SDSS AGNs with X-ray Emission from ROSAT PSPC Pointed Observations
A. A. Suchkov 1 , R. J. Hanisch 2 , W. Voges 3 , and T. M. Heckman
ABSTRACT
We present a sample of 1744 of Type 1 AGNs from the Sloan Digital Sky Survey
Data Release 4 (SDSS DR4) spectroscopic catalog with X-ray counterparts in the White-
Giommi-Angelini catalog (WGACAT) of ROSAT PSPC pointed observations. Of 1744
X-ray sources, 1410 (80.9%) are new AGN identifications. Of 4574 SDSS DR4 AGNs
for which we found radio matches in the catalog of radio sources from the Faint Images
of the Radio Sky at Twenty cm Survey (FIRST), 224 turned up in our sample of SDSS
X-ray AGNs. The sample objects are given in a catalog that contains optical and X-
ray parameters and supporting data, including redshifts; it also contains radio emission
parameters where available. We illustrate the content of our catalog and its potential for
AGN science by providing statistical relationships for the catalog data. The potential
of the morphological information is emphasized by confronting the statistics of optically
resolved, mostly low-redshift AGNs with unresolved AGNs that occupy a much wider
redshift range. The immediate properties of the catalog objects include significant
correlation of X-ray and optical fluxes, which is consistent with expectations. Also
expected is the decrease of X-ray flux toward higher redshifts. The X-ray-to-optical flux
ratio for the unresolved AGNs exhibits a decline toward higher redshifts, in agreement
with previous results. The resolved AGNs, however, display the opposite trend. The
X-ray hardness ratio shows a downward trend with increasing low-energy X-ray flux
and no obvious dependence on redshift. At a given optical brightness, X-ray fluxes of
radio- loud AGNs are on average higher than those of radio-quiet AGNs by a factor of 2.
We caution, however, that because of the variety of selection effects present in both the
WGACAT and the SDSS, the interpretation of any relationships based on our sample
of X-ray AGNs requires a careful analysis of these effects.
Subject headings: catalogs — galaxies: active — galaxies: starbursts — quasars: general —
radio continuum: galaxies — X-ray: galaxies
department of Physics and Astronomy, Johns Hopkins University, 3400, North Charles Street, Baltimore, MD
21218.
2 Space Telescope Science Institute, operated by AURA Inc., under contract with NASA, 3700 San Martin Dr.,
Baltimore, MD 21218
3 Max-Planck-Institut fur extraterrestrische Physik, 85748 Garching, Germany
- 2 -
1. Introduction
Active galactic nuclei (AGNs) 1 produce X-ray emission as the gaseous material of a galaxy is
accreted onto the massive black hole in its center. This emission provides powerful diagnostics to
probe the inner working of the AGN central engines, and X-ray studies of AGNs at different redshifts
have always been of fundamental importance for understanding the physics, origin, and evolution
of these objects. X-ray data become especially informative when combined with optical data and
redshift information, and there is much effort to build and analyze such samples. Thus, Fiore et al.
(2003), presented optical identifications from the HELLAS2XMM survey for 122 sources detected
with XMM-Newton in the surveyed area of 0.9 square degree. Green et al. (2004) found 335 unique
optical counterparts for X-ray sources in the fields included in Chandra Multiwavelength Project
(ChaMP), which is a survey of serendipitous Chandra X-ray sources in a ~ 14 square degree field
(Kim et al. 2004a). Half of the 125 counterparts that Green et al. (2004) classified spectroscopically
proved to be broad-line AGNs and 40% turned out to be galaxies with narrow emission lines or
absorption-line galaxies.
Along with designated optical follow-ups, much work was done to identify X-ray counterparts
for AGNs detected in various optical and infrared surveys. Some of these surveys, such as that
presented by Wolf et al. (2004) for the Chandra Deep Field South, specifically targeted the fields
observed in X rays. Others serendipitously overlap with the fields from various X-ray missions.
Particularly important for AGN studies is the Sloan Digital Sky Survey, which not only increased
the number of known AGNs by an order of magnitude but also furnished redshifts and spectra.
Many thousands of AGNs were spatially resolved, so a large amount of morphological information
on underlying galaxies became available, too. A catalog of ~ 1200 SDSS AGNs from the Early
Data Release with X-ray counterparts from the ROSAT All Sky Survey (RASS; Voges et al.,
1999; 2000) was presented by Anderson et al. (2003); this number nearly tripled for SDSS DR3
AGNs (Shen et al. 2005). The online system ClassX, developed by McGlynn et al. (2004; see
also http://heasarc.gsfc.nasa.gov/classx), uses Virtual Observatory protocols 2 to retrieve optical
information from other major surveys, such as 2MASS and USNO Bl, and provides object type
classifications using artificial intelligence algorithms. Most of the previously unidentified ROSAT
X-ray sources are classified by ClassX as AGNs. Yet most of the X-ray sources detected by ROSAT,
Chandra, XMM-Newton, and other large X-ray missions lack optical identifications.
There is a wide range of issues that can be addressed with samples of AGNs that have both
X-ray and optical data, including redshifts (for recent review see, e. g., Brandt Sz Hasinger 2005).
Hasinger, Miyaji, & Schmidt (2005) compiled a sample of ~ 1000 X-ray sources with energies of
0.5-2 keV from ROSAT, Chandra, and XMM-Newton observations that were optically identified
1 Both pointlike QSOs and resolved galaxies with active nuclei are referred to in this paper as AGNs.
2 In this work we have made use of data obtained from the US National Virtual Observatory, which is sponsored
by the National Science Foundation
-3-
with AGNs. With that sample they derived for the first time reliable space densities of X-ray
selected AGNs at cosmological distances. They quantified in much detail previous findings on the
strong luminosity-dependent cosmological evolution of these objects, confirming the density peaks
at z ~ 2 and z < 1 near the high and low limits, respectively, of the AGN X-ray luminosity range
( \ogL x ~ 45 and 42).
Barger et al. (2005) looked into cosmological evolution of AGNs using a few hundred optically
identified hard Chandra X-ray sources with redshifts from spectroscopy or, in the case of too
faint objects, estimated from optical photometry. The study found a decrease in the AGN X-ray
luminosity by an order of magnitude since redshift z = 1.2. That decrease was shown to be due
to AGN downsizing rather than to declining accretion rates onto the central black holes. Another
result from this study suggests that the widely discussed unification model, according to which
the broad-line (Seyfert 1) and narrow-line (Seyfert 2) AGNs differ only due to the line-of-sight
effects (different inclinations of the black hole accreting disk) needs to be modified by including
luminosity-dependent effects.
Brand et al. (2005) used a sample of ~3000 Chandra ACIS sources with redshifts from optical
photometry to study the accretion history of active nuclei. The sources were identified with red
galaxies in the Bootes area. Over the redshift range z = 0.3 — 0.9 the accuracy of photometric red-
shifts sufficed to show that the mean X-ray luminosity increases toward higher redshifts, interpreted
as evidence for a substantial decline in the nuclear accretion since the epoch z ~ 1.
A rapidly growing body of evidence suggests the existence of a connection between the star-
burst and AGN phenomena (e.g., Kauffmann et al. 2003; Imanishi & Wada 2004). This fact is of
fundamental significance, because it implies that the same physics underlies processes leading to
the central burst of star formation and the AGN phenomenon, e.g., formation of a central molec-
ular torus feeding both the starburst and the central black hole. Both phenomena are associated
with a prodigious output of X rays and thus can be studied through the properties of the galaxy
X-ray emission. Deep insights into intimate details of the AGN-starburst connections, hence the
aforementioned underlying physics, can be gained from a study of combined X-ray and optical
properties of a large sample of galaxies. If redshifts are available, the evolution of these connections
can be tracked on a cosmological timescale. Recent efforts along these lines, which are based on the
Chandra X-ray Observatory data in the first place, have already brought significant results. Horn-
schemeier et al. (2005) cross-correlated the Chandra sources from the archive of ACIS observations
with SDSS galaxies. The resulting sample of 42 galaxies with redshifts within z = 0.03 — 0.25
was used to compare optical spectroscopic diagnostics of galaxy activity, both star formation and
nuclear accretion. Although the small size of the sample limited the scope of issues that could be
addressed with it, a number of substantial results were obtained. In particular, all X-ray-luminous,
X-ray-hard galaxies were found to have AGN spectroscopic signatures. This supports earlier claims
that simple observational effects may explain the puzzle of the class of X-ray-bright, optically
normal galaxies (XBONGs) seen in the Chandra and XMM-Newton observations of galaxies at
moderate redshifts. But the sample used is rather small, and statistically larger samples of both
-4 -
normal and AGN X-ray galaxies are still needed in order to either firmly rule out or confirm high
X-ray luminosities that are due to mechanisms other than an AGN.
Optical spectra and redshifts will continue to be the main challenge in achieving the full science
potential of modern X-ray missions such as Chandra and XMM-Newton. This is also true for the
ROSAT data, although the situation here is in some sense much better. The reason is that the
ROSAT data can be supported to a much larger degree by the rich SDSS optical data. Anderson
et al. (2003) emphasized that the ROSAT All Sky Survey and the SDSS survey are well matched
in many respects, most importantly in terms of sensitivity. The sensitivity match is even better
for the PSPC pointed observations, where a typical flux limit of logf x ~ —13.2 samples very well
the peak in the SDSS AGN brightness distribution at the AGN sample i-band limiting magnitude
of ~ 19 (see the logf x — i relationship below). Therefore, although less deep and having lower
positional accuracy than the Chandra and XMM-Newton, and sampling only the soft (< 2.5 keV)
X-ray band, the ROSAT catalogs remain indispensable for many scientific issues of current interest.
There are a few thousand ROSAT sources, mainly from the shallow All-Sky Survey, identified with
resolved SDSS AGNs and/or starburst galaxies in the low-redshift Universe, z < 0.6. The number
counts in the similarly soft samples of the Chandra objects at these redshifts are typically limited
to less than a hundred (see, e.g., Green et al. 2004; Barger et al. 2005; Hornschemeier et al.
2005). Prior to the SDSS barely more than a thousand of ROSAT sources were identified with
optically known AGNs (Brinkmann, Yuan, & Siebert 1997; Yuan, et al. 1998). The initial catalog
by Anderson et al. (2003) of SDSS AGNs with X-ray emission from the RASS contains ~ 1200
objects; this number was increased to 3366 by Shen et al. (2005) using SDSS DR3 AGNs.
In this paper we present a catalog of 1744 AGNs from SDSS DR4 with X-ray emission from
ROSAT PSPC pointed observations, which are substantially deeper than those in the RASS. Half
of these AGNs are in the low-redshift Universe, z < 0.6, similar to the AGNs in the Andersen et
al. (2003) catalog. A significant fraction, ~ 20%, are resolved SDSS objects, so their morphology
can be studied from SDSS images. Statistically large samples of such AGNs will be especially
useful to study AGN-starburst connections, the role of galaxy mergers in the AGN phenomenon,
the relationships between the galaxy bulges and the central black holes (e.g., Heckman et al. 2004),
and other problems for which the knowledge of the details of galaxy morphology is essential. The
sample's usefulness for some other problems is limited, however, by selection effects present in both
the WGACAT and SDSS. For instance, both the WGACAT sources and the SDSS QSO (AGN)
sample are biased against obscured AGNs. Soft X-ray emission that can be detected with ROSAT
is absorbed in such objects, and instruments operating in harder X-ray bands, such as Chandra
and XMM-Newton, are needed to observe them. Obscuration also depresses AGN spectral features
in the optical. Because of that the SDSS QSO (AGN) sample misses the population of heavily
obscured AGNs. Due to these effects the catalog we present is essentially limited to Type 1 AGNs.
The above selection effects are not the only ones, and in this paper we discuss more of them.
- 5 -
2. The catalog
2.1. The WGACAT data
In the 4 yr period of ROSAT pointed observations with the PSPC (Position Sensitive Pro-
portional Counter), ~ 18% of the sky was covered, more or less randomly, to various degrees of
sensitivity. The WGACAT was created by reprocessing data from these observations (see, e.g.,
White et al. 2002 at http://wgacat.gfsc.nasa.gov). The exposure times were typically 100 times
longer than in the 6 month ROSAT All-Sky Survey program. The source detection algorithm was
optimized for point sources, although sources in which extended emission is present are also found.
The catalog provides count rates in the 0.1 — 2.4 keV band and three narrow bands along with sup-
porting information for 76,763 unique sources detected in the PSPC field of view within 60 arcmin
from the detector axis. Some sources were observed repeatedly; the WGACAT contains separate
entries for each individual observation, so there are a total of 88,579 catalog entries. Since both
the sensitivity and positional accuracy of ROSAT PSPC decrease with increasing offset from the
detector axis, we use in our studies only the data obtained at the smallest off-axis angles; if the
source off-axis angles in different exposures are the same then the data from the largest exposure
time is used.
Less than 15% of the sources, mostly the brightest ones, were identified in the WGACAT with
stars, galaxies, AGNs, and other objects as a result of positional cross-correlation against 92 catalogs
of optical, radio, X-ray, and infrared sources. The sample of identified sources was employed by
McGlynn et al. (2004) to build the automated classifier ClassX that determines the likely object
type for any WGACAT source. The vast majority of the previously unidentified sources were
classified by ClassX as AGNs (object classes QSO and AGN in the classification scheme discussed
by McGlynn et al. 2004). Many of them are expected to be X-ray counterparts of SDSS AGNs in
the fields where the two surveys overlap, so optical identification for these sources can be derived
from positional cross-correlation of SDSS AGNs with the WGACAT sources.
Because of the design of the ROSAT PSPC instrument and the way the ROSAT PSPC pointed
observations program was carried out, the WGACAT represents a data set that is in general neither
flux limited nor complete to any given flux level. One of the reasons is that the exposure times at
different pointings were not the same, so the detection limit varies between different pointings. But
even at the same pointing the sources detected at larger offsets from the detector axis are skewed
toward higher fluxes due to decline in the detector sensitivity with increasing off-axis distance.
The respective selection effects would transfer to any sample of objects with the WGACAT X-
ray, impacting the sample's statistics. Therefore, interpretation of the relationships obtained from
such samples would require a careful analysis of these effects. We addressed them at some level
while illustrating below the properties of the SDSS X-ray AGNs presented in this paper. However,
depending on the scientific problem at hand, a deeper analysis would typically be needed.
-6-
2.2. Cross-correlation of SDSS DR4 AGNs with the WGACAT sources
The sample of SDSS AGNs cross-correlated with WGACAT includes all 57,800 objects from
Data Release 4 (Abazajian et al. 2005) that have SDSS spectral type 3, which corresponds to
AGNs, and have measured brightness in all five bands. Because of the AGN selection algorithm
used in SDSS, almost all of them are Type 1 AGNs, i.e., those with broad emission lines (QSOs
and Type 1 Seyfert galaxies). Following Anderson et al. (2003), positional cross-correlation of the
SDSS DR4 AGNs with the WGACAT sources was performed using a search radius of 60 arcsec.
The first run of the cross-correlation resulted in 1720 X-ray counterparts. Analysis of that run
revealed, however, that the X-ray source coordinate offset from the respective position of the SDSS
object has a systematic component as a function of the source off-axis angle 6, which signaled the
presence of a systematic error in the source coordinates in the WGACAT {6 is the angular offset
of a source position from the central axis of the PSPC field of view). Figures 1 and 2 illustrate
the problem. This systematic error can be reduced, if not eliminated, by applying coordinate
corrections derived from a regression analysis such as shown in Figures 1 and 2. Specifically, based
on these Figures, we corrected the original X-ray source coordinates, RA or i and Dec or i, as follows:
In this way the SDSS data helped to improve the WGACAT coordinates. The second run of cross-
correlation using new coordinates resulted in 1744 X-ray counterparts, 24 more than in the first
run. There were only four cases in which the same WGACAT source was within the search radius
of more than one SDSS AGN; in these cases the AGN closer to the X-ray source was selected.
All the correlations that we studied with this new set of SDSS X-ray AGNs yielded better results,
which indicates that the coordinate improvement, even though small, was real.
The reliability of the identification of the WGACAT sources with the AGNs in our sample can
be directly assessed as follows. Given the footprint area of the SDSS DR4 spectroscopic sample of
4783 deg 2 , and assuming the WGACAT sky coverage of 18%, we get the estimated footprint area
of the PSPC pointed observations £ xr ay = 861 deg 2 . The estimated numbers of AGNs and X-ray
sources in this area are Ni = 10396 and A xray = 8853, respectively. The ratio of the area within
the cross-correlation radius to the footprint area gives an estimate of the probability for an X-ray
source to turn up spuriously within the cross-correlation radius of an AGN. For the 1 arcmin radius
this ratio is P s ~ 10 -6 . For the radius of 30 arcsec, where we find the majority of the X-ray-optical
matches, 1252 of 1744 (72%), this ratio is correspondingly 4 times smaller, P s ~ 2.5 x 10~ 7 . So
the expected number of spurious X-ray-optical associations is N s = N XTay x N{ x P s = 92 and
23, respectively, or ~ 5% and ~ 2% of the actually detected matches within these two radii. The
summary of the cross-correlation statistics is shown in Table 1.
The estimated percentage of spurious positional coincidences, 5% for the 60 arcsec radius,
proves to be the same as in Anderson et al. (2003), where it was obtained for the RASS sources
RA = RA ori - (1.372 + 0.227 #)/3600,
(1)
Dec = Dec ori - (2.119 - 0.260 #)/3600.
(2)
- 7-
Fig. 1. — SDSS-WGACAT positional offset in right ascension, 5RA, vs. X-ray source off-axis
angle, 0. The ^-dependent positive bias seen in this diagram indicates the presence of a systematic
coordinate error.
-8-
10 20 30 40 50 60
off — axis angle 6, arcmin
Fig. 2. — Same as Figure 1 but for positional offset in declination. An indication of a systematic
source coordinate error is evidenced by the increasing negative bias in S Dec toward larger off-axis
angles.
-9-
cross-correlated with SDSS AGNs in the 1400 deg 2 sky area. Another similarity between the
RASS and WGACAT cross-correlation results is in the shape of the X-ray-optical positional offset
distribution. Both the SDSS-WGACAT and SDSS-RASS offset distributions peak at ~10 arcsec,
(compare our Figure 3 with Figure 8 in Anderson et al. 2003). The high reliability of our SDSS-
WGACAT cross-identifications is consistent with the results of classification of WGACAT sources
with ClassX X-ray classifiers 3 . The classifier trained only on X-ray data yielded 1414 (81.1%) of
sources as QSOs or AGNs; within classification uncertainties of the classifier, this is consistent with
all the sources being actually AGN or QSO objects.
Of 1744 SDSS X-ray AGNs 334 (19.1%) are identified in the WGACAT with previously known
QSO or AGN objects, while 1410 (80.9%) are new AGN identifications. The bulk of the AGNs,
79.4%, have redshift confidence P z > 0.95. In comparison, the sample of all SDSS DR4 AGNs has
only 70.6% of objects with high-confidence redshifts.
2.3. The catalog data
The catalog presented in this paper contains optical, X-ray, and radio parameters as shown in
Table 2. The format, name, and short description of the parameters are as follows:
1 14 ID : ID number
2 A19 SDSS ID : SDSS ID
3 F11.6 RA : R.A. from SDSS (J2000.0, degrees)
4 F11.6 Dec : Dec. from SDSS (J2000.0, degrees)
5 F6.3 Redshift : Redshift
6 F4.2 P z : Redshift confidence
7 12 MT : SDSS morphology type: resolved (3) or unresolved (6)
8 F7 . 3 u : Dereddened u magnitude
9 F7 . 3 e u : Error on u magnitude
10 F7.3 g : Dereddened g magnitude
11 F7.3 e s : Error on g magnitude
12 F7.3 r : Dereddened r magnitude
13 F7.3 e r : Error on r magnitude
14 F7 . 3 i : Dereddened i magnitude
15 F7.3 e{ : Error on u magnitude
16 F7.3 z : Dereddened z magnitude
17 F7.3 e z : Error on z magnitude
18 A 18 WGACAT ID : WGACAT ID
19 14 Sr : Offset of the X-ray source from the SDSS position (arcsec)
20 14 9 : X-ray source offset from the PSPC axis (off-axis angle, arcmin)
21 F8.4 C . 1-2.4 : Count rate in the broadband, 0.1 - 2.4 keV
22 F8.4 eco.i_2.4 : Error on the broadband count rate
23 F8.4 log f x : Logarithm of broadband X-ray flux corrected for Galactic absorption ( ergs cm" 2 s" 1 )
24 F8 . 4 log L x : Logarithm of broadband X-ray luminosity corrected for Galactic absorption ( ergs s" 1 )
25 F8 . 4 HR1 : Hardness ratio 1 (corrected for Galactic absorption)
26 F8 . 4 HR2 : Hardness ratio 2 (no Galactic absorption correction applied)
27 F8.4 FIRST ID : FIRST ID
3 ht t p : / / heasarc . gsf c . nasa. gov /ciassx
-10-
150
me dian= 18.05
/ tx=22.15±0.357
ct=14.93±0.252
#=1744
100
43
3
50
10 20 30 40 50 60
SDSS-WGACAT positional offset, arcsec
Fig. 3. — Distribution of the SDSS-WGACAT positional offsets. The legend in this and similar
figures gives the histogram mean, /j, standard deviation, a, and the median; dark shading indicates
resolved AGNs only. The distribution is very similar to that obtained by Anderson et al. (2003)
for the SDSS-RASS positional offsets, which was argued to be as expected if the respective SDSS
AGNs were statistically proper identifications of ROSAT sources.
- 11 -
28 F8.4 / poak : Logarithm of the peak 1.4 GHz flux (mJy)
29 F8.2 /i n t : Logarithm of the integrated 1.4 GHz flux (mJy)
30 F8 . 2 rms : Flux rms
31 F8.2 log.Li.4GHzpeak : Logarithm of the peak 1.4 GHz luminosity ( ergs s _1 Hz _1 )
32 F8.2 logLi.4GHzint : Logarithm of the integrated 1.4 GHz luminosity ( ergs s _1 Hz _1 )
The SDSS parameters include dereddened SDSS magnitudes, u,g,r,i, and z, magnitude errors,
redshift, and redshift confidence (although we denote both redshift and z-band magnitude as z,
the meaning of z is always clear from the context). Also included are SDSS coordinates, RA and
Dec, and SDSS morphological type, which in our case is either 3 (resolved objects) or 6 (unresolved
objects).
The X-ray parameters include the positional offset between the SDSS and ROSAT sources,
5r = [((5RA) 2 + (<5Dec) 2 ] 1//2 , X-ray source off-axis angle, 9 (source angular distance from the PSPC
axis), and broadband (0.1 — 0.4 keV) count rate corrected for vignetting and PSF variation across
the PSPC field. Derived parameters include broadband X-ray flux, f x , and X-ray luminosity,
L x , corrected for galactic absorption. These two parameters were computed using photon index
T = 2 (see, e.g., Brinkmann et al. 1997; 2000). The cosmology adopted in this paper is Hq =
70 km s- 1 Mpc' 1 , fl = 0.3, and A = 0.7.
As seen in Figure 4, at low energies, 0.1 — 0.4 keV (soft band), absorption by the Galactic
neutral hydrogen can be quite significant for our sources. But no absorption is evident at 0.4 — 0.9
(mid band) and 0.9 — 2.4 keV (hard band) for the range of HI column densities shown in Figure 4.
The column density of the absorbing material, Njji, shown in Figure 4 was computed based on the
Dickey k, Lockman (1990) and Stark et al. (1992) data. Absorption effects in the soft-band count
rate can be accounted for on a purely empirical basis by removing the trend revealed by Figure 4:
where s\ = —0.815. Equation (3) assumes that soft-band absorption becomes negligible at log Ahi <
19.8 cm' 2 , so the latter value was used for normalization in Equation (3). The parameter C so ft
from this equation is used to compute the hardness ratio HR1. The catalog includes two hardness
ratios, HR1 and HR2, which are denned as follows:
log C soft = log c;
' s a b f r b + ^i(19.8-logiVHi),
(3)
HR1 =
(4)
C m id + Cgoft
and
HR2 =
(5)
-12-
Fig. 4. — Soft-band count rate vs. Galactic HI column density. The legend in this and similar
figures gives the linear regression parameters: correlation coefficient, c, intercept, cto, and slope, Si,
with uncertainty in the slope value given in percent. The downward trend in this diagram reflects
the varying Galactic absorption, implying an increase in the absorption by an order of magnitude
as logA^Hi increases from 20 to 21 cm~ 2 .
-13-
3. Statistics of the SDSS X-ray AGNs
3.1. X-ray detection rate and optical brightness
The bulk of the SDSS DR4 AGNs are from a sample that is limited in the i band to i < 19.1.
Therefore, one may expect that, similar to the i-brightness distribution of the entire SDSS AGN
sample, the number of X-ray AGNs would increase toward fainter i and then sharply drop at the
limiting magnitude. The actual shape of the brightness distribution depends on the detectable
flux level in the ROSAT PSPC pointed observations for AGNs of various optical brightnesses. The
X-ray AGN detection rate as a function of optical brightness is illustrated in Figure 5. It shows
the number ratio of the AGNs detected in the X-ray and all AGNs in the sky footprint area of
the PSPC pointed observations. As expected, the rate of X-ray detections is very high for bright
AGNs, up to 50% and more at magnitudes i < 16. It declines to ~ 10% at the limiting magnitude
of the SDSS main quasar sample, i = 19.1. Despite this decline, the brightness distribution of the
X-ray AGNs in our sample sharply increases toward the limiting magnitude (Figure 6), which is
obviously due to an even sharper increase in the number of SDSS AGNs toward faint magnitudes.
The distribution is nearly flat within i ~ 19 — 20, where AGNs from outside of the SDSS AGN
main sample dominate. This implies that a large number of X-ray sources matching yet-unidentified
AGNs with i > 19.1 in the SDSS photometric database should be present in the WGACAT. As
indicated below, these are expected to be mostly weak sources, below the detection limit of the
ROSAT All Sky Survey.
3.2. Redshift distribution
Figure 7 displays the redshift distribution of the SDSS X-ray AGNs. Similar to the Anderson
et al. (2003) results, the overall distribution peaks at z ~ 0.3, gradually declining to almost zero
at z ~ 2.5. This is very different from the distribution of all SDSS AGNs (Figure 8), which peaks
at z ~ 1.7 and then sharply declines toward z ~ 2.5. This difference has to do with the evolution
of the X-ray luminosity function, the depth of the WGACAT, and the fact that the spectrum in
the X-ray is steeper than in the optical, with the value of the energy index typically ranging within
a x = —1.0 to —1.5 for the power-law spectrum vs. a Q ~ —0.5 for the optical power-law spectrum
(see, e.g., Brinkmann et al. 2000; Anderson et al. 2003).
Most of the resolved AGNs are within the redshift range z = — 0.3, and their distribution
peaks at z ~ 0.2. The distribution of resolved AGNs in Figure 8 is similar to that in Figure 7.
3.3. X-ray flux and optical brightness
PSPC pointed observation exposures are, on average, substantially deeper than those in the
ROSAT All-Sky Survey, so the WGACAT contains many sources below the RASS detection limit.
-14-
0)
u
o
cd
1.2
1.0
0.8
- x
o
<u 0.6
<v
0.4
0.2
0.0
15
x
16 17 18
i— band magnitude
19
Fig. 5. — Fraction of AGNs detected in the X-ray as a function of brightness in the i-band. At
bright magnitudes, % < 16, more than half of the SDSS AGNs in the footprint area of the ROSAT
PSPC pointings are detected in the X-ray at typical exposure times of ROSAT PSPC pointed
observations. This fraction drops below 20% near the brightness limit of the SDSS main quasar
sample, % = 19.1.
43
3
140
120
100
80
60
40
me cZmn= 18.42
/ tx=18.23±0.022
ct = 0.930±0.016
#=1685
20
15
16 17 18
i-band magnitude
19
20
Fig. 6. — Brightness distribution of SDSS X-ray AGNs in the i band. The distribution gradually
increases to the peak near the brightness limit of the SDSS AGN main sample, i « 19, and then
sharply drops. It is nearly flat within i ~ 19 — 20, where AGNs from outside of the SDSS AGN
main sample dominate. So the WGACAT should contain a large number of sources matching yet
unidentified AGNs with i > 19.1 in the SDSS photometric database.
-16-
median=0 .73 1 _
/x=0.851±0.014 _
ct = Q.587±0.QQ9 _
#=1732
0.0 0.5 1.0 1.5 2.0 2.5
redshift
100
80
Fig. 7. — Redshift distribution of SDSS X-ray AGNs within the redshift range z = — 2.5 with
X-ray emission from the WGACAT.
-17-
2000
1500
median= 1 .272 _
yuj= 1.2 18 ±0.002 _
ct = Q.609±0.QQ1 _
#=55177 _
s-,
| 1000
500
0.0
0.5
1.0 1.5
redshift
2.0
u
2.5
Fig. 8.-
Redshift distribution of all SDSS DR4 AGNs within redshift the range z = 0- 2.5.
-18-
Comparison of our Figure 9 with Figure 3 in Shen et al. (2005) illustrates this. While in the sample
with the RASS counterparts less than 1% of the AGNs are fainter than log/ z = —13, these are
over 25% in our sample. As seen in Figure 10, they become dominant at redshifts z > 1.5. So
these low-flux objects provide an important probe for the AGN X-ray emission at redshifts z > 1.5.
Among the resolved AGNs the fraction of weak X-ray sources is small (see Figure 11), consistent
with a similarly low fraction of weak unresolved AGNs at low redshifts.
Figure 12 reveals a strong correlation between X-ray flux and optical brightness, which is
consistent with previous studies. The sparse but uniformly distributed population in the area at
the faint end of the diagram, i = 19 — 20, suggests that many more SDSS AGNs with log f x from
— 13.0 to —13.5 can be found among the WGACAT sources 1.0-1.5 mag fainter than the 19.1
mag brightness limit of the SDSS main AGN/QSO sample. Given that only ~ 20% of the X-ray
sources are identified with AGNs in the SDSS-WGACAT footprint area (1744 of 8853 sources),
the number of X-ray AGNs fainter than i = 19.1 can be a factor of 5 larger than our present
sample. So our current sample can be substantially expanded in the low-flux, high-redshift domain
{log f x < —13, z > 1.5) by going 1.0-1.5 mag fainter in the i band. We plan to do that in the
future, using the ClassX technique (Suchkov, Hanisch, and Margon 2005) to identify these faint
AGNs in the SDSS photometric database and estimate their redshift.
3.4. Hardness ratio
The relationship between hardness ratio and flux can provide useful information regarding
X-ray absorption (see, e.g., Green et al. 2004).
The diagrams in Figures 13 and 14 present the hardness ratio for our sample plotted against
the soft-band flux. In this band, 0.1 — 0.4 keV, both the resolved and unresolved AGNs exhibit
essentially the same tendency to have a hardness ratio that is somewhat higher toward weaker
sources. But at higher energies, in the 0.4-0.9 and 0.9-2.4 keV bands, the hardness ratio shows
no trend whatsoever. This is consistent with Kim et al. (2004b) and Green et al. (2004), who
concluded that spectral hardening toward faint fluxes is substantial only in the soft bands and is
thus most likely due to absorption.
3.5. X-ray-to-optical flux ratio
Figures 15 and 16 show the X-ray-to-optical flux ratio as dependent on redshift. Resolved
and unresolved AGNs exhibit opposite behavior with respect to how the X-ray-to-optical flux ratio
varies with redshift. Consistent with previous studies, Figure 15 shows log(f x / fi) trending down,
so that the flux ratio decreases on average by a factor of ~ 3 over the redshift range z = 0-2. In
contrast, for the resolved AGNs, on a much smaller redshift scale, this ratio increases rather than
decreases with redshift. The ratio becomes on average larger by a factor of ~ 3 as redshift increases
150
median— - 12.7
/z=-12.6±0.011
ct = Q.464±0.QQ7 -
#=1741 _
100 -
43
3
14.0
13.5
13.0
12.5
12.0
11.5
11.0
log / x (ergs s 1 cm 2 )
Fig. 9. — X-ray flux distribution for the SDSS-WGACAT sample. The sample is substantially-
deeper than the sample of SDSS AGNs with the RASS X-ray counterparts. While in the latter
sample < 1% of the sources are weaker than logf x = —13, these are ~25% in our SDSS-WGACAT
sample.
-20-
o
-1 1.0
-11.5
-12.0
-12.5
-13.0
-13.5
-14.0
> c
o o
'$ a Z <x o
*tl o-
<*> Oct;
o
o o <
o
> ° V - s
o «
<*> o
f?.8
o o
. o
Unresolved
0.5
1.0
1.5
2.0
redshift
2.5
Fig. 10. — X-ray flux as a function of redshift for unresolved AGNs with redshifts < z < 2.5. As
expected, high-redshift AGNs are weaker in the X-ray, with flux declining on average by an order
of magnitude within this redshift range. Weak sources, log f x < —13, become dominant at z > 1.5.
- 21 -
10
-11
12
13
o o o
o o<> o§o «>o o 8 v
^^Vf-V* o
• o «
ooS<*>o« o o
14
0.
Resolved
0.1
0.2 0.3
redshift
0.4
0.5
Fig. 11. — Same as Figure 10, but for resolved AGNs only and a narrower redshift range, z = — 0.5.
X-ray flux drops on average by an order of magnitude within this range.
-22-
Fig. 12. — X-ray flux vs. i-band magnitude for AGNs brighter than i = 16 mag. The diagram
reveals the expected correlation between the two parameters (linear regression is displayed sepa-
rately for radio-loud and radio-quiet AGNs) and shows that at any given brightness the radio-loud
AGNs have fluxes higher than those of the radio-quiet AGNs, on average by a factor of ~2. It also
suggests that a large number of additional WGACAT counterparts can be found for SDSS AGNs
below the brightness limit of the SDSS main spectroscopic sample, i = 19.1 mag.
-23-
-0.2
■0.4 -
c = 0.419, aQ = -0.93
s^-o.iaois.gi^o
N-1342
0.6
-0.5
0.0
log C.
soft
Fig. 13. — Hardness ratio HR1 vs. the 0.1 — 0.4 Kev count rate corrected for Galactic absorption.
There is a tendency for hardness ratio to be higher for fainter sources, which is likely due to a
larger, on average, absorption in the case of faint sources. Almost all these sources are from the
area on the PSPC detector close to the detector axis, 6 < 20 arcmin, where the detector sensitivity
is highest.
-24-
Fig. 14. — Same as Figure 13, but for resolved AGNs. Resolved AGNs exhibit essentially the same
behavior as the unresolved AGNs.
15
15
14
X
X
O
6 14
13
13
12
5
5
5
5
c = 0.398, a0=14.38
si = -0.282±6.24%
N-1359
o t \ I $ o o o * o
°<A <> . * »« « o o
o 8
o @
o o>
«^«8%^* X> o o
*8
O o
O
»o o
O
o ? o
°° *
o o
* o
3£>
o
Unresolved
0.5
1.0 1.5
redshift
2.0
2.5
Fig. 15. — X-ray-to-optical flux ratio for unresolved AGNs. f x ( ergs s _1 cm -2 ) is the X-ray broad-
band flux, 0.1 — 2.4 keV, and fa is spectral energy flux density in the % band ( ergs s _1 cm -2 Hz -1 ),
log fi = — 0A(i + 48.6). The diagram shows the known decline of the ratio with redshift.
-26-
O
15.5
15.0
14.5
14.0
13.5
13.0
12.5
c = 0.326, a0=13.71
si=1.316±15.0%
N = 371
Resolved
0.1
0.2 0.3
redshift
0.4
0.5
Fig. 16. — Same as Figure 15, but for resolved AGNs. In contrast with unresolved AGNs, the
X-ray-to-optical flux ratio of resolved AGNs exhibits a substantial upward trend.
-27-
from to 0.4 (unresolved AGNs exhibit no trend whatsoever within this narrow redshift range).
This difference should be due to the fact that the optical luminosity of low-redshift active galaxies,
which is dominated by stellar light, occupies a much narrower range than X-ray luminosity. Toward
higher redshifts, X-ray flux can be detected only from galaxies with the highest X-ray luminosities,
but optical luminosities will not be much different from those at low redshifts. As a result, the
average X-ray-to-optical flux ratio gets higher. It is worthwhile to note that the flux ratio of the
resolved AGNs is substantially smaller than that of the unresolved AGNs within the same range,
by a factor of ~ 3.
3.6. Radio sources
Confronting AGN radio properties with those in the optical and X-ray can provide important
insights into the diversity of AGNs. Best et al. (2005a, b) analyzed a large sample of radio-loud
SDSS Type 2 AGN galaxies at redshifts 0.03 < z < 0.3, which was constructed using the NRAO
VLA Sky Survey (NVSS; Condon et al. 1998) and the Faint Images of the Radio Sky at Twenty
cm survey (FIRST; Becker et al. 1995). They concluded that the optical and low radio- luminosity
AGN phenomena are independent and are triggered by different physical mechanisms. The data
suggested that while the integrated [OIII] luminosity density from emission-line AGNs is produced
mainly by black holes with masses below 10 8 Mq, the integrated radio luminosity density comes
from the most massive black holes in the Universe. X-ray emission is another manifestation of active
galactic nuclei, independent of the optical and the radio. Therefore, one may expect that combining
X-ray data with radio and optical data would result in further insights into the differences between
various AGNs.
Brinkmann et al. (2000) cross-correlated the ROSAT All-Sky Survey with the FIRST sources.
This resulted in a sample of 843 sources, of which slightly over 100 sources were identified as various
types of AGNs. But many of the X-ray-radio matches proved to be very faint in the optical.
Because of the lack of spectra, and hence redshifts, these faint objects remained unclassified, which
substantially impacted the scope of analysis of ROSAT radio-loud AGNs in that paper.
In this paper we also address the issue of AGN radio emission and include in our catalog the
data for the AGNs for which we found radio counterparts in the FIRST catalog (the FIRST catalog
can be found at http://sundog.stsci.edu/first/catalogs.html; see also White et al. 1997.) As Best
et al. (2005a) showed, most of the FIRST sources with a 10 arcsec separation from SDSS AGNs
are reliable radio identifications. So we cross-correlated 57,800 SDSS DR4 Type 1 AGNs with the
FIRST sources and found 4,792 AGNs (8.3%) that have a matching radio source within 10 arcsec.
We will refer to them as "radio-loud AGNs" . The FIRST sources are limited to fluxes > 1 mJy, so
discrimination between our radio-loud and radio-quiet samples occurs at the flux level of ~ 1 mJy.
Among the identified radio- loud AGNs, 224 objects were in our list of AGNs with X-ray
emission from the WGACAT. The fraction of radio-loud AGNs in our sample, 12.8%, is thus
-28-
600
500
median=0 .340 E
/ a=0.457±Q.QQ5 E
ct = Q.377±Q.QQ3 E
#=4465 4
400
| 300
200
100
0.0 0.5 1.0 1.5
SDSS-FIRST position offset, arcsec
2.0
Fig. 17.— SDSS-FIRST positional offset distribution. Of a total of 4792 FIRST sources matching
the position of SDSS DR4 AGNs, 4465 (-93.2%) sources are within 2 arcsec of the SDSS AGN
position.
-29-
noticeably larger than the average for SDSS AGNs, by a factor of ~1.5%.
As seen in Figure 17, positional offsets of the FIRST counterparts are in general very small,
with ~93% of them within 2 arcsec of the SDSS position. The discrepancy between the SDSS and
FIRST positions within this radius is, on average, larger for sources with larger integrated-to-peak
flux ratio. This suggests that much of it is due to extended sources, for which peak radio emission
does not necessarily coincide with the central optical object and source coordinates in general are
less certain.
Consistent with earlier work, the X-ray emission of the radio-loud AGNs in our sample is
stronger than that of the radio-quiet AGNs, on average by a factor of ~ 2 at any given optical
brightness (see Figure 12).
In Figure 18 the 1.4 GHz flux is plotted against the broadband X-ray flux. At low redshifts,
there are many sources that are strong in the X-ray and weak in the radio. At high redshifts, there
is a tendency for AGNs with stronger radio emission to also be strong in the X-ray. The sample
has essentially no strong radio sources that are weak in the X-ray (upper- left corner in Figure 18).
4. Summary
We presented a sample of 1744 SDSS DR4 AGNs with X-ray emission from ROSAT PSPC
pointed observations. The sample was obtained by positional cross-correlation of all SDSS DR4
AGNs/QSOs (SDSS spectroscopic type 3) with all unique X-ray sources in the catalog of ROSAT
pointed observations (WGACAT). Of 1744 X-ray sources, 1,410 (80.9%) are new AGN identifica-
tions while the rest are the sources identified in the WGACAT with previously known AGNs. Of
4574 SDSS DR4 AGNs for which we found radio matches in the catalog of the FIRST radio sources,
224 turned up in our sample of X-ray emitting AGNs. We provided the major parameters for the
sample objects from the respective catalogs, as well as a number of derived parameters, such as
X-ray broadband luminosity, 1.4 GHz radio luminosity, and X-ray hardness ratios. We presented a
series of statistical relationships for the sample objects that illustrate their overall properties and
demonstrate a high reliability of the X-ray counterpart identification. We plan to make use of the
sample data in our subsequent studies of AGNs and starburst galaxies for which the combined
optical and X-ray information is crucial.
Funding for the creation and distribution of the SDSS Archive has been provided by the
Alfred P. Sloan Foundation, the Participating Institutions, the National Aeronautics and Space
Administration, the National Science Foundation, the U.S. Department of Energy, the Japanese
Monbukagakusho, and the Max Planck Society. The SDSS Web site is http://www.sdss.org/. The
SDSS is managed by the Astrophysical Research Consortium (ARC) for the Participating Insti-
tutions. The Participating Institutions are The University of Chicago, Fermilab, the Institute for
Advanced Study, the Japan Participation Group, The Johns Hopkins University, the Korean Sci-
-30-
4
3
(0
CD
k 2
O
o
O
O
o
rP
8 Oo
o°
o
30
o
O
o o^o
^ o
O o-d
_ °^ O
o OOO Oo°
o
0° <»o o d
o
o o
O
O o
o
o
O
o o
o
Oo
13.5
13.0
12.5
log f x
12.0
1 1.5
1 1.0
Fig. 18. — The 1.4 GHz flux density (mJy) vs. broadband X-ray flux ( ergs s _1 cm~ 2 ) for radio-loud
X-ray AGNs with counterparts from the FIRST catalog. The symbol size is linearly scaled with
redshift, with z < 0.02 and z > 2.3 corresponding to the smallest and largest symbols, respectively.
At low redshifts, there are many weak radio sources that are strong in the X-ray. At high redshifts,
the AGNs with stronger radio emission tend to also be strong in the X-ray. Only two weak X-ray
emitters are found among strong radio sources.
-31 -
entist Group, Los Alamos National Laboratory, the Max-Planck-Institute for Astronomy (MPIA),
the Max-Planck- Institute for Astrophysics (MPA), New Mexico State University, the University of
Pittsburgh, the University of Portsmouth, Princeton University, the United States Naval Observa-
tory, and the University of Washington.
We thank the referee for useful comments and recommendations, which helped us to improve
the paper. This research has made use of data obtained from or software provided by the US
National Virtual Observatory, which is sponsored by the National Science Foundation.
REFERENCES
Abazajian, K., et al. 2005, AJ, 129, 1755
Anderson, S. F. et al. 2003, AJ, 126, 2209
Barger, A. J., Cowie, L. L., Mushotzky, R. F. et al. 2005, AJ, 129, 578
Becker, R. H., White, R. L., & Helfand, D. J. 1995, ApJ, 450, 559
Best, N. P., Kauffmann, G., Heckman, T. M., & ivezic, Z. 2005a, MNRAS, 362, 9
Best, N. P., Kauffmann, G., Heckman, T. M., Brinchmann, J., Chariot, S., Ivezic, Z., & White, S.
D. M. 2005b, MNRAS, 362, 25
Brand, K. et al. 2005, ApJ, 626, 723
Brandt, W. N. & Hasinger, G. 2005, ARA&A, 43, 827
Brinkmann, W. et al. 2000, A&A, 356, 445
Brinkmann, W., Yuan, W. & Siebert, J. 1997, A&A, 319, 413
Condon, J.J. et al. 1998, AJ, 115, 1623
Dickey, J. M. & Lockman, F. J. 1990, ARA&A, 28, 215
Fiore, F. et al. 2003, A&A, 409, 79
Green, P. J. et al. 2004, ApJS, 150, 43
Hasinger, G. Miyaji, T., & Schmidt, M. 2005, A&A, 441, 417
Heckman, T. M., Kauffmann, G., Brinchmann, J., Chariot, S., Tremonti, C, & White, S. 2004,
ApJ, 613, 109
Hornschemeier, A. E., Heckman, T M., Ptak, A. F., Tremonti, C. R., & Colbert, E. J. M. 2005,
AJ, 129, 86
Imanishi, M. & Wada, K. 2004, ApJ, 617, 214
Kauffmann, G. et al. 2003, MNRAS, 346, 1055
Kim, D.-W. et al. 2004a, ApJS, 150, 19
Kim, D.-W. et al. 2004b, ApJ, 600, 59
-32-
McGlynn, T. et al. 2004, ApJ, 616, 1284
Shen, S., White, S. M. D., Mo, H. J., Voges, W., Kauffmann, G., Tremonti, Ch., Anderson, S. F.
2005, Astro-ph/05 12645
Stark, A. A. et al. 1992, ApJS, 79, 77
Suchkov, A. A., Hanisch, R. J., & Margon, B. 2005, AJ, 130, No. 6
Voges, W. et al. 1999, A&A, 349, 386
Voges, W. et al. 2000, IAU Circ. 7432
White, R. L., Becker, R. H., Helfand, D. J., & Gregg, M. D. 1997, ApJ, 475, 479
Wolf, C. et al. 2004, A&A, 421, 913
Yuan, W., Brinkmann, W., Siebert, J., & Voges, W. 1998, A&A, 330, 108
This preprint was prepared with the AAS IATjrjX macros v5.2.
-33-
Table 1. Statistics of the SDSS-WGACAT-FIRST Matches
Source Catalog Sky Footprint Wwgacat 3, ^match b Spurious
(deg 2 )
SDSS 4783 10396 1744 <5%
WGACAT 861 8853 1744
FIRST d 4797 862 224 <5%
Note. — The WGACAT sky footprint area is estimated as 0.18 of the
footprint area of the SDSS DR4 Spectroscopic Catalog. The FIRST footprint
area is assumed to be the same as that of the SDSS.
a Estimated number counts in the WGACAT footprint area.
b Numbcr of the SDSS-WGACAT and FIRST d -WGACAT matching
sources in the WGACAT footprint area.
c Percentage of spurious associations with X-ray sources; < 2% for 72% of
the AGNs, which are within = 30 arcmin.
d Only sources matching the SDSS DR4 AGNs are counted.
Table 2. SDSS AGNs with X-Ray Emission from ROSAT PSPC Pointed Observations
SDSS Parameters
ID
SDSS ID
RA
(deg)
Dec
(deg)
Redshift
p a
MT b
9°
r c
e r
i°
z"
56
587725473883357292
162.552597
62.884251
0.201
1.00
3
20.329
0.099
19.033
0.015
18.014
0.009
17.443
0.008
17.293
0.024
57
587725489986535603
253.006348
62.535847
1.633
0.98
6
19.281
0.027
19.087
0.011
19.158
0.013
19.000
0.018
18.969
0.049
58
587725489988042772
255.566803
59.448818
1.169
1.00
6
19.394
0.029
19.289
0.013
19.006
0.012
18.991
0.017
18.980
0.048
59
587725489988108517
255.583603
59.260723
1.798
0.97
6
20.154
0.046
19.739
0.017
19.261
0.014
18.743
0.014
18.449
0.032
68
587725503945310523
254.581360
62.639751
0.703
0.97
6
20.279
0.054
19.762
0.016
19.779
0.022
19.801
0.030
19.726
0.103
X-Ray Source Parameters
ID
WGACAT ID
Sr
(arcscc)
(arcmin)
Co. 1-2.4
ec 0.1-2.4
log/a
( ergs cm ~ 2 s ~ 1 )
log Ii
( ergs s -1 )
HR1
HR2
56
1WGA .11050.1+6253
3
34
0.0169
0.0021
-12.75
43.16
-0.66
0.47
57
1WGA J1651. 9+6232
22
34
0.0439
0.0026
-12.09
45.31
-0.72
-0.23
58
1WGA J1702. 2+5926
38
23
0.0038
0.0007
-13.24
43.98
-0.70
-0.06
59
1WGA J1702. 2+5916
59
30
0.0041
0.0009
-13.20
44.25
-0.81
0.10
68
1WGA J1658. 3+6238
9
10
0.0168
0.0012
-12.52
44.36
-0.39
0.19
FIRST Radio Source Parameters
ID
FIRST ID
/peak
/int
rms
log il.4GHzpoak
log £l.4GHzint
(mjy)
(mjy)
(mjy)
( ergs s~ 1 Hz~ 1
( ergs s~ 1 Hz~ 1
56
10480+63011J
1.710
1.500
0.143
30.178
30.121
57
16540+62339J
31.830
39.830
0.151
33.117
33.215
58
17000+59294J
27.240
29.230
0.156
32.818
32.848
59
17000+59042J
8.820
8.920
0.142
32.623
32.628
68
17000+62339J
1.490
1.970
0.148
31.169
31.290
Note. — Only a few lines of the table arc shown here as a guide to the table's content and format. The table is available in its entirety in a machine-readable form in the electronic edition of the Astronomical
Journal
a R.cdshift confidence.
b SDSS morphology type, 3 for resolved and 6 for unresolved objects.
c Corrcctcd for Galactic reddening (SDSS dcrcddcncd magnitudes).
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Books to Borrow
Open Library
TV News
Understanding 9/11