Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://hea-www.harvard.edu/~pgreen/Papers/hs1216_04.ps Äàòà èçìåíåíèÿ: Wed Sep 14 21:41:23 2005 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 00:56:53 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: m 5

Mon. Not. R. Astron. Soc. 349, 1261--1266 (2004) doi:10.1111/j.1365­2966.2004.07592.x
HS 1216+5032: a physical quasar pair with one radio­loud broad
absorption line quasar
P. J. Green, 1# Thomas L. Aldcroft, 1 Warren R. Brown, 1 Olga Kuhn 2 and Abhijit Saha 3
1 Harvard­Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA
2 Joint Astronomy Center, 660 N. A'ohoku Place, University Park, Hilo, HI 96720, USA
3 Kitt Peak National Observatory, National Optical Astronomy Observatory, PO Box 26372, Tucson, AZ 85726, USA
Accepted 2004 January 5. Received 2003 December 31; in original form 2003 October 20
ABSTRACT
We report on new multiwavelength observations of HS 1216+5032, a pair of quasi­stellar
objects (QSOs) at z = 1.45 separated by 9.1 arcsec, which has been a perennial candidate
for a massive dark lens. We explore high signal­to­noise ratio optical spectra from the MMT
of both components, which show that aside from the effects of absorption, the emission­line
profiles are quite similar. Near­infrared spectra from the United Kingdom Infrared Telescope
show identical velocities in H# emission, but a significant difference in the strength of the
narrow component, which is difficult to explain in a lens scenario. We highlight that, based
on data from the Very Large Array Faint Images of the Radio Sky survey, HS 1216+5032B
is a radio­loud broad absorption line QSO, which certifies HS 1216+5032 as a physical
quasar pair. Intriguingly, both quasars show spectroscopic evidence for high accretion rates
and large Eddington ratios L/L Edd , supporting the hypothesis that close galaxy interactions
trigger nuclear activity.
Key words: gravitational lensing -- quasars: absorption lines -- quasars: individual: HS
1216+5032 -- X­rays: general -- X­rays: individual: HS 1216+5032.
1 I NTRODUCT I ON
Most gravitational lens candidates with separations ## < 3 arc­
sec have identifiable primary lens galaxies in deep Near­Infrared
Camera and Multi­Object Spectrometer (NICMOS) observations. 1
There are 16 wide separation quasar pairs (WSQPs, with 3 <## <
10 arcsec), which all require large lensing masses. Three are bona
fide gravitational lenses, each with similar optical/radio flux ratios,
negligible spectroscopic differences, and a lens galaxy. One shows
a host in the faint but not the bright NICMOS image, so could be
a binary (Mu”noz et al. 2001). Three more systems are pairs with
strongly discrepant flux ratios, or >3# velocity differences, so are
very probably binary quasars. The remaining nine are WSQPs with
essentially identical redshifts for which no lens galaxy or cluster has
been detected. If these WSQPs are gravitationally lensed, then they
are produced by cluster­scale masses that failed to make substantial
galaxies (i.e. `dark clusters'). However, without direct detection of
lens baryons, proof of lensing is challenging. Detection of time de­
lays between the light curves of lensed image components requires
intensive monitoring over time­scales which may span years. The
required time­scales depend strongly on lens geometries, which are
unknown in the dark lens case. Direct detection of shear in the quasar
# E­mail: pgreen@cfa.harvard.edu
1 See http://cfa­www.harvard.edu/castles.
host galaxy images can serve as proof of lensing, but is also difficult
(Keeton et al. 2000). It is proving easier to disprove lensing for any
particular case than to prove it (Green et al. 2002; Aldcroft & Green
2003).
Of the nine potentially lensed WSQPs, we here examine the
widest (9.1 arcsec) pair, HS 1216+5032, which was discovered
in the Hamburg Quasar Survey (Hagen et al. 1996). The compo­
nents are bright (B A = 17.2 and B B = 19.0), and the redshifts of
the two components (z = 1.455 ± 0.005) are identical within the
errors. Given the wide separation of the pair, an intervening lensing
cluster mass should be easily detectable (i.e. in optical or X­rays)
even with an unusually low baryon fraction (>1/3 normal). 2 No
lensing galaxy or cluster is evident to an imaging depth of R #
22.5 mag (Hagen et al. 1996). Nevertheless, the case for lensing
was invigorated by a detailed ultraviolet (UV) spectroscopic study
(Lopez, Hagen & Reimers 2000) with the Hubble Space Telescope
(HST) Faint Object Spectrograph (FOS), which found a complex
of intervening C IV absorber systems. At the z = 0.72 redshift of
the absorber system, the minimum velocity dispersion required of
an object causing the observed image separation in a standard lens
2 For a singular isothermal sphere model of a lensing cluster, the minimum
flux redshift is about 0.7, for which the expected X­ray flux of the lensing
cluster would be #2 â 10 -14 erg cm -2 s -1 , and an L # elliptical has R #
21.
C
# 2004 RAS

1262 P. J. Green et al.
model 3 is only 640 km s -1 . The three absorbers are observed to span
1500 km s -1 along the line of sight, which could be taken as strong
evidence for a cluster lensing potential, as noted by the authors.
HS 1216+5032 with its cluster of intervening absorbers provides
a particularly interesting case for a dark lens, so we obtained op­
tical and infrared (IR) spectroscopy to further investigate spectral
differences, as described in Sections 2 and 3. We also analyse mea­
surements from the Very Large Array (VLA) Faint Images of the
Radio Sky (FIRST) survey (Section 4). In Section 5, we show that
the radio data effectively rule out the lens scenario, and we briefly
discuss HS 1216+5032 in the context of models for interaction­
triggering of active galactic nuclei (AGN).
2 OPT I CAL SPECTRA AND COLOURS
Both ground­based (Hagen et al. 1996) and HST FOS spectra (Lopez
et al. 2000) show that only the fainter component (HS 1216+5032B)
shows strong broad absorption line (BAL) systems in Ly#, C I, C II,
C III], C IV, N III, N V and O VI. This has led several authors to con­
clude that the pair is a true binary. However, the existence of BALs
in only one quasar of a wide pair such as HS 1216+5032 may not
be adequate to dismiss the lens hypothesis for several reasons. First,
some spectral differences are always seen between images in lenses.
The degree to which spectral similarity should be used as an argu­
ment for lensing has been a continuing subject of debate. Beyond
consistent redshifts, no quantitative spectral similarity criterion has
been devised to label a pair as definitely lensed or not (i.e. Falco et
al. 1999; Peng et al. 1999). Secondly, the spectral differences might
be explained in a lensing scenario by (1) narrow BAL wind/cloud
geometry combined with viewing angles (Chelouche 2003), (2) vari­
ability (Barlow et al. 1992) and time delay, and/or (3) microlensing
and dust (e.g. Lewis et al. 2002).
We obtained optical spectroscopy of the two components at the
Multiple Mirror Telescope (MMT) on 2003 April 03 with the blue
channel of the MMT spectrograph. We exposed for 35 min with
the 300 l mm -1 grating centred at 6000 å and a 1.25­arcsec slit,
thereby achieving 2 å pixel -1 resolution. Fig. 1 shows an overplot
of the flux­calibrated spectra, with the spectrum of HS 1216+5032B
scaled up by a constant factor of 4. We note that strong BALs affect
the emission­line profiles of HS 1216+5032B. Using the `balnicity
index' (BI) 4 of Weymann et al. (1991), we measure B I = 1870, with
maximum outflow
velocityv max = 9500 ± 1500 km s -1 (including
a conservative redshift error of ±0.005). A revised definition from
Hall et al. (2002), the absorption index (AI), yields AI = 4178 ±
9 km s -1 .
The Mg II # 2800 emission line of HS 1216+5032A shows a
strong (W rest
# = 0.8 ± 0.04 å) narrow absorption line (FWHM =
480 ± 50 km s -1 ) at very nearly the systemic redshift. For HS
1216+5032B, we identify broad shallow absorption on the blue
3 The image separation ## =
8#(#v /c) 2 D LS /D OS depends only on the
velocity dispersion of the potential
#v and the ratio of the comoving distances
between the lens and the source, D LS , and the observer and the source, D OS .
Here we assume a single isothermal sphere (SIS) model (Schneider, Ehlers
& Falco 1992) and throughout use H 0 = 70 km s -1 Mpc -1 , #M = 0.3 and
## = 0.7.
4 The BI is a measure of the equivalent width of absorption measured in km
s -1 , between 3000 and 25 000 km s -1 blueward of the expected systemic
redshift z sys of the emission line, where z sys is preferably based on a low­
ionization narrow emission line. The integration includes only contiguous
troughs at 90 per cent of the continuum level, and spanning >2000 km s -1 .
A positive BI indicates a classic BAL. The AI is a revised calculation that
admits measurements of narrower troughs, and provides for error estimates.
Relative
Flux
MgII
CIV
AlIII/CIII]
B
A
Fe blend
sky
line
Figure 1. The optical MMT spectrum of HS 1216+5032A is shown at the
top, with the spectrum of HS 1216+5032B shown at the bottom, scaled
upward by a constant factor of 4 for display purposes. The continuum
shapes are not reliable. Strong BALs affect the emission­line profiles of HS
1216+5032B. The Mg II# 2800 emission line of HS 1216+5032A shows
a narrow absorption line at 6871 å (observed). The small spike at 5577 å
results from poor night­sky line subtraction there.
half of the Mg II # 2800 emission line, and clear broad troughs in
the Al III/C III] complex. This is consistent with the finding that two
of three BAL systems detected in the HST FOS spectra (-1000,
-3000 and -4000 km s -1 , respectively; Lopez et al. 2000) show
smooth broad absorption in low­ionization lines of Ly#, C II#1334
and C III]#977. We measure AI = 0 for both troughs, so HS
1216+5032B may thus be a borderline member of the class of low­
ionization BAL (loBAL) QSOs.
The component continuum slopes in Fig. 1 appear to be dis­
crepant. Excess reddening of the BALQSO is expected if there is
dust associated with the absorber, and some reddening has been
demonstrated in recent samples from the Sloan Digital Sky Survey
(SDSS; Reichard et al. 2003). However, we obtained the MMT spec­
tra of both objects simultaneously, with the slit necessarily at
34# from the parallactic angle. At the mean airmass 1.14, differential
atmospheric refraction would cause a 0.45­arcsec offset between
light at 3500 and 7000 å across the 1.25­arcsec slit. The apparent
continuum shapes may not be trustworthy, or may be due to differing
line­of­sight absorption.
Because no published CCD photometry is available for the pair
that might confirm the difference in optical spectral slopes, we ob­
tained images at the WIYN 3.5­m on Kitt Peak. On 2003 June 5
using the MiniMosaic camera, we obtained two exposures of 50 s
each in Harris B and R filters. We find B A = 17.32 and B B = 19.4,
and in the red R A = 16.79 and R B = 18.5, with random photomet­
ric errors <0.04 mag and nightly solution errors #0.05 mag. The
colours of the two components are (B--R) A = 0.52 and (B--R) B =
0.90 mag (both ±0.05 mag), so they thus differ significantly, with
#(B--R) = 0.37 ± 0.06.
While the possibility of slit losses in the MMT spectra remain,
the photometric colours are consistent with the spectral differences
shown in Fig. 1. Contributing to the colour difference is not only the
relatively redder continuum in HS 1216+5032B, but also the fact
that the strong emission lines which fall into the B band are severely
weakened by the BALs.
3 I NFRARED SPECTRA
Near­IR spectra of HS 1216+5032 were obtained on UT 2003 March
20andUT2003 April 04 through the service programme at the United
C
# 2004 RAS, MNRAS 349, 1261--1266

HS 1216+5032 1263
Relative
Flux
A
B
A-B
Figure 2. The H­band UKIRT spectrum of HS 1216+5032A. The next
lowest spectrum is that of HS 1216+5032B, scaled upward by a constant
factor of 4, after which the continuum shape and broad H# line wings of the
two components are identical. At the bottom is the residual spectrum, which
shows the strong narrow line emission absent in HS 1216+5032B.
Kingdom Infrared Telescope Facility (UKIRT) on MaunaKea, using
CGS4 with a 1.2­arcsec slit on to the 40 l mm -1 grating. The H­band
spectra were taken in first order, centred at 1.6 µm, with the B2 order
separating filter (>1.43 m), while the J­band spectra were obtained
in second order centred at 1.18 m, with the B1 filter (>0.99 m).
Total exposure times for the spectra are 80 min in the J band and 48
min in the H band from each night. We used the two­dimensional
(2D) spectra produced by the ORAC­DR pipeline (Economou et al.
2001). Our extraction and analysis uses standard procedures in the
IRAF 5 software package. The spectra have dispersions of 12 and 24 å
pixel -1 at J and H, respectively, or resolutions of about 600 and 900
km s -1 . To remove atmospheric absorption features, we obtained the
spectrum of the F7V star BS 4761. We divided the stellar spectrum
into the quasar spectrum, after each had been wavelength calibrated
and normalized to the same integration time. To flux calibrate, we
multiplied the resulting quotient by a blackbody spectrum having
an effective temperature of 6250 K and normalized to the V =
6.21 magnitude of BS 4761. Given that the slit may not have been
perfectly aligned on both components, we eschew detailed study of
the continuum, and here concern ourselves only with relative line
profiles.
In Fig. 2 we show the H­band UKIRT spectrum of HS
1216+5032A, along with that of HS 1216+5032B, scaled upward
by a constant factor of 4. The continuum shape and broad H# line
wings of the two components are identical. The redshifts we mea­
sure are z A = 1.4569 ± 0.0006 and z B = 1.4556 ± 0.0015, which
is consistent with no significant velocity difference. The remaining
equivalent widths are W # = 630 and W # = 350 å respectively (with
errors of ± 50 å dominated by choice of continuum placement).
A strong narrow emission­line component in HS 1216+5032A is
effectively absent in HS 1216+5032B, as can be seen from the
residual spectrum at the bottom.
The lack of a strong narrow H# component in HS 1216+5032B
is consistent with the fact that BALQSOs are known to have re­
duced narrow­line emission (Turnshek et al. 1997). More generally,
X­ray weak quasars such as BALQSOs have relatively weaker nar­
5 IRAF is distributed by the National Optical Astronomy Observatory, which
is operated by the Association of Universities for Research in Astronomy,
Inc., under cooperative agreement with the National Science Foundation.
Relative Flux [OIII]
##
##
Figure 3. The J­band UKIRT spectrum of HS 1216+5032A. H# and
[O III]# 5007 are clearly visible at 1.196 and 1.230 µm respectively. H#
can be seen at 1.069 µm. We derive a redshift of 1.456 from [O III]#5007.
The J spectrum of HS 1216+5032B had insufficient S/N to detect emission
lines.
row components to their emission lines (Green 1998). This has been
proposed to be due to absorbers blocking the ionizing flux from
reaching the larger narrow­line region (NLR). If the NLR is spa­
tially extended (over several hundred pc; i.e. Kraemer & Crenshaw
2000), then small sightline differences should cause little change to
total NLR flux. These differing H# profiles thus strongly discourage
the lens hypothesis.
Fig. 3 shows the J­band UKIRT spectrum of HS 1216+5032A.
For HS 1216+5032A, we derive a redshift of 1.456 from [O III]
#5007 and [O III]#4959. The J­band spectra of HS 1216+5032B
from UT 2003 March 20 had insufficient signal­to­noise ratio (S/N)
to detect any emission lines. We tried again with longer exposures
on UT 2003 April 20, but with only marginal improvement.
4 RAD I O PROPERT I ES
4.1 FIRST data
We searched for catalogued radio sources within 20 arcsec of HS
1216+5032A in the 20­cm VLA survey for FIRST. A source in
the FIRST catalogue at 12 h 18 m 40. s 462 +
50# 15 43. ## 30 is just 0.5
arcsec from the United States NavalObservatory B1.0(USNO­B1.0;
Monet et al. 2003) optical catalogue position of HS 1216+5032B. 6
A sequence of tiled 1­arcmin images of the field in photographic
B and R optical bands [Digitized Sky Survey (DSS) scans of the
Second Palomar Observatory Sky Survey (POSS­II)], J, H and K
near­IR bands [Two­Micron All­Sky Survey (2MASS)] and 20­cm
radio bands (VLA FIRST) are shown in Fig. 4. The integrated radio
flux of HS 1216+5032B is 3.92 ± 0.143 mJy. This is identical to
the peak flux in the beam, indicating that it is consistent with a point
source at the #5 arcsec spatial resolution of the FIRST. This flux
corresponds at z = 1.455 to a (log) radio luminosity of 32.53 erg
s -1 Hz -1 . In contrast, no source is catalogued at the position of
HS 1216+5032A, nor is one evident in the image. The rms of the
6 Source positions derived from FIRST images have an error of 0.5 arcsec
rms for the weakest discernible point sources (#0.75 mJy) and substantially
smaller uncertainties for brighter sources. McMahon et al. (2002) report
FIRST positional uncertainties of <1 arcsec (radius of 90 per cent confi­
dence). The quoted position errors for the USNOB1.0 sources are about 0.1
arcsec.
C
# 2004 RAS, MNRAS 349, 1261--1266

1264 P. J. Green et al.
B R J
H K 20cm
B
A
Figure 4. A mosaic of multiwavelength images of the HS 1216+5032 field
gathered from the web. Each image is 1 arcmin on a side, with 4­arcsec
diameter circles marking the USNO­B1.0 positions for the two components.
The flux ratios in the DSSII B and R images, and those in the 2MASS J­,
H­ and K­band images are all roughly consistent with HS 1216+5032A,
being brighter by a factor of 5--6. In the FIRST 20­cm (1.4 GHz) image, HS
1216+5032B is at least five times brighter.
background in the vicinity of HS 1216+5032 is 0.16 mJy, so that
a 5# upper limit for point source detection corresponds to 0.8 mJy,
or log L R < 31.83 erg s -1 Hz -1 .
4.2 Radio­loudness differences
Combining optical with radio fluxes, we can constrain the radio­
loudness log R, which is the log ratio of the emitted monochromatic
luminosity at 1.4 GHz and 4410 å, log R # log(L 1.4 GHz /L 4410 å ).
To the observed B mag from Section 2, we apply a correction for
extinction. 7 We assume an intrinsic optical continuum f # # # -0.5 ,
with specific optical normalization from Marshall et al. (1984).
We derive log R = 1.74, with a conservative error estimate of
±0.4. 8 A radio­loudness parameter greater than unity is typically the
criterion for a radio­loud classification, and HS 1216+5032B surely
qualifies. 9 This marks HS 1216+5032B as one of a handful of bona
fide radio­loud BALQSOs. Although BALQSOs were originally
thought to be radio quiet as a rule (Stocke et al. 1992), by now
more than 50 BAL quasars have been identified via radio selection,
primarily from the FIRST Bright Quasar Survey (FBQS; Becker
et al. 2000; Menou et al. 2001).
The limit we derive for HS 1216+5032A is log R < 0.20, so the
radio­loudness difference between A and B is significant at >3# .
5 D I SCUSS I ON
5.1 No hope for the lens hypothesis
As we mention in Section 2, variability or sightline differences
might conceivably explain some of the large optical spectral and
7 We use E(B--V ) = 0.0184 mag from the maps of Schlegel, Finkbeiner &
Davis (1998), and assume an R V = 3.1 absorbing medium, with A(#)/A(B)
from Cardelli, Clayton, & Mathis (1989)
8 This error estimate allows for the worst­case combination of 1­# photo­
metric errors with errors ±0.2 in the assumed continuum slopes.
9 A different definition of radio loudness, which divides QSOs into two
populations at a radio luminosity of about 10 25 WHz -1 , does not affect any
of the results we present here.
colour differences seen between the two image components of HS
1216+5032. Could sightline differences or variability explain the
differing radio fluxes?
We first examine the properties of known lenses and lens candi­
dates to see if radio flux ratios of #5 are common. The Cosmic Lens
All­Sky Survey (CLASS) requires a double detection and a com­
ponent flux density ratio of <10 as a selection criterion (Browne
et al. 2003). Four out of seven pairs in the CLASS statistically
well­defined sample have flux ratios above 5. In a lens scenario,
we may choose to assume that the optical/near­IR flux ratio of #5
represents the ratio of the magnification factors due to the lens. The
difference in radio flux in the opposite sense by another factor of #5
requires a factor of #25 intrinsic radio flux difference to be due to the
9.1­arcsec sightline shift. If the jet is Doppler boosted, the observed
flux f # is related to the source rest­frame flux f #
via
f # = # D
1 + z # 3+#
f #
#
where # is the radio spectral index and D is the Doppler boosting
factor
D =
1
# (1 - # cos # )
(Rybicki & Lightman 1979). Even assuming a very fast jet (# =
0.9), for any line­of­sight angle # to the jet direction, the maximum
change in Doppler boosting achieved by a 9­arcsec change in # is
a factor of about 2 â 10 -4 , so nowhere near capable of explaining
the observed radio flux ratio.
Some radio­quiet or radio­intermediate objects flare by factors of
#30 at high frequencies (i.e. III Zw2 at 22 GHz; Brunthaler et al.
2003). Among flat spectrum (typically core­dominated) quasars
(#40 per cent) variability of radio sources on scales of 1--5 yr is com­
mon. The spectral and variability characteristics of such quasars are
very similar (Valtaoja et al. 1992) to those of BL Lacs. However,
variability rarely exceeds about 20 per cent -- at 151 MHz (Riley
1993) or 1.4 GHz (Rys & Machalski 1990) -- and few variations of
>5â are known. Therefore, variability combined with time delays
is very unlikely to account for the differing radio loudness of HS
1216+5032A and B.
5.2 The interaction/merger hypothesis
Because the number of WSQPs is #100â that expected from simple
extrapolations of the quasar--quasar correlation function, galaxy in­
teractions may be important in creating binary quasars (Kochanek,
Falco & Munoz 1999; Mortlock, Webster & Francis 1999) and can
be used as a tool to study the triggering of nuclear activity in galaxies
(Osterbrock 1993).
Because it appears most convincingly from the radio­loudness
differences that HS 1216+5032 is not lensed, the two quasars are
separated by at least their projected separation of #65 h 70 kpc.
HS 1216+5032 could represent a high­z example of interaction­
triggered but as­yet unmerged luminous AGN. Approach velocities
less than the internal host galaxy velocities induce significant angu­
lar momentum loss (Fang & Saslaw 1997) and merging over 6--10
crossing times. Assuming an encounter space velocity of 500 km
s -1 and 50­kpc separation implies that the last encounter occurred
#100 Myr ago in the QSO rest frame.
This would support the view of quasar activity as a short ac­
tive phase of supermassive black holes accreting at rate —
M #
1-100 M# yr -1 (Rees 1984). Cavaliere & Vittorini (2000), Menci
et al. (2003) and others have proposed that earlier than z #3,gas­rich
C
# 2004 RAS, MNRAS 349, 1261--1266

HS 1216+5032 1265
protogalaxies grow by merging, simultaneously inducing growth of
central holes accreting at their full Eddington rates. More recently
(for z < 3) the black holes are triggered by the encounters of a
gas­rich host with its companions in a group, which destabilize the
gas and induce accretion. These models link hierarchical structure
formation, the observed evolution of the quasar luminosity function
(space density peaking near z # 3), and the correlation between
central black hole and bulge masses in nearby galaxies (Ferrarese
& Merritt 2000; Gebhardt et al. 2000). Higher rates of interaction
and triggered activity in close pairs must be demonstrated to verify
the premises of these models.
The weak [O III] lines in the J­band spectrum of HS 1216+5032A
suggest that it has similarities to HS 1216+5032B. A principal com­
ponent analysis of low­redshift optical spectra reveals that the princi­
pal eigenvector of a quasar spectral sample (PC1, the linear combina­
tion of parameters that represents the largest variance in the sample
spectra) links the strength of Fe II emission, [O III] emission, and H#
line asymmetry (Boroson & Green 1992). In a common interpreta­
tion, objects with weak [O III]/H# ratios have high Eddington ratios
L/L Edd . The second principal eigenvector (PC2), dominated by lu­
minosity and its anticorrelation with the strength of He II#4686, is
presumed to be driven by accretion rate —
M . The J­band spectrum of
HS 1216+5032A is clearly on the low PC1, low PC2 end (cf. fig. 2
of Boroson 2002), which is the region also inhabited by BALQSOs
(their fig. 1). Only #10 per cent of quasars inhabit this region of
observational parameter space, which is proposed to map physically
to high L/L Edd and —
M . Because HS 1216+5032 is not lensed, then
these quasars probably have both high L and L/L Edd in common,
perhaps related to a common dense environment (i.e. a cluster) or to
the ongoing effects of their interaction. While the current example
is anecdotal, its properties should help guide directed searches for
other examples.
6 SUMMARY
The HS 1216+5032 quasar pair, particularly with evidence for an
intervening cluster potential, held out promise of being a dark lens
system. We obtained improved optical (MMT) and IR (UKIRT)
spectra of HS 1216+5032A and HS 1216+5032B, away from the
strongest (UV) effects of BALs, and measured the emission lines,
finding significant differences in the H# profiles. Optical B, V and
R colours measured at the WIYN telescope confirm different con­
tinuum shapes.
From FIRST data, we confirm that HS 1216+5032B is radio loud,
to a degree that we find is not plausibly explained in a lens + time­
delay scenario by either variability or line­of­sight differences. The
HS 1216+5032 pair is thus certainly a physical quasar binary rather
than a lensed system. The system remains interesting because there
are only a handful of bona fide radio­loud BALQSOs known to date,
and because BALQSOs have been proposed to be high Eddington
ratio accretors that are triggered by interactions.
ACKNOWLEDGMENTS
Thanks to Chris Kochanek, Dan Schwartz and Bryan Gaensler for
useful comments, and to Deborah Freedman for attempting imaging
for us at the F. L. Whipple Observatory (FLWO) 1.2­m on Mt Hop­
kins. We thank the referee, Paul Hewett, for advice and pruning.
This work was supported by CXO grant GO2­3132X and NASA
grant NAS8­39073. PJG and TLA gratefully acknowledge support
through NASA Contract NAS8­39073 (CXC).
The IR spectra were obtained as part of the UKIRT Service Pro­
gramme. UKIRT is operated by the Joint Astronomy Centre on
behalf of the UK Particle Physics and Astronomy Research Coun­
cil. ORAC­DR, the UKIRT data reduction pipeline, was developed at
the Joint Astronomy Centre by Frossie Economou and Tim Jenness
in collaboration with the UK Astronomy Technology Centre.
Optical spectra were obtained by WRB at the MMT Observatory,
a joint facility of the Smithsonian Institution and the University
of Arizona. WIYN photometry was obtained at Kitt Peak, part of
the National Optical Astronomy Observatory, operated by the As­
sociation of Universities for Research in Astronomy, Inc., under
cooperative agreement with the National Science Foundation.
The DSS was produced at the Space Telescope Science Insti­
tute under US Government grant NAGW­2166. The images of these
surveys are based on photographic data obtained using the Oschin
Schmidt Telescope on Palomar Mountain and the UK Schmidt Tele­
scope. The plates were processed into the present compressed digi­
tal form with the permission of these institutions. The POSS­II was
made by the California Institute of Technology with funds from the
National Science Foundation, the National Aeronautics and Space
Administration, the National Geographic Society, the Sloan Foun­
dation, the Samuel Oschin Foundation, and the Eastman Kodak
Corporation.
This research has made use of the USNOFS Image and
Catalogue Archive operated by the USNO, Flagstaff Station
(http://www.nofs.navy.mil/data/fchpix/).
REFERENCES
Aldcroft T. L., Green P. J., 2003, ApJ, 592, 710
Barlow T. A., Junkkarinen V. T., Burbidge E. M., Weymann R. J., Morris S.
L., Korista K. T., 1992, ApJ, 397, 81
Becker R. H., White R. L., Gregg M. D., Brotherton M. S., Laurent­
Muehleisen S. A., Arav N., 2000, ApJ, 538, 72
Boroson T. A., 2002, ApJ, 565, 78
Boroson T. A., Green R. F., 1992, ApJS, 80, 109
Browne I. W. A. et al., 2003, MNRAS, 341, 13
Brunthaler A., Falcke H., Bower G. C., Aller M. F., Aller H. D., Ter˜asranta
H., Krichbaum T. P., 2003, PASA, 20, 126
Cardelli J. A., Clayton G. C., Mathis J. S., 1989, ApJ, 345, 245
Cavaliere A., Vittorini V., 2000, ApJ, 543, 599
Chelouche D., 2003, ApJ, 596, L43
Economou F., Jenness T., Cavanagh B., Wright G. S., Bridger A. B., Kerr
T. H., Hirst P., Adamson A. J., 2001, in Harnden F. R. Jr., Primini F. A.,
Payne H. E., eds, ASP Conf. Ser. Vol. 238, Astronomical Data Analysis
and Systems X. Astron.