,
1502: 636, L133
1503: \reference{} Gouliermis, D.~A., Dolphin, A.~E., Brandner, W., \&
1504: Henning, T.\ 2006b, ApJS, 166, 549 (Paper I)
1505: \reference{} Gouliermis, D., Brandner, W., \& Henning, T.\ 2005, \apj,
1506: 623, 846
1507: %\reference{} Gouliermis, D., Brandner, W., \& Henning, T.\ 2006c, ApJ,
1508: %641, 838
1509: \reference{} Gouliermis, D.~A., Quanz, S.~P., \& Henning, T.\ 2007, ApJ,
1510: 665, 306
1511: \reference{} Gouliermis, D., et al.\ 2007b, submitted to ApJL
1512: \reference{} Grebel, E. K., Richtler, T., \& de Boer, K. S. 1992, A\&A,
1513: 254, L5
1514: \reference{} Haberl, F., \& Sasaki, M.\ 2000, \aap, 359, 573
1515: \reference{} Henize, K.~G.\ 1956, ApJS, 2, 315
1516: \reference{} Herbst, W., Herbst, D. K., Grossmann, E. J., Weinstein, D.
1517: 1994, AJ, 108, 1906
1518: \reference{} Johnson, H. L., Morgan, W. W. 1953, ApJ, 117, 313
1519: \reference{} Keller, S. C., Wood, P. R., \& Bessell, M. S. 1999, A\&AS,
1520: 134, 489
1521: \reference{} K{\"o}hler, R., Petr-Gotzens, M.~G., McCaughrean, M.~J.,
1522: Bouvier, J., Duch{\^e}ne, G., Quirrenbach, A., \& Zinnecker, H.\ 2006,
1523: A\&A, 458, 461
1524: \reference{}Kontizas, E., Kontizas, M., Gouliermis, D., Dapergolas, A.,
1525: Korakitis, R., Morgan, D. H. 1999, IAUS, 190, 410
1526: \reference{} Larson, R.~B.\ 1995, \mnras, 272, 213
1527: \reference{} Massey, P., Parker, J. W., Garmany, C. D. 1989, AJ, 98, 1305
1528: \reference{} Mathieu, R.~D.\ 1994, ARA\&A, 32, 465
1529: \reference{} Nakajima, Y., et al. 1998, ApJ, 497, 721
1530: \reference{} Naz\'{e}, Y., et al. 2002, ApJ, 580, 225
1531: \reference{} Niemela, V. S., Marraco, H. G., \& Cabanne, M. L. 1986, PASP,
1532: 98, 1133
1533: \reference{} Nota, A., et al. 2006, ApJ, 640, L29
1534: \reference{} Palla, F., \& Stahler, S.~W.\ 1999, ApJ, 525, 772
1535: \reference{} Palla, F., \& Stahler, S.~W.\ 2000, ApJ, 540, 255
1536: \reference{} Petr, M.~G., Coude Du Foresto, V., Beckwith, S.~V.~W.,
1537: Richichi, A., \& McCaughrean, M.~J.\ 1998, ApJ, 500, 825
1538: \reference{} Pollock, A. M. T. 2002, in ASP Conf. Ser., 260, Interacting
1539: Winds from Massive Stars, ed. A. F. J. Moffat \& N. St Louis (San
1540: Francisco: ASP), 363
1541: \reference{} Preibisch, T., \& Zinnecker, H., 1999, AJ, 117, 2381
1542: \reference{} Prosser, C.~F., Stauffer, J.~R., Hartmann, L., Soderblom,
1543: D.~R., Jones, B.~F., Werner, M.~W., \& McCaughrean, M.~J.\ 1994, ApJ,
1544: 421, 517
1545: \reference{} Pudritz, R.~E.\ 2002, Science, 295, 68
1546: \reference{} Rela\~{n}o, M., Peimbert, M., \& Beckmann, J. 2002, ApJ, 564,
1547: 704
1548: \reference{} Rochau, B., Gouliermis, D.~A., Brandner, W., Dolphin,
1549: A.~E., \& Henning, T.\ 2007, ApJ, 664, 322
1550: \reference{} Rubio, M., Contursi, A., Lequeux, J., Probst, R., Barbá, R.,
1551: Boulanger, F., Cesarsky, D., Maoli, R. 2000, A\&A, 359, 1139
1552: \reference{} Sabbi, E., et al. 2007, AJ, 133, 44
1553: \reference{} Sanduleak, N.\ 1968, AJ, 73, 34
1554: \reference{} Scalo, J. M.\ 1985 in Protostars and Planets II, University
1555: of Arizona Press, 201
1556: \reference{} Sherry, W. H.\ 2003, Ph.D.~Thesis, State University of New
1557: York at Stony Brook
1558: \reference{} Sherry,W. H., Walter, F. M., Wolk, S. J. 2004 AJ, 128, 2316
1559: \reference{} Shu, F.~H., Adams, F.~C., \& Lizano, S.\ 1987, \araa, 25, 23
1560: \reference{} Siess, L., Dufour, E., \& Forestini, M. 2000, A\&A, 358, 593
1561: \reference{} Simon, J. D., et al.\ 2007, accepted for publication in ApJ
1562: ({\tt arXiv:astro-ph/0707.3998})
1563: \reference{} Stahler, S. W., \& Palla, F. 2005, The Formation of Stars,
1564: ISBN 3-527-40559-3, Wiley-VCH
1565: \reference{} Tout, C. A. 1991, MNRAS, 250, 701
1566: \reference{} Vanhala, H. A. T., \& Cameron, A. G. W. 1998, ApJ, 508, 291
1567: \reference{} V\'azquez-Semadeni, E. 2004, Ap\&SS, 292, 187
1568: \reference{} Vel\'{a}zquez, P. F., Koenigsberger, G., \& Raga, A. C.
1569: 2003, ApJ, 584, 284
1570: \reference{} Walborn, N. R., et al. 2000, PASP, 112, 1243
1571: \reference{} Walborn, N. R. 1978, APJ, 224, L133
1572: \reference{} Walborn, N. R., \& Blades, J. C. 1986, ApJ, 304, L17
1573: \reference{} Wessolowski, U. 1996, in MPE Rep. 263, R\"{o}ntgenstrahlung
1574: from the Universe, ed. H. U. Zimmerman, J. E. Tr\"{u}mper, \& H. Yorke,
1575: 75
1576: \reference{} Whitney, B. A., Wood, K., Bjorkman, J. E., \& Wolff, M. J.
1577: 2003, ApJ, 591, 1049
1578: \reference{} Whitney, B. A., Wood, K., Bjorkman, J. E., \& Cohen, M. 2003,
1579: ApJ, 598, 1079
1580: \reference{} Whitney, B. A., Indebetouw, R., Bjorkman, J. E., \& Wood, K.
1581: 2004, ApJ, 617, 1177
1582: \reference{} Woitas, J., Leinert, C., K\"{o}hler, R.\ 2001, A\&A, 376, 982
1583: \reference{} Ye, T., Turtle, A. J., \& Kennicutt, R. C., Jr. 1991,
1584: MNRAS, 249, 722
1585: \end{references}
1586: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1587:
1588:
1589: \end{document}
1590:
1591: