0709.4400/ms.tex
1: \documentclass{emulateapj}
2: 
3: \def\OH{$\log({\rm O/H})+12$}
4: 
5: \begin{document}
6: \slugcomment{Accepted to AJ}
7: 
8: \title{ MMT Extremely Metal Poor Galaxy Survey I.  An Efficient Technique to 
9: Identify Metal Poor Galaxies}
10: 
11: \author{Warren R.\ Brown}
12: \affil{Smithsonian Astrophysical Observatory}
13: % \authoraddr{60 Garden St, Cambridge, MA 02138}
14: \email{wbrown@cfa.harvard.edu}
15: 
16: \author{Lisa J. Kewley\altaffilmark{1}}
17: \affil{University of Hawaii}
18: % \authoraddr{2680 Woodlawn Drive, Honolulu, HI 96822}
19: 
20: \and
21: 
22: \author{Margaret J.\ Geller}
23: \affil{Smithsonian Astrophysical Observatory}
24: % \authoraddr{60 Garden St, Cambridge, MA 02138}
25: 
26: \altaffiltext{1}{Hubble Fellow}
27: 
28: \shorttitle{ MMT Extremely Low Metallicity Galaxy Survey I.}
29: \shortauthors{Brown, Kewley \& Geller}
30: 
31: \begin{abstract}
32: 
33: 	We demonstrate a successful strategy for identifying extremely metal poor
34: galaxies.  Our preliminary survey of 24 candidates contains 10 metal poor galaxies
35: of which 4 have \OH$<7.65$, some of the lowest metallicity blue compact galaxies
36: known to date.  Interestingly, our sample of metal poor galaxies have systematically
37: lower metallicity for their luminosity than comparable samples of blue compact
38: galaxies, dIrrs, and normal star-forming galaxies.  Our metal poor galaxies share
39: very similar properties, however, with the host galaxies of nearby long-duration
40: gamma-ray bursts (GRBs), including similar metallicity, stellar ages, and star
41: formation rates.  We use H$\beta$ to measure the number of OB stars present in our
42: galaxies and estimate a core-collapse supernova rate of $\sim$10$^{-3}$ yr$^{-1}$.  
43: A larger sample of metal poor galaxies may provide new clues into the environment
44: where GRBs form and may provide a list of potential GRB hosts.
45: 
46: \end{abstract}
47: 
48: \keywords{
49: 		galaxies: abundances ---
50: 		galaxies: starburst ---
51: 		gamma rays: bursts }
52: 
53: \clearpage
54: 
55: \section{INTRODUCTION}
56: 
57: 	Metal poor galaxies are the key to understanding star formation and gas
58: enrichment in a nearly pristine interstellar medium, and may provide a template for
59: understanding the formation of the first stars.
60: 	Extremely metal poor galaxies (XMPGs) are extremely rare:  fewer than 1\% of
61: dwarf galaxies are XMPGs, with a gas-phase oxygen abundance \OH $\leq 7.65$
62: \citep{kunth00, kniazev03}.  Known XMPGs are mostly gas-rich, blue compact galaxies
63: with spectra dominated by emission lines.  The first surveys to search for XMPGs
64: were objective prism surveys \citep{macalpine77, kunth81, terlevich91}.  Abundance
65: studies of these surveys revealed up to a dozen XMPGs \citep{kunth83, campbell86,
66: masegosa94}, but none so extreme as I Zw 18 \citep{searle72}.  I Zw 18, now along
67: with SBS 0335-052W and DDO 68, are the most metal poor local galaxies known, with
68: \OH\ ranging 7.12 - 7.17 \citep{izotov05, izotov07}.  \citet{papaderos06} found 2
69: new XMPGs in the 2dF survey.  Among the 1,000,000 spectra released by the Sloan
70: Digital Sky Survey (SDSS), there are 19 identified XMPGs \citep{kniazev03, izotov04,
71: izotov06b, izotov07}, emphasizing the rarity of such objects.
72: 
73: 	\citet{kewley07} recently announced the serendipitous discovery of a new
74: XMPG, SDSS J080840.85+172856.48 (hereafter SDSS J0809+1729).  The object is a
75: stellar point source in the SDSS catalog.  It was observed as part of the
76: \citet{brown06} hypervelocity star survey on the basis of its stellar B-type
77: photometric colors.  Spectroscopy reveals that the object is not a star but rather a
78: compact blue galaxy at $cz=13232$ km s$^{-1}$.  Our re-analysis shows that the
79: galaxy has \OH$ =7.48 \pm 0.1$.  Interestingly, the observed electron density, star
80: formation rate, and total luminosity of this XMPG are remarkably similar to nearby
81: GRB host galaxies \citep{kewley07}.
82: 
83: 	Nearby $z<0.2$ long duration GRBs are observed in metal poor galaxies
84: \citep{prochaska04, sollerman05, fruchter06, stanek06, wolf07, wiersema07,
85: margutti07}.  \citet{fruchter06} argue that the link between metal poor galaxies and
86: GRBs originates in the atmospheres of massive stars.  Massive metal poor stars lack
87: the opacity to support significant stellar winds, and thus can produce the anomalous
88: Type 1c supernovae associated with nearby GRBs.  \citet{berger07} argue that GRBs
89: are linked to young starburst populations, which at low redshift happen to be found
90: predominantly in low mass, metal-poor galaxies.  At large redshift $z>0.2$, the
91: metallicity of GRB hosts is much harder to determine \citep{prochaska06}.  At least
92: one GRB host galaxy is an extremely red and probably metal rich object
93: \citep{berger07}, and other GRB hosts $0.2<z<1$ appear consistent with normal
94: metallicity-luminosity relations \citep{wolf07, margutti07}.  GRBs clearly occur in
95: different types of host galaxies; they are not tied exclusively to the most metal
96: poor galaxies.  Yet finding and studying metal poor galaxies with properties similar
97: to nearby GRB host galaxies may yield critical clues about the environment where
98: nearby GRBs form.
99: 
100: 	Inspired by the discovery of SDSS J0809+1729, we designed a survey to find
101: metal poor galaxies.  We use a technique similar to photometric redshifts to
102: identify metal poor galaxies in the SDSS galaxy catalog.  We test this technique and
103: target $g'\sim20$ galaxies with very blue colors, a region of parameter space not
104: well probed by previous surveys.  Our strategy uncovers 10 metal poor galaxies from
105: a sample of 24 candidates, 4 of which are new XMPGs.
106: 
107: 	In \S 2 we present our technique to find new metal poor galaxies and discuss
108: the efficacy of our survey.  In \S 3 we describe the properties of the entire set of
109: metal poor galaxies, and compare the galaxies with samples of blue compact galaxies
110: and nearby GRB hosts.  In \S 4 we estimate the expected core-collapse supernova rate
111: in our galaxies.  We conclude in \S 5.
112: 
113: \begin{figure}		% FIGURE 1: SPECTRUM * FILTER
114:  \includegraphics[width=3.25in]{f1.eps}
115:  \caption{ \label{fig:filters} Observed spectrum of the XMPG SDSS J0809+1729
116: \citep{kewley07} plotted against the total system throughput of the SDSS
117: $u'~g'~r'~i'$ filter bandpasses (dotted lines).  The strong emission line
118: contribution to the galaxy's broadband magnitudes results in unusual colors that
119: vary with redshift. }
120:  \end{figure}
121: 
122: \section{DATA}
123: 
124: \subsection{Technique to Find Extremely Metal Poor Galaxies}
125: 
126: 	Known XMPGs outside the Local Group are starburst galaxies characterized by
127: very low internal extinction, high ionization parameter, and large gas-phase
128: electron density \citep[e.g.][]{kniazev03, izotov05, izotov06b, papaderos06,
129: kewley07, izotov07}.  As a result, XMPGs have steep blue continua, large [O{\sc
130: iii}]/[O{\sc ii}] ratios, and strong hydrogen Balmer emission lines.  These
131: characteristics significantly affect the broadband colors of XMPGs.  For example,
132: Figure \ref{fig:filters} shows the observed spectrum of SDSS J0809+1729 plotted
133: against the total system throughput of the SDSS $u'g'r'i'$ filter
134: bandpasses\footnote{http://www.sdss.org/dr5/instruments/imager/index.html}.  The
135: contribution of emission lines to the broadband magnitudes of SDSS J0809+1729 is
136: approximately 3\%, 32\%, 6\%, and 12\% at $u'$, $g'$, $r'$, and $i'$ respectively.
137: 
138: 	The strong emission line contribution to XMPGs' broadband magnitudes results
139: in unusual colors that change with redshift.  For example, at lower redshift, SDSS
140: J0809+1729 does not satisfy the B-star criteria in \citet{brown06b} because
141: H$\alpha$ drops into the $r'$ band and produces a much redder $(g'-r')$ color.  
142: Conversely, at higher redshift, H$\beta$ and \ion{O}{3} move from $g'$ into $r'$ and
143: also produce a much redder $(g'-r')$ color (see Figure \ref{fig:filters}).  We
144: quantify these effects using the RVSAO package \citep{kurtz98} to shift the XMPG
145: spectra to different redshifts. We predict the galaxy's colors at different
146: redshifts by adding the {\it relative} change in color to the observed photometry.  
147: In effect, we are calculating $k$-corrections for the XMPG.
148: 
149: 
150: \begin{figure}		% FIGURE 2: COLOR-REDSHIFT TRACKS
151:  \includegraphics[width=3.25in]{f2.eps}
152:  \caption{ \label{fig:tracks}
153: 	Target selection for our MMT survey for XMPGs. a) Color-color plot showing
154: all SDSS DR4 galaxies (dots) and the color-redshift track for the XMPG J0809+1729
155: (solid line).  b) - d) Color-redshift plots showing SDSS objects (dots) that
156: simultaneously match all three color-redshift tracks (solid lines) within 0.1 mag.  
157: After removing saturated stars and nearby H{\sc ii} regions by visual inspection, we
158: are left with a sample of 38 XMPG candidates (x's).  Four new XMPGs discovered by
159: this survey are indicated by stars. }
160:  \end{figure}
161: 
162: 
163: 	Figure \ref{fig:tracks} plots the resulting color-redshift tracks for SDSS
164: J0809+1729 (solid lines).  Over the range $0.0<z<0.10$, the XMPG's $(u'-g')_0$ and
165: $(g'-r')_0$ colors vary by $\sim0.3$ mag and the $(r'-i')_0$ color varies by $\sim$1
166: mag.  The subscript 0 indicates colors corrected for Galactic extinction following
167: \citet{schlegel98}.  We note that the Johnson/Cousins passbands are much less
168: sensitive than their SDSS equivalents because the strong H$\beta$/[\ion{O}{3}] and
169: H$\alpha$ emission lines remain in their respective $V$ and $R$ passbands for
170: $0.0<z<0.10$.
171: 
172: 	Knowing how the broadband colors of an XMPG change with redshift, we can
173: search for new XMPGs at other redshifts in the SDSS photometric catalog.  
174: Unfortunately, searching for new XMPGs like SDSS J0809+1729 in the stellar catalog
175: is not feasible because of immense contamination from white dwarfs with similar
176: colors.  Instead, we search the SDSS galaxy catalog.
177: 
178: 	We use the color-redshift track of our newly discovered XMPG, SDSS
179: J0809+1729, as the basis for the sample of XMPGs published here.  We begin by
180: selecting all galaxies in SDSS DR4 \citep{adelman06} with $g'<20.5$.  Figure
181: \ref{fig:tracks}a plots these galaxies with similar $(u'-g')_0$ and $(g'-r')_0$ to
182: SDSS J0809+1729.  Only by combining two or more colors can we meaningfully select
183: XMPG candidates.  We find $\sim10^3$ galaxies simultaneously within 0.1 mag of any
184: pair of color-redshift tracks, and a mere 107 galaxies (small squares, Figure
185: \ref{fig:tracks}b-d) simultaneously within 0.1 mag of all three color-redshift
186: tracks.  We visually inspect the objects and find that many are near saturated stars
187: or that they are HII regions in nearby galaxies.  After eliminating the unwanted
188: objects, we are left with a sample of 38 photometrically-selected XMPG candidates
189: (marked by x's, Figure \ref{fig:tracks}b-d).  Our survey is based on this sample of
190: 38 XMPG candidates.
191: 
192: 	We use a restrictive color selection as the first demonstration of our
193: technique.  If our XMPG selection strategy is successful, we can easily broaden the
194: search parameters to identify many more faint XMPG candidates.
195: 
196: \subsection{Observations}
197: 
198: 	We obtained spectroscopy of the 24 candidates available on the nights of
199: 2006 May 25-27 and 2006 June 19-20.  Table \ref{tab:other} lists the 24 candidates.  
200: Observations were obtained with the 6.5m MMT telescope and the Blue Channel
201: spectrograph.  We operated the spectrograph with the 300 line mm$^{-1}$ grating and
202: a 1$\arcsec$ slit.  These settings provide a wavelength coverage of 3400 \AA\ to
203: 8600 \AA\ and a spectral resolution of 6.2 \AA.  Exposure times were typically 30
204: minutes.  We obtained comparison lamp exposures after every exposure.
205: 
206: 	We reduce the spectra using standard IRAF\footnote{
207:         IRAF is distributed by the National Optical Astronomy Observatories, which
208: are operated by the Association of Universities for Research in Astronomy, Inc.,
209: under cooperative agreement with the National Science Foundation.}
210: 	spectral reduction tasks and measure recession velocities from emission
211: lines using the package RVSAO \citep{kurtz98}.  We flux calibrate using
212: spectrophotometric standards \citep{massey88} and the standard Kitt Peak atmospheric
213: extinction correction.  For objects obtained in non-photometric conditions, we scale
214: the spectra by the flux ratio of the observed spectroscopic and SDSS broadband
215: magnitudes.  We estimate that absolute flux calibration is accurate to $\sim$25\%.
216: 
217: 	We measure emission line fluxes using IRAF {\it splot} and {\it fitprof}
218: tasks, and find no significant offset between the two methods.  Table
219: \ref{tab:lineflux} presents the observed line strengths with their measurement
220: uncertainties.  The statistical uncertainties are formally a few percent, but the 
221: true error is dominated by uncertainties in the reddening correction, the stellar 
222: absorption correction, and the absolute flux calibration.
223: 
224: 	For our analysis, we correct the observed emission line fluxes for reddening
225: using the Balmer decrement and the \citet{cardelli89} reddening curve.  We assumed
226: an $R_{V}=A_{V}/{\rm E}(\bv) = 3.1$ and an intrinsic H$\alpha$/H$\beta$ ratio of
227: 2.85 \citep[the Balmer decrement for case B recombination at T$=10^4$K and $n_{e}
228: \sim 10^2 - 10^4 {\rm cm}^{-3}$;][]{osterbrock89}.
229: 
230: 	We find no evidence for stellar absorption in the expected sense.  While one
231: would expect some underlying stellar absorption, we do not see Stark-broadened
232: absorption in the wings of the Balmer emission lines.  The de-reddened Balmer
233: emission line ratios exhibit a 7\% scatter around the \citet{osterbrock89} values
234: for case B recombination at T$=10^4$ K.  If we apply a constant 2 \AA\ equivalent
235: width correction, appropriate for the young 4-5 Myr stellar age of the galaxies (see
236: \S 3.4), the scatter of the Balmer line ratios around the \citet{osterbrock89}
237: values remains unchanged at 7\%.  We conclude that stellar absorption is smaller
238: than the uncertainty in the line ratio measurements.
239: 
240: 	We calculate electron densities with the \ion{S}{2} $\lambda 6717$ /
241: \ion{S}{2} $\lambda 6731$ line ratio, when present, in conjunction with a 5 level
242: model atom using the Mappings photoionization code \citep{sutherland93}.  We derive
243: the gas-phase oxygen abundance following the procedure outlined in \citet{izotov06a}
244: within the framework of the classical two-zone HII-region model \citep{stasinska80}.  
245: This procedure utilizes the electron-temperature T$_{\rm e}$ calibrations of
246: \citet{aller84} and the atomic data compiled by \citet{stasinska05}.
247: 
248: 	The gas-phase oxygen abundance depends on line ratios and thus is
249: independent of the uncertainties in our absolute flux calibration.  We propagate the
250: errors from the line ratio measurement, the extinction correction, and the stellar
251: absorption correction, and find that the relative errors in metallicities derived
252: using the same method are formally $\le 0.07$ dex.  However, the absolute error is
253: at least $\sim 0.1$ dex.  Thus systematics dominate the errors; our metallicities
254: are accurate at the $\pm0.1$ dex level.
255: 
256: \begin{figure}          % FIGURE 3: XMPG SPECTRA and PICS
257:   \includegraphics[scale=0.89]{f3aa.eps}
258:   \includegraphics[scale=0.43]{f3b.eps}
259:  \caption{ \label{fig:spectra}
260:         MMT spectra of our 4 new XMPGs.  [\ion{O}{3}] $\lambda$4363 is well-detected
261: in all four objects (see insets).  The extremely weak [N{\sc ii}] and [S{\sc ii}]
262: lines visibly indicate the extremely low abundance of these objects.  The thumbnail
263: images, courtesy of SDSS, are 48\arcsec on a side and show that all 4 XMPGs appear
264: to be compact dwarfs.}
265:  \end{figure}
266: 
267: \subsection{Survey Efficiency}
268: 
269: 	Our initial survey of 24 candidates contains 10 metal poor galaxies
270: with \OH $<8$ of which 4 are new XMPGs.  Thus our strategy is $\sim$20\% efficient
271: for selecting XMPGs.  Spectra and thumbnail images of the 4 new XMPGs are displayed
272: in Figure \ref{fig:spectra} (see also Table \ref{tab:gals}).  Three of the new XMPGs
273: (SDSS J142250.72+514516.5, SDSS J144158.32+291434.2 and SDSS J225900.86+141343.5)
274: have lower abundance than SDSS J0809+1729.  The remaining objects in our survey are
275: either A stars in the Milky Way, galaxies with modest emission lines, and a few odd
276: objects (one E+A, one possible BL Lac, one possible quasar) listed in Table
277: \ref{tab:other}.
278: 
279: 
280: \section{GALAXY PROPERTIES}
281: 
282: 	We now open our discussion to include all the metal poor galaxies in our
283: survey.  Including SDSS J0809+1729, our survey contains 5 XMPGs and 6 metal poor
284: galaxies.  Table \ref{tab:gals} summarizes the spectroscopic measurements for the 11
285: galaxies.  Because the galaxies are compact objects, our spectra provide reasonable
286: global estimates of their properties.  We compare the properties of our galaxies
287: with similar samples of blue compact galaxies (BCGs) and with nearby GRB hosts.  
288: Throughout this paper we adopt a flat, $\Lambda$-dominated cosmology with $H_o=70$
289: and $\Omega_{m}=0.3$.
290: 
291: \begin{figure}		% FIGURE 4:  COLOR-MAG PLOT
292:  \includegraphics[width=3.0in]{f4.eps}
293:  \caption{ \label{fig:colormag}
294: 	Color-magnitude distribution of nearby BCGs \citep{paz03} and our 11 metal
295: poor galaxies, for which we estimate $(B-R)$ from SDSS photometry
296: \citep{fukugita96}.}
297:  \end{figure}
298: 
299: \subsection{Color and Redshift Distribution}
300: 
301: 	Our 11 metal poor galaxies are systematically bluer and more luminous than
302: other known BCGs.  Figure \ref{fig:colormag} plots the \citet{paz03} sample of
303: nearby BCGs, for which $(B-R)_0$ and $M_B$ are all available.  We estimate $(B-R)$
304: for our galaxies from SDSS photometry \citep{fukugita96}, and shift the observed
305: magnitudes to the rest-frame using $k$-corrections we calculate for $B$ and $R$
306: passbands as described in \S 2.  It is clear that our metal poor galaxies are more
307: than 0.5 mag bluer in $(B-R)_0$ than most nearby BCGs.  I Zw 18 and UCM 1612+1308
308: \citep{rego98} are the two BCGs with colors comparable to our metal poor galaxies.  
309: However, our metal poor galaxies are systematically more luminous than the BCGs with
310: similar colors.
311: 
312: 	Our 11 metal poor galaxies are also at greater redshift than most known
313: BCGs.  Figure \ref{fig:zmag} plots the redshift distribution of BCGs
314: \citep{kong02,paz03}, metal poor galaxies from 2dF \citep{papaderos06} and SDSS
315: \citep{kniazev03}, our 11 metal poor galaxies, and the 3 nearest GRB host galaxies
316: \citep{stanek06}.  Note that we calculate $M_B$ for the 2dF galaxies assuming the
317: average $(B-V)=0.5$ for that sample \citep{papaderos06}.  We estimate $M_B$ for the
318: SDSS galaxies from $g'$ and $r'$ photometry in \citet{kniazev03}.  Figure
319: \ref{fig:zmag} shows that the vast majority of known BCGs have redshifts $z<0.02$.  
320: The 2dF and SDSS surveys access fainter magnitudes than earlier BCG samples and thus
321: contain a number of very low luminosity galaxies, plus a few higher luminosity metal
322: poor galaxies at $z\sim0.04$.  Our sample of metal poor galaxies, by construction,
323: spans the range $0.02<z<0.08$.  Searching a large volume of space enables the
324: discovery of rare objects like XMPGs.
325: 
326: \begin{figure}		% FIGURE 5:  REDSHIFT PLOT
327:  \includegraphics[width=3.0in]{f5.eps}
328:  \caption{ \label{fig:zmag}
329: 	Redshift and luminosity distribution of BCGs \citep{kong02,paz03},
330: metal poor galaxies from 2dF \citep{papaderos06} and SDSS \citep{kniazev03},
331: our 11 metal poor galaxies, and the 3 nearest GRB host galaxies
332: \citep{stanek06}.  $M_B$ is calculated for the 2dF galaxies assuming 
333: $(B-V)=0.5$ \citep{papaderos06}.  $M_B$ is estimated for the SDSS galaxies from 
334: SDSS photometry \citep{kniazev03}.}
335:  \end{figure}
336: 
337: \subsection{Luminosity-Metallicity Relation}
338: 
339: 	Luminosity-metallicity relations are well determined for galaxies ranging
340: from large star-forming galaxies \citep[e.g.][]{tremonti04} to dwarf irregulars
341: \citep[e.g.][]{richer95}.  The physical basis for the luminosity-metallicity
342: relation is a mass-metallicity relation:  low mass galaxies are thought to sustain
343: less star formation and retain fewer metals than high mass galaxies.  We now compare
344: the luminosity and metallicity of our galaxies with samples of comparable metal-poor
345: dwarf galaxies.  Because there is some disagreement about the calibration of
346: different metallicity-estimate methods, we only consider galaxy samples with
347: metallicities derived with the T$_{\rm e}$ method.
348: 
349: 	Figure \ref{fig:lz} plots samples of BCGs \citep{kong02, shi05}, metal poor
350: galaxies from 2dF \citep{papaderos06} and SDSS \citep{kniazev03}, our metal poor
351: galaxies, and four nearby GRB host galaxies \citep{stanek06}.  We use metallicities
352: for the GRB hosts calculated using the T$_{\rm e}$ method as described in
353: \citet{kewley07}.  Figure \ref{fig:lz} shows that our sample of galaxies has either
354: 1) lower metallicity by $\sim$0.5 dex than the \citet{richer95}
355: luminosity-metallicity relation for normal dIrrs, or 2) higher luminosity by 3 - 5
356: mag in $M_B$.
357: 
358: 	Other samples of nearby, metal poor galaxies exhibit a large scatter around
359: the luminosity-metallicity relation \citep{kunth00}.  Yet our galaxies appear
360: unusual because of their large {\it systematic} offset from the
361: luminosity-metallicity relation.  We will use population synthesis models to address
362: the evolutionary paths of our metal poor galaxies in a future paper.
363: 
364: \begin{figure}		% FIGURE 6:  L-Z PLOT --- T_e
365:  \includegraphics[width=3.0in]{f6.eps}
366:  \caption{ \label{fig:lz}
367: 	Luminosity-metallicity plot for BCGs \citep{kong02,shi05}, metal poor
368: galaxies from 2dF \citep{papaderos06} and SDSS \citep{kniazev03}, our metal poor
369: galaxies, and four nearby GRB host galaxies \citep{stanek06}.  Our sample of metal
370: poor galaxies fills in the region defined by the GRB hosts; all these galaxies have
371: lower metallicity by $\sim$0.5 dex than the \citet{richer95} luminosity-metallicity
372: relation ({\it dashed line}) for normal dIrrs.}
373:  \end{figure}
374: 
375: 	Remarkably, our sample of metal poor galaxies fills the region of the
376: luminosity-metallicity diagram outlined by nearby GRB host galaxies.  The only other
377: galaxies with similar properties are the five metal poor 2dF and SDSS galaxies at
378: $z\sim0.04$:  2dF 169299, 2dF 115901, SDSSJ051902.64+000730.0,
379: SDSSJ104457.84+035313.2, and SDSSJ084030.00+470710.2.  The nearby BCG II Zw 70 also
380: appears to fall in the region defined by the GRB hosts.
381: 
382: 	The relative distribution of galaxies in the luminosity-metallicity plot
383: does not change with different metallicity estimators.  To illustrate this point, we
384: calculate strong line metallicities using R$_{23}$ \citep{mcgaugh91} for the entire
385: set of galaxies.  The results are plotted in Figure \ref{fig:lz2}.  The R$_{23}$
386: metallicities have an average offset of +0.2 dex from the T$_{\rm e}$ metallicities,
387: and so we shift the \citet{richer95} luminosity-metallicity relation in Figure
388: \ref{fig:lz2} by a constant +0.2 dex for consistency.  Yet the relative distribution
389: of galaxies remains the same: our metal poor galaxies (and the GRB host galaxies)  
390: maintain a large, systematic offset from the luminosity-metallicity relation defined
391: by normal dIrrs and BCGs.
392: 
393: \subsection{Extinction}
394: 
395: 	Our sample of 11 metal poor galaxies suffer from very little internal
396: extinction.  Table \ref{tab:gals} lists the Balmer H$\alpha$/H$\beta$ ratios, which
397: average $2.8\pm0.2$.  The intrinsic H$\alpha$/H$\beta$ ratio is 2.86 for case B
398: recombination at T$=10^4$ K and $n_e\sim10^2 - 10^4$ cm$^{-3}$ \citep{osterbrock89}.  
399: Low extinction is expected in low metallicity galaxies, and is also observed in
400: nearby GRB hosts \citep[e.g.][]{kewley07}.
401: 
402: \subsection{Star Formation Age and Rates}
403: 
404: 	XMPGs remain a puzzle because they may be pristine galaxies undergoing their
405: first burst of star formation or they may contain older stellar population from
406: previous episodes of star formation.  {\it Hubble Space Telescope} images resolve
407: old stellar populations in the nearest metal poor galaxies.  I Zw 18, for example,
408: has stars with ages ranging from $\sim 500$~Myr \citep{izotov04b} to $\sim 1$~Gyr
409: \citep{aloisi99}.  Although we cannot estimate the age of old stellar populations
410: (if any) in our metal poor galaxies, stellar population synthesis models provide an
411: estimate of the age of the young stellar population.
412: 
413: 	We estimate the stellar age of our metal poor galaxies from the H$\beta$
414: equivalent width following \citet{schaerer98}.  Under the assumption of a Salpeter
415: initial mass function and an instantaneous burst of star formation, the stellar ages
416: of our metal poor galaxies are in the range 4 - 5 Myr (see Table \ref{tab:gals}).
417: 
418: 	Star formation rates (SFRs) are more difficult to estimate because
419: traditional calibrations are not applicable to extremely metal poor objects.  We use
420: the calibration derived by \citet{kewley07} based on the stellar population
421: synthesis models of \citet{bicker05}.  Our metal poor galaxies then have SFRs in the
422: range 0.1 - 0.5 M$_{\sun}$ yr$^{-1}$ (Table \ref{tab:gals}).  Both the stellar ages
423: and SFRs of our metal poor galaxies are remarkably similar to those found for nearby
424: GRB hosts \citep{kewley07}.
425: 
426: \begin{figure}		% FIGURE 7:  L-Z PLOT --- R23
427:  \includegraphics[width=3.0in]{f7.eps}
428:  \caption{ \label{fig:lz2}
429: 	Same as Figure \ref{fig:lz}, except we plot strong line metallicities
430: calculated using R$_{23}$ \citet{mcgaugh91}.  The strong line metallicities are
431: systematically +0.2 dex larger than the T$_{\rm e}$ metallicities, but the overall
432: distribution remains the same:  our sample of metal poor galaxies and the GRB hosts
433: are systematically offset from the \citet{richer95} luminosity-metallicity relation
434: ({\it dashed line}), which we shift by +0.2 dex for consistency.}
435:  \end{figure}
436: 
437: 
438: \section{SUPERNOVA RATES AND THE GRB CONNECTION}
439: 
440: 	We next ask how frequently GRBs might occur in our sample of metal poor
441: galaxies, if GRBs are indeed associated with core-collapse supernova in metal poor
442: galaxies.  We start by estimating the number of massive O and B stars in our sample
443: of 11 galaxies, and then estimate the rate of core-collapse supernovae.
444: 
445: 	Under the assumption of an ionization-bounded nebula, the H$\beta$ line
446: luminosity provides an estimate of the ionizing flux present in the galaxies.  The
447: metal poor galaxies suffer from very little extinction, as measured by their
448: H$\alpha$/H$\beta$ ratios; thus we make no correction for internal extinction.
449: 	We follow \citet{schaerer98} and convert the observed H$\beta$ ionizing flux
450: to an equivalent number of O stars.  We select the ``equivalent O7V to O star''
451: ratio $\eta_o$ based on the stellar age and metallicity of the galaxy; $\eta_o$
452: values range from 0.25 - 0.5.  The minimum O star mass is $\sim$13.3 M$_\sun$ at our
453: metallicities \citep{vacca94}.  However, core-collapse supernovae will result from B
454: stars with masses as low as $\sim$8 M$_\sun$.  A Salpeter initial mass function has
455: the same number of O stars with 13 - 120 M$_\sun$ as B stars with 8 - 13 M$_\sun$.  
456: Thus, the total number of core-collapse supernova progenitors is roughly twice the
457: \citet{schaerer98} number of equivalent O stars.  We list our estimate of the total
458: number of core-collapse supernova progenitors in the column n(OB) (Table
459: \ref{tab:gals}).  On average, there are $\sim$25,000 potential core-collapse
460: supernova progenitors per galaxy in our sample.  This number may appear relatively
461: small, but the metal poor galaxies are dwarfs, not massive galaxies.
462: 
463: 	We estimate core-collapse supernova rates in our metal poor galaxies by
464: assuming supernovae occur uniformly over the lifetime of the longest-lived
465: progenitor.  The \citet{schaller92} stellar evolution track for a 9 M$_\sun$ star
466: with $Z=0.001$ has a lifetime of 30 Myr.  Thus the average rate of core-collapse
467: supernovae in our metal poor galaxies is $\sim$10$^{-3}$ yr$^{-1}$, with an
468: uncertainty of a factor of a few.  In other words, a couple thousand such metal poor
469: galaxies must be monitored to witness one core-collapse supernova per year.  A
470: typical spiral galaxy, in comparison, has a typical SFR of $\sim$10 M$_\sun$
471: yr$^{-1}$ \citep[e.g.][]{brinchmann04} and a core-collapse supernova rate $\sim$100
472: times larger than the metal poor galaxies, assuming a constant SFR and Salpeter
473: initial mass function.
474: 
475: 	The ratio of GRBs to core-collapse supernovae events depends on how narrowly
476: beamed GRBs are.  If 1 out of 100 core-collapse supernovae appear as GRBs, then a
477: sample of a $\sim$10$^5$ metal poor galaxies must be monitored to witness one GRB
478: per year.  There is currently little constraint on the space density of metal poor
479: galaxies, in part because they are such low-luminosity systems.  Surveys reaching
480: faint magnitude limits may uncover large numbers of new metal poor galaxies.  
481: However, it may be difficult to detect a supernova coincident on the high
482: surface-brightness core of these galaxies.  Figure \ref{fig:spectra} shows that the
483: four XMPGs, for example, are compact systems.  Although metal poor galaxies are
484: potential GRB hosts, a survey to find GRBs by monitoring a sample metal poor
485: galaxies appears impractical currently.  Future deep imaging surveys, such as
486: Pan-STARRS and the Large Synoptic Survey Telescope (LSST), may be able to detect
487: supernovae in these low-luminosity galaxies.
488: 
489: 
490: \section{CONCLUSIONS}
491: 
492: 	We have designed a successful strategy to find new metal poor galaxies.  We
493: calculate the expected colors of metal poor galaxies at different redshifts based on
494: the spectrum of the newly discovered XMPG, SDSS J0809+1729 \citep{kewley07}.  We
495: observed an initial sample of 24 candidates with the MMT telescope and find 4 new
496: XMPGs with \OH $\leq 7.65$, a $\sim$20\% selection efficiency.
497: 
498: 	Our full set of 11 metal poor galaxies are systematically bluer and more
499: luminous than comparable samples of BCGs.  Our galaxies are also systematically more
500: metal poor by 0.5 dex (or more luminous by 3 - 5 mag) than samples of BCGs, dIrrs,
501: and other metal-poor dwarf galaxies.  Remarkably, our galaxies share the same region
502: of the luminosity-metallicity diagram with nearby GRB hosts.  The similarity in
503: extinction, stellar age, and star formation rates suggests that our metal poor
504: galaxies are potential hosts for GRBs.
505: 
506: 	We estimate an average core-collapse supernova rate $\sim$$10^{-3}$
507: yr$^{-1}$ in our metal poor galaxies.  This estimate comes from an estimate of the 
508: number of O and B stars in the galaxies.  If GRBs are indeed linked to core-collapse 
509: supernova in metal poor galaxies, future surveys such as Pan-STARRS or LSST may be 
510: able to find GRBs by monitoring a large sample of metal poor galaxies.
511: 
512: 	The success of our XMPG selection strategy allows us to expand our survey.  
513: For example, using the new XMPGs (Figure \ref{fig:spectra}) as additional templates,
514: we identify a total of 335 XMPG candidates in SDSS Data Release 5.  Spectroscopic
515: observations of these XMPG candidates are underway.
516: 
517: 
518: \acknowledgements
519: 
520: 	W.~R.~Brown was supported in part by a Clay Fellowship during this work.  
521: L.~J.~Kewley was supported by a Hubble Fellowship.  We thank S.\ Kenyon and the
522: anonymous referee for comments that greatly improved this paper.  This research has
523: made use of NASA's Astrophysics Data System Bibliographic Services.  This project
524: makes use of data products from the Sloan Digital Sky Survey, which is managed by
525: the Astrophysical Research Consortium for the Participating Institutions.  We thank
526: the Smithsonian Institution for partial support of this research.
527: 
528: {\it Facilities:} MMT (Blue Channel Spectrograph)
529: 
530: 
531: 	% REFERENCES 
532: % \clearpage
533: % \bibliographystyle{/home/wbrown/lib/apj} \bibliography{/home/wbrown/text/RefHS}
534: \begin{thebibliography}{52}
535: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
536: 
537: \bibitem[{{Adelman-McCarthy} {et~al.}(2006)}]{adelman06}
538: {Adelman-McCarthy}, J.~K. {et~al.} 2006, \apjs, 162, 38
539: 
540: \bibitem[{{Aller}(1984)}]{aller84}
541: {Aller}, L.~H. 1984, in Astrophysics \& Space Science Library vol.\ 112, ed.
542:   L.~H. {Aller}
543: 
544: \bibitem[{{Aloisi} {et~al.}(1999){Aloisi}, {Tosi}, \& {Greggio}}]{aloisi99}
545: {Aloisi}, A., {Tosi}, M., \& {Greggio}, L. 1999, \aj, 118, 302
546: 
547: \bibitem[{{Berger} {et~al.}(2007){Berger}, {Fox}, {Kulkarni}, {Frail}, \&
548:   {Djorgovski}}]{berger07}
549: {Berger}, E., {Fox}, D.~B., {Kulkarni}, S.~R., {Frail}, D.~A., \& {Djorgovski},
550:   S.~G. 2007, \apj, 660, 504
551: 
552: \bibitem[{{Bicker} \& {Fritze-v.~Alvensleben}(2005)}]{bicker05}
553: {Bicker}, J. \& {Fritze-v.~Alvensleben}, U. 2005, \aap, 443, L19
554: 
555: \bibitem[{{Brinchmann} {et~al.}(2004){Brinchmann}, {Charlot}, {White},
556:   {Tremonti}, {Kauffmann}, {Heckman}, \& {Brinkmann}}]{brinchmann04}
557: {Brinchmann}, J., {Charlot}, S., {White}, S.~D.~M., {Tremonti}, C.,
558:   {Kauffmann}, G., {Heckman}, T., \& {Brinkmann}, J. 2004, \mnras, 351, 1151
559: 
560: \bibitem[{{Brown} {et~al.}(2006{\natexlab{a}}){Brown}, {Geller}, {Kenyon}, \&
561:   {Kurtz}}]{brown06}
562: {Brown}, W.~R., {Geller}, M.~J., {Kenyon}, S.~J., \& {Kurtz}, M.~J.
563:   2006{\natexlab{a}}, \apjl, 640, L35
564: 
565: \bibitem[{{Brown} {et~al.}(2006{\natexlab{b}}){Brown}, {Geller}, {Kenyon}, \&
566:   {Kurtz}}]{brown06b}
567: ---. 2006{\natexlab{b}}, \apj, 647, 303~(Paper I)
568: 
569: \bibitem[{{Campbell} {et~al.}(1986){Campbell}, {Terlevich}, \&
570:   {Melnick}}]{campbell86}
571: {Campbell}, A., {Terlevich}, R., \& {Melnick}, J. 1986, \mnras, 223, 811
572: 
573: \bibitem[{{Cardelli} {et~al.}(1989){Cardelli}, {Clayton}, \&
574:   {Mathis}}]{cardelli89}
575: {Cardelli}, J.~A., {Clayton}, G.~C., \& {Mathis}, J.~S. 1989, \apj, 345, 245
576: 
577: \bibitem[{{Fruchter} {et~al.}(2006)}]{fruchter06}
578: {Fruchter}, A.~S. {et~al.} 2006, \nat, 441, 463
579: 
580: \bibitem[{{Fukugita} {et~al.}(1996){Fukugita}, {Ichikawa}, {Gunn}, {Doi},
581:   {Shimasaku}, \& {Schneider}}]{fukugita96}
582: {Fukugita}, M., {Ichikawa}, T., {Gunn}, J.~E., {Doi}, M., {Shimasaku}, K., \&
583:   {Schneider}, D.~P. 1996, \aj, 111, 1748
584: 
585: \bibitem[{{Gil de Paz} {et~al.}(2003){Gil de Paz}, {Madore}, \&
586:   {Pevunova}}]{paz03}
587: {Gil de Paz}, A., {Madore}, B.~F., \& {Pevunova}, O. 2003, \apjs, 147, 29
588: 
589: \bibitem[{{Izotov} {et~al.}(2006{\natexlab{a}}){Izotov}, {Papaderos}, {Guseva},
590:   {Fricke}, \& {Thuan}}]{izotov06b}
591: {Izotov}, Y.~I., {Papaderos}, P., {Guseva}, N.~G., {Fricke}, K.~J., \& {Thuan},
592:   T.~X. 2006{\natexlab{a}}, \aap, 454, 137
593: 
594: \bibitem[{{Izotov} {et~al.}(2004){Izotov}, {Stasi{\'n}ska}, {Guseva}, \&
595:   {Thuan}}]{izotov04}
596: {Izotov}, Y.~I., {Stasi{\'n}ska}, G., {Guseva}, N.~G., \& {Thuan}, T.~X. 2004,
597:   \aap, 415, 87
598: 
599: \bibitem[{{Izotov} {et~al.}(2006{\natexlab{b}}){Izotov}, {Stasi{\'n}ska},
600:   {Meynet}, {Guseva}, \& {Thuan}}]{izotov06a}
601: {Izotov}, Y.~I., {Stasi{\'n}ska}, G., {Meynet}, G., {Guseva}, N.~G., \&
602:   {Thuan}, T.~X. 2006{\natexlab{b}}, \aap, 448, 955
603: 
604: \bibitem[{{Izotov} \& {Thuan}(2004)}]{izotov04b}
605: {Izotov}, Y.~I. \& {Thuan}, T.~X. 2004, \apj, 616, 768
606: 
607: \bibitem[{{Izotov} \& {Thuan}(2007)}]{izotov07}
608: ---. 2007, preprint astro-ph/0704.3842, 704
609: 
610: \bibitem[{{Izotov} {et~al.}(2005){Izotov}, {Thuan}, \& {Guseva}}]{izotov05}
611: {Izotov}, Y.~I., {Thuan}, T.~X., \& {Guseva}, N.~G. 2005, \apj, 632, 210
612: 
613: \bibitem[{{Kewley} {et~al.}(2007){Kewley}, {Brown}, {Geller}, {Kenyon}, \&
614:   {Kurtz}}]{kewley07}
615: {Kewley}, L.~J., {Brown}, W.~R., {Geller}, M.~J., {Kenyon}, S.~J., \& {Kurtz},
616:   M.~J. 2007, \aj, 133, 882
617: 
618: \bibitem[{{Kniazev} {et~al.}(2003){Kniazev}, {Grebel}, {Hao}, {Strauss},
619:   {Brinkmann}, \& {Fukugita}}]{kniazev03}
620: {Kniazev}, A.~Y., {Grebel}, E.~K., {Hao}, L., {Strauss}, M.~A., {Brinkmann},
621:   J., \& {Fukugita}, M. 2003, \apjl, 593, L73
622: 
623: \bibitem[{{Kong} \& {Cheng}(2002)}]{kong02}
624: {Kong}, X. \& {Cheng}, F.~Z. 2002, \aap, 389, 845
625: 
626: \bibitem[{{Kunth} \& {{\"O}stlin}(2000)}]{kunth00}
627: {Kunth}, D. \& {{\"O}stlin}, G. 2000, \aapr, 10, 1
628: 
629: \bibitem[{{Kunth} \& {Sargent}(1983)}]{kunth83}
630: {Kunth}, D. \& {Sargent}, W.~L.~W. 1983, \apj, 273, 81
631: 
632: \bibitem[{{Kunth} {et~al.}(1981){Kunth}, {Sargent}, \& {Kowal}}]{kunth81}
633: {Kunth}, D., {Sargent}, W.~L.~W., \& {Kowal}, C. 1981, \aaps, 44, 229
634: 
635: \bibitem[{{Kurtz} \& {Mink}(1998)}]{kurtz98}
636: {Kurtz}, M.~J. \& {Mink}, D.~J. 1998, \pasp, 110, 934
637: 
638: \bibitem[{{MacAlpine} {et~al.}(1977){MacAlpine}, {Smith}, \&
639:   {Lewis}}]{macalpine77}
640: {MacAlpine}, G.~M., {Smith}, S.~B., \& {Lewis}, D.~W. 1977, \apjs, 34, 95
641: 
642: \bibitem[{{Margutti} {et~al.}(2007)}]{margutti07}
643: {Margutti}, R. {et~al.} 2007, \aap, preprint astroph/0709.0198
644: 
645: \bibitem[{{Masegosa} {et~al.}(1994){Masegosa}, {Moles}, \&
646:   {Campos-Aguilar}}]{masegosa94}
647: {Masegosa}, J., {Moles}, M., \& {Campos-Aguilar}, A. 1994, \apj, 420, 576
648: 
649: \bibitem[{{Massey} {et~al.}(1988){Massey}, {Strobel}, {Barnes}, \&
650:   {Anderson}}]{massey88}
651: {Massey}, P., {Strobel}, K., {Barnes}, J.~V., \& {Anderson}, E. 1988, \apj,
652:   328, 315
653: 
654: \bibitem[{{McGaugh}(1991)}]{mcgaugh91}
655: {McGaugh}, S.~S. 1991, \apj, 380, 140
656: 
657: \bibitem[{{Osterbrock}(1989)}]{osterbrock89}
658: {Osterbrock}, D.~E. 1989, {Astrophysics of gaseous nebulae and active galactic
659:   nuclei} (Mill Valley, CA: University Science Books)
660: 
661: \bibitem[{{Papaderos} {et~al.}(2006){Papaderos}, {Guseva}, {Izotov}, {Noeske},
662:   {Thuan}, \& {Fricke}}]{papaderos06}
663: {Papaderos}, P., {Guseva}, N.~G., {Izotov}, Y.~I., {Noeske}, K.~G., {Thuan},
664:   T.~X., \& {Fricke}, K.~J. 2006, \aap, 457, 45
665: 
666: \bibitem[{{Prochaska}(2006)}]{prochaska06}
667: {Prochaska}, J.~X. 2006, \apj, 650, 272
668: 
669: \bibitem[{{Prochaska} {et~al.}(2004)}]{prochaska04}
670: {Prochaska}, J.~X. {et~al.} 2004, \apj, 611, 200
671: 
672: \bibitem[{{Rego} {et~al.}(1998){Rego}, {Cordero-Gracia}, {Gallego}, \&
673:   {Zamorano}}]{rego98}
674: {Rego}, M., {Cordero-Gracia}, M., {Gallego}, J., \& {Zamorano}, J. 1998, \aap,
675:   330, 435
676: 
677: \bibitem[{{Richer} \& {McCall}(1995)}]{richer95}
678: {Richer}, M.~G. \& {McCall}, M.~L. 1995, \apj, 445, 642
679: 
680: \bibitem[{{Schaerer} \& {Vacca}(1998)}]{schaerer98}
681: {Schaerer}, D. \& {Vacca}, W.~D. 1998, \apj, 497, 618
682: 
683: \bibitem[{{Schaller} {et~al.}(1992){Schaller}, {Schaerer}, {Meynet}, \&
684:   {Maeder}}]{schaller92}
685: {Schaller}, G., {Schaerer}, D., {Meynet}, G., \& {Maeder}, A. 1992, \aaps, 96,
686:   269
687: 
688: \bibitem[{{Schlegel} {et~al.}(1998){Schlegel}, {Finkbeiner}, \&
689:   {Davis}}]{schlegel98}
690: {Schlegel}, D.~J., {Finkbeiner}, D.~P., \& {Davis}, M. 1998, \apj, 500, 525
691: 
692: \bibitem[{{Searle} \& {Sargent}(1972)}]{searle72}
693: {Searle}, L. \& {Sargent}, W.~L.~W. 1972, \apj, 173, 25
694: 
695: \bibitem[{{Shi} {et~al.}(2005){Shi}, {Kong}, {Li}, \& {Cheng}}]{shi05}
696: {Shi}, F., {Kong}, X., {Li}, C., \& {Cheng}, F.~Z. 2005, \aap, 437, 849
697: 
698: \bibitem[{{Sollerman} {et~al.}(2005){Sollerman}, {{\"O}stlin}, {Fynbo},
699:   {Hjorth}, {Fruchter}, \& {Pedersen}}]{sollerman05}
700: {Sollerman}, J., {{\"O}stlin}, G., {Fynbo}, J.~P.~U., {Hjorth}, J., {Fruchter},
701:   A., \& {Pedersen}, K. 2005, New Astronomy, 11, 103
702: 
703: \bibitem[{{Stanek} {et~al.}(2006)}]{stanek06}
704: {Stanek}, K.~Z. {et~al.} 2006, Acta Astronomica, 56, 333
705: 
706: \bibitem[{{Stasi{\' n}ska}(2005)}]{stasinska05}
707: {Stasi{\' n}ska}, G. 2005, \aap, 434, 507
708: 
709: \bibitem[{{Stasinska}(1980)}]{stasinska80}
710: {Stasinska}, G. 1980, \aap, 84, 320
711: 
712: \bibitem[{{Sutherland} \& {Dopita}(1993)}]{sutherland93}
713: {Sutherland}, R.~S. \& {Dopita}, M.~A. 1993, \apjs, 88, 253
714: 
715: \bibitem[{{Terlevich} {et~al.}(1991){Terlevich}, {Melnick}, {Masegosa},
716:   {Moles}, \& {Copetti}}]{terlevich91}
717: {Terlevich}, R., {Melnick}, J., {Masegosa}, J., {Moles}, M., \& {Copetti},
718:   M.~V.~F. 1991, \aaps, 91, 285
719: 
720: \bibitem[{{Tremonti} {et~al.}(2004)}]{tremonti04}
721: {Tremonti}, C.~A. {et~al.} 2004, \apj, 613, 898
722: 
723: \bibitem[{{Vacca}(1994)}]{vacca94}
724: {Vacca}, W.~D. 1994, \apj, 421, 140
725: 
726: \bibitem[{{Wiersema} {et~al.}(2007)}]{wiersema07}
727: {Wiersema}, K. {et~al.} 2007, \aap, 464, 529
728: 
729: \bibitem[{{Wolf} \& {Podsiadlowski}(2007)}]{wolf07}
730: {Wolf}, C. \& {Podsiadlowski}, P. 2007, \mnras, 375, 1049
731: 
732: \end{thebibliography}
733: 
734: 
735: 
736: % FIGURES
737: 
738: % TABLES
739: 
740: \begin{deluxetable}{lcrl}           % TABLE OF XMPG CANDIDATES
741: \tablecolumns{4} 	\tablewidth{0pt}
742: \tablecaption{XMPG CANDIDATES\label{tab:other}}
743: \tablehead{
744:   \colhead{ID} & \colhead{$g'$} & \colhead{$cz$} & \colhead{Comment} \\
745:   \colhead{} & \colhead{(mag)} & \colhead{(km s$^{-1})$} & \colhead{}
746: }
747: 	\startdata
748: SDSS J100539.35+315441.6 & 18.70 &   ... & quasar?           \\
749: SDSS J120955.68+142155.7 & 20.17 & 23340 & metal poor galaxy \\
750: SDSS J123944.58+145612.8 & 19.77 & 21534 & metal poor galaxy \\
751: SDSS J124638.82+350115.1 & 20.36 & 19522 & metal poor galaxy \\
752: SDSS J124709.24+325118.6 & 19.10 & 27450 & galaxy            \\
753: SDSS J133424.53+592057.0 & 20.55 & 21890 & metal poor galaxy \\
754: SDSS J135641.64+654748.8 & 19.72 & 10742 & galaxy            \\
755: SDSS J140439.28+542136.9 & 19.98 &   318 & H{\sc ii} region  \\
756: SDSS J141333.54+463234.0 & 19.43 &    90 & A-star            \\
757: SDSS J142250.72+514516.5 & 20.22 & 11654 & metal poor galaxy \\
758: SDSS J143345.99-025602.2 & 19.69 &    65 & A-star            \\
759: SDSS J144158.32+291434.2 & 20.13 & 13741 & metal poor galaxy \\
760: SDSS J145621.69+503523.0 & 20.49 &   ... & BL Lac?           \\
761: SDSS J150316.52+111056.9 & 19.41 & 23405 & metal poor galaxy \\
762: SDSS J150535.89+314639.4 & 20.55 & 15817 & galaxy            \\
763: SDSS J151221.08+054911.2 & 20.15 & 24025 & metal poor galaxy \\
764: SDSS J152802.62+240425.6 & 18.65 &    -6 & A-star            \\
765: SDSS J154742.23-005554.2 & 20.57 & 16594 & E+A galaxy        \\
766: SDSS J160238.71+444923.8 & 19.65 & 12350 & galaxy            \\
767: SDSS J165835.08+192415.3 & 18.93 &   -23 & A-star            \\
768: SDSS J172955.61+534338.8 & 19.55 & 24308 & metal poor galaxy \\
769: SDSS J211613.97-000851.3 & 20.39 & 11916 & galaxy            \\
770: SDSS J221912.56+140602.8 & 20.41 &    36 & A-star            \\
771: SDSS J225900.86+141343.5 & 19.09 &  8918 & metal poor galaxy \\
772: 	\enddata
773:  \end{deluxetable}
774: 
775: \begin{deluxetable}{lcccccccccc}
776: \tablecolumns{11}	\tablewidth{0pt}
777: % \rotate
778: \tabletypesize{\scriptsize}
779: \tablecaption{OBSERVED LINE INTENSITIES\label{tab:lineflux}}
780: \tablehead{
781:  \colhead{F($\lambda_0$ Ion)\tablenotemark{a}} &
782:  \colhead{J120955.67} & \colhead{J123944.58} & \colhead{J124638.82} &
783:  \colhead{J133424.53} & \colhead{J142250.72} & \colhead{J144158.32} &
784:  \colhead{J150316.52} & \colhead{J151221.08} & \colhead{J172955.61} & 
785:  \colhead{J225900.86} \\
786:  \colhead{} &
787:  \colhead{+142155.9} & \colhead{+145612.8} & \colhead{+350115.1} &
788:  \colhead{+592057.0} & \colhead{+514516.5} & \colhead{+291434.2} &
789:  \colhead{+111056.9} & \colhead{+054911.2} & \colhead{+534338.8} &
790:  \colhead{+141343.5}
791: }
792: 	\startdata
793: 3727 [O{\sc ii}]  &  8.20 $\pm$ 0.22 &  11.8 $\pm$  0.2 &  16.4 $\pm$  0.4 &  22.5 $\pm$  0.3 &  7.80 $\pm$ 0.14 &  5.70 $\pm$ 0.13 &  43.8 $\pm$  0.5 &  9.00 $\pm$ 0.12 &  29.9 $\pm$  0.5 &  15.1 $\pm$  0.2 \\
794: 3798 H10          &      \nodata     &  0.84 $\pm$ 0.10 &  0.49 $\pm$ 0.06 &  0.62 $\pm$ 0.07 &  0.35 $\pm$ 0.07 &  0.27 $\pm$ 0.05 &  0.72 $\pm$ 0.09 &  0.36 $\pm$ 0.10 &  1.25 $\pm$ 0.11 &  1.20 $\pm$ 0.16 \\
795: 3835 H9           &      \nodata     &  0.29 $\pm$ 0.08 &  0.34 $\pm$ 0.06 &  0.55 $\pm$ 0.07 &  0.91 $\pm$ 0.10 &  0.45 $\pm$ 0.07 &  1.35 $\pm$ 0.12 &  0.35 $\pm$ 0.06 &  1.38 $\pm$ 0.11 &  1.31 $\pm$ 0.15 \\
796: 3868 [Ne{\sc iii}]&  1.75 $\pm$ 0.20 &  3.38 $\pm$ 0.14 &  4.09 $\pm$ 0.13 &  5.62 $\pm$ 0.11 &  6.11 $\pm$ 0.27 &  1.47 $\pm$ 0.12 &  14.5 $\pm$  0.3 &  2.98 $\pm$ 0.23 &  13.5 $\pm$  0.4 &  7.07 $\pm$ 0.12 \\
797: 3889 He {\sc i} + H8    &  0.66 $\pm$ 0.08 &  2.12 $\pm$ 0.17 &  1.86 $\pm$ 0.21 &  2.22 $\pm$ 0.11 &  3.14 $\pm$ 0.11 &  0.84 $\pm$ 0.10 &  5.43 $\pm$ 0.11 &  1.44 $\pm$ 0.11 &  5.09 $\pm$ 0.29 &  5.52 $\pm$ 0.11 \\
798: 3968 [Ne{\sc iii}] + H7 &  1.07 $\pm$ 0.11 &  2.50 $\pm$ 0.11 &  2.39 $\pm$ 0.12 &  3.84 $\pm$ 0.14 &  4.78 $\pm$ 0.28 &  0.89 $\pm$ 0.09 &  8.67 $\pm$ 0.15 &  1.90 $\pm$ 0.11 &  8.21 $\pm$ 0.11 &  6.80 $\pm$ 0.12 \\
799: 4101 H$\delta$    &  1.40 $\pm$ 0.11 &  2.50 $\pm$ 0.12 &  2.40 $\pm$ 0.23 &  3.60 $\pm$ 0.18 &  3.80 $\pm$ 0.11 &  0.90 $\pm$ 0.11 &  8.50 $\pm$ 0.19 &  2.30 $\pm$ 0.13 &  6.90 $\pm$ 0.14 &  7.90 $\pm$ 0.11 \\
800: 4340 H$\gamma$    &  2.70 $\pm$ 0.10 &  4.30 $\pm$ 0.11 &  4.40 $\pm$ 0.10 &  6.10 $\pm$ 0.10 &  8.30 $\pm$ 0.11 &  1.90 $\pm$ 0.11 &  15.9 $\pm$  0.3 &  4.00 $\pm$ 0.11 &  12.4 $\pm$  0.2 &  15.0 $\pm$  1.0 \\
801: 4363 [O{\sc iii}] &  0.30 $\pm$ 0.04 &  1.10 $\pm$ 0.13 &  1.00 $\pm$ 0.11 &  1.10 $\pm$ 0.10 &  2.00 $\pm$ 0.20 &  0.40 $\pm$ 0.07 &  3.50 $\pm$ 0.14 &  0.70 $\pm$ 0.09 &  2.90 $\pm$ 0.19 &  2.90 $\pm$ 0.11 \\
802: 4471 He {\sc i}   &      \nodata     &      \nodata     &  0.36 $\pm$ 0.06 &  0.32 $\pm$ 0.04 &  0.73 $\pm$ 0.09 &  0.29 $\pm$ 0.04 &  1.31 $\pm$ 0.11 &      \nodata     &      \nodata     &  0.66 $\pm$ 0.08 \\
803: 4686 He {\sc ii}  &      \nodata     &      \nodata     &  0.19 $\pm$ 0.05 &  0.23 $\pm$ 0.04 &  0.73 $\pm$ 0.09 &      \nodata     &      \nodata     &  0.37 $\pm$ 0.05 &  0.41 $\pm$ 0.09 &  0.68 $\pm$ 0.08 \\
804: 4861 H$\beta$     &  5.50 $\pm$ 0.11 &  10.4 $\pm$  0.1 &  10.3 $\pm$  0.1 &  14.1 $\pm$  0.1 &  18.3 $\pm$  0.2 &  4.00 $\pm$ 0.10 &  37.9 $\pm$  0.4 &  9.80 $\pm$ 0.15 &  29.2 $\pm$  0.4 &  33.4 $\pm$  0.3 \\
805: 4959 [O{\sc iii}] &  5.50 $\pm$ 0.24 &  15.1 $\pm$  0.4 &  14.7 $\pm$  0.1 &  20.8 $\pm$  0.2 &  21.8 $\pm$  0.2 &  4.50 $\pm$ 0.21 &  62.8 $\pm$  0.6 &  11.9 $\pm$  0.2 &  62.5 $\pm$  0.7 &  33.5 $\pm$  0.3 \\
806: 5007 [O{\sc iii}] &  16.3 $\pm$  0.2 &  45.0 $\pm$  0.5 &  44.6 $\pm$  0.5 &  61.8 $\pm$  0.7 &  65.5 $\pm$  0.7 &  13.6 $\pm$  0.2 & 190.0 $\pm$  1.9 &  35.2 $\pm$  0.4 & 188.0 $\pm$  2.2 & 100.0 $\pm$  1.0 \\
807: 5876 He {\sc i}   &  0.46 $\pm$ 0.06 &  2.60 $\pm$ 0.11 &  1.18 $\pm$ 0.11 &  1.73 $\pm$ 0.12 &  2.08 $\pm$ 0.32 &  0.37 $\pm$ 0.05 &  3.73 $\pm$ 0.12 &  1.03 $\pm$ 0.11 &  3.76 $\pm$ 0.13 &  2.80 $\pm$ 0.13 \\
808: 6563 H$\alpha$    &  14.8 $\pm$  0.2 &  30.7 $\pm$  0.4 &  31.5 $\pm$  0.3 &  40.3 $\pm$  0.7 &  46.2 $\pm$  0.5 &  11.3 $\pm$  0.1 & 102.0 $\pm$  1.2 &  25.8 $\pm$  0.3 &  94.2 $\pm$  1.0 &  91.9 $\pm$  0.9 \\
809: 6584 [N{\sc ii}]  &      \nodata     &      \nodata     &  0.70 $\pm$ 0.08 &  1.30 $\pm$ 0.21 &      \nodata     &  0.20 $\pm$ 0.03 &  1.90 $\pm$ 0.11 &  0.40 $\pm$ 0.05 &  1.40 $\pm$ 0.11 &  0.40 $\pm$ 0.09 \\
810: 6678 He {\sc i}   &      \nodata     &      \nodata     &      \nodata     &  0.39 $\pm$ 0.06 &      \nodata     &      \nodata     &  0.94 $\pm$ 0.25 &      \nodata     &  0.83 $\pm$ 0.09 &  0.42 $\pm$ 0.07 \\
811: 6717 [S{\sc ii}]  &  1.40 $\pm$ 0.26 &  1.00 $\pm$ 0.16 &  1.40 $\pm$ 0.11 &  2.20 $\pm$ 0.11 &  0.70 $\pm$ 0.08 &  0.50 $\pm$ 0.09 &  3.70 $\pm$ 0.12 &  0.90 $\pm$ 0.10 &  3.20 $\pm$ 0.14 &  1.40 $\pm$ 0.14 \\
812: 6731 [S{\sc ii}]  &  0.50 $\pm$ 0.08 &  1.00 $\pm$ 0.15 &  0.90 $\pm$ 0.13 &  1.30 $\pm$ 0.12 &  0.50 $\pm$ 0.09 &  0.30 $\pm$ 0.06 &  3.20 $\pm$ 0.11 &      \nodata     &  2.20 $\pm$ 0.12 &  1.00 $\pm$ 0.11 \\
813: EW(H$\beta$)      &  54.4 $\pm$  3.0 &  91.2 $\pm$  5.0 &  70.7 $\pm$  3.9 &  72.5 $\pm$  4.0 &  95.9 $\pm$  5.3 &  76.7 $\pm$  4.2 &    80 $\pm$  4.4 &  61.6 $\pm$  3.4 &   158 $\pm$  8.7 &   134 $\pm$  7.4 \\
814: 	\enddata
815:  \tablenotetext{a}{Flux in units of $\times10^{-16}$ ergs s$^{-1}$ cm$^{-2}$.  
816: Uncertainties are statistical errors only, and do not include the errors due to 
817: reddening, stellar absorption, and absolute flux calibration.}
818:  \end{deluxetable}
819: 
820: 
821: \begin{deluxetable}{lccccccccrc}           % TABLE OF METAL POOR GALAXIES
822: \tablecolumns{11} 	\tablewidth{0pt}
823: % \rotate
824: \tabletypesize{\scriptsize}
825: \tablecaption{OUR METAL POOR GALAXIES\label{tab:gals}}
826: \tablehead{
827:   \colhead{ID} & \colhead{$cz$} & \colhead{$M_B$} & \colhead{T$_e$} & \colhead{\OH}
828:   & \colhead{Age} & \colhead{H$\alpha$/H$\beta$}
829:   & \colhead{L(H$\beta$)} & \colhead{SFR (T$_e$)} & \colhead{N(OB)}
830:   & \colhead{SN rate} \\
831:   \colhead{} & \colhead{(km s$^{-1})$} & \colhead{(mag)} & \colhead{(K)} & \colhead{}
832:   & \colhead{(Myr)} & \colhead{}
833:   & \colhead{(erg s$^{-1}$)} & \colhead{(M$_{\sun}$ yr$^{-1}$)}
834:   & \colhead{} & \colhead{(10$^{-4}$ yr$^{-1}$)}
835: }
836: 	\startdata
837: SDSS J225900.86+141343.51 &  8918 & -16.4 & 18600 & 7.37 & 4.0 & 2.75 & 1.80$\times10^{40}$ & 0.21 &  13900 &  4 \\
838: SDSS J142250.72+514516.49 & 11654 & -15.9 & 19100 & 7.41 & 4.0 & 2.52 & 8.46$\times10^{39}$ & 0.09 &   6500 &  2 \\
839: SDSS J144158.32+291434.22 & 13741 & -16.3 & 19400 & 7.47 & 4.5 & 2.80 & 1.08$\times10^{40}$ & 0.13 &  10400 &  3 \\
840: SDSS J080840.85+172856.48\tablenotemark{a} & 13233 & -17.1 & 19000 & 7.48 & 4.5 & 2.64 & 1.94$\times10^{40}$ & 0.18 &  18700 &  6 \\
841: SDSS J123944.58+145612.80 & 21534 & -17.7 & 17000 & 7.65 & 4.5 & 2.95 & 3.69$\times10^{40}$ & 0.54 &  37400 & 12 \\
842: SDSS J151221.08+054911.21 & 24025 & -17.6 & 15300 & 7.68 & 5.5 & 2.64 & 2.63$\times10^{40}$ & 0.32 &  40600 & 14 \\
843: SDSS J120955.67+142155.91 & 23340 & -17.4 & 15000 & 7.71 & 5.5 & 2.72 & 2.26$\times10^{40}$ & 0.29 &  34900 & 12 \\
844: SDSS J124638.82+350115.11 & 19522 & -16.8 & 16200 & 7.75 & 5.0 & 3.06 & 1.61$\times10^{40}$ & 0.26 &  24900 &  8 \\
845: SDSS J150316.52+111056.93 & 23405 & -18.3 & 14800 & 7.85 & 4.5 & 2.69 & 6.31$\times10^{40}$ & 0.80 &  65800 & 22 \\
846: SDSS J133424.53+592057.04 & 21890 & -16.8 & 14500 & 7.86 & 5.0 & 2.86 & 1.73$\times10^{40}$ & 0.23 &  27800 &  9 \\
847: SDSS J172955.61+534338.80 & 24308 & -18.3 & 13900 & 8.00 & 4.0 & 3.23 & 9.46$\times10^{40}$ & 1.98 &  60800 & 20 \\
848:         \enddata
849: \tablenotetext{a}{\citet{kewley07}}
850:  \end{deluxetable}
851: 
852: \end{document}
853: