astro-ph0005545/wr.tex
1: \documentclass{aastex}
2: \usepackage{emulateapj5}
3: 
4: \begin{document}
5: 
6: \title{UV excess galaxies: Wolf-Rayet galaxies}
7: \author{Warren R.~Brown, Scott J.~Kenyon, Margaret J.~Geller, and Daniel
8: G.~Fabricant}
9: \affil{Harvard-Smithsonian Center for Astrophysics, 60 Garden St.,
10: Cambridge, MA 02139}
11: 
12: \email{wbrown@cfa.harvard.edu,
13: skenyon@cfa.harvard.edu,
14: mgeller@cfa.harvard.edu, 
15: and dfabricant@cfa.harvard.edu
16: }
17: 
18: \shorttitle{Wolf-Rayet Galaxies}
19: \shortauthors{Brown et al.}
20: 
21: \begin{abstract}
22: 
23: 	We discuss $V$ and $R$ band photometry for 67\% of the
24: \citet{S2000} SA57 ultraviolet-selected galaxy sample.  In a sample
25: of 176 UV-selected galaxies, \citet{S2000} find that 24\% have (UV-$B$)
26: colors too blue for consistency with starburst spectral synthesis models.
27: We propose that these extreme blue, UV excess galaxies are Wolf-Rayet (WR)
28: galaxies, starburst galaxies with strong UV emission from WR stars. We
29: measure a median $(\vr)=0.38\pm0.06$ for the UV-selected sample,
30: bluer than a sample optically selected at $R$ but consistent with
31: starburst and WR galaxy colors.  We demonstrate that redshifted WR
32: emission lines can double or triple the flux through the UV bandpass at
33: high redshifts.  Thus the (UV$-B)$ color of a WR galaxy can be up to 1.3
34: mag bluer at high redshift, and the expected selection function is skewed
35: to larger redshifts.  The redshift distribution of the extreme blue, UV
36: excess galaxies matches the selection function we predict from the
37: properties of WR galaxies.
38: 
39: \end{abstract}
40: 
41: 
42: \keywords{galaxies: starburst, galaxies: photometry, stars: Wolf-Rayet,
43: ultraviolet: galaxies}
44: 
45: \section{INTRODUCTION}
46: 
47: 	\citet[hereafter S2000]{S2000} published an ultraviolet (UV)
48: flux-limited catalog of galaxies, imaged with the FOCA balloon-borne UV
49: camera of \citet{mil92}.  The catalog is an extension of work by
50: \citet{tre98}, and contains 273 galaxies with redshifts $z\lesssim0.4$.
51: Vacuum UV imaging enables the direct detection of radiation from hot,
52: massive stars and, by association, actively star forming galaxies. The
53: luminosity function (LF) of the UV-selected galaxies is somewhat
54: surprising.  S2000 find a faint end slope in the \citet{sch76}
55: parametrization of the LF, $\alpha_{uv}=-1.55\pm0.11$, steeper than
56: typical optical LF faint end slopes \citep[etc.]{fol99, mar98, gel97}.  
57: As a result, S2000 find no strong decrease in UV luminosity density
58: between $z\simeq0.4$ and $z=0$ and thus little decline in the density of
59: star formation over this range of cosmic epochs.
60: 
61: 	In contrast, \citet{mad96} conclude that there is a factor of
62: $\simeq10$ decline in the comoving density of star formation between
63: redshifts $z=1$ and $z=0$. This conclusion is based on a combination of
64: rest frame near-UV luminosities from the $I$-selected Canada France
65: Redshift Survey \citep{lil95} and local H$\alpha$ measures from Gallego et
66: al.'s (1995) objective prism survey.  
67: 
68: 	Curiously, a substantial fraction of the S2000 UV-selected
69: galaxies have (UV$-B)$ colors bluer than the bluest color that can be
70: produced by starburst spectral synthesis models \citep{fio97,lei99}. S2000
71: could therefore not explain the color of these galaxies.
72: 
73: 	Here we propose that the extremely blue galaxies are Wolf-Rayet
74: (WR) galaxies.  WR galaxies show direct signatures of WR stars in their
75: spectra, usually a broad He {\small II} $\lambda4686$ feature originating
76: in the stellar winds of WR stars.  This subtle feature is not seen in the
77: fiber-spectra of S2000 (M. Sullivan 2000, private communication), but
78: requires good S/N and is spatially dependent on the inclusion of starburst
79: regions \citep{sch99}.  The starburst regions dominate the broadband UV
80: flux and, though short-lived ($<10^{7}$ yrs), WR stars can dominate the UV
81: light of young starbursts.
82: 
83: 	In this paper we describe optical photometry for 67\% of the S2000
84: Selected Area 57 (SA57) UV-selected galaxy sample.  Our goal is to compare
85: the optical and UV colors, and to suggest that WR emission lines are the
86: source of the extremely blue colors found by S2000.  In section 2 we
87: discuss our data, in section 3 we discuss the (\vr) colors, and in section
88: 4 we discuss the impact of Wolf-Rayet UV emission lines on the
89: interpretation of the UV sample.  We conclude in section 5.
90: 
91: 
92: \section{DATA}
93: 
94: 
95: 	We obtained broadband $V$ and $R$ images in 1998 December and 1999
96: February with the MOSAIC camera on the Kitt Peak National Observatory 0.9m
97: telescope.  MOSAIC is an 8 CCD array with pixel scale 0.424 arcsec
98: pixel$^{-1}$ and field of view $59^\prime $ x $59^\prime $ on the 0.9m. We
99: obtained three offset 200 sec $R$ and 333 sec $V$ images at each position
100: along a strip in RA.  Our data serendipitously overlap with 148 of the 222
101: S2000 UV-selected galaxies in the SA57 region.
102: 
103: 	We processed the images in the standard way with the IRAF MSCRED
104: package.  Total magnitudes are obtained using SExtractor \citep{ber96}.
105: Seven percent of the magnitudes are obtained with a modified version of
106: IRAF-based GALPHOT \citep{fre93}.  GALPHOT allows photometry of galaxies
107: poorly fit by SExtractor, including low surface brightness galaxies or
108: galaxies with close companions.  A comparison between the two software
109: packages yields a 1-$\sigma$ scatter in the total galaxy magnitudes of
110: $\pm0.05$ mag.  We use Landolt fields \citep{lan92} and the M67 cluster
111: \citep{mon93} for flux calibration.  Depending on the nightly seeing
112: conditions, the error in the standard star calibration is $\pm0.02-0.03$
113: mag.  A direct measure of our intrinsic error is the RMS scatter of
114: measurements of identical galaxies taken on different nights: $\pm0.04$
115: mag.
116: 
117: 	The photometry includes galactic extinction corrections from
118: \citet{sch98}, who combine $COBE$/DIRBE and $IRAS$/ISSA data to map dust
119: emission in the Milky Way.  Typical extinction values are $A_R=0.025$ and
120: $A_V=0.03$ mag.  The photometry also includes ``k corrections'' for the
121: changing rest frame bandpass \citep{fre94}.  Because we know the (\vr)
122: color and the redshift of each galaxy, we can determine accurate k
123: corrections. Typical values for the $V$ and $R$ bandpasses are $-0.1 < k_R
124: < 0.1$ and $-0.15 < k_V < 0.15$ mag.
125: 
126: \subsection{Existing UV and $B$ measurements}
127: 
128: 	The S2000 UV measurements were made from a balloon-borne, 40 cm
129: Cassegrain telescope stabilized to 2 arcsec RMS.  The filter response
130: approximates a Gaussian centered at 2015 \AA, with a FWHM of 188 \AA.  
131: The limiting magnitude in the UV catalog is m$_{UV}=18.5$.  S2000
132: state that the UV zero-point is accurate to $\le 0.2$ mag; the uncertainty
133: in the relative photometry may reach $\simeq 0.5$ mag at the limiting
134: magnitude.
135: 
136: 	With the UV images in hand, S2000 searched for optical
137: counterparts in digitized POSS plates.  S2000 obtained accurate positions
138: and $B$ photometry, with uncertainty $\pm0.20$ mag, from the POSS plates.
139: We follow S2000 and only use galaxies with a single optical counter-part
140: when discussing UV colors.  S2000 used the resulting (UV$-B)$ colors to
141: calculate k corrections.  Finally, S2000 apply a reddening correction
142: based on H$\alpha$ and H$\beta$ line strengths and the empirical
143: \citet{cal97} extinction law for starburst galaxies.  They applied a mean
144: reddening correction $A_V=0.97$ to galaxies lacking H$\alpha$ and/or
145: H$\beta$ emission.
146: 
147: 	The (UV$-B)_{\rm o}$ colors we quote from S2000 include k
148: corrections and reddening corrections.  Our (\vr) colors include k
149: corrections but no reddening corrections, because S2000 do not provide
150: enough information to reproduce their reddening corrections.  Even if this
151: reddening data were provided, 72\% of the galaxies with (\vr) colors lack
152: both H$\alpha$ and H$\beta$ emission.  For these galaxies, the mean
153: reddening correction would add a uniform $\simeq-0.2$ mag offset to the
154: (\vr) colors.
155: 
156: 
157: \section{OPTICAL COLORS}
158: 
159: 
160: 	Figure 1 shows that the (\vr) colors of the UV-selected galaxies
161: are significantly bluer than those from an $R$-selected galaxy sample, the
162: Century Survey \citep{gel97}.  The Century Survey is a complete redshift
163: survey of 102 square degrees to a limiting magnitude $R=16.13$, containing
164: 1762 galaxies.  The Century Survey galaxies have median
165: $(\vr)_{CS}=0.58\pm0.04$; the UV-selected galaxies have median
166: $(\vr)=0.38\pm0.06$.
167: 
168: 	The distribution in Figure 1 appears double-peaked, but
169: application of the non-parametric estimator of \citet{pis93} shows the
170: second peak is not significant.  The \citet{pis93} kernel probability
171: estimator allows estimation of the probability density distribution
172: underlying a data sample.  In applying this we assume an average
173: 1-$\sigma$ (\vr) error $\pm0.066$ mag.
174: 
175: 	The S2000 (UV$-B)_{\rm o}$ colors provide a measure of the star
176: formation activity, but there is no correlation between the (\vr) and
177: (UV$-B)_{\rm o}$ colors.
178: 
179: 	Unlike the extreme blue (UV$-B)_{\rm o}$ colors found by S2000,
180: however, the optical (\vr) colors are consistent with starburst galaxy
181: colors.  For example, the template starburst-galaxy spectral energy
182: distributions SB1 and
183: 
184: 
185:         \includegraphics*[scale=0.44]{fig1b.ps}
186:  
187: \figcaption[fig1b.ps]{Histogram of (\vr) colors for UV-selected galaxies  
188: (solid line) and $R$-selected Century Survey (dashed line).}
189: 
190:  
191: 	~
192: 
193: \noindent SB2 of \citet{kin96} have synthetic $(V-R)=0.26$.  All but
194: two of the S2000 UV-selected galaxies are redder than $(V-R)=0.23$, and we
195: note that these blue outliers appear atypically faint and point-like in
196: our images.  Thus $\sim95\%$ of the UV galaxy (\vr) colors are consistent
197: with starburst galaxy colors. If the extreme blue galaxies were AGN, their
198: average color should be $(\vr)=0.60$ \citep{kot93}, which is clearly not
199: the case.
200: 
201: 
202: \section{DISCUSSION}
203: 
204: 
205: 	An unresolved issue from the S2000 paper is the existence of
206: galaxies with (UV$-B)_{\rm o}$ colors $\lesssim-2$, colors too blue to be
207: explained by standard spectral synthesis models.  In our subsample of the
208: S2000 UV-selected galaxies, 48 of 148 galaxies have (UV$-B)_{\rm o}<-2$,
209: yet the optical (\vr) colors show no significant departure from the rest
210: of the sample.  A KS test gives an 80\% probability of the extreme
211: (UV$-B)_{\rm o}<-2$ and the other UV-selected galaxies being drawn from
212: the same distribution of (\vr) colors.
213: 
214: 	Next we examine the redshift distribution of the extremely blue
215: galaxies.  The lower panel of Figure 2 shows the redshift distribution of
216: the extreme blue (UV$-B)_{\rm o}<-2$ galaxies, and the upper panel the
217: remaining (UV$-B)_{\rm o}>-2$ galaxies.  Figure 2 shows that the extreme
218: blue galaxies are preferentially found at high redshift. A KS-test yields
219: a $10^{-11}$ probability of the two populations being drawn from the same
220: distribution.
221: 
222: 	The peaks in the redshift histograms (Figure 2) are well matched
223: to the $z\sim0.3$ limit of the Medium Deep Survey \citep{wil96}, which
224: includes the SA57 region.  The Medium-Deep Survey is $\sim70\%$ complete
225: to $B_J\sim20.5$ mag, and gives a sense of the existing large scale
226: structure in the direction of UV survey.  Thus the extremely blue galaxies
227: appear to trace the same large-scale structure as ``normal'' galaxies.
228: 
229: 
230: \subsection{Wolf-Rayet galaxies.}
231: 
232: 	The very blue (UV$-B)_{\rm o}<-2$ galaxies are an extreme
233: population of galaxies found mostly at larger redshift.  An explanation
234: probably requires emission lines redshifting into the UV bandpass.  
235: Because S2000 and our $(\vr)$ color distribution rule out AGN, it seems
236: plausible the emission lines come from a stellar emission line source,
237: i.e.~Wolf-Rayet (WR) stars, planetary nebulae, or symbiotic stars.  
238: Because there are known galaxies with WR features, we suggest that the
239: extreme blue (UV$-B)_{\rm o}<-2$ 
240: 
241: 
242:         \includegraphics[scale=0.43]{fig2b.ps}
243:  
244: \figcaption[fig2b.ps]{Redshift histograms for the (UV$-B)_{\rm o}>-2$  
245: galaxies (upper panel) and the extreme blue (UV$-B)_{\rm o}<-2$ galaxies
246: (lower panel). Note the excess of extreme blue galaxies at large
247: redshift.  The selection function curves are calculated with the S2000 LF
248: parameters (dashed-line) and the SSRS2 Sm-Irr LF parameters
249: (dash-dot, solid lines) as described in section 4.1.}
250: 
251: 
252: 	~
253: 
254: \noindent galaxies are WR galaxies.  The strongest
255: UV features come from ionized C, O, He, Si, and Fe in the stellar winds of
256: WR stars (e.g.~Hillier \& Miller 1999).
257: 
258: 	Our optical spectra of 10 extreme blue galaxies all show narrow
259: emission lines indicative of active star formation.  WR galaxies are
260: characterized by the direct signatures of WR stars in their spectra, but
261: our S/N is too poor to detect a possible $\lambda4686$ WR feature
262: convincingly.
263: 
264: 	To calculate a typical (\vr) color for WR galaxies, we searched
265: the NASA/IPAC Extragalactic Database for photometery of known WR galaxies.  
266: The average optical color of 13 WR galaxies is
267: $\overline{(\vr)}=0.27\pm0.34$ mag.  This average color is in agreement
268: with the blue color we find for the (UV$-B)_{\rm o}<-2$ sample,
269: $\overline{(\vr)}=0.35\pm0.11$ mag.
270: 
271: 	Figure 3 illustrates how WR emission lines are redshifted into the
272: S2000 UV bandpass.  We plot a sample WR ultraviolet spectrum, WC star HD
273: 165763 \citep{hil99}, at redshift $z=0$ and at redshift $z=0.285$ with the
274: S2000 UV and $B$ bandpasses overlaid. Integration of the UV bandpass over
275: the redshifted spectrum shows that the WR emission lines can boost the UV
276: flux by a factor of $\lesssim3$ compared to the flux at $z=0$.  We also
277: integrate the $B$ bandpass over the \citet{hil99} WC stellar spectrum, and
278: find the (UV$-B)$ color is shifted up to 1.3
279: mag bluewards over the redshift range z=0 to z=0.5 (see 
280: 
281: 
282: 	\includegraphics*[scale=0.43]{fig3b.ps}
283:  
284: \figcaption[fig3b.ps]{WR spectrum of star HD 165763 at $z=0$ and at
285: $z=0.285$, with rest-frame UV and $B$ bandpasses.}
286:  
287: 
288: \noindent Figure 4), blue enough to explain $2/3$ of the extreme blue
289: (UV$-B)_{\rm o}$ galaxies in the SA57 UV-selected sample.
290: 
291: 	The WC stellar spectrum represents a maximal effect; known WR
292: galaxies have typical WC/WN ratios of 0.2 - 0.4 \citep{sch99}.  We perform
293: the (UV$-B)$ calculation on the WN stellar spectrum, WR152 \citep{ham98},
294: and find the (UV$-B)$ color shifts up to 1.1 mag bluewards over the same
295: redshift range (see Figure 4).
296: 
297: 	We acknowledge that strong UV emission lines are not common
298: features of known low redshift WR galaxies (Leitherer 2000, private
299: communication).  We calculate the shifts in (UV$-B)$ for a dozen of the
300: \citet{mcq95} ``UV to optical spectral distributions of northern
301: star-forming galaxies,'' and find an average $\Delta({\rm UV}-B)\sim-0.3$
302: mag (mostly due to the continua, not the emission lines). Yet good
303: examples of the very blue objects may not be found locally.  The UV
304: selection should pick out extreme objects; known WR galaxies are optically
305: selected. We note the very blue UV galaxies have median z=0.2. By
306: comparison, the \citet{sch99b} catalog of known WR galaxies drops off
307: sharply at z=0.02.
308: 
309: 	We can also take advantage of $HST$ FOS observations of the Small
310:  Magellanic Cloud H {\small II} region N88A \citep{kur99} as a UV
311:  spectral templates for WR galaxies.  When the S2000 bandpass is centered
312:  on the strong C{\small III} and Si{\small III} lines in this object, the
313:  UV flux doubles compared to the surrounding continuum level.
314: 
315: 	S2000 compute their luminosity function (LF) with the full UV
316: sample; here we construct a model LF for the WR galaxies alone.  The
317: impact of the emission lines on the observed UV flux changes with
318: redshift.  We explore the selection function for finding WR galaxies in a
319: magnitude limited redshift survey by computing the change in absolute
320: magnitude resulting from the UV emission lines.
321: 
322: 	WR galaxies are blue and show active star formation.  We thus use
323: the Second Southern Sky Redshift Survey \citep[SSRS2]{mar98} LF for the
324: bluest galaxies, the Magellanic spiral and irregulars, as the basis of our
325: model. The SSRS2 is a $B$-selected survey complete to $m_b\le15.5$
326: that contains 5404 galaxies, all with morphological classifications.  The
327: SSRS2 Sm-Irr LF has a steep faint-end slope, $\alpha=-1.81$, similar to
328: the S2000 LF, $\alpha=-1.55$.
329: 
330: 	We compute the selection function for the WR galaxies according to
331: equation (1),
332:  \begin{equation}
333: N(<M_{lim}) \propto \int_z D_C(z)^2 dD_C(z)
334: \int_{M=-\infty}^{M_{lim}(z)} \phi(M,z) dM
335:  \end{equation}
336: 	where $D_C(z)$ is the co-moving distance.  We assume a flat
337: cosmology and $H_o$=100 km s$^{-1}$ Mpc$^{-1}$, $q_o$=0.5.  $\phi(M,z)$ is
338: the luminosity function in the \citet{sch76}
339: 
340: 
341: 	\includegraphics*[scale=0.43]{fig4b.ps}
342:  
343: \figcaption[fig4b.ps]{Change in (UV$-B)$ color vs.~redshift, normalized to
344: $z=0$.  The calculations are based on the HD 165763 WC and the WR152 WN
345: stellar spectra.}
346: 
347: 
348: \noindent
349: parameterization, $\phi(M) = 0.4 ln10 \phi^*$
350: $[10^{0.4(M^*-M)}]^{(1+\alpha)}$ $exp(-10^{0.4(M^*-M)})$, where $\phi^*$
351: is the normalization (galaxies per unit volume), $\alpha$ is the faint end
352: slope, and $M^*$ is the characteristic absolute magnitude.  To convert the
353: SSRS2 Sm-Irr $M_B^*=-19.78$ to the UV bandpass, we add the median
354: (UV$-B)_{\rm o}=-0.91$ color for the UV galaxies.  To simulate WR
355: galaxies, we add the relative change in (UV$-B)$ over redshift (Figure 4):
356:  \begin{equation}
357: M_{\rm UV}^* = M_B^* + ({\rm UV}-B)_{\rm o} + \Delta({\rm UV}-B)[z]
358:  \end{equation}
359: 	To find the number of galaxies in redshift bin $(z,z+\delta z)$,
360: we integrate $\phi(M,z)$ to the limiting absolute magnitude $M_{lim}(z)$
361: over the volume of the redshift interval $z$ to $z+\delta z$. The
362: selection functions represent the number of galaxies expected per unit
363: redshift if galaxies are uniformly distributed.
364: 
365: 	The curves in Figure 2 are the UV selection functions normalized
366: to the observed number of galaxies.  The selection function for the S2000
367: LF ($M^*=-20.59, \alpha=-1.51$) is very similar to the SSRS2 Sm-Irr LF
368: ($M^*=-19.78+($UV$-B)_{\rm o},$ $\alpha=-1.81$), and matches the
369: (UV$-B)_{\rm o}>-2$ galaxy redshift distribution (Figure 2, upper panel)
370: quite well.
371: 
372: 	In contrast, the UV selection function we calculate for WR
373: galaxies predicts more WR galaxies at large redshifts.  The redshift
374: distribution of (UV$-B)_{\rm o}<-2$ galaxies (Figure 2, lower panel)
375: matches the WR selection function curves much better than the S2000 curve.
376: The shifts we make in $M^*$ reflect only the change in the flux through
377: the UV bandpass as a function of redshift.  Thus the UV emission lines of
378: WR galaxies can explain the extreme blue (UV$-B)_{\rm o}$ colors {\it and}
379: the redshift distribution of the (UV$-B)_{\rm o}<-2$ galaxies.
380: 
381: 	We calculate the number density of Century Survey (which includes
382: the SA57 region) galaxies to the same effective limiting magnitude and
383: solid angle as the UV sample, and find that the extreme blue objects are
384: $\sim$1\% of a ``normal'' ($R$-selected) galaxy population.  It is
385: difficult to compare the known WR galaxy fraction.  We estimate a lower
386: limit by assuming the WR catalog of \citet{sch99b} is complete to $B$=14
387: mag over the northern sky above 15 deg galactic latitude.  The resulting
388: WR galaxy fraction lower limit is 0.75\%.  Thus the extreme blue galaxy
389: number density is consistent with the abundance of known WR galaxies.
390: 
391: 
392: \section{CONCLUSIONS}
393: 
394: 
395: 	We show that the extreme blue (UV$-B)_{\rm o}<-2$ galaxies
396: observed by S2000 may be WR galaxies, starburst galaxies with
397: strong UV emission from WR stars.  The model explains the observed
398: (UV$-B)_{\rm o}<-2$ colors and the redshift distribution for the bluest
399: objects.
400: 
401: 	The average (\vr) color of 13 known WR galaxies taken from the
402: literature, $\overline{(\vr)}=0.27\pm0.34$, is consistent with the range
403: of (\vr) colors we observe for the extreme blue (UV$-B)_{\rm o}<-2$
404: galaxies.  We measure a median optical color $(\vr)=0.38\pm0.06$ for 67\%
405: of the S2000 SA57, UV-selected sample.  
406: 
407: 	The UV emission lines can double or triple the flux through the
408: S2000 UV bandpass at large redshift.  As a result, the (UV$-B)$ color of a
409: WR galaxy can become up to 1.3 mag bluer at larger redshift.  This
410: changing (UV$-B)$ as a function of redshift predicts a selection function
411: weighted toward large redshifts.  Not only do we find an excess of
412: (UV$-B)_{\rm o}<-2$ galaxies at large redshift, compared to the
413: (UV$-B)_{\rm o}>-2$ sample, but our model matches the observed redshift
414: distribution well. We plan to test this model by obtaining high quality
415: spectra of the suspected WR galaxies and looking for the telltale broad WR
416: emission features.
417: 
418: 
419: \acknowledgements
420: 
421: 	We thank R.~Ellis, M.~Sullivan, A.~Barth, C.~Leitherer, and the
422: referee, H.~Ferguson, for their thoughtful comments and careful reading of
423: the paper.  We thank A.~Mahdavi for the use of his ``kernel'' code to do
424: the \citet{pis93} statistical calculation.  This research has made use of
425: the NASA/IPAC Extragalactic Database (NED) which is operated by the Jet
426: Propulsion Laboratory, California Institute of Technology, under contract
427: with NASA.  This research was supported by the Harvard-Smithsonian Center
428: for Astrophysics.
429: 
430: 
431: \begin{thebibliography}{}
432: 
433: \bibitem[Bertin \&Arnouts(1996)]{ber96} Bertin, E., \& Arnouts, S. 1996,
434: \aap, 117, 393
435: 
436: \bibitem[Calzetti(1997)]{cal97} Calzetti, D. 1997, AIP Conf. Proc., 408,
437: 403
438: 
439: \bibitem[Dufour et al.(1993)]{duf93} Dufour, R. J., Skillman, E. D.,
440: Garnett, D. R., Shields, G. A., Peimbert, M., Torres-Peimbert, S.,
441: Terlevish, E., \& Terlevich, R. 1993, RMAA, 27, 115
442: 
443: \bibitem[Fioc \& Rocca-Volmerange(1997)]{fio97} Fioc, M., \&
444: Rocca-Volmerange, B. 1997, \aap, 362, 950
445: 
446: \bibitem[Folkes et al.(1999)]{fol99} Folkes, S., et al. 1999, \mnras, 308,
447: 459
448: 
449: \bibitem[Frei \& Gunn(1994)]{fre94} Frei, Z., \& Gunn, J. E. 1994, \aj,
450: 108, 1476
451: 
452: \bibitem[Freudling(1993)]{fre93} Freudling, W. 1993, in 5th ESO/ST-ECF
453: Data Analysis Workshop, ed. P.J. Grosbol \& R.C.E. de Ruijsscher (Garching
454: bei M\"{u}nchen: ESO), 27
455: 
456: \bibitem[Gallego et al.(1995)]{gal95} Gallego, J., Zamarano, J.,
457: Arag\'on-Salamanca, A., \& Rego, M. 1995, \apjl, 455, L1-L4
458: 
459: \bibitem[Geller et al.(1997)]{gel97} Geller, M. J., Kurtz, M. J., Wegner,
460: G., Thorstensen, J. R., Fabricant, D. G., Marzke, R. O., Huchra, J. P.,
461: Schild, R. E., \& Falco, E. E. 1997, \aj, 114, 2205
462: 
463: \bibitem[Hamann \& Koesterke(1998)]{ham98} Hamann, W.-R., \& Koesterke, L.
464: 1998, \aap, 333, 251
465: 
466: \bibitem[Hillier \& Miller(1999)]{hil99} Hillier, D. \& Miller, D. 1999,
467: \apj, 519, 354
468: 
469: \bibitem[Kurt et al.(1999)]{kur99} Kurt, C. M., Dufour, R. J., Garnett, D.
470: R., Skillman, E. D., Mathis, J. S., Peimbert, M., Torres-Peimbert, S., \&
471: Ruiz, M.-T. 1999, \apj, 518, 246
472: 
473: \bibitem[Landolt(1992)]{lan92} Landolt, A. 1992, \aj, 104, 340
474: 
475: \bibitem[Leitherer et al.(1999)]{lei99} Leitherer C., et al. 1999, \apjs,
476: 123, 3
477: 
478: \bibitem[Lilly et al.(1995)]{lil95} Lilly, S. J., Le F\`evre, O.,
479: Crampton, D., Hammer, F., \& Tresse, L. 1995, \apj, 455, 108
480: 
481: \bibitem[Kinney et al.(1996)]{kin96} Kinney, A. L., Calzetti, D., Bohlin,
482: R. C., McQuade, K., Storchi-Bergmann, T., \& Schmitt, H. R. 1996, \apj,
483: 467, 38
484: 
485: \bibitem[Kotilainen et al.(1993)]{kot93} Kotilainen, J. K., Ward, M. J.,
486: \& Williger, G. M. 1993, \mnras, 263, 655
487: 
488: \bibitem[McQuade et al.(1995)]{mcq95} McQuade, K., Calzetti, D., \&
489: Kinney, A. L. 1995, \apjs, 97, 331
490: 
491: \bibitem[Madau et al.(1996)]{mad96} Madau, P., Pozzetti, L., Dickinson, M.
492: E., Giavalisco, M., Steidel, C. C., \& Fruchter, A. 1996, \mnras, 283, 1388
493: 
494: \bibitem[Marzke et al.(1998)]{mar98} Marzke, R. O., da Costa, L. N.,
495: Pellegrini, P. S., Willmer, C. N., \& Geller, M. J. 1998, \apj, 503, 617
496: 
497: \bibitem[Milliard et al.(1992)]{mil92} Milliard, B., Donas, J., Laget, M.,
498: Armand, C., \& Vuillemin, A. 1992, \aap, 257, 24
499: 
500: \bibitem[Montgomery, Marschall, \& Janes(1993)]{mon93} Montgomery, K. A.,
501: Marschall, L. A., \& Janes K. A.  1993, \aj, 106, 181
502: 
503: \bibitem[Pisani(1993)]{pis93} Pisani, A. 1993, \mnras, 265, 706
504: 
505: \bibitem[Schaerer et al.(1999)]{sch99} Schaerer, D., Contini, T.,
506: \& Kunth, D. 1999, \aap, 341, 399
507: 
508: \bibitem[Schaerer et al.(1999b)]{sch99b} Schaerer, D., Contini, T.,
509: \& Pindao, M. 1999b, AASS, 136, 35
510: 
511: \bibitem[Schechter(1976)]{sch76} Schechter, P. 1976, \apj, 203, 297
512: 
513: \bibitem[Schlegel, Finkbeiner, \& Davis(1998)]{sch98} Schlegel, D. J.,
514: Finkbeiner, D. P., \& Davis, M. 1998, \apj, 500, 525
515: 
516: \bibitem[Sullivan et al.(2000)]{S2000} Sullivan, M., Treyer, M. A., Ellis,
517: R. S., Bridges, T. J., Milliard, B., \& Donas, J. 2000, \mnras, 312, 442
518: 
519: \bibitem[Treyer et al.(1998)]{tre98} Treyer, M. A., Ellis, R. S.,
520: Milliard, B., Donas, J., \& Bridges, T. J. 1998, \mnras, 300, 303
521: 
522: \bibitem[Willmer et al.(1996)]{wil96} Willmer, C. N. A., Koo D. C.,
523: Ellman, N., Kurtz, M. J., \& Szalay, A. S. 1996, \apjs, 104, 199
524: 
525: \end{thebibliography}
526: 
527: \end{document}
528: