Content uploaded by Francois Hammer
Author content
All content in this area was uploaded by Francois Hammer
Content may be subject to copyright.
Content uploaded by Francois Hammer
Author content
All content in this area was uploaded by Francois Hammer
Content may be subject to copyright.
THE ASTROPHYSICAL JOURNAL, 499:112È133, 1998 May 20
1998. The American Astronomical Society. All rights reserved. Printed in U.S.A.(
HUBBL E SPACE T EL ESCOPE IMAGING OF THE CFRS AND LDSS REDSHIFT SURVEYS.
I. MORPHOLOGICAL PROPERTIES1
JARLE ROBERTO DAVID LAURENCE RICHARD S. SIMONBRINCHMANN,2ABRAHAM,2SCHADE,3TRESSE,2ELLIS,2LILLY,2,4
OLIVIER KARL MATTHEWLEFE`VRE,5,7 GLAZEBROOK,6FRANCÓOIS HAMMER,7COLLESS,8
DAVID AND TOMCRAMPTON,3BROADHURST9
Received 1997 August 8; accepted 1998 January 5
ABSTRACT
We analyze Hubble Space Telescope images of a complete sample of 341 galaxies drawn from the
CFRS and LDSS ground-based redshift surveys. In this, the Ðrst paper in the series, each galaxy has
been morphologically classiÐed according to a scheme similar to that developed for the Medium Deep
Survey. We discuss the reproducibility of these classiÐcations and quantify possible biases that may arise
from various redshift-dependent e†ects. We then discuss automated classiÐcations of the sample and con-
clude, from several tests, that we can expect an apparent migration with redshift to later Hubble types
that corresponds to a misclassiÐcation in our adopted machine classiÐcation system of D24% ^11 of
the true ““ spirals ÏÏ as ““ peculiars ÏÏ at a redshift z^0.9. After allowing for such biases, the redshift dis-
tribution for normal spirals, together with their luminosity function derived as a function of redshift,
indicates approximately 1 mag of luminosity evolution in by z^1. The elliptical sample is too smallBAB
for precise evolutionary constraints. However, we Ðnd a substantial increase in the proportion of galaxies
with irregular morphology at large redshift from 9% ^3% for 0.3 ¹z¹0.5 to 32%^12% for
0.7 ¹z¹0.9. These galaxies also appear to be the dominant cause of the rapid rise with redshift in the
blue luminosity density identiÐed in the redshift surveys. Although galaxies with irregular morphology
may well comprise a mixture of di†erent physical systems and might not correspond to present-day
irregulars, it is clear that the apparently declining abundance and luminosities of our distant
““ irregulars ÏÏ holds an important key to understanding recent evolution in the star formation history of
normal galaxies.
Subject headings: galaxies: fundamental parameters È galaxies : structure È surveys
1.INTRODUCTION
The refurbished Hubble Space Telescope (HST ) o†ers the
exciting prospect of addressing the morphological evolution
of galaxies directly from systematic studies of galaxies
imaged at various redshifts. The angular resolution of the
Wide Field Planetary Camera 2 (WFPC2) is which0A.1,
corresponds to a physical scale of less than 2 kpc at all
redshifts in most popular cosmological world models (we
have assumed km s~1 Mpc~1 here and will adoptH0\50
this value and in the rest of the paper).q0\0.5
Considerable progress in understanding the morphologi-
cal mixture of the faint-galaxy population has already been
achieved through the Medium Deep Survey (MDS;
et al. Windhorst, & GriffithsGrifÐths 1994 ; Driver, 1995a),
an extensive imaging program using WFPC2 in parallel
mode. Counts of galaxies classed by visual morphology
1Based on observations with the NASA/ESA Hubble Space Telescope
obtained at the Space Telescope Science Institute, which is operated by the
Association of Universities for Research in Astronomy, Inc., under NASA
contract NAS 5-26555.
2Institute of Astronomy, Madingley Road, Cambridge CB3 0HA,
England.
3Dominion Astrophysical Observatory, Victoria, Canada.
4Department of Astronomy, University of Toronto, Toronto, Canada.
5Laboratoire dÏAstronomie Spatiale, Traverse du Siphon, B.P.8, 13376
Marseille Cedex 12, France.
6Anglo-Australian Observatory, Siding Spring Observatory, Coona-
barabran, NSW 2357, Australia.
7Observatoire de Paris, Section de Meudon, DAEC, 92195 Meudon
Principal Cedex, France.
8Mount Stromlo and Siding Spring Observatories, Australian Nation-
al University, Weston Creek, ACT 2611, Australia.
9Astronomy Department, University of California, Berkeley, CA
94720.
et al. et al. and by other(Glazebrook 1995a; Driver 1995a)
means et al. et al. have(Abraham 1996b; Odewahn 1996)
been compared to model predictions, and an apparent
excess of ““ irregular/peculiar/merger ÏÏ galaxies is noted
when compared to models based on no evolution. Deeper
HST images of smaller areas have been taken using
WFPC2 in primary mode. et al. haveDriver (1995b)
analyzed a single deeper pointed exposure of 5.7 hr, con-
Ðrming and extending the MDS analysis to I\24.5.
et al. have likewise categorized galaxies inAbraham (1996b)
the Hubble Deep Field to I\25. Given the magnitude
limits over which these changes are seen, the above studies
point to fairly recent changes in the morphological charac-
teristics of the galaxy population.
Similar progress has been made from large systematic
ground-based spectroscopic surveys that serve to delineate
the Ðeld galaxy luminosity function (LF) and its evolution
out to redshifts z^1. The I-selected Canada-France Red-
shift Survey (CFRS; et al. Fe` vre et al.Lilly 1995a ; Le 1995,
and references therein) comprises a complete spectroscopic
sample of 591 galaxies in the magnitude range 17.5 ¹
with determined redshifts. The rest-frameIAB¹22.5 BAB-
LF has been determined for various redshifts andband
color-selected components of the population. Strong evolu-
tion with redshift is found in the luminosity function of the
bluer galaxies. These trends are supported by those from
the less deep but also extensive AutoÐb/Low-Dispersion
Survey Spectrograph (hereafter LDSS) Redshift Survey
et al. and references therein). This survey is(Ellis 1996 bJ-
and spans a wide apparent magnitude rangeselected enabling the shape of the star-forming(11.5 \bJ\24),
component of the LF to be monitored to z\0.75. Strong
112
CFRS AND LDSS SURVEY GALAXIES 113
evolution is seen in terms of the space density and lumi-
nosity of galaxies with intense star formation categorized
via their [O II] emission. Using the same data, et al.Heyl
showed that it is the late spectral types that dominate(1997)
the evolution out to z^0.5. Both surveys are consistent
with a large decline in the luminosity density of star forming
galaxies since a redshift of 1 et al. et al.(Lilly 1996; Ellis
1996).
The HST imaging and ground-based spectroscopic
surveys present di†erent but complementary views of the
evolving galaxy population since z^1. Indeed, it is tempt-
ing to connect the rapid increase with look-back time in the
proportion of galaxies in the irregular/peculiar/merger cate-
gory with the strong evolution seen in the star-forming blue
sources in the redshift surveys. However, until recently there
has been surprisingly little overlap between the wealth of
HST data and the ground-based redshift surveys, largely
because of the mismatch in Ðeld size between WFPC2 and
ground-based multiobject spectrographs.
In this series of papers we plan to remedy this deÐciency
via a HST imaging program of 341 galaxies targeted in
either the CFRS or LDSS surveys. The goals of the study
are to employ techniques developed in the analysis of both
the HST and ground-based data sets to understand physi-
cally the origin of the remarkably recent evolutionary
trends identiÐed in the independent data sets.
This paper is concerned with describing the survey
parameters and selection criteria and the techniques used to
analyze the HST data. Further details of the ground-based
spectroscopic data sets can be found in the CFRS and
LDSS articles et al. et al. By(Crampton 1995 ; Ellis 1996).
bringing together HST and redshift data for a large com-
plete sample of distant galaxies, we address in this paper the
question of whether the rapid evolution in the morphologi-
cally peculiar population can be identiÐed with the star-
forming blue sources in the redshift surveys. As shown
below, the correspondence is convincing, as originally con-
jectured by the MDS studies.
A plan of the paper follows. In we discuss the basic°2
features of the LDSS and CFRS redshift surveys. As the
photometric systems and the treatment of k-corrections
di†er between the two surveys, we compare and align these
prior to further analysis. The HST imaging data are intro-
duced in and both visual and automated morphological°3,
classiÐcations are presented. A major question is the extent
to which the apparent morphological type recognized in the
HST data is a†ected by redshift-dependent biases. Through
simulations based on local multicolor imaging data, we
address this point in detail in and obtain statistical°4
correction factors that are applied in the subsequent
analyses. presents the redshift distributions andSection 5
luminosity functions for three broad morphological classes.
We also discuss the associated blue luminosity density for
each type as a function of redshift and interpret this in the
context of simple models put forward to explain the evolu-
tionary trends found in the redshift survey data. Our main
conclusions are summarized in °6.
Later papers in the series use structural parameters for
each HST image to discuss the physical processes that drive
this evolution. et al. hereafter examineLilly (1998, Paper II)
the surface brightness characteristics of the largest disk gal-
axies in order to constrain the extent to which evolution of
massive galaxies may be important. et al.Schade (1997,
hereafter address the question of evolution in thePaper III)
number density and photometric properties of the spher-
oidal population. Papers and both concentrate onII III
quantitative morphological measures for the galaxies,
extending and improving the analysis in et al.Schade
Fe` vre et al. hereafter perform(1995). Le (1998, Paper IV)
quantitative measures of clustering on small physical scales
and use this to study the rate of merging as a function of
redshift.
2.SURVEY DESCRIPTION
Upon completion of the CFRS and LDSS redshift
surveys, both CFRS and LDSS teams independently sought
HST WFPC2 time in Cycle 4. Early results from these
programs have been described by et al. andSchade (1995)
From 1994 the two teams agreed to merge theirEllis (1995).
e†orts, and further allocations of HST time to the com-
bined team were made in Cycles 5 and 6.
In Cycle 4, prior to the merged e†ort, the strategy
adopted by two teams was somewhat di†erent. Although
both teams sought F814W imaging for morphological clas-
siÐcations, the CFRS team chose to supplement these
images with ones in F450W, whereas the LDSS team
explored the visibility of their samples in F336WbJ-selected
and F218W. Only in Cycle 5 and 6 did a common strategy
emerge based on F814W images augmented by F450W
images for a subset of the targets.
It is important to recognize that the ground-based stra-
tegies adopted by the CFRS and LDSS survey teams dif-
fered in several respects. Foremost is the fact that the CFRS
survey is I-selected, whereas the LDSS survey is bJ-selected.
The strong di†erential e†ects of k-correction with galaxy
type mean that, although the LDSS survey is shallower, it is
more sensitive to the presence of star-forming galaxies that
apparently dominate the evolutionary trends. By contrast,
the CFRS survey is less a†ected by k-correction e†ects
overall and probes to higher redshift. Both teams chose to
present their data in terms of rest-frame blue magnitudes,
although CFRS presented LFs on a photometric scale
based on B(AB), whereas the LDSS team did so in ThebJ.
transformation between these two magnitude scales is
studied below.
summarizes the HST survey data obtained forTable 1
this analysis. The total data set consists of 341 objects
drawn from 25 WFPC2 HST frames. Six Ðelds are drawn
from the contiguous ““ Groth strip ÏÏ et al.(Groth 1994)
imaged by WFPC2, for which supplementary spectroscopy
of 59 objects was obtained to a magnitude limit identical to
that adopted for the main CFRS survey using LDSS-2 on
the 4.2 m William Herschel Telescope. Galaxies have only
been included in the Ðnal catalog if they have been spectro-
scopically targeted in either the CFRS, LDSS, or Groth
strip surveys. Thirty-seven objects were excised, as their
HST images are partially or completely obscured by edges
of the WFPC chips or cosmetic defects, precluding accurate
morphological analysis. The catalog contains seven objects
from the AutoÐb Ðbre survey et al. as well as 22(Ellis 1996),
LDSS-1 et al. and references therein) objects.(Colless 1993
The spectroscopic completeness for these objects is only
62%, and they will not be discussed further in the present
paper, since little is gained in terms of numbers. We will
instead concentrate on galaxies from the CFRS and
LDSS-2 et al. surveys. The absolute(Glazebrook 1995b)
magnitudeÈredshift distributions of the objects from the
CFRS and LDSS-2 samples and the subsamples for which
114 BRINCHMANN ET AL. Vol. 499
TABLE 1
HST IMAGING SURVEY
Integration Integration
Time Time
FieldaRed Filter (s) NCOMBbBlue Filter (s) NCOMBbNtargetscNzdNstare
cfrs–03–1 ...... F814W 6700 5 . . . . . . . . . 14 13 0
cfrs–03–2 ...... F814W 6700 5 . . . . . . . . . 15 14 0
cfrs–03–3 ...... F814W 6700 5 . . . . . . . . . 13 10 0
cfrs–03–4 ...... F814W 6400 6 F450W 6600 6 11 11 0
cfrs–03–5 ...... F814W 6400 6 F450W 6600 6 5 5 0
cfrs–10–1 ...... F814W 6700 5 . . . . . . . . . 17 16 1
cfrs–10–2 ...... F814W 6700 5 . . . . . . . . . 12 11 1
cfrs–10–3 ...... F814W 5302.5 4 . . . . . . . . . 24 20 2
cfrs–14–1 ...... F814W 7400 6 F450W 7800 6 21 21 0
cfrs–14–2 ...... F814W 7400 6 F450W 7800 6 13 10 3
cfrs–22–1 ...... F814W 6700 5 . . . . . . . . . 8 7 0
cfrs–22–2 ...... F814W 6700 5 . . . . . . . . . 14 12 0
cfrs–22–3 ...... F814W 6700 5 . . . . . . . . . 15 13 1
grth–14–1...... F814W 4400 4 F606W 2800 4 36 31 4
grth–14–2...... F814W 4400 4 F606W 2800 4 16 12 4
grth–14–3...... F814W 4400 4 F606W 2800 4 21 17 1
grth–14–4...... F814W 7400 6 F450W 7800 6 20 20 0
grth–14–5...... F814W 4400 4 F606W 2800 4 18 14 3
grth–14–6...... F814W 4400 4 F606W 2800 4 11 10 1
ldss–10a ....... F814W 5800 5 F380W 6000 6 13 10 1
ldss–10b ....... F814W 5800 5 F380W 6000 6 18 13 2
ldss–10c........ F814W 5800 5 F218W 6000 6 10 7 2
ldss–13a ....... F814W 5800 5 F380W 6000 6 13 6 2
ldss–13b ....... F814W 5800 5 F380W 6000 6 21 15 1
ldss–13c........ F814W 5800 5 F380W 6000 6 15 8 2
Totalf.......... 341 269 44 28
aHere, cfrs–03 \CFRS 3hÐeld; grth–14 \Groth 14hÐeld; ldss–10a \LDSS-2 10h.
bThe number of individiual exposures combined to form the Ðnal image.
cThe number of spectroscopic targets in the Ðeld.
dThe number of objects with reliable redshift measurement (note [1).
eThe number of stars.
fThere is some overlap between the di†erent frames, so the total is not the sum of the columns.
HST data are available are compared in It can beFigure 1.
seen that the two surveys span di†erent regions in this
parameter space. The CFRS survey is particularly useful in
probing all classes of luminous galaxies in the interval
0.5 \z\1, whereas the LDSS survey is e†ective in probing
less-luminous star-forming galaxies in the interval
0.2 \z\0.7. The complete catalog of galaxies comprising
the HST survey is given together with detailed comments in
Table 2.
All observed Ðelds were imaged through the F814W Ðlter,
which ensures that a self-consistent and uniform photo-
metric scale can be provided across both the CFRS and
LDSS galaxies. Only the CFRS survey galaxies currently
have reliable ground-based photometry. To achieve aIAB
uniform photometric scale, the raw HST images were pro-
cessed using the standard STScI pipeline, and photometry
in the F814W system was performed using the IRAF
APPHOT package with 3Adiameter apertures. The zero-
point calibration was taken from et al.Holtzman (1995).
The color transformation between and was cal-IAB IF814W
culated for the CFRS galaxies by interpolating within the
observed V[Icolors according to a set of spectral energy
distributions (SEDs); good agreement with the ground-
based magnitudes was found.
As it was not possible to image all the ground-based
Ðelds in both surveys within the HST time allocated, the
LDSS group initially selected their Ðelds on the basis of
maximizing the fraction of targets for which redshifts had
been secured, whereas the CFRS group originally imaged
Ðelds to maximize the fraction of high-redshift objects. After
the two groups joined forces, it was agreed that no particu-
lar criterion should be used to select the remaining Ðelds.
We have, retrospectively, veriÐed that the earlier selection
criteria have not unduly weighted the HST survey to an
unrepresentative sample of galaxies.
The completeness in the original ground-based surveys
varies from Ðeld to Ðeld primarily because of the vagaries of
weather at the time of the original observations. The com-
pleteness statistics are given in for the subset ofTable 3
galaxies in the survey that were either in the CFRS or
LDSS-2 redshift surveys. The completeness is only margin-
ally higher than that appropriate for the parent survey.
shows the redshift distribution separately for theFigure 2
CFRS and LDSS-2 objects in the HST survey. Clearly a
greater fraction of the CFRS galaxies have been selected for
study with HST than is the case for the LDSS survey. The
median redshift for the HST -selected CFRS objects,
SzT\0.61, can be compared to SzT\0.56 for the entire
CFRS ground-based survey. For the LDSS-2 objects the
di†erence is similarly smallÈSzT\0.46 in the ground-
based survey, compared to SzT\0.43 in the HST imaged
subset.
Clearly, it is important to construct a uniform absolute-
magnitude scale across the two surveys. This is an impor-
tant problem, not only because of observational di†erences
in the photometric selection criteria used by the two groups,
but also because of procedural di†erences used in estimat-
ing the k-corrections. To check the photometric di†erences,
we transformed the LDSS-2 photometry to thebJBAB
system used by the CFRS. The required color wasbJ[BAB
No. 1, 1998 CFRS AND LDSS SURVEY GALAXIES 115
FIG. 1.ÈAbsolute magnitudeÈredshift distribution for the combined CFRS and LDSS-2 surveys (assuming kms s~1 Mpc~1). Open symbols:H0\50
CFRS sample, limited at I(AB) \22.5; Ðlled symbols: LDSS-2 survey, limited at circles: objects in the HST survey.bJ\24;
found by using the LDSS-2 to deÐne an SED frombJ[rF
which the color o†set was located by interpolation. A histo-
gram of the o†sets obtained in this manner is shown in the
top panel of where it is seen that the o†set andFigure 3,
scatter are both quite small compared to the bin widths that
we will use in discussing evolutionary trends, e.g., in the LF.
To investigate systematic di†erences in the k-corrections
used between the two groups, we compared the k-
corrections derived for a subset of CFRS galaxies using
their set of SEDs located via the V[Icolor to those deÐned
similarly based on the LDSS-2 set of SEDs. The agreement
is again surprisingly good, indicating a dispersion of only
and no systematic o†set of signiÐcance middle0m. 1 (Fig. 3,
panel).
There is one further procedural di†erence concerning the
surveys. In the case of the CFRS survey, the absolute mag-
nitudes are derived from isophotal magnitudes, whereasIAB
the LDSS-2 group used corrected aperture magnitudes. To
determine the o†sets involved, we note that the 3Aaperture
FIG. 2.ÈRelative contributions of the LDSS-2 and CFRS redshift
surveys to the total redshift distribution for the sample used in the analysis.
HST magnitudes for the faint CFRS objects agree well with
their ground-based isophotal magnitudes. We therefore
measured 3Aaperture magnitudes for the LDSS-2 objects
on the HST images and calculated absolute magnitudes
from these that were then compared to the published
LDSS-2 absolute magnitudes. This comparison is shown in
the bottom panel of and reveals no signiÐcant shiftFigure 3
with a dispersion of only enabling us to conclude that0m.2,
the LDSS-2 corrected aperture magnitudes are closely
equivalent to the CFRS isophotal magnitudes. TheIAB
photometric systems can thus be aligned by adding the
color term to the LDSS absolute magnitudes.
In conclusion, within the typical photometric error of
there is no evidence of a serious systematic shift0m.2,
between the two absolute-magnitude scales. For consis-
tency, all absolute magnitudes for galaxies drawn from both
surveys were calculated with the same program, using IAB
isophotal magnitudes for CFRS objects and 3AI-aperture
magnitudes from HST for the LDSS objects. In the sub-
sequent analyses virtually identical results are obtained
when the published absolute magnitudes are used.
3.CLASSIFICATIONS
To interpret the evolution of the galaxy LF physically as
delineated by the original ground-based redshift surveys,
both the CFRS and LDSS analyses subdivided the samples
on the basis of spectroscopic and photometric classes.
CFRS analyses of the LF based on rest-frame color found
luminosity evolution to be stronger for galaxies bluer than
an Sbc. LDSS samples selected according to the rest-frame
equivalent width of the [O II] 3727 line found that the
evolutionary trends arose almost exclusively from galaxies
with strong emission lines.
The availability of the HST data for 341 galaxies allows
us to investigate these changes in more detail. Ideally, gal-
axies should be classiÐed according to a label that is not
modiÐed by any of the physical processes responsible for
TABLE 2
DATA FOR OBJECTS IN THE SURVEY
AC EW*OII+
ID zF814W MBAB NoteaClassbAcCcClassd(1 p)eOrigin Old IDf
03.0035 ...... 0.880 21.50 [22.06 1 4 0.080 0.319 4 . . . CFRS . . .
03.0315 ...... 0.223 20.54 [18.86 4 5 0.079 0.257 4 46(3) CFRS . . .
03.0316 ...... 0.815 22.26 [20.91 3 4 0.081 0.202 4 12(2) CFRS . . .
03.0321 ...... ... 21.87 . . . 0 2 0.123 0.418 1 . . . CFRS . . .
03.0327 ...... 0.606 21.86 [20.34 3 6 0.119 0.349 4 27(4) CFRS . . .
03.0332 ...... 0.188 21.92 [17.62 4 4 0.065 0.216 4 14(3) CFRS . . .
03.0337 ...... 0.360 22.31 [18.86 3 1 0.085 0.474 1 22(10) CFRS . . .
03.0346 ...... ... 21.79 . . . 0 4 0.165 0.418 4 . . . CFRS . . .
03.0358 ...... 0.088 17.31 [19.95 3 2 0.416 0.550 1 . . . CFRS . . .
03.0365 ...... 0.219 19.32 [20.08 4 3 0.060 0.449 1 32(5) CFRS . . .
03.0384 ...... ... 21.63 . . . 0 6 0.237 0.441 6 . . . CFRS . . .
03.0443 ...... 0.118 19.55 [19.50 4 1 0.103 0.564 1 . . . CFRS . . .
03.0445 ...... 0.530 20.80 [21.55 3 5 0.060 0.220 4 10(1) CFRS . . .
03.0466 ...... 0.534 23.14 [19.59 3 4 0.034 0.198 4 29(4) CFRS . . .
03.0480 ...... 0.608 22.24 [20.32 3 4 . . . . . . 4 99(13) CFRS . . .
03.0485 ...... 0.606 21.53 [20.24 4 6 0.148 0.146 6 78(16) CFRS . . .
03.0488 ...... 0.607 21.43 [20.86 4 6 0.276 0.100 6 66(11) CFRS . . .
03.0523 ...... 0.651 21.28 [21.22 4 6 0.313 0.382 4 38(10) CFRS . . .
03.0528 ...... 0.714 21.36 [21.23 3 3 0.058 0.370 4 . . . CFRS . . .
03.0560 ...... 0.697 21.34 [21.06 3 2 0.079 0.425 1 11(1) CFRS . . .
03.0579 ...... 0.660 22.12 [20.47 2 4 0.107 0.236 4 5(5) CFRS . . .
03.0595 ...... 0.606 21.57 [20.80 4 8 0.276 0.133 6 17(1) CFRS . . .
03.0599 ...... 0.480 21.19 [20.63 4 5 0.111 0.158 4 41(5) CFRS . . .
03.0717 ...... 0.607 20.93 [21.18 3 5 0.048 0.254 4 4(4) CFRS . . .
03.0982 ...... 0.195 21.30 [18.12 4 2 0.070 0.429 1 34(2) CFRS . . .
03.0983 ...... 0.370 21.03 [19.90 3 5 0.091 0.275 4 0(5) CFRS . . .
03.0992 ...... 0.262 22.74 [17.46 2 3 0.149 0.271 4 . . . CFRS . . .
03.0999 ...... 0.704 21.49 [21.39 3 5 0.308 0.233 6 11(2) CFRS . . .
03.1014 ...... 0.197 19.32 [20.84 3 5 0.068 0.232 4 6(5) CFRS . . .
03.1016 ...... 0.705 22.38 [20.47 3 6 0.145 0.161 6 85(10) CFRS . . .
03.1027 ...... 1.038 22.06 [21.64 39 8 0.178 0.239 4 107(7) CFRS . . .
03.1031 ...... 0.422 20.62 [20.12 3 2 0.105 0.568 1 0(12) CFRS . . .
03.1032 ...... 0.618 20.33 [21.53 4 1 0.110 0.593 1 15(1) CFRS . . .
03.1034 ...... ... 22.21 . . . 0 2 0.057 0.266 1 . . . CFRS . . .
03.1035 ...... 0.635 21.25 [20.84 3 3 0.095 0.435 1 5(2) CFRS . . .
03.1050 ...... 0.264 21.49 [18.83 4 4 0.068 0.211 4 67(24) CFRS . . .
03.1051 ...... 0.155 21.01 [18.16 4 5 0.045 0.272 4 29(8) CFRS . . .
03.1056 ...... 0.944 21.97 [21.23 3 6 0.130 0.303 4 85(2) CFRS . . .
03.1060 ...... 0.480 20.72 [20.39 3 3 0.095 0.431 1 6(4) CFRS . . .
03.1077 ...... 0.938 21.57 [22.86 3 1 0.143 0.372 4 0(3) CFRS . . .
03.1319 ...... 0.620 21.62 [20.61 4 1 0.171 0.425 1 32(7) CFRS . . .
03.1347 ...... 0.562 20.60 [21.45 3 3 . . . . . . 4 15(2) CFRS . . .
03.1373 ...... 0.482 20.73 [20.42 3 3 0.123 0.536 1 0(5) CFRS . . .
03.1375 ...... 0.637 22.12 [20.38 3 5 0.130 0.188 6 24(2) CFRS . . .
03.1381 ...... 0.636 20.23 [22.02 4 0 0.074 0.551 1 0(5) CFRS . . .
03.1384 ...... 0.785 21.61 [21.22 2 1 0.101 0.414 1 54(6) CFRS . . .
03.1387 ...... 0.222 20.72 [18.60 1 [1 0.174 0.622 1 . . . CFRS . . .
03.1392 ...... 0.605 20.89 [21.22 4 4 0.070 0.389 1 0(3) CFRS . . .
03.1393 ...... 0.852 22.24 [21.21 9 5 0.108 0.220 4 0(3) CFRS . . .
03.1395 ...... 0.708 21.79 [20.55 2 3 0.092 0.340 4 7(0) CFRS . . .
03.1413 ...... 0.487 20.64 [20.42 2 2 0.091 0.454 1 0(7) CFRS . . .
03.1416 ...... 0.488 20.22 [21.20 4 1 0.075 0.484 1 2(2) CFRS . . .
03.1426 ...... ... 22.00 . . . 0 6 0.098 0.233 6 . . . CFRS . . .
03.1499 ...... 0.827 21.62 [21.41 93 5 0.071 0.292 4 . . . CFRS . . .
03.1531 ...... 0.715 22.06 [20.70 3 6 0.093 0.134 6 38(5) CFRS . . .
03.1540 ...... 0.690 21.04 [21.56 3 4 0.202 0.286 4 18(2) CFRS . . .
03.1650 ...... 0.637 21.86 [20.17 3 5 0.092 0.169 6 22(6) CFRS . . .
03.9003 ...... 0.619 20.88 [21.16 4 5 0.283 0.201 6 50(2) CFRS . . .
10.0747 ...... 0.340 20.40 [20.19 3 4 0.059 0.281 4 . . . CFRS . . .
10.0761 ...... 0.983 22.16 [21.63 3 6 0.061 0.171 4 8(1) CFRS . . .
10.0763 ...... 0.671 21.16 [21.85 3 5 0.109 0.232 4 11(2) CFRS . . .
10.0765 ...... 0.536 22.35 [19.93 4 6 . . . . . . 6 60(6) CFRS . . .
10.0769 ...... 0.669 21.17 [21.28 4 2 0.092 0.306 4 4(1) CFRS . . .
10.0771 ...... 0.787 22.76 [20.58 8 6 0.116 0.102 6 70(14) CFRS . . .
10.0793 ...... 0.577 21.31 [20.86 4 4 0.196 0.289 4 70(9) CFRS . . .
10.0794 ...... 0.580 21.16 [20.50 3 1 0.124 0.583 1 0(5) CFRS . . .
10.0802 ...... 0.309 21.88 [19.28 4 6 0.293 0.147 6 47(4) CFRS . . .
10.0805 ...... 0.147 21.78 [17.44 4 4 0.076 0.164 6 . . . CFRS . . .
10.0811 ...... 0.738 21.37 [21.60 93 4 0.102 0.223 4 . . . CFRS . . .
10.0812 ...... 0.385 20.10 [20.57 4 2 0.048 0.496 1 0(7) CFRS . . .
10.0813 ...... 0.467 22.27 [19.45 4 4 0.118 0.210 4 70(14) CFRS . . .
10.0818 ...... 0.000 19.23 . . . 4 [2 ... ... [2 . . . CFRS . . .
116
TABLE 2ÈContinued
AC EW*OII+
ID zF814W MBAB NoteaClassbAcCcClassd(1 p)eOrigin Old IDf
10.0826 ...... 0.643 20.57 [21.95 3 5 0.135 0.232 4 7(2) CFRS . . .
10.0829 ...... 0.526 21.73 [20.24 93 6 0.071 0.212 4 . . . CFRS . . .
10.1017 ...... 0.816 21.66 [21.50 2 4 0.058 0.195 4 0(4) CFRS . . .
10.1153 ...... 0.552 21.03 [20.64 4 4 0.099 0.308 4 0(3) CFRS . . .
10.1155 ...... 0.507 21.13 [20.89 4 5 0.206 0.346 4 123(13) CFRS . . .
10.1161 ...... 0.200 20.03 [18.91 4 2 0.074 0.457 1 5(2) CFRS . . .
10.1178 ...... 0.197 21.63 [17.94 4 4 0.127 0.217 4 . . . CFRS . . .
10.1180 ...... 0.465 20.19 [21.15 3 3 0.046 0.402 1 0(1) CFRS . . .
10.1182 ...... 0.471 22.26 [19.38 3 6 0.086 0.204 4 48(18) CFRS . . .
10.1183 ...... 0.649 20.60 [21.92 4 3 0.156 0.277 4 57(16) CFRS . . .
10.1189 ...... 0.949 21.71 [21.98 2 3 0.163 0.351 4 0(5) CFRS . . .
10.1203 ...... 0.686 22.28 [20.29 3 6 0.129 0.255 4 62(2) CFRS . . .
10.1207 ...... 0.706 21.42 [20.64 3 1 0.155 0.367 4 48(16) CFRS . . .
10.1209 ...... 0.841 21.32 [21.97 3 1 0.107 0.499 1 0(7) CFRS . . .
10.1213 ...... 0.817 21.95 [21.04 3 6 0.072 0.175 4 27(2) CFRS . . .
10.1220 ...... 0.909 22.28 [21.10 3 6 0.394 0.357 4 38(4) CFRS . . .
10.1222 ...... 0.519 21.42 [20.15 3 4 0.180 0.287 4 25(16) CFRS . . .
10.1231 ...... 0.473 21.09 [20.07 3 1 0.067 0.464 1 46(13) CFRS . . .
10.1233 ...... ... 21.45 . . . 0 5 0.078 0.221 4 . . . CFRS . . .
10.1236 ...... 0.750 22.12 [20.91 3 4 0.215 0.226 6 14(4) CFRS . . .
10.1243 ...... 0.585 20.92 [20.84 3 3 0.055 0.449 1 0(10) CFRS . . .
10.1255 ...... 0.467 19.41 [21.95 4 0 0.045 0.475 1 9(6) CFRS . . .
10.1257 ...... 0.777 21.42 [21.29 3 4 0.023 0.375 1 0(4) CFRS . . .
10.1262 ...... 0.578 21.65 [20.20 3 1 0.116 0.438 1 0(5) CFRS . . .
10.1270 ...... 0.670 21.26 [21.08 4 4 0.071 0.423 1 . . . CFRS . . .
10.1281 ...... 0.111 21.80 [15.79 3 6 0.119 0.210 4 . . . CFRS . . .
10.1313 ...... ... 22.37 . . . 0 0 0.122 0.319 1 . . . CFRS . . .
10.1349 ...... 0.468 20.56 [20.91 4 3 0.081 0.356 4 13(5) CFRS . . .
10.1423 ...... 0.724 22.59 [20.23 2 1 0.100 0.465 1 14(3) CFRS . . .
10.1502 ...... ...g22.04 . . . 0 4 0.099 0.407 4 . . . CFRS . . .
10.1612 ...... 0.073 21.18 [15.92 3 5 0.077 0.189 4 . . . CFRS . . .
10.1613 ...... 0.076 21.33 [16.11 1 6 0.117 0.205 4 . . . CFRS . . .
10.1614 ...... 0.000 18.46 . . . 4 [2 ... ... [2 . . . CFRS . . .
10.1631 ...... 0.000 20.14 . . . 4 [2 ... ... [2 . . . CFRS . . .
10.1637 ...... 0.497 20.56 [20.79 4 1 0.085 0.519 1 0(6) CFRS . . .
10.1643 ...... 0.234 20.61 [18.90 4 4 0.146 0.354 4 . . . CFRS . . .
10.1644 ...... 0.077 19.90 [17.09 4 6 0.066 0.259 4 . . . CFRS . . .
10.1650 ...... 0.007 19.21 [14.88 4 6 0.353 0.199 6 . . . CFRS . . .
10.1651 ...... 0.197 20.28 [18.53 4 2 0.093 0.502 1 . . . CFRS . . .
14.0147 ...... 1.181 22.26 [21.76 9 6 0.122 0.158 6 . . . CFRS . . .
14.0163 ...... 0.000 18.99 . . . 4 [2 ... ... [2 . . . CFRS . . .
14.0198 ...... 1.603 20.06 . . . 14 [1 0.159 0.847 [1 . . . CFRS . . .
14.0207 ...... 0.546 19.67 [22.19 4 2 0.056 0.612 1 0(15) CFRS . . .
14.0293 ...... 0.761 21.25 [21.82 3 4 0.082 0.312 4 12(0) GRTH . . .
14.0310 ...... 0.238 20.93 [19.40 4 6 0.054 0.236 4 36(0) CFRS . . .
14.0312 ...... 0.746 21.80 [20.86 3 4 0.161 0.268 4 19(0) GRTH . . .
14.0377 ...... 0.260 20.81 [19.61 4 6 0.218 0.130 6 79(30) CFRS . . .
14.0384 ...... 0.000 21.97 . . . 2 [2 ... ... [2 . . . CFRS . . .
14.0393 ...... 0.602 20.49 [21.93 4 5 0.205 0.147 4 22(1) CFRS . . .
14.0400 ...... 0.674 21.38 [21.19 4 4 0.089 0.284 4 12(0) GRTH . . .
14.0411 ...... 0.836 21.45 [21.65 3 6 0.212 0.399 1 57(0) GRTH . . .
14.0422 ...... 0.421 20.39 [20.64 2 1 0.073 0.479 1 3(1) CFRS . . .
14.0435 ...... 0.068 18.39 [18.83 3 3 0.051 0.454 1 . . . CFRS . . .
14.0443 ...... 0.000 20.14 . . . 4 [2 ... ... [2 . . . CFRS . . .
14.0462 ...... 0.000 22.22 . . . 3 [2 ... ... [2 . . . CFRS . . .
14.0485 ...... 0.654 22.20 [20.47 3 4 0.037 0.238 4 34(5) CFRS . . .
14.0501 ...... 0.372 21.66 [20.00 4 6 0.077 0.205 4 44(0) GRTH . . .
14.0516 ...... ... 22.22 . . . 0 1 0.116 0.432 1 . . . GRTH .. .
14.0528 ...... 0.064 20.61 [16.89 4 1 0.107 0.479 1 . . . CFRS . . .
14.0529 ...... 0.000 18.45 . . . 4 [2 ... ... [2 . . . CFRS . . .
14.0547 ...... 1.160 21.40 [23.09 3 6 0.223 0.115 6 13(0) GRTH . . .
14.0574 ...... 0.000 21.66 . . . 2 [2 ... ... [2 . . . CFRS . . .
14.0593 ...... 0.614 22.48 [20.41 3 6 0.163 0.176 6 29(4) CFRS . . .
14.0608 ...... 0.969 22.16 [21.43 2 6 0.118 0.180 6 10(0) GRTH . . .
14.0620 ...... 0.000 22.25 . . . 3 [2 ... ... [2 . . . CFRS . . .
14.0651 ...... 0.637 22.02 [20.30 1 2 0.066 0.437 1 . . . CFRS . . .
14.0665 ...... 0.809 22.97 [20.78 2 6 . . . . . . 6 22(0) GRTH . . .
14.0666 ...... 0.000 21.31 . . . 4 [2 ... ... [2 . . . CFRS . . .
14.0685 ...... 0.081 17.85 [19.96 4 2 0.065 0.524 1 . . . CFRS . . .
14.0695 ...... 0.266 21.39 [19.08 4 1 0.131 0.347 4 40(0) GRTH . . .
14.0700 ...... 0.643 20.42 [21.60 4 2 0.082 0.590 1 0(0) GRTH .. .
14.0725 ...... 0.582 22.11 [19.74 3 6 0.190 0.230 4 42(3) CFRS . . .
14.0743 ...... 0.986 22.19 [22.05 2 6 0.105 0.182 6 28(0) GRTH . . .
117
TABLE 2ÈContinued
AC EW*OII+
ID zF814W MBAB NoteaClassbAcCcClassd(1 p)eOrigin Old IDf
14.0746 ...... 0.675 21.43 [20.68 3 2 0.068 0.423 1 0(5) CFRS . . .
14.0749 ...... 0.818 22.41 [20.61 2 6 0.126 0.185 6 18(0) GRTH . . .
14.0760 ...... ... 22.07 . . . 0 8 . . . . . . 6 . . . CFRS . . .
14.0807 ...... 0.985 21.88 [21.82 2 3 0.088 0.344 4 9(0) GRTH . . .
14.0846 ...... 0.989 21.81 [21.91 92 6 0.313 0.193 6 . . . CFRS . . .
14.0848 ...... 0.662 22.60 [20.31 3 3 0.136 0.263 4 37(9) CFRS . . .
14.0851 ...... ... 21.99 . . . 0 2 0.121 0.293 1 . . . CFRS . . .
14.0854 ...... 0.992 21.63 [22.31 2 1 0.096 0.423 1 0(5) CFRS . . .
14.0899 ...... 0.875 21.71 [21.73 9 3 0.054 0.288 4 9(2) CFRS . . .
14.0916 ...... 0.325 20.95 [19.94 3 3 0.085 0.328 4 18(3) CFRS . . .
14.0922 ...... 0.000 22.29 . . . 4 [2 ... ... [2 . . . CFRS . . .
14.0939 ...... 0.918 21.96 [22.24 1 6 0.268 0.340 4 0(0) GRTH . . .
14.0972 ...... 0.674 21.15 [21.53 4 6 0.148 0.378 4 66(1) CFRS . . .
14.0983 ...... 0.286 21.26 [19.59 4 3 0.103 0.373 4 34(0) GRTH . . .
14.0985 ...... 0.807 22.29 [20.62 3 4 0.094 0.204 4 27(5) CFRS . . .
14.1008 ...... 0.433 20.66 [20.51 3 4 0.090 0.397 1 9(0) GRTH . . .
14.1012 ...... 0.479 21.46 [20.41 3 1 0.054 0.381 1 29(2) CFRS . . .
14.1028 ...... 0.988 21.63 [22.22 3 2 0.159 0.415 1 31(3) CFRS . . .
14.1037 ...... 0.549 21.27 [20.63 3 6 0.099 0.223 4 24(3) CFRS . . .
14.1039 ...... 0.079 19.57 [18.51 4 6 0.153 0.423 1 . . . CFRS . . .
14.1042 ...... 0.722 21.33 [21.07 3 2 0.137 0.480 1 13(1) CFRS . . .
14.1043 ...... 0.641 20.22 [22.20 4 3 0.043 0.404 1 0(3) CFRS . . .
14.1052 ...... 0.000 17.58 . . . 4 [2 ... ... [2 . . . CFRS . . .
14.1071 ...... 0.359 22.32 [18.67 3 4 0.199 0.094 6 57(12) CFRS . . .
14.1079 ...... 0.901 21.82 [21.45 9 6 0.086 0.284 4 38(3) CFRS . . .
14.1087 ...... 0.659 21.95 [20.58 3 8 0.137 0.134 6 52(4) CFRS . . .
14.1103 ...... 0.209 22.41 [17.75 4 [1 0.124 0.564 1 0(17) CFRS . . .
14.1126 ...... 0.743 22.20 [20.72 3 6 0.154 0.164 6 62(6) CFRS . . .
14.1129 ...... 0.831 21.12 [22.10 3 6 0.251 0.115 6 28(0) GRTH . . .
14.1136 ...... 0.640 22.02 [20.98 3 8 0.131 0.346 4 73(5) CFRS . . .
14.1139 ...... 0.660 20.48 [22.31 3 6 0.159 0.227 4 16(1) CFRS . . .
14.1143 ...... 0.673 22.44 [20.10 93 3 0.063 0.256 4 . . . CFRS . . .
14.1146 ...... 0.744 21.68 [21.12 3 8 0.185 0.274 4 53(4) CFRS . . .
14.1158 ...... 0.000 20.81 . . . 3 [2 ... ... [2 . . . CFRS . . .
14.1164 ...... 0.671 21.76 [20.94 3 6 0.120 0.211 4 34(0) GRTH . . .
14.1166 ...... 1.015 22.28 [21.36 3 1 0.089 0.358 4 56(6) CFRS . . .
14.1178 ...... ... 22.22 . . . 0 1 0.101 0.326 1 . . . GRTH .. .
14.1179 ...... 0.434 21.44 [19.73 2 2 0.142 0.483 1 0(14) CFRS . . .
14.1189 ...... 0.753 22.06 [20.80 3 4 0.152 0.302 4 43(6) CFRS . . .
14.1190 ...... 0.754 21.06 [21.84 3 3 0.102 0.276 4 9(2) CFRS . . .
14.1193 ...... 0.078 21.67 [16.61 4 6 0.220 0.179 6 . . . CFRS . . .
14.1200 ...... 0.235 21.84 [18.21 2 6 0.096 0.163 6 0(42) CFRS . . .
14.1209 ...... 0.234 20.98 [19.43 4 3 0.045 0.283 4 . . . CFRS . . .
14.1234 ...... 0.000 22.16 . . . 4 [2 ... ... [2 . . . CFRS . . .
14.1239 ...... 0.362 21.74 [19.68 3 6 0.099 0.220 4 47(2) CFRS . . .
14.1242 ...... 0.290 21.69 [19.03 3 6 0.339 0.099 6 33(25) CFRS . . .
14.1251 ...... 0.814 22.17 [20.42 3 4 0.076 0.411 1 0(0) GRTH . . .
14.1257 ...... 0.291 20.68 [19.82 3 4 0.066 0.248 4 55(0) GRTH . . .
14.1258 ...... 0.645 22.39 [20.25 3 1 0.116 0.323 4 62(9) CFRS . . .
14.1264 ...... 0.703 22.80 [20.00 91 8 . . . . . . 6 . . . CFRS . . .
14.1273 ...... 0.257 21.94 [18.64 4 6 0.111 0.217 4 57(33) CFRS . . .
14.1277 ...... 0.810 21.33 [21.93 2 8 0.131 0.387 1 19(10) CFRS . . .
14.1281 ...... 0.141 21.07 [18.07 3 3 0.091 0.390 1 30(0) GRTH . . .
14.1311 ...... 0.806 20.58 [22.90 3 1 0.087 0.442 1 0(1) CFRS . . .
14.1321 ...... 0.106 21.40 [17.39 4 8 0.107 0.278 4 . . . CFRS . . .
14.1356 ...... 0.831 22.23 [21.10 3 8 0.081 0.199 6 47(7) CFRS . . .
14.1371 ...... 0.000 18.70 . . . 3 [2 ... ... [2 . . . CFRS . . .
14.1395 ...... 0.530 21.78 [20.20 4 4 0.134 0.210 4 63(8) CFRS . . .
14.1415 ...... 0.745 21.06 [21.62 2 1 0.138 0.445 1 0(0) GRTH . . .
14.1419 ...... 0.236 22.72 [16.21 93 [1 0.195 0.374 4 . . . CFRS . . .
14.1427 ...... 0.860 21.54 [21.50 9 4 0.114 0.278 4 33(0) GRTH . . .
14.1446 ...... 0.348 20.07 [21.10 4 1 0.058 0.468 1 24(1) CFRS . . .
14.1464 ...... 0.462 21.06 [19.94 2 0 0.171 0.514 1 11(8) CFRS . . .
14.1496 ...... 0.899 21.93 [21.61 3 [1 0.105 0.405 1 54(6) CFRS . . .
14.1501 ...... 0.989 22.02 [21.98 2 6 0.241 0.187 6 60(0) GRTH . . .
14.1502 ...... ... 22.26 . . . 0 3 0.087 0.260 4 . . . GRTH .. .
14.1524 ...... 0.427 19.87 [21.89 3 3 0.124 0.243 4 15(0) GRTH . . .
14.9025 ...... 0.155 19.16 [20.28 4 3 0.084 0.453 1 . . . CFRS . . .
14.9987 ...... 0.420 22.48 [18.86 92 4 0.110 0.373 4 25(0) CFRS . . .
22.0377 ...... ... 22.28 . . . 0 3 0.077 0.327 4 . . . CFRS . . .
22.0434 ...... 0.094 19.90 [18.83 4 5 0.062 0.268 4 . . . CFRS . . .
22.0453 ...... 0.623 22.12 [20.25 3 6 0.150 0.234 4 . . . CFRS . . .
22.0497 ...... 0.470 19.42 [22.75 4 0 0.171 0.369 4 0(5) CFRS . . .
118
TABLE 2ÈContinued
AC EW*OII+
ID zF814W MBAB NoteaClassbAcCcClassd(1 p)eOrigin Old IDf
22.0501 ....... 0.424 20.48 [20.77 4 0 0.053 0.436 1 0(9) CFRS . . .
22.0541 ....... ... 22.69 . . . 0 3 0.099 0.217 4 . . . CFRS . . .
22.0576 ....... 0.890 21.93 [21.12 9 6 0.225 0.445 1 63(25) CFRS . . .
22.0583 ....... 0.431 21.63 [19.36 3 5 0.075 0.189 4 36(21) CFRS . . .
22.0585 ....... 0.294 20.74 [19.40 2 2 0.155 0.348 4 0(24) CFRS . . .
22.0599 ....... 0.889 21.62 [21.66 9 6 0.092 0.330 4 64(11) CFRS . . .
22.0609 ....... 0.475 20.60 [20.84 3 3 0.085 0.357 4 0(11) CFRS . . .
22.0618 ....... 0.830 22.28 [20.78 1 2 0.054 0.333 4 . . . CFRS . . .
22.0622 ....... 0.325 21.92 [18.58 3 4 0.211 0.401 1 13(8) CFRS . . .
22.0671 ....... 0.319 20.87 [20.24 4 3 0.115 0.371 4 31(5) CFRS . . .
22.0676 ....... 0.141 20.69 [18.24 4 2 0.063 0.422 1 . . . CFRS . . .
22.0758 ....... 0.294 19.54 [20.80 3 0 0.050 0.491 1 0(5) CFRS . . .
22.0764 ....... 0.819 21.98 [20.91 3 6 0.128 0.148 6 19(3) CFRS . . .
22.0779 ....... 0.925 21.89 [21.64 9 3 0.169 0.370 4 12(2) CFRS . . .
22.0819 ....... 0.293 20.86 [19.61 4 4 0.087 0.239 4 45(3) CFRS . . .
22.0890 ....... ... 21.28 . . . 0 2 0.088 0.399 1 . . . CFRS . . .
22.0919 ....... 0.474 21.29 [20.25 4 6 0.320 0.487 1 8(1) CFRS . . .
22.0923 ....... ... 22.28 . . . 0 4 0.149 0.182 4 . . . CFRS . . .
22.0944 ....... 0.249 18.84 [21.59 3 4 0.061 0.388 1 0(57) CFRS . . .
22.0945 ....... 0.676 21.93 [20.71 3 5 0.164 0.186 6 17(3) CFRS . . .
22.0953 ....... 0.977 22.33 [21.39 8 6 0.084 0.154 6 34(6) CFRS . . .
22.0988 ....... 0.477 22.83 [19.13 93 5 0.157 0.116 6 . . . CFRS . . .
22.1015 ....... 0.231 23.38 [15.50 94 3 . . . . . . 4 . . . CFRS . . .
22.1037 ....... 0.550 21.91 [19.91 2 2 0.218 0.471 1 163(29) CFRS . . .
22.1078 ....... 0.671 22.13 [20.65 1 1 0.241 0.477 1 . . . CFRS . . .
22.1279 ....... 0.594 21.28 [20.70 3 1 0.089 0.358 1 0(20) CFRS . . .
22.1313 ....... 0.819 22.24 [21.45 3 6 0.154 0.131 6 73(5) CFRS . . .
22.1374 ....... 0.093 18.21 [20.45 4 4 0.054 0.334 4 . . . CFRS . . .
22.1406 ....... 0.818 21.97 [20.97 4 [1 0.121 0.319 4 100(4) CFRS . . .
22.1453 ....... 0.816 21.61 [21.59 3 6 0.340 0.163 6 0(3) CFRS . . .
22.1466 ....... ... 21.85 . . . 0 3 0.135 0.273 4 . . . CFRS . . .
22.1486 ....... 0.953 22.58 [21.32 8 [1 ... ... [1 12(6) CFRS . . .
22.1507 ....... 0.820 21.37 [21.48 3 1 0.098 0.446 1 0(2) CFRS . . .
10.10116 ...... ... 19.75 . . . 0 4 0.085 0.332 4 . . . AutoÐb 10f–14
10.11699 ...... 0.000 19.76 . . . 4 [2 ... ... [2 . . . LDSS-1 10.2.9HI
10.11702 ...... 0.168 19.36 . . . 4 2 0.118 0.511 1 10([9) LDSS-1 10.2.12HI
10.11703 ...... 0.437 19.63 . . . 4 3 0.050 0.429 1 5([9) LDSS-1 10.2.13HI
10.11706 ...... 0.151 21.27 . . . 4 6 0.230 0.133 6 21([9) LDSS-1 10.2.16HI
10.11709 ...... 0.179 20.94 . . . 4 3 0.046 0.276 4 16([9) LDSS-1 10.2.19HI
10.12058 ...... ... 23.36 . . . 0 6 0.268 0.081 6 . . . LDSS-2 10.21.227
10.12059 ...... 0.307 21.15 . . . 4 5 0.049 0.218 4 20(2) LDSS-2 10.21.233
10.12060 ...... 0.294 21.23 [18.39 4 1 0.120 0.493 1 12(4) LDSS-2 10.21.262
10.12062 ...... 0.634 21.85 [20.21 4 0 0.110 0.475 1 55(1) LDSS-2 10.21.279
10.12063 ...... 1.108 22.27 [21.36 4 [1 0.161 0.397 1 65(10) LDSS-2 10.21.288
10.12065 ...... 0.207 20.97 [17.85 4 1 0.103 0.587 1 19(6) LDSS-2 10.21.328
10.12066 ...... 0.924 22.30 [20.36 2 6 0.104 0.155 6 17(8) LDSS-2 10.21.22
10.12071 ...... 0.177 21.25 [17.84 4 3 0.082 0.283 4 0(8) LDSS-2 10.21.88
10.12073 ...... 0.492 20.35 [20.61 4 6 0.074 0.354 4 0(0) LDSS-2 10.21.109
10.12076 ...... 0.323 21.91 [18.40 4 5 0.067 0.208 4 25(5) LDSS-2 10.21.301
10.12078 ...... 0.296 22.82 . . . 4 3 0.072 0.196 4 58(8) LDSS-2 10.22.223
10.12080 ...... 0.314 20.81 [18.96 4 [1 0.182 0.643 1 0(3) LDSS-2 10.22.248
10.12081 ...... 0.563 22.23 [19.42 4 6 0.221 0.142 6 17(2) LDSS-2 10.22.260
10.12085 ...... ... 23.02 . . . 0 6 0.259 0.109 6 . . . LDSS-2 10.22.315
10.12086 ...... 0.324 22.13 [18.24 4 6 0.105 0.159 6 50(3) LDSS-2 10.22.330
10.12087 ...... 2.749 22.81 . . . 4 [1 ... ... [1 . . . LDSS-2 10.22.25
10.12089 ...... 0.384 20.39 [19.99 4 1 0.060 0.500 1 0(3) LDSS-2 10.22.61
10.12091 ...... 0.476 20.25 [20.76 4 3 0.104 0.456 1 10(2) LDSS-2 10.22.71
10.12092 ...... 0.436 20.23 . . . 4 2 0.090 0.490 1 0(2) LDSS-2 10.22.77
10.12095 ...... 0.724 21.75 [20.59 2 6 0.073 0.218 4 30(2) LDSS-2 10.22.122
10.12519 ...... 0.097 19.74 [18.45h4 3 0.182 0.315 4 0(0) LDSS-2 10.23.218
10.12520 ...... 1.999 22.42 . . . 4 [1 ... ... [1 . . . LDSS-2 10.23.222
10.12522 ...... 0.000 19.02 . . . 4 [2 ... ... [2 . . . LDSS-2 10.23.235
10.12524 ...... 0.149 21.51 [17.16 4 2 0.108 0.326 4 42(5) LDSS-2 10.23.255
10.12525 ...... 0.435 19.79 [20.81 4 1 0.291 0.069 1 3(1) LDSS-2 10.23.273
10.12527 ...... 0.000 19.87 . . . 4 [2 ... ... [2 . . . LDSS-2 10.23.332
10.12528 ...... 0.582 20.69 [21.11 4 5 0.088 0.317 4 8(1) LDSS-2 10.23.28
10.12529 ...... ... 23.16 . . . 0 6 0.243 0.163 6 . . . LDSS-2 10.23.32
10.12530 ...... 0.476 20.50 [20.42 4 3 0.123 0.468 1 0(0) LDSS-2 10.23.40
10.12534 ...... ... 21.31 . . . 0 [1 0.238 0.549 [1 . . . LDSS-2 10.23.92
10.12535 ...... 0.000 18.07 . . . 4 [2 ... ... [2 . . . LDSS-2 10.23.105
10.12536 ...... 1.256 21.00 . . . 4 [1 0.158 0.622 1 . . . LDSS-2 10.23.116
10.12537 ...... 0.000 21.69 . . . 4 [2 ... ... [1 . . . LDSS-2 10.23.126
10.12786 ...... ... 21.88 . . . 0 6 . . . . . . 6 . . . LDSS-1 10.2.2FB
119
120 BRINCHMANN ET AL. Vol. 499
TABLE 2ÈContinued
AC EW*OII+
ID zF814W MBAB NoteaClassbAcCcClassd(1 p)eOrigin Old IDf
10.12787 ...... 0.283 22.32 . . . 4 [1 ... ... [1 . . . LDSS-1 10.2.5FB
13.10222 ...... 0.052 18.98 [17.23i[1 3 0.059 0.249 4 0([9) AutoÐb 13b–14
13.10379 ...... ... 19.35 . . . 0 [2 ... ... [2 . . . AutoÐb 13m–16
13.11753 ...... 0.198 19.08 . . . 2 4 . . . . . . 4 28([9) LDSS-1 13.2.1HI
13.11772 ...... 0.512 20.90 . . . 4 4 0.098 0.353 4 63([9) LDSS-1 13.2.20HI
13.11874 ...... ... 19.33 . . . 0 3 0.064 0.415 4 . . . AutoÐb 13f1–1
13.11924 ...... 0.281 18.86 [19.83i[1 1 0.085 0.634 1 0([9) AutoÐb 13f1–63
13.11925 ...... 0.256 20.34 [18.83i[1 3 0.061 0.320 4 46([9) AutoÐb 13f1–64
13.11976 ...... 0.336 21.37 [18.93i[1 5 0.090 0.247 4 100([9) AutoÐb 13xf–64
13.12099 ...... 0.385 21.00 [19.69 4 5 0.095 0.345 4 19(2) LDSS-2 13.21.323
13.12106 ...... 0.556 21.60 [20.09 4 6 0.103 0.200 4 14(3) LDSS-2 13.21.465
13.12107 ...... 0.556 21.22 [20.47 2 2 0.143 0.389 1 34(5) LDSS-2 13.21.480
13.12109 ...... 0.462 22.25 [18.97 2 5 0.153 0.143 4 36(7) LDSS-2 13.21.517
13.12111 ...... 0.089 22.31 [15.42 4 6 0.191 0.154 6 41(7) LDSS-2 13.21.27
13.12112 ...... 0.424 21.09 [19.92 4 2 0.102 0.474 1 12(2) LDSS-2 13.21.38
13.12116 ...... 0.187 21.18 [17.94 4 1 0.058 0.472 1 59(14) LDSS-2 13.21.106
13.12117 ...... 0.536 22.05 [19.55 2 5 0.094 0.224 4 32(3) LDSS-2 13.21.123
13.12118 ...... 0.335 21.86 [18.61 4 5 0.135 0.333 4 35(4) LDSS-2 13.21.160
13.12538 ...... ... 21.71 . . . 0 2 0.085 0.490 1 . . . LDSS-2 13.21.311
13.12539 ...... 0.000 19.97 . . . 4 [2 ... ... [2 . . . LDSS-2 13.22.325
13.12540 ...... 0.452 22.33 [18.99 2 6 0.148 0.178 6 0(0) LDSS-2 13.22.344
13.12542 ...... ... 22.03 . . . 0 6 0.200 0.153 6 . . . LDSS-2 13.22.367
13.12545 ...... 0.830 20.14 [22.57 4 6 0.195 0.289 4 19(1) LDSS-2 13.22.400
13.12546 ...... 0.283 20.01 [19.44 4 3 0.112 0.528 1 0(0) LDSS-2 13.22.417
13.12549 ...... 0.493 21.02 [20.26 4 3 0.105 0.328 4 13(1) LDSS-2 13.22.469
13.12550 ...... 0.000 18.44 . . . 4 [2 ... ... [2 . . . LDSS-2 13.22.484
13.12551 ...... ... 22.12 . . . 0 5 0.077 0.224 4 . . . LDSS-2 13.22.492
13.12552 ...... 0.566 20.42 [21.14 4 3 0.093 0.268 4 7(1) LDSS-2 13.22.510
13.12553 ...... 0.278 20.47 [19.42 4 3 0.086 0.310 4 8(3) LDSS-2 13.22.519
13.12554 ...... ... 20.90 . . . 0 4 0.318 0.233 4 . . . LDSS-2 13.22.12
13.12555 ...... 0.426 21.99 [19.09 2 [1 0.165 0.385 1 4(0) LDSS-2 13.22.28
13.12556 ...... ... 22.57 . . . 0 [2 ... ... [2 . . . LDSS-2 13.22.34
13.12559 ...... ... 22.73 . . . 0 [1 ... ... [1 . . . LDSS-2 13.22.98
13.12560 ...... 0.363 20.49 [19.90 4 1 0.067 0.528 1 7(2) LDSS-2 13.22.116
13.12561 ...... 0.326 19.51 [20.73 4 3 0.106 0.519 1 4(1) LDSS-2 13.22.131
13.12566 ...... 0.000 19.59 . . . 4 [2 ... ... [2 . . . LDSS-2 13.22.180
13.12567 ...... ... 22.28 . . . 0 1 0.152 0.422 1 . . . LDSS-2 13.22.186
13.12759 ...... 0.000 21.07 . . . 4 [2 ... ... [2 . . . LDSS-1 13.2.9LO
13.12764 ...... ... 21.97 . . . 0 2 0.156 0.457 1 . . . LDSS-1 13.2.14LO
13.12767 ...... ... 21.84 . . . 0 [1 ... ... [1 . . . LDSS-1 13.2.17LO
13.12783 ...... ... 22.45 . . . 0 8 0.210 0.205 6 . . . LDSS-1 13.2.35LO
13.12795 ...... 2.934 21.85 . . . 4 [1 ... ... [1 . . . LDSS-1 13.2.1FB
13.12797 ...... 0.627 22.69 . . . 2 [1 ... ... [1 . . . LDSS-1 13.2.3FB
13.12798 ...... 0.297 22.78 . . . 2 [1 ... ... [1 . . . LDSS-1 13.2.4FB
13.12801 ...... ... 22.47 . . . 0 6 0.157 0.167 6 . . . LDSS-1 13.2.7FB
13.12802 ...... 0.667 22.63 . . . 2 2 0.120 0.233 4 . . . LDSS-1 13.2.9FB
13.12803 ...... ... 23.10 . . . 0 6 . . . . . . 6 . . . LDSS-1 13.2.11FB
13.12808 ...... 0.000 22.18 . . . 0 [2 ... ... [2 . . . LDSS-1 13.2.28LO
13.12810 ...... ... 22.30 . . . 0 [1 ... ... [1 . . . LDSS-1 13.2.10FB
13.12811 ...... 0.550 21.90 . . . 2 4 0.137 0.193 4 . . . LDSS-1 13.2.13FB
aThe conÐdence class for the redshift. For the LDSS objects this has been transformed to the CFRS system by assigning note \4 to conÐdent
redshifts, note \2 to less secure redshifts, and 0 to failures. For the few LDSS objects for which there is no conÐdence class, we have assigned
note \[1.
bThe eyeball classiÐcation for the object.
cThe A- and C-parameters (uncorrected); see °3.2.
dThe AC classiÐcation for the object using the division lines in Fig. 8.
eThe equivalent width of [O II]. For the CFRS objects this is from et al. for the LDSS objects it is from the AutoÐb surveyHammer (1997) ;
et al.(Ellis 1996).
fThe identiÐcation given in the original LDSS paper.
gThe object is clearly extended, but was given z\0 in in the CFRS survey.
hThe HST photometry here is uncertain, and is based on the original photometry.MBAB bJ
iThe absolute magnitude is the original AutoÐb absolute magnitude based on transformed to AB.bJ
the evolution. Part of the difficulty with color and emission-
line strength is that populations deÐned according to these
criteria may well be transient, and thus detailed compari-
sons of luminosities and volume densities at various red-
shifts will be confused.
Although the same criticisms can no doubt be applied to
galaxy morphology the morphologically(White 1996),
dependent number magnitude counts derived from the
Medium Deep Survey and Hubble Deep Field (Glazebrook
et al. et al. et al. raise1995a; Driver 1995a ; Abraham 1996b)
important questions concerning the apparent rapid evolu-
tion of the irregular/peculiar/merger galaxies, in compari-
son with the slower trends noted for the spheroidal and
regular spiral classes. How do these classiÐcations map
No. 1, 1998 CFRS AND LDSS SURVEY GALAXIES 121
TABLE 3
HST SURVEY COMPLETENESS
COMPLETENESS
SURVEY LIMITS
SURVEY (deg2) Geometric Spectroscopic EFFECTIVE AREAa
CFRS ............... 17.5 \IAB\22.5 0.5477 90.44% 0.01377
LDSS-2 (10h)....... 22.5 \bJ\24.0 0.7941 85.19% 0.002513
LDSS-2 (13h)b...... 22.5 \bJ\23.3 0.4516 85.71% 0.001576
aDeÐned as surveyed area ]geometric completeness.
bThe LDSS-2 13hÐeld has less deep spectroscopy; we have adopted the completeness limits as
discussed in et al.Glazebrook (1995b).
onto the redshift survey data plane? In this section we
discuss the various ways in which we have classiÐed the
galaxy morphologies, taking care to note these uncer-
tainties and systematic changes that may occur because our
survey samples galaxies over a large range of redshifts.
3.1. V isual ClassiÐations
The Ðrst technique we used to investigate the morpho-
logical characteristics of our sample follows the visual
approach adopted by the Medium Deep Survey team. Fol-
lowing the precepts discussed by et al.Glazebrook (1995b),
FIG. 3.ÈVerifying the absolute-magnitude scales of the CFRS and
LDSS-2 redshift surveys. Top panel: color for the LDSS-2 gal-BAB[bJ
axies. Middle panel: Di†erence in for CFRS galaxies using the CFRSMB
and LDSS-2 SEDs. Bottom panel: Redshift dependence of di†erencesMB
obtained by calculating k-corrections for the LDSS-2 galaxies using andbJ
IF814W.
three of us (R. S. E., S. J. L., O. L. F.) have classiÐed all the
galaxies by eye according to a scheme illustrated in Figure
The scheme that we have adopted here di†ers slightly4.
from that utilized by the MDS team in that we decided to
separate compact objects with faint extensions (so-called
tadpoles) from the irregular/merger/peculiar and compact
objects. This was done for objects for which classiÐcations
were either compact or peculiar and for which no consensus
could be reached.
An comparison of the eyeball classiÐcations between the
three observers is shown in and indicates a scatterFigure 5
of around 1.2 classes. This is similar to the scatter found by
the MDS team et al. at the same magni-(Abraham 1996a)
tude limit. The individual classiÐcations were then merged
by taking the median value of the di†erent classiÐers. The
Ðnal number of objects in various classes is also indicated in
The resulting diagram is shown inFigure 4. MB-zFigure 6
with the fractional redshift distributions of the various mor-
phological classes inlaid.
In order to determine whether the morphological mixture
is robustly estimated, we can compare the morphologically
segregated N(m) for our sample with that for the MDS
survey et al. et al.(Abraham 1996b; Glazebrook 1995a).
Since the MDS survey is I-selected, we can only compare it
to that subset of our galaxies drawn from the CFRS survey.
As the CFRS survey did not target all objects lying between
the photometric limits, we must correct for those objects
that are in the HST Ðeld within the magnitude limits that
were not targeted spectroscopically. We have done this by
multiplying the counts in each Ðeld with the ratio of photo-
metrically to spectroscopically observed objects for each
HST frame, taken from the CFRS survey. The resulting
counts are shown in It can be seen that there is anFigure 7.
apparent lack of bright irregular galaxies. This is in part
because the bright irregulars by coincidence happen to be in
Ðelds with high completeness. By assigning the bright
irregulars to random Ðelds, we Ðnd that the di†erence
between our irregular counts and the MDS counts is gener-
ally less than 2 p, and we do not consider this as a potential
problem, as we will see that the abundance of low-redshift
irregulars is almost exactly as expected from the local lumi-
nosity function (see We Ðnd an integrated count from°5.1).
to from our survey of 394, whereasIAB \17.5 IAB\22.5
that expected from the MDS survey is 409, i.e., in close
agreement. The Ðeld-to-Ðeld variation of the morphological
composition is also satisfactorily constant to within the
uncertainties.
3.2. Automated ClassiÐcations
A major difficulty with visual classiÐers is their subjective
nature et al. Accordingly, as an objective route(Naim 1995).
NStar=27
Star=-2
NCompact=22
Compact=-1
NE/S0=45
E=0, E/S0=1
NS0/a=33
S0/a=2
NSab=49
Sab=3
NSbc=47
Sbc=4
NScd=32
Scd=5
NIrr=75
Irregular=6
NTadpole=11
Tadpole=11
122 BRINCHMANN ET AL. Vol. 499
FIG. 4.ÈExamples of the morphological types used in this paper, with the total number of objects in each class in the survey. Each image is 6A]6A.
forward, we have also performed machine-based classi-
Ðcations using the procedures adopted by Abraham et al.
The technique is based on measurements of a(1994, 1996b).
central concentration index Cand a rotational asymmetry
factor A. The Ðrst of these parameters tracks the bulge-to-
disk ratio, while the second traces the degree of irregularity.
At low redshift the positions of galaxies on the log Aversus
log Cplane can be used to distinguish between early-type
systems, spirals earlier than type Sd, late-type spirals, and
irregulars/peculiars/mergers. As will be shown below, at
higher redshifts these classes can also be distinguished in
principle using an A-Cdiagram. However, uncertainties
arise because galaxies of a given type may move into a
region of the A-Cplane occupied by a di†erent class
because of redshift-dependent e†ects. These biases must be
carefully studied before quantitative comparisons can be
made over a range in redshift. We defer a discussion of these
studies until the next section.
The starting point for measuring the two parameters is a
““ segmented ÏÏ galaxy image, constructed by isolating those
pixels that lie above a surface brightness threshold Np
above the sky brightness, where pis the sky variance, and N
is a constant, typically 1.5. For the present work (see dis-
cussion below and in we have chosen pso thatAppendix A),
the measurements go to a uniform limiting surface bright-
ness. The C-parameter represents the ratio of light within an
inner and outer elliptical aperture determined from the sky-
subtracted, intensity-weighted, second-order moment of the
resulting image. The major and minor axes of the outer
aperture are normalized so that the total area within the
ellipse is the isophotal area of the galaxy. The inner aperture
is deÐned by scaling these axes down by a linear factor of 3.
Whereas this deÐnition of central concentration is ade-
quate for local galaxies, the value determined unfortunately
depends on redshift, since the threshold is deÐned relative to
the sky. Thus, less of the galaxy is sampled at high redshift,
because of cosmological dimming. Measuring Cto a Ðxed
rest-frame surface brightness isophote is not really practical,
so it is necessary to consider how to correct Cfor this e†ect.
A procedure to do this is discussed in and inAppendix A,
the rest of the paper all Cvalues have been corrected using
the minimal correction deÐned in equation (7).
No. 1, 1998 CFRS AND LDSS SURVEY GALAXIES 123
FIG. 5.ÈCorrelation between the di†erent classiÐers. The radii of the
circles are proportional to the value in the correlation matrix. The corre-
lation between the di†erent classiÐers is only slightly larger than the inter-
nal scatter of one classiÐer.
The rotational asymmetry parameter Ais deÐned via:
A\;
ij NoIij [Iij
Ro;
ij Iij [kA, (1)
where is the intensity in pixel (i,j), and is the corre-Iij Iij
R
sponding intensity after image rotation by 180¡ about the
centroid of the segmented galaxy image. The term inkA
is a small correction accounting for signalequation (1)
introduced into Aby noise in the sky background; iskA
determined by measuring the asymmetry within a rotated
and self-subtracted region of sky equal in area to that of the
galaxy being analyzed.
The measurement of these parameters is only practical
for that subset of galaxy images with more than 64 contig-
uous WFPC2 pixels above the surface brightness threshold
(see also the discussion in et al. Twenty-ÐveAbraham 1994).
galaxies are too compact for reliable classiÐcation via this
approach. For these objects, only the visual estimates are
available. The upper panel of shows the correctedFigure 8
Aand Cdistributions for our sample. Dashed lines deÐne
three morphological bins (ellipticals/spirals/peculiars),
drawn on the basis of similar measurements made on a local
sample of galaxies of known morphology Guhatha-(Frei,
kurta, & Gunn Detailed consideration of this local1996).
data set is deferred to To avoid confusion with the°4.
visual classiÐcations, we will refer to the automated classes
as AC ellipticals (AC-E), AC spirals (AC-S) and AC pecu-
liars (AC-P). The angle from the intersection point of the
dashed line in the upper panel of to each point inFigure 8
the A-Cplane deÐnes a continuous classiÐcation angle #.
The correlation between this angle and the eyeball classi-
Ðcations is shown in the lower panel of the Ðgure. Overall,
the agreement is satisfactory, but a sizeable scatter is appar-
ent. Note in particular that the angle does not distinguish
well between early-type spirals and E/S0 galaxies.
4.REDSHIFT-DEPENDENT BIASES
High-zgalaxies imaged by HST di†er in appearance
from their local counterparts because of their reduced
apparent size and sampling characteristics, a lower signal-
to-noise ratio and reduced surface brightness with respect
to the sky background, and a shift in the rest wavelength of
the observations. We will refer to the latter term as
““ bandpass shifting.ÏÏ These e†ects will combine to give some
uncertainty in the morphological classiÐcation of galaxies,
generally in the sense of shifting objects to apparently later
Hubble types. et al. attempted to address this in theGlazebrook (1995b)
context of their visual classiÐcation procedure via a blind
classiÐcation of local galaxies whose appearance was care-
fully simulated as viewed at a redshift z\0.7. For machine
classiÐcations this has been addressed to some extent by
et al. using the parameters Aand C.AAbraham (1996b),
beneÐt of working in the framework of Aand Cclassi-
Ðcations is that it enables more quantitative statements
about these biases.
In this paper we will extend the discussion begun by
Abraham et al., utilizing the known redshifts of our sample
to correct for these redshift-dependent biases. To do so, we
have relied on extensive simulations based on a set of multi-
color CCD images of local galaxies from Guhatha-Frei,
kurta, & Gunn Our approach will be to assess (in(1996).
various ways) the extent to which the A- and C-parameters
for local galaxies of known type are likely to shift when they
are placed at larger redshift. We then compare the distribu-
tion of classes statistically, over a number of galaxies, as
viewed at a given redshift and with the intrinsic value at
zB0 in order to determine a ““ misclassiÐcation fraction ÏÏ
for each type that is a function of redshift. The method
assumes that the et galaxies of a given type areFrei al.
representative. Note, in particular, that we do not require
an accurate sampling of the local morphological mix.
Indeed, in principle, only a few representative galaxies of
each type are required. Our goal and procedure di†er from
124 BRINCHMANN ET AL. Vol. 499
FIG. 6.ÈDistribution of galaxies in the plane for the HST survey. Inlaid histograms show the fractional contribution of the four di†erent visualM(BAB)-z
classes as a function of redshift. In this plot the tadpoles have been grouped with the irregular galaxies.
the synthetic creation of faint no-evolution samples (see,
e.g., Broadhurst, & Silk for which theBouwens, 1997) Frei
et sample is not well suited.al.
4.1. T he et Calibration SampleFrei al.
The et sample consists of 82 galaxies with andFrei al. BJ
RCCD images, uniform in quality, with foreground stars
removed. It is important to note, however, that the etFrei
galaxies were not chosen with the intention of samplingal.
the luminosity function of local systems uniformly. Indeed,
et chose galaxies that are (a) bright, (b) have well-Frei al.
resolved morphological structures, and (c) span a wide
range of Hubble system classiÐcation classes. It is therefore
important to assess whether this sample is appropriate for
calibrating redshift-dependent morphological trends.
The absolute magnitude distribution of the etFrei al.
sample is shown in The plot is based on dataFigure 9. MB
published in the Revised Shapely Ames Catalog (RSA). We
have applied an additional 05 shift to take account of cor-
rections applied for internal extinction in the RSA that we
ignore in the CFRS ]LDSS samples. We also indicate
characteristic M* values for local elliptical and Scd galaxies
et al. The et data peak near M*, and,(Marzke 1994). Frei al.
as expected, are deÐcient in systems D2 mag or more fainter
than M*. This could be a drawback for our purposes,
depending on how the morphological biases are corrected.
If, for example, we correct morphologies back to those
appropriate for the rest-frame I-band, a proper match
between the CFRS/LDSS and et luminosity dis-Frei al.
tributions is more important at high redshift. In this case,
the majority of our survey galaxies beyond z^0.3 are
drawn from within 2 mag of M*. In we show linesFigure 9
corresponding to the apparent magnitude limits for the
CFRS survey based on k-corrections for early- and late-
type systems. Clearly, the underluminous galaxies that are
deÐcient in the et sample are undetectable. For ourFrei al.
approach this leads to a low number of objects and hence to
a large statistical uncertainty. But since we cover the whole
range of morphological types, we do not introduce any sys-
tematic biases.
One might worry that selecting the galaxies to be nice
and regular looking might lead us to infer less bandshifting
than is observed. However, we do not think that this is a
major problem for our approach, since the comparison with
the low-redshift data e†ectively takes out the actual values
of Aand C. Thus, although the et sample is usefulFrei al.
for estimating the apparent shift in morphology with red-
shift for regular objects of suitable luminosity, it could be
improved. Future imaging surveys of local systems should
aim to sample fairly the distribution of morphological types
within both the luminosity selection criteria of the survey
and the regions on the Aversus Cdiagram. Such data sets
would also allow compilation of no-evolution and mild-
evolution simulated data sets for comparison with high-
redshift data (see also et al.Bouwens 1997).
4.2. Wavelength-dependent Trends
The simplest test that we can perform is to measure the
A- and C-parameters deÐned earlier in both the R- and
images of the et galaxies. Since our faintBJ-band Frei al.
HST images are taken with the F814W Ðlter, we e†ectively
see the Rband at a redshift of 0.2 and the band at aBJ
redshift of 0.87. The shift in Aand Cacross the etFrei al.
sample is thus a crude but simple measure of the shift
arising from the bandpass e†ect over much of the redshift
range sampled.
In the top panel of we plot the change in asym-Figure 10
metry Aobserved using the Rand images of the sameBJ
local galaxy. Late-type systems are denoted by Ðlled circles,
and, as expected, there is a clear trend for such systems to
have larger asymmetry at shorter wavelengths where the
star formation signatures are more irregular. However, in
quantitative detail, the size of the e†ect is quite small. The
corresponding shift in concentration C, shown in the
bottom panel, is somewhat larger. The change in Crequired
for a galaxy to cross the AC-peculiar and AC-elliptical
No. 1, 1998 CFRS AND LDSS SURVEY GALAXIES 125
FIG. 7.ÈNumber-magnitude counts for the survey (circles) compared to those from the larger MDS survey (triangles). Error bars are Poissonian (Gehrels
In this plot the tadpoles have been grouped with the irregulars.1986).
boundary is also indicated on the Ðgure; for the AC pecu-
liars this is strongly dependent on asymmetry.
This simple comparison indicates that only a small frac-
tion of the et galaxies would cross into the areasFrei al.
deÐning di†erent morphological types. However, the com-
parison is crude and takes no account of more complex
sampling and noise e†ects.
4.3. Results from Detailed Simulations
To quantify the redshift biases more precisely, we decided
to incorporate each of the e†ects that combine to give the
Ðnal HST appearance. The method adopted is based on
that described in greater detail by Freedman, &Abraham,
Madore and can be summarized as follows:(1997)
1. For each pixel we calculated the colors ; toBJ[R
avoid edge e†ects we smoothed the images slightly before
calculating the colors.
2. The pixel colors were then used to select anBJ[R
SED for each pixel (as for the integrated colors discussed
earlier). This SED was then used to determine the
k-correction applicable to each pixel.
3. The Ðnal step was to rebin the image, using the known
and wanted redshift, applying (1 ]z)4surface brightness
126 BRINCHMANN ET AL. Vol. 499
FIG. 8.ÈUpper panel: Distribution of galaxies in the A-Cplane. Only
objects with area larger than 64 contiguous pixels are shown. Bottom
panel: ClassiÐcation angle #(see text) plotted vs. eyeball classiÐcation for
each object. Large circles show the median of each eyeball class.
dimming and adding noise corresponding to the character-
istics of WFPC2 for our mean exposure time.
For each redshift we only analyzed those redshifted gal-
axies that would have been selected into either the CFRS or
LDSS-2 redshift surveys. Beyond z^0.9, however, only a
FIG. 9.ÈAbsolute-magnitude distribution of the et galaxiesFrei al.
with the selection limits appropriate for the deep HST survey overlaid. The
selection function for Sbc galaxies is indicated.
FIG. 10.ÈTop panel: Change in asymmetry Abetween the Rand BJ
images of the local et galaxies. Filled circles denote late-typeFrei al.
systems for which the trends are somewhat more pronounced. Bottom
Panel: As above, but for the concentration Cillustrating the inÑuence of
the bandshifting e†ects on Aand CclassiÐcations. Labeled lines deÐne the
boundary at which the change in Cwould move an object to a di†erent
class as indicated (assuming Ais at the median of the distribution).
few of the et galaxies would be seen. This is becauseFrei al.
the bright galaxies in the Frei sample tend to be Sab gal-
axies whose k-corrections are substantial when the 4000 A
break enters the IÐlter (this is also indicated in ThisFig. 9).
leads to a small number of galaxies and hence to an increase
in the statistical uncertainties. To rectify this problem, we
adopted an alternative approach. We selected the etFrei al.
galaxies that would have had i.e., 1 mag tooIAB \23.50,
faint, and brightened these galaxies so that they fell within
our selection criteria. This led to an average brightening of
mag, and kept This enabled us toSMBTB1.5 MBº[22.5.
get a satisfactory number of objects without introducing
any overluminous objects. Provided that the morphological
characteristics do not vary signiÐcantly over this small
luminosity range, this should not a†ect our conclusions.
The aim of the detailed simulations is to determine the
fraction of a given type of et galaxy that appears toFrei al.
be of a di†erent morphological type (as measured by Aand
C) at a chosen redshift z. For convenience we will denote the
number of objects in a given category by Nwith superscript
““ obs ÏÏ for the observed number and no superscript for the
true number in that class. We connect these two numbers
via a drift coefficient that characterizes the drift fromDXY
category Xto category Y, deÐned as
DXY \NX?Y
NX
, (2)
No. 1, 1998 CFRS AND LDSS SURVEY GALAXIES 127
where is the number of objects of class Xthat areNX?Y
classed as Yat a higher redshift. From the simulations we
can estimate for various redshifts.DXY
This methodology is similar to that of the k-correction
applied to convert an observed magnitude into a rest-frame
value. The ultimate aim here is to recover the rest-frame
morphology. Although the analogy fails in detail because of
the added complications of distance-dependent resampling
and surface brightness dimming, our simulations indicate
that these are second-order e†ects. If the central concentra-
tion is corrected for surface brightness e†ects as described in
the dominant cause of a migration to a di†erent mor-°3.2,
phological type is the pixel-by-pixel k-correction.
The analogy with the k-correction suggests two
approaches for analyzing our data. Given that we have
R-band images of the Frei galaxies and the local morpho-
logical studies are done in B, it seems natural to correct our
HST morphologies to those appropriate for the rest-frame
band. As we observe this rest wavelength directly atBJ
zB0.9, the corrections would be small at high redshift but
larger at low redshift. A problem with this approach is that
our benchmark sample is not ideal for studying the mor-
phological shifts at the low luminosities appropriate for the
nearby objects, owing to the low number of faint galaxies.
The alternative approach is to correct our HST morphol-
ogies to rest-frame R, assuming, as seems reasonable, that
the morphology changes little between rest-frame Iand R.
In this case, nearby intrinsically faint galaxies need no cor-
rection, and although a larger correction is needed at high
redshifts, in this case the et sample is well matchedFrei al.
in luminosity. Neither approach is perfect, but we consider
the latter to be more reliable until larger, multiband local
samples are available. We will therefore concentrate on
applying corrections to rest-frame Rmorphologies
(although, for completeness, we tabulate drift coefficients
appropriate for rest-frame BJ).
The results of this exercise are summarized in Table 4.
The numbers here are consistent with the earlier discussion
of the change in Cfrom the R- to images, but weBJ-band
consider the results to be more reliable, given the more
detailed treatment of the e†ects of sampling and back-
ground noise. In particular, the drift coefficients calculated
from R- to images agree excellently with the onesBJ-band
found here. The drift coefficients not listed were all found to
be zero in the simulations.
We can relate the observed number of objects in class X
to the true number through
NX
obs \ NX];
YEX
ClassesNYDYX [NX;
YEX
ClassesDXY . (3)
can be readily solved, given the observedEquation (3)
number of galaxies.
Broadly speaking, there are two dominant e†ects. First,
there is an apparent migration from AC spirals to AC pecu-
liars, if we classify galaxies from observed I-band images.
This bias is expected to occur also for eyeball classiÐcation.
At z\0.7 we can expect only 13% of the AC spirals to be
misclassiÐed as AC peculiars, whereas by z\0.9 the mis-
classed fraction grows to 24%. Presumably, this trend con-
tinues at higher redshift. Due to the restrictions in the local
data set discussed above, we have not extended the simula-
tions beyond z\0.9. U-band imaging would be highly
desirable to quantify bandshifting beyond z\1 (see also
& VaccaHibbard 1997)
In addition, there is a strong trend for AC ellipticals to
migrate into the AC-spiral category. The net result is some-
what less clear, as there is also a drift in the opposite sense
(from AC-S into AC-E) resulting from random measure-
ment errors on Cwhere the boundary is nearly vertical.
From the errors in it is evident that the driftTable 4
coefficients are uncertain. A larger sample of calibrating
galaxies is clearly required to make more precise statements
about the drift coefficients. Nevertheless, the principal con-
clusion of the simulations is a reasonably precise measure of
the proportion of high-redshift AC peculiars that are likely
to be genuine spiral galaxies. The mixture of regular spirals
and spheroidal galaxies should be faithfully observed with
HST out to redshifts z^1, but with a likely loss of AC-E at
high redshift that might be able to make up for the loss of
AC-S to the AC-P category. A montage of the AC peculiars
sorted by redshift and rest-frame [O II] equivalent width is
shown in (Plates 4È6).Figure 11
5.ANALYSIS
The Ðrst question we address relates to luminosity evolu-
tion as a function of galaxy class inferred independently
from the MDS and ground-based surveys. We wish to
understand these results in the context of the morphologi-
cally segregated redshift distributions and luminosity func-
tions (LFs) now available to us. Such results are of
considerable interest, as recent semianalytical models
TABLE 4
THE MOVEMENT OF FREI GALAXIES IN THE A-CPLANE
Redshift AC-E AC-S AC-P DSPbDPSbDSEbDESb
Drift CoefÐcients for Corrections to RRest-Frame Morphologies
z\0.0a...... 33 41 5 0 0 0 0
z\0.2 ...... 29 30 6 0 0 0.06 ^0.04 0
z\0.7 ...... 30 29 5 0.13 ^0.09 0 0.20 ^0.12 0.32 ^0.13
z\0.9 ...... 16 22 6 0.24 ^0.11 0 0.10 ^0.07 0.33 ^0.12
Drift CoefÐcients for Corrections to BJRest-Frame Morphologies
z\0.0a...... 33 41 5 0 0.67 ^0.21 0.29 ^0.08 0.05 ^0.05
z\0.2 ...... 29 30 6 0 0.67 ^0.24 0.35 ^0.10 0
z\0.7 ...... 30 29 5 0.05 ^0.05 0.67 ^0.47 0.43 ^0.14 0.30 ^0.17
z\0.9 ...... 16 22 6 0.07 ^0.05 0.20 ^0.20 0.26 ^0.10 0.21 ^0.12
aThe numbers for z\0 are for all Frei galaxies.
bThe errors are 1 pPoisson errors.
128 BRINCHMANN ET AL. Vol. 499
Cole, & Frenk & Fukugita(Baugh, 1996; Shimasaku 1997)
already claim to reproduce observed trends in the global
star formation history based on ground-based redshift
surveys and Lyman-limitÈselected samples in the Hubble
Deep Field However, a physical understand-(Madau 1997).
ing of these trends in the context of these models demands a
more detailed comparison, such as is now possible for each
of the various morphological types. Although hierarchical
assembly may transform late-type systems into more
regular spheroidal and disk galaxies, in principle, these
e†ects can be incorporated in such models.
In discussing the observational results, as mentioned in
we will concentrate only on the subset of the survey°2,
containing objects from the LDSS-2 and CFRS surveys.
This provides a total of 249 galaxies with secure redshifts
z\1.2. In order to implement the quantitative results on
redshift-dependent biases from we will restrict dis-°4,
cussion to types based on the AC classiÐcations. For 12
objects whose isophotal area is less than 64 pixels, or whose
images lie within the planetary camera, the Aand Cmea-
surements were replaced by eyeball classiÐcations. One
galaxy among the 249 was classed as compact. Given the
uncertainty associated with dealing with such an object, we
increased the error bar in the relevant redshift range accord-
ingly. In total, there are 24 spectroscopically conÐrmed
stars, four QSOs with z[1.2 and 35 objects with uncertain
redshift (note \1) or no redshift estimate at all. The failures
have been ignored in the analysis (and would not change
the main conclusions below if included).
5.1. Redshift Distributions
It will be helpful in discussing the observed type-
dependent redshift distributions N(z) to have no-evolution
predictions based on local LFs. To make these predictions
we adopted type-dependent Schechter LFs listed in Table 5.
For the spirals and ellipticals, these are updated versions of
those used by Glazebrook et al. (1995b) with /* adjusted to
give the observed fractions given by et al. inShanks (1984)
their sample. For the irregular/peculiar galaxiesbJ\16.7
we have adopted the late-type/irregular LF given by
et al. Using these local LFs and theMarzke (1994).
k-corrections discussed earlier, we calculate N(z) for the
LDSS-2 and CFRS galaxies using the selection criteria and
areas listed in For the spiral and spheroidal gal-Table 3.
axies, we are primarily interested in the luminosity scales at
the bright end, and so, even though there are considerable
uncertainties in the local LFs, the predictions based upon
them are nonetheless a useful guide.
As well as indicating the expected N(z) distributions for
the combined CFRS and LDSS-2 magnitude limits, appro-
priately weighted for the sample sizes involved, we also cal-
TABLE 5
ADOPTED LOCAL LUMINOSITY FUNCTION
km s~1 Mpc~1)(H0\50
/*
Hubble Type M*(bJ)a(Mpc~3)
E/S0 ......... [21.21 [1.00 1.39 ]10~3
Sab .......... [20.90 [1.00 6.0 ]10~4
Sbc........... [20.90 [1.00 1.1 ]10~3
Scd........... [20.90 [1.00 4.5 ]10~4
Sdm.......... [20.90 [1.00 2.6 ]10~4
Irr............ [20.29 [1.87 7.5 ]10~5
culate the e†ect on the distribution of a luminosity
evolution equivalent to a linear shift in M* with redshift
that amounts to 1 mag at z\1.0. We will refer to this
prediction as ““ mild evolution.ÏÏ We then determined the
observed number of objects as a function of AC class in
each redshift bin fully incorporating the e†ects of morpho-
logical bias as determined in the using°4, equation (3).
Since the local morphological mix for the luminosity func-
tions is based on B-band morphologies, we use the DXY
values for correction to morphologies, taken fromBJ-band
The N(z) distributions for the various types of ACTable 4.
class are compared to the no-evolution and mild-evolution
predictions in Figure 12.
shows that, taking into account the biases dis-Figure 12
cussed above, the number of high-zellipticals and spirals is
broadly consistent with the expectations based on N(m)
counts. Given the limited size of our survey, the possibility
of incompleteness and, especially, the reliance that has to be
made on the local LF normalization in such comparisons, it
is difficult to make a precise statement on the extent of any
luminosity evolution. A K-S test applied to the elliptical
N(z) is unable to distinguish between no evolution and mild
evolutionary predictions. The size of the spiral sample is
larger, but precise conclusions are difficult to make because
FIG. 12.ÈRedshift distribution of the three broad AC classes, com-
pared to theoretical predictions for no evolution (dashed line) and for mild
evolution (corresponding to 1 mag of luminosity evolution at z\1; dash-
dotted line). The models have been corrected to observed numbers using
the method outlined in the text.
No. 1, 1998 CFRS AND LDSS SURVEY GALAXIES 129
of the number of spirals without redshifts. Even so, the mild
evolutionary case is preferred: the no-evolution N(z)is
rejected at the 97% level, and this conÐdence level would be
stronger if the failures are at high redshift, as is most likely
the case. The drop in numbers past z\1 is most likely due
to incompleteness in the redshift determinations (see also
et al.Cowie 1996).
The most convincing result apparent from isFigure 12
the very considerable excess population of AC peculiars
beyond z^0.4, which cannot be explained through
residual uncertainties in the misclassiÐcation fraction as
applied to the regular spirals. The Ðgure shows quite clearly
how the large excess population recognized from early
ground-based studies arises primarily from these sources.
Note also that the form of the redshift distribution is skewed
to high redshift, and thus a mistaken normalization in the
local LF would not be helpful in reconciling the data with
the simple model predictions. The result is highly suggestive
of an evolutionary e†ect. Although it is important to
remember that the AC-peculiar category may include a
variety of physical types (see, in particular, the discussion of
I[Kcolors of these objects in et al. andGlazebrook 1997)
that, in principle, morphology itself may be a transient phe-
nomenon, understanding the strong redshift dependence in
the abundance of the AC peculiars is clearly a crucial goal
for making progress.
5.2. L uminosity Functions
exploits one aspect of the plane toSection 5.1 MB-z
discuss the properties of the survey galaxies; a disadvantage
in the interpretation of however, is the need toFigure 12,
assume a local LF and, particularly, its normalization. In
order to understand how the population of AC peculiars
evolves so dramatically to provide the excess population,
and to complement the study of AC-S and AC-E, it is there-
fore valuable to consider the form of the LF at various
redshifts. This can be calculated for the HST survey gal-
axies using a formalism.Vmax
As the CFRS and LDSS surveys have quite di†erent
selection criteria, we utilize an approach similar to that
adopted by et al. For each galaxy in the survey,Ellis (1996).
was calculated usingVmax
Vmax,i\;
j
SurveysVij ,
where is the normal accessible volume of galaxy iinVij
survey j. To cope with the di†erent selection criteria in the
two surveys, estimates of the magnitudes of the CFRSbJ
galaxies and magnitudes of the LDSS-2 galaxies areIAB
required. For the LDSS-2 objects we used the HST 3Amag-
nitudes and converted these to following methods dis-IAB
cussed above. For the CFRS objects we calculated bJ
magnitudes synthetically from Vphotometry similarly. The
method was then checked for a subset for which photo-BAB
metry was available and a good agreement was found. For
consistency, color terms were calculated using the same
interpolated SEDs that were used for the calculation of the
absolute magnitudes. The e†ective areas were taken from
The LFs estimated in this way are plotted for threeTable 3.
redshift bins in Error bars were determined usingFigure 13.
bootstrap resampling techniques.
The redshift biases determined in were incorporated°4.3
in a simple fashion. We only took account of the shift of AC
spirals into the AC-peculiar category in the highest redshift
bin. We then assigned 24% of the AC-peculiar class to the
AC-spiral class for each bootstrap repetition. The e†ects are
not signiÐcant in the following discussion.
A major disadvantage of multiobject redshift surveys
with narrow ranges in apparent magnitude is that they
provide very little overlap in LF across the di†erent redshift
bins. Nonetheless, the Ðgure shows that the LFs for AC
ellipticals do not exhibit strong evolution. The AC-spiral
LFs show a shift of B1 mag to z\1, but the greater com-
ponent occurs between the two lower redshift intervals. The
LFs for the AC peculiars clearly does show substantial evo-
lution from z\0toz\1, in the sense of either a dramatic
brightening with redshift or a substantial increase in the
volume density of luminous examples. It is difficult to quan-
tify this evolution precisely because of the lack of overlap
between the various redshift bins. In particular, we do not
observe any luminous AC-peculiar galaxies in the lower
redshift bin, even though they should be detected, given our
magnitude limits. However, as with the redshift distribu-
tions, it seems that the apparently declining population of
luminous AC peculiars has the dominant e†ect (as orig-
inally proposed by the MDS team), but with a signiÐcant
contribution also from the AC spirals.
5.3. L uminosity Densities
We now attempt to quantify the extent to which the evol-
ution that we have found for the AC peculiars contributes
to that observed for the overall population. Numerous
authors et al. et al.(Lilly 1996; Madau 1996 ; Madau 1997)
have expressed the overall evolution of the galaxy popu-
lation in terms of a rest-frame luminosity density or a
volume-averaged star formation rate as a function of red-
shift. Over the redshift range 0 \z\1, these articles inter-
pret the observations in terms of a substantial decline in star
formation activity at recent times. A crucial question is
whether virtually all of this decline arises from the rapidly
evolving irregular component discussed above.
We can address this by examining the rest-frame B(AB)
luminosity density of the galaxies in our HST survey as a
function of morphological type. The results of this exercise
are shown in where the errors bars are obtainedFigure 14,
through bootstrap resampling as before. To correct for the
fact that we observe only a limited magnitude range, we
Ðtted Schechter functions to the luminosity functions in
with the faint-end slope Ðxed to a\[0.5 forFigure 13,
AC-E and a\[1 for AC-S and AC-P. It is clear from the
Ðgure that this leads to considerable uncertainties. In par-
ticular, the Ðt for AC-P in the low-redshift bin is too uncer-
tain to be useful. The corrected values have been plotted as
open symbols in Figure 14.
Since the correction for AC peculiars at low redshift is
unknown, one could argue that their rapid rise in the blue-
light density is just an artifact of the limited magnitude
range sampled at low redshift. However, it is interesting to
note from that we do not see any bright ACFigure 13
peculiars, even though they would lie within the selection
criteria. This constrains the luminosity function for this
region, and a conservative upper limit to the correction
factor is 1.5. Thus, we would argue that the rapidly increas-
ing contribution to the blue light by the AC peculiars is a
robust result. Over z^0.3È0.9, this class provides an order
of magnitude increase in the detected luminosity density in
a given magnitude-limited sample, consistent with their
dominant e†ect in the HST galaxy counts over the range
130 BRINCHMANN ET AL. Vol. 499
FIG. 13.ÈLuminosity functions as a function of AC class binned as in the CFRS analysis et al. The dotted line is the CFRS 0.2\z\0.5(Lilly 1995b).
luminosity function.
detected by the surveys. The AC spirals contribute a smaller
amount to the overall evolution. Their contribution is not
so evident in the HST galaxy counts and can be attributed
to our being able to correct for redshift-dependent e†ects in
the classiÐcations. At high redshift galaxies would be assign-
ed too low C-values because of surface brightness dimming.
The correction in rectiÐes for this and henceAppendix A
demonstrates the importance of allowing for redshift-
dependent e†ects.
5.4. T he Physical Nature of the Star-forming Galaxies
Given that galaxies of irregular morphology are rapidly
evolving with redshift in their abundance and/or mean
luminosity (and hence detected luminosity density), and
that this trend appears to dominate the evolution that we
see, at least in the context of galaxy counts selected accord-
ing to apparent magnitude, these remarkable systems
clearly hold the key to understanding the rapid demise in
star formation activity since z^1.
A basic question that we must address is whether this
category represents a single type of object evolving in iso-
lation, perhaps fading in luminosity after an energetic burst
of activity, or that we are witnessing the gradual transform-
ation of galaxies, rendered irregular in form simply by
virtue of their enhanced star formation, into more regular
systems? Progress might be made on this question if it
could be demonstrated robustly that the rapid evolution
seen in the AC peculiars does or does not occur at the
No. 1, 1998 CFRS AND LDSS SURVEY GALAXIES 131
rest-frame luminosity density of galaxies detected in theFIG. 14.ÈBAB
survey as a function of redshift and AC class. The values plotted are o†set
in redshift slightly for clarity. The downward arrow indicates the e†ect of a
bandshifting correction of 24% for the AC-P class. The open symbols
represent the corrected luminosity densities, as discussed in the text.
expense of accompanying changes in the regular spirals.
That the high-redshift volume density of spirals is consistent
with local estimates does imply, at Ðrst sight, that the AC
peculiars might evolve independently of the more modest
changes that a†ect the spirals. Of course, a decline in star
formation activity in a long-lived class, e.g., the spirals, will
most likely be accompanied by a drop in luminosity. Such
transformations may not always occur above the detection
limits of the survey, so it is difficult to use this argument
with conÐdence. Also, conclusions derived on the basis of
number conservation are uncertain in our case, given our
small sample size and the remaining e†ects of redshift
incompleteness. Greater progress may be possible in con-
straining the growth of the spiral population by attempting
to examine a well-deÐned subset, taking due care to allow
for the e†ects of size and surface brightness. This is the
approach adopted in Paper II.
Finally, as we have emphasized, in morphological terms
the AC peculiars represent a mixture of very di†erent
objects The category includes some very late-type(Fig. 11).
spirals whose asymmetries and central concentrations place
them in the A-Cplane normally occupied by low-zirregu-
lars, double systems most likely in the act of merging and
other peculiar systems that defy classiÐcations in the
normal Hubble sequence. in this series examinesPaper IV
the structural properties and merger statistics in an attempt
to quantify the dominant subprocesses that may drive the
evolution in this class.
Examination of suggests that a signiÐcant frac-Figure 11
tion of the AC peculiars with high [O II] equivalent widths
appear to be rather compact. These could be the more
extreme examples of the star-forming population that domi-
nates the evolutionary trends discussed above. In this
context it is interesting to consider the et al.Guzman (1997)
claim that a considerable fraction of the star formation
activity seen at high redshift occurs in ““ compactÏÏ galaxies.
The compact galaxies in their study were selected within the
Hubble Deep Field Ñanking Ðelds on the basis of apparent
magnitude, angular size, and average surface brightness.
Although their sample is somewhat smaller than that
analyzed here, both their magnitude and surface brightness
limits are generally fainter. However, our HST exposure
times are generally longer.
A key point in understanding the et result inGuzman al.
the context of this paper, where the bulk of the evolution
occurs in galaxies of irregular morphology, is the precise
deÐnition of a ““ compact ÏÏ galaxy. et adopted aGuzman al.
half-light radius as well as a surface brightnessr1@2¹0A.5,
selection criterion et al. We can ask how(Phillips 1997).
many of our regular and irregular galaxies would fulÐll this
compactness criterion at various redshifts. From our data
we Ðnd that as many as 37% of the AC spirals and 26% of
AC peculiars beyond z\0.5 have andr1@2\0A.5 kF814 [
22.24.
We may continue the comparison by examining the
volume-averaged star formation rate (SFR) for the galaxies
detected in our survey, based, as in et al. onGuzman (1997),
the measurements of the equivalent width (EW) of [O II]
and the relation between EW[O II] and SFR et al.(Guzman
1997),
SFR(M_yr~1)B2.5 ]10~12 ]10~0.4(MB~MB_)EW*OII+ .
(4)
This corresponds to in the simple model of(U[V)AB \0.7
et al. color that is representative for theHammer (1997)Èa
galaxies in the higher redshift interval.
Adopting the same redshift intervals as for the luminosity
density and LFs, we show the results based on equation (4)
in It is interesting to note how closely this ÐgureFigure 15.
corresponds to Relative to low redshift, the AC-Figure 14.
peculiar category causes the strongest evolution, even
though the evolution in the AC-S and AC-P categories is
comparable at high redshift. As before, the absolute values
are highly uncertain, both because they refer only to the
detected population and because of the approximate con-
version of [O II] to SFR (see, in particular, the extensive
discussion of the properties of the CFRS objects in Hammer
et al. Nonetheless, the strong morphological trends1997).
are clear. Finally, we note that the galaxies in our sample
satisfying the et selection criteria contributeGuzman al.
26% of the star formation rate between z\0.5 and z\1.0.
This is in reasonable agreement with the etGuzman al.
results when the fact that their survey goes 1.7 mag fainter
than ours is taken into account.
In summary, there seems to be reasonable agreement
between the present work and et al. in theGuzman (1997)
physical properties of the faint galaxies claimed to dominate
FIG. 15.ÈStar formation density for detected sources as a function of
redshift for the three AC classes calculated using The upwardequation (4).
arrows show the total in each redshift bin. The downward arrow indicates
a change of 24% in for AC peculiars.oSFR
132 BRINCHMANN ET AL. Vol. 499
the evolutionary trends. The overlap with our own result is
understandable, although we would argue that no great
signiÐcance can be attached to the label ““ compact,ÏÏ e.g., in
considering the present-day equivalent of such systems or in
describing the sources in Figure 11.
6.CONCLUSIONS
We have analyzed the HST images of 341 galaxies drawn
from both the CFRS and AutoÐb/LDSS ground-based red-
shift surveys. Our catalog includes new spectroscopy of a
magnitude-limited sample in the Groth strip. In this, the
Ðrst paper in this series, we have analyzed the HST data in
conjunction with the available spectroscopic redshifts in
order to understand the evolutionary trends identiÐed inde-
pendently from the morphological studies in the Medium
Deep Survey and from the redshift-dependent luminosity
functions derived from our two extensive ground-based red-
shift surveys.
We summarize our Ðndings as follows:
1. We have extended the automated classiÐcation scheme
developed by et al. which places galaxiesAbraham 1996b,
in three broad categories (ellipticals/spirals/peculiars), and
have quantiÐed, via simulations and other tests, the likeli-
hood of misclassiÐcation when such systems are viewed at
high redshift. For the typical HST exposure times involved
in this survey, systematic misclassiÐcations occur in the
sense of shifting normal galaxies to apparently later Hubble
types at Ðxed observed wavelength. We quantify how this
can be taken into account when redshifts are available.
2. Taking these biases into account, we demonstrate that
the number-redshift relation for regular AC ellipticals is
consistent with expectations on the basis of no evolution or
““ mild ÏÏ evolution (corresponding to 1 mag luminosity evo-
lution by a redshift 1). However, the numbers are too small
to di†erentiate these possibilities. In the case of the AC
spirals, models incorporating mild evolution are preferred,
and the luminosity function indicates luminosity evolution
of about 1 mag to zD1.
3. The number of galaxies with irregular morphology
increases with redshift well beyond what is reasonably
expected on the basis of systematic misclassiÐcations of
spirals. We conclude there is a signiÐcant evolutionary
signal conÐned to this population. Analysis based on the
luminosity density and luminosity functions divided by
morphological class conÐrms that the demise in the AC-
peculiar population is a dominant component in the recent
evolution of the galaxy population.
4. There is no obvious decline with redshift in the abun-
dance of regular galaxies, as might be the case if the AC-
peculiar population is transforming into more familiar
systems. However, this conclusion is not particularly
robust, given that it relies on uncertain volume densities in
our modest sample. Such quantitative interpretations are
also hampered by the fact that the classiÐcations may well
be transient, and that luminosity fading may remove
systems from the sample at lower redshift. A variety of dif-
ferent physical sources may contribute to the peculiar cate-
gory (including misclassed late-type spirals, genuine
irregulars, merging systems, and starburst galaxies).
We acknowledge useful discussions with Simon White,
Carlos Frenk, Joe Silk, and Alan Dressler. We also acknow-
ledge the invaluable contributions provided by all STScI
sta† involved in the HST project. J. B. was supported by
The Research Council of Norway, project number 107798/
431. B. A. acknowledges support from PPARC.
APPENDIX A
CORRECTION OF C
A primary advantage of the use of the asymmetry and concentration indices (Aand C) in our survey, compared to its
equivalent application to the Medium Deep Survey data, is that redshifts are available for all the galaxies in the sample. It is
therefore of interest to consider how to optimize the measurement of these parameters to account for the variation of the
limiting rest isophote.
In measuring C, a limiting isophote is selected relative to the background sky. Since the HST images generally havekl
similar limiting isophotes, regardless of the galaxy redshift, in the rest frame there are shifts with redshift that scale as
10 log (1 ]z). To correct for this e†ect, one can either attempt to measure Cwith the same rest-frame isophote, or one can
correct the measurements with an average shift calibrated with simulations. The former approach has the disadvantage of
discarding information at low redshift, since, to maintain uniformity, the adopted rest-frame threshold must be set fairly high
to accommodate the high-zimages. Given the large redshift range in our sample, the latter approach is preferred. For
symmetric light proÐles the central concentration is given by
C\f(0.3R)
f(R), (5)
where Ris the radius at which the surface brightness is equal to the limiting isophote selected, and f(R) is given by
f(R)\2nP0
RI(r)rdr .
can be calculated exactly for a pure de Vaucouleurs law,Equation (5)
I(r)\Ieexp M[7.67[(r/re)1@4[ 1]N.
No. 1, 1998 CFRS AND LDSS SURVEY GALAXIES 133
The resultant expression for the central concentration is then
CdV \g(V)
g(0.31@4V),V\7.67Akl[ke
8.3276 ]1B, (6)
with g(V) being given by
g(V)\P0
Vv7e~vdv \7!A1[e~x;
k/0
7xk
k!B.
It is worth noting that the expression for Cis independent of details of the proÐle, except for the central surface brightness,
or, equivalently, the surface brightness at the e†ective radius. This statement remains true for exponential disks but breaks
down for mixed proÐles, where both the exponential scale length and the e†ective radius come into play.
Accordingly, we therefore have a relation that can be used to calculate the change in Cthat occurs as a function of redshift:
*C\g(z\0, This is a function of the details of the proÐle, but not strongly so. In practice, we used thek0)[g(z\z,k0).
functional form and calibrated the correction from the redshifted et images (see This enabled us to adopt theFrei al. °4.2).
following relation to correct C:
*C(z)\0.6[CdV(0, k0\15.0) [CdV(z,k0\15.0)] , (7)
with given in above. We will refer to this as a minimal correction, as it does not take into account bandshiftingCdV equation (6)
e†ects, and it has been applied to all Cvalues in this paper.
REFERENCES
B., Freedman, W., & Madore, B. 1997, in Proc. 36th Herstmon-Abraham,
ceux Conf., HST and the High-Redshift Universe, ed. N. Tanvir,
A. Aragon-Salamanca, & J. V. Wall (Singapore: World ScientiÐc)
R. G., Tanvir, N. R., Santiago, B. X., Ellis, R. S., Glazebrook, K.,Abraham,
& van den Bergh, S. 1996a, MNRAS, 279, L47
R. G., Valdes, F., Yee, H. K. C., & van den Bergh, S. 1994, ApJ,Abraham,
432, 75 R. G., van den Bergh, S., Ellis, R. S., Glazebrook, K., Santiago,Abraham,
B. X., Griffiths, R. E., & Surma, P. 1996b, ApJS, 107, 1
C. M., Cole, S., & Frenk, C. S. 1996, MNRAS, 282,Baugh, L27
R., Broadhurst, T., & Silk, J. 1997, preprintBouwens, astro-ph/9710291
M., Ellis, R. S., Broadhurst, T. J., Taylor, K., & Peterson, B. A.Colless,
1993, MNRAS, 261, 19
L. L., Songaila, A., Hu, E. M., & Cohen, J. G. 1996, AJ, 112,Cowie, 839
D., Le Fe` vre, O., Lilly, S., & Hammer, F. 1995, ApJ,Crampton, 455
S. P., Windhorst, R. A., & Griffiths, R. E. 1995a, ApJ, 453,Driver, 48
S. P., Windhorst, R. A., Ostrander, E. J., Keel, W. C., Griffiths,Driver,
R. E., & Ratnatunga, K. U. 1995b, ApJ, 449, L23
R. S. 1995, in IAU Symp. 164, Stellar Populations, ed. P. van derEllis,
Kruit & G. Gilmore (Dordrecht: Kluwer), 291
R. S., Colless, M., Broadhurst, T., Heyl, J., & Glazebrook, K. 1996,Ellis,
MNRAS, 280, 235
Z., Guhathakurta, P., & Gunn, J. E. 1996, AJ, 111,Frei, 174
N. 1986, ApJ, 303,Gehrels, 336
K., Abraham, R., Santiago, B., Ellis, R., & Griffiths, R. 1997,Glazebrook,
preprint K., Ellis, R. S., Colless, M., Broadhurst, T., Allington-Smith,Glazebrook,
J., & Tanvir, N. 1995b, MNRAS, 273, 157
K., Ellis, R., Santiago, B., & Griffiths, R. 1995a, MNRAS, 275,Glazebrook,
L19 R. E., et al., A. J. 1994, ApJ, 435,GrifÐths, L19
E. J., Kristian, J. A., Lynds, R., OÏNeil, Earl J., J., Balsano, R.,Groth,
Rhodes, J., & Idt, W. 1994, BAAS, 185, 5309
R., Gallego, J., Koo, D. C., Phillips, A. C., Lowenthal, J. D.,Guzma n,
Faber, S. M., Illingworth, G. D., & Vogt, N. P. 1997, ApJ, 489, 559
F., et al. 1997, ApJ, 481,Hammer, 49
J., Colless, M., Ellis, R. S., & Broadhurst, T. 1997, MNRAS, 285,Heyl, 613
J. E., & Vacca, W. D. 1997, AJ, 114,Hibbard, 1741
J. A., Burrows, C. J., Casertano, S., Hester, J. J., Trauger, J. T.,Holtzman,
Watson, A. M., & Worthey, G. 1995, PASP, 137, 1065
Fe` vre, O., Crampton, D., Lilly, S. J., Hammer, F., & Tresse, L. 1995,LeApJ, 455, 60
Fe` vre, O., et al. 1998, in preparation (PaperLe IV)
S., et al. 1998, ApJ, 499, in press (PaperLilly, II)
S. J., Le Fe` vre, O., Crampton, D., Hammer, F., & Tresse, L. 1995a,Lilly,
ApJ, 455, 50
S., Le Fe` vre, O., Hammer, F., & Crampton, D. 1996, ApJ, 460,Lilly, L1
S. J., Tresse, L., Hammer, F., Crampton, D., & Le Fe` vre, O. 1995b,Lilly,
ApJ, 455, 108
P. 1997, in The Hubble Deep Field, ed. M. Livio, S. M. Fall, & P.Madau,
Madau (Baltimore : STScI)
P., Ferguson, H. C., Dickinson, M. E., Giavalisco, M., Steidel,Madau,
C. C., & Fruchter, A. 1996, MNRAS, 283, 1388
R. O., Geller, M. J., Huchra, J. P., Harold, G., & Corwin, J. 1994,Marzke,
AJ, 108(2), 437
A., et al. 1995, MNRAS, 274,Naim, 1107
S. C., Windhorst, R. A., Driver, S. P., & Keel, W. C. 1996, ApJ,Odewahn,
472, L13
A. C., Guzma n, R., Gallego, J., Koo, D. C., Lowenthal, J. D., Vogt,Phillips,
N. P., Faber, S. M., & Illingworth, G. D. 1997, preprint
D., et al. 1997, in preparation (PaperSchade, III)
D., Lilly, S. J., Crampton, D., Hammer, F., Le Fe` vre, O., & Tresse,Schade,
L. 1995, ApJ, 451, L1
T., Stevenson, P. R. F., Fong, R., & Macgillivray, H. T. 1984,Shanks,
MNRAS, 206, 767
K., & Fukugita, M. 1997,Shimasaku, preprint
S. D. M. 1996, in Science with Large Millimetre Arrays, ed.White,
P. Shaver, ESO Astrophysics Symposia (Berlin: Springer), 33
FIG. 11.ÈObjects classed as AC peculiar sorted by increasing redshift upward. The left column contains objects with EW[O II]\0 or unknown. The
galaxies in the middle column have 0 \EW[O II]\40, and the right-column galaxies contain objects with EW[O II][40. Every image is 6A]6A, with the
exception of 10.1650, which is 12A]12A.
BRINCHMANN et al. (see 499, 127)
PLATE 4
FIG. 11.ÈContinued
BRINCHMANN et al. (see 499, 127)
PLATE 5
FIG. 11.ÈContinued
BRINCHMANN et al. (see 499, 127)
PLATE 6
