0708.3510/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: %% manuscript produces a one-column, double-spaced document:
3: %% \documentclass[manuscript]{aastex}
4: %% preprint2 produces a double-column, single-spaced document:
5: %\documentclass[preprint2]{aastex}
6: %% Sometimes a paper's abstract is too long to fit on the
7: %% title page in preprint2 mode. When that is the case,
8: %% use the longabstract style option.
9: %% \documentclass[preprint2,longabstract]{aastex}
10: %% If you want to create your own macros, you can do so
11: %% using \newcommand. Your macros should appear before
12: %% the \begin{document} command.
13: \newcommand{\vdag}{(v)^\dagger}
14: \newcommand{\myemail}{skywalker@galaxy.far.far.away}
15: \newcommand{\ha}{H$\alpha$}
16: \newcommand{\hb}{H$\beta$}
17: \newcommand{\oiiir}{[O~{\scriptsize III}]~$\lambda$5007}
18: \newcommand{\hii}{H~{\scriptsize II}}
19: \newcommand{\hi}{H~{\scriptsize I}}
20: \newcommand{\oiii}{[O~{\scriptsize III}]}
21: \newcommand{\teff}{$T_{\mathrm{eff}}$}
22: \newcommand{\ebv}{ $E(B-V)$}
23: \newcommand{\cho}{{\sc\footnotesize  CHORIZOS}}
24: \newcommand{\stb}{{\sc\footnotesize Starburst99}}
25: \newcommand{\rv}{$R_{5495}$}
26: \newcommand{\ecc}{$E(4405-5495)$}
27: \newcommand{\logoh}{$12 + \log(\mathrm{O/H})$}  
28: \newcommand{\hrd}{ Hertzsprung--Russell diagram}
29: \newcommand{\mbol}{$M_\mathrm{bol}$}
30: \slugcomment{Published by The Astronomical Journal}
31: \shorttitle{The young stellar population of NGC 4214.II}
32: \shortauthors{\'Ubeda et al.}
33: 
34: 
35: \begin{document}
36: 
37: \title{The young stellar population of NGC 4214 as observed with 
38:  HST. II. Results.\footnote{Based on observations made with the NASA/ESA 
39: {\it Hubble Space Telescope,} obtained at the Space Telescope Science
40:  Institute, which is operated by the Association of Universities for 
41:  Research in Astronomy, Inc., under NASA contract NAS 5-26555.}}
42: 
43: \author{ Leonardo \'{U}beda, Jes\'us  
44: Ma\'{i}z Apell\'aniz\footnote{Affiliated with the Space Telescope Division of the 
45: European Space Agency, ESTEC, Noordwijk, Netherlands.} , \& John W. MacKenty}
46: \email{lubeda@stsci.edu, jmaiz@stsci.edu, mackenty@stsci.edu}
47: \affil{Space Telescope Science Institute, 3700 San 
48: Martin Drive, Baltimore, MD 21218, U.S.A.}
49: 
50: 
51: \begin{abstract} 
52: We present the results of a detailed UV--optical study of the
53: nearby dwarf starburst galaxy NGC~4214 using multifilter HST/WFPC2+STIS photometry.
54: %The data show the existence of both young and old stellar populations that can be resolved
55: %at the distance of NGC 4214 (2.94 Mpc). 
56: The stellar extinction is found to be quite patchy, with some areas
57: having values of $\ecc < 0.1$ mag and others, associated with star forming regions, much more heavily obscured, 
58: a result which is consistent with previous studies of the nebular extinction. We determined the ratio of
59: blue--to--red supergiants and found it to be consistent with theoretical models for the metallicity of the SMC.
60: The stellar IMF of the field in the range $20-100 \,\,M_\odot$ is found to be steeper than $\gamma=-2.8$
61:  ($\gamma=-2.35$  for a Salpeter IMF). 
62: % when considering the main sequence population
63: %and not accounting for the evolved massive stars.
64: A number of massive clusters and associations  with ages between a few
65: and 200 million years are detected and their properties are discussed.
66: \end{abstract}
67: 
68: \keywords{galaxies: individual (\objectname{NGC~4214}) --- galaxies: star clusters ---
69: galaxies: stellar content --- stars: early-type --- 
70: stars: luminosity function, mass function}
71: 
72: 
73: \section{Introduction}
74: In  this paper we present the results of
75: a comprehensive study of NGC~4214. In \cite{Ubedetala06}  (hereafter Paper~I) we 
76: presented and discussed the deep UV--optical images from two HST instruments:
77: WFPC2 and STIS.
78: We described  the methods that we had employed to 
79: reduce and analyze our HST images, and we briefly discussed  its stellar populations
80: by means of the analysis of 
81: two
82: Hertzsprung--Russell   [ $\log(T_{\mathrm{eff}}) $, $M_{\mathrm{bol}}$ ]  diagrams. 
83: 
84: In Paper~I we made a thorough   description of 
85: how we used \cho\  ~\citep{Maiz04},   
86: an IDL package
87: that fits an arbitrary family of spectral energy  distribution 
88: models (SEDs),  to analyze our multi--color photometric data. 
89: We applied \cho\ to stars and clusters separately, using different 
90: input SEDs. In this present work we show and discuss  the results.
91: 
92:   
93: NGC~4214 (see Figures 1, 2 and 4 in Paper~I)  is a nearby ($2.94 \pm 0.18 $ Mpc; \cite{Maizetal02a})  dwarf 
94: galaxy \citep{Vauc91} with a low metallicity content $(Z=0.006)$  \citep{Ubedetala06}.
95: It has two main regions of star formation which we studied using 
96: high--resolution images obtained from three HST archival proposals:
97: 6716 (P.I.: Theodore Stecher), 6569  (P.I.: John MacKenty), and
98: 9096 (P.I.: Jes\'us Ma\'{i}z-Apell\'aniz).
99: 
100:  
101: NGC~4214 possesses a combination of properties (high star formation rate, proximity, 
102: low extinction, spatially well resolved) that  make it a perfect candidate  for astrophysical studies
103: of  young stellar populations,
104: which we investigate in the present work. 
105: We analyze the ratio of blue to red (B/R)
106: supergiants, the initial stellar mass function (IMF), and the properties of its young-- and
107: intermediate--age cluster population. We also study the variable extinction across the galaxy.
108: 
109: 
110: In order to advance the understanding of massive stars
111: it is important to have an observational database with which the 
112: predictions of stellar evolutionary theory may be compared and refined.
113: In particular, the ratio B/R of the blue to red supergiants is a major characteristic of the luminous 
114: star population in galaxies.
115: %
116: \Citet{Bergh68}  first suggested the B/R ratio varied among nearby galaxies as a 
117: result of the effect of metallicity on massive star evolution. Several studies confirmed that 
118: the B/R ratio is  an increasing function of metallicity   \citep{LaMa94,Eggeetal02}. This
119: fact is known as  a result of studies in the Galaxy \citep{HumpMcel84} and  the Magellanic Clouds
120: \citep{Hump83,MassOlse03}.  
121: This ratio has also been studied in M33  \citep{Walk64,HumpSand80} and more recently 
122: in Sextans A by \cite{DohmSkil02}.  In this paper we examine the B/R ratio    for NGC~4214.
123: 
124: \cite{Krou02} presents an updated review
125: of our current knowledge of the IMF for different types of stars.
126: This function is particularly important for the most
127: massive stars \citep{Garmetal82}    
128: which evolve very quickly and strongly influence their environment via stellar winds and 
129: mass loss.  \cite{Massetal95, Massetal95b} studied the massive star content of 
130: OB associations in the Milky Way and the Magellanic Clouds 
131: and found that there is no difference in the IMF slope between the MW, the LMC
132: and the SMC; therefore, to first order, metallicity does not appear to affect the IMF of massive stars, 
133: at least over the factor of 5 spanned
134: between the SMC ($Z=0.004$), the LMC ($Z=0.008$)  and the MW ($Z=0.02$).
135: OB associations have  been discovered in some Local Group galaxies, and with the 
136: fine spatial resolution provided by the HST, it is possible to obtain the stellar IMF
137: of those systems. Recent works include M31  \citep{Veltetal04}, M33  \citep{GonDPere00}, 
138: IC10 \citep{Hunt01}, and  IC1613  \citep{Georetal99}. The previous work on the IMF of 
139: NGC~4214 has been based on the integrated spectrum of the galaxy 
140: \citep{Leitetal96,Chanetal05}. Here we present the first work on its IMF based on high--resolution
141: HST imaging.
142: 
143:  
144: Stellar clusters have long been recognized as important laboratories for astrophysical
145: research. They are extremely useful in different aspects of Astronomy: they
146: provide classical tests of stellar evolution, they can be used for studies of stellar dynamics
147: and they provide an
148: understanding of the galactic structure \citep{LadaLada03}.
149: Young clusters are tracers of recent star formation; they form in 
150: giant molecular clouds (GMCs) and they remain embedded or close to
151: their parent cloud during 
152: the first $\approx 1$ Myr, making it difficult to study them in their early stages.
153: However, recent developments in infrared imaging cameras have led 
154: to the conclusion that young embedded clusters are numerous, and that
155: a significant fraction of all stars may form in such systems. 
156: %
157: %
158: Massive Young Clusters (MYCs) can be divided into 
159: Super Star Clusters (SSCs), which are organized  
160: around a compact core (size $\approx 1-3$ pc), and 
161: Scaled OB Associations (SOBAs) which lack such a
162: structure and are more extended objects (size $> 10$ pc). 
163: SSCs are bound objects and represent the high--mass 
164: end of young stellar clusters while SOBAs are 
165: unbound and are the massive relatives of regular OB associations \citep{Maiz01b}.  
166: MYCs have been discovered in various environments, including our Galactic Center,
167: in the nuclei of late--type galaxies, in nearby starburst galaxies and  in merging galaxies.
168:  
169:  
170:  
171:   
172: The present paper is organized as follows: In Section 2  we present 
173: the study of the extinction throughout  NGC~4214; in Section 3 we analyze the B/R ratio;
174: the IMF is addressed in Section 4; in Section 5 we perform a detailed  study of 
175: some interesting clusters. 
176: Finally, in Section 6 we summarize our results and provide the conclusions.
177: 
178: \section{Extinction}
179: 
180:  A comparison between stellar atmosphere models and 
181:  observed colors can be used to infer the extinction toward 
182: individual stars. For our study, we used the IDL code 
183:   \cho ~\citep{Maiz04};
184: this program (see Section 2.5 in Paper~I) 
185: reads unreddened SED models, extinguishes them and obtains the 
186: synthetic photometry
187: that is compatible with  the input colors.    In this way
188: we obtained the extinction law  that best agrees  with the data,
189: and the extinction value   [ \ecc ~] ~for each star in our main photometric
190: lists:  {\sc\footnotesize  LIST336} (built with F336W as reference filter) and
191: {\sc\footnotesize  LIST814}  (built with F814W as reference filter) as described in Paper~I.
192: For the extinction law, we considered  
193: the $R_{5495}-$dependent family of 
194:  \citet{Cardetal89} (Galactic extinction); the average LMC and 
195: LMC2 laws of  \citet{Missetal99}   (extinction of the LMC), and the SMC
196: law of  \citet{GordClay98}  (extinction of the SMC).  \cho\ also provides
197: the uncertainties $\sigma  $  of the calculated parameters. 
198: 
199: In  previous studies of NGC~4214, the extinction correction 
200: was performed using different values of  $E(B-V)$:  \citet{Maizetal98} 
201: measured a variable $E(B-V)$ with values between 0.0 and 0.6 mag from the ratio of the nebular emission
202: lines \ha\ to \hb. \citet{Maizetal02a} measured an average value of $E(B-V)=0.09$ mag from the optical
203: colors of the young stellar population far from the main star--forming complexes. \citet{Droz02} and 
204: \citet{Calzetal04} adopted the value $E(B-V)=0.02$ mag provided by the IRAS DIRBE map 
205: of \citet{Schletal98}. All those estimates assume a \citet{Cardetal89} extinction law with $R_{5495}=3.1$.
206: 
207: In order to develop an extinction map of NGC~4214, we made a selection of
208: objects from  {\sc\footnotesize  LIST336}, using the following criteria:
209: We considered all the objects with
210: $\sigma_{E(4405-5495)}  \leqslant 0.1 $ mag  for those stars with $E(4405-5495)  \leqslant 0.4 $ mag.
211: This selection yields 855 objects with $<{\mathrm F336W}>=20.93$ and  $<\sigma_{\mathrm  F336W}>=1.22$.
212: Objects with  $E(4405-5495)  > 0.4 $  mag
213:  were selected on the basis of their F336W magnitude 
214:  (F336W $< 21$ mag)  and their location within the galaxy.
215:  We made no restriction according to photometric or
216:  extinction error in this case. We found 43 objects with $E(4405-5495)  > 0.4 $
217:  and they present $<E(4405-5495) >= 0.54$ mag and 
218:  $<\sigma_{E(4405-5495)} > = 0.21 $ mag.
219: % If we had made a restriction to their extinction error (like   
220: % $\sigma_{E(4405-5495)}  \leqslant 0.1 $ mag) we would not have found any objects 
221: % in this category.  
222: % We understand that our sample is selection biased, and that 
223: %we may be missing some blue objects, and
224: %
225: Using this biased sample of objects, we
226:  may be underestimating the extinction values and probably 
227: missing some blue objects.
228: It is important to note that extinction is a complicated three-dimensional effect which 
229: involves stars, nebulae and dust and which varies at  small  ($  \lesssim 1$ pc)  scales.
230: All these facts contribute  to  make extinction--modelling 
231: quite difficult.
232: %
233: Finally, we added the results of the extinction analysis (Section 5.1)
234: for clusters I--As, I--Es, IIIs, and IVs to our list. 
235: The left panel of Figure~\ref{fig01} is an F656N  mosaic of the surveyed region of NGC~4214,
236: where we have marked the  objects in our final list with a color scheme
237: to indicate the extinction   distribution throughout the galaxy.  
238: To build the \ha\ (filter F656N) mosaic, we used  images u3n8010fm + gm from
239: proposal 6569. The total exposure time is 1600 sec. This mosaic clearly shows the patchy 
240: nature of the extinction. 
241:  
242: To build the extinction map, we created a  spatial grid $M$ and 
243: assigned a weighted value of the extinction
244: to each pixel in the grid. The procedure follows:
245: We first calculated the distance $d_{ijk}$ (in  WF pixels of $0\farcs1$) between each pixel in the grid $M[i,j]$
246: and all the stars $k$ in our list.  We then 
247: calculated a weight with
248: \begin{equation} 
249: w_k = \frac{\exp\left( - \frac{ d_{ijk}^2}{30}    \right)}{\sigma_{E_k}^2}  ,
250: \end{equation}
251: 
252: \noindent where  $\sigma_{E_k}$  is the error in $ E =E(4405-5495) $ for star $k$, and
253: the value 30 is a scale factor. 
254: We used several values of the scale factor to represent the extinction map, and 
255: we present here the map with the one  that best showed the different extinction  regions. 
256: The value associated 
257: to the pixel $M[i,j]$ is given by 
258: 
259: \begin{equation}
260: \label{eqn-0}  
261: M[i,j] = \frac{\sum_k E_k \cdot w_k}{\sum_k w_k} 
262: \end{equation}
263: 
264: The extinction map is shown in the right panel
265: of Figure~\ref{fig01}. The most prominent feature in this map is the fact that
266: NGC~4214 is characterized by low values of the extinction, except for some well defined
267: regions with high values,  which is in agreement
268: with   \cite{Maizetal98}, \cite{Droz02}, and  \cite{Calzetal04}.
269: In Section 5  we discuss in more detail the regions around some of the clusters.
270: 
271: \cite{Maizetal98} and \citet{Maiz00} used the Balmer ratio (H$\alpha$/H$\beta$) as a tracer of the
272: reddening that affects the ionized gas and produced maps of this ratio.
273: Their analysis showed a significant difference between the two most prominent 
274: complexes in the galaxy: NGC~4214--I and  NGC~4214--II. Their conclusions are
275: that the nebular emission and stellar continuum are produced in co--spatial 
276: or close regions in NGC~4214--II, while the emitting gas is clearly spatially offset with 
277: respect to the stellar cluster in  the brightest knots on the NGC~4214--I complex. They also find 
278: that the reddening in NGC 4214--II is, on average, higher than in NGC 4214--I. Our results, derived
279: from the stellar colors, validate those points. The two main cavities in NGC 4214--I show low 
280: extinction surrounded by higher values while for NGC 4214--II the extinction is higher overall.
281: 
282: \cite{Waltetal01} present an interferometric study of the molecular gas in NGC~4214. They
283: detect three regions of molecular emission, in the northwest, southeast, and center of 
284: the galaxy. 
285: These authors compared the structure of  the molecular tracer, CO, with
286: tracers of star formation like \ha.
287: Two of the three CO complexes are associated directly with star--forming regions.
288: The southeastern CO clump appears co--spatial  with  NGC~4214--II. The peak 
289: of the CO emission is almost on top of one of the clusters. Our extinction map shows 
290: a high extinction region co--spatial with this molecular cloud. 
291: The central CO emission is associated with the largest region in \ha\ emission, 
292: spanning most of the NGC~4214--I region. 
293: This CO complex is diffuse instead of centrally concentrated, and the peak 
294: of the CO emission is shifted to the west of the peak of the \ha\ emission, with little CO  seen at
295: the location of the two main cavities. Overall, our extinction map traces the molecular cloud in this 
296: part of the galaxy too, although with less detail. 
297:  
298:  The extinction derived from the stellar continuum  
299: is similar to the extinction derived via the analysis of 
300: nebular lines across the galaxy, and this is true on a star by star basis.
301: The coincidence is fairly good throughout the galaxy.
302: 
303: 
304: 
305: \section{The ratio of blue--to--red supergiants }
306: 
307: The ratio B/R of the blue to red supergiants of initial masses
308:  larger than 
309: 15 $M_{\sun}$    is an important observable  of the luminous 
310: star population in galaxies and it is one of the stellar properties that can
311: be easily measured beyond the Local Group. 
312: This quantity depends  strongly on the model parameters 
313: and it can be used to constrain the model physics very accurately. This ratio has
314: been calculated in Galactic clusters and  in some clusters in the LMC and SMC. 
315: Both \cite{Eggeetal02} and  \cite{LaMa94}  present comprehensive reviews of past studies.  
316: 
317: The main result from previous observational studies 
318:  is that, for a given luminosity range, B/R steeply increases with 
319: increasing metallicity $Z$, by a factor of about 10 between the SMC and the inner 
320: Galactic regions. Current theoretical models of massive stars are unable to 
321: correctly reproduce the changes of B/R with metallicity, from solar to SMC 
322: value. It is known that the B/R ratio is a sensitive quantity  to mass loss,
323: rotation, convection and mixing processes, hence it constitutes  an important 
324: and sensitive test for stellar evolution models if it were fully understood.
325: 
326: In this paper we  present our results for NGC~4214. 
327: The B/R value is dependent on how stars are counted, and thus disagreement
328: with the predictions of stellar evolutionary models have to be carefully evaluated. 
329: Notice that the definition of B/R is not always the same. We followed the method 
330: suggested  by  \cite{MaedMeyn01}. We count in the B/R ratio the B star 
331: models from the end of the main sequence to type B9.5 I, which corresponds 
332: to $\log(T_{\mathrm{eff}})  =3.99 $ according to the calibration by \cite{Flow96}.
333: We count as red supergiants all star models below  $\log(T_{\mathrm{eff}})  = 3.70 $. 
334: In both cases we considered stars located between the 15 and 25   $M_{\sun}$
335: evolutionary  tracks.
336: Figures~8 and 9 in Paper~I  show the detailed evolutionary tracks 
337:  of non--rotating stellar models for initial 
338: masses between 5, 7, 10, 12, 15, 20, 25, 40, and 120   $M_{\sun}$,  LMC--like metallicity $(Z=0.008) $
339: from \cite{Schaetal93}.
340: Two vertical lines (at $\log(T_{\mathrm{eff}})  =3.99 $ and
341: $\log(T_{\mathrm{eff}})  =3.70 $) mark the limits between  the blue 
342: supergiants locus and the position of the red supergiants.
343: The  location of the blue and  red supergiants is marked in those
344: Figures with two  polygons  
345: between the 15 and 25   $M_{\sun}$ tracks.
346: We used Figure 8 in Paper~I to count red supergiants and Figure~9 
347: in Paper~I  to count blue supergiants.
348: 
349: Several studies identify  supergiants by considering objects brighter than  
350: $M_{\mathrm{bol}} = -7.5 $ mag  which corresponds to masses larger than 
351: 15 $M_{\sun}$  for red supergiants. If a lower luminosity is chosen there is a 
352: chance of contamination by intermediate asymptotic giant branch (AGB) stars 
353:   \citep{Brunetal86}.
354: We avoided this problem by selecting objects located above
355: the 15 $M_{\sun}$ evolutionary track in all cases.
356: We  performed an analysis to study how the  
357: stars located below the 15 $M_{\sun}$ evolutionary track
358: would contaminate  those which we consider red supergiants.
359: For this contamination analysis we generated artificial stars with masses that
360: place them below the 15 $M_{\sun}$ evolutionary track and used the typical uncertainties 
361: obtained from \cho\ for  \teff\ and \mbol\ to modify their positions in the H--R diagram accordingly. 
362: Then, using the observed distribution below the 15 $M_{\sun}$ track, we tested whether a significant 
363: number could have ``leaked'' to a higher mass bin due to the experimental uncertainties. 
364: The contamination thus calculated was found to be less than 1\%, so a correction was not applied.
365: For our analysis, we assumed that we are considering single stars and therefore no correction for 
366: blends of single stars was performed. 
367: 
368: In order to compare our observational results with what theory predicts, 
369: we calculated the   B/R   ratio values using three grids of theoretical models
370: from the Geneva database: [1] the evolutionary tracks of non--rotating 
371: ($v_{ini} = 0$ km~s$^{-1}$) stellar models for initial masses between 
372: 9 and 60   $M_{\sun}$ and SMC--like metallicity $(Z=0.004) $  \citep{MaedMeyn01}; 
373: [2]    the evolutionary tracks of rotating ($v_{ini} = 300$ km~s$^{-1}$)
374: stellar models for initial masses between 9 and 60   $M_{\sun}$ 
375: and SMC--like metallicity $(Z=0.004)  $   \citep{MaedMeyn01}; and  [3]   
376: the evolutionary tracks of non--rotating stellar models for initial 
377: masses between 10 and 120   $M_{\sun}$,  LMC--like metallicity $(Z=0.008) $
378: \citep{Schaetal93}. We did not find any major difference between models
379: with high and normal mass--loss rates, so we adopted normal mass--loss
380: rates evolutionary tracks. 
381: Unfortunately, evolutionary tracks of rotating ($v_{ini} = 300$ km~s$^{-1}$)
382: stellar models with LMC--like metallicity are not available in the literature.
383: 
384: The ratio of the densities of blue  supergiants to red supergiants at a given initial
385: mass is equal to the ratio of the lifetimes along the evolutionary tracks in the
386: corresponding $\log(T_{\mathrm{eff}})  $  intervals. The 
387: resulting  theoretical B/R values are given in  Table~\ref{tbl01}.
388: 
389: 
390: In order to count  the number of blue and red supergiants in our lists, 
391: we had to apply corrections based  on our completeness tests.  
392: We used the completeness
393: values  described in Section 2.3 of Paper~I, which we summarized in  
394: its Table~3.
395: We   calculated the completeness values for two 
396: intervals of $\log(T_{\mathrm{eff}}) $  along each evolutionary track: 
397: from the end of the main-sequence to $\log(T_{\mathrm{eff}}) = 3.99 $ 
398: and for the interval  $\log(T_{\mathrm{eff}})  \le  3.70 $. For each of the points 
399: that define the evolutionary tracks  we interpolated linearly  in the completeness tables 
400: to calculate their completeness value. We also computed a weight defined as the
401:  difference between the age of this point on the evolutionary track, and the age 
402:  of the same point in the immediately older evolutionary track.  The adopted 
403:  completeness value is the weighted mean in each interval. 
404: 
405: Counting the blue supergiants in our sample presented  no problem; we simply 
406: calculated the number of stars within each mass range provided by the evolutionary 
407: tracks. However, if we compare the distribution of red supergiants in the \hrd ~to
408: that of the various stellar evolutionary models, we find that none of the models
409: produce RSGs as cool and luminous as what is actually observed. This fact 
410: was noted by  \cite{Massey03}, and \cite{MassOlse03}. They 
411: found the same problem while trying
412: to account for the RSG content in the Magellanic Clouds.
413: In their work, the sample of RSGs includes both stars with known spectral type and
414: a group of objects with known photometry but no spectroscopy.
415: They plot those supergiants in    [$\log(T_{\mathrm{eff}}) $, $M_{\mathrm{bol}}$] planes
416: and overplot  several sets of evolutionary tracks. They use a new calibration 
417: between spectral type and  $\log(T_{\mathrm{eff}}) $ to place stars with
418: known spectral type. For the rest of the stars, they use the intrinsic color $(V-R)_0$.
419: Whatever the method to place the stars in the diagram, we can argue 
420: that there is no significant difference.
421: Their Figure 5(a)  for the LMC RSGs is quite similar to our Figure~8 in Paper~I.
422: It is important to note that our method is strictly photometric, because \cho\
423: fits the best known SEDs from \cite{Kuru04} to a set of observed  photometric colors.
424: 
425: In order to count the
426: RSGs in our sample we had to artificially extend   the evolutionary 
427: tracks at constant  values of   $M_{\mathrm{bol}}$, and used the region
428: marked by a small rectangle located between the 15 and 25  $M_{\sun}$
429: tracks in Figures~8 and 9 in Paper~I  to the right of $\log(T_{\mathrm{eff}})  = 3.70 $.
430: After counting the red and blue supergiants in our lists we corrected these 
431: numbers for incompleteness. A summary of the results  of the ratio B/R is presented in  
432: Table~\ref{tbl01}. There, the columns labeled ``theory'' show the numbers 
433: derived from the ratio of time spent by a
434: star in each region and, therefore, assumes a constant star formation rate. 
435: The ``observation'' columns show the
436: numbers previously described in this paragraph.
437: 
438: Figure~\ref{fig02} shows  the distribution of blue and red
439: supergiants on an F814W  mosaic of NGC~4214.
440: The filled circles represent the confirmed supergiants, both blue and red that
441: we used to calculate the B/R ratio.
442: With open red circles we represent the group of stars  that follow the criteria
443: $\log(T_{\mathrm{eff}})  \leqslant 3.70 $ and  $M_{\mathrm{bol}}  \leqslant  -6.0$.
444: The latter objects may be either RSGs, AGB stars, or even bright red giants that 
445: have ``leaked'' into higher--mass
446: regions in the H--R diagram due to observational uncertainties.
447: 
448: Despite the   difficulty in counting RSG because the
449: computed grids do not produce RSGs as cool  and luminous as what 
450: we observe, we find a good agreement between our data and the non--rotating 
451: models with low metallicity $(Z=0.004).  $
452: These models predict a B/R value of 24 in the $ 15-20 \, M_{\sun}$ mass--range, and 
453: we measure $34 \pm 10$. Some of the stars below the  15 $M_{\sun}$ track
454: may be  RSGs, so we are  underestimating the  real number of 
455: RSGs.
456: % by a factor of two  in our  $ 15-20 M_{\sun}$  sample.
457: In the mass--range $ 20-25 \, M_{\sun}$ the
458: theoretical prediction is 47 and we obtain $46 \pm 23$.
459: In both cases, our results agree with  the theoretical ones  within Poisson errors. 
460: 
461: %15 Ð 20	24	34 ± 10	0.4	29 ± 10	1.0	13 ± 4
462: %20 Ð 25	47	46 ± 23	0.4	45 ± 31	2.7	67 ± 66
463: 
464: 
465: There are two caveats regarding this result that should be mentioned. 
466: First, the small quantity of confirmed 
467: RSGs in our sample in NGC~4214 imply that stochastic effects due 
468: to small--number statistics may be present.
469: However, given the large differences in the theoretical ratios 
470: shown in Table~\ref{tbl01}, this effect is likely to
471: be unimportant. The second caveat is related to the conversion from 
472: observed colors to effective temperatures and
473: bolometric magnitudes. \cite{Leveetal05} present a new effective temperature scale for Galactic 
474: RSGs  by fitting MARCS stellar atmosphere models \citep{Gustetal75, Plezetal92} which include an 
475: improved treatment of molecular opacity to 74 Galactic $(Z=0.020)  $ 
476:  RSGs of known distance. They compare their location on the  
477: [$\log(T_{\mathrm{eff}}) $, $M_{\mathrm{bol}}$] plane with theoretical 
478: evolutionary models from \cite{MaedMeyn03} and find a much better agreement
479: between theory and observation.
480: Their main result is that RSGs appear to be warmer than previously thought.
481: This effect shifts the stars to the left in the diagrams making them coincide with the end of the tracks. 
482: It would be interesting to compare the change of temperature
483: from the fitting of observed optical colors using \cho\ or a similar code 
484: from Kurucz to MARCS atmospheres
485: to verify if it accounts for the $350-400$ K discrepancy detected in our data. 
486: More importantly, such a change should
487: also shift the stars downwards in the H--R diagram, since a higher
488:  temperature implies a lower
489: bolometric correction. 
490:  
491: 
492: At the time in which we performed our fits, \cho\ did not
493: have the capability to fit  stellar models other than Kurucz or Lejeune. 
494: It is in our future plans to reanalyze this data 
495: using the MARCS stellar atmospheres models.
496: 
497: 
498: 
499: \section{Initial mass function}
500: 
501: \subsection{The IMF of the resolved population}
502: A lot of effort has been put into obtaining the IMF of the Milky Way and the Magellanic Clouds
503:  in the past decades.
504: Star counts  in clusters/associations in those local galaxies reveal an IMF with a slope close to Salpeter
505: $(\gamma = -2.35) $ above $\sim 1 M_{\odot}$  \citep{Scha03} when  a power law of the
506: form $\frac{dN}{dm}  = A \, m^{\gamma} $ is used.
507: \cite{Masse98} concludes that there is no difference in IMF slopes found between the Milky Way, the LMC
508: and SMC. He claims that the weighted average of the IMF slopes of the MW is $\gamma = -2.1\pm0.1$;
509: and that of the MCs is $\gamma = -2.3\pm0.1$. This shows that 
510: metallicity does not affect the IMF
511: slopes of massive stars.
512: 
513: It has long been known that some very massive stars are not currently
514: found in clusters or  associations, but they seem to be part of the field \citep{Masse98}.
515: Studies of the MCs showed that very massive stars can be found in highly
516: isolated regions, and that  they are the result of small star--forming events. 
517: An IMF study of these field objects led to surprising results: the actual IMF slope
518: is quite steep, with  $\gamma \sim -5$  \citep{Massetal95b}. This trend is found 
519: in the MCs and the Milky Way as well, and is easily detected in the spectral
520: type distribution \citep{Bergh04}.
521: It is currently debated whether the difference in slopes is due to
522: differences between in situ formation styles or to
523: a majority of field O stars being runaways \citep{deWietal05}.
524: 
525: 
526: What do we know about the stellar IMF above $\sim 1 M_{\odot}$ for galaxies beyond the MCs?
527: There are several problems  involved in the determination of IMFs in extragalactic systems. 
528: Only a few  galaxies are close enough for star counts to be carried out with any reliability, 
529: and even these are so distant that only the very brightest part of the luminosity function 
530: can usually be sampled. In addition, there are practical problems with star counts in 
531: external galaxies, including crowding, incompleteness, and corrections for foreground stars. 
532: Among other studies,
533: \cite{Veltetal04} studied the IMF of M~31 and derived a slope of $\gamma = -2.59\pm0.09$; 
534: \cite{Jameetal04} 
535: found an IMF slope of  $\gamma  = -2.37 \pm  0.16$ for the ionizing cluster of NGC~588 
536: in the outskirts of the nearby galaxy M~33; \cite{Annietal03} analyzed the star formation history
537: of NGC~1705 and inferred  an IMF slope close to Salpeter.
538: %
539: 
540: The  most direct and reliable method of obtaining the IMF of a certain
541: stellar population is based on counts of stars as a function
542: of their luminosity/mass. However, there are other indirect methods
543: which do not employ star counts, but still yield some information on the IMF \citep{Scalo86}.
544: One of these is the method of population synthesis, which attempts to match the
545: observed galaxy colors, spectrum or line strengths by finding the best mixture
546: of stars of various spectral types and luminosity classes. The main problems that arise
547: with this method are the uniqueness of the solution and the types of
548: astrophysical constrains that need to be imposed on the data.
549: 
550: \cite{Chanetal05}  apply this method to estimate the IMF slope in the field stars of NGC~4214.
551: They  assume that  the faint intra--cluster light in NGC~4214 corresponds to the 
552: field stars in the galaxy. They 
553: compared the spectroscopic signature with \stb\ evolutionary synthesis models
554: and found a lack of   strong O--star wind features in the spectra, which led them to conclude that
555: the field light originates primarily  from a different stellar population, and not from scattering 
556: of UV photons leaking out of the massive clusters.
557: Fitting IMF slopes in continuous--formation \stb\ models they infer that the best value 
558: for NGC~4214 would be   $\gamma =  -3.5  $.
559: Their work also provides similar values of the IMF slope for other local starburst galaxies
560: such as NGC~1741, NGC~3310, and NGC~5996.
561: 
562: Comparing   low resolution spectra taken 
563:  with the International Ultraviolet Explorer (IUE) 
564: in the 1100--3200 \AA\ range with the predictions of 
565: evolutionary population synthesis models, \cite{MasHKunt99} 
566: obtain $ \gamma = -3.0$ 
567: for the sum of field and clusters in NGC~4214.
568: 
569: 
570: 
571: \subsection{Method and sample selection}
572: Determining the IMF of a stellar
573: population with mixed ages is a difficult problem.
574: Stellar masses cannot be weighed directly in most instances, so the mass
575: has to be deduced indirectly by measuring the star's luminosity  and evolutionary state.
576: We followed the photometric/spectroscopic method provided by  \cite{Lequ79}: 
577: we estimated bolometric magnitudes 
578: (\mbol), and effective temperatures (\teff) for each star in the sample, as we 
579: explain in Paper~I.
580:  The present--day mass function was  
581: estimated by counting the number of stars in the \hrd  ~between 
582: theoretical evolutionary tracks computed
583: for models with different masses. 
584: In the case of a coeval star--formation region, the obtained function
585: {\em  is}  the IMF  modified by evolution at the high--mass end. 
586: %
587: For a continuous
588: star formation region, the IMF followed by division of the number of objects 
589: in each mass range
590: by the main sequence lifetime $(\tau_{MS})$ of the corresponding mass.
591:   
592: 
593: For the IMF determination, we used the theoretical evolutionary tracks from   \cite{LeSc01} 
594: in the mass range 5  to 120 $M_{\odot}$.
595:  These  theoretical models    provide 51 values of \teff, $L$ and age for 
596: each evolutionary track for a given metallicity. 
597: The first 11 points of each track define their main--sequence sections.
598: A line connecting the first point of each track defines the 
599: zero age main--sequence (ZAMS). 
600: This line is the left boundary of our main--sequence. A line connecting the 
601: 11th point of each track defines the right boundary of the main--sequence.
602: 
603: 
604: Here we present a detailed analysis of the IMF of NGC~4214 using two key
605: assumptions: (i) we selected only main--sequence stars from our sample, and 
606: (ii) we assumed  that the  IMF has remained constant
607:   as a function of time.
608: %
609: Figure~2 in our Paper~I shows 
610: three regions of strong  \ha\  emission, which can be matched to active recent star--forming regions. 
611: These regions are I--A, I--B and II. 
612: In   {\sc\footnotesize  LIST336}  we found a total of  3061  objects located along  the main--sequence band.  
613: We considered  three lists of masses to calculate the IMF: (1) {\sc\footnotesize  LIST1}: a list including
614: all objects  (3061 stars), (2)  {\sc\footnotesize  LIST2}: a list with objects in regions  
615: I--A, I--B and II (active star--forming regions)
616:  (900 stars), and (3)  {\sc\footnotesize  LIST3}: a list with all objects  except those of  I--A, I--B and II  (2161 stars.)
617:  %
618: We assumed all binary and higher--order stellar systems are resolved into individual stars. See
619: Section 4.3.3  for a detailed discussion on multiplicity.
620: 
621: 
622: 
623: We started by determining the stellar masses and their errors
624: of all objects in the lists
625: using the estimated values of \mbol ~and \teff    ~obtained with \cho\ as described in Paper~I.
626: We plotted a \hrd  ~[$\log$ \teff , \mbol ] on which we superimposed
627: theoretical evolutionary tracks from   \cite{LeSc01} in the mass range $5-120 M_{\odot}$.
628: We then isolated  the main--sequence sections of those tracks and we obtained interpolated mass values 
629: between the tracks using splines. We built an  IDL code that triangulates the whole grid and interpolates 
630: the data to obtain a finely spaced grid of masses throughout the whole \hrd. This procedure gave us a 
631: two variable function $M = f($~\teff,~\mbol~) which assigns a mass $(M)$ value 
632: from the input values \mbol ~and \teff. 
633: The individual stellar masses $(M_i)$ followed directly from this function.
634: To obtain the uncertainties $(\sigma_{M_i})$  for the individual masses $(M_i)$, 
635: we created a randomly distributed
636: set of 10000 points around the mean values of \teff~   and \mbol~   for 
637: each star and  inferred
638:  their individual masses. Those points were distributed using 
639:  a bidimensional gaussian distribution according to the parameters provided
640:  by the output of {\sc\footnotesize  CHORIZOS}.
641: The standard deviation of these masses gave us  a reliable value of the error 
642:  $(\sigma_{M_i})$  in the mass  $(M_i)$ of each star.  
643:  %
644:  How dependent are our results on the evolutionary models used for the mass tracks in the \hrd ~? 
645: We found that the locus  of an  evolutionary track   of a given mass on the \hrd, changes with initial
646: conditions  such as metallicity, mass--loss, rotation, etc. Most of them show loops which 
647: give them a  complex shape.
648: However, these changes are minor and may be neglected when we consider the uncertainty 
649: in the mass determination for the objects in our lists, especially since 
650: we are considering objects that lie within the limits of the main sequence where
651: those changes are insignificant. 
652: The covariance ellipse for each single
653: object, usually spans a wide range of evolutionary tracks (due to the intrinsic errors
654: in the determination of  \teff\ and \mbol\ from \cho.) This would have not changed if
655: we had chosen a different set of theoretical models. 
656: It is important to note that the most massive objects in our lists have 
657: smaller relative errors
658: than the less massive ones. 
659: How does that translate into the determination of the error in the mass of each star? 
660: Even though the most massive objects have smaller errors in  \teff\ and \mbol, 
661: their errors in mass 
662:  may still be large 
663: because their covariance ellipses extend over  a wider range of masses.
664: On the other hand, objects with lower masses  have bigger errors in    \teff\ and \mbol, but
665: their location in the \hrd ~is such that their ellipses span a narrower  range of 
666: evolutionary tracks.
667: 
668: 
669: Regions  I--A, I--B and II  are young clusters with recent bursts of star formation.
670: They all have different ages as we will show in Section~5. For these regions, we 
671: considered that 
672: %We treated    regions I--A, I--B and II  as places where star formation took place coevally and therefore
673: the present--day mass function {\em is} the IMF.
674: The rest of the galaxy is  composed of a large number of star--forming regions
675: of different ages, and a fairly good approximation would be to 
676: consider it as a region of  continuous star formation.  For this case we corrected the star counts for 
677: the main--sequence life--time $\tau_{MS}$ at each  particular mass, using values from Table~4 in Paper~I.
678: 
679: 
680: \subsection{Correction of systematic effects}
681: 
682: In order to determine the IMF of our sample of stars, we 
683: analyzed four sources of systematic effects, as pointed out by  \cite{Maizetal05}:
684: (1) data incompleteness,
685: (2) appropriate  bin selection,  (3) unresolved objects in our sample,
686: and  (4)  mass diffusion.  
687: 
688: \subsubsection{Data incompleteness}
689: 
690: 
691: In the derivation of an IMF, a quantitative 
692: evaluation of completeness
693: of the photometric data is required.  
694: To obtain the completeness values along the main--sequence section of each evolutionary track,
695: we calculated its value on each point that defines the main--sequence using
696: the results from Section 2.3  in Paper~I; we also 
697: computed a weight defined as the difference between the age of this point on the evolutionary 
698: track, and the age of the corresponding point in the  nearest more massive   evolutionary track. 
699: The adopted  completeness value is the weighted mean in each interval. All values are
700: summarized in  Table~4 in Paper~I.
701: We determined the main--sequence lifetime of each available track by interpolating among the ages 
702: provided by the models.
703: %Table~\ref{tbl-5} gives the adopted values of main--sequence lifetime.
704: 
705: 
706: 
707: \subsubsection{Bin selection}
708: 
709: To calculate the slope of the IMF we followed the technique suggested
710: by      \cite{MaizUbed05}  
711: and used a weighted least square method
712: to fit  a power law $\frac{dN}{dm}  = A \cdot m^{\gamma} $
713: using   5 bins of  variable size, so that the number of stars in
714: each bin is approximately constant. This method guarantees that  the binning biases
715: will be minimum 
716: and that the uncertainty estimates are correct.
717: For each   bin, we used the weight $w_i$ derived from
718: a binomial distribution as  suggested by  \cite{MaizUbed05}.
719: 
720: 
721: \subsubsection{Unresolved objects}
722: 
723: 
724: In distant stellar systems like the MCs or NGC~4214, we encounter a serious problem: namely, 
725: crowding of stellar images which makes the IMF determination somewhat complicated.
726: At the distance of NGC~4214 (2.94 Mpc), 1 arcsec corresponds to $\approx$14 pc. We therefore
727: expect some of 
728: the regions in the galaxy to be  crowded from the observational point of view.
729: We may find  unresolved objects in our sample, i.e., merged stellar images, irrespective of
730: their nature as a physical system or a chance coincidence.
731: The reliability of the highest mass stars is questionable because of  stochastic effects due to  
732: small number statistics  and evolutionary effects. Some stellar evolution has 
733: certainly taken place for these 
734: stars and some of them may fall out of the main--sequence band.
735: 
736: The effect of unresolved objects in a sample used to determine a stellar IMF has been 
737: already analyzed by several authors.  
738: \cite{SaRi91} studied the effect of actual binaries
739: in the mass range $2-14 ~M_{\odot}$ by performing several Monte-Carlo experiments
740: using initial  values $\gamma = -3.5, -2.5,$ and $ -1.5$ and several binary fractions.
741: They come to the conclusion that the general effect 
742: is a flattening of the mass function.
743: With an intrinsic slope of $\gamma = -3.5$ there is hardly any influence
744: on the MF slope; and for $\gamma = -2.5$    the  MF slope undergoes a 
745: flattening of 0.34 if the binary fraction increases to 50\%.
746: They made the
747: assumption that stellar masses are distributed randomly among the components.
748: %They claim that any visible effect will be more pronounced in the low mass domain, since
749: %a faint companion to a bright star does not alter the total luminosity of the bright 
750: %object appreciably. 
751: \cite{Krou91} show that unresolved binary stars have a significant effect on any 
752: photometrically determined luminosity function. Their research leads to the conclusion
753: that when binaries are not taken into account, there is an underestimate on the 
754: number of low mass stars, leading to a flattening of the IMF slope.
755: Given the large multiplicity fractions 
756: observed for OB stars, this effect must be significant for intermediate-- and high--mass stars 
757: \citep{Maizetal05}. However, since the binary fraction is unknown  in our sample,
758: it is not possible to correct  the estimated  values of the IMF slope for the presence of binaries. 
759: We used $M_{up} = 100  M_{\odot}   $ as the high--mass end of the integration which  allowed us to 
760: discard some  unresolved objects in our sample, but 
761: this effect is difficult to solve for a galaxy located at the distance of 
762: NGC~4214 because, even with 
763: the high--resolution images provided by HST, we still observe 
764: multiple systems. 
765: %Unresolved objects produce a systematic 
766: %flattening of the IMF slope, which 
767: 
768: One additional  source of multiple objects  is  the blends of stars caused by 
769: chance alignments. The greater the distance to the object of study,
770: the more likely one is prone to encounter  these kind of objects.
771: %
772: %
773: Our analysis of multiple systems 
774: leads us to conclude that
775: the real value of the slope would be steeper than the 
776: value that we obtain with our study.
777: 
778: 
779: For the low--mass end of the integration, we used several  mass values  in the range 
780: $6, 6.5, ..., 25  M_{\odot}   $. The main problem regarding the low--mass end is the completeness
781: of our data, because the observational sensitivity limit is reached here.
782: We corrected the number of counts for incompleteness using the values given in  Table~4 in Paper~I,
783: as described above. The IMF slope values that we present were calculated for 
784:  $M_{low} = 20  M_{\odot} $  where the completeness is higher than 85 \%.
785: 
786: 
787: 
788: \subsubsection{Mass diffusion }
789: 
790: Using \cho\ we  translated uncertainties in the measured magnitudes and colors
791: into uncertainties in temperature (or bolometric magnitudes) and 
792: luminosity as we explained in Paper~I.  
793: We then  used the results from Section~4.2  to obtain the 
794: uncertainties in mass.
795: For our stellar mass range, the IMF has a negative slope, meaning
796: that there are more low--mass stars than high--mass ones.
797: Therefore, if 
798: we measure a star to have magnitude $m'$, there should be a higher probability 
799: that its real magnitude $m$ is dimmer than $m'$ than it is  brighter (i.e. there 
800: are more dim stars disguising as bright ones at a given measured magnitude 
801: than bright stars disguising as dim ones).
802: \cite{Maizetal05}  study this  artifact that appears when dealing with real data.
803: %
804: This  effect basically
805: smoothes the IMF slope by shifting dim objects (low--mass stars)
806: to where bright objects  (high--mass stars) are, and it is basically
807: due to the effect of uncertainties in temperature and luminosity of individual
808: stars.
809: %
810: To take care of this effect, we proceeded in the following way: First, a
811: function  $(\sigma_{M}(M))$ was computed from the values for the individual 
812: stars  $(M_i , \sigma_{M_i})$
813: by fitting a third degree polynomial. Then,  we used a random number generator to produce 
814: 50 lists
815: of masses in the range $  [10,200]  M_{\odot}$  for different  input values of  the IMF slope
816: which we called $\gamma_{real}$. Those values are in the range $-1.2 $ to $-4.0$ with step of 0.1.
817: We assigned  errors randomly  to the masses in our artificial lists to smooth the values
818: using   the function   $\sigma_{M}(M)$, and 
819:  simulated the incompleteness of the data using the values given in  Table~4 in Paper~I.
820: We then  fitted these lists with a power law using   5 bins of  variable size, so that the 
821: number of stars in
822: each bin is approximately constant. For the integration, we  used the same mass
823: limits   that
824: we had used for the real data.
825: This process allowed  us to fit  a polynomial to $\gamma_{fit} - \gamma_{real}$
826: as a function of $\gamma_{fit}$  which is the correction that needs to be applied
827: in order to obtain $ \gamma_{real}$ from the estimated $\gamma_{fit} $.
828: 
829: 
830: 
831: 
832: 
833: \subsection{IMF  Results}
834: 
835: 
836: The result of the process described in Sections 4.2 and 4.3
837:  is shown in Tables~\ref{tbl02} and ~\ref{tbl03}.
838: For each set of stars, we give the computed values 
839:  $\gamma_{fit}$, and $ \gamma_{real}$ that we obtained for six values of 
840:  the low mass   ($M_{low}$ ) used in the integration. We also present 
841:  the uncertainties  ($\sigma$) in $\gamma_{real}$  derived from the $\chi ^2$  fit 
842: as well as the number $(N)$ of stars used for each fit. 
843: Table~\ref{tbl02} gives the results for    
844: a continuous star--formation scenario.
845: Table~\ref{tbl03}  gives the results for    
846: a burst star--formation scenario.
847: %
848: We adopt  $ \gamma_{real} = -2.83 \pm 0.07$  as a 
849: representative value of the IMF slope of NGC~4214,  which we calculated
850: using  $M_{low} = 20 \, M_{\odot}   $ and $ M_{up} = 100 \, M_{\odot}$.
851: We used IDL's CURVEFIT  to fit the power law and the $\chi^2$ value
852: provided by the fit in the [20, 100] $M_{\odot}$ mass range is 1.94.
853: 
854: 
855: %An example of the fit polynomial used to correct the calculated  $\gamma_{fit} $ 
856: %for  $M_{low} = 20 M_{\sun}$ is given in the left panel of Figure~\ref{fig-52}.
857: %In the right panel of Figure~\ref{fig-52} we present an example of 
858: %a histogram of the counts
859: %as a function of the mass for   {\sc\footnotesize  LIST1}. This is a linear fit of $N=1363$ stars with masses
860: %in the range $20-100 M_{\odot}$. The fit yields the value $\gamma_{fit} =-2.58 $;
861: %once this value is corrected with the method described above, we obtain  
862: %$ \gamma_{real} = -2.83$. 
863: %
864: %
865: %Figure~\ref{fig-41} [left panel] shows the $\gamma_{fit}$ (thin line), and 
866: %$ \gamma_{real}$ (thick lines) values 
867: %obtained as a function of $M_{low}$ for the continuous star--formation models. 
868: %The green lines represent the values obtained using stars from {\sc\footnotesize  LIST1}; blue lines 
869: %show  the values calculated for the  objects in {\sc\footnotesize  LIST2}; and finally,
870: %the red lines represent   all stars in {\sc\footnotesize  LIST3}.  
871: %The  mean value of $ \gamma_{real}$  ( considering  $M_{low}$  in [15, 25] $M_{\odot}$  ) 
872: %for all the objects is $ \gamma_{real} = -2.78$.
873: %The  mean value of $ \gamma_{real}$ in the same interval of $M_{low}$   
874: %for all the stars in {\sc\footnotesize  LIST3}  is $ \gamma_{real} = -2.87$.
875: %
876: Since the correction for unresolved objects is uncertain  (given the large distance to
877: NGC~4214), and knowing that unresolved objects produce a systematic 
878: flattening of the IMF slope, we conclude that 
879: the real value of the slope calculated for {\sc\footnotesize  LIST1}
880:  would be steeper than $ \gamma= -2.83$.
881: %
882: The burst star--formation model may be only applied to the stars included in  
883: regions I--A, I--B and II ({\sc\footnotesize  LIST2}),
884: because these are compact star--forming regions produced as a single burst.
885: %these are the star--forming regions in NGC~4214.
886: %Figure~\ref{fig-41} [right panel] shows the $\gamma_{fit}$ (thin line), and 
887: %$ \gamma_{real}$ (thick lines) values 
888: %obtained as a function of $M_{low}$ for such group of stars in this particular scenario. 
889: %The  mean value of $ \gamma_{real}$  ( considering  $M_{low}$  in [15, 25] $M_{\odot}$  ) 
890: %for all the objects is $ \gamma_{real} = -3.57$.
891: %
892: Note that the correction supplied by the  polynomial is a decreasing  function of 
893: $M_{low}$. This correction is greater for low values of $M_{low}$  where the incompleteness of the
894: data plays a strong role. 
895: %
896: %
897: %
898: %
899: When considering objects from {\sc\footnotesize  LIST2}, one has to face two problems:
900: (1)  the  uncertainty 
901: in the  correction for unresolved objects as described above for  {\sc\footnotesize  LIST1},
902: and (2) the fact that these regions are composed of a mixture of stars with ages in
903: the range $0-5$ Myr.  This means that some of the
904: most  massive stars must have evolved and disappear
905: in the form of supernovae  emptying the high--mass bins and producing a steeper slope.
906: %
907: We studied this problem by considering stars in the following
908: mass ranges (in units of solar masses): [20,  40],  [20,  100],  and  [40,  100].
909: The values of the IMF slope are  $ \gamma_{real} = -4.41 \pm 0.28$,  $ \gamma_{real} = -3.56 \pm 0.13$,
910: and  $ \gamma_{real} = -2.97 \pm 0.37$  respectively.
911: %
912: %
913: These different values could be explained by claiming that
914: the fraction of multiple and blended systems is larger in the   [40,  100] range than in the 
915:  [20,  100].
916: This makes the slope to appear shallower than when using  $ M_{up} = 40  M_{\odot}$. 
917: In summary,  $ \gamma_{real} = -3.56$ is a lower value for the IMF slope
918: when using the burst star--forming model.
919: 
920: 
921: 
922: %Our estimation of the IMF slope in the  continuous star--formation 
923: %scenario is   shallower but close to  the one provided by  
924: \cite{Chanetal05}  estimate the value $ \gamma = -3.5$
925: for the field stars in NGC~4214  by matching integrated SEDs
926: to STIS spectra, 
927: and using $M_{low} = 1  M_{\odot} $ and 
928: $M_{up} = 100  M_{\odot}$. Their extracted field regions include
929: star clusters, and hence some additional massive stars that do not belong to the field.
930: This implies that their slope value represents a lower limit to the field IMF slope.
931: As a consequence of this, they argue that the field is much less likely 
932: to produce massive stars than the cluster environment.
933: Using a similar approach, but with $IUE$ spectra, 
934: \cite{MasHKunt99}  could constrain the IMF slope  of NGC~4214 to
935: $ \gamma = -3.0$. Their spectra clearly includes stars in clusters as well as field stars.
936: We expect the ``real'' IMF slope to be shallower than Chandar's and 
937: steeper than our calculated value ($ \gamma= -2.83$).
938: 
939: 
940: \section{Analysis of extended objects}
941: 
942: In Paper~I we described the process that we used to obtain aperture photometry 
943: of 13  stellar clusters ( I--As, I--Es, IIIs, IVs, I--A, I--B,
944:  I--Ds, II, II--A, II--B, II--C, II--D, and II--E )
945: in NGC~4214. The magnitudes are listed in  Tables~3 and 4
946: of that paper. We also   explained 
947: how we translated the observed magnitudes and photometric colors
948: into clusters  physical properties using \cho\ ~\citep{Maiz04}. 
949: We based our  work on clusters on   
950: \stb\ \citep{Leitetal99}  models  of integrated stellar populations.
951: Here we analyze the results.
952:  
953: 
954: \subsection{Unresolved clusters}
955: \subsubsection{Cluster I--As}
956: 
957: We start our analysis with cluster I--As. 
958: \cho\ was executed using the clusters
959: magnitudes from Table~6 in Paper~I, and  
960: leaving  all three parameters [$\log({\rm age}) $, \ecc ,  $R_{5495}$] 
961: unconstrained, which  gave us a well defined value 
962: of the clusters age: $\log(\rm {age/yr})=6.60     \pm 0.06 $, and a low value for the extinction 
963: $E(4405-5495) =  0.07 \pm 0.02$ mag.
964: The reddening vs. age likelihood contour plot (not shown) clearly indicates
965: that  there is only one solution compatible with the available photometry.
966: However, the low value obtained for the reddening 
967: implies a degeneracy in the \rv ~values, since the effect of  extinction  becomes
968: nearly  independent 
969: of the choice of extinction law   when \ecc $\ll 1.0$ mag.  This means that even though
970: we are able to restrict the age and extinction of this cluster, 
971: the appropiate value of \rv ~remains undefined using our set
972: of observed magnitudes.
973: 
974: 
975: Running the code with the same set of magnitudes, but constraining the extinction law
976: to a MC--type (average LMC, LMC2, or SMC) we obtained similar values, 
977: as shown in  Table~\ref{tbl04}.
978: Since we are fitting two parameters
979: and using four colors (derived from five magnitudes), the problem has 2 degrees of
980: freedom. 
981: The best fit, which is certainly degenerate,  appears to be 
982: the one obtained using the SMC extinction law,
983: according to the $\chi^2$ reduced value. 
984: Figure~\ref{fig03}    shows the SEDs that best agree with  the data for the LMC2
985: and SMC  fits. In Figure~\ref{fig04}  we   show the reddening vs. age likelihood contour
986: plots using the LMC2 and SMC extinction laws. Both plots show that 
987: the age of this cluster spans  the range  $6.50 \le   \log(\rm {age/yr}) \le  6.70 $.
988: 
989: \cho\ provides a time--dependent correction that can   
990: be applied to transform the current magnitude of a cluster to the one at zero age.
991: We used this quantity to derive  the  clusters zero age masses. For our
992: analysis, we used \stb\ models 
993: which   only include stars more massive than  $ 1 \, M_{\odot}.$ 
994: Since most of a cluster's mass is contained in the low--mass stellar population, we had to 
995: further correct these values  considering a realistic IMF that spans the mass 
996: range $0.1-100 \, M_{\odot}.$
997: %For this correction, we used a Kroupa IMF \citep{Krou02}. 
998: All our mass estimates in this paper
999: were  therefore calculated using a Kroupa IMF \citep{Krou02}
1000: which considers stars in the range $0.1-100 \, M_{\odot}.$
1001:   
1002:  
1003: %One of the  \cho\ outputs is  $C(t)$, the  time--dependent correction that has to 
1004: %be applied to transform the current magnitude of a cluster to the one at zero age.
1005: %We used it to calculate the  clusters zero age masses, which for 
1006: The derived mass of cluster I--As is $\approx 27\,000 M_{\odot}.$ 
1007: Once the approximate mass
1008: of the cluster was known, we used the \stb\ output to calculate the stellar content of a cluster
1009: with  this mass,   an SMC--like  metallicity $(Z=0.004)$  and the  appropriate age.
1010: The results of the evolutionary synthesis models are  shown in Table~\ref{tbl04}.
1011: It is  known that cluster I--As is a SSC with a compact core and a strong massive halo,
1012: very similar in structure to 30 Doradus in the LMC, as shown by 
1013: \cite{Maiz01b}. This implies that aperture effects are 
1014: very important when performing any kind of photometric measurements
1015: for this cluster, and the derived quantities such as the mass of the cluster
1016: and the stellar population will depend on the chosen aperture. 
1017:  
1018: Using an SMC extinction law, our models yield an age of $4.0 \pm 0.6$ Myr for 
1019: cluster I--As from the optical--UV
1020: photometry. This value agrees with the ones obtained by \citet{Leitetal96} 
1021: using UV spectroscopy (4--5 Myr) and
1022: by \citet{Maizetal98} using $W$(H$\beta$)  and the strength of the WR 4660 \AA\ blend (3--4 Myr).
1023: The models that best fit our data yield the values\footnote{Note that the value of \ecc ~is 
1024: slightly different from the one without restraining the extinction law
1025: but that  both  are separated by only $\approx 1 \sigma.$} $A_V = 0.19  \pm 0.06,$ and  
1026: $E(4405-5495) =  0.04  \pm 0.01 $ mag . This low extinction is expected, since I--As is located within a 
1027: heart--shaped \ha\ cavity created by the kinetic energy input of the cluster into its surrounding medium
1028: \citep{Maizetal99a,MacKetal00} and, therefore, one would expect little gas or dust in the line 
1029: of sight. 
1030: The quite large number of  massive stars formed here have apparently wiped 
1031: out the dust particles from their surroundings, leaving free paths through which
1032: the stellar continuum can emerge.
1033: The  low extinction agrees with the value measured at this location by \citet{Maizetal98} 
1034: from the nebular  ratio of \ha\ to \hb. 
1035: 
1036: From the detection of a broad emission feature from Wolf--Rayet stars (WRs), several authors   
1037: \citep{Leitetal96,SargFili91,MasHKunt91a,Maizetal98}
1038: indicate the presence of WR stars inside cluster I--As. 
1039: \cite{Leitetal96} reexamined the results of \cite{SargFili91}.
1040: Their  results were calculated from  ultraviolet 
1041: spectra of cluster I--As  obtained with
1042: the Faint Object Spectrograph onboard the HST in 1993. They used the $1  \farcs 0  $ 
1043: circular aperture
1044: and assumed a distance of 4.1 Mpc to NGC~4214.
1045: They estimated that approximately 
1046: 15  WN and 15 WC  stars are present inside cluster I--As, keeping
1047: in mind uncertainties of a factor of 2. These results translate to 5 WR stars 
1048: using the improved distance of 2.94 Mpc.
1049: %
1050: \cite{Maizetal98}  study four regions with WR stars in NGC~4214, including cluster I--As. 
1051: Using their  observed values 
1052: of  the equivalent widths of the WR blend we can estimate 
1053: 60 WR stars inside an area of radius $1  \farcs 8 $ assuming  a distance of 
1054: 4.1 Mpc.  This result is equivalent to  3 WR stars
1055: inside the aperture radius that we had used for cluster I--As at a distance
1056: of 2.94 Mpc.
1057: The \cho ~results  predict 4 WRs 
1058: using the same metallicity $(Z=0.020)$, and 1 WR using  $(Z=0.004)$,
1059: which is consistent with the observations.
1060: It is important to note that WR stars may be present both in the nucleus of the cluster, 
1061: and  they may also 
1062: be part of the massive halo. We want to emphasize that aperture effects are of great significance 
1063: in dealing with this cluster. This  makes it hard to compare different works performed with 
1064: different techniques; however, our results are within Poisson errors of previous published values.
1065: 
1066: 
1067: 
1068: \subsubsection{Cluster I--Es}
1069: We executed   \cho ~for cluster I--Es using  seven observed magnitudes (six colors)
1070: from Table~6 in Paper~I and  leaving all parameters unconstrained.
1071: The likelihood map produced by \cho\  shows two solutions in the [$\log({\rm age})$, \ecc] plane,  
1072: a young one around $\log({\rm age/yr}) \approx 6.85$, and an old one around $\log({\rm age/yr}) \approx 8.25$.
1073: We reran the code to isolate these two solutions using those age intervals,
1074: and the properties of both solutions are shown in Table~\ref{tbl04}. 
1075: The old solution (age of $189\pm53$  Myr)
1076: is the one that has the highest probability, but the young solution cannot be immediately rejected.
1077: The likelihood contour plots produced by \cho\ for this cluster are shown
1078:  in Figure~\ref{fig05}   for both solutions. The young solution shows 
1079: very well defined ellipses in the plots. The old solution yields a high value of
1080: \rv ~which is clearly visible in Figure~\ref{fig05}  (lower row, center and right.)
1081: The SED for both models  are given in Figure~\ref{fig03}.
1082: Both solutions show that cluster I--Es  is older that cluster I--As
1083: which  agrees with 
1084: \cite{MacKetal00} who indicate that 
1085:  the continuum colors of I--Es  are significantly redder than for I--As,
1086:  indicating a greater age. 
1087:  We executed \stb\   to estimate 
1088: the stellar content of a cluster with mass, age and metallicity appropiate 
1089: for both  solutions. The young solution  ($\approx 6\,000 M_{\odot}$ )  predicts the
1090:  presence of 1 red supergiant
1091: and 2 blue  supergiants, while the old solution ($\approx 129\,000 M_{\odot}$ )  predicts
1092: 9 red supergiants and zero blue supergiants. 
1093: 
1094: 
1095: Trying to discern between the two possible solutions leads us to give
1096: a word of caution about the predictions of stellar synthesis models: in general, theoretical
1097: models have intrinsic uncertainties, especially when dealing with fast--evolving phases such as the RSG one.
1098: Furthermore, a more subtle effect is caused by incomplete sampling.
1099: Most current synthesis codes predict the average values of observables for an {\em infinitely large population}
1100: of stars, which samples a given IMF completely. \cite{CervVall03} show that for an integrated property which
1101: originates from an effective number of stars $N$, there is
1102: a critical value of $N\approx 10$ below which the results of the codes must be taken with caution, since they 
1103: can be biased and underestimate the actual dispersion of the observables. 
1104: %
1105: Cluster I--Es is an example of this sampling problem, since its F814W magnitude is dominated by the presence of
1106: a few RSGs (both the old and young solutions predict this). Therefore, the integrated colors are not enough to 
1107: differentiate between the two solutions, since stochastic effects should be larger than evolutionary ones.
1108: Is there a way to overcome this issue? One could do it by looking at the resolved stellar population of the
1109: cluster, which is what we attempt in Figure~\ref{fig06} by
1110: representing cluster I--Es in detail in three filters: F336W,
1111: F555W, and F814W. A close inspection of these plots reveals that the cluster
1112: has some  non--radial  substructure.
1113: These  images show two objects that blend with the structure of cluster I--Es,
1114: which are located inside the aperture radius that we used for the photometry
1115: marked in green in Figure~\ref{fig06}.
1116: The  object located towards the north   is clearly visible in  three filters (and is likely to be a 
1117: BSG), while the other is not detectable in filter F336W (and is likely to be a RSG).
1118: The presence of a single bright red object favors the young solution over the old one, given that the first one
1119: predicts 1 RSG and the second one 9 of them. On the other hand, the bright red object could simply be a RSG 
1120: unassociated with the cluster that happens to lie in the same direction (a look at Figure~\ref{fig02} shows that
1121: this possibility is not unlikely). Therefore, we have to conclude that the  available data does not allow us to 
1122: discern between the two solutions.
1123: 
1124: 
1125: Regarding the extinction that affects I--Es, both solutions (young and old) yield values much higher than for I--As.
1126: This is consistent with the \ha/\hb\ results of \citet{Maizetal98} and \citet{Maiz00} and with the existence of a
1127: small peak in the CO distribution near the position of the cluster \citep{Waltetal01}. Also, it is interesting to
1128: note that the $R_{5495}$ obtained from either solution is higher than the standard 3.1. However, given the
1129: uncertainty in the characteristics of this cluster deduced from its integrated colors caused by its relatively low
1130: number of stars, we should regard this measurement of $R_{5495}$ as rather uncertain.
1131:  
1132: \subsubsection{Cluster IIIs}
1133: 
1134: To analyze cluster IIIs, we used all seven
1135: magnitudes   listed in Table~6 from Paper~I. We executed   \cho\ without limiting any  parameter 
1136: and using F336W as the reference filter. This gave us 7 magnitudes, 6 colors and
1137: 3 free parameters.
1138: We found an excellent fit of the spectrum with two possible solutions: a weak (high $\chi^2$) young solution
1139: around $\log({\rm age/yr}) \approx 7.20$ and a stronger (low $\chi^2$) around $\log({\rm age/yr}) \approx 8.20$.
1140: We reran the code twice to isolate these two solutions and 
1141: the results for both models are presented in Table~\ref{tbl04}.
1142: The likelihood contour plots for both solutions are presented in Figure~\ref{fig07}.
1143: Both solutions are very good fits, with small values of $\chi^2$.
1144: The slight difference between the two SEDs (Figure~\ref{fig08}) lies in 
1145: the depth of the bump located
1146: around  2175 \AA.This spectroscopic 
1147: feature  is known as  the graphite bump.
1148: Our study cannot disentangle between both solutions, since we do not have 
1149: any measured magnitude in this part of the spectrum. However, the relative position of this cluster
1150: in the galaxy suggests that  a low value of the extinction is to be expected.
1151: Also, the image of the cluster appears to be smooth, with no salient features 
1152: throughout its structure.
1153: These two properties suggest that the old solution 
1154:  (age of $168 \pm 61$  Myr, mass of  $(6 \pm 2) \times 10^5$  $M_{\odot}$ )
1155: should be preferred over the young solution.
1156: 
1157: 
1158: \subsubsection{Cluster IVs} 
1159: 
1160: To analyze cluster IVs, we used all seven
1161: magnitudes   listed in Table~6 from Paper~I. We executed   \cho\ 
1162: without limiting any  parameter 
1163: and using F555W as the reference filter.
1164: There are two distinct  solutions compatible with our set 
1165: of magnitudes:  a weak young solution
1166: around $\log({\rm age/yr}) \approx 7.20$
1167: and a more conspicuous old 
1168: around $\log({\rm age/yr}) \approx 8.20$.
1169: Figure~\ref{fig08} represents the best fit spectra for both solutions, and  Figure~\ref{fig09}
1170: shows the probability contour plots for the young (upper row) and old (lower row) 
1171: solutions. 
1172: %
1173: With our current magnitudes, it is not possible to decide between both models. 
1174: However, for  cluster IVs we expect a low value of the extinction since it is located 
1175: in a   region of the galaxy  with little gas and dust. Also, as in cluster IIIs, its smooth
1176: appearence favors the old solution (age of $150 \pm 34$  Myr) over the young model.
1177: As in cluster I--As, we find   large values of  \rv ~(spanning the range
1178: $2.0  \leqslant R_{5495}  \leqslant 6.0$), indicating  that 
1179: for a low value of the extinction, \rv ~is degenerated.
1180: In order to disentangle this problem,
1181: it would be useful to measure a magnitude near the right side of  the 
1182: Balmer discontinuity  such as F439W (WFPC2 $B$).
1183: 
1184: 
1185: In  Table~\ref{tbl04} we present the number of K+M stars of types I and II, obtained from 
1186: \stb.
1187: It is noticeable  that in   clusters IIIs and IVs, the young solution provides a much lower quantity 
1188: of RSGs than the old solution. The larger number of RSGs would explain the smooth 
1189: appearance observed in the images, and this is another fact that favors the older solution
1190: for clusters IIIs, and IVs.
1191: \cite{Larsetal04} estimate  $\sim (200 \pm 52 ) $ Myr as  the age of clusters IIIs and IVs,
1192: which is consistent with our results.
1193: 
1194: \subsection{Resolved clusters}
1195: We used five  magnitudes from   Table~7 in Paper~I
1196: to execute \cho ~for cluster I--Ds. We left 
1197:  all parameters
1198: [  $\log({\rm age}) $, $E(4405-5495)$, $R_{5495}$ ] unconstrained, which 
1199: means that we had only one degree of freedom in this run.
1200: Figure~\ref{fig10} shows the best fit spectrum to the measured data.
1201: The contour plots  for this cluster are presented in 
1202: Figure~\ref{fig11}  (left column.)
1203: There appears to be two possible solutions, both of them very young, with a strong peak 
1204: at   $\log({\rm age/yr}) \approx 6.5 $.  The output of the code yields
1205: a mean age of  $2.6 \pm 1.5  $ Myr, and a mean  extinction  of 
1206: $ E(4405-5495) =  0.28 \pm 0.05$ mag. The results are listed in Table~\ref{tbl05}.
1207: 
1208: 
1209: The rest of the clusters which are part of the structure of  complex II,
1210: were analyzed in a similar way. We used the set of magnitudes
1211:  listed in  Table~7 from Paper~I  with the exception of magnitudes F555W and F702W
1212:  for the reason explained below. We obtain fairly good fits 
1213:  (highest $\chi^2 = 2.53 $ ) which we display in 
1214:  Figure~\ref{fig10}. Note that the points that correspond to 
1215:  magnitudes F555W and F702W are plotted, but they were not considered 
1216:  during the  execution of \cho\ because those magnitudes are heavily contaminated by nebular emission. 
1217: %Figures~\ref{fig11} and ~\ref{fig-48}
1218: %present the contour plots for these clusters. 
1219: 
1220: 
1221: The equivalent width of the Balmer lines can be used to estimate the
1222: age of a star--forming region \citep{CervMasH94}. The expected values for 
1223: $W$(H$\alpha$) are $1000-2500$ \AA\ for very young clusters (age $ \lesssim 3$ Myr) 
1224: and $500-1000$ \AA\ for ages in the range between 3 and 4 Myr.
1225: \cite{MacKetal00}  estimated this parameter using three extinction corrections for all
1226: the clusters in NGC~4214. 
1227: 
1228: The first execution for cluster II--A showed two possible solutions
1229: compatible with our data. One in the range 
1230:   $0.1  \leqslant   E(4405-5495)     \leqslant 0.26$ (in mag) and the other
1231: in the range   $  0.26  \leqslant   E(4405-5495)   \leqslant 0.6$ (in mag). 
1232: We reran the code to isolate the solutions individually, and 
1233: here, we present   the diagrams corresponding 
1234: to the second  case only.
1235: \cite{MacKetal00} provide values of $W$(H$\alpha$) for all the clusters
1236: within NGC~4214--II.  These are close to or higher than $1000$ \AA\ 
1237: indicating a very young  age of this complex.  This fact led us to 
1238: favor the solution with the highest $E(4405-5495)$.
1239: With this method we can restrict the age of this cluster to 
1240: $3.1 \pm 1.4  $ Myr. This means that there may be WR stars
1241: within its structure. We also obtain a large value of \rv, compatible 
1242: to the one  calculated for cluster II.
1243: 
1244: 
1245: As for cluster II--B, the first execution showed two 
1246: potential solutions with the same mean value of \rv.  
1247: We isolated a young and an old solution.  Here we present 
1248: the plots that correspond to   the younger one. 
1249: Following the same reasoning as for II--A, we favor the younger solution
1250: because high values of $W$(H$\alpha$)  are measured in this region of the galaxy.
1251: This solution yields a very small age: $2.0 \pm 0.8  $ Myr.
1252:  
1253: 
1254:  
1255: 
1256: The results of the best fits for the other three clusters are given in Table~\ref{tbl05}.
1257: %and their contour plots are presented in Figure~\ref{fig-48}.
1258: Cluster II--C  yields a single  young solution, while II--D
1259: and II--E have  more complicated outputs; however, their reddenings are 
1260: very similar, as well as their  \rv\ values.
1261: 
1262: 
1263: 
1264: 
1265: 
1266: 
1267: \subsection{Large complexes}
1268: 
1269: For complexes  I--A, I--B, and II we ran \cho~ 
1270: leaving all three parameters
1271: [ $\log({\rm age}) $, \ecc,  $R_{5495}$ ] unconstrained.
1272: We used the estimated magnitudes from  Table~6 in Paper~I.
1273: Note that for complex~II,
1274: magnitudes F555W and F702W are listed, but they were not considered
1275: for the $\chi^2-$fit. 
1276: The inclusion of the 2MASS magnitudes did not change the output of the code 
1277: significantly in any case.
1278: %
1279: The best--fit  SEDs are presented in Figure~\ref{fig12}.
1280: The fits of complexes I--A, and I--B are excellent. The F170W magnitude 
1281: is the one that contributes the most to $\chi^2 = 4.25$ in the case of cluster I--B.
1282: Figure~\ref{fig11} shows the contour plots provided as output by \cho\ for cluster I--B.
1283: It is clear from the  plots that for cluster  I--B there is a single solution 
1284: compatible with our photometry. We obtain the same age (5.0 Myr) for both of them 
1285: and, as expected, a low value for the extinction $ E(4405-5495) =  0.07$ mag.
1286: This value is very close to the one obtained for cluster I--As, as we
1287: had expected.
1288: It is important to note that the aperture that we used in both cases is large enough
1289: to include  a part of the galaxy that contains a mixed population of stars. 
1290: Figure~2 in Paper~I  clearly shows both late-- and early--type stars are present
1291: within the apertures, something unexpected for an age of 5.0 Myr.  
1292: Our extinction analysis showed that these large complexes contain stars with  variable
1293: extinction as well. 
1294: This leads to conclude that
1295: the SEDs for clusters I--A and I--B are composite spectra of different complex 
1296: stellar populations. 
1297: Again, the degeneracy in the extinction law for low values of the reddening gives 
1298: high values or \rv.
1299: An estimate of the total mass of these clusters gives:  $(156 \pm 19) \times 10^3$  
1300: $M_{\odot}$ for cluster I--A,
1301: and  $(34 \pm 4) \times 10^3$  $M_{\odot}$ for cluster I--B. All the results
1302: are presented in Table~\ref{tbl06}.
1303: 
1304: 
1305: For cluster II we used the largest aperture radius: 200 PC pixels. This means that we 
1306: are considering a mixed population of stars embedded in a gas--rich environment.
1307: As a result, we obtained
1308: a composite spectrum characterized by an age of  $1.9 \pm 0.9$ Myr. Its estimated mass
1309: is    $(9 \pm 3) \times 10^5$  $M_{\odot}$.  The right column in Figure~\ref{fig11}
1310: shows the contour plots obtained for cluster II. The top plot indicates that this cluster is 
1311: younger than cluster I--A, and that it is actually extinct. The middle and 
1312: bottom diagrams imply that the law in the \rv--dependent family of \citet{Cardetal89}
1313:  that best fits  our photometry has a very high value or \rv.
1314: 
1315: Complex II is expected to house objects with high extinction, since these are located within or very near 
1316: filled \ha\ regions that show little evidence of wind--blown 
1317: or supernova created bubbles \citep{Maizetal98,MacKetal00}.
1318: The extinction map represents very well this part of NGC~4214, showing the
1319: highest values of extinction in the southern part of cluster II--A and between clusters II--B and II--C.
1320: The measured values for the stellar \ecc\ are in good agreement with those obtained from \ha/\hb\ 
1321: \citep{Maizetal98}.
1322: We also calculated the mean value of the stellar extinction in each of 
1323: the clusters  (II--ABCDE) 
1324:  inside complex NGC~4214--II and found excellent agreement with the values obtained 
1325: by fitting \stb\ models to the spatially--integrated magnitudes.
1326: One remarkable  result from   Table~\ref{tbl06} is that  our models yield high values
1327: of \rv\ when we analyze large complexes. When using a large aperture 
1328: one includes a mixed population of stars; we analyzed the \rv\ output for 
1329: the brightest stars in this region and found that individual stars have indeed large values of 
1330: extinction, specially those in cluster II--B. 
1331: It is well known that the value of \rv\ depends upon the environment
1332: along the line of sight. 
1333: A direction  through
1334:  low--density ISM usually has rather low value of extinction (about 3.1). Lines of 
1335:  sight penetrating into  dense molecular clouds like
1336:  Orion, Ophiuchus
1337:  or Taurus  yield  $4 < R_V< 6$  \citep{Math90}.
1338: For example, star Her~36 in M~8 has $R_{5495} = 5.39 \pm 0.09$ \citep{Ariaetal05}.
1339: 
1340: 
1341: 
1342: \section{Summary and conclusions}
1343: 
1344: \subsection{Extinction}
1345: 
1346: Using the estimated values of stellar extinction from the output 
1347: of {\sc\footnotesize  CHORIZOS}, we built  an  
1348: extinction map in the field of view where we performed our study.
1349: The most noticeable  characteristic  in this map are the
1350: low values of the extinction scattered throughout the map, with
1351: the exception of  some well defined
1352: regions with high values.   
1353: %
1354: We compared our extinction map with results provided by 
1355: \cite{Maizetal98} and \citet{Maiz00} who used the Balmer 
1356: ratio (H$\alpha$/H$\beta$) as a tracer of the
1357: reddening.
1358: They  found  
1359: that the reddening in NGC 4214--II is, on average, higher than in NGC 4214--I. 
1360: Our results, derived
1361: from the stellar colors, are in agreement.  
1362: The two main cavities in NGC 4214--I show low 
1363: extinction surrounded by higher values while for NGC 4214--II the extinction is overall higher.
1364: %
1365: Our extinction 
1366: map traces fairly well  the molecular clouds studied by  
1367: \cite{Waltetal01} which are directly associated to the star--forming regions
1368: in NGC~4214.
1369: 
1370: 
1371: 
1372: Studying the stellar content of seven blue compact galaxies,  
1373: \cite{Faneetal88} found a discrepancy between the  extinction derived from 
1374: the  UV continuum of starburst galaxies and that derived from the Balmer lines.
1375: The extinction derived from UV continuum was systematically lower. 
1376: %
1377: \cite{Calzetal94}   analyzed $IUE$ UV and optical spectra of 39 starburst and blue 
1378: compact galaxies
1379: in order to study the average properties of dust extinction. They derived 
1380: the UV and optical extinction law
1381: under the hypothesis  that the dust is a screen in front of the source.
1382: The characteristics of the extinction law are different from the 
1383: MW and LMC laws: the overall slope is more gray than the 
1384: MW or LMC slopes, and, most remarkably, the 2\,175 \AA\ dust feature is 
1385: absent within the observational uncertainties.
1386: The different slope explained the differences observed between stellar and nebular extinction.
1387: %
1388: %
1389: However, \cite{MasHKunt99} showed that the observed UV to optical SEDs of 
1390: 17 starburst  galaxies can be very well reproduced by reddening the 
1391: corresponding synthetic spectra with one of the three extinction laws 
1392: (Galactic, LMC, and SMC) with no need to invoke an additional universal law.  
1393: As proposed by \cite{MasHKunt99} and \cite{MacKetal00}, 
1394: the   effect was merely geometrical: while the continuum 
1395: flux comes from the stars, the nebular emission originates in 
1396: extended regions adjacent to  the original molecular cloud. 
1397: Stellar winds and supernova explosions might wipe out the dust 
1398: from the neighborhood of the massive clusters and concentrate it in 
1399: filaments and dust patches located within the ionizing region.
1400: %
1401: Depending on the specific geometrical distribution of dust and stars, 
1402: the extinction could affect mainly the nebular gas emission but 
1403: only weakly the stellar continuum. 
1404: This implies that the attenuation law by \cite{Calzetal94}  applies only to regions 
1405: like I--A  in NGC~4214, where the stars have evolved long enough to disrupt the ISM; 
1406: meanwhile, this law should not be applied to NGC~4214--II, because 
1407: the distribution of stars and  dust are co--spatial in this part of the galaxy.
1408: %
1409: %
1410: 
1411:   
1412: \cite{MacKetal00} analyzed the differential extinction of the gas in NGC~4214
1413: and found that  once 
1414: a cluster becomes older than $\approx 2$ Myr, 
1415:  the stellar components are mostly concentrated in those 
1416: regions where the gas shows very low extinction.
1417: In our current work, we studied the differential extinction of the continuum from the 
1418: stars in the galaxy.
1419: With this research we arrive to the conclusion that  the extinction derived from the stellar continuum  
1420: is similar to the extinction derived via the analysis of 
1421: nebular lines across the galaxy, and we find this to be  true on a pixel by pixel  basis.
1422: The coincidence is fairly good throughout the galaxy.
1423: This confirms the idea advanced by \cite{MasHKunt99} and \cite{MacKetal00} 
1424: that the differences between the   \cite{Calzetal94}  attenuation law and other extinction laws
1425: are caused by  the differences in the spatial distribution
1426: of the ionized gas and the young stellar population. 
1427: 
1428:  
1429: \subsection{The ratio of blue--to--red supergiants}
1430: 
1431: We approached the problem of determining the ratio    of blue to
1432: red supergiants (B/R) in NGC~4214. 
1433: We compared our observational results with three sets of theoretical models
1434: calculated with the MC metallicities and arrived at the conclusion that 
1435: the best fit to  our data is the non--rotating,  $Z=0.004  $, Geneva set
1436:   \citep{MaedMeyn01}.
1437: These models predict a B/R value of 24 in the $ 15-20 \,M_{\sun}$ mass--range, and 
1438: we measure  $34 \pm 10$. In the mass--range $ 20-25 \,M_{\sun}$ the
1439: theoretical prediction is 47 and we obtain  $46 \pm 23$.
1440: In both cases, our results agree with  the theoretical ones  within Poisson errors. 
1441: We discussed two caveats in the determination of B/R: the
1442:  stochastic effects due 
1443: to small--number statistics of  RSGs in our sample in NGC~4214, and
1444: the conversion from observed colors to effective temperatures and
1445: bolometric magnitudes. We observe a discrepancy of $350-400$ K in our data
1446: which may be accounted for by 
1447: fitting the  observed optical colors using \cho\ or a similar code 
1448: with  MARCS atmospheres instead of   Kurucz atmospheres.
1449: 
1450: \subsection{The initial mass function} 
1451: We studied the initial mass function of NGC~4214 following the 
1452: method provided by  \cite{Lequ79} with several  improvements, and 
1453: taking into account four sources of systematic effects
1454: (incompleteness of the data, optimum bin size, mass diffusion, and
1455: unresolved multiple systems)  presented by 
1456:  \cite{Maizetal05}. 
1457: We obtained  a mean value of $ \gamma= -2.83$  for the IMF slope
1458: of NGC~4214 in a continuous star--formation scenario.
1459: Since the correction for unresolved objects is uncertain  (given the large distance to
1460: NGC~4214), and knowing that unresolved objects produce a systematic 
1461: flattening of the IMF slope, we conclude that 
1462: the real value of the slope would be steeper than $ \gamma= -2.83$. 
1463: We expect the ``real'' IMF slope to be shallower than the one
1464: provided by  \cite{Chanetal05} ($ \gamma = -3.5$) and 
1465: steeper than our calculated value.
1466: Our estimation is   closer to 
1467: the one calculated by 
1468: \cite{MasHKunt99} ($ \gamma = -3.0$) for the sum of field and clusters
1469: in NGC~4214.
1470: %
1471: Some OB associations in the Magellanic Clouds and the Milky Way
1472: \citep{Masse98} 
1473: have comparable values of the IMF slope 
1474: to the one  that we obtained for NGC~4214.
1475: 
1476: 
1477: \subsection{Clusters}
1478:  
1479: We searched for the best fits of \stb\ models to photometric colors
1480: of 13 clusters in NGC~4214.
1481: We chose four unresolved compact clusters, three large complexes and six
1482: small resolved clusters.
1483: 
1484: We find that the best models that are compatible with
1485:  our data of cluster I--As are those with a MC--like extinction law. 
1486: These models predict that  cluster I--As is a young  ($4.0 \pm 0.6$ Myr) 
1487: massive  ($\approx 27\,000 M_{\odot}$) star cluster
1488: with low ( $E(4405-5495) = 0.04 $ mag ) extinction.
1489: Our study is consistent with  the presence of WR stars in this cluster, with 
1490: a number close to  the value previously obtained by \cite{Leitetal96}
1491: and \cite{Maizetal98}.
1492: %
1493: %
1494: \cite{MacKetal00}  estimated the equivalent widths of the Balmer lines
1495: $W$(H$\alpha$) 
1496:  using three extinction corrections for all
1497: the clusters in NGC~4214, in order to restrict the clusters ages.
1498: %
1499: For clusters NGC~4214--I--A and  I--B they estimate an age in the range between 
1500: 4 and 5  Myr, which
1501: agrees with our photometric method. 
1502: Their  age estimate for cluster  I--Ds is 7 Myr, while we obtain a much smaller
1503: value: $2.6 \pm 1.5$ Myr.
1504: Cluster I--Es provides two very different solutions, but with 
1505:  the current data
1506: we prefer not to favor any of the solutions over the other. 
1507: %
1508: 
1509: \cite{MacKetal00}   find equivalent widths of \ha\
1510: very close or higher than 1000 \AA\ for all the knots in NGC~4214--II,
1511: except for II--C, indicating that this large complex  is very young. We consistently obtain
1512: very small ages for all the clusters within  NGC~4214--II, including cluster II--C. 
1513: The overall value of the $W$(H$\alpha$) for NGC~4214--II is quite lower than 1000 \AA.
1514: \cite{MacKetal00}  explain that this may be due to the existence of an underlying population.
1515: They took this effect into account
1516: and estimated the  value $2.5-3.0$ Myr for the age of   cluster  II. Our 
1517: age estimate is $1.9 \pm 0.9$ Myr which agrees with their value.
1518: %
1519: The  $W$(H$\alpha$)  for NGC~4214--I is in general lower than those
1520: of  for NGC~4214--II, and these authors suggest an average age in the
1521: range $3.0-4.0 $ Myr for cluster I. 
1522: 
1523: 
1524: Our analysis for clusters IIIs and IVs includes six photometric colors
1525: covering wavelengths from the UV to the IR. This allowed us to find restrictions 
1526: to the age and mass of these clusters.
1527: Using several arguments (extinction values, number of red supergiants, image appearance)
1528: we suggest that the older solutions should be preferred 
1529: over the younger ones in the case of these two clusters.
1530: \cite{Billeetal02}   study clusters IIIs and IVs as part of a survey of compact star
1531: clusters in nearby galaxies which include NGC~4214. Unfortunately, their aperture photometry 
1532: includes  errors such as not taking into consideration the contamination effect 
1533: for filter F336W and the conversion of HST flight magnitudes
1534: to the Johnson--Cousins system, which  is not recommended 
1535: because of  their limited precision \citep{Gonzetal03,deGretal05}. 
1536: Using an evolutionary track taken from the  \stb\ models on a color--color diagram,
1537: they infer age estimates to these clusters. The evolutionary track on this diagram
1538: is clearly degenerate, and shows 
1539: several possible solutions for the age of these clusters.
1540: They used only two photometric colors, and this  translates
1541: into a very poor determination of the clusters age  \citep{deGretal03a}. 
1542: Using  the photometry produced by   \cite{Billeetal02} and the cluster 
1543: evolutionary models of  \cite{BruzChar03},    \cite{Larsetal04} infer the age 
1544: and mass of clusters IIIs and IVs. The mass is inferred  using a dynamical method
1545: for clusters in virial equilibrium.
1546: 
1547: 
1548: 
1549: %
1550: We find that  NGC~4214--IIIs and --IVs  are compact old massive clusters with age $> 100$ Myr.
1551: Ten other clusters (one unresolved compact cluster,  three large complexes and six
1552: small resolved clusters) yield ages  $< 10$ Myr.
1553: Cluster luminosity functions and color distributions are the most important tools in
1554: the study of cluster populations in nearby galaxies.
1555: The use of individual cluster spectroscopy, acquired with 8--m--class telescopes
1556: is very time consuming, because observations of large numbers of clusters
1557: are needed in order to obtain statistically significant results.
1558: Multipassband  imaging is a useful alternative. \cite{deGretal03a} analyze 
1559: the systematic uncertainties in age, extinction, and metallicity determinations
1560: for young stellar clusters, inherent to the use of broad--band, integrated colors. 
1561: They studied clusters within NGC~3310 and  found that  red--dominated 
1562: passband combinations result in significantly different
1563: age solutions, while blue--selected passband combinations tend to result in age 
1564: estimates that are slightly skewed towards lower ages. 
1565: Their advise is to use at least four filters including both blue and red optical
1566: passbands. This choice leads to the most representative age distribution. 
1567: See also \cite{Andeetal04} and  \cite{deGretal05}.
1568: 
1569: For our  cluster analysis in NGC~4214, we employed at least five  filters in each case,
1570: and we determined all the free parameters individually for
1571: each cluster as suggested by \cite{deGretal03b}.
1572: The accuracy to which the ages can be estimated depends on the number of 
1573: different broad--band filters and, crucially, on the actual wavelengths range covered by 
1574: the observations. 
1575: %
1576: We found some degeneracies that could have been disentangled, 
1577: provided that we had a measurement of a
1578: magnitude near the right side of  the 
1579: Balmer discontinuity  such as F439W (WFPC2 $B$).
1580: However, we are confident with the age estimate of our sample of clusters
1581: in NGC~4214.
1582: 
1583: 
1584: 
1585: Employing multicolor images of the Antennae galaxies, \cite{Falletal05}
1586: studied the age distribution of the  population of star clusters.
1587: They estimated the cluster ages, by fitting SEDs from the \cite{BruzChar03} models and found
1588: the age distribution declines steeply, starting at very young ages. The median age of the clusters
1589: is  $\sim10$ Myr. According to their study, after $\sim10$ Myr, the surface brightness of the clusters
1590: would be 5 mag fainter than initially (at $\sim1$ Myr), and therefore
1591: the cluster would disappear among the statistical fluctuations in the foreground 
1592: and background 
1593: of field stars. They call this effect the ``infant mortality" of clusters.
1594: Our sample of clusters in NGC~4214 includes 10 objects younger than  10 Myr,
1595: with NGC~4214--IA and --IB already showing effects of disruption. 
1596: What could cause the disruption of the clusters?
1597: \cite{Falletal05} claim that the momentum output 
1598: from massive stars comes in the form of ionizing radiation, stellar winds, jets, and 
1599: supernovae, and all these processes could easily remove much of the ISM from the
1600: protocluster leaving the stars within it gravitationally unbound. 
1601: Another explanation is provided by \cite{Claretal05} who performed
1602: simulations and showed that unbound giant
1603: molecular clouds manage to form a series of star clusters and disperse 
1604: in $\sim10$ Myr, making them transient features. At later
1605: ages they would not be recognized as clusters.
1606: This result is further confirmed  by several empirical and theoretical 
1607: studies. See   \cite{LadaLada03} for a review and the references therein.
1608: %
1609: \cite{Lameetal05a}  present a simple analytical description
1610: of the disruption star clusters in a tidal field, and found that about half of the
1611: clusters in the solar neighborhood become unbound within about 
1612: 10 Myr.
1613:  
1614:  
1615: \subsection{Implications}
1616: Nearby galaxies provide ideal laboratories  to test how stars form, how 
1617: star formation is triggered, and details of how galaxies assembled. 
1618: The study of nearby galaxies like NGC~4214 is rather important because 
1619: most of the information that we infer about  high--redshift  galaxies relies on
1620: what we observe in galaxies in the local universe. Therefore, understanding
1621: nearby galaxies is of extreme importance to comprehend what
1622: is going on in more distant ones.
1623: NGC~4214 is a low--metallicity galaxy, and this gave us the possibility
1624: to study the physical conditions   in an environment 
1625: of current astrophysical interest.
1626: %with the closest resemblance  to the 
1627: %early universe. 
1628: In particular, we could infer that the B/R supergiant ratio
1629: is close to the one of the SMC
1630: and that its IMF is steeper that Salpeter. 
1631: 
1632: %New F330W, F555W, and F814W ACS images have recently become available in the STScI Archive
1633: %and they are valuable observations which might help to  constrain the age and mass  
1634: %of the clusters.
1635: %These observations will also be very useful to accurately 
1636: %determine  the star formation history of this galaxy by matching synthetic CMDs.
1637: 
1638: 
1639: 
1640: 
1641: \acknowledgments
1642: %We want to thank Rupali Chandar for useful comments
1643: %that helped improve this paper.
1644: We want to thank the referee, Dr. Richard de Grijs, whose constructive and detailed 
1645: report helped immensely to  improve the quality of this paper.
1646: We also want to thank Rupali Chandar and Claus Leitherer  for useful comments on the
1647: original draft of this work. 
1648:  
1649: 
1650: Support for this work was provided by NASA through grants GO--06569.01--A, 
1651: GO--09096.01--A, GO--09419.01--A, and 
1652: AR--09553.02--A from the Space Telescope Science Institute, Inc., under NASA contract 
1653: NAS5--26555.
1654: This research has made use of the VizieR catalogue access tool, CDS, Strasbourg, France.
1655: This publication makes use of data products from the Two Micron All Sky Survey, which
1656:  is a joint project of the University of Massachusetts and the Infrared Processing and 
1657:  Analysis Center/California Institute of Technology, funded by the National Aeronautics 
1658:  and Space Administration and the National Science Foundation.
1659: 
1660: Facilities: \facility{HST(WFPC2)}, \facility{HST(STIS)}.
1661: 
1662: %\bibliography{ms}
1663: %\bibliographystyle{apj}
1664: %\bibliography{../general}
1665: %\bibliographystyle{apj}
1666: \begin{thebibliography}{80}
1667: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
1668: 
1669: \bibitem[{{Anders} {et~al.}(2004){Anders}, {Bissantz}, {Fritze-v.~Alvensleben},
1670:   \& {de Grijs}}]{Andeetal04}
1671: {Anders}, P., {Bissantz}, N., {Fritze-v.~Alvensleben}, U., \& {de Grijs}, R.
1672:   2004, \mnras, 347, 196
1673: 
1674: \bibitem[{{Annibali} {et~al.}(2003){Annibali}, {Greggio}, {Tosi}, \&
1675:   {Leitherer}}]{Annietal03}
1676: {Annibali}, F., {Greggio}, L., {Tosi}, M., \& {Leitherer}, C. 2003, \aj, 126,
1677:   2752
1678: 
1679: \bibitem[{{Arias} {et~al.}(2006){Arias}, {Barb{\'a}}, {Apell{\'a}niz},
1680:   {Morrell}, \& {Rubio}}]{Ariaetal05}
1681: {Arias}, J.~I., {Barb{\'a}}, R.~H., {Apell{\'a}niz}, J.~M., {Morrell}, N.~I.,
1682:   \& {Rubio}, M. 2006, \mnras, 366, 739
1683: 
1684: \bibitem[{{Billett} {et~al.}(2002){Billett}, {Hunter}, \&
1685:   {Elmegreen}}]{Billeetal02}
1686: {Billett}, O.~H., {Hunter}, D.~A., \& {Elmegreen}, B.~G. 2002, \aj, 123, 1454
1687: 
1688: \bibitem[{{Brunish} {et~al.}(1986){Brunish}, {Gallagher}, \&
1689:   {Truran}}]{Brunetal86}
1690: {Brunish}, W.~M., {Gallagher}, J.~S., \& {Truran}, J.~W. 1986, \aj, 91, 598
1691: 
1692: \bibitem[{{Bruzual} \& {Charlot}(2003)}]{BruzChar03}
1693: {Bruzual}, G. \& {Charlot}, S. 2003, \mnras, 344, 1000
1694: 
1695: \bibitem[{{Calzetti} {et~al.}(2004){Calzetti}, {Harris}, {Gallagher}, {Smith},
1696:   {Conselice}, {Homeier}, \& {Kewley}}]{Calzetal04}
1697: {Calzetti}, D., {Harris}, J., {Gallagher}, J.~S., {Smith}, D.~A., {Conselice},
1698:   C.~J., {Homeier}, N., \& {Kewley}, L. 2004, \aj, 127, 1405
1699: 
1700: \bibitem[{{Calzetti} {et~al.}(1994){Calzetti}, {Kinney}, \&
1701:   {Storchi-Bergmann}}]{Calzetal94}
1702: {Calzetti}, D., {Kinney}, A.~L., \& {Storchi-Bergmann}, T. 1994, \apj, 429, 582
1703: 
1704: \bibitem[{{Cardelli} {et~al.}(1989){Cardelli}, {Clayton}, \&
1705:   {Mathis}}]{Cardetal89}
1706: {Cardelli}, J.~A., {Clayton}, G.~C., \& {Mathis}, J.~S. 1989, \apj, 345, 245
1707: 
1708: \bibitem[{{Cervi{\~ n}o} \& {Valls-Gabaud}(2003)}]{CervVall03}
1709: {Cervi{\~ n}o}, M. \& {Valls-Gabaud}, D. 2003, \mnras, 338, 481
1710: 
1711: \bibitem[{{Cervi{\~n}o} \& {Mas-Hesse}(1994)}]{CervMasH94}
1712: {Cervi{\~n}o}, M. \& {Mas-Hesse}, J.~M. 1994, \aap, 284, 749
1713: 
1714: \bibitem[{{Chandar} {et~al.}(2005){Chandar}, {Leitherer}, {Tremonti},
1715:   {Calzetti}, {Meurer}, \& {de Mello}}]{Chanetal05}
1716: {Chandar}, R., {Leitherer}, C., {Tremonti}, C.~A., {Calzetti}, D., {Meurer},
1717:   G.~R., \& {de Mello}, D. 2005, \apj, 628, 210
1718: 
1719: \bibitem[{{Clark} {et~al.}(2005){Clark}, {Bonnell}, {Zinnecker}, \&
1720:   {Bate}}]{Claretal05}
1721: {Clark}, P.~C., {Bonnell}, I.~A., {Zinnecker}, H., \& {Bate}, M.~R. 2005,
1722:   \mnras, 359, 809
1723: 
1724: \bibitem[{{de Grijs} {et~al.}(2005){de Grijs}, {Anders}, {Lamers}, {Bastian},
1725:   {Fritze-v.~Alvensleben}, {Parmentier}, {Sharina}, \& {Yi}}]{deGretal05}
1726: {de Grijs}, R., {Anders}, P., {Lamers}, H.~J.~G.~L.~M., {Bastian}, N.,
1727:   {Fritze-v.~Alvensleben}, U., {Parmentier}, G., {Sharina}, M.~E., \& {Yi}, S.
1728:   2005, \mnras, 359, 874
1729: 
1730: \bibitem[{{de Grijs} {et~al.}(2003{\natexlab{a}}){de Grijs}, {Bastian}, \&
1731:   {Lamers}}]{deGretal03b}
1732: {de Grijs}, R., {Bastian}, N., \& {Lamers}, H.~J.~G.~L.~M. 2003{\natexlab{a}},
1733:   \mnras, 340, 197
1734: 
1735: \bibitem[{{de Grijs} {et~al.}(2003{\natexlab{b}}){de Grijs},
1736:   {Fritze-v.~Alvensleben}, {Anders}, {Gallagher}, {Bastian}, {Taylor}, \&
1737:   {Windhorst}}]{deGretal03a}
1738: {de Grijs}, R., {Fritze-v.~Alvensleben}, U., {Anders}, P., {Gallagher}, J.~S.,
1739:   {Bastian}, N., {Taylor}, V.~A., \& {Windhorst}, R.~A. 2003{\natexlab{b}},
1740:   \mnras, 342, 259
1741: 
1742: \bibitem[{{de Vaucouleurs} {et~al.}(1991){de Vaucouleurs}, {de Vaucouleurs},
1743:   {Corwin}, {Buta}, {Paturel}, \& {Fouque}}]{Vauc91}
1744: {de Vaucouleurs}, G., {de Vaucouleurs}, A., {Corwin}, H.~G., {Buta}, R.~J.,
1745:   {Paturel}, G., \& {Fouque}, P. 1991, {Third Reference Catalogue of Bright
1746:   Galaxies} (Volume 1-3, XII, 2069 pp.~7 figs..~ Springer-Verlag Berlin
1747:   Heidelberg New York)
1748: 
1749: \bibitem[{{de Wit} {et~al.}(2004){de Wit}, {Testi}, {Palla}, {Vanzi}, \&
1750:   {Zinnecker}}]{deWietal05}
1751: {de Wit}, W.~J., {Testi}, L., {Palla}, F., {Vanzi}, L., \& {Zinnecker}, H.
1752:   2004, \aap, 425, 937
1753: 
1754: \bibitem[{{Dohm-Palmer} \& {Skillman}(2002)}]{DohmSkil02}
1755: {Dohm-Palmer}, R.~C. \& {Skillman}, E.~D. 2002, \aj, 123, 1433
1756: 
1757: \bibitem[{{Drozdovsky} {et~al.}(2002){Drozdovsky}, {Schulte-Ladbeck}, {Hopp},
1758:   {Greggio}, \& {Crone}}]{Droz02}
1759: {Drozdovsky}, I.~O., {Schulte-Ladbeck}, R.~E., {Hopp}, U., {Greggio}, L., \&
1760:   {Crone}, M.~M. 2002, \aj, 124, 811
1761: 
1762: \bibitem[{{Eggenberger} {et~al.}(2002){Eggenberger}, {Meynet}, \&
1763:   {Maeder}}]{Eggeetal02}
1764: {Eggenberger}, P., {Meynet}, G., \& {Maeder}, A. 2002, \aap, 386, 576
1765: 
1766: \bibitem[{{Fall} {et~al.}(2005){Fall}, {Chandar}, \& {Whitmore}}]{Falletal05}
1767: {Fall}, S.~M., {Chandar}, R., \& {Whitmore}, B.~C. 2005, \apjl, 631, L133
1768: 
1769: \bibitem[{{Fanelli} {et~al.}(1988){Fanelli}, {O'Connell}, \&
1770:   {Thuan}}]{Faneetal88}
1771: {Fanelli}, M.~N., {O'Connell}, R.~W., \& {Thuan}, T.~X. 1988, \apj, 334, 665
1772: 
1773: \bibitem[{{Flower}(1996)}]{Flow96}
1774: {Flower}, P.~J. 1996, \apj, 469, 355
1775: 
1776: \bibitem[{{Garmany} {et~al.}(1982){Garmany}, {Conti}, \& {Chiosi}}]{Garmetal82}
1777: {Garmany}, C.~D., {Conti}, P.~S., \& {Chiosi}, C. 1982, \apj, 263, 777
1778: 
1779: \bibitem[{{Georgiev} {et~al.}(1999){Georgiev}, {Borissova}, {Rosado}, {Kurtev},
1780:   {Ivanov}, \& {Koenigsberger}}]{Georetal99}
1781: {Georgiev}, L., {Borissova}, J., {Rosado}, M., {Kurtev}, R., {Ivanov}, G., \&
1782:   {Koenigsberger}, G. 1999, \aaps, 134, 21
1783: 
1784: \bibitem[{Gonzaga(2003)}]{Gonzetal03}
1785: Gonzaga, S. 2003, WFPC2 Data Analysis: A Tutorial, version3.0 (STScI)
1786: 
1787: \bibitem[{{Gonz{\'a}lez Delgado} \& {P{\'e}rez}(2000)}]{GonDPere00}
1788: {Gonz{\'a}lez Delgado}, R.~M. \& {P{\'e}rez}, E. 2000, \mnras, 317, 64
1789: 
1790: \bibitem[{{Gordon} \& {Clayton}(1998)}]{GordClay98}
1791: {Gordon}, K.~D. \& {Clayton}, G.~C. 1998, \apj, 500, 816
1792: 
1793: \bibitem[{{Gustafsson} {et~al.}(1975){Gustafsson}, {Bell}, {Eriksson}, \&
1794:   {Nordlund}}]{Gustetal75}
1795: {Gustafsson}, B., {Bell}, R.~A., {Eriksson}, K., \& {Nordlund}, A. 1975, \aap,
1796:   42, 407
1797: 
1798: \bibitem[{{Humphreys}(1983)}]{Hump83}
1799: {Humphreys}, R.~M. 1983, \apj, 265, 176
1800: 
1801: \bibitem[{{Humphreys} \& {McElroy}(1984)}]{HumpMcel84}
1802: {Humphreys}, R.~M. \& {McElroy}, D.~B. 1984, \apj, 284, 565
1803: 
1804: \bibitem[{{Humphreys} \& {Sandage}(1980)}]{HumpSand80}
1805: {Humphreys}, R.~M. \& {Sandage}, A. 1980, \apjs, 44, 319
1806: 
1807: \bibitem[{{Hunter}(2001)}]{Hunt01}
1808: {Hunter}, D.~A. 2001, \apj, 559, 225
1809: 
1810: \bibitem[{{Jamet} {et~al.}(2004){Jamet}, {P{\' e}rez}, {Cervi{\~ n}o},
1811:   {Stasi{\' n}ska}, {Gonz{\' a}lez Delgado}, \& {V{\'{\i}}lchez}}]{Jameetal04}
1812: {Jamet}, L., {P{\' e}rez}, E., {Cervi{\~ n}o}, M., {Stasi{\' n}ska}, G.,
1813:   {Gonz{\' a}lez Delgado}, R.~M., \& {V{\'{\i}}lchez}, J.~M. 2004, \aap, 426,
1814:   399
1815: 
1816: \bibitem[{{Kroupa}(2002)}]{Krou02}
1817: {Kroupa}, P. 2002, Science, 295, 82
1818: 
1819: \bibitem[{{Kroupa} {et~al.}(1991){Kroupa}, {Gilmore}, \& {Tout}}]{Krou91}
1820: {Kroupa}, P., {Gilmore}, G., \& {Tout}, C.~A. 1991, \mnras, 251, 293
1821: 
1822: \bibitem[{Kurucz(2004)}]{Kuru04}
1823: Kurucz, R.~L. 2004, in http://kurucz.harvard.edu
1824: 
1825: \bibitem[{{Lada} \& {Lada}(2003)}]{LadaLada03}
1826: {Lada}, C.~J. \& {Lada}, E.~A. 2003, \araa, 41, 57
1827: 
1828: \bibitem[{{Lamers} {et~al.}(2005){Lamers}, {Gieles}, {Bastian}, {Baumgardt},
1829:   {Kharchenko}, \& {Portegies Zwart}}]{Lameetal05a}
1830: {Lamers}, H.~J.~G.~L.~M., {Gieles}, M., {Bastian}, N., {Baumgardt}, H.,
1831:   {Kharchenko}, N.~V., \& {Portegies Zwart}, S. 2005, \aap, 441, 117
1832: 
1833: \bibitem[{{Langer} \& {Maeder}(1995)}]{LaMa94}
1834: {Langer}, N. \& {Maeder}, A. 1995, \aap, 295, 685
1835: 
1836: \bibitem[{{Larsen} {et~al.}(2004){Larsen}, {Brodie}, \& {Hunter}}]{Larsetal04}
1837: {Larsen}, S.~S., {Brodie}, J.~P., \& {Hunter}, D.~A. 2004, \aj, 128, 2295
1838: 
1839: \bibitem[{{Leitherer} {et~al.}(1999){Leitherer}, {Schaerer}, {Goldader},
1840:   {Delgado}, {Robert}, {Kune}, {de Mello}, {Devost}, \& {Heckman}}]{Leitetal99}
1841: {Leitherer}, C., {Schaerer}, D., {Goldader}, J.~D., {Delgado}, R.~M.~G.,
1842:   {Robert}, C., {Kune}, D.~F., {de Mello}, D.~F., {Devost}, D., \& {Heckman},
1843:   T.~M. 1999, \apjs, 123, 3
1844: 
1845: \bibitem[{{Leitherer} {et~al.}(1996){Leitherer}, {Vacca}, {Conti},
1846:   {Filippenko}, {Robert}, \& {Sargent}}]{Leitetal96}
1847: {Leitherer}, C., {Vacca}, W.~D., {Conti}, P.~S., {Filippenko}, A.~V., {Robert},
1848:   C., \& {Sargent}, W.~L.~W. 1996, \apj, 465, 717
1849: 
1850: \bibitem[{{Lejeune} \& {Schaerer}(2001)}]{LeSc01}
1851: {Lejeune}, T. \& {Schaerer}, D. 2001, \aap, 366, 538
1852: 
1853: \bibitem[{{Lequeux}(1979)}]{Lequ79}
1854: {Lequeux}, J. 1979, \aap, 80, 35
1855: 
1856: \bibitem[{{Levesque} {et~al.}(2005){Levesque}, {Massey}, {Olsen}, {Plez},
1857:   {Josselin}, {Maeder}, {Meynet}, \& {White}}]{Leveetal05}
1858: {Levesque}, E.~M., {Massey}, P., {Olsen}, K.~A.~G., {Plez}, B., {Josselin}, E.,
1859:   {Maeder}, A., {Meynet}, G., \& {White}, N. 2005, ApJ, 628, 973
1860: 
1861: \bibitem[{{MacKenty} {et~al.}(2000){MacKenty}, {Ma{\'{\i}}z-Apell{\' a}niz},
1862:   {Pickens}, {Norman}, \& {Walborn}}]{MacKetal00}
1863: {MacKenty}, J.~W., {Ma{\'{\i}}z-Apell{\' a}niz}, J., {Pickens}, C.~E.,
1864:   {Norman}, C.~A., \& {Walborn}, N.~R. 2000, \aj, 120, 3007
1865: 
1866: \bibitem[{{Maeder} \& {Meynet}(2001)}]{MaedMeyn01}
1867: {Maeder}, A. \& {Meynet}, G. 2001, \aap, 373, 555
1868: 
1869: \bibitem[{{Ma{\'{\i}}z-Apell{\' a}niz}(2001)}]{Maiz01b}
1870: {Ma{\'{\i}}z-Apell{\' a}niz}, J. 2001, \apj, 563, 151
1871: 
1872: \bibitem[{{Ma{\'{\i}}z-Apell{\' a}niz}(2004)}]{Maiz04}
1873: ---. 2004, \pasp, 116, 859
1874: 
1875: \bibitem[{{Ma{\'{\i}}z-Apell{\' a}niz} {et~al.}(2002){Ma{\'{\i}}z-Apell{\'
1876:   a}niz}, {Cieza}, \& {MacKenty}}]{Maizetal02a}
1877: {Ma{\'{\i}}z-Apell{\' a}niz}, J., {Cieza}, L., \& {MacKenty}, J.~W. 2002, \aj,
1878:   123, 1307
1879: 
1880: \bibitem[{{Ma{\'{\i}}z-Apell{\' a}niz} {et~al.}(1999){Ma{\'{\i}}z-Apell{\'
1881:   a}niz}, {Mu{\~ n}oz-Tu{\~ n}{\' o}n}, {Tenorio-Tagle}, \&
1882:   {Mas-Hesse}}]{Maizetal99a}
1883: {Ma{\'{\i}}z-Apell{\' a}niz}, J., {Mu{\~ n}oz-Tu{\~ n}{\' o}n}, C.,
1884:   {Tenorio-Tagle}, G., \& {Mas-Hesse}, J.~M. 1999, \aap, 343, 64
1885: 
1886: \bibitem[{{Ma{\'{\i}}z-Apell{\' a}niz} {et~al.}(2005){Ma{\'{\i}}z-Apell{\'
1887:   a}niz}, {\'{U}beda}, Walborn, \& Nelan}]{Maizetal05}
1888: {Ma{\'{\i}}z-Apell{\' a}niz}, J., {\'{U}beda}, L., Walborn, N.~R., \& Nelan,
1889:   E.~P. 2005, in ASP Conf. Ser. Vol. TBA: Resolved Stellar Populations
1890: 
1891: \bibitem[{{Ma{\'{\i}}z-Apell{\'a}niz}(2000)}]{Maiz00}
1892: {Ma{\'{\i}}z-Apell{\'a}niz}, J. 2000, \pasp, 112, 1138
1893: 
1894: \bibitem[{{Ma\'{\i}z-Apell\'aniz} {et~al.}(1998){Ma\'{\i}z-Apell\'aniz},
1895:   {Mas-Hesse}, {Mu\~noz-Tu\~n\'on}, {V\'{\i}lchez}, \&
1896:   {Casta\~neda}}]{Maizetal98}
1897: {Ma\'{\i}z-Apell\'aniz}, J., {Mas-Hesse}, J.~M., {Mu\~noz-Tu\~n\'on}, C.,
1898:   {V\'{\i}lchez}, J.~M., \& {Casta\~neda}, H.~O. 1998, \aap, 329, 409
1899: 
1900: \bibitem[{{Ma{\'{\i}}z Apell{\'a}niz} \& {{\'U}beda}(2005)}]{MaizUbed05}
1901: {Ma{\'{\i}}z Apell{\'a}niz}, J. \& {{\'U}beda}, L. 2005, \apj, 629, 873
1902: 
1903: \bibitem[{{Mas-Hesse} \& {Kunth}(1991)}]{MasHKunt91a}
1904: {Mas-Hesse}, J.~M. \& {Kunth}, D. 1991, \aaps, 88, 399
1905: 
1906: \bibitem[{{Mas-Hesse} \& {Kunth}(1999)}]{MasHKunt99}
1907: ---. 1999, \aap, 349, 765
1908: 
1909: \bibitem[{{Massey}(1998)}]{Masse98}
1910: {Massey}, P. 1998, in ASP Conf. Ser. 142: The Stellar Initial Mass Function
1911:   (38th Herstmonceux Conference), 17
1912: 
1913: \bibitem[{{Massey}(2003)}]{Massey03}
1914: {Massey}, P. 2003, \araa, 41, 15
1915: 
1916: \bibitem[{{Massey} {et~al.}(1995{\natexlab{a}}){Massey}, {Johnson}, \&
1917:   {Degioia-Eastwood}}]{Massetal95}
1918: {Massey}, P., {Johnson}, K.~E., \& {Degioia-Eastwood}, K. 1995{\natexlab{a}},
1919:   \apj, 454, 151
1920: 
1921: \bibitem[{{Massey} {et~al.}(1995{\natexlab{b}}){Massey}, {Lang},
1922:   {Degioia-Eastwood}, \& {Garmany}}]{Massetal95b}
1923: {Massey}, P., {Lang}, C.~C., {Degioia-Eastwood}, K., \& {Garmany}, C.~D.
1924:   1995{\natexlab{b}}, \apj, 438, 188
1925: 
1926: \bibitem[{{Massey} \& {Olsen}(2003)}]{MassOlse03}
1927: {Massey}, P. \& {Olsen}, K.~A.~G. 2003, \aj, 126, 2867
1928: 
1929: \bibitem[{{Mathis}(1990)}]{Math90}
1930: {Mathis}, J.~S. 1990, \araa, 28, 37
1931: 
1932: \bibitem[{{Meynet} \& {Maeder}(2003)}]{MaedMeyn03}
1933: {Meynet}, G. \& {Maeder}, A. 2003, \aap, 404, 975
1934: 
1935: \bibitem[{{Misselt} {et~al.}(1999){Misselt}, {Clayton}, \&
1936:   {Gordon}}]{Missetal99}
1937: {Misselt}, K.~A., {Clayton}, G.~C., \& {Gordon}, K.~D. 1999, \apj, 515, 128
1938: 
1939: \bibitem[{{Plez} {et~al.}(1992){Plez}, {Brett}, \& {Nordlund}}]{Plezetal92}
1940: {Plez}, B., {Brett}, J.~M., \& {Nordlund}, A. 1992, \aap, 256, 551
1941: 
1942: \bibitem[{{Sagar} \& {Richtler}(1991)}]{SaRi91}
1943: {Sagar}, R. \& {Richtler}, T. 1991, \aap, 250, 324
1944: 
1945: \bibitem[{{Sargent} \& {Filippenko}(1991)}]{SargFili91}
1946: {Sargent}, W.~L.~W. \& {Filippenko}, A.~V. 1991, \aj, 102, 107
1947: 
1948: \bibitem[{{Scalo}(1986)}]{Scalo86}
1949: {Scalo}, J.~M. 1986, Fundamentals of Cosmic Physics, 11, 1
1950: 
1951: \bibitem[{{Schaerer}(2003)}]{Scha03}
1952: {Schaerer}, D. 2003, in IAU Symposium, 642
1953: 
1954: \bibitem[{{Schaerer} {et~al.}(1993){Schaerer}, {Meynet}, {Maeder}, \&
1955:   {Schaller}}]{Schaetal93}
1956: {Schaerer}, D., {Meynet}, G., {Maeder}, A., \& {Schaller}, G. 1993, \aaps, 98,
1957:   523
1958: 
1959: \bibitem[{{Schlegel} {et~al.}(1998){Schlegel}, {Finkbeiner}, \&
1960:   {Davis}}]{Schletal98}
1961: {Schlegel}, D.~J., {Finkbeiner}, D.~P., \& {Davis}, M. 1998, \apj, 500, 525
1962: 
1963: 
1964: \bibitem[{{\'{U}beda} {et~al.}(2007){\'{U}beda}, {Ma{\'{\i}}z-Apell{\' a}niz},
1965:   \& {MacKenty}}]{Ubedetala06}
1966: {\'{U}beda}, L., {Ma{\'{\i}}z-Apell{\' a}niz}, J., \& {MacKenty}, J.~W. 2007,
1967:  \aj, 133, 917 (Paper I)
1968:  
1969: 
1970: \bibitem[{{van den Bergh}(1968)}]{Bergh68}
1971: {van den Bergh}, S. 1968, \jrasc, 62, 219
1972: 
1973: \bibitem[{{van den Bergh}(2004)}]{Bergh04}
1974: ---. 2004, \aj, 128, 1880
1975: 
1976: \bibitem[{{Veltchev} {et~al.}(2004){Veltchev}, {Nedialkov}, \&
1977:   {Borisov}}]{Veltetal04}
1978: {Veltchev}, T., {Nedialkov}, P., \& {Borisov}, G. 2004, \aap, 426, 495
1979: 
1980: \bibitem[{{Walker}(1964)}]{Walk64}
1981: {Walker}, M.~F. 1964, \aj, 69, 744
1982: 
1983: \bibitem[{{Walter} {et~al.}(2001){Walter}, {Taylor}, {H{\" u}ttemeister},
1984:   {Scoville}, \& {McIntyre}}]{Waltetal01}
1985: {Walter}, F., {Taylor}, C.~L., {H{\" u}ttemeister}, S., {Scoville}, N., \&
1986:   {McIntyre}, V. 2001, \aj, 121, 727
1987: 
1988: \end{thebibliography}
1989: 
1990: 
1991: 
1992: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%  FIGURE01
1993: \clearpage
1994: \begin{figure}
1995: %\centerline{\includegraphics*[width=0.47\linewidth]{f1a.eps}
1996: %                  \includegraphics*[width=0.47\linewidth]{f1b.eps}}
1997: \includegraphics[width=0.9\linewidth]{f1.jpg}
1998: \caption{ 
1999:  [Left panel] F656N mosaic of NGC~4214 where we have over--plotted the selection of
2000:  stars that we used to make the extinction map. The color scale represents 
2001:  values of \ecc\
2002:  [Right panel] Extinction map of the central region of NGC~4214. 
2003:  The  orientation  is north pointing up and east pointing to the left. 
2004: The field dimensions are  875 pc $\times$ 972 pc or  $ 61 \farcs 4 \times 68 \farcs 3$.
2005:  The cluster apertures are also
2006:  drawn for reference. See the text for a description of
2007:  how  we made this map.
2008: }
2009:  \label{fig01}
2010: \end{figure}
2011: 
2012: 
2013: 
2014: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%  FIGURE02
2015: \clearpage
2016: \begin{figure}
2017: \includegraphics[width=0.9\textwidth]{f2.jpg}
2018: \caption{Distribution of blue and red
2019: supergiants on an F814W  mosaic of NGC~4214.
2020: The field shown is the same as in Figure~\ref{fig01}.
2021: The filled circles represent the confirmed supergiants, both blue and red that
2022: we used to calculate the B/R ratio.
2023: With open red circles we represent the group of stars  that follow 
2024: the criteria
2025: $\log(T_{\mathrm{eff}})  \leqslant 3.70 $ and  $M_\mathrm{bol} \leqslant  -6.0$.
2026: and which might  be RSGs,  AGBs/ bright RSGs.
2027: The cluster apertures are also drawn for reference.
2028: }
2029:  \label{fig02}
2030: \end{figure}
2031: 
2032: 
2033:  
2034: 
2035: 
2036: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE03
2037: \clearpage
2038: \begin{figure}
2039: %\centerline{\includegraphics*[bb = 51 47 535 565, width=0.47\linewidth]{f3a.ps}
2040:  %               \includegraphics*[bb = 60 47 545 565, width=0.47\linewidth]{f3b.ps}}
2041: %\centerline{\includegraphics*[bb = 40 25 521 505, width=0.47\linewidth]{f3c.ps}
2042: %                 \includegraphics*[bb = 75 25 556 505, width=0.47\linewidth]{f3d.ps}}
2043: \includegraphics[width=0.9\textwidth]{f3.jpg}
2044: 
2045: \caption{SEDs for  the best fits of clusters I--As and I--Es, based on \stb\ models
2046: of integrated stellar populations. The photometry is shown by the symbols
2047: with error bars (vertical ones for uncertainties and horizontal ones for 
2048: the approximate length coverage of each filter.) Star symbols indicate
2049: the calculated magnitude of the model SED for each filter.
2050: [Top left] Best fit SED for cluster I--As using an LMC2--like extinction law.
2051: [Top right] Best fit SED for cluster I--As using an SMC--like extinction law.
2052: [Bottom left] Best fit SED for cluster I--Es for the old (189 Myr) solution.
2053: [Bottom right] Best fit SED for cluster I--Es for the young (7 Myr) solution.  }
2054: \label{fig03}
2055: \end{figure}
2056:  
2057: %%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%% FIGURE04
2058: \clearpage
2059: \begin{figure}
2060: %\centerline{\includegraphics*[bb = 45 30 550 565, width=0.47\linewidth]{f4a.ps}
2061: %                  \includegraphics*[bb = 45 30 550 565, width=0.47\linewidth]{f4b.ps}}
2062: \includegraphics[width=0.9\textwidth]{f4.jpg}
2063: 
2064: \caption{  Reddening vs. age likelihood  contour plots for cluster I--As provided as output by \cho\ 
2065: and obtained using \stb\ models.
2066: [Left] Contour plot obtained using the LMC2 extinction law. 
2067: [Right] Contour plot obtained using the SMC extinction law. 
2068: Note that the age of this cluster is in the range 3.0--4.5 Myr, and that it 
2069: is characterized by a very small extinction. The white circle marks the most
2070: likely value (mode).
2071:   }
2072: \label{fig04}
2073: \end{figure}
2074: 
2075: 
2076: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%  FIGURE05
2077:     
2078:  \clearpage
2079: \begin{figure}              
2080: %\centerline{\includegraphics*[bb = 28 28 556 566, width=0.38\linewidth]{f5a.ps}\hspace{-0.8cm} 
2081: %                 \includegraphics*[bb = 28 28 556 566, width=0.38\linewidth]{f5b.ps}\hspace{-0.8cm}
2082: %                   \includegraphics*[ bb = 28 28 556 566, width=0.38\linewidth]{f5c.ps}}
2083: %\centerline{\includegraphics*[bb = 28 28 556 496,  width=0.38\linewidth]{f5d.ps}\hspace{-0.8cm}
2084: %                   \includegraphics*[ bb = 28 28 556 496, width=0.38\linewidth]{f5e.ps} \hspace{-0.8cm}
2085: %                 \includegraphics*[bb = 28 28 556 496, width=0.38\linewidth]{f5f.ps}}  
2086: \includegraphics[width=0.9\textwidth]{f5.jpg}
2087: 
2088: \caption{ Likelihood contour plots for cluster I--Es provided as output by \cho\ and obtained using \stb\ models.
2089: The upper row corresponds to the young solution and the lower row to the old one.
2090: [Left] Reddening vs. age plot. [Center] Extinction law vs. age. [Right] Extinction
2091: law vs. reddening. Age is expressed in years. The white circle marks the most
2092: likely value (mode). Note that the mode in 3D parameter space (shown here) 
2093: does not necessarily coincide with the mode
2094: in 2D parameter space. 
2095: }
2096: \label{fig05}
2097: \end{figure}
2098: 
2099: 
2100: 
2101: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%   FIGURE06
2102: \clearpage
2103: \begin{figure}
2104: \includegraphics[width=0.9\textwidth]{f6.jpg}
2105: \caption{Images of cluster I--Es obtained with HST/WFPC2 using filters 
2106: F336W, F555W, and F814W. The images have been 
2107: resampled in order to use the same linear scale in all cases, with the field
2108:  sizes being $15.9  \times 15.9$ pc$^2$.
2109: The orientation in each case is north pointing up and east pointing to the left.
2110: Special attention must be drawn to the two objects located 
2111: off center from  the cluster. See the text for a complete discussion. }
2112:  \label{fig06}
2113: \end{figure}
2114: 
2115: %%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%% FIGURE07
2116:      \clearpage
2117: \begin{figure}              
2118: %\centerline{\includegraphics*[bb = 28 28 556 566, width=0.38\linewidth]{f7a.ps}\hspace{-0.8cm} 
2119: %                 \includegraphics*[bb = 28 28 556 566, width=0.38\linewidth]{f7b.ps}\hspace{-0.8cm}
2120: %                   \includegraphics*[ bb = 28 28 556 566, width=0.38\linewidth]{f7c.ps}}%
2121: %\centerline{\includegraphics*[bb = 28 28 556 496,  width=0.38\linewidth]{f7d.ps}\hspace{-0.8cm}
2122: %                   \includegraphics*[ bb = 28 28 556 496, width=0.38\linewidth]{f7e.ps} \hspace{-0.8cm}
2123: %                 \includegraphics*[bb = 28 28 556 496, width=0.38\linewidth]{f7f.ps}}  
2124: \includegraphics[width=0.9\textwidth]{f7.jpg}
2125: 
2126: \caption{ Likelihood contour plots for cluster IIIs provided as output by \cho\ and obtained using \stb\ models.
2127: The upper row corresponds to the young solution and the  lower row to the old one.
2128: [left] Reddening vs. age plot. [Center] Extinction law vs. age. [Right] Extinction
2129: law vs. reddening. Age is expressed in years. The white circle marks the most
2130: likely value (mode). }
2131: \label{fig07}
2132: \end{figure}
2133: 
2134: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%% FIGURE08
2135: \clearpage
2136: \begin{figure}
2137: %\centerline{\includegraphics*[bb = 42 47 533 565, width=0.47\linewidth]{f8a.ps}
2138:  %              \includegraphics*[bb = 66 47 553 565, width=0.47\linewidth]{f8b.ps}}
2139:  %              \centerline{\includegraphics*[bb = 51 25 546 505, width=0.47\linewidth]{f8c.ps}
2140: %                 \includegraphics*[bb = 51 25 544 505, width=0.47\linewidth]{f8d.ps}}
2141: \includegraphics[width=0.9\textwidth]{f8.jpg}
2142: 
2143: \caption{ Same as Figure~\ref{fig03} for  clusters IIIs and IVs.
2144: [Top left] Best fit SED for cluster IIIs for the old (167 Myr) solution.
2145: [Top right] Best fit SED for cluster IIIs  for the young (16 Myr) solution.
2146: [Bottom left] Best fit SED for cluster IVs for the old (150 Myr) solution.
2147: [Bottom right]  Best fit SED for cluster IVs for the young (16 Myr) solution.
2148:  }
2149: \label{fig08}
2150: \end{figure}
2151:  
2152: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%% FIGURE09
2153: 
2154: \clearpage
2155: \begin{figure}
2156: %\centerline{\includegraphics*[bb = 28 28 556 566, width=0.38\linewidth]{f9a.ps}\hspace{-1cm} 
2157: %                 \includegraphics*[bb = 28 28 556 566, width=0.38\linewidth]{f9b.ps}\hspace{-1cm}
2158: %                   \includegraphics*[ bb = 28 28 556 566, width=0.38\linewidth]{f9c.ps}}
2159: %\centerline{\includegraphics*[bb = 28 28 556 496,  width=0.38\linewidth]{f9d.ps}\hspace{-1cm}
2160: %                   \includegraphics*[ bb = 28 28 556 496, width=0.38\linewidth]{f9e.ps} \hspace{-1cm}
2161: %                 \includegraphics*[bb = 28 28 556 496, width=0.38\linewidth]{f9f.ps}}  
2162: \includegraphics[width=0.9\textwidth]{f9.jpg}
2163: 
2164: \caption{ Likelihood contour plots for cluster IVs provided as output by \cho\ and obtained using \stb\ models.
2165: The upper row corresponds to the young solution and the lower row to the old one.
2166: [Left] Reddening vs. age plot. [Center] Extinction law vs. age. [Right] Extinction
2167: law vs. reddening. Age is expressed in years. The white circle marks the most
2168: likely value (mode).   }
2169: \label{fig09}
2170: \end{figure}
2171: 
2172: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % FIGURE10
2173: 
2174: \clearpage
2175: \begin{figure}
2176: %\centerline{\includegraphics*[bb = 28 28 556 566, width=0.38\linewidth]{f10a.ps}\hspace{-0.8cm} 
2177: %                 \includegraphics*[bb = 28 28 556 566, width=0.38\linewidth]{f10b.ps}\hspace{-0.8cm}
2178: %                   \includegraphics*[ bb = 28 28 556 566, width=0.38\linewidth]{f10c.ps}}
2179: %\centerline{\includegraphics*[bb = 28 28 556 477,  width=0.38\linewidth]{f10d.ps}\hspace{-0.8cm}
2180: %                   \includegraphics*[ bb = 28 28 556 477, width=0.38\linewidth]{f10e.ps} \hspace{-0.8cm}
2181: %                 \includegraphics*[bb = 28 28 556 477, width=0.38\linewidth]{f10f.ps}}  
2182: \includegraphics[width=0.9\textwidth]{f10.jpg}
2183: 
2184: \caption{ SEDs for  the best fits of clusters I--Ds [Top--left], II--A [Top--center],
2185: II--B [Top--right],  II--C [Bottom--left], II--D [Bottom--center], and II--E [Bottom--right],   
2186: based on \stb\ models
2187: of integrated stellar populations. The photometry is shown by the symbols
2188: with error bars  (vertical ones for uncertainties and horizontal ones for 
2189: the approximate length coverage of each filter.) Star symbols indicate
2190: the calculated magnitude of the model SED for each filter. 
2191: Magnitudes in 
2192: filters F555W and F702W are displayed in the spectrum of NGC~4214--IIABCDE, but
2193: they were not considered during the execution of \cho\ because of possible
2194: contamination.   }
2195: \label{fig10}
2196: \end{figure}
2197: 
2198: %%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%% FIGURE11
2199: 
2200: \clearpage
2201: \begin{figure}
2202: \includegraphics[width=0.9\textwidth]{f11.jpg}
2203: 
2204: \caption{ Likelihood contour plots for clusters I--Ds [Left], I--B [Middle], and II [Right]
2205:  provided as output by \cho\ and obtained using \stb\ models.
2206: [Top] Reddening vs. age plot. [Center] Extinction law vs. age. [Bottom] Extinction
2207: law vs. reddening. Age is expressed in years. The white circle marks the most
2208: likely value (mode). }
2209: \label{fig11}
2210: \end{figure}
2211: 
2212:  
2213: 
2214: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%    FIGURE12
2215: \clearpage
2216: \begin{figure}
2217: \includegraphics[width=0.9\textwidth]{f12.jpg}
2218: \caption{SEDs for  the best fits of clusters I--A [Top--left], I--B [Top--right], and II [Bottom], 
2219: based on \stb\ models
2220: of integrated stellar populations. The photometry is shown by the symbols
2221: with error bars  (vertical ones for uncertainties and horizontal ones for 
2222: the approximate length coverage of each filter.) Star symbols indicate
2223: the calculated magnitude of the model SED for each filter.  Magnitudes in 
2224: filters F555W and F702W are displayed in the spectrum of NGC~4214--II, but
2225: they were not considered during the execution of \cho\ because of possible
2226: contamination. See text for details.}
2227: \label{fig12}
2228: \end{figure}
2229: 
2230:  
2231: 
2232: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%    TABLE01
2233: \clearpage
2234: 
2235: \begin{table}[htbp]
2236: %\centering
2237: \caption{Comparison between theoretical  and observational values of 
2238: the ration B/R of the number of blue supergiants and red supergiants.
2239:  \label{tbl01}
2240: \vspace{5mm}}
2241: \begin{tabular}{@{} c|c|c|c|c|c|c @{}}   \tableline\tableline
2242:            &\multicolumn{2}{c|}{SMC  $v_{ini} \tablenotemark{a}=0$}      &\multicolumn{2}{c|}{SMC  $v_{ini} =300$}    
2243:            & \multicolumn{2}{c}{LMC  $v_{ini} =0$}    {\rule [0mm]{0mm}{5mm} }    \\   
2244:   \raisebox{1.8ex}[0pt]{mass range}         & theory  &   obs   &   theory  &   obs   & theory  &   obs  {\rule [-3mm]{0mm}{0mm} }       \\   \tableline
2245:    $   15-20 $   & 24 & $34 \pm 10$ & 0.4 & $29 \pm 10$& 1.0 & $13 \pm 4$  {\rule [0mm]{0mm}{5mm} }\\
2246:     $  20-25 $   & 47 & $46 \pm 23$ & 0.4 & $45 \pm 31$& 2.7 & $67  \pm 66${\rule [-3mm]{0mm}{0mm} }\\ \tableline
2247: \end{tabular}
2248: \tablenotetext{a}{  $v_{ini}$ is the initial rotational velocity of the theoretical model, in km s$^{-1}$.  }
2249: 
2250: \end{table}
2251: 
2252: 
2253:  
2254: 
2255: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%    TABLE02
2256: \clearpage
2257: 
2258: \begin{deluxetable}{lccccccccccc} 
2259: \tabletypesize{\scriptsize}
2260: \tablecaption{ IMF slope values  with their errors for the three lists calculated for several  values
2261: of the low--mass end  ($M_{low}$) and $M_{up} = 100 M_{\odot}$. These results would apply
2262: to a continuous star--formation region. \label{tbl02}}
2263: \tablewidth{0pt}
2264: \tablehead{  \colhead{} 
2265: & \multicolumn{3}{c}{ I--A +  I--B +  II}  &  \colhead{}
2266: & \multicolumn{3}{c}{ Other regions} &  \colhead{}
2267: & \multicolumn{3}{c}{NGC~4214}  \\
2268: \cline{2-4}\cline{6-8}\cline{10-12} \\    
2269: \colhead{ $M_{low} [M_{\odot}]$}    &  \colhead{$\gamma_{fit}$} &  \colhead{$\gamma_{real}$}   & \colhead{$N$}  &\colhead{}  &
2270: \colhead{$\gamma_{fit}$} &  \colhead{$\gamma_{real}$}  &  \colhead{$N$}  &\colhead{} &
2271:  \colhead{$\gamma_{fit}$} &  \colhead{$\gamma_{real}$}  &   \colhead{$N$}    }
2272: \startdata
2273:   15 &    -2.17 &   $ -2.67 \pm     0.08 $&  752 &  &  -2.15 &  $  -2.64 \pm     0.06 $& 1680 &  &  -2.15 & $   -2.64 \pm      0.05  $ & 2432\\
2274:    16 &    -2.20 &  $  -2.61 \pm     0.09 $&  655 &   & -2.28 &  $  -2.72 \pm     0.06 $& 1494 &  &  -2.26 & $   -2.70 \pm      0.05 $  & 2149\\
2275:    17 &    -2.24 &    $-2.60 \pm     0.10 $&  573 &   & -2.43 &  $  -2.83 \pm    0.07 $& 1357 &  &  -2.39 & $   -2.78 \pm         0.06 $  & 1930\\
2276:    18 &    -2.21 &   $ -2.51 \pm     0.10 $&  501 &   & -2.59 &$    -2.95 \pm     0.08 $& 1216 &  &  -2.48 & $   -2.82  \pm        0.06  $ & 1717\\
2277:    19 &    -2.30 &   $ -2.57 \pm     0.11 $&  449 &  &  -2.72 & $   -3.03 \pm     0.08$ & 1098 &  &  -2.56 &$    -2.86 \pm     0.06  $ & 1547 \\
2278:    20 &    -2.36 &   $ -2.60 \pm     0.12$ &  397 &  &  -2.71 &$    -2.98 \pm     0.09 $&  961 &  &  -2.58 & $   -2.83 \pm        0.07$   & 1358   \\
2279: \enddata
2280: \end{deluxetable}
2281: 
2282: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%    TABLE03
2283:  
2284: 
2285: \begin{deluxetable}{llccc} 
2286: \tabletypesize{\scriptsize}
2287: \tablecaption{IMF slope values  with their errors for  regions   I--A,  I--B, and  II. 
2288: These results would apply
2289: to a burst star--formation region.   \label{tbl03}}
2290: \tablewidth{0pt}
2291: \tablehead{  \colhead{} &\colhead{} 
2292: & \multicolumn{3}{c}{ I--A +  I--B +  II}    \\
2293: \cline{3-5}\\  
2294: \colhead{ $M_{low} [M_{\odot}]$}    & \colhead{ $M_{up} [M_{\odot}]$}    & \colhead{$\gamma_{fit}$} &  \colhead{$\gamma_{real}$}    & \colhead{$N$}  }
2295: \startdata
2296:    15 & 100 &   -3.30 &$    -4.41 \pm     0.09 $&  752   \\
2297:    16 &  100 &  -3.25 &  $  -4.06 \pm     0.10 $&  655  \\
2298:    17 & 100 &   -3.24 & $   -3.87 \pm    0.10 $&  573  \\
2299:    18 &  100 &  -3.14 & $   -3.61 \pm    0.11 $&  501 \\
2300:    19 & 100 &   -3.20 & $   -3.59 \pm     0.12 $&  449  \\
2301:    20 & 100 &   -3.23 & $   -3.56 \pm     0.13 $&  397   \\
2302:       40 & 100 &   -2.89 & $   -2.97 \pm     0.37 $&  85   \\
2303:    20 & 40 &   -3.73 & $   -4.41 \pm     0.28 $&  312   \\
2304: 
2305: \enddata
2306: \end{deluxetable}
2307: 
2308: 
2309: 
2310: 
2311: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%    TABLE04
2312: \clearpage
2313: \begin{deluxetable}{llllllllllll}
2314: \tabletypesize{\scriptsize}
2315: \rotate
2316: \tablecolumns{12}
2317: \tablewidth{0pt}
2318: \tablecaption{Results from  \cho ~fitting for clusters I--As, I--Es, IIIs, and IVs. The mass was estimated using a  Kroupa IMF \citep{Krou02}.
2319:  The number of stars is obtained from \stb\ models.
2320: \label{tbl04} }
2321: \tablehead{
2322:  \multicolumn{1}{l}{Parameter}     
2323: &  \multicolumn{2}{c}{Cluster I--As} &  \colhead{}  
2324: &  \multicolumn{2}{c}{Cluster I--Es} &  \colhead{}
2325: &  \multicolumn{2}{c}{Cluster IIIs}  &  \colhead{}   
2326: &   \multicolumn{2}{c}{ Cluster IVs}     \\[-00mm]
2327: \cline{2-3}\cline{5-6}\cline{8-9}\cline{11-12} \\[-2mm]
2328:  \colhead{}
2329: & \multicolumn{1}{c}{LMC2}    
2330: &  \multicolumn{1}{c}{SMC}    
2331: & \colhead{} 
2332: &  \multicolumn{1}{c}{Young}    
2333: & \multicolumn{1}{c}{Old}    
2334: & \colhead{}   
2335: & \multicolumn{1}{c}{Young}    
2336: &  \multicolumn{1}{c}{Old}    
2337: & \colhead{} 
2338: &  \multicolumn{1}{c}{Young}       
2339: & \multicolumn{1}{c}{Old}     % \\[-14mm]
2340: }
2341: \startdata
2342: Age (Myr)  &      $4.0 \pm 0.6$ &  $4.2 \pm 0.6$ &  &   $ 7.1\pm0.1$    & $189\pm53$   
2343:            &        &      $15.8\pm 0.2$ &  $168 \pm 61$ & &   $15.8 \pm 0.0$    & $150 \pm 34$    \\ %[-2mm]
2344: \ecc ~(mag)  &   $0.05 \pm 0.01$    &    $0.04 \pm 0.01$  &   &  $0.46\pm0.01$   & $0.32\pm0.03$   
2345:             &          &   $0.44 \pm 0.01$    &    $0.26 \pm 0.03$  &   &  $0.22\pm0.01$   & $0.06\pm0.03$   \\ %[-2mm]
2346: Extinction law  / \rv  &   LMC2    &  SMC    & &     $4.21\pm0.25$  & $5.72\pm0.26$   
2347:              &            &   $3.95\pm0.20$     &  $4.98\pm0.60$   &   &     $3.39\pm0.52$  & $4.35\pm1.12$   \\ %[-2mm]
2348: Mass ($10^3 M_{\odot} $)   & $ 27\pm   4  $ & $ 27 \pm  4  $ & &    $6\pm1$     &  $129\pm26$   
2349:                                 &             & $ 134\pm   18  $ & $ 626 \pm   175  $ & &    $27\pm4$     &  $114\pm20 $   \\ %[-2mm]
2350: $\chi^2  $ per degree of freedom   &    1.33   &   0.91  &  &  2.58 &  2.19 & &    0.39   &   0.66   & &  1.77  &  2.18   \\ %[-2mm]
2351: O+B stars, types I and II  &     11  &  11 &   &2 &  0  &   &     31  &  0 &   &7 &  0  \\ %[-2mm]
2352: K+M stars, types I and II  &   0    & 0  &  &  1 & 9   &   &   6   &  45  & &  1 & 26 \\ %[-2mm]
2353: WR stars  & 1     & 1  & &   0 &  0   &   & 0     & 0&  &  0 &  0  \\[0mm]  
2354: \enddata
2355: \end{deluxetable}
2356:  
2357:   
2358: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%  TABLE05
2359: \clearpage
2360:  
2361:  
2362: \begin{deluxetable}{lllllll}
2363: \tabletypesize{\scriptsize}
2364: %\rotate
2365: \tablecolumns{7}
2366: \tablewidth{0pt}
2367: %\tablewidth{15cm} 
2368: \tablecaption{Results from  \cho ~fitting for resolved clusters.The mass was estimated using a  Kroupa IMF \citep{Krou02}.
2369: \label{tbl05} }
2370: \tablehead{
2371: \multicolumn{1}{l}{Parameter}   & \multicolumn{1}{c}{Cluster I--Ds}    
2372: &  \multicolumn{1}{c}{Cluster II--A}    
2373: &  \multicolumn{1}{c}{Cluster II--B}    
2374: &  \multicolumn{1}{c}{Cluster II--C}    
2375: & \multicolumn{1}{c}{Cluster II--D}    
2376: & \multicolumn{1}{c}{Cluster II--E}    
2377: }
2378: \startdata
2379: Age (Myr)  &      $2.6 \pm 1.5$ &  $3.1 \pm 1.4$   &   $ 2.0\pm0.8$    & $1.7\pm0.6$   &      $4.0\pm 4.0$ &  $3.1 \pm 1.4$   \\ 
2380: \ecc ~(mag)  &   $0.28 \pm 0.05$    &    $0.35 \pm 0.08$     &  $0.35\pm0.01$   & $0.28\pm0.01$ & $0.22 \pm 0.16$    &    $0.20 \pm 0.05$     \\  
2381: Extinction law / \rv   &       $4.05\pm0.19$  & $3.79\pm0.33$ &   $4.11\pm0.11$     &  $4.82\pm0.13$      &     $4.05\pm0.91$  & $2.92\pm0.43$   \\ 
2382: Mass ($10^3 M_{\odot} $)   & $ 20 \pm   9  $ & $ 34 \pm   23  $  &    $43\pm10$     &  $63\pm14$ & $ 7\pm   5  $ & $ 7 \pm 3  $     \\ %[-2mm]
2383: $\chi^2  $ per degree of freedom   &    1.31   &   0.41  &     2.02 &  2.53 &    0.15   &   0.22   \\ %[-2mm]
2384: \enddata
2385: \end{deluxetable}
2386: 
2387: 
2388: 
2389: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%  TABLE06
2390: 
2391: \begin{deluxetable}{llll}
2392: \tabletypesize{\scriptsize}
2393: %\rotate
2394: \tablecolumns{4}
2395: \tablewidth{0pt}
2396: %\tablewidth{15cm} 
2397: \tablecaption{Results from  \cho ~fitting for large complexes.The mass was estimated using a  Kroupa IMF \citep{Krou02}.
2398: \label{tbl06} }
2399: \tablehead{
2400: \multicolumn{1}{l}{Parameter}   &
2401: \multicolumn{1}{c}{Cluster I--A}    
2402: &  \multicolumn{1}{c}{Cluster I--B}    
2403: &  \multicolumn{1}{c}{Cluster II}    
2404: }
2405: \startdata
2406: Age (Myr)  &      $5.0 \pm 0.0$ &  $5.0 \pm 0.0$   &   $ 1.9\pm0.9$       \\ 
2407: \ecc ~(mag)  &   $0.07 \pm 0.01$    &    $0.07 \pm 0.01$     &  $0.43\pm0.03$  \\
2408: Extinction law / \rv   &       $5.88\pm0.16$  & $5.81\pm0.19$ &   $4.99\pm0.11$    \\  
2409: Mass ($10^3 M_{\odot} $)   & $ 156 \pm   19  $ & $ 34 \pm   4  $  &    $923\pm331$    \\
2410: $\chi^2  $ per degree of freedom   &    0.95   &   4.25  &   0.21     \\ %[-2mm]
2411: \enddata
2412: \end{deluxetable}
2413: 
2414: 
2415: \end{document}
2416: 
2417: