1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[12pt,authoryear]{aastex}
3:
4: % PUNCTUATION SPACING
5: % By default, punctuation [.?!:;,] is followed by extra space EXCEPT
6: % when the punctuation follows an upper case letter. The following
7: % removes the exception, i.e., punctuation will produce extra space
8: % regardless of what character precedes the punctuation. If you
9: % don't want the extra space, follow the offending punctuation mark
10: % with '\ ' or '~'. \frenchspacing and \nonfrenchspacing work as
11: % usual to turn extra spacing off and back on, respectively.
12:
13: \sfcode`A=1000 \sfcode`B=1000 \sfcode`C=1000 \sfcode`D=1000
14: \sfcode`E=1000 \sfcode`F=1000 \sfcode`G=1000 \sfcode`H=1000
15: \sfcode`I=1000 \sfcode`J=1000 \sfcode`K=1000 \sfcode`L=1000
16: \sfcode`M=1000 \sfcode`N=1000 \sfcode`O=1000 \sfcode`P=1000
17: \sfcode`Q=1000 \sfcode`R=1000 \sfcode`S=1000 \sfcode`T=1000
18: \sfcode`U=1000 \sfcode`V=1000 \sfcode`W=1000 \sfcode`X=1000
19: \sfcode`Y=1000 \sfcode`Z=1000
20:
21: % Misc macros
22:
23: %\def\deg{\hbox{deg}}
24: %\def\Mpc{\hbox{Mpc}}
25:
26: \newcommand{\js}{}
27: \newcommand{\etal}{{\js et~al.\/}}
28: \newcommand{\eg}{{\js e.g.\/}}
29: \newcommand{\ie}{{\js i.e.\/}}
30: \newcommand{\Lsol}{$L_{\odot}$}
31: \newcommand{\Msun}{\mbox{$M_\sun$}}
32: \newcommand{\Lsun}{\mbox{$L_\sun$}}
33: \newcommand{\0}{\phantom{0}} % for spacing in tables and such
34: \newcommand{\no}{\nodata}
35:
36: \newcommand{\sst}{{\it Spitzer Space Telescope}}
37: \newcommand{\s}{{\it Spitzer}}
38: \newcommand{\hii}{{\sc h~ii}}
39: %%%%%
40:
41: \slugcomment{v2 \today}
42: \shorttitle{The 8.0~\micron\ Luminosity Function}
43: \shortauthors{Huang et al.}
44: \received{2007 January 24}
45: \begin{document}
46: \title{The Local Galaxy 8\,\micron\ Luminosity Function}
47: \author{J.-S.~Huang,$\!$\altaffilmark{1}
48: M. L. N. Ashby,$\!$\altaffilmark{1}
49: P.~Barmby,\altaffilmark{1}
50: M.~Brodwin,\altaffilmark{2}
51: M.~J.~I.~Brown,\altaffilmark{3}
52: N.~Caldwell,\altaffilmark{1}
53: R.~J.~Cool,\altaffilmark{4}
54: P.~Eisenhardt,\altaffilmark{2}
55: D.~Eisenstein,\altaffilmark{4}
56: G.~G.~Fazio,\altaffilmark{1}
57: E.~Le Floc'h,\altaffilmark{4}
58: P.~Green,\altaffilmark{1}
59: C.~S.~Kochanek,\altaffilmark{6}
60: N.~Y.~Lu,\altaffilmark{5}
61: M.~A.~Pahre,\altaffilmark{1}
62: D.~Rigopoulou,\altaffilmark{7}
63: J.~L.~Rosenberg,\altaffilmark{1,8}
64: H.~A.~Smith,\altaffilmark{1}
65: Z.~Wang,\altaffilmark{1}
66: C.~N.~A. Willmer,\altaffilmark{4}
67: \&
68: S.~P.~Willner\altaffilmark{1}
69: }
70:
71: \altaffiltext{1}{Harvard-Smithsonian Center for Astrophysics, 60 Garden Street,
72: Cambridge, MA 02138}
73: \altaffiltext{2}{Jet Propulsion, California Institute of Technology,
74: Mail Stop 169-506, Pasadena, CA91109}
75: \altaffiltext{3}{School of Physics, Monash
76: University, Clayton 3800, Victoria, Australia}
77: \altaffiltext{4}{Steward Observatory, University of Arizona, 933 N
78: Cherry Ave., Tucson, AZ 85121}
79: \altaffiltext{5}{Spitzer Science Center,
80: Caltech, 1200 E. California, Pasadena, CA 91125}
81: \altaffiltext{6}{Department of Astronomy, The Ohio State University,
82: 140 West 18th Ave., Columbus, OH 43210}
83: \altaffiltext{7}{Department of Astrophysics, Oxford University, Keble
84: Rd, Oxford, OX1 3RH, U.K.}
85: \altaffiltext{8}{Current address: Department of Physics and
86: Astronomy, George Mason University, 4400 University Drive, MS 3F3,
87: Fairfax, VA 22030}
88: \begin{abstract}
89: A \sst\ survey in the NOAO Deep-Wide Field in Bo\"otes provides a
90: complete, 8~\micron-selected sample of galaxies to a limiting (Vega)
91: magnitude of 13.5. In the 6.88~deg$^2$ field sampled, 79\% of the
92: 4867 galaxies have spectroscopic redshifts, allowing an accurate
93: determination of the local ($z<0.3$) galaxy luminosity
94: function. Stellar and dust emission can be separated on the basis of
95: observed galaxy colors. Dust emission (mostly PAH)
96: accounts for 80\% of the 8~\micron\ luminosity, stellar
97: photospheres account for 19\%, and AGN emission accounts for roughly
98: 1\%. A sub-sample of the 8~\micron-selected galaxies have blue,
99: early-type colors, but even most of these have significant PAH
100: emission. The luminosity functions for the total 8~\micron\
101: luminosity and for the dust emission alone are both well fit by
102: Schechter functions. For the 8~\micron\ luminosity function, the
103: characteristic luminosity is $\nu L_{\nu}^*(8.0~\micron) = 1.8 \times
104: 10^{10}$~\Lsun\ while for the dust emission alone it is $1.6 \times
105: 10^{10}$~\Lsun\null. The average 8~\micron\ luminosity density at
106: $z<0.3$ is $3.1 \times 10^7$~\Lsun~Mpc$^{-3}$, and the average luminosity
107: density from dust alone is $2.5 \times 10^7$~\Lsun~Mpc$^{-3}$. This luminosity
108: arises predominantly from galaxies with 8~\micron\ luminosities ($\nu
109: L_{\nu}$) between $2\times 10^9$ and $2\times10^{10}$~\Lsun, i.e.,
110: normal galaxies, not LIRGs or ULIRGs.
111:
112: \end{abstract}
113: \keywords{cosmology: observations --- galaxies:dust emission --- galaxies:
114: Survey --- galaxies: mid-infrared}
115:
116:
117: \section{INTRODUCTION}
118:
119: Galaxy emission in the mid-infrared (3\micron--10\micron) arises from
120: three main mechanisms \citep[\eg][]{Piovan2006}: 1) the
121: Rayleigh-Jeans tail of stellar emission; 2) emission from heated dust
122: in the interstellar medium \citep[ISM,][]{ces96, roe96, ver96,
123: ver01}; and 3) a power law component powered by accreting black holes
124: \citep{moo86, roc91, Elvis1994}. The ISM component displays
125: prominent, broad emission features at 3.3, 6.2, 7.7, 8.6, and
126: 11.3~\micron\ \citep{gil73,Willner1977,Phillips1984} often attributed
127: to polycyclic aromatic hydrocarbons (PAH, \citealt{leg84, all85,
128: pug89}). Elliptical galaxies typically have little or no dust, and
129: therefore their mid-infrared SEDs are expected to be dominated by
130: stellar emission \citep{lu00, lu03, pah04a, pah04b}. Star-forming
131: galaxies, including the extreme Luminous Infrared Galaxies (LIRG,
132: $L_{IR} > 10^{11}$~\Lsun) and Ultra-Luminous Infrared Galaxies
133: (ULIRG, $L_{IR} > 10^{12}$~\Lsun), show the PAH emission
134: features \citep{lu03,pah04a,eng05,gen98,rig99,Armus2007}. Active nuclei can
135: occur in any galaxy type. Galaxy spectral energy distributions
136: (SEDs) are the sum of radiation from all three mechanisms, and
137: because the proportions vary widely, galaxy SEDs in the mid-infrared
138: are very diverse.
139:
140: Mid-infrared dust emission luminosity is a good indicator of a
141: galaxy's star formation rate. It is insensitive to dust obscuration,
142: and it can be measured for large samples of galaxies at both
143: $z\approx0$ and $z\approx2$. \citet[and earlier papers]{Roche1985}
144: showed that mid-infrared feature emission is prominent in star
145: forming galaxies, and ISO spectroscopic studies showed
146: that PAH feature strength increases with other measures of star
147: formation rate \citep[SFR --- \eg][]{Vigroux1999,elb02}.
148: \citet{FS2004} argued for a quantitative
149: relationship between PAH strength and SFR, and \citet{Wu2005}
150: confirmed a relationship using \s\ IRAC 8~\micron\ data to measure
151: the PAH emission.
152:
153: Although PAH strength is a good SFR indicator, it is not perfect.
154: Galaxies with powerful AGN may have very weak or no PAH emission
155: features visible in their SEDs \citep{moo86, roc91, gen98,lau00,
156: ega04, alo05}. This doesn't mean active star formation is absent
157: because, for example, intense ultraviolet radiation produced by
158: the accreting black hole may dissociate PAH \citep{rig99}, especially
159: near the nucleus.
160: PAH features are also weak or absent in the SEDs of galaxies with low
161: metallicities \citep{eng05,Rosenberg2006,Wu2006,Madden2006} or low
162: luminosities \citep{hog05}. At an extreme, \citet{hou04} found that
163: the SED of the blue compact galaxy SBS~0335$-$052 shows no PAH emission
164: at all. One possible explanation is that the formation and existence
165: of PAH depends on metallicity.
166: Whatever the limitations of PAH emission as an SFR indicator, it is
167: still important to know how much PAH emission there is in the local
168: Universe and what kinds of galaxies that emission comes from. The
169: results should at least give the SFR from typical galaxies, if not
170: from unusual galaxy types. Understanding the local population is an
171: essential step in determining how the SFR has evolved with time.
172:
173: The \sst\ offers the opportunity to observe large samples of galaxies
174: at 8~\micron. At low redshift, {\it Spitzer's\/} Infrared Array
175: Camera (IRAC, Fazio et al. 2004) can detect the 6.2, 7.7, and
176: 8.6~\micron\ PAH emission features with its 5.8 and 8.0~\micron\
177: detector arrays, and IRAC's wide field-of-view ($5' \times 5'$) and
178: high sensitivity allow IRAC to map large areas of sky efficiently.
179: Maps at 24~\micron\ made with the Multi-band Imaging Photometer
180: \citep[MIPS ---][]{Rieke2004} are an ideal way of selecting large
181: samples of galaxies with PAH emission at $z \sim 2$ \citep{pap04,
182: yan04, hou05, pap06, web06, Caputi2006}, and many authors
183: \citep[eg][]{hou05, yan05, luz05, Lagache2004} have used the Infrared
184: Spectrograph \citep[IRS ---][]{hou04} to study the PAH emission
185: features in the high-redshift $2 \la z \la 3$ population.
186:
187: This paper focuses on the 7.7 plus 8.6~\micron\ PAH emission features
188: from galaxies at low redshift detected in the IRAC 8.0~\micron\
189: band. By combining redshifts and IRAC photometry, we obtain an IRAC
190: 8.0~\micron-selected local galaxy sample and derive a local PAH
191: luminosity function. The IRAC 8.0~\micron-selected galaxy sample is
192: selected at the same rest wavelength as a MIPS 24~\micron-selected
193: sample at $z=2$ \citep[e.g.,][]{Caputi2006} or an {\it AKARI}
194: 16~\micron-selected sample at $z=1$. A
195: comparison between these populations can thus reveal the effects of
196: evolution between $z=2$ and the present. The paper is
197: arranged as follows: \S 2 describes the IRAC data and the sample
198: selection. \S 3 discusses K-corrections for the IRAC bands and shows
199: the color-luminosity relations of the sample galaxies. The luminosity
200: function is presented in \S 4, and \S 5 summarizes the results.
201: Distances throughout this paper are based on $H_0 =
202: 70$~km~s$^{-1}$~Mpc$^{-1}$, $\Omega_M = 0.3$, $\Omega_\Lambda = 0.7$.
203:
204: \section{OBSERVATIONS AND SAMPLE SELECTION}
205:
206: The IRAC data come from the IRAC Shallow Survey \citep{eis04}, which
207: is one of the major IRAC GTO programs. The data cover 8.5~square
208: degrees in the Bo\"otes field of the NOAO Deep-Wide Field Survey
209: \citep[NDWFS][]{Jannuzi1999}. Each sky location was observed with
210: three 30~s frames at 3.6, 4.5, 5.8, and 8.0~\micron. To create the
211: images in each waveband, the Basic Calibrated Data (BCD) were
212: mosaiced using the Spitzer Science Center software MOPEX
213: \citep{mor05}. An 8~\micron\ catalog was created using SExtractor
214: \citep{ber96}. The 5$\sigma$ limiting magnitude at 8.0~\micron\ is
215: 14.9 in the Vega system, equivalent to a flux density of 69~$\mu$Jy
216: \citep{bro06}. The galaxy surface density at this limit is about
217: 10600 galaxies deg$^{-2}$ \citep{Fazio2004}.
218:
219: Photometry in all four IRAC bands was performed using the SExtractor
220: double-image mode. For each source in the 8~\micron\ catalog, two
221: magnitudes were derived \citep{bro06}: (1) an aperture magnitude
222: within a 5\arcsec\ diameter, corrected to total magnitude using the
223: point spread function (PSF) growth curve for point sources, and (2) a
224: SExtractor ``auto'' magnitude. Figure~\ref{f:mags} shows the
225: difference between these two magnitudes as a function of the aperture
226: magnitude. For objects with angular diameters larger than a few
227: arcseconds, the aperture correction does not fully account for flux
228: outside the aperture, and the aperture magnitude therefore
229: underestimates the source flux. For faint point sources and also for
230: extended sources with low surface brightness, SExtractor chooses too
231: small a size for its ``auto'' aperture, and the auto magnitudes
232: underestimate the flux \citep[e.g., Fig.~5
233: of][]{Labbe2005,Brown2007}. In order to account for these biases, we
234: used the brighter of the two magnitude measurements. For the largest
235: galaxies the auto aperture encircles almost all of the flux and
236: therefore gives total magnitude directly, while for faint galaxies,
237: which are nearly pointlike, the aperture magnitude calibrated for
238: point sources likewise lead to a total magnitude. Flux densities for
239: low surface brightness galaxies may still be underestimated, but
240: these are a small fraction of the sample.\footnote{Less than 5\% of
241: galaxies have surface brightness lower than the equivalent of
242: 13.5~mag in a 5\arcsec\ diameter. This is a strong upper limit on
243: the fraction of galaxies that could be affected.} All magnitudes in
244: this paper are reported in the Vega magnitude system.
245:
246: In addition to the photometric data, the AGN and Galaxy Evolution
247: Survey (AGES, \citealt{koc06}) has measured 15,052 redshifts in the
248: Bo\"otes field with the Hectospec multi-object spectrograph
249: \citep{Fabricant1998, Roll1998, Fabricant2005} on the MMT. Redshifts
250: have been measured for 92\% of galaxies to a magnitude limit of
251: $[8.0]=13.2$ and 65\% of galaxies to $[8.0]=13.8$ in a 6.88~deg$^2$
252: sub-field indicated in Figure~\ref{f:field}. We limited the sample
253: to this subfield and to $[8.0]<13.5$ (251~$\mu$Jy) to minimize the
254: redshift incompleteness and to $z<0.3$ to minimize uncertainties in
255: the K-corrections.\footnote{The magnitude limits for the AGES
256: spectroscopy, $15\le I \le 20$, had no significant effect on the
257: sample selection.} Beyond $z=0.3$, the strongest PAH feature at
258: 7.7~\micron\ shifts out of the IRAC 8.0~\micron\ band. The sample
259: was also limited to $z>0.02$ to avoid statistical noise from small
260: numbers of nearby galaxies.
261: Figure~\ref{f:redshift} shows the redshift completeness in the final
262: sample. The sample limit corresponds to a limiting 8~\micron\
263: absolute magnitude of $-27.7$ at $z=0.3$ (neglecting the K-correction).
264:
265: The 8.0~\micron-selected sample includes normal galaxies, AGN, and
266: stars. Figure~\ref{f:stern} shows the $[5.8]-[8.0]$ color as a
267: function of $[3.6]-[4.5]$ for all of the objects in the sample. At
268: low redshift, AGN
269: and galaxies with PAH emission both have red IRAC colors, but the two
270: classes can be separated using IRAC color criteria
271: \citep{lac04,ste05}. Galactic stars and elliptical galaxies are
272: mixed together at colors near (0,0); morphology was needed to
273: distinguish between them. This sample is nearby enough that
274: elliptical galaxies are detectably extended in the NDWFS data. The
275: number of objects of each type are listed in Table~1 --- 964 objects,
276: 14\% of the photometric sample, are identified as AGN.\footnote{
277: Some galaxies identified as ``normal'' probably have weak or obscured
278: AGN, but these AGN are not detectable in either the IRAC colors or the AGES
279: spectra and are therefore contribute little if anything to the
280: observed light.}
281: However, only 42 of the 549 AGN with a measured redshift are at
282: $z<0.3$ whereas 2556 out of the 3667 normal galaxies are at $z<0.3$.
283: The AGN fraction at $z<0.3$ is therefore about 1.6\% before allowing
284: for incompleteness. Because AGES preferentially observed AGN
285: candidates, the true AGN fraction is likely to be even lower than
286: this. Known AGN were excluded from the sample, but separate results
287: for them are explicitly mentioned in a few places below. As the
288: results show, unknown AGN should have an insignificant effect on the
289: derived luminosity functions.
290:
291:
292: \section{K-CORRECTIONS}
293:
294: K-corrections for the galaxies in this sample are based on a simple
295: two-component model for the SEDs. In this model, the SED for every
296: galaxy is a linear combination of two fixed SEDs: (1) an old
297: ``early-type'' stellar
298: population, as might be found in an elliptical galaxy or spiral
299: bulge, and (2) a mix of stars and interstellar emission as might be
300: found in a ``late-type'' spiral galaxy disk. For each of these
301: components, the SED
302: was determined from an average of the observed SEDs for nearby
303: galaxies that have both 2MASS
304: photometry and ISO spectroscopy \citep{lu03}. Ellipticals were used
305: to define the early-type SED, and disk galaxies were used to
306: define the late-type SED.\footnote{
307: We use ``early-type'' and ``late-type'' as convenient terms to refer to the
308: two SED components despite having no direct knowledge of the actual
309: morphology of individual galaxies. For example, irregular galaxies
310: may have a ``late-type SED'' but not disk morphology. Mid-infrared
311: colors for the SED components are the opposite of those in visible
312: light: late types are red, and early types are blue.}
313: The SED for each of these components was normalized at the rest-frame
314: $K$~band. Eleven model SEDs, ranging from a pure early-type template to a pure
315: late-type template,
316: were created. The late/early ratio for these model SEDs is defined as
317: the ratio of the flux densities of the late-type and early-type components in
318: the rest-frame $K$~band. In addition, a twelfth template matching the M82
319: SED was included to represent galaxies with high
320: star-formation rates.\footnote{M82 has an IR luminosity of
321: $4\times10^{10}$~\Lsol\ \citep{Bell2003} corresponding to a SFR of
322: about 7~\Msun~yr$^{-1}$. Its SED is similar to
323: SEDs of galaxies of even higher luminosity and SFR.}
324: Figure~\ref{f:templates} shows the templates from which the model
325: SEDs were constructed.
326: Convolving each model with the IRAC filter functions gave the
327: K-correction for each of the four IRAC bands as a function of redshift.
328: Figure~\ref{f:k} shows the results. For $z<0.3$, the 3.6 and 4.5~\micron\
329: K-corrections are almost independent of the models, and the
330: 8.0~\micron\ correction is also nearly model-independent unless the galaxy is
331: early-type-dominated. In contrast, the 5.8~\micron\
332: K-correction is highly variable because it depends critically on the
333: strength of the 6.2~\micron\ PAH feature.
334:
335: The suite of SED models spans the observed range of galaxy colors
336: with very few outliers and almost none at $z<0.3$.
337: Figure~\ref{f:cm8} shows the observed $[4.5]-[8.0]$ galaxy colors as
338: a function of redshift. Given the observed color and redshift for
339: each galaxy, the proportion of early- and late-type SED components
340: was determined by interpolation. This proportion defines the
341: rest-frame colors for the galaxy and thus the K-correction for each
342: of the IRAC magnitudes. The few galaxies with $[4.5]-[8.0]$ colors
343: redder than M82 were assigned the M82 K-corrections. AGN (when
344: considered at all) were assigned K-corrections based on a power law
345: SED between 5.8 and 8.0~\micron.
346:
347: \section{RESULTS}
348: \subsection{Mid-Infrared Colors of the Sample Galaxies}
349:
350: The galaxy sample divides into two populations in a color-magnitude
351: diagram. Figure~\ref{f:cm8} shows the 8.0~\micron\ absolute magnitude
352: as a function of rest-frame $[4.5]-[8.0]$ color. Galaxies in the red,
353: late-type population ($[4.5]-[8.0]_{\rm rest} > 1.1$, cf.\
354: Fig.~\ref{f:color}) are star-forming systems with prominent
355: 8~\micron\ PAH emission. According to \citet{pah04b}, this color
356: corresponds to Sa and later morphological types. The blue, early-type
357: population consists of galaxies with weak or no PAH
358: emission corresponding to S0/a and earlier types \citep{pah04b}.
359: Figure~\ref{f:cm36} shows a different color-magnitude diagram, this
360: one based on 3.6~\micron\ magnitude, which is a measure of the stellar
361: luminosity. The early-type population generally has higher stellar
362: luminosity than the late-type population, and galaxies fainter than
363: $[3.6]_{\rm abs}=-24$ are uncommon in the early-type population but common
364: in the late-type
365: one. Within the late-type population, the lack of correlation between the
366: rest-frame $[4.5]-[8.0]$ color and $[3.6]_{\rm abs}$ implies that the
367: 3.6~\micron\ luminosity is little affected by the 3.3~\micron\ PAH
368: emission feature. In contrast, $[8.0]_{\rm abs}$ is strongly
369: correlated with the rest-frame $[3.6]-[4.5]$ color (Fig.~\ref{f:cm8short}),
370: especially for $[8.0]_{\rm abs}<-27$. This correlation suggests that
371: dust emission at 4.5~\micron\ is detectable in the star-forming late-type
372: population, \ie, galaxies with strong PAH emission also show dust
373: continuum emission at 4.5~\micron. This is consistent with ISOPHOT
374: spectroscopy: \citet{lu03} detected 4~\micron\ dust emission from
375: nearby spiral galaxies and suggested it arises from a dust component
376: closely related to the PAH. The same UV sources that heat PAH molecules
377: may also heat the dust that produces the excess 4.5~\micron\
378: emission. Dust emission at 4.5~\micron\ is relatively unimportant
379: for the least luminous galaxies, those with $[8.0]_{\rm abs}>-26$,
380: despite the presence of 8~\micron\ emission in a subset of the
381: low-luminosity galaxies (Fig.~\ref{f:cm8}).
382:
383: The 8~\micron\ number counts are strongly dominated by the late-type
384: population: only 3\% of the spectroscopic sample have early-type colors.
385: Even for the early-type-dominated galaxies, most of the
386: 8~\micron\ emission comes from dust rather than stellar light
387: (Fig.~\ref{f:color}). The late-type galaxies are even more dominant in
388: 8~\micron\ luminosity, as shown in Fig.~\ref{f:cm8}. All the early-type
389: galaxies have 8~\micron\ absolute magnitudes at least three magnitudes fainter
390: than the brightest late-type galaxies.
391:
392: \subsection{The 8.0~\micron\ Luminosity Function}
393:
394: There are many ways to estimate the galaxy luminosity function from
395: survey data. When examining a new wavelength regime, a
396: non-parametric method --- one not assuming a specific functional form
397: --- is desirable because the true luminosity function may not follow
398: a Schechter or other typical distribution. For the present survey,
399: where galaxy distances vary by a factor of ten, it is also better to
400: avoid methods that are sensitive to possible galaxy
401: clustering. \citet{Willmer1997} and \citet{Takeuchi2000} have
402: presented detailed comparisons of many methods. We have computed the
403: luminosity functions with both the familiar $1/V_{\rm max}$ method
404: \citep{Schmidt1968}
405: and the \citet{LB1971} C$^-$~method. The C$^-$~method does not
406: require binning the data, thus avoiding the possible bias in
407: $1/V_{\rm max}$ found by \citet{Page2000}, and it is insensitive to
408: possible galaxy clustering. The results of the two methods are
409: consistent. A further verification that large scale structure has
410: little effect was to ignore data in narrow but heavily populated
411: redshift ranges such as $z\approx0.13$ and $z\approx0.19$
412: (Fig.~\ref{f:color}). The resulting luminosity functions do not
413: change significantly.
414:
415: The luminosity functions have to be corrected for spectroscopic
416: incompleteness in the sample. Because this sample has spectra for
417: nearly all galaxies (Table~1, Fig.~\ref{f:redshift}), the
418: incompleteness correction has an insignificant effect on the results
419: except in the faintest luminosity bins.\footnote{Incompleteness
420: for bright {\em apparent} magnitudes --- mostly galaxies at small
421: distances --- is distributed across all luminosity bins and has
422: little effect on most of them because it affects only small numbers
423: of galaxies in each bin.}
424: This was checked by
425: recomputing the luminosity function with two different incompleteness
426: corrections. One method was to use only galaxies with spectroscopic
427: redshifts but apply a weight to each one according to the reciprocal
428: probability it would have been observed in the spectroscopic sample.
429: The second method included all galaxies. Galaxies were divided into
430: bins based on $[4.5]-[8.0]$ colors and $[8.0]$ magnitudes, and
431: galaxies without redshifts were assigned the same redshift
432: distribution as the galaxies with known redshifts in the same
433: bin.\footnote{In practice, this was achieved by weighting the
434: galaxies with known redshifts according to (total number)/(number
435: with known redshift) where the ratio was calculated for
436: each bin.} The two methods give luminosity functions that are equal
437: within the Poisson uncertainties.
438:
439: The non-parametric 8.0~\micron\ luminosity functions for various
440: galaxy classes are shown in Figure~\ref{f:lfdiff} and Tables~2 and~3. As
441: shown in this figure and also Figure~\ref{f:cm8}, blue
442: galaxies contribute little to the luminosity function and are rare
443: for $[8.0]_{\rm abs}<-26$. AGN are insignificant except in
444: the brightest luminosity bins, where the uncertainties are large
445: because few objects are in the survey. The luminosity function in
446: the range $-29.5 \la [8.0]_{\rm abs} \la -22$ is dominated by
447: galaxies with significant dust contribution to the 8~\micron\ flux.
448:
449: In addition to total luminosity, the observations allow us to measure
450: the 8.0~\micron\ PAH (dust) luminosity alone. For each galaxy in the
451: sample, the rest-frame 3.6~\micron\ flux density was scaled by a
452: factor of 0.227 \citep[an estimate of the stellar flux density at
453: 8~\micron\ ---][]{pah04a} and subtracted from the rest-frame 8.0~\micron\
454: flux density. This subtracts the starlight, leaving the PAH flux
455: density. Because the subtraction is done on {\em rest-frame} flux
456: densities, the procedure is equivalent to the two-component
457: decomposition illustrated in Fig.~\ref{f:color}. After subtracting
458: the continuum, galaxies with residual 8.0~\micron\ magnitudes fainter
459: than the limiting magnitude $[8.0]=13.5$ were excluded from the
460: sample, and the luminosity function was re-calculated for the
461: remaining galaxies to give a PAH-only luminosity function. The
462: result is shown in Figure~\ref{f:lfdiff} and Tables~2 and~3.
463:
464: The uncertainties in the luminosity functions are set both by Poisson
465: statistics of the sample and by cosmic variance. The Poisson
466: uncertainties for the $1/V_{\rm max}$ method are given in Table~2 and
467: statistical uncertainties for
468: the $C^-$ method in Table~3. Cosmic variance uncertainty can
469: be estimated \citep{dh} from the volume sampled plus an estimate of
470: the galaxy two-point correlation function. This is not known
471: directly for the 8~\micron\ sample, but we assume it's the same as
472: for optically-selected galaxy samples. (If anything, the correlation
473: function is likely to be smaller for 8~\micron-selected galaxies,
474: which are predominantly late-type, than for optically-selected ones,
475: more of which are early-type --- \citealt{Norberg2002}. This will
476: lead to less cosmic variance for an 8~\micron-selected sample.) In
477: practice, a correlation function based on the fluctuation power
478: spectrum from WMAP-1 \citep{Spergel2003} was extrapolated to the
479: present via the method of \citet{Seljak1996} then transformed via a
480: spherical Bessel function and smoothed with a 1~Mpc radius. The
481: volume integral \citep[e.g.,][Eq.~1]{Newman2002} was then evaluated
482: via a Monte Carlo approach. The resulting cosmic variance
483: uncertainty is $\approx$15\% in bins $[8.0]_{\rm abs}<-27.7$, where
484: the survey samples the full volume of $z\le0.3$. This is a single
485: uncertainty for the entire luminosity function, not an independent
486: uncertainty in each bin. In the fainter bins, the uncertainty should
487: in principle be larger, but the entire method is based on the
488: assumption that the luminosity function is constant within the volume
489: surveyed. In the faintest bin, for example, the survey magnitude
490: limit corresponds to 135~Mpc, and the data give no information
491: whatsoever on the luminosity function outside this distance. Density
492: peaks in the sample are evident in Figure~\ref{f:color} but do not
493: affect the derived luminosity function.
494:
495: As it turns out, a Schechter function is an excellent fit to both the
496: total 8.0~\micron\ and the PAH luminosity functions, as shown in
497: Figure~\ref{f:lfdiff}. Table~4 gives the best-fit parameters as
498: found using the STY maximum likelihood method \citep{sty79}, and
499: Figure~\ref{f:ellipse} shows the interdependence of the derived
500: characteristic magnitudes $M^*$\footnote{The characteristic absolute magnitude
501: values refer to luminosity $\nu L_\nu$
502: emerging from the galaxy near 8~\micron\
503: measured in units of the Sun's bolometric luminosity.}
504: and faint-end slopes $\alpha$.
505: The derived Schechter exponents for the middle to faint
506: end slopes are $\alpha = -1.38$ and $\alpha = -1.26$
507: respectively. The slopes are similar to those
508: found in the Sloan
509: $r$~band \citep{Loveday2004,Xia2006} and the infrared $K$ band
510: \citep{Huang2003,Jones2006} over a similar redshift range, though
511: \citet{Blanton2003} found a slightly shallower slope in $r$.
512: The steeper slope for the total
513: luminosity shows that fainter galaxies tend to have relative less PAH
514: emission compared to their stellar emission (though the slope
515: uncertainties overlap at the 2$\sigma$ level, as seen in
516: Fig.~\ref{f:ellipse} and Table~4). The inactive ``blue''
517: galaxies have just as steep a faint-end slope ($\alpha=-1.39$) as the
518: entire sample, in contrast to the result in blue light, where
519: inactive galaxies show a much shallower
520: faint-end slope \citep{Madgwick2002}. At the bright end,
521: $L>L^*$, there are no excess galaxy numbers above the Schechter
522: function prediction, in contrast to the results of \citet{sau90} for
523: the IRAS 60~\micron\ luminosity function. The high luminosity
524: population, \ie, LIRGs and ULIRGs, consists of galaxies with
525: far-infrared color temperatures $>$36~K \citep{sau90}, characteristic
526: of Seyfert and starburst galaxies. Our exclusion of AGN in the
527: 8.0~\micron\ sample, which removed galaxies from the bright end of
528: the luminosity function, does not explain why a Schechter function
529: can fit the remainder of the sample.
530:
531: Integrating the luminosity functions over the whole sample gives the
532: total 8~\micron\ luminosity density in the local Universe. The
533: result is $3.1 \times 10^7~$\Lsun~Mpc$^{-3}$ for total 8~\micron\
534: luminosity density (not including AGN), $2.5 \times
535: 10^7~\Lsun$~Mpc$^{-3}$ for PAH luminosity density, and a highly
536: uncertain $0.05 \times 10^7~\Lsun$~Mpc$^{-3}$ for AGN luminosity
537: density. (The sample is too small to determine reliable AGN
538: numbers.) Thus over the entire galaxy population, stellar emission
539: contributes about 19\% of the 8.0~\micron\ luminosity while dust
540: emission contributes 80\% and AGN roughly 1\%. Galaxies with
541: $2\times 10^9~\Lsun <\nu L_{\nu}(8.0\micron) < 2\times10^{10}$~\Lsun\
542: contribute 50\% of the total 8~\micron\ luminosity, and galaxies
543: above and below this range contribute about 25\% each.
544: Thus the dominant contribution to the PAH luminosity density comes from
545: normal galaxies, not LIRGs or ULIRGs.
546: \label{s:pah}
547:
548: \subsection{Star Formation Rates}
549:
550: Star formation rates in the local universe have been measured by a
551: variety of techniques. (See the review by \citealt{Kennicutt1998a}.)
552: \citet{hop04} compiled star formation rate densities (SFRD) from a
553: variety of tracers. According to Hopkins' compilation, over the
554: $z<0.3$ range of the present survey, the volume-weighted
555: average ${\rm SFRD} \approx 0.031$~\Msun~yr$^{-1}$~Mpc$^{-3}$, though
556: some tracers tend to be systematically higher (emission lines) or lower
557: (ultraviolet) than this value.
558: \citet{Martin2005} combined ultraviolet with
559: far infrared data and found a local ($z<0.04$) ${\rm SFRD} \approx
560: 0.021$~\Msun~yr$^{-1}$~Mpc$^{-3}$, about 20\% higher than but
561: consistent with Hopkins' result at this much smaller redshift.
562:
563: Dust emission at 8.0~\micron\ is correlated with the total infrared
564: luminosity $L_{IR}$ in galaxies \citep{rig99,elb02,lu03} although the
565: relation is non-linear \citep{lu03} and has large scatter \citep[\eg,
566: Fig.~17 of][]{smith07}. In normal star-forming galaxies, the
567: integrated flux between 5.8 and 11.3~\micron\ is between 9\% (for
568: galaxies \citeauthor{lu03} call ``FIR-quiescent'') and 5\% (for
569: \citeauthor{lu03} ``FIR-active'' galaxies) of $L_{IR}$ \citep{lu03}.
570: Recent results from \s\ are consistent with these values;
571: \citet{smith07} find that in the central areas of normal galaxies,
572: the 7.7~\micron\ PAH complex can contribute up to 10\% of $L_{IR}$.
573: In ULIRGs, PAH emission in the 5.8--11.3~\micron\ range is lower at
574: $\la$1\% of $L_{IR}$
575: \citep{rig99,Armus2007}; this could be partly due to an AGN contributing to
576: $L_{IR}$. Despite the uncertainties, some authors
577: \citep[e.g.,][]{FS2004,Wu2005} have suggested that $\nu L_\nu ({\rm
578: PAH})$ can be used to measure star formation rates for galaxies, and
579: \citet{Wu2005} derived ${\rm SFR} \approx \nu L_\nu ({\rm PAH}) /
580: (1.5 \times 10^9~\Lsun)$~\Msun~yr$^{-1}$. This result was based on a
581: sample of Sloan galaxies showing H$\alpha$ emission. The PAH
582: luminosity density found in \S\ref{s:pah} combined with the
583: local SFRD \citep[Fig.~2]{hop04} integrated from $z=0$ to 0.3
584: gives a volume average ${\rm SFR} \approx
585: \nu L_\nu ({\rm PAH}) / (0.84 \times 10^9~\Lsun)$~\Msun~yr$^{-1}$,
586: about double that found by \citeauthor{Wu2005} for individual
587: galaxies. In other words, the volume-averaged PAH luminosity density
588: is half as much as expected, given the volume-averaged
589: SFRD and ${\rm SFRD} / \nu L_\nu({\rm PAH})$ ratios of individual galaxies.
590: The difference suggests that the galaxy samples have been weighted to
591: galaxies with unusually large PAH emission for their SFRs, but this
592: is hard to understand if the dominant contribution to star formation
593: comes from the normal galaxy population \citep[e.g.,][]{flo05}.
594:
595:
596: \section{SUMMARY}
597:
598: An 8.0~\micron-selected sample of low redshift ($z<0.3$) galaxies
599: finds primarily star-forming galaxies. AGN represent less than 2\%
600: of the sample both by number and by fraction of the
601: 8~\micron\ luminosity. The luminosity function of the remaining
602: galaxies can be described as a Schechter function with
603: characteristic magnitude $M^*=-28.46$
604: (corresponding to $\nu L_{\nu} = 1.8 \times
605: 10^{10}$~\Lsun) and $\alpha=-1.38$.
606: Subtracting starlight continuum emission from each galaxy gives a
607: luminosity function for emission from the interstellar
608: medium component (mostly PAH) alone. This can also be approximated
609: by a Schechter
610: function with characteristic magnitude $M^*=-28.34$ ($\nu L_{\nu} = 1.6
611: \times 10^{10}$~\Lsun) and $\alpha=-1.26$.
612:
613: The total 8.0~\micron\ luminosity density for this 8.0~\micron-selected
614: sample is $3.1 \times 10^7$~\Lsun~Mpc$^{-3}$. Ignoring the luminosity
615: contributed by AGN, about 81\% of the 8~\micron\ luminosity is from
616: interstellar dust emission, presumably PAH, and about 19\% is from
617: stellar photospheres. The observed, volume-averaged ratio of PAH
618: luminosity density to star formation rate density is about half of
619: an estimate based on local, individual galaxies.
620:
621:
622:
623: \acknowledgements
624:
625: This work is based in part on observations made with the Spitzer
626: Space Telescope, which is operated by the Jet Propulsion Laboratory,
627: California Institute of Technology under a contract with
628: NASA. Support for this work was provided by NASA through an award
629: issued by JPL/Caltech. JLR has received support from an NSF
630: Astronomy and Astrophysics Postdoctoral Fellowship under grant
631: AST-0302049. Observations reported here were obtained at the MMT
632: Observatory, a joint facility of the Smithsonian Institution and the
633: University of Arizona. This work made use of images and/or data
634: products provided by the NOAO Deep Wide-Field Survey (Jannuzi and Dey
635: 1999), which is supported by the National Optical Astronomy
636: Observatory (NOAO). NOAO is operated by AURA, Inc., under a
637: cooperative agreement with the National Science Foundation.
638:
639: Facilities: \facility{Spitzer(IRAC),MMT(Hectospec)}
640:
641:
642:
643: \begin{thebibliography}{}
644: \frenchspacing
645:
646: \bibitem[Allamandola et al.(1985)]{all85}Allamandola, L. J., Tielens,
647: A. G. G. M., \& Barker, J. R. 1985, \apj, 290, L25
648:
649: \bibitem[Allamandola et al.(1989)]{all89}Allamandola, L. J., Tielens,
650: A. G. G. M., \& Barker, J. R. 1989, \apjs, 71, 733
651:
652: \bibitem[Alonso-Herrero et al.(2004)]{alo04}Alonso-Herrero, A., et
653: al. 2004, \apjs, 154, 155
654:
655: \bibitem[Alonso-Herrero et al.(2005)]{alo05}
656: Alonso-Herrero, A., et al.\ 2006, \apj, 640, 167
657:
658: \bibitem[Armus et al.(2007)]{Armus2007}
659: Armus, L., et al.\ 2007, \apj, 656, 148
660:
661: \bibitem[Bell(2003)]{Bell2003}
662: Bell, E.~F.\ 2003, \apj, 586, 794
663:
664: \bibitem[Bertin \& Arnouts(1996)]{ber96}Bertin, E. \& Arnouts,
665: S. 1996, \aap,117,393
666:
667: \bibitem[Blanton et al.(2003)]{Blanton2003}
668: Blanton, M.~R., et al.\ 2003, \apj, 592, 819
669:
670: \bibitem[Brodwin et al.(2007)]{bro06}Brodwin, M., et al. 2007, \apj,
671: submitted
672:
673: \bibitem[Brown et al.(2007)]{Brown2007}
674: Brown, M.~J.~I., Dey, A., Jannuzi, B.~T., Brand, K., Benson, A.~J.,
675: Brodwin, M., Croton, D.~J., \& Eisenhardt, P.~R.\ 2007, \apj, 654, 858
676:
677: \bibitem[Caputi et al.(2006)]{Caputi2006}
678: Caputi, K.~I., et al.\ 2006, \apj, 637, 727
679:
680: \bibitem[C\'esarsky et al.(1996)]{ces96}C\'esarsky, D., Lequeux, J.,
681: Abergel, A., Perault, M., Palazzi, E., Madden, S., \& Tran, D. 1996,
682: \aap, 315, L309
683:
684: \bibitem[Crawford, Marr, Partridge, \& Strauss(1996)]{Crawford1996}
685: Crawford, T., Marr, J., Partridge, B., \& Strauss, M.~A.\ 1996, \apj, 460,
686: 225
687:
688: \bibitem[Dale et al.(2003)]{dal03}Dale, D., et al. 2003, \aj, 120,
689: 583
690:
691: \bibitem[Davis \& Huchra(1982)]{dh}
692: Davis, M. \& Huchra, J. 1982, \apj, 254, 437.
693:
694: \bibitem[Egami et al.(2004)]{ega04} Egami, E., et al. 2004, \apjs,
695: 154, 130
696:
697: \bibitem[Eisenhardt et al.(2004)]{eis04} Eisenhardt, P., et al. 2004,
698: \apjs, 154, 48
699:
700: \bibitem[Elbaz et al.(2002)]{elb02}
701: Elbaz, D., Cesarsky, C.~J., Chanial, P., Aussel, H., Franceschini, A.,
702: Fadda, D., \& Chary, R.~R.\ 2002, \aap, 384, 848
703:
704: \bibitem[Elvis et al.(1994)]{Elvis1994} Elvis, M., et al.\ 1994,
705: \apjs, 95, 1
706:
707: \bibitem[Engelbracht et al.(2005)]{eng05} Engelbracht, C. W., et
708: al. 2005, \apj, 628, L29
709:
710: \bibitem[Fabricant et al.(1998)]{Fabricant1998}
711: Fabricant, D.~G., Hertz, E.~N., Szentgyorgyi, A.~H., Fata, R.~G., Roll,
712: J.~B., \& Zajac, J.~M.\ 1998, \procspie, 3355, 285
713:
714: \bibitem[Fabricant et al.(2005)]{Fabricant2005}
715: Fabricant, D., et al.\ 2005, \pasp, 117, 1411
716:
717: \bibitem[Fazio et al.(2004a)]{faz04}Fazio, G.~G., et al. 2004, \apjs,
718: 154, 10
719:
720: \bibitem[Fazio et al.(2004b)]{Fazio2004}
721: Fazio, G.~G., et al.\ 2004, \apjs, 154, 39
722:
723: \bibitem[Flores et al.(1999)]{flo99}Flores, H., et al. 1999, \apj, 517,
724: 148
725:
726: \bibitem[F{\"o}rster Schreiber et al.(2004)]{FS2004}
727: F{\"o}rster Schreiber, N.~M., et al.\ 2004, \apj, 616, 40
728:
729: \bibitem[Genzel et al.(1998)]{gen98}Genzel, R., et al. 1998, \apj,
730: 498, 579
731:
732: \bibitem[Gillett et al.(1973)]{gil73}Gillett, F. C., Forrest, W. J.,
733: \& Merrill, K. 1973, \apj, 183, 87
734:
735: \bibitem[Haarsma et al.(2000)]{haa00}
736: Haarsma, D.~B., Partridge, R.~B., Windhorst, R.~A., \& Richards, E.~A.\
737: 2000, \apj, 544, 641
738:
739: \bibitem[Hammer et al.(1997)]{ham97} Hammer, F., et al. 1997, \apj,
740: 481, 49
741:
742: \bibitem[Hogg et al.(1998)]{hog98}
743: Hogg, D.~W., Cohen, J.~G., Blandford, R., \& Pahre, M.~A.\ 1998, \apj, 504,
744: 622
745:
746: \bibitem[Hogg et al.(2005)]{hog05}Hogg, D., et al. 2005, \apj, 624,
747: 162
748:
749: \bibitem[Hopkins(2004)]{hop04}Hopkins, A. M. 2004,
750: \apj, 615, 209
751:
752: \bibitem[Houck et al.(2004)]{hou04}Houck, J.~R., et~al. 2004,
753: \apjs, 154, 211
754:
755:
756: \bibitem[Houck et al.(2005)]{hou05}Houck, J.~R., et~al. 2005,
757: \apj, 622, L105
758:
759: \bibitem[Huang, Glazebrook, Cowie, \& Tinney(2003)]{Huang2003}
760: Huang, J.-S., Glazebrook, K., Cowie, L.~L., \& Tinney, C.\ 2003, \apj, 584,
761: 203
762:
763: \bibitem[Jannuzi \& Dey(1999)]{Jannuzi1999}
764: Jannuzi, B.~T., \& Dey, A.\ 1999, ASP Conf.~Ser.~191: Photometric Redshifts
765: and the Detection of High Redshift Galaxies, 191, 111
766:
767: \bibitem[Jones, Peterson, Colless, \& Saunders(2006)]{Jones2006}
768: Jones, D.~H., Peterson, B.~A., Colless, M., \& Saunders, W.\ 2006, \mnras,
769: 369, 25
770:
771: \bibitem[Kennicutt(1998a)]{Kennicutt1998a}
772: Kennicutt, R.~C., Jr.\ 1998, \araa, 36, 189
773:
774: \bibitem[Kennicutt(1998b)]{Kennicutt1998}
775: Kennicutt, R.~C.\ 1998, \apj, 498, 541
776:
777: \bibitem[Kochanek et al.(2006)]{koc06}Kochanek, C., et~al.\ 2007, in
778: preparation
779:
780: \bibitem[Labb{\'e} et al.(2005)]{Labbe2005}
781: Labb{\'e}, I., et al.\ 2005, \apjl, 624, L81
782:
783: \bibitem[Lacy et al.(2004)]{lac04}
784: Lacy, M., et~al.\ 2004, \apjs, 154, 166
785:
786: \bibitem[Lagache et al.(2004)]{Lagache2004}
787: Lagache, G., et al.\ 2004, \apjs, 154, 112
788:
789: \bibitem[Laurent et al.(2000)]{lau00}
790: Laurent, O., Mirabel, I.~F., Charmandaris, V., Gallais, P., Madden, S.~C.,
791: Sauvage, M., Vigroux, L., \& Cesarsky, C.\ 2000, \aap, 359, 887
792:
793: \bibitem[Le Floc'h et al.(2005)]{flo05}Le Floc'h, E., et al. 2005,
794: \apj, 632, 169
795:
796: \bibitem[L\'eger \& Puget(1984)]{leg84}L\'eger, A., \& Puget,
797: J. L. 1984, \aap, 137, L5
798:
799: \bibitem[Lilly et al.(1996)]{lil96}
800: Lilly, S.~J., Le Fevre, O., Hammer, F., \& Crampton, D.\ 1996, \apjl, 460,
801: L1
802:
803: \bibitem[Loveday(2004)]{Loveday2004}
804: Loveday, J.\ 2004, \mnras, 347, 601.
805:
806: \bibitem[Lu et al.(2000)]{lu00}Lu, N., \& Hur, M. 2000, BAAS, 196,
807: 2702
808:
809: \bibitem[Lu et al.(2003)]{lu03}
810: Lu, N., et al.\ 2003, \apj, 588, 199
811:
812: \bibitem[Lutz et al.(2005)]{luz05}
813: Lutz, D., Valiante, E., Sturm, E., Genzel, R., Tacconi, L.~J., Lehnert,
814: M.~D., Sternberg, A., \& Baker, A.~J.\ 2005, \apjl, 625, L83
815:
816: \bibitem[Lynden-Bell(1971)]{LB1971}
817: Lynden-Bell, D.\ 1971, \mnras, 155, 95
818:
819: \bibitem[Madden, Galliano, Jones, \& Sauvage(2006)]{Madden2006}
820: Madden, S.~C., Galliano, F., Jones, A.~P., \& Sauvage, M.\ 2006, \aap, 446,
821: 877
822:
823: \bibitem[Madgwick, Hewett, Mortlock, \& Lahav(2002)]{Madgwick2002}
824: Madgwick, D.~S., Hewett, P.~C., Mortlock, D.~J., \& Lahav, O.\ 2002,
825: \mnras, 334, 209
826:
827: \bibitem[Makovoz \& Khan(2005)]{mor05}
828: Makovoz, D., \& Khan, I.\ 2005, ASP Conf.~Ser.~347: Astronomical Data
829: Analysis Software and Systems XIV, 347, 81
830:
831: \bibitem[Martin et al.(2005)]{Martin2005}
832: Martin, D.~C., et al.\ 2005, \apjl, 619, L59
833:
834: \bibitem[Moorwood(1986)]{moo86}Moorwood, A.\ F\. M\. 1986, \aap, 166, 4
835:
836: \bibitem[Newman \& Davis(2002)]{Newman2002}
837: Newman, J.~A.\ \& Davis, M.\ 2002, \apj, 564, 567
838:
839: \bibitem[Norberg et al.(2002)]{Norberg2002}
840: Norberg, P., et al.\ 2002, \mnras, 332, 827
841:
842: \bibitem[Page \& Carrera(2000)]{Page2000}
843: Page, M.~J.\ \& Carrera, F.~J.\ 2000, \mnras, 311, 433
844:
845: \bibitem[Pahre et al.(2004a)]{pah04a}Pahre, M. A., Ashby, M. L. N.,
846: Fazio, G. G., \& Willner, S. P. 2004, \apjs, 154, 229
847:
848: \bibitem[Pahre et al.(2004b)]{pah04b}Pahre, M. A., Ashby, M. L. N.,
849: Fazio, G. G., \& Willner, S. P. 2004, \apjs, 154, 235
850:
851: \bibitem[Papovich et al.(2004)]{pap04}Papovich, C., et al. 2004,
852: \apjs, 154, 70
853:
854: \bibitem[Papovich et al.(2006)]{pap06}
855: Papovich, C., et al.\ 2006, \apj, 640, 92
856:
857: \bibitem[Phillips, Aitken, \& Roche(1984)]{Phillips1984}
858: Phillips, M.~M., Aitken, D.~K., \& Roche, P.~F.\ 1984, \mnras, 207,
859: 25
860:
861: \bibitem[Piovan, Tantalo, \& Chiosi(2006)]{Piovan2006}
862: Piovan, L., Tantalo, R., \& Chiosi, C.\ 2006, \mnras, 366, 923
863:
864: \bibitem[Puget \& L\'eger(1989)]{pug89}Puget, J. L. \& L\'eger,
865: A. 1989, ARA\&A, 27, 161
866:
867: \bibitem[Rieke et al.(2004)]{Rieke2004}
868: Rieke, G.~H., et al.\ 2004, \apjs, 154, 25
869:
870: \bibitem[Rigopoulou et al.(1999)]{rig99}
871: Rigopoulou, D., Spoon, H.~W.~W., Genzel, R., Lutz, D., Moorwood, A.~F.~M.,
872: \& Tran, Q.~D.\ 1999, \aj, 118, 2625
873:
874: \bibitem[Roelfsema et al.(1996)]{roe96}
875: Roelfsema, P.~R., et al.\ 1996, \aap, 315, L289
876:
877: \bibitem[Roche \& Aitken(1985)]{Roche1985}
878: Roche, P.~F., \& Aitken, D.~K.\ 1985, \mnras, 213, 789
879:
880: \bibitem[Roche et al.(1991)]{roc91}Roche, P. F., Aitken, D. K.,
881: Smith, C. H., \& Ward, M. J. 1991, \mnras, 248, 606
882:
883: \bibitem[Roll, Fabricant, \& McLeod(1998)]{Roll1998}
884: Roll, J.~B., Fabricant, D.~G., \& McLeod, B.~A.\ 1998, \procspie,
885: 3355, 324
886:
887: \bibitem[Rosenberg, Ashby, Salzer, \& Huang(2006)]{Rosenberg2006}
888: Rosenberg, J.~L., Ashby, M.~L.~N., Salzer, J.~J., \& Huang, J.-S.\ 2006,
889: \apj, 636, 742
890:
891: \bibitem[Rush, Malkan, \& Spinoglio(1993)]{Rush1993}
892: Rush, B., Malkan, M.~A., \& Spinoglio, L.\ 1993, \apjs, 89, 1
893:
894: \bibitem[Sandage, Tammann, \& Yahil(1979)]{sty79}
895: Sandage, A., Tammann, G.~A., \& Yahil, A.\ 1979, \apj, 232, 352
896:
897: \bibitem[Sanders \& Mirabel(1996)]{san96}Sanders, D. \& Mirabel,
898: I. F. 1996, ARA\&A, 24, 749
899:
900: \bibitem[Saunders et al.(1990)]{sau90}
901: Saunders, W., Rowan-Robinson, M., Lawrence, A., Efstathiou, G., Kaiser, N.,
902: Ellis, R.~S., \& Frenk, C.~S.\ 1990, \mnras, 242, 318
903:
904: \bibitem[Schiminovich et al.(2005)]{Schiminovich2005}
905: Schiminovich, D., et al.\ 2005, \apjl, 619, L47
906:
907: \bibitem[Schmidt(1968)]{Schmidt1968}
908: Schmidt, M. 1968, \apj, 151, 393
909:
910: \bibitem[Seljak \& Zaldarriaga(1996)]{Seljak1996}
911: Seljak, U., \& Zaldarriaga, M.\ 1996, \apj, 469, 437
912:
913: \bibitem[Smith \etal(2007)]{smith07}
914: Smith, J.~D.~T., et al.\ 2007, \apj, 656, 770
915:
916: \bibitem[Soifer et al.(1987)]{soi87}
917: Soifer, B.~T., Sanders, D.~B., Madore, B.~F., Neugebauer, G., Danielson,
918: G.~E., Elias, J.~H., Lonsdale, C.~J., \& Rice, W.~L.\ 1987, \apj, 320, 238
919:
920: \bibitem[Spergel et al.(2003)]{Spergel2003}
921: Spergel, D.~N., et al.\ 2003, \apjs, 148, 175
922:
923: \bibitem[Stern et al.(2005)]{ste05}
924: Stern, D., et al.\ 2005, \apj, 631, 163
925:
926: \bibitem[Sullivan et al.(2000)]{sul00}
927: Sullivan, M., Treyer, M.~A., Ellis, R.~S., Bridges, T.~J., Milliard, B., \&
928: Donas, J.\ 2000, \mnras, 312, 442
929:
930: \bibitem[Takeuchi, Yoshikawa, \& Ishii(2000)]{Takeuchi2000}
931: Takeuchi, T.~T., Yoshikawa, K., \& Ishii, T.~T.\ 2000, \apjs, 129, 1
932:
933: \bibitem[Vermeij et al.(2002)]{ver01}Vermeij, R., Peeters, E.,
934: Tielens, A. G. G. M., \& van der Hulst, J. M. 2002, \aap, 382, 1042
935:
936: \bibitem[Verstraete et al.(1996)]{ver96}Verstraete, L., Puget, J. L.,
937: Falgarone, E., Drapatz, S., Wright, C. M., \& Timmermann, R. 1996,
938: \aap, 315, L337
939:
940: \bibitem[Vigroux et al.(1999)]{Vigroux1999}
941: Vigroux, L., et al.\ 1999, ESA SP-427: The Universe as Seen by ISO, 805
942:
943: \bibitem[Webb et al.(2006)]{web06}
944: Webb, T.~M.~A., et al.\ 2006, \apjl, 636, L17
945:
946: \bibitem[Willmer(1997)]{Willmer1997}
947: Willmer, C.~N.~A.\ 1997, \aj, 114, 898
948:
949: \bibitem[Willner et al.(1977)]{Willner1977}
950: Willner, S.~P., Soifer, B.~T., Russell, R.~W., Joyce, R.~R., \& Gillett,
951: F.~C.\ 1977, \apjl, 217, L121
952:
953: \bibitem[Wilson et al.(2002)]{wil02}Wilson, G., et al. 2002, \aj,
954: 124, 1258
955:
956: \bibitem[Wu et al.(2005)]{Wu2005}
957: Wu, H., Cao, C., Hao, C.-N., Liu, F.-S., Wang, J.-L., Xia, X.-Y., Deng,
958: Z.-G., \& Young, C.~K.-S.\ 2005, \apjl, 632, L79
959:
960: \bibitem[Wu et al.(2006)]{Wu2006}
961: Wu, Y., Charmandaris, V., Hao, L., Brandl, B.~R., Bernard-Salas, J., Spoon,
962: H.~W.~W., \& Houck, J.~R.\ 2006, \apj, 639, 157
963:
964: \bibitem[Xia et al.(2006)]{Xia2006}
965: Xia, L., Zhou, X., Yang, Y., Ma, J., \& Jiang, Z.\ 2006, \apj, 652, 249
966:
967: \bibitem[Xu(2000)]{Xu2000}
968: Xu, C.\ 2000, \apj, 541, 134
969:
970: \bibitem[Yan et al.(2004)]{yan04}Yan, L., et al.\ 2004, \apjs, 154, 75
971:
972: \bibitem[Yan et al.(2005)]{yan05}Yan, L., et al.\ 2005, \apj, 628, 604
973:
974: \bibitem[Yun, Reddy, \& Condon(2001)]{Yun2001}
975: Yun, M.~S., Reddy, N.~A., \& Condon, J.~J.\ 2001, \apj, 554, 803
976:
977: \end{thebibliography}
978:
979:
980:
981:
982: \clearpage
983: \begin{figure}
984: %\plotone{f1.ps}
985: \caption{Difference between the SExtracter total (``auto'')
986: magnitudes and 5\arcsec\ aperture magnitudes. Offsets are mostly
987: positive, indicating that aperture magnitudes underestimate the total
988: flux of resolved galaxies. We selected the sample using the aperture
989: magnitudes to avoid bias against lower surface brightness galaxies,
990: and set the magnitude limit at $[8.0]=13.5$. The derived luminosity
991: functions are based on the auto magnitudes for the resolved galaxies
992: and the aperture magnitudes for the unresolved galaxies.
993: \label{f:mags}}
994: \end{figure}
995:
996: \clearpage
997: \begin{figure}
998: %
999:
1000: %\plotone{f2.ps}
1001: \caption{The B\"ootes field showing the approximate area of 8~\micron\
1002: coverage outlined in red. Positions are J2000. Black dots denote
1003: denote the 8.0~\micron-detected
1004: galaxies that have AGES spectroscopic redshifts. The black lines enclose
1005: the area of the highly complete redshift sample used for this paper.
1006: \label{f:field}}
1007: \end{figure}
1008:
1009: \clearpage
1010: \begin{figure}
1011: \plotone{f3.eps}
1012: \caption{Redshift identification fraction in each
1013: half-magnitude bin (filled circles, left ordinate label). Error bars
1014: show the Poisson
1015: uncertainty. The solid line shows the histogram of the 8~\micron\
1016: magnitude distribution of all galaxies in the catalog (right ordinate
1017: label). Only galaxies with $[8.0]<13.5$ are in the final sample.
1018: \label{f:redshift}}
1019: \end{figure}
1020:
1021: \clearpage
1022: \begin{figure}
1023: \begin{center}
1024: %\includegraphics[scale=0.85]{f4.ps}
1025: \end{center}
1026: \caption{$[5.8]-[8.0]$ vs $[3.6]-[4.5]$ color-color diagram for
1027: separation of AGN, star-forming galaxies, and early type galaxies
1028: \citep{ste05}. Data are plotted as observed, \ie, before
1029: K-correction. Black dots denote normal galaxies, red dots denote
1030: stars, and blue dots denote AGN, all as classified by optical
1031: spectroscopy. The two thick lines show two SED models from $0<z<0.6$,
1032: one with and the other without PAH emission. (The latter is almost
1033: hidden behind the data near zero color.) The thin solid lines show
1034: the \citet{ste05} AGN-galaxy separation criterion; AGN lie within the
1035: quadrilateral extending to the top of the plot. The color cut at
1036: $[5.8]-[8.0]=0.5$
1037: separates stars and elliptical galaxies from galaxies with
1038: significant PAH emission. Galaxies with PAH emission cannot enter
1039: the ``no PAH'' area unless their redshifts are $>$0.6.
1040: \label{f:stern}}
1041: \end{figure}
1042:
1043: \clearpage
1044: \begin{figure}
1045: \plotone{f5.ps}
1046: \caption{Galaxy SED templates. The upper curve shows the M82
1047: template, the middle curve the spiral (``late-type'') template, and
1048: the bottom curve the elliptical (``early-type'') template. The
1049: vertical axis is in arbitrary units. The early-type template is an
1050: average ISOPHOT-S spectrum of six classical elliptical galaxies
1051: compiled by \citet{lu00}. It is very similar to the elliptical
1052: spectrum shown by \citet{lu03} except at wavelengths longer than
1053: 8~\micron, where an improved ISOPHOT-S calibration made the current
1054: spectrum slightly steeper (i.e., bluer). The late-type template is
1055: the average for spiral galaxies from \citet{lu00}. Nine additional
1056: templates used in this paper are mixtures of the early- and late-type
1057: templates.
1058: \label{f:templates}}
1059: \end{figure}
1060:
1061: \clearpage
1062: \begin{figure}
1063: \plotone{f6.ps}
1064: \vspace*{-48pt}
1065: \caption{K-corrections for the four IRAC channels for each of the
1066: twelve SED models. The bottom line in each panel shows the
1067: early-type model. The 3.6 and 4.5~\micron\ bands sample
1068: mainly stellar emission longward of the Wien peak and therefore show
1069: a negative K-correction. The 5.8~\micron\ band is relatively narrow
1070: and shows a positive K-correction, depending on the amount of PAH
1071: present, as the 6.2~\micron\ PAH feature leaves the band at
1072: $z\approx0.05$. As redshift increases further, the band samples only
1073: stellar emission, and the K-correction decreases. The 8~\micron\
1074: band is wide, and the 7.7~\micron\ feature remains inside the band at
1075: low redshift. As redshift increases, the 6.2~\micron\ PAH feature
1076: enters the band before the 7.7~\micron\ feature leaves, leading to a
1077: nearly constant K-correction. At $z\ga0.2$, the 7.7~\micron\ feature
1078: begins to leave the band, and the K-correction increases according to
1079: the PAH strength.
1080: \label{f:k}}
1081: \end{figure}
1082:
1083: \clearpage
1084: \begin{figure}
1085: \plotone{f7.ps}
1086: \caption{
1087: Color-magnitude diagram showing $[4.5]-[8.0]$ color versus 8~\micron\
1088: absolute magnitudes. Points denote galaxies with spectroscopic
1089: redshifts. Colors and magnitudes are both corrected to the rest
1090: frame. Most galaxies show strong PAH emission ($[4.5]-[8.0]>2$) and
1091: are therefore in the ``late-type'' category as we have defined
1092: it. ``Early-type'' galaxies with weak or no PAH emission
1093: ($[4.5]-[8.0]\la0.4$) are well separated in this diagram from the
1094: much larger late-type population, but galaxies span the entire color range.
1095: None of the early-type galaxies has high luminosity at
1096: 8~\micron.
1097: \label{f:cm8}}
1098: \end{figure}
1099:
1100: \clearpage
1101: \begin{figure}
1102: \plotone{f8.ps}
1103: \caption{\footnotesize The $[4.5]-[8.0]$ color versus redshift for
1104: galaxies with spectroscopic redshifts. Points denote the observed
1105: galaxies with AGN excluded. Lines show colors for each of twelve
1106: model SEDs as a function of redshift. The reddest model is derived
1107: from the M~82 SED. Other models are a linear combination of late-type
1108: and early-type templates as shown in Fig.~\ref{f:templates}. These
1109: SEDs are normalized at rest 2.2~\micron\ with the
1110: stellar fraction going from 0 to 1 in steps of 0.1 from the bluest
1111: model to the reddest one (second-reddest including the M82 model).
1112: Only galaxies with $z<0.3$ are used in this paper.
1113: \label{f:color}}
1114: \end{figure}
1115:
1116: \clearpage
1117: \begin{figure}
1118: \plotone{f9.ps}
1119: \caption{Color-magnitude diagram showing $[4.5]-[8.0]$ color vs
1120: 3.6~\micron\ absolute magnitudes. Points denote galaxies with
1121: spectroscopic redshifts. Colors and magnitudes are both corrected
1122: to the rest frame. The upper luminosity envelope is populated by
1123: both blue ($[4.5]-[8.0]<1.1$) and star-forming galaxies, but there
1124: are no blue galaxies with low 3.6~\micron\ luminosities.
1125: For the strongly star-forming galaxies, $[4.5]-[8.0] \ga 2$, color
1126: does not correlate with the absolute 3.6~\micron\ magnitude.
1127: \label{f:cm36}}
1128: \end{figure}
1129:
1130: \clearpage
1131: \begin{figure}
1132: \plotone{f10.eps}
1133: \caption{Color-magnitude diagram showing $[3.6]-[4.5]$ color vs
1134: 8~\micron\ absolute magnitudes. Points denote galaxies with
1135: spectroscopic redshifts. Colors and magnitudes are both corrected
1136: to the rest frame. The correlation between color and magnitude
1137: indicates that dust emission may constitute about 30\% of the total
1138: 4.5\micron\ emission for galaxies with $[8.0]_{\rm abs} \la -27$.
1139: \label{f:cm8short}}
1140: \end{figure}
1141:
1142: \clearpage
1143: \begin{figure}
1144: \begin{center}
1145: \includegraphics[height=5.5in, trim=108 108 0 54]{f11.eps}
1146: \end{center}
1147: \caption{\footnotesize
1148: Derived differential luminosity functions corrected for
1149: spectroscopic incompleteness. Black dots denote the 8~\micron\
1150: luminosity function for all galaxies excluding AGN, and red squares
1151: denote the luminosity function for PAH alone, \ie, after
1152: subtracting the stellar continuum emission for each galaxy. These
1153: two luminosity functions were derived from the non-parametric C$^-$
1154: method. Blue triangles
1155: denote the luminosity function
1156: for early-type galaxies, specifically those with
1157: $[4.5]-[8.0]<1.1$. Even for these galaxies, much of the emission
1158: is PAH despite the color selection. Green crosses denote the
1159: luminosity function for AGN. These latter two luminosity functions
1160: were derived from the $1/V_{max}$ method because of the smaller
1161: numbers of galaxies. Because of the small numbers, the blue galaxy
1162: luminosity function is only reliable for $-27\la [8.0]_{abs} \la
1163: -25$. The AGN luminosity function is highly uncertain throughout;
1164: only one bin has as many as 8 galaxies in it. All points are
1165: slightly offset horizontally for clarity; see Tables~2 and~3 for actual
1166: magnitude bins. Error bars show the Poisson uncertainties in each
1167: bin. Lines show the respective parametric Schechter functions that best fit
1168: the data for ``all 8~\micron\ emission'' and ``PAH only.'' Both were derived
1169: from the STY method. The bottom axis is labeled in 8~\micron\ Vega
1170: magnitudes, while the top axis is labeled in the equivalent $\nu
1171: L_\nu$ units.
1172: \label{f:lfdiff}}
1173: \end{figure}
1174:
1175: \clearpage
1176: \begin{figure}
1177: \includegraphics[angle=-90]{f12.ps}
1178: \caption{Uncertainty ellipses for characteristic magnitude $M^*$ and
1179: faint end slope $\alpha$ of the Schechter luminosity functions. The
1180: ellipses were calculated using a grid in $M^*$--$\alpha$ space and
1181: the STY formalism. The curves represent contours of constant $\Delta
1182: \chi^2$ corresponding to 1$\sigma$ (inner), 2$\sigma$ (middle), and
1183: 3$\sigma$ (outer) confidence levels assuming three fitting parameters.
1184: Statistical uncertainty in the overall normalization $\Phi^*$ is very
1185: small, but the true uncertainty is set by cosmic variance (\S4.2).
1186: Faint end slopes probably differ but are consistent at the 2$\sigma$ level.
1187: \label{f:ellipse}}
1188: \end{figure}
1189:
1190:
1191: \clearpage
1192:
1193: \begin{table*}[hbt]
1194: {%\scriptsize
1195: \begin{center}
1196: \centerline{\sc Table 1}
1197: \vspace{0.1cm}
1198: \centerline{\sc Number of Objects at $[8.0]<13.5$}
1199: \vspace{0.3cm}
1200: \begin{tabular}{lccc}
1201: \hline\hline
1202: \noalign{\smallskip}
1203: Object & Total & Redshift Sample & $z<0.3$\cr
1204: \hline
1205: \noalign{\smallskip}
1206: Star & 1894 & 0 & 0\cr
1207: Blue Galaxies & 236 & 109 & 100\cr
1208: Star-Forming Galaxies & 3667 & 3174 & 2556\cr
1209: AGN & 964 & 549 & 42\cr
1210: \noalign{\hrule}
1211: \noalign{\smallskip}
1212: \end{tabular}
1213: \end{center}
1214: }
1215: \label{tab1}
1216: \end{table*}
1217:
1218:
1219: \clearpage
1220: \begin{table*}[hbt]
1221: {%\scriptsize
1222: \begin{center}
1223: \centerline{\sc Table 2}
1224: \vspace{0.1cm}
1225: \centerline{\sc Luminosity Functions Calculated by the $1/V_{max}$ Method }
1226: \vspace{0.3cm}
1227: \begin{tabular}{ccrccrcc}
1228: \hline\hline
1229: \noalign{\smallskip}
1230: [8.0]$_{abs}$\tablenotemark{a}
1231: & $\nu L_\nu(8~\micron)$
1232: & $N$
1233: & $\phi$(8.0\micron)
1234: & $\Delta\phi$(8.0\micron)
1235: & $N$
1236: & $\phi$(PAH)
1237: & $\Delta\phi$(PAH)\\
1238:
1239: \hline
1240: \noalign{\smallskip}
1241: &&&\multicolumn{2}{c}{$10^{-3}$~Mpc$^{-3}$~mag$^{-1}$}&&
1242: \multicolumn{2}{c}{$10^{-3}$~Mpc$^{-3}$~mag$^{-1}$}\\
1243: -30.27& $9.5\times10^{10}$ & 2&\00.003 & \00.002& 2&\00.003 &\00.002 \\
1244: -29.77& $6.0\times10^{10}$ & 10&\00.018 & \00.006& 9&\00.016 &\00.005 \\
1245: -29.27& $3.8\times10^{10}$ & 37&\00.073 & \00.012& 31&\00.062 &\00.011 \\
1246: -28.77& $2.4\times10^{10}$ &133& 0.25 & 0.02 &116& 0.22 & 0.02 \\
1247: -28.27& $1.5\times10^{10} $&294& 0.56 & 0.03 &268& 0.50 & 0.03 \\
1248: -27.77& $9.5\times10^{9\0}$&447& 1.06 & 0.05 &384& 0.86 & 0.04 \\
1249: -27.27& $6.0\times10^{9\0}$&429& 1.45 & 0.07 &416& 1.34 & 0.07 \\
1250: -26.77& $3.8\times10^{9\0}$&413& 2.07 & 0.10 &367& 1.75 & 0.09 \\
1251: -26.27& $2.4\times10^{9\0}$&327& 2.49 & 0.14 &270& 1.94 & 0.12 \\
1252: -25.77& $1.5\times10^{9\0}$&218& 2.71 & 0.18 &179& 2.04 & 0.15 \\
1253: -25.27& $9.5\times10^{8\0}$&137& 3.2\0 & 0.3\0 & 98& 2.02 & 0.20 \\
1254: -24.77& $6.0\times10^{8\0}$& 63& 2.5\0 & 0.3\0 & 64& 2.17 & 0.27 \\
1255: -24.27& $3.8\times10^{8\0}$& 41& 3.0\0 & 0.5\0 & 27& 1.73 & 0.33 \\
1256: -23.77& $2.4\times10^{8\0}$& 23& 3.7\0 & 0.8\0 & 13& 1.60 & 0.44 \\
1257: -23.27& $1.5\times10^{8\0}$& 9& 2.4\0 & 0.8\0 & 16& 3.0\0 & 0.8\0 \\
1258: -22.77& $9.5\times10^{7\0}$& 12& 7.4\0 & 2.1\0 & 7& 3.2\0 & 1.2\0 \\
1259: -22.27& $6.0\times10^{7\0}$& 7& 7.7\0 & 2.9\0 & 6& 5.8\0 & 2.4\0 \\
1260: -21.77& $3.8\times10^{7\0}$& 1& 2.4\0 & 2.4\0 &\no& \no & \no \\
1261: -21.27& $2.4\times10^{7\0}$& 2& 10.1\0\0& 7.6\0 &\no& \no & \no \\
1262: \noalign{\hrule}
1263: \noalign{\smallskip}
1264: \end{tabular}
1265: \end{center}
1266: }
1267: \label{tab2}
1268: \tablenotetext{a}
1269: {Absolute Vega magnitude bin centers.}
1270: \end{table*}
1271:
1272: \clearpage
1273: \begin{table*}[hbt]
1274: {%\scriptsize
1275: \begin{center}
1276: \centerline{\sc Table 3}
1277: \vspace{0.1cm}
1278: \centerline{\sc Luminosity Functions Calculated by the C$^-$ Method }
1279: \vspace{0.3cm}
1280: \begin{tabular}{cccccc}
1281: \hline\hline
1282: \noalign{\smallskip}
1283: [8.0]$_{abs}$\tablenotemark{a}
1284: & $\nu L_\nu(8~\micron)$
1285: & $\phi$(8.0\micron)
1286: & $\Delta\phi$(8.0\micron)
1287: & $\phi$(PAH)
1288: & $\Delta\phi$(PAH)\\
1289:
1290: \hline
1291: \noalign{\smallskip}
1292: &&\multicolumn{2}{c}{$10^{-3}$~Mpc$^{-3}$~mag$^{-1}$}&
1293: \multicolumn{2}{c}{$10^{-3}$~Mpc$^{-3}$~mag$^{-1}$}\\
1294: -30.13& $8.5\times10^{10}$ &\00.006 & \00.002&\00.006 &\00.002 \\
1295: -29.63& $5.3\times10^{10}$ &\00.049 & \00.007&\00.048 &\00.007 \\
1296: -29.13& $3.4\times10^{10}$ &\00.121 & \00.010&\00.112 &\00.009 \\
1297: -28.63& $2.1\times10^{10}$ & 0.37 & 0.02 & 0.35 & 0.02 \\
1298: -28.13& $1.5\times10^{10} $& 0.67 & 0.02 & 0.63 & 0.02 \\
1299: -27.63& $8.5\times10^{9\0}$& 0.89 & 0.02 & 0.76 & 0.02 \\
1300: -27.13& $5.3\times10^{9\0}$& 0.90 & 0.02 & 0.80 & 0.02 \\
1301: -26.63& $3.4\times10^{9\0}$& 1.24 & 0.03 & 1.06 & 0.03 \\
1302: -26.13& $2.1\times10^{9\0}$& 1.82 & 0.06 & 1.54 & 0.05 \\
1303: -25.63& $1.5\times10^{9\0}$& 1.92 & 0.08 & 1.53 & 0.07 \\
1304: -25.13& $8.5\times10^{8\0}$& 3.1\0 & 0.2\0 & 1.92 & 0.12 \\
1305: -24.63& $5.3\times10^{8\0}$& 2.8\0 & 0.2\0 & 2.10 & 0.17 \\
1306: -24.13& $3.4\times10^{8\0}$& 3.3\0 & 0.4\0 & 1.91 & 0.27 \\
1307: -23.63& $2.1\times10^{8\0}$& 4.5\0 & 0.6\0 & 4.1\0 & 0.5\0 \\
1308: -23.13& $1.5\times10^{8\0}$& 3.4\0 & 0.6\0 & 2.8\0 & 0.5\0 \\
1309: -22.63& $8.5\times10^{7\0}$& 5.0\0 & 0.9\0 & 3.6\0 & 0.8\0 \\
1310: -22.13& $5.3\times10^{7\0}$& 7.3\0 & 1.5\0 & 6.6\0 & 1.7\0 \\
1311: -21.63& $3.4\times10^{7\0}$& 11.7\0\0& 4.5\0 & \no & \no \\
1312: -21.13& $2.1\times10^{7\0}$& 7.2\0 & 3.9\0 & \no & \no \\
1313: \noalign{\hrule}
1314: \noalign{\smallskip}
1315: \end{tabular}
1316: \end{center}
1317: }
1318: \label{tab3}
1319: \tablenotetext{a} {Approximate bin centers in absolute Vega
1320: magnitudes. The actual first bin center is $-30.15$ for the ``8.0''
1321: columns and $-30.11$ for the ``PAH'' columns. Successive bin centers
1322: differ by half-magnitude intervals.}
1323: \end{table*}
1324:
1325: \clearpage
1326: \begin{table*}[hbt]
1327: {%\scriptsize
1328: \begin{center}
1329: \centerline{\sc Table 4}
1330: \vspace{0.1cm}
1331: \centerline{\sc Schechter Function Parameters}
1332: \vspace{0.3cm}
1333: \begin{tabular}{ccc}
1334: \hline\hline
1335: \noalign{\smallskip}
1336: & all 8~\micron\ emission & PAH only \\
1337: \hline
1338: \noalign{\smallskip}
1339: $M^*(8~\micron)$ & $-28.46\pm0.08$ & $-28.34\pm{0.08}$ \\
1340: $L^*(8~\micron)$ & $1.8\times10^{10}$ & $1.6\times10^{10}$~\Lsun \\
1341: $\Phi^*$~\tablenotemark{a} & $1.10\times10^{-3}$ &
1342: $1.17\times10^{-3}$ \\
1343: $\alpha$ & $-1.38\pm0.04$ & $-1.26\pm0.04$ \\
1344: Integrated luminosity\tablenotemark{b} & $3.1\times10^7$
1345: & $2.5\times10^7$~\Lsun~Mpc$^{-3}$\\
1346: \noalign{\hrule}
1347: \noalign{\smallskip}
1348: \end{tabular}
1349: \end{center}
1350: }
1351: \label{tab4}
1352: \tablenotetext{a}
1353: {Fiducial galaxy density in galaxies~Mpc$^{-3}$~mag$^{-1}$.
1354: Statistical uncertainty is very small, but the real uncertainty is
1355: about 15\% set by cosmic variance.}
1356: \tablenotetext{b}
1357: {Integral of $\phi(L) dL$, where $L \equiv \nu L_\nu$ for $\nu$
1358: corresponding to 8~\micron.}
1359: \end{table*}
1360:
1361: \end{document}
1362: