The Morphological Butcher-Oemler effect in the SDSS Cut&Enhance Galaxy Cluster Catalog
Extracted source text ยท arXiv astro-ph/0301302
\documentclass[proof]{pasj00}
\draft % this ensures double-spaced
%\documentclass[twocolumn]{pasj00}
\begin{document}
%%--- DRAFTCOPY: COMMENT OUT IF NOT NEEDED -------
%% Prints a large "DRAFT" diagonally across each page
%% Does not show up in TeXview
% \typeout{Prints "DRAFT" on each page; does not show in TeXView}
% \special{!userdict begin /bop-hook{gsave 200 30 translate
% 65 rotate /Times-Roman findfont 216 scalefont setfont
% 0 0 moveto 0.90 setgray (DRAFT) show grestore}def end}
%%------------------------------------------------
\SetRunningHead{Goto et al.}{Morphological Butcher-Oemler Effect}
%\Received{2002/03/21}%{yyyy/mm/dd}
%\Accepted{2002/03/21}%{yyyy/mm/dd}x
%\title{Morphological Butcher-Oemler effect\\ in the SDSS Cut~\&~Enhance Galaxy Cluster Catalog}
\title{Morphological Butcher-Oemler effect\\ in the SDSS Cut~\&~Enhance Galaxy Cluster Catalog}
%%% begin:list of authors
%
%\author{Tomotsugu \textsc{Goto}
%% \thanks{Present Address is
%%Department of Physics, Carnegie Mellon University,
%%5000 Forbes Avenue, Pittsburgh, PA 15213-3890 }
%}
%\affil{Institute for Cosmic Ray Research, University of
%Tokyo, Kashiwanoha, Kashiwa, Chiba 277-0882, Japan\\
%Department of Physics, Carnegie Mellon University,
%5000 Forbes Avenue, Pittsburgh, PA 15213-3890, USA }
%\email{tomo@cmu.edu}
\author{%
Tomotsugu \textsc{Goto}\altaffilmark{1,2,3},
Sadanori \textsc{Okamura}\altaffilmark{2},
Masafumi \textsc{Yagi}\altaffilmark{4},
% Michael L. \textsc{Balogh}\altaffilmark{6},\\
Ravi K. \textsc{Sheth}\altaffilmark{5},\\
Neta A. \textsc{Bahcall}\altaffilmark{6},
Shane A. \textsc{Zabel}\altaffilmark{7},
Michael S. \textsc{Crouch}\altaffilmark{7},
Maki \textsc{Sekiguchi}\altaffilmark{1} ,\\
James \textsc{Annis}\altaffilmark{8},
Mariangela \textsc{Bernardi}\altaffilmark{7},
% J. \textsc{Brinkmann}\altaffilmark{8},
Shang-Shan \textsc{Chong}\altaffilmark{7},
Percy L. \textsc{G{\' o}mez}\altaffilmark{7},\\
Sarah \textsc{Hansen}\altaffilmark{9},
Rita S. J. \textsc{Kim}\altaffilmark{10},
Adam \textsc{Knudson}\altaffilmark{7},
Timothy A. \textsc{Mckay}\altaffilmark{9},\\
and Christopher J. \textsc{Miller}\altaffilmark{7}
% Robert C. \textsc{Nichol}\altaffilmark{2},
% Kathy A. \textsc{Romer}\altaffilmark{2},\\
}
\altaffiltext{1}{Institute for Cosmic Ray Research, University of
Tokyo,\\ Kashiwanoha, Kashiwa, Chiba 277-0882, Japan}
\altaffiltext{2}{Department of Astronomy and Research Center for the
Early Universe,\\ School of Science, University of Tokyo, Tokyo
113-0033, Japan}
\altaffiltext{3}{yohnis@icrr.u-tokyo.ac.jp}
\altaffiltext{4}{
National Astronomical Observatory, 2-21-1 Osawa, Mitaka, Tokyo 181-8588,
Japan.}
\altaffiltext{5}{ Department of Physics and Astronomy
University of Pittsburgh\\
3941 O'Hara Street Pittsburgh, PA 15260 }
% \altaffiltext{6}{Department of Physics, University of Durham, South
% Road, Durham DH1 3LE, UK}
\altaffiltext{6}{Princeton University Observatory, Princeton, NJ 08544,
USA}
\altaffiltext{7}{Department of Physics, Carnegie Mellon University, \\
5000 Forbes Avenue, Pittsburgh, PA 15213-3890, USA}
\altaffiltext{8}{Fermi National Accelerator Laboratory, P.O. Box 500,
Batavia, IL 60510, USA }
\altaffiltext{9}{University of Michigan, Department of Physics,\\ 500
East University, Ann Arbor, MI 48109, USA}
\altaffiltext{10}{Department of Physics and Astronomy, The Johns Hopkins
University,\\ 3400 North Charles Street, Baltimore, MD 21218-2686, USA}
% \altaffiltext{8}{Apache Point Observatory,\\
% 2001 Apache Point Road,
% P.O. Box 59, Sunspot, NM 88349-0059, USA}
%%% end:list of authors
%%% Please use the following style in case that sorting by
%%% affilation is impossible.
%
% \author{%
% D-Firstname \textsc{D-Familyname}\altaffilmark{1}
% E-Firstname \textsc{E-Familyname}\altaffilmark{1,2}
% and
% F-Firstname \textsc{F-Familyname}\altaffilmark{2}}
% \altaffiltext{1}{Address of Institute}
% \email{ddddd@xxx.xxx.xx.xx}
% \email{eeeee@xxx.xxx.xx.xx}
% \altaffiltext{2}{Address of Institute}
%% `\KeyWords{}' always has to be placed before `\maketitle'.
\KeyWords{galaxies: clusters: general} %Do NOT move this preamble from here!
\maketitle
\begin{abstract}\label{abstract}
We investigate the evolution of the
%present an analysis of the
fractions of late type
cluster galaxies as a function of redshift, using one of the largest,
most uniform cluster samples available. The sample consists of 514 clusters of galaxies in the range
0.02$\leq$z$\leq$0.3 from the Sloan Digital Sky Survey Cut \& Enhance
galaxy cluster catalog.
This catalog was created using a single
automated cluster finding algorithm applied to uniform data from a single
telescope, with accurate CCD photometry, thus,
minimizing selection biases.
%, which were prevalent in the previous Butcher-Oemler type of
% analyses.
We use four independent methods to analyze the evolution of the late
type galaxy fraction.
% We analyze the fraction of late type galaxies in four independent ways.
Specifically, we select late type galaxies based on: restframe
$g-r$ color, $u-r$ color, galaxy profile fitting and concentration
index.
The first criterion corresponds to the one used in
the classical Butcher-Oemler analyses. The last two criteria are more sensitive to the
morphological type of the galaxies. In all the four cases, we find an
increase in the fraction of late type galaxies with increasing
redshift, significant at the 99.9\% level.
%with greater than a 99.9\%
%significance level.
The results confirm that cluster galaxies do
change colors with redshift (the Butcher-Oemler effect) and, in addition,
they change their morphology to later-type toward higher redshift --- indicating a
morphological equivalent of the Butcher-Oemler effect.
We also find a tendency of richer clusters to have lower fractions of late type
galaxies. The trend is consistent with a ram pressure
stripping model,
% where richer clusters have more effective ram
% pressure due to their higher temperature.
where galaxies in richer clusters are affected by stronger ram
pressure due to higher temperature of clusters.
%
% Although a ram-pressure stripping model is consistent with all of our
% observational results, there remains a possibility for other mechanisms
% to be responsible for the evolution of cluster galaxies. Thus further
% work is needed to conclude about an underlying physical mechanism
% causing these effects.
% The underlying physical mechanism causing these effects is
% unclear, but our results may suggest mechanisms which affect the gas supply
% (e.g. ram-pressure stripping, galaxy infall) rather than
% which are accompanied by an increased star formation such as
% dynamic processes such as
% galaxy harassment, galaxy-galaxy interaction/mergers,
% or gas depletion through enhanced star formation.
%The uniformity of the catalog throughout the redshift
% range used combined with the accurate photometry of the Sloan Digital Sky Survey data.
% As well as the blue galaxy fraction used in
% Butcher-Oemler analysis,
\end{abstract}
\section{Introduction}\label{intro}
The Butcher-Oemler effect was first reported by Butcher \& Oemler
(1978, 1984) as an increase in the fraction of blue galaxies ($f_b$) toward
higher redshift in 33 galaxy clusters over the redshift range
0$<$z$<$0.54.
% They calculated the fraction of blue galaxies as those whose
% $B-V$ colors are bluer than by 0.2 mag bluer than the cluster E/S0
% ridge line and within the radius containing 30\% of cluster galaxies.
% in 33 galaxy clusters over the redshift range
% 0$<$z$<$0.54 and found the fraction of blue galaxies increase with
% increasing redshift.
Butcher and Oemler's work made a strong impact since it showed direct evidence for the
evolution of cluster galaxies. Much work regarding the nature of
these blue galaxies followed. Rakos \& Schombert (1995) found that the
fraction of blue galaxies increases from 20\% at z=0.4 to 80\% at
z=0.9, suggesting that the evolution in clusters is even stronger than
previously thought. Margoniner \& De Carvalho (2000) studied 48
clusters in the redshift range of 0.03$<$z$<$0.38 and detected a strong
Butcher-Oemler effect consistent with that of Rakos \& Schombert (1995).
Despite the trend with redshift, almost all previous work has reported
a wide range of blue fraction values.
In particular, in a large sample of 295 Abell clusters, Margoniner et al. (2001)
not only confirmed the existence of the Butcher-Oemler effect,
but also found the blue fraction depends on cluster richness.
% Kodama \& Bower (2001) used 7 clusters
% to find an excess of blue galaxies.
Although the detection of the Butcher-Oemler effect has been claimed in
various studies, %and many evolutionary theories have been proposed,
there have been some suggestions of strong selection biases in the cluster samples.
Newberry, Kirshner \& Boroson (1988) measured velocity
dispersions and surface densities of galaxies in clusters and found a marked difference between local clusters and intermediate
redshift clusters. More recently, Andreon \& Ettori (1999)
measured X-ray surface brightness profiles, sizes, and luminosities of the
Butcher-Oemler sample of clusters and concluded that the sample is not
uniform. The selection bias, thus, could mimic evolutionary
effects. Smail et al. (1998) used 10 X-ray bright clusters in the
redshift range of 0.22$\leq$z$\leq$0.28 and found
that the clusters have only a small fraction of blue galaxies. The
Butcher-Oemler effect was not observed with their sample.
Similarly, galaxies in radio selected groups are not significantly
bluer at higher redshifts (Allington-Smith et al. 1993).
Garilli et al. (1996) observed 67 Abell and X-ray selected clusters and
found no detectable Butcher-Oemler effect until z=0.2. Fairley et
al. (2002) studied eight X-ray selected galaxy clusters and found no
correlation of blue fraction with redshift. Rakos \& Schombert (1995)'s
sample were selected from the catalog compiled by Gunn, Hoessel \& Oke
(1986) using photographic plates taken only in two color bands.
% only two color photographic plates.
The sample
thus have a possible bias against red, high redshift clusters.
In addition to the possible sample selection biases, with the exception of Margoniner et
al. (2001), the number of clusters in the previous works was small, consisting of a few to dozens of clusters. Therefore the statistical
uncertainty was large. Many authors also noted
that cluster-to-cluster variation of the fraction of blue galaxies is considerable.
The need for a larger, more uniform sample of clusters has been evident.
There have been various attempts to find another
physical mechanism causing the large scatter which has been seen in
almost all previous work.
Wang \& Ulmer (1997) claimed the existence of a correlation between the blue fraction
and the ellipticity of the cluster X-ray emissions in their sample of
clusters at 0.15$\leq$z$\leq$0.6. Metevier, Romer \& Ulmer (2000) showed that two clusters with a bimodal X-ray surface brightness profile have an unusually high blue fraction
value and thus do not follow the typical Butcher-Oemler relation. They
claimed that the Butcher-Oemler effect is an environmental phenomenon
as well as an evolutionary phenomenon. Margoniner et al. (2001) found
a richness dependence in the sense that richer
clusters have smaller blue fractions. They claimed that this
richness dependence causes a large scatter in the blue fraction--redshift
diagram. Therefore, it is of extreme interest to explore an origin
of the scatter in the blue fraction despite the redshift trend.
At the same time, various studies using morphological information have
reported a similar evolution effect in cluster galaxies.
Dressler et al. (1997) studied 10 clusters at
0.37$<$z$<$0.56 and found a steep increase in the S0 fraction toward $lower$
redshift, compared to nearby clusters studied earlier (Dressler 1980).
%They proposed that most of the cluster S0s are formed through the
%morphological transformation of spirals.
Couch et
al. (1994,1998) studied three clusters at z=0.31 and found their S0 fraction to be consistent with the
trend observed by Dressler et
al. (1997). Fasano et al. (2000) observed nine clusters at intermediate
redshifts (0.09$<$z$<$0.26) and also found an increase in the S0 fraction
toward lower redshift. %, which was consistent with Dressler et al. (1997).
It has been proposed that the increase in the S0 fraction is caused by the
transformation of spiral galaxies into S0 galaxies through a process
yet unknown. These studies, however, need to be pursued further, considering that most
of the previous work was based on morphological galaxy
classification by eye. Although it is an excellent tool to
classify galaxies, manual classification could potentially have unknown
biases. (A detailed comparison of human classifiers can be found in
Lahav et al. 1995).
A machine based, automated classification
would better control biases and would allow a reliable determination of
% might not be as sophisticated as human eyes, but is preferred since it is possible and simple to
the completeness and false positive rate. A further reason to
investigate the evolution of cluster galaxies is the sample size.
The morphological fraction studies of clusters reported so
far are based on only dozens of clusters. Furthermore, the clusters
themselves have intrinsic variety in the fraction of blue/spiral
galaxies as reported by various Butcher-Oemler and morphological analyses
listed above. Since several authors have suggested that the fraction of
blue galaxies depends on cluster richness, it is important
to use a uniform cluster sample, preferably selected by an automated
method with a well known selection function.%, as opposed to one selected by human eyes.
Various theoretical models have been proposed to explain the Butcher-Oemler
effect and the increase of the S0 fraction. These models include ram pressure stripping
of gas (Spitzer \& Baade 1951, Gunn \& Gott 1972, Farouki \& Shapiro
1980; Kent 1981, Fujita 1998; Abadi, Moore \& Bower 1999, Fujita \& Nagashima 1999,
Quilis, Moore \& Bower
2000), galaxy infall (Bothun \& Dressler 1986, Abraham et al. 1996a,
Ellingson et al. 2001),
galaxy harassment (Moore et al. 1996, 1999), cluster tidal forces (Byrd
\& Valtonen 1990, Valluri 1993), enhanced star formation (Dressler \&
Gunn 1992),
and
removal and consumption of the gas (Larson,
Tinsley \& Caldwell 1980, Balogh et al. 2001, Bekki et al. 2002). It is, however, yet
unknown exactly what processes play major roles in changing the color and
morphology of cluster galaxies.
%E+A galaxies, the only class of galaxies thought to
% be associated with the Butcher-Oemler effect, are seen in
% the local field region as well (Zabludorff et al. 1996).
To derive a clear model explaining
the evolution of cluster galaxies, it is important to clarify both the
Butcher-Oemler effect and the S0 increase, using a large and uniform
cluster sample in conjunction with a machine based morphological classification.
With the advent of the Sloan Digital Sky Survey (SDSS; York et al. 2000),
which is an imaging and spectroscopic survey of 10,000 deg$^2$ of the
sky, we now have the opportunity to overcome these limitations. The SDSS Cut
\& Enhance galaxy cluster catalog (Goto et al. 2002a) provides a large
uniform cluster catalog with a well defined selection function.
% created by
%an automated method, from the photometric catalog alone.
The CCD-based, accurate photometry of the SDSS (Fukugita
et al. 1996, Hogg et al. 2002, Smith et al. 2002) and the wide coverage of the SDSS
on the sky allow accurate estimation of blue fraction with robust
local background subtraction. Although the SDSS is a ground based
observation, the state-of-the-art reduction software and the accuracy
of CCD data make it possible to derive morphological classification in an
automated way (Lupton et al. 2001, 2002). By using the SDSS data
set, we are able to study one of the largest samples to date --- 514 clusters --- to
the depth of $M_r$=-19.44 ($h$=0.75).
The purpose of this paper is as follows. We aim to confirm or disprove the existence
of the Butcher-Oemler effect using one of the largest, most
uniform cluster samples. At the same time, we
hope to shed light on the morphological properties of the Butcher-Oemler
galaxies using morphological parameters derived from the SDSS data.
%If we find a signal in the Butcher-Oemler effect or morphological
%transformation,
Finally we investigate the origin of the scatter in the galaxy type
fraction versus redshift
relation, in hope of gaining some understanding
% Once we confirm the Butcher-Oemler effect, our % parallel aim
% goal is to gain some understanding %glean clues
about the physical processes responsible for the scatter.
% If we find large scatter in the blue fraction as previous work
% reported, we investigate the cause of the scatter.
Since several people reported that field galaxies also evolve both
morphologically (Schade et al. 1996; Brinchmann et al. 1998;Lilly et
al. 1998; Kajisawa et al. 2001;Abraham et al. 2001;) and
spectroscopically
(Madau et al. 1996; Lilly et al.1996; Hammer et al. 1997; Treyer et al. 1998; Cowie et al. 1999; Sullivan et al. 2000; Wilson et al. 2002), it is of extreme
importance to compare the evolution of cluster galaxies with that of field galaxies to
specify a responsible physical mechanism.
It is possible that the Butcher-Oemler effect and morphological
transition of cluster galaxies are more commonly happening including
the field region of the universe, thus a cluster specific mechanism is
not responsible for the evolution of galaxies. However, since the SDSS spectroscopic data
are not deep enough to probe cosmologically interesting time scale, we
leave it to future work.
The paper is organized as follows: In
Section 2, we describe the SDSS data and the Cut \& Enhance cluster
catalog. In Section 3, we analyze the late type fraction,
both spectrally and morphologically. In Section 4, we discuss
the possible caveats and underlying physical processes in the evolution of galaxies. In
Section 5, we summarize our work and findings.
% Understanding how cluster populations can be so different at recent
% look-back times and determining the physical processes responsible for
% their subsequent transformation remain an important challenge in
% observational cosmology.
% We hope to shed some light on morphological properties of
% Butcher-Oemler galaxies. A parallel aim is to glean clues about the
% physical processes responsible for these changes.
The cosmological parameters adopted throughout this paper are $H_0$=75 km
s$^{-1}$ Mpc$^{-1}$, and ($\Omega_m$,$\Omega_{\Lambda}$,$\Omega_k$)=(0.3,0.7,0.0).
\section{Data}\label{data}
% In this section, we outline the data used in this paper.
The galaxy catalog used here is taken from the Sloan Digital Sky Survey (SDSS)
Early Data Release (see Fukugita et al. 1996, Gunn et al. 1998, Lupton
et al. 1999, 2001, 2002, York et al. 2000, Hogg et al. 2001, Pier et al. 2002, Stoughton et al. 2002 and Smith et al. 2002 for more detail of the SDSS data).
We use equatorial scan data, a contiguous area of 250 deg$^2$
(145.1$<$RA$<$236.0, -1.25$<$DEC$<$+1.25) and 150 deg$^2$
(350.5$<$RA$<$56.51, -1.25$<$DEC$<$+1.25).
The SDSS imaging survey observed the region to depths of 22.3, 23.3, 23.1, 22.3
and 20.8 in the $u,g,r,i$ and $z$ filters, respectively. (See
Fukugita et al. 1996 for the SDSS filter system, Hogg et al. 2002
and Smith et al. 2002 for its calibration).
Since the SDSS photometric system is not yet finalized, we refer to the
SDSS photometry presented here as $u^*,g^*,r^*,i^*$ and $z^*$. We
correct the data for galactic extinction determined from the maps given
by Schlegel, Finkbeiner \& Davis (1998).
%The seeing varied from 1.1'' to 2.5'' during the observation.
We include galaxies to $r^*$=21.5 (petrosian magnitude), which is the
star/galaxy separation limit (studied in detail by Scranton et
al. 2002) in the SDSS data.
% Among imaging galaxies, bright galaxies
% are also observed by the SDSS spectroscopic survey (Strauss et al. 2002, Eisenstein et
% al. 2002). The redshift information, thus, is available for about 40000
% galaxies brighter than $r^*\sim$17.77 in the region.
The galaxy cluster catalog used here is a subset of the SDSS Cut \& Enhance
galaxy cluster catalog (Goto et al. 2002a). There are 4638 clusters in
the equatorial region (See Kim et al. 2002, Annis et al. 2002, Bahcall
et al. 2002 and Miller et al. 2002 for other works on the SDSS
galaxy clusters). Besides the uniformity of the catalog with its well
defined selection function, the catalog has very good photometric redshifts,
$\delta z$=0.015 at $z<$0.3, which enables us to use a large sample
of clusters (Goto et al. 2002a; see also Gal et al. 2000 and Annis et
al. 2002 for photometric redshift methods for clusters). We use
clusters in the redshift range of 0.02$\leq$z$\leq$0.3
and galaxies brighter than $M_r=-19.44$, which corresponds to $r^*$=21.5
at z=0.3.
Since several authors in previous work claimed that
biases in sample cluster selection can mimic the evolutionary effect, it
is important to control cluster richness well. We use clusters with
more than 25 member galaxies between $M_{r^*}=-24.0$ and $M_r=-19.44$
within 0.7 Mpc
after fore/background subtraction, as explained in the next
Section. The large areal coverage of the SDSS data
enables us to subtract the fore/background counts reliably. We can thus
control the richness of the sample clusters well. The criteria
leave 514 clusters in the region.
% we assign the
% spectroscopic redshift using the galaxies observed by the SDSS
% spectroscopic survey since it is the key quantity in this study.
%We assign the redshift of clusters as the redshift of a galaxy if $i$, the galaxy
% exists within the radius given in Goto et al. (2002) and $ii$ the
% redshift of the galaxy is within $\pm$0.02 from the photometric redshift
% of the cluster. If multiple galaxies satisfy these criteria, we select
% the closest spectroscopic (galaxy) redshift to the photometric (cluster)
% redshift as the spectroscopic redshift of the cluster. 1282 galaxy clusters have
% redshift assigned in this way.
We stress the importance of the uniformity of the cluster
catalog. Although the Abell cluster catalog (Abell 1958, Abell, Corwin
and Olowin 1989) has been used in many studies, it was constructed by eye,
and is sensitive to projection effects.
% Cluster catalogs selected by eye could potentially
% contain unpredictable biases, which could affect scientific results.
When comparing clusters at different redshifts, it is particularly important
to ensure that the data quality and cluster selection techniques are
comparable, to avoid the introduction of potential selection biases.
% Also people sometimes combine the data from two different telescopes.
% For example, to study cluster evolution, data from a larger telescope
% with a high redshift sample might be combined with low redshift data
% from a smaller telescope. Furthermore, clusters selected using
% different algorithms may not agree with each other. Both these
% techniques bring unknown bias into a sample.
%
% the data from the bigger telescope for high redshift sample are sometimes
% combined with the
% data for low redshift sample from the smaller telescope to study the
% evolution. Often the clusters are selected using the different
% algorithm for low and high redshift data. Cluster catalogs selected using
% different algorithm not necessary agree with each other (Goto et
% al. 2002a), thus, could bring bias into the study.
% The data from different telescopes also have potential to contain the unknown
% biases,
% Ideally the better one can study the data from the same
% telescope for both low redshift and high redshift. Even better if the
% clusters are selected using a same algorithm.
The SDSS Cut \& Enhance cluster catalog used here is constructed using
only the
data taken with the SDSS telescope. Also, clusters
are detected using a single algorithm (CE) throughout the
entire redshift range (0.02$\leq$z$\leq$0.3). Combined with the well controlled
richness criteria, our cluster sample is not only one of the biggest
but also one of the most statistically uniform cluster catalogs.
To study colors of galaxies in clusters, it is also important to use a
cluster catalog created without targeting the red sequence of color
magnitude relation of cluster galaxies. For example, Gladders \&
Yee(2000) and Annis et
al. (2002) use a color filter targeting the red sequence of clusters
and find galaxy clusters efficiently without suffering from
projection effects. These techniques, however, can potentially have
biases with regards to the colors of detected galaxy clusters, since
clusters with a strong red sequence
is more easily detected. They may not, therefore, be ideally
suited to a Butcher-Oemler type of analysis. In contrast, the SDSS Cut \& Enhance
cluster catalog does not pick red galaxies selectively, and is therefore more
suitable for this study. (Note that the the Cut \& Enhance method does
use generous color cuts. Although the color cut is designed to be wide
enough to include blue galaxies in clusters,
therefore it is not completely free from color originated biases.) In
previous work,
clusters have often been detected using data from only one or two
color bands. This can introduce a bias, since
higher redshift clusters are redder and fainter than lower
redshift clusters.
The SDSS Cut \& Enhance cluster catalog detects clusters using
four bands of the SDSS data ($g,r,i$ and $z$), which minimizes the bias
against redshift.
% which not only overcomes this bias,
%but also helps to minimize projection effects, since we can
%use the colors to separate clusters aligned along the line of sight.
% and could potentially have had a bias towards systems more active in the band.A
% cluster catalog created using only one passband also suffer from a
% projection effect, while the SDSS Cut \& Enhance cluster catalog can
% separate projected clusters by using colors.
% These are advantages of using the SDSS Cut \& Enhance cluster catalog.
% Since the cluster catalog is constructed from the same
% data using the automated cluster detection method with even selection
% function, thus,
% it is free from human bias in eye selected cluster
% catalogs. In the previous works, sometimes photometric redshift of
% clusters are used to increase the sample. in this study, all sample
% clusters have spectroscopic redshifts.
\section{Analysis and Results}\label{method}
\subsection{Fore/Background subtraction}\label{bg}
%In the following four sections,
We compute fractions of late-type
galaxies in four different ways.
First, we describe the statistical fore/background
subtraction method we use.
In counting galaxies, all galaxies are assumed to be at the cluster
redshift to calculate absolute magnitudes. Then galaxies only between
$M_{r^*}=-24.0$ and $M_{r^*}=-19.44$ are used in the analysis.
We count the number of late-type/total galaxies within 0.7 Mpc from each
cluster.
%This ensures that the sizes
%of areas used for a cluster region and for fore/background estimation
% is the same.
Valotto et al. (2001) claimed that the global background
correction can not correct background contamination appropriately.
Following the claim, we use a local background correction. The
number of late-type/total galaxies in fore/background is estimated
using an annular area around each cluster with an inner
radius of 2.1 Mpc and an outer radius of 2.21 Mpc in
the same absolute magnitude range.
The annulus fore/background subtraction enables us to
estimate the fore/background locally, minimizing variations in galaxy
number counts due to the large scale structure.
When an outer annulus touches the boundary of the region, a
fore/background count is globally subtracted using galaxy number counts
in the entire 400 deg$^2$ region by adjusting it to the angular area each
cluster subtends.
This fore/background subtraction is used in the analyses described in
sections 3.2-3.5.
The fraction of blue/late-type galaxies and its error are computed
according to the following equations.
\begin{equation}
f_{late}=\frac{N_{c+f}^{late}-N_{f}^{late}}{N_{c+f}^{all}-N_{f}^{all}},\label{frac}
\end{equation}
% \begin{equation}
% \delta f_{late}= f_{late}\times \Bigg(\frac{\sqrt{ N_{c+f}^{late}+N_{f}^{late}}}{ N_{c+f}^{late}-N_{f}^{late} } + \frac{\sqrt{N_{c+f}^{all}+N_{f}^{all}}}{N_{c+f}^{all}-N_{f}^{all}}\Bigg),\label{frac_err}
% \end{equation}
\begin{equation}
\delta f_{late}= f_{late}\times \sqrt{\frac{ N_{c+f}^{late}+N_{f}^{late}}{( N_{c+f}^{late}-N_{f}^{late} )^2} + \frac{N_{c+f}^{all}+N_{f}^{all}}{(N_{c+f}^{all}-N_{f}^{all})^2} - \frac{2(\sqrt{N_f^{late}\times N_f^{all}}+ \sqrt{N_{c+f}^{late} \times N_{c+f}^{all}})}{(N_{c+f}^{late}-N_f^{late})(N_{c+f}^{all}-N_f^{all})} }
\label{frac_err}
\end{equation}
\noindent where $N_{c+f}^{late}$ and $N_{f}^{late}$ represent numbers
of blue/late-type galaxies in a cluster region and a field region,
respectively. $N_{c+f}^{all}$ and $N_{f}^{all}$ represent numbers
of all galaxies in a cluster region and a field region, respectively.
The equation (\ref{frac_err}) assumes that $N^{late}$ and $N^{all}$ are not
independent. We explain the derivation of equation (\ref{frac_err}) in
Appendix \ref{error}.
%Note that our error estimation is larger than some of previous work
%since we include uncertainty in fore/background subtraction correctly.
\subsection{Blue Fraction}\label{f_b}
The blue fraction of galaxy clusters ($f_b$) is measured as the fraction
of galaxies bluer in $g-r$ rest frame color than the color of the ridge
line of the cluster by
0.2 mag. This color criterion is equivalent to
Butcher \& Oemler's (1984) 0.2 mag in $B-V$ and Margoniner et al.'s (2000,
2001) 0.2 mag in $g-r$.
The color of the ridge line is measured from the color-magnitude
diagram using the same color-magnitude box used in measuring
photometric redshift (Goto et al. 2002a).
% determined from the redshift of the cluster as the model color of
% elliptical galaxies at that redshift in Fukugita et al. (1995).
The colors of the ridge lines are confirmed
to agree with empirical color of elliptical galaxies observed in the
SDSS at the same redshift with less than 0.05 difference in $g-r$ color
(Eisenstein, private communication).
We also use Fukugita et
al.'s (1995) model of an elliptical galaxy and a galaxy bluer than it by
0.2 mag in $g-r$. By redshifting these two galaxies, we measure
$\delta (g-r)$ in the observed frame, which corresponds to the
restframe $\delta (g-r)$=0.2 . In calculating $f_b$, we count
% $f_b$ is defined as the ratio of the number of bluer galaxies than
% ridge line color by more than 0.2 mag to the
% total number of the
galaxies within 0.7 Mpc from the center of each cluster, which is the
same radius as Margoniner et al. (2000,
2001), and corresponds to the average radius of Butcher \& Oemler
(1984). (We explore possible caveats in using fixed radius in Appendix.)
Galaxies between $M_{r^*}=-24.0$ and $M_{r^*}=-19.44$ are counted, which
corresponds to $r^*$=21.5 at $z$=0.3 for an average $k$-correction of all types of
galaxies (Fukugita et al. 1995). Compared to the field luminosity function
of Blanton et al. (2001), this includes galaxies as faint as $M^*_r$+1.36.
% The number of blue/total galaxies in fore/background is estimated in
% the same way using an annular area around the cluster of inner
% radius of 2.1 Mpc and outer radius of 2.21 Mpc. When annuli touch
% the boundary of the region, the fore/background is globally subtracted using the number counts in the
% entire region.
% In counting galaxies, all galaxies are assumed to be at the cluster redshift and galaxies
% between $M_{r^*}$=-24.0 and $M_{r^*}$=-19.44 are counted.
% To confirm the reality of the clusters and reduce the statistical
% uncertainty, we only use clusters with their total number of galaxies
% more than 25 after the background subtraction.
% The number of clusters in this richness range is 514.
Fore/background galaxies are statistically subtracted in the way described in Section \ref{bg}.
The lower left panel of Figure \ref{fig:bo} shows $f_b$ as a function of redshift.
The error in $f_b$ is estimated using equation (\ref{frac_err}) and
the median values of the errors in $f_b$ and z are shown in the upper
left corner of the plot.
% Note that our errors look larger than previous works since we take the
% errors from fore/background subtraction into consideration properly,
% while some previous works did not.
Dashed line shows the weighted least-squares fit to the data.
Solid lines and stars show the median values of the data.
The scatter is considerable, but both of the lines show a clear increase of $f_b$
toward higher redshift. The Spearman's correlation coefficient is
0.238 with significance of more than 99.99\% as shown in Table
\ref{tab:correlation}. The correlation is weak, but
of high significance. The lower left panel of Figure \ref{fig:kstest}
further clarifies the evolution effect. A dashed and a solid line show
normalized distributions of $f_b$ for clusters with z$\leq$0.15 and
0.15$<$z$\leq$0.3, respectively. The two distributions are significantly
different at the 98\% level, as determined by a Kolomogorov-Smirnov test.
% Although the origin of the scatter is unknown, we note that previous works also found large
% scatter in their $f_b$ v.s. z data (e.g. Butcher \& Oemler 1984, Rakos \&
% Schombert 1995, Garilli et al. 1996, Ellingson et al. 2001, Margoniner
% et al. 2000, 2001).
% Metevier, Romer \& Ulmer (2000) showed that clusters with
% a bimodal X-ray surface brightness profile have an unusually high $f_b$
% value. Since the Cut \& Enhance galaxy cluster catalog used here is
% designed to not miss any unusual clusters, the variety of the
% cluster sample could be one of the reasons.
%It is, however, inconsistent with zero with more than 99\%
% confidence level. To confirm the reality of the correlation, we shuffled the position of th%e cluster randomly and measured $f_b$ in the same way in the lower left panel of Figure \ref%{fig:random}. The correlation coefficience in this case is 0.0,
% assuring the reality of the relation in Figure \ref{fig:bo}.
The slope shown with the dashed line
rises up to $f_b\sim$=0.2 at $z$=0.3 (look back time of $\sim$3.5 Gyr),
which is consistent with previous
work such as Butcher \& Oemler (1978,1984), Rakos \& Schombert (1995),
Margoniner et al. (2000, 2001), within the scatter.
We conclude that the Butcher-Oemler effect is seen in
the SDSS Cut \& Enhance galaxy cluster catalog.
We caution readers on the systematic uncertainties in measuring $f_b$.
Marzke et al.(1994,1997,1998), Lin et al. (1999) and Blanton et
al. (2001) showed that luminosity functions of galaxy clusters depend
significantly on galaxy type, in such a way that
the bright end of the cluster luminosity function is dominated by
redder galaxies and the faint end is dominated by bluer galaxies.
Boyce et al.(2001) and Goto et al. (2002b) showed that a similar
tendency exists for cluster luminosity functions.
%Since spiral galaxies are generally bluer than elliptical galaxies,
This difference in luminosity functions leads to a different blue fraction depending
on the absolute magnitude range used. Furthermore, if the radial
distributions of blue and red galaxies are different ($e.g.$ Kodama et
al. 2001), the $f_b$ measurement
depends heavily on the radius. When comparing with previous work, therefore,
it is important to take account of the exact method used to calculate $f_b$.
%Since each author measures $f_b$ in a
% slightly different way, careful attention
% is necessary when comparing to previous works.
We discuss the
uncertainty in measuring blue fractions further in Section \ref{discussion}.
\subsection{Late Type Fraction Using $u-r<$2.2}\label{u-r}
% Several authors suggested morphological transition in clusters in the past (Dressler et
% al 1997, Fasano et al 2002.), the data are limited by the small number
% of sample clusters and by the difficulties in measuring galaxy
% morphology.
% Besides we have much larger number (514) of clusters in our catalog
% than previous works,
% Recently SDSS $u-r$ color has been proposed as a powerful classifier of galaxy morphology
% (Shimasaku et al. 2001, Strateva et al. 2001).
Recently Shimasaku et al. (2001) and Strateva et al. (2001) showed that
the SDSS $u-r$ color correlates well with galaxy morphologies.
In this section we use $u-r$
color to separate early($u-r\geq$2.2) and late ($u-r<$2.2) type galaxies
as proposed by Strateva et al. (2001). Note that although $u-r$ color
is claimed to correlate well with galaxy types, it is still a color
classifier and thus different from the morphological parameters
we investigate in the following two sections.
The methodology used to measure late type fraction is similar to the one we use
to measure $f_b$. We regard every galaxy with $u-r<$2.2 as a late type
galaxy. We define $f_{u-r}$ as the ratio of the number of late type
galaxies to the total number of galaxies within 0.7 Mpc.
Fore/background subtraction is performed in a way described in Section \ref{bg}.
% The fore/background counts are taken from the 2.1-2.21 Mpc annuli
% around each cluster in the same way as in the last section. If an
% annulus touches the boundary of the whole region, global fore/background
% subtraction is performed instead.
The upper left panel of figure \ref{fig:bo} shows $f_{u-r}$
as a function of redshift. The error in $f_{u-r}$ is calculated using equation
(\ref{frac_err}) and the median values of the errors in $f_{u-r}$ and z
are shown in the upper left corner
of the
panel. The dashed line shows the least square fit to the data.
The solid lines and stars show the median values of the data.
As in the case of $f_b$, the scatter is considerable, but the weak
increase of the late type galaxies is seen. The Spearman's correlation coefficient
is 0.234 and is inconsistent with zero at greater than 99.99\% confidence
level (Table \ref{tab:correlation}). Again, weak but significant
correlation is found. The upper left panel of Figure \ref{fig:kstest}
shows distributions of $f_{u-r}$ for z$\leq$0.15 clusters and
0.15$<$z$\leq$0.3 clusters with a dashed and solid line, respectively. A Kolomogorov-Smirnov
test shows that the distributions are different with more than a 99\%
significance.
In addition to the increase in $f_b$ shown in the last Section,
the increase in $f_{u-r}$ provides further evidence of color
evolution of cluster galaxies. Furthermore, since $u-r$ color of galaxies is
sensitive to a galaxy's morphology as shown in Figure 6 of Strateva et
al (2001), it suggests possible evolution of morphological types of galaxies
as well.
% Although the result has too low of a statistical weight to be
% entirely conclusive, this figure suggests that the number of morphological late type galaxies
% increases toward higher redshift as well as the blue galaxies seen in
% the Butcher-Oemler effect in the previous section.
% We also
% note that slope of the fitted line is almost as same as the one of blue
% fraction.
We investigate the morphological evolution of galaxies in clusters in the next subsection.
\subsection{Late Type Fraction Using Profile fitting} \label{exp}
One of the purposes of this paper is to determine if there is a
morphological change of galaxies in clusters as a function of
redshift.
The SDSS photometric pipeline (PHOTO; Lupton et al. 2002) fits a de Vaucouleur
profile and an exponential profile to every object detected in the SDSS imaging
data and returns the likelihood of the fit. By comparing the likelihoods
of having an exponential profile against that of a de Vaucouleur profile, we can classify
galaxies into late and early types. In this section, we
regard every galaxy that has an exponential likelihood higher than a de Vaucouleur
likelihood in $r$ band as a late type galaxy. A galaxy with higher
de Vaucouleur likelihood in $r$ band is regarded as an early type
galaxy.
% We discard galaxies whose de Vaucouleur and exponential likelihood are the same
% (14.9\% of total galaxies are affected by this).
We define $f_{exp}$ in the same way as in previous subsections. $i.e.$
$f_{exp}$ is the ratio of the number of late type galaxies to
the total number of galaxies within 0.7 Mpc. Fore/background counts are
corrected using the method described in Section \ref{bg}.
%2.1-2.21 Mpc annular region around each cluster,
% except at places where annuli touch the boundary of the survey data when global
% fore/background correction is applied.
%the global background correction when annuli touches the boundary.
The resulting $f_{exp}$ is plotted in the lower right panel of Figure
\ref{fig:bo}.
The error in $f_{exp}$ is estimated using equation (\ref{frac_err}) and
the median values of the errors in $f_{exp}$ and z are shown in the upper
left corner of the plot.
%The error plotted in the upper left corner is from equation (\ref{frac_err}).
The dashed line shows the weighted least-squares fit.
The solid lines and stars show the median values of the data.
The scatter is considerable, but we see the increase of $f_{exp}$ toward the higher redshift. The
Spearman's correlation coefficient is 0.194, which is inconsistent with zero at
more than a 99.99\% confidence level (Table \ref{tab:correlation}).
The upper right panel of Figure \ref{fig:kstest} shows the
distributions of clusters with z$\leq$0.15 and
with 0.15$<$z$\leq$0.3 with a dashed and solid line, respectively.
The two distributions show a difference of more than 99\% significance
in a Kolomogorov-Smirnov test.
We emphasize that the galaxy
classification used here is purely morphological --- independent
of colors of galaxies. The fact that we still see the increase of the
late type galaxies toward higher redshift suggests that these
Butcher-Oemler type blue galaxies also change their morphological
appearance as well as their colors. We also point out that the slope of the
change is similar to the panel in the lower left of Figure
\ref{fig:bo}, which is $\sim$30\% between z=0.02 and z=0.3.
We note that there is a potential bias associated with the use of $r$
band profile fitting throughout
the redshift range, since the $r$ band wavelength range at $z$=0.3 is almost that of $g$
band at restframe.
%The change in restframe wavelength to fit the profiles
% could invite a systematic bias in our analysis.
We investigate this effect in Section \ref{discussion}, and conclude that it is small.
Like the blue fraction, the morphological late-type fraction is also sensitive to the
magnitude range considered.
% The same uncertainty from the magnitude range used can be found in
% morphological classifications of galaxies, too.
Binggeli et al. (1988), Loveday et al. (1992), Yagi et al. (2002a,b) and Goto et al. (2002b) reported
luminosity functions of elliptical galaxies have brighter
characteristic magnitudes and flatter faint end tails compared to those
of spiral galaxies in both field and cluster regions. Careful attention should be paid to
the magnitude range used in an analysis when fractions of spiral galaxies are compared.
We discuss the uncertainty further in Section \ref{discussion}.
\subsection{Late Type Fraction Using Concentration Parameter} \label{Cin}
As another morphological galaxy classification method, we use the inverse of
the concentration parameter ($C_{in}$) advocated by Shimasaku et
al. (2001) and Strateva et al. (2001). We define $C_{in}$ as the
ratio of Petrosian 50\% radius to Petrosian 90\% radius in $r$
band. They are the radii which
contain 50\% and 90\% of Petrosian flux, respectively. (See Stoughton
et al. 2002 for more details of Petrosian parameters). Since $C_{in}$
is the inverse of a conventional concentration parameter, spiral
galaxies have a higher value of $C_{in}$. Following Strateva et al. (2001), we
use $C_{in}$=0.4 to divide galaxies into early and late type
galaxies. Readers are referred to Morgan (1958,1959), Doi, Fukugita \& Okamura (1993) and
Abraham et al. (1994, 1996) for previous usage of concentration of light
as a classification parameter.
$f_{Cin}$ is defined as the ratio of the number of galaxies
with $C_{in}>$0.4 to the total number of galaxies within 0.7 Mpc as
in the previous subsections. Note that our early type galaxies with
$C_{in}<$0.4 include S0 galaxies in addition to elliptical galaxies
since discerning elliptical and S0 galaxies is very difficult with
the SDSS data, in which the seeing is typically 1.5''. (See Shimasaku et
al. 2001 and Strateva et al. 2001 for the correlation of $C_{in}$
with an eye classified morphology).
Fore/background
number counts are corrected as described in Section \ref{bg}.
% using the 2.1-2.21 Mpc annulus if possible. Otherwise they are corrected
% from the global number counts.
The absolute magnitude range used is
$-24<M_r<-19.44$. The upper right panel of Figure \ref{fig:bo} shows $f_{Cin}$ as a function
of redshift. Since the classification using $C_{in}$=0.4 leans
toward late type galaxies, the overall fraction is higher than the
other panels in the figure. The increase of late type fraction,
however, is clearly seen. The dashed line shows the weighted least-squares
fit. The solid lines and stars show the median values of the data.
% The median error in the upper left comes from equation (\ref{frac_err}).
The error in $f_{Cin}$ is estimated using equation (\ref{frac_err}) and
the median values of the errors in $f_{Cin}$ and z are shown in the upper
left corner of the plot.
The Spearman's correlation coefficient is
0.223 with significance of more than 99.99\% as shown in Table
\ref{tab:correlation}. The upper right panel of Figure
\ref{fig:kstest} further clarifies the evolution effect. The
distribution of z$\leq$0.15 clusters in a dashed line and the
distribution of 0.15$<$z$\leq$0.3 clusters in a solid line show a
difference with more than a 99\% significance level.
We stress that the galaxy classification
based on this concentration parameter is purely a morphological one.
% Nevertheless,
In this morphological classification, we still see the increase of the late type fraction just
like the increase of $f_b$ --- as if observing the morphological equivalence of the
Butcher-Oemler effect.
% the late type fraction just in the same way as the increase of blue
% fraction in the lower left panel of the figure.
The increase in $f_{Cin}$ combined with the increase in $f_{exp}$
provides rather firm evidence of morphological change in the
Butcher-Oemler type
galaxies. Possible caveats in the usage of $C_{in}$ and comparisons
with previous works are discussed in Section \ref{discussion}.
\subsection{On the Origin of the Scatter} \label{scatter}
In the last four sections, we observed the increase of late type
fractions toward higher redshift in all the four cases. At the same time, we see
a significant amount of scatter around the late type fraction
v.s. redshift relations. Although the errors on these measurements are
also large, this scatter might suggest that there might be one or more physical
properties which determine the amount of late-type galaxies in
clusters. Table \ref{tab:error_comparison} compares median error sizes
of $f_b,f_{u-r},f_{exp}$ and $f_{Cin}$ with scatters around the
best-fit lines (dotted lines in Figure \ref{fig:bo}).
In fact, in all cases, real scatters are larger than the statistical errors.
%except for $f_{Cin}$. Perhaps, a straight line does not represent the
% relation very well in case for $f_{Cin}$ (Figure \ref{fig:bo}).
In the literature, several
correlations are proposed such as with
X-ray shapes of clusters (Wang et al. 2001, Metevier et al. 2000), and
cluster richness (Margoniner et al. 2000).
% Note that the errors are also quite large for these measurements.
Our cluster richnesses are plotted against redshift in Figure \ref{fig:z_rich}.
Richnesses are measured as numbers of galaxies between $M_r=-24.0$ and
$M_r=-19.44$
within 0.7 Mpc after fore/background subtraction, as explained in
Section \ref{bg}. In figure \ref{fig:z_rich}, this richness has no
apparent bias with redshift.
In Figure \ref{fig:okamura_rich}, the difference of the late type
fraction from the best-fit line is plotted against cluster richness.
The circles and solid lines show median values.
In all the panels, there is a clear tendency of richer clusters
having a lower fraction of late type galaxies. This tendency is in
agreement with Margoniner et al. (2000) who found richer clusters
had lower blue fractions. We further discuss the richness
dependence of the late type fraction in Section \ref{discussion} and Appendix.
As an alternative parameter to X-ray shape, we plot the difference from
the best-fit line against cluster elongation in Figure
\ref{fig:okamura_elong}. The elongation parameter is taken from Goto
et al. (2002a), which measured the ratio of major and minor axes
in their enhanced map to find clusters. Circles and solid lines show
median values. No obvious trend is seen here. Our result
seems to agree with Smail et al.'s (1997) caution that a correlation between $f_b$
and cluster ellipticity found by Wang et al. (1997) could be due to a small
and diverse sample.
However, since distribution of
galaxy positions might not represent cluster ellipticities well, we do not
conclude that there is no dependence on cluster ellipticities. The
dependence should be pursued further in the future, ideally using X-ray
profile shape with a large sample of clusters.
\section{Discussion}\label{discussion}
\subsection{Morphological k-correction}
In the upper right panel of Figure \ref{fig:bo}, we use $C_{in}$
(inverse of concentration index) in the $r$ band to classify galaxies throughout our redshift
range (0.02$\leq$z$\leq$0.3). This could potentially cause redshift dependent
biases in our calculation of $C_{in}$. Since the universe is expanding,
by analyzing the observed $r$ band data, we are analyzing bluer
restframe wavelengths in the higher redshift galaxies.
In fact, the $r$ band at $z$=0.3 is almost $g$ band in the
restframe. Various authors have pointed out that galaxy
morphology significantly changes according to the wavelength used
($e.g.$ Abraham et al. 2001). To estimate how large this bias is, we
plot the normalized distributions of $C_{in}$ in $g$ and $r$ bands in
Figure \ref{fig:cin_g_r} using
the galaxies with 0.02$\leq$z$\leq$0.03 in the SDSS spectroscopic data (1336
galaxies in total; See Eisenstein et al. 2001, Strauss et al. 2002 and
Blanton et al. 2002 for the SDSS spectroscopic data). In this redshift range, the color shift due to the
expansion of the universe is small. We use this to study the
dependence of the $C_{in}$ parameter on the restframe wavelength. At
$z$=0.3, $r$ band corresponds to restframe $g$ band. The solid
and dashed lines show the distribution for $g$ and $r$ bands,
respectively. The two distributions are not exactly the same, but the
difference between the two distributions is small. We summarize the
statistics in Table \ref{tab:cin_g_r}.
There are 802/1336 galaxies with $C_{in}>$0.4 in $g$ band, and 787/1336
galaxies have $C_{in}>0.4$ in $r$ band. The difference is 15/1336 galaxies, which is 1.1\% of the
sample. In section \ref{Cin}, the change in $f_{Cin}$ is $\sim$30\%.
The effect of the morphological k-correction is therefore much smaller.
We also point out that this analysis assumes the largest difference in
redshift (0.02$\leq$z$\leq$0.3),
therefore it gives the upper limit of the bias.
Since the majority of our clusters are at
$z\sim$0.2, the wavelength difference between the observed and
restframe bands is typically much smaller.
We conclude that the effect of the morphological k-correction
is much smaller than the change in $f_{Cin}$ we observed in Section \ref{Cin}.
In Section \ref{exp}, we use the $r$ band fit for all galaxies in our
sample. The same redshift effect could potentially bring bias to our
analysis. In Table \ref{tab:exp_g_r}, we limit our galaxies to
0.02$\leq$z$\leq$0.03 and count the fraction of late type galaxies in
the $g$ and $r$ bands corresponding to observed frame $r$ band at z=0.0 and z=0.3.
We list the number of galaxies with
exponential likelihood higher than de Vaucouleur likelihood in column 1, the
total number of galaxies in column 2, and the ratio of columns 1 to 2 in column
3. As shown in the 3rd row, the difference in the fraction of late type
galaxies between $g$ band data and $r$ band data is only 2.5\%, which
is much smaller than the $f_{exp}$ change we see in the upper right panel of
Figure \ref{fig:bo} ($\sim$30\%).
% We also point out that the fact we
% do not see any redshift dependence in cluster shuffling test in Figure
% \ref{fig:random} is re-assuring the effect in Figure \ref{fig:bo} is
% not caused by redshift bias.
We conclude that the change of
$f_{exp}$ and $f_{Cin}$ is not caused by the small redshift bias in using $r$ band
data throughout the redshift range.
%
% Another possible sources of bias in measuring $f_{exp}$ and $f_{Cin}$
% is the decrease of the size of galaxy with increasing redshift. As we
% study higher redshifts, the sizes of galaxies become smaller and
% smaller, and thus, the profile fit or the radius measurement could be
% affected. In such a case, however, higher redshift galaxies
% are likely to be classified more as elliptical galaxies since their sizes are
% smaller and more compact.
% The resulting effect should decrease the $f_{exp}$ and $f_{Cin}$.
% We, however, find an increase of both $f_{exp}$ and $f_{Cin}$ toward
% higher redshift.
% Even if these biases exist, they can not mimic the morphological
% Butcher-Oemler effect witnessed here.
%\subsection{Comparison with Previous Works}
\subsection{Seeing Dependence}
Another possible source of bias in measuring $f_{exp}$ and
$f_{Cin}$ is the dependence on the seeing, relative to the size of the galaxies. At
higher redshift, the size of a galaxy is smaller and a seeing convolution
could be more problematic. Especially for the concentration
parameter ($C_{in}$), galaxy light becomes less concentrated when
the seeing size is comparable to the galaxy size, and thus, the effect could
cause a bias towards higher $C_{in}$ values.
To check this, we plot $f_{Cin}$ against the
point-spread function (PSF) size in the $r$ band for two redshift limited samples in Figure
\ref{fig:seeing_cin}. Open squares and solid lines show
the distribution and medians of low z clusters (z$\leq$0.15). Filled
triangles and dashed lines show the distributions and medians of high z
clusters (0.24$<$z$\leq$0.3).
For the median measurements, bins are chosen so that equal
numbers of galaxies are included in each bin. 1 $\sigma$ errors are shown as vertical bars.
As expected, lower
redshift clusters show almost negligible dependence on seeing
size. Higher redshift clusters show about a 5\% increase in $f_{Cin}$ between the best
and worst seeing size. The evolution effect we see in the upper right
panel of Figure \ref{fig:bo} is more than 20\%. Furthermore,
as the distribution of seeing in Figure \ref{fig:seeing_hist} shows,
87\% of our sample galaxies have seeing better than 2.0''.
Therefore we conclude that varying seeing causes
a small bias which is significantly weaker than
%it is not so significant as to mimic
the evolution we find in Section \ref{method}.
The effect of varying seeing is less significant for the $f_{exp}$
parameter. In Figure \ref{fig:seeing_exp}, we plot $f_{exp}$ against
seeing size for two redshift samples with redshift ranges and symbols
as in Figure \ref{fig:seeing_cin}. 1 $\sigma$ errors, shown as
vertical bars, are dominant. There is no significant correlation of
$f_{exp}$ with seeing size.
\subsection{Radius, Fore/background Subtraction and Cluster Centroids}
Throughout the analyses in Section \ref{method}, we use
% 0.7 Mpc as a cluster radius and
a 2.1-2.21 Mpc annular region for fore/background
subtraction. In return for
taking cosmic variance into account, annular (local) background subtraction
has larger errors than global background subtraction due to its smaller
angular area coverage. However the difference is not so large. In
case of blue galaxy counts ($f_b$) in the background, the
median Poisson (1 $\sigma$) uncertainty for global background is 12.2\%, whereas 1
$\sigma$ variation of local background is 12.6\%. This increases
the errors, but only by 0.4 points. The actual effect to the late-type
fraction is plotted in Figure \ref{fig:bg_test}. Solid lines show
distributions for our default choice of 0.7 Mpc radii
and 2.1-2.21 Mpc annuli. Dashed lines show distributions for global
fore/background subtraction, where fore/background subtraction is
performed using global number counts of galaxies for all the clusters
in the sample. A Kolomogorov-Smirnov test between two samples does not show any
significant difference.
% Although this enables us to subtract the fore/background
% locally without suffering from the large scale variance, we need to
% clarify how sensitive our results are to the choice of the
% fore/background subtraction.
For cluster radius, we use 0.7 Mpc, since we do not have
information about the virial radii of each system.
It is, however, ideal to use virial radii since, for example, in a standard cold dark
matter cosmology, virial radii
at a fixed mass scales as $\propto$ (1+z)$^{-1}$.
Another possible cause of uncertainty is the accuracy in deciding cluster
centers. % We use cluster centers given in Goto et al. (2002).
% Although the SDSS Cut \& Enhance cluster catalog
In this work, a center position of each cluster is taken from
Goto et al. (2002a), and is estimated from the position of the peak in their
enhanced density map. Although, from Monte-Carlo simulations, cluster centroids are expected
to be determined with an accuracy better than $\sim$40 arcsec,
the offsets have a possibility to introduce a bias in our analyses.
We test
different choices of these parameters in Figure \ref{fig:bg_test}.
Dotted lines show distributions where radii change as
0.7 $\times$ (1+z)$^{-1}$ Mpc assuming a standard cold dark matter
cosmology. Long dashed lines show distributions when the position of
brightest cluster galaxy (within 0.7 Mpc and $Mr<$-24.0) is used as a cluster center.
%(A brightest cluster galaxy is chosen as the brightest galaxies in 0.7 Mpc with $Mr<$-24.0).
%In none of the cases,
Kolomogorov-Smirnov tests show no significant difference in any of the above cases.
In all cases, the probability that the distributions are different is
less than 26\%. Our results in Section \ref{method} are thus not
particularly sensitive to our choice of annuli, radii or cluster centers.
We further pursue the effect of radius dependence of blue/late type fractions in
Appendix, and show that it does not change our main results.
\subsection{Comparison with Late-type Fraction from Spectroscopy}
To further test our late-type fraction measurement, we compare
the late-type fraction obtained from the SDSS spectroscopic data with
that obtained from the SDSS imaging data in figure
\ref{fig:comparison_with_spectroscopy}. Since the SDSS spectroscopic
data are limited to $r^*<$17.77 (Strauss et al. 2002), the comparison
can be done only for clusters with z$<$0.06. In the literature, three
clusters are found to satisfy these criteria in the region used in this
study. The clusters include ABELL 295, RXC
J0114.9+0024, and ABELL 957. For these clusters, late-type fractions
are measured in the same way as in section \ref{method}. Late-type
fractions from spectroscopy are measured using galaxies within 0.7 Mpc
and $\delta$z=$\pm$0.005 from the redshift of each cluster. Note that there is no
fore/background correction for spectroscopic late-type fraction. In figure
\ref{fig:comparison_with_spectroscopy}, all points agree with each
other within the error. The good agreement suggests that our fore/background
subtraction technique described in section \ref{bg} works
properly. It would be ideal to perform the same test for high redshift
clusters as well. However, the SDSS spectroscopic data are not deep
enough to perform the test for higher redshift clusters.
\subsection{The Butcher-Oemler Effect}
The Butcher-Oemler effect--- an increase in the ratio of blue galaxies in
clusters as a function of redshift--- is strong evidence of direct evolution of
the stellar populations in galaxies; it has been studied by numerous
authors in the past.
% originating from the work of Butcher \& Oemler (1978, 1984).
In this section,
we compare our results with previous work.
% and discuss the consistency.
% Since each author of the previous papers, however, used a
Since different authors use
different cluster samples, color bands, cosmology, absolute
magnitude ranges and methods of fore/background subtraction,
which could affect the comparison,
%Since these
% different data and ways of analysis potentially could change the results significantly,
we emphasize the differences in analysis by each author. Note that one
important difference is that some previous work used
a sample of quite rich clusters. e.g. clusters with more than 100 members
in magnitude and radius ranges comparable to those adopted in this
study. Poorest systems in our
sample have only 25 member galaxies after fore/background subtraction.
Thus, difference in cluster samples could cause a difference in results.
Butcher \& Oemler (1978, 1984) studied 33 clusters between z=0.003 and
z=0.54. They used galaxies brighter than $M_V$=-20 ($h$=0.5 and $q_0$=0.1) within the circular area
containing the inner 30\% of the total cluster population. They found
$f_b$ increases with redshift for z $\geq$ 0.1. Their $f_b$ at z=0.3 is
$\sim$0.15, which is slightly lower than our value.
Considering the large scatter in both their and our samples, we do not claim that our results
are inconsistent with their value. Note that Andreon \& Ettori (1999)
found a trend of increasing X-ray luminosity with increasing redshift
in the sample clusters of Butcher \& Oemler (1984).
Rakos \& Schombert (1995) studied 17 clusters using Stromgren $uvby$
filters. Due to the usage of the narrow band filters redshifted to the
cluster distance, their study did need to use model-dependent k-corrections.
However, their high-redshift cluster sample is drawn from that
of Gunn, Hoessel \& Oke (1986) which is based on IIIa-J and IIIa-F
photographic plates. At $z>0.5$, these plates measure the rest-frame
ultraviolet to blue region of the spectrum. Thus the cluster catalog will
be biased toward clusters rich in blue galaxies at high redshift.
Rakos \& Schombert found $f_b\sim$0.25 at z=0.3, which is
slightly higher than the estimation of Butcher \& Oemler (1984) but
in agreement with our results.
Margoniner et al. (2000) studied 44 Abell clusters between z=0.03 and z=0.38. They used galaxies
between $M_r=-21.91$ and $M_r=-17.91$ ($h=0.75$) within 0.7 Mpc of the cluster center.
The fore/background counts are subtracted using five control fields.
Their results are more consistent with the steeper relation estimated in
1995 by Rakos and Schombert than with the original one by Butcher and
Oemler in 1984. The results are also consistent with ours.
Margoniner et al. (2001) extended their work to 295 Abell clusters
% between z=0.03 and z=0.38. They used galaxies from $Mr$= -21.16 to $Mr$=-18.16 ($h$=0.67 and
% $q_0$=0.1) within 0.7 Mpc around each cluster. Their fore/background
% counts are subtracted using those of the control field.
and found $f_b$=(1.34$\pm$0.11)$\times$z-0.03 with a {\it r.m.s.}
of 0.07, which is in agreement with our fitted function shown in Figure
\ref{fig:bo}.
%They also found that $f_b$ is systematically higher for
% poor clusters and claimed this dependence is the chief cause of scatter in
% Butcher-Oemler effect in previous works. In our work, we also have a
% large scatter in Figure \ref{fig:bo}.
%In our case, however, selecting
% subsample of our clusters using our richness, $N_{gal}$, does not remove
% the scatter in Figure \ref{fig:bo}. The possible reason is that our
% sample clusters are selected much smaller area than Margoniner et
% al. (2001) , thus, most of them are not as rich as Abell
% clusters. Since the data used in this work is only 5\% of the total
% SDSS data, the study using the richer and larger sample of clusters
% will verify the richness dependence in the near future.
Ellingson et al. (2001) studied 15 CNOC1 clusters (Yee, Ellingson, \&
Carlberg 1996) between z=0.18 and z=0.55. Since they used spectroscopically observed
galaxies, they do not suffer from the fore/background
correction (but see Diaferio et al. 2001). They used galaxies brighter than $M_{r}=-19.0$ within
$r_{200}$ from the cluster center (with an average of 1.17$h^{-1}$
Mpc). Their best fit shows $f_b\sim$0.15 at z=0.3. The scatter in their Fig. 1 and our data are both substantial.
Thus, we can not
conclude this is inconsistent with our results.
All of these authors found considerable scatter
in $f_b$ v.s. z plot as is seen in our Figure \ref{fig:bo}.
It is promising that our results are consistent with the previous
work within the scatter, despite the differences in the radial coverage
and magnitude ranges used.
% In each author's $f_b$ v.s. z plot, however,
%the scatter is large at z$<$0.3.
% Although our results shows weak
% increase of $f_b$ between z=0.02 and z=0.3, we stress that our results
% are consistent with previous authors including the presence of large
% scatter.
\subsection{Morphological Butcher-Oemler effect}
In sections \ref{u-r}, \ref{exp}, and \ref{Cin}, we found an
increase in the fraction of late type galaxies selected by morphological parameters
with increasing redshift --- as if the Butcher-Oemler effect is happening morphologically.
Perhaps revealing this morphological Butcher-Oemler effect is the most striking result of this study.
% Butcher-Oemler effect $i.e$ we found the increase of morphologically
% late type galaxies as a function of redshift as well as the increase of
% blue galaxies.
It suggests that the Butcher-Oemler blue galaxies
change their morphology from late to early type at the same time that they
change their color from blue to red. Although accurately quantifying the fraction of
galaxies which experience the morphological Butcher-Oemler effect is
difficult due to the considerable scatter in the data, our best-fit lines suggest
that $\sim$30\% of galaxies in clusters go through this transition
between z=0.3 and z=0.02.
In previous work, Dressler et al. (1997)
found a deficit of S0 galaxies in 10 intermediate (z$\sim$0.5)
clusters by classifying galaxy morphology in the HST image
by eye. They claimed that many S0s needed to be added to
reach the fraction of S0s found in present clusters (Dressler 1980).
Couch et al. (1994, 1998) also found an indication of morphological
transformation in the Butcher-Oemler galaxies by studying three rich
clusters at z=0.31. Later, Fasano et al. (2000) showed that spiral
galaxies are, in fact, turning into S0 galaxies
by observing nine clusters at intermediate redshifts and analyzing them
together with higher redshift clusters in the literature. Their galaxy morphology was
also based on eye classification.
Our SDSS data is taken using ground based telescopes with moderate
seeing ($\sim$1.5''), and thus does not allow us to separate S0 galaxies
from elliptical galaxies as the HST does.
%space telescope data with excellent seeing condition does.
The advantage of our classification is its automated nature,
which allows accurate reproducibility and quantification of
systematic biases.
%free from the unpredictable bias in human classification.
% which could potentially be unpredictable by its nature.
In particular, it is easy to compute the completeness and contamination rate for
the automated classification, based on simulations; for the present sample,
the completeness and contamination rate of the parameters
are given in Shimasaku et al. (2001) and Strateva et
al. (2001). Furthermore, an automated galaxy classification is easier to reproduce in
future observational work and in detailed computer simulations.
%Although we do not have as detailed morphological classification as
%theirs ($e.g.$ we can not distinguish S0s from ellipticals.),
Although we can not distinguish S0s from ellipticals, the increase of blue
fraction and increase of late type galaxies toward higher redshift is
qualitatively consistent with the process of S0 production over the
interval in cosmic time suggested by previous investigations.
Various physical mechanisms could be the cause of the morphological
and spectral Butcher-Oemler effects.
Possible causes include ram pressure stripping of gas (Gunn \& Gott 1972, Farouki
\& Shapiro 1980; Kent 1981, Abadi, Moore \& Bower 1999, Quilis, Moore \& Bower
2000), galaxy infall (Bothun \& Dressler 1986, Abraham et al. 1996a, Ellingson et al. 2001), galaxy harassment via high speed impulsive encounters (Moore et al. 1996, 1999), cluster
tidal forces (Byrd \& Valtonen 1990, Valluri 1993) which distort
galaxies as they come close to the center, interaction/merging of
galaxies (Icke 1985, Lavery \& Henry 1988, Bekki 1998), and removal \& consumption of the gas due to the cluster environment (Larson, Tinsley \& Caldwell 1980, Balogh et
al. 2001, Bekki et al. 2002). Mamon (1992) and Makino \& Hut (1997) showed that
interactions/mergers can occur in a rich cluster environment despite the
high relative velocities. Shioya et al. (2002) showed that the
truncation of star formation can explain the decrease of S0 with
increasing redshift.
It has been known that preheating of intergalactic medium
can effect morphologies of galaxies by strangling the gas accretion (Mo
\& Mao 2002; Oh \& Benson 2002). In fact, Finoguenov et al. (2003)
found the filamentary gas in Coma cluster and predicted quiescent star
formation in galaxy disks around the filament.
%Our results does not give a conclusive evidence that decides which one
%of these processes are responsible to the BO effects.
Although our results provide some important clues, pinpointing what
processes are responsible in the morphological and spectral
Butcher-Oemler effect is a more difficult challenge.
Our results suggest that the cause will be one that affects both
color and morphological appearance of galaxies at the same time.
%In a high redshift
% cluster MS1054, van Dokkum et al. (1999) found many cluster galaxies with red
% color with spiral morphology $i.e$ 'passive spirals'.
Couch et al. (1998), Dressler et al. (1999) and Poggianti et al. (1999) found ``passive
spirals'', which are galaxies with spiral morphology but without star
formation. They are probably belong to the same population as ``anemic
spirals'' found by van den Bergh (1976).
The mechanism creating ``passive spirals'' or ``anemic
spirals'', however, affects only the
color of galaxies and, thus,
probably is not the main mechanism that accounts for the entire effect.
The increase of morphologically late type galaxies toward higher
redshifts at the same time as the
increase of blue galaxies is consistent with mechanisms which affect the gas supply
(e.g. ram-pressure stripping, galaxy infall). However if the infalling
rate of field galaxies (mostly blue/late type) is higher in the past,
almost any of the mechanisms mentioned above can explain our
observational results.
%
%
% Furthermore, if we assume the infall rate of field galaxies are smaller
% than the effects we found in this work, the increase of morphologically
% late type galaxies toward higher redshifts at the same time as the
% increase of blue galaxies might favor mechanisms which affect the gas supply
% (e.g. ram-pressure stripping, galaxy infall) rather than dynamical
% processes such as galaxy harassment, galaxy-galaxy interaction or galaxy
% mergers since such dynamical processes disturb the galaxy morphology
% (making a galaxy less concentrated)
% and accelerates star formation (making a galaxy bluer) which should be
% seen as an increase of blue galaxies toward lower redshift. On the
% contrary, we do not see any increase of $f_{cin}$ or $f_b$ toward lower redshift.
%
% Also, the gas depletion through accelerated star
% formation due to the cluster environment might be less favored because
% in such a case, we would see the accelerated star formation as an
% increase in blue galaxies toward lower redshift. On the contrary, we
% see a clear
% decrease of blue galaxies toward lower redshift (the Butcher-Oemler
% effect) in the lower left panel of Figure \ref{fig:bo}. Balogh et
% al. (1997) also supports this idea since they did not find any excess of
% star formation in a sample of luminous X-ray selected clusters and
% concluded that the cluster environment only works in the direction of
% suppressing star formation. However we caution readers that there still
% remain many uncertainties and possibilities. If the infalling rate
% changes dramatically,
% it can ruin/mimic the above arguments.
% In the first place,
Furthermore, although we
discussed about cluster specific phenomena, it is also known that
galaxies in the field region evolve as a function of redshift as well.
(e.g. Hammmer et al. 1996; Lilly et al. 1996; Balogh et al. 1997,
2002). The evolution of field galaxies needs to be compared
with that of cluster galaxies further in detail.
Therefore, it is still an open question what mechanism causes spectral
and morphological evolution of cluster galaxies.
% we only suggest that a mechanism which affect the gas
% supply is consistent with our results, and we do not reject a
% possibility for other mechanisms to be responsible for the evolution of
% galaxies, including the one that happens in the field region.
% In the previous section, we see the galaxies in clusters change their
% color and morphology. In this subsection, we discuss about the time
% scale of morphological and color change of galaxies.
The finding of a 30\% change of the fraction during the look back time
of $\sim$3.5 Gyr could also give us an additional hint in finding an
underlying physical process.
If the gas in spiral galaxies is removed very efficiently by some
physical processes ($e.g.$ ram-pressure stripping) or consumed rapidly by
star formation, the spiral arms will disappear after several disk
rotation periods, $\sim$ 1 Gyr (Sellwood \& Carlberg 1984).
Interaction/merger processes are quicker than gas removal processes
($\sim$0.5 Gyr; Mihos 1995).
Moore et al.'s (1996) simulation showed that the galaxy harassment phase
lasts for several Gyr. Kodama et al. (2001) used the phenomenological simulations to show that the timescale of the morphological transformation from spiral to S0 is 1$\sim$3 Gyr.
For spectral change, Shioya et al. (2002) showed that a disk needs 2-3 Gyr after the removal of gas (or truncation of star formation) to show a k spectrum. Poggianti
et al. (1999) compared the spectral and morphological properties of
cluster galaxies and suggested that the timescale of the
morphological transition is longer than that of the spectral transition.
This difference in timescale is interesting since if one process is
significantly quicker than the other, we might be able to see the time
difference in the decreases in the fraction of late type galaxies and
blue galaxies, which will provide a strong constraint in the evolution history of the Butcher-Oemler galaxies.
In Figure \ref{fig:bo}, we see a $\sim$30\%
of change in both the photometric and morphological Butcher-Oemler effect
between z=0.02 and z=0.30 ($\sim$3.5 Gyr). The
scatter in our measurement, however, is considerable and our choices of
criteria between late and early type galaxies do not necessarily coincide
with each other. It is thus not straightforward to convert the
information to the time scale of the responsible physical process.
In addition, to understand change in fraction of morphological and
spectral late-type galaxies, the change in infalling rate of field
galaxies needs to be understood as well.
%We also note that numerical simulations with $\sim$10$^4$
% particles suffers from numerical heating and potentially
% underestimates the time scale of morphological transformation.
%We, thus, leave the constrains on galaxy evolution to the future paper.
%do not conclude our results are inconsistent
%with the difference in timescale of the spectral and morphological
%(dynamical) change.
Since computer simulations have recently made dramatic progress,
in the near future it will become possible for
state-of-the-art simulations to simulate both dynamical and
spectral evolutions of cluster galaxies, plus infalling rate of field
galaxies in order to compare the results with the
observed trend. For example, such a simulation can be done by combining
dynamical simulations (e.g.
Evrard 1991, Kauffmann et al. 1995, Bekki, Shioya \& Couch 2001,
Vollmer et al. 2001, Bekki et al. 2002) with cluster
phenomenological simulations (e.g. Abraham et al. 1996, Fujita
1998,2001, Balogh et al. 1999, Stevens, Acreman, \& Ponman 1999,
Balogh, Navarro, \& Morris 2000, Kodama \& Bower 2001).
Figure \ref{fig:bo} in this work provides the interesting observational data
to tackle with using such a simulation of cluster galaxy formation.
%
%\subsection{On the Time Scale of Morphological Change}
%
% Although Butcher and Oemler (1984) claimed that the fraction of spiral
% galaxies are much smaller than the fraction of blue galaxies with
% typical ration of 1:4, we do not conclude anything about the relative
% fraction of $f_b$, $f_{u-r}$, $f{Cin}$ and $f_{exp}$ since the relative
% fraction of spiral galaxies easily changes if we change the criteria to
% separate galaxies $e.g$ If we lower concentration criteria from
% $C_{in}$=0.4 to $C_{in}$=0.3, it increases the fraction of late type galaxies.
%
%
%Is this mergering time scale ?
%Is this young star life time scale ?
%
%
%Passive spiral.
%evidence of S- S0 transition rather than S-E transition.
%%key feature in 1/3 of Hubble time.
%
%
%\section{Field Butcher-Oemler effect}
% In the previous section, we see the galaxies in clusters change their
% color and morphology. The next interesting question is about the
% origin of this color and morphological change. In the literature,
% several people report that not only cluster galaxies but also field
% galaxies might change their color or morphology (Balogh et al., Hogg et
% al. ADD REFERENCE). In the previous studies, however, the number of
% galaxies studied was limited because there have not been a big volume
% limited sample of field galaxies. To clarify the Butcher-Oemler effect
% is cluster origin or it happens in the field region as well. We
% studied field galaxies in the SDSS spectroscopic survey.
%
\subsection{Richness Dependence}
In Section \ref{scatter}, we observe the tendency of richer clusters
to have smaller fractions of late type galaxies, by measuring the
residuals from the best fit relations as a function of cluster
richness. Our result is consistent with Margoniner et al. (2001), who
used a similar optical richness to find that poorer clusters tend to
have larger blue fractions than
richer clusters at the same redshift. Figure \ref{fig:okamura_rich},
however, still shows a significant amount of scatter, which might be
suggesting the existence of another factor in determining the blue
fraction in addition to redshift and richness.
The dependence of the late type fraction on cluster richness, however, provides
another hint on the underlying physical processes. Since ram pressure
is stronger in clusters with higher temperature at the same gas density, Fujita \&
Nagashima(1999) theoretically predicted that if ram pressure is the only
mechanism responsible for the evolution of galaxies in clusters, the
fraction of blue galaxies will always be higher in lower X-ray luminosity
clusters, which usually have low temperatures. Our data shown in Figure
\ref{fig:okamura_rich} are consistent with the prediction from their ram
pressure stripping model. Although
our richness is from numbers of galaxies in optical imaging data, it is reasonable to assume it correlates well with X-ray luminosity (Bahcall 1977;
Bower et al. 1994). Then, the optical
richness can be related to the gas temperature through the well known $L_{X}-T$
relation (Mitchell, Ives, and Culhane 1977; Henry \& Arnaud 1991; Edge
\& Stewart 1991; David et al. 1993; White, Jones, and Forman 1997; Allen
\& Fabian 1998; Markevitch 1998; Arnaud \& Evrard 1999; Jones \& Forman
1999; Reichart, Castander, \& Nichol 1999; Wu, Xue, and Fang 1999; Xue
\& Wu 2000; and see the references therein).
In a simple estimation, ram pressure is proportional to $\rho
v^2$. $L_X$ is proportional to $\rho^2$. From the virial theorem, $v^2\propto T$.
The $L_{X}-T$ relation studied by Xue \& Wu (2000) is $L_X\propto T^{2.8}$.
Therefore, ram pressure is proportional to $\sim L_X^{0.86}$.
Combined with an assumption that optical richness scales with X-ray
luminosity (see, e.g. Bahcall 1974, Jones \& Forman 1978, Bower et
al. 1994, and Miller et al. in preparation),
Figure \ref{fig:okamura_rich} provides another hint that ram
pressure stripping induces the evolution of cluster galaxies.
In the literature, however, the dependence of blue fractions on cluster richness has been controversial. Bahcall (1977) studied 14 X-ray clusters and found that
the fraction of spiral galaxies decreases with increasing X-ray
luminosity.
Lea \& Henry(1988) observed 14 clusters in X-ray and found that
the percentage of blue objects in the clusters seems to increase with the X-ray luminosity.
On the other hand, Fairley et al. (2002) studied eight X-ray selected
clusters and did not find any dependence of blue fractions on X-ray
luminosities. Balogh et al. (2002) studied 10 clusters at z=0.25 with low X-ray
luminosity and found similar morphological and spectral properties of
galaxies compared with clusters with high X-ray luminosity (Balogh et
al. 1997).
In all cases, the results were based on a small sample
of clusters. We also point out that although our results are consistent with
a ram-pressure stripping model, there is a possibility that the other
mechanisms could explain the phenomena. For example, richer clusters
might have higher rate of merger/interaction due to their higher galaxy
density. The same argument holds true for galaxy harassment. Thus, more
study is needed to conclude about the physical mechanism causing the phenomena.
In the near future, confirming the richness
dependence using X-ray luminosities or velocity dispersions with a larger
sample of clusters would offer us more insight on the subject.
\section{Conclusions}\label{conclusion}
In this paper, we have investigated the fraction of late type galaxies in
four different ways using one of the largest, most uniform samples of
514 clusters between 0.02$\leq$z$\leq$0.3 from the SDSS Cut \& Enhance galaxy cluster catalog.
% We include galaxies brighter than $M_r$=-19.44.
All the clusters selected here have more than 25 member galaxies within
0.7 Mpc and between $M_r=-24.0$ and $M_r=-19.44$ after statistical
local fore/background
subtraction. The following four different ways to estimate the fractions of late
type galaxies are adopted: restframe $g-r$ color (a classical Butcher-Oemler estimator), $u-r$ color, concentration index and de Vaucouleur/exponential profile fit.
%The former two
% methods are based on colors of galaxies and the latter two are based
% on the morphology of the galaxies.
The last three parameters are known to correlate well with galaxy
morphologies (Shimasaku et al. 2001, Strateva et al. 2001).
In all four cases, we observe an
increase of the fraction of late type galaxies toward higher redshift
with a significance of more than 99.99\% (Table \ref{tab:correlation}).
We draw the following conclusion from this work.
1. We confirm the presence of the Butcher-Oemler effect using $g-r$ color. The Butcher-Oemler effect is real and exists in our sample clusters as seen in the lower left panel of Figure \ref{fig:bo}. The slope of the increase is consistent with previous work although the scatter in the blue fraction is considerable.
Previous work also noted a large scatter in the fraction of blue galaxies.
The fraction of late type galaxies also shows a similar increase when
we use a $u-r$ color cut.
%The increase
% of late type galaxies selected in $u-r$ color further support the
% reality of the Butcher-Oemler effect as we see in the upper left panel
% of Figure \ref{fig:bo}.
2. We observe a morphological Butcher-Oemler effect as an
increase of late type galaxies toward higher redshift, using pure morphological
parameters such as a concentration parameter and de
Vaucouleur/exponential profile fit. The rates of increase
are consistent with previous work on the spiral to S0
transition, albeit with considerable scatter (Figure \ref{fig:bo}).
%(two right side panels in Figure \ref{fig:bo}).
The increase is also in agreement with the original Butcher-Oemler effect
from $g-r$ color. Our results are consistent with the evolutionary
scenario proposed by Dressler et al. (1997), Smail et al. (1997), Couch
et al. (1998), and Kodama \& Smail (2002),
%involving the disk galaxy% populations
in which there is a progressive morphological conversion in
clusters from spirals into E/S0's.
%Underlying physical processes are yet unknown.
3. We find a slight tendency for richer clusters to have lower values of
the late type fraction (Figure \ref{fig:okamura_rich}). This trend
agrees with the ram pressure stripping model proposed by Bahcall (1977)
and Fujita et al. (1999), in
which galaxies in richer clusters are more affected by ram pressure due to
their high temperature.
%The trend is consistent with ram pressure
%stripping being responsible for the evolution of cluster
%galaxies.
%
% 4. Combining the results 1 and 2, we propose that the underlying
% physical mechanism responsible for the change of colors and morphologies of
% the Butcher-Oemler galaxies would be one that affects both star formation and
% morphological appearance of galaxies at the same time. It is important to note that
% we do not see any sudden increase of blue galaxies toward lower
% redshift during the look-back time
% observed ($\sim$ 3.5 Gyr), which should be seen as an increase of star
% formation activity if merger/interactions or the rapid gas consumption
% through the accelerated star formation are the main causes of the effect, although the data are too noisy to be conclusive.
% We also do not see an increase of less concentrated galaxies resulting from galaxy-galaxy merger/interactions. Interpreting with conclusion 3, we propose that a mechanism which
% works on gas supply, such as ram-pressure stripping or galaxy infall is preferable to
% dynamical mechanisms such as merger/interaction and galaxy harassment or
% rapid gas depletion through accelerated star formation.
% a dynamical mechanism which enhances star formation (e.g
% merge/interaction, galaxy harassment) or rapid gas depletion through
% accelerated star formation.
Although our results 1,2, and 3 are all consistent with a ram-pressure
stripping model, there still remains a possibility that other physical
mechanisms are responsible for the evolution of cluster galaxies.
Thus, further study is needed both theoretically and observationally to
reveal the underlying physical mechanism responsible for the evolution
of cluster galaxies.
Since this work is based on only 5\% of the whole SDSS data, an
increase in the data will improve the statistical accuracy as the SDSS
proceeds. Extending the work to higher redshifts using 4-8 m class
telescopes will offer more insight on the origin and evolution of
cluster galaxies.
%
% As the possible causes of morphological Butcher-Oemler effect, our
% results favor the mechanisms that affect the gas supply
% (e.g. ram-pressure stripping, galaxy infall) rather than dynamical
% processes such as galaxy harassment, galaxy-galxy interaction or galaxy
% mergers. The conpletion of the gas
% through accelerated star formation, however, is less favored since we
% do not see the increase of blue galaxeis toward the lower redshift as
% a sign of accelerated star formation activity.
%
\bigskip
%\acknowledgments
We are grateful to Nell Hana Hoffman, Ricardo Colon, Michael L. Balogh,
A. Kathy Romer and
Robert C. Nichol for valuable comments, which contributed to improve
the paper.
We thank anonymous referee for valuable comments, which improved the
paper significantly.
T.G. acknowledges financial support from the Japan Society for the
Promotion of Science (JSPS) through JSPS Research Fellowships for Young
Scientists.
Funding for the creation and distribution of the SDSS Archive has been provided by the Alfred P. Sloan Foundation, the Participating Institutions, the National Aeronautics and Space Administration, the National Science Foundation, the U.S. Department of Energy, the Japanese Monbukagakusho, and the Max Planck Society. The SDSS Web site is http://www.sdss.org/.
The SDSS is managed by the Astrophysical Research Consortium (ARC) for the Participating Institutions. The Participating Institutions are The University of Chicago, Fermilab, the Institute for Advanced Study, the Japan Participation Group, The Johns Hopkins University, Los Alamos National Laboratory, the Max-Planck-Institute for Astronomy (MPIA), the Max-Planck-Institute for Astrophysics (MPA), New Mexico State University, Princeton University, the United States Naval Observatory, and the University of Washington.
\appendix
\section{Varying Radius}\label{varying}
In Section \ref{method}, we used a fixed 0.7 Mpc radius to measure
blue/spiral fractions among cluster galaxies since it was difficult to
measure virial radius for relatively poor clusters in our sample from the
SDSS imaging data.
However it is known that
virial radius changes according to cluster richness; i.e. richer
clusters have larger virial radius than poorer clusters. Therefore
using a fixed radius could bring some bias associated with cluster
richness. In this section we try to rectify this problem using cluster
richness to calculate virial radius under a simple assumption.
% We assume that cluster richness is proportional to volume of a cluster,
% and therefore, proportional to virial $radius^3$.
% We
% assume the following; (i) all clusters in our sample
% have the same radial profile of galaxy number density. (ii) galaxy
% number density is the same for all clusters in our sample.
% Under the assumption,
We assume that our cluster richness (number of galaxies between $M_{r^*}=-24.0$ and $M_r=-19.44$
within 0.7 Mpc
after fore/background subtraction) is proportional to volume of a cluster, and
therefore proportional to $radius^3$. Since richness is a relatively
easy parameter to measure from the imaging data, we use the following
equation to calculate radius for each cluster.
\begin{equation}
radius = 0.7\times(Richness/32)^{1/3}
\end{equation}
where median richness of our sample cluster is 32. The coefficient of
the equation is adjusted so that median clusters in our sample have
radius of 0.7 Mpc, which corresponds to the mean radius used in Butcher et
al. (1978, 1984) and Margoniner et al. (2000, 2001). The distribution of
radius calculated in this way is presented in figure \ref{fig:new_rad}. As expected
it has a peak at 0.7 Mpc.
Using this varying radius, we re-calculated all figures in Section
\ref{method}. Sample clusters are still required to have more than 25
galaxies after fore/background subtraction within the new
radius. Therefore the number of sample clusters are somewhat reduced to
413 clusters.
Results are presented in figures
\ref{fig:bo_new}-\ref{fig:okamura_elong_new}.
Reassuringly, all
figures have the same trend as presented in Section
\ref{method}. Therefore the discussion in Section \ref{discussion}
still holds. Although it is ideal
to use virial radius to measure blue/spiral fractions of clusters, we
regard that our analysis using fixed 0.7 Mpc radius is not hampered to
the extend where our main conclusions change.
\section{Errors on Blue/Late Type Fractions}\label{error}
In this section we summarize how we derived eq. (\ref{frac_err}) to
calculate errors on blue/late type fractions.
As a starting point, we assume the following.
\begin{itemize}
\item Number of galaxies in a cluster region ($N_{c+f}^{all}$) follows
Poisson statistics.
\item Number of galaxies in a certain area of field region ($N_{f}^{all}$) follows
Poisson statistics.
\item $N_{c+f}^{*}$ and $N_{f}^{*}$ are independent of each other.
\item Number of blue/late type galaxies in a cluster region ($N_{c+f}^{late}$) is
strongly correlated with number of all galaxies in that region
($N_{c+f}^{all}$).
\item Number of blue/late type galaxies in a certain field region ($N_{f}^{late}$) is
strongly correlated with number of all galaxies in that region
($N_{f}^{all}$).
\end{itemize}
And we clarify the definition of our notation.
In this appendix, $\delta A $ means a deviation of a sampled value
from an expectation value, $E(A)$. \\
$\delta A$=$A-E(A)$, where A is each data value.\\
$E(A)$ and $\delta A$ satisfy the following relations.\\
$E(\delta A)$=0, $E(\delta A)^2$=$\sigma^2$,
and if A and B are independent, $E(\delta A \delta B)$=0.
Note that the equation \ref{frac_err} is not a deviation of
a single sample but the expectation value estimated from the sample,
and should be written as $E(\delta f_{late}^2)$ if we write rigidly.
Under these assumptions, the error of late type fraction,
$\delta f_{late}$, become
\begin{equation} \label{xy}
\delta f_{late}^2 = (X/Y)^2 \times ((\delta X^2/X^2) + (\delta Y^2/Y^2) -2(\delta X \delta Y/XY)),
\end{equation}
where $X$=$N_{c+f}^{late}-N_{f}^{late}$,
$Y$=$N_{c+f}^{all}-N_{f}^{all}$, and $f_{late}$ = $X/Y$.
Since $N_{c+f}^{all}$ and $N_{f}^{all}$ follow Poisson statistics,
and they are independent.
\begin{equation}
E(\delta Y^2) = N_{c+f}^{all}+N_{f}^{all}\end{equation}
Similarly, when $N_{c+f}^{all}$ follows Poisson statistics,
$N_{c+f}^{late}$ also follows Poisson statistics
since $N_{c+f}^{late} \sim N_{c+f}^{all}\times f_{late}$.
Therefore,
\begin{equation}
E(\delta X^2) = N_{c+f}^{late}+N_{f}^{late}.\end{equation}
%since both
% $N_{c+f}^{late}$ and $N_{f}^{late}$ follow Poisson statistics and
% they are independent.
Deriving the cross term at the end of the equation is not
so straightforward.
%We explain it rather verbosely below.
The cross term is expanded as
\begin{equation}
\delta X \delta Y %% \end{equation}\begin{equation}
= \delta(N_{c+f}^{late}-N_{f}^{late})\delta(N_{c+f}^{all}-N_{f}^{all})\end{equation}\begin{equation}
=(\delta(N_{c+f}^{late})-\delta(N_{f}^{late}))(\delta(N_{c+f}^{all})-\delta(N_{f}^{all}))\\\end{equation}\begin{equation}
=\delta(N_{c+f}^{late})\delta(N_{c+f}^{all})
-\delta(N_{f}^{late})\delta(N_{c+f}^{all})
-\delta(N_{c+f}^{late})\delta(N_{f}^{all})
+\delta(N_{f}^{late})\delta(N_{f}^{all})\end{equation}
Since we assume $N_{f}^{late}$ and $N_{c+f}^{all}$,
$N_{c+f}^{all}$ and $N_{f}^{late}$ are both independent,
\begin{equation}
E(-\delta(N_{f}^{late})\delta(N_{c+f}^{all}))=0 \end{equation}
and
\begin{equation}
E(-\delta(N_{c+f}^{late})\delta(N_{f}^{all}))=0. \end{equation}
Therefore,
\begin{equation}
E(\delta X \delta Y) =
E(\delta(N_{c+f}^{late})\delta(N_{c+f}^{all})+\delta(N_{f}^{late})\delta(N_{f}^{all}))\\
\end{equation}
Since we assume that $N_{c+f}^{late}$ and $N_{c+f}^{all}$, or
$N_{f}^{late}$ and $N_{f}^{all}$ strongly correlate,
%dependent since we
% assumed that $N_{c+f}^{late} = f_{late} \times N_{c+f}^{all}$ (or
% $N_{f}^{late} = f_{late} \times N_{f}^{all}$).
we can approximate that\\
\begin{equation}
E(\delta(N_{c+f}^{late})\delta(N_{c+f}^{all}))=
\sigma(N_{c+f}^{late})\sigma(N_{c+f}^{all}) =
\sqrt{N_{c+f}^{late}}\sqrt{N_{c+f}^{all}} \end{equation}
and,\\
\begin{equation}\label{field_approximation}
E(\delta(N_{f}^{late})\delta(N_{f}^{all}))=
\sigma(N_{f}^{late})\sigma(N_{f}^{all}) =
\sqrt{N_{f}^{late}}\sqrt{N_{f}^{all}}\\ \end{equation}
Therefore we obtain,
\begin{equation}\label{deltaxy}
E(\delta X \delta Y) = \sqrt{N_{c+f}^{late}}\sqrt{N_{c+f}^{all}} +\sqrt{N_{f}^{late}}\sqrt{N_{f}^{all}}
\end{equation}
By substituting eq. (\ref{deltaxy}) for $\delta X \delta Y$
in eq. (\ref{xy}) , we derive
eq. (\ref{frac_err}).
However this is not the only way to estimate the error. Actually, the
correlation between $N_{f}^{late}$ and $N_{f}^{all}$ is not so
obvious since late type fraction in the field and that in the cluster
region might be different. Although we regard the difference is so small
that we can assume the eq.(\ref{field_approximation}),
if we assume that $N_{f}^{late}$, $N_{f}^{early} (= N_{f}^{all} - N_{f}^{late})$,
$N_{c+f}^{late}$, and $N_{c+f}^{early} (= N_{c+f}^{all} - N_{c+f}^{late})$
are independent, we derive,
\begin{equation}
E(\delta(N_{c+f}^{late})\delta(N_{c+f}^{all}))=
E(\delta(N_{c+f}^{late})(\delta(N_{c+f}^{late})+\delta(N_{c+f}^{early})))=
\sigma(N_{c+f}^{late})^2 = N_{c+f}^{late} \end{equation}
and,\\
\begin{equation}
\delta(N_{f}^{late})\delta(N_{f}^{all})=
\sigma(N_{f}^{late})^2 = N_{f}^{late}. \end{equation}
Then, the expectation value of $\delta X \delta Y$ becomes,
\begin{equation} E(\delta X \delta Y) = N_{c+f}^{late}+N_{f}^{late}\end{equation}
In this case, eq. (\ref{frac_err}) becomes,
\begin{equation}
E(\delta f_{late}) = f_{late}\times \sqrt{\frac{ N_{c+f}^{late}+N_{f}^{late}}{( N_{c+f}^{late}-N_{f}^{late} )^2} + \frac{N_{c+f}^{all}+N_{f}^{all}}{(N_{c+f}^{all}-N_{f}^{all})^2} - \frac{2(N_{c+f}^{late}+N_{f}^{late})}{(N_{c+f}^{late}-N_f^{late})(N_{c+f}^{all}-N_f^{all})} }
\label{frac_err_referee}
\end{equation}
%
%\section{Errors on Blue/Late Type Fractions}\label{error}
%
% In this section we summarize how we derived eq. (\ref{frac_err}) to
% calculate errors on blue/late type fractions.
% As a starting point, we assume the following.
%
% \begin{itemize}
% \item Number of galaxies in a cluster region ($N_{c+f}^{all}$) follows
% Poisson statistics.
% \item Number of galaxies in a certain area of field region ($N_{f}^{all}$) follows
% Poisson statistics. $N_{c+f}^{all}$ and $N_{f}^{all}$ are not dependent each
% other.
% \item Number of blue/late type galaxies in a cluster region ($N_{c+f}^{late}$) is
% strongly correlated with number of all galaxies in that region
% ($N_{c+f}^{all}$).
% \item Number of blue/late type galaxies in a certain field region ($N_{f}^{late}$) is
% strongly correlated with number of all galaxies in that region
% ($N_{c+f}^{all}$).
% \end{itemize}
%
% Under these assumptions, $\delta f_{late}$ become
% \begin{equation}
% \delta f_{late}^2 = (X/Y)^2 \times ((\delta X^2/X^2) + (\delta Y^2/Y^2) -2(\delta X \delta Y/XY))
% \end{equation}
% where $X$=$N_{c+f}^{late}-N_{f}^{late}$,
% $Y$=$N_{c+f}^{all}-N_{f}^{all}$, and $f_{late}$ = $X/Y$.
% Then, \begin{equation}
% \delta Y^2 = N_{c+f}^{all}+N_{f}^{all}\end{equation}
%since both
% $N_{c+f}^{all}$ and $N_{f}^{all}$ follow Poisson statistics and
% they are independent.
% Similarly, when $N_{c+f}^{all}$ follows Poisson statistics,
% $N_{c+f}^{late}$ also follows Poisson statistics
% since $N_{c+f}^{late}$=$N_{c+f}^{all}\times f_{late}$.
% Therefore,
% \begin{equation}
% \delta X^2 = N_{c+f}^{late}+N_{f}^{late}\end{equation}
%since both
% $N_{c+f}^{late}$ and $N_{f}^{late}$ follow Poisson statistics and
% they are independent.
%
% Deriving the cross term at the end of the equation is not
% so straightforward. We explain it rather verbosely below.
% First,
% $\delta X$ means a deviation from expectation value,
% $E(X)$. \\
% $\delta X$=X-E(X), where X is each data value.\\
% $E(X)$ and $\delta X$ satisfy the following relations.\\
% $E(\delta X)$=0\\
% $E(\delta X)^2$=$\sigma^2$\\
% $E(\delta X \delta Y)$=0 if X and Y are
% independent.\\
% The cross term is expanded as
%\begin{equation}
% \delta X \delta Y \end{equation}\begin{equation}
%= \delta(N_{c+f}^{late}-N_{f}^{late})\delta(N_{c+f}^{all}-N_{f}^{all})\end{equation}\begin{equation}
% =(\delta(N_{c+f}^{late})-\delta(N_{f}^{late}))(\delta(N_{c+f}^{all})-\delta(N_{f}^{all}))\\\end{equation}\begin{equation}
% =\delta(N_{c+f}^{late})\delta(N_{c+f}^{all})
% -\delta(N_{f}^{late})\delta(N_{c+f}^{all})
% -\delta(N_{c+f}^{late})\delta(N_{f}^{all})
% +\delta(N_{f}^{late})\delta(N_{f}^{all})\end{equation}
%
%
% Since we assume $N_{f}^{late}$ and $N_{c+f}^{all}$,
% $N_{c+f}^{all}$ and $N_{f}^{late}$ are both independent,
% \begin{equation}
% -\delta(N_{f}^{late})\delta(N_{c+f}^{all})=0 \end{equation}
%
% and
%\begin{equation}
%-\delta(N_{c+f}^{late})\delta(N_{f}^{all})=0. \end{equation}
%
% Therefore,
%\begin{equation}
% \delta X \delta Y =
% \delta(N_{c+f}^{late})\delta(N_{c+f}^{all})+\delta(N_{f}^{late})\delta(N_{f}^{all})\\
%\end{equation}
%
% Here we assume that $N_{c+f}^{late}$ and $N_{c+f}^{all}$, or
% $N_{f}^{late}$ and $N_{f}^{all}$ are strongly dependent since we
% assumed that $N_{c+f}^{late} = f_{late} \times N_{c+f}^{all}$ (or
% $N_{f}^{late} = f_{late} \times N_{f}^{all}$).
% We approximate that\\
%\begin{equation}
% \delta(N_{c+f}^{late})\delta(N_{c+f}^{all})=
% \sigma(N_{c+f}^{late})\sigma(N_{c+f}^{all}) =
%\sqrt{N_{c+f}^{late}}\sqrt{(N_{c+f}^{all})} \end{equation}
%
% and,\\
%\begin{equation}\label{field_approximation}
% \delta(N_{f}^{late})\delta(N_{f}^{all})=
% \sigma(N_{f}^{late})\sigma(N_{f}^{all}) =
%\sqrt{N_{f}^{late}}\sqrt{(N_{f}^{all})}\\ \end{equation}
%
% Therefore we obtain,
% \begin{equation}\label{deltaxy}
% \delta X \delta Y = \sqrt{N_{c+f}^{late}}\sqrt{N_{c+f}^{all}} +\sqrt{N_{f}^{late}}\sqrt{N_{f}^{all}}
% \end{equation}
%
% By substituting eq. (\ref{deltaxy}) for $\delta X \delta Y$ , we derive
% eq. (\ref{frac_err}).
%
%
%
%
% However this is not the only way to estimate the error. Actually, the
% correlation between $N_{f}^{late}$ and $N_{f}^{all}$ is not so
% obvious since late type fraction in the field and that in the cluster
% region might be different. Although we regard the difference is so small
% that we can assume the eq.(\ref{field_approximation}),
% if we assume that $N_{f}^{late}$ and $N_{f}^{all}$ are
% independent, we derive,
%\begin{equation}
% \delta(N_{c+f}^{late})\delta(N_{c+f}^{all})=
% \sigma(N_{c+f}^{late})^2 = N_{c+f}^{late} \end{equation}
% and,\\
%\begin{equation}
% \delta(N_{f}^{late})\delta(N_{f}^{all})=
% \sigma(N_{f}^{late})^2 =
%N_{f}^{late} \end{equation}
% Then, $\delta X \delta Y$ becomes,
%
%\begin{equation}\delta X \delta Y = N_{c+f}^{late}+N_{f}^{late}\end{equation}
% In this case, eq. (\ref{frac_err}) becomes,
%\begin{equation}
% \delta f_{late}= f_{late}\times \sqrt{\frac{ N_{c+f}^{late}+N_{f}^{late}}{( N_{c+f}^{late}-N_{f}^{late} )^2} + \frac{N_{c+f}^{all}+N_{f}^{all}}{(N_{c+f}^{all}-N_{f}^{all})^2} - \frac{2(N_{c+f}^{late}+N_{f}^{late})}{(N_{c+f}^{late}-N_f^{late})(N_{c+f}^{all}-N_f^{all})} }
%\label{frac_err_referee}
% \end{equation}
%
%
%%
%% $\sqrt{N_{c+f}^{late}-N_{f}^{late}}\sqrt{N_{c+f}^{all}-N_{f}^{all}}$\\
%% However there is no correlation between $N_{c+f}^{late}$ and
%% $N_{f}^{all}$, nor $N_{c+f}^{all}$ and
%% $N_{f}^{late}$. Therefore only the term,
%% $(\sqrt{N_f^{late}\times N_f^{all}}+ \sqrt{N_{c+f}^{late} \times
%% N_{c+f}^{all}})$, remains. Combining all of these terms
%% together, we derive eq. (\ref{frac_err}).
\begin{thebibliography}{DUM}\label{refrence}
\bibitem[Abadi, Moore, \& Bower(1999)]{1999MNRAS.308..947A} Abadi, M.~G.,
Moore, B., \& Bower, R.~G.\ 1999, \mnras, 308, 947
\bibitem{Abell58} Abell, G. 1958, APJS, 3, 211
\bibitem{Abell89} Abell, G., Corwin, H. \& Olowin, R. 1989, APJS, 70, 1
\bibitem[Abraham et al.(1996)]{1996ApJS..107....1A} Abraham, R.~G., van den
Bergh, S., Glazebrook, K., Ellis, R.~S., Santiago, B.~X., Surma, P., \& Griffiths, R.~E.\ 1996, \apjs, 107, 1
\bibitem[Abraham, Valdes, Yee, \& van den Bergh(1994)]{1994ApJ...432...75A}
Abraham, R.~G., Valdes, F., Yee, H.~K.~C., \& van den Bergh, S.\ 1994,
\apj, 432, 75
\bibitem[Abraham et al.(1996)]{1996ApJ...471..694A} Abraham, R.~G.~et al.\
1996, \apj, 471, 694
\bibitem[Abraham \& van den Bergh(2001)]{2001Sci...293.1273A} Abraham,
R.~G.~\& van den Bergh, S.\ 2001, Science, 293, 1273
\bibitem[Allen \& Fabian(1998)]{1998MNRAS.297L..57A} Allen, S.~W.~\&
Fabian, A.~C.\ 1998, \mnras, 297, L57
\bibitem[Allington-Smith, Ellis, Zirbel, \& Oemler(1993)]{1993ApJ...404..521A} Allington-Smith, J.~R., Ellis, R.,
Zirbel, E.~L., \& Oemler, A.~J.\ 1993, \apj, 404, 521
\bibitem[Andreon \& Ettori(1999)]{1999ApJ...516..647A} Andreon, S.~\& Ettori, S.\ 1999, \apj, 516, 647
\bibitem{}Annis, J. et al. 2002 $in$ $preparation$
\bibitem[Arnaud \& Evrard(1999)]{1999MNRAS.305..631A} Arnaud, M.~\& Evrard,
A.~E.\ 1999, \mnras, 305, 631
\bibitem[Bahcall(1974)]{1974ApJ...193..529B} Bahcall, N.~A.\ 1974, \apj,
193, 529
\bibitem[Bahcall(1977)]{1977ApJ...218L..93B} Bahcall, N.~A.\ 1977, \apjl,
218, L93
\bibitem{}Bahcall, N. A. et al. 2002 $in$ $preparation$
\bibitem[Balogh et al.(1997)]{1997ApJ...488L..75B} Balogh, M.~L., Morris,
S.~L., Yee, H.~K.~C., Carlberg, R.~G., \& Ellingson, E.\ 1997, \apjl, 488,
L75
\bibitem[Balogh et al.(1999)]{1999ApJ...527...54B} Balogh, M.~L., Morris,
S.~L., Yee, H.~K.~C., Carlberg, R.~G., \& Ellingson, E.\ 1999, \apj, 527, 54
\bibitem[Balogh, Navarro, \& Morris(2000)]{2000ApJ...540..113B} Balogh,
M.~L., Navarro, J.~F., \& Morris, S.~L.\ 2000, \apj, 540, 113
\bibitem[Balogh, Christlein, Zabludoff, \& Zaritsky(2001)]{2001ApJ...557..117B} Balogh, M.~L., Christlein, D.,
Zabludoff, A.~I., \& Zaritsky, D.\ 2001, \apj, 557, 117
% \bibitem{}Balogh, M.~L., Bower, R.G., Smail, I., Ziegler, B.L.,Davies, R.L., Gazelu, A., Fritz, A. 2002, submitted to MNRAS
\bibitem[Balogh et al.(2002)]{2002MNRAS.337..256B} Balogh, M., Bower,
R.~G., Smail, I., Ziegler, B.~L., Davies, R.~L., Gaztelu, A., \& Fritz, A.\
2002, \mnras, 337, 256
\bibitem[Bekki(1998)]{1998ApJ...502L.133B} Bekki, K.\ 1998, \apjl, 502,
L133
\bibitem[Bekki, Shioya, \& Couch(2001)]{2001ApJ...547L..17B} Bekki, K.,
Shioya, Y., \& Couch, W.~J.\ 2001, \apjl, 547, L17
\bibitem[Bekki, Couch, \& Shioya(2002)]{2002ApJ...577..651B} Bekki, K.,
Couch, W.~J., \& Shioya, Y.\ 2002, \apj, 577, 651
\bibitem[Binggeli, Sandage, \& Tammann(1988)]{1988ARA&A..26..509B}
Binggeli, B., Sandage, A., \& Tammann, G.~A.\ 1988, \araa, 26, 509
\bibitem{}Binney, J., \& Tremaine, S. 1987, Galactic Dynamics
(Princeton: Princeton Univ. Press)
\bibitem[Blanton et al.(2001)]{2001AJ....121.2358B} Blanton, M.~R.~et al.\
2001, \aj, 121, 2358
\bibitem[Blanton et al.(2002)]{2002AJ} Blanton, M.R., Lupton, R.H.,
Maley, F.M., Young, N., Zehavi,
I., and Loveday, J. 2002, AJ,
submitted
\bibitem[Bower et al.(1994)]{1994MNRAS.268..345B} Bower, R.~G., Bohringer,
H., Briel, U.~G., Ellis, R.~S., Castander, F.~J., \& Couch, W.~J.\ 1994,
\mnras, 268, 345
\bibitem[Boyce et al.(2001)]{2001MNRAS.328..277B} Boyce, P.~J., Phillipps,
S., Jones, J.~B., Driver, S.~P., Smith, R.~M., \& Couch, W.~J.\ 2001,
\mnras, 328, 277.
\bibitem[Bothun \& Dressler(1986)]{1986ApJ...301...57B} Bothun, G.~D.~\& Dressler, A.\ 1986, \apj, 301, 57
\bibitem[Butcher \& Oemler(1978)]{1978ApJ...226..559B} Butcher, H.~\& Oemler, A.\ 1978, \apj, 226, 559
\bibitem[Butcher \& Oemler(1984)]{1984ApJ...285..426B} Butcher, H.~\& Oemler, A.\ 1984, \apj, 285, 426
\bibitem[Byrd \& Valtonen(1990)]{1990ApJ...350...89B} Byrd, G.~\& Valtonen,
M.\ 1990, \apj, 350, 89
\bibitem[Cavaliere, Menci, \& Tozzi(1997)]{1997ApJ...484L..21C} Cavaliere,
A., Menci, N., \& Tozzi, P.\ 1997, \apjl, 484, L21
\bibitem[Couch, Ellis, Sharples, \& Smail(1994)]{1994ApJ...430..121C}
Couch, W.~J., Ellis, R.~S., Sharples, R.~M., \& Smail, I.\ 1994, \apj, 430,
121
\bibitem[Couch et al.(1998)]{1998ApJ...497..188C} Couch, W.~J., Barger,
A.~J., Smail, I., Ellis, R.~S., \& Sharples, R.~M.\ 1998, \apj, 497, 188
\bibitem[Cowie, Songaila, and Barger(1999)]{1999AJ....118..603C} Cowie,
L.~L., Songaila, A., \& Barger, A.~J.\ 1999, \aj, 118, 603
\bibitem[David et al.(1993)]{1993ApJ...412..479D} David, L.~P., Slyz, A.,
Jones, C., Forman, W., Vrtilek, S.~D., \& Arnaud, K.~A.\ 1993, \apj, 412,
479
\bibitem[Diaferio et al. (2001)]{} Diaferio, A., Kauffmann, G., Balogh, M.L., White, S.D.M., Schade, D. \& Ellingson, E. 2001, \mnras, 323, 999.
\bibitem[Doi, Fukugita, \& Okamura(1993)]{1993MNRAS.264..832D} Doi, M.,
Fukugita, M., \& Okamura, S.\ 1993, \mnras, 264, 832
\bibitem[Dressler(1980)]{1980ApJ...236..351D} Dressler, A.\ 1980, \apj,
236, 351
\bibitem[Dressler \& Gunn(1992)]{1992ApJS...78....1D} Dressler, A.~\& Gunn,
J.~E.\ 1992, \apjs, 78, 1
\bibitem[Dressler et al.(1997)]{1997ApJ...490..577D} Dressler, A.~et al.\
1997, \apj, 490, 577
\bibitem[Dressler et al.(1999)]{1999ApJS..122...51D} Dressler, A., Smail,
I., Poggianti, B.~M., Butcher, H., Couch, W.~J., Ellis, R.~S., \& Oemler,
A.~J.\ 1999, \apjs, 122, 51
\bibitem[Edge \& Stewart(1991)]{1991MNRAS.252..414E} Edge, A.~C.~\&
Stewart, G.~C.\ 1991, \mnras, 252, 414
\bibitem[Eisenstein et al.(2001)]{2001AJ....122.2267E} Eisenstein, D.~J.~et
al.\ 2001, \aj, 122, 2267
\bibitem[Ellingson, Lin, Yee, \& Carlberg(2001)]{2001ApJ...547..609E}
Ellingson, E., Lin, H., Yee, H.~K.~C., \& Carlberg, R.~G.\ 2001, \apj, 547,
609
\bibitem[Evrard(1991)]{1991MNRAS.248P...8E} Evrard, A.~E.\ 1991, \mnras,
248, 8P
\bibitem[Fairley et al.(2002)]{2002MNRAS.330..755F} Fairley, B.~W., Jones,
L.~R., Wake, D.~A., Collins, C.~A., Burke, D.~J., Nichol, R.~C., \& Romer,
A.~K.\ 2002, \mnras, 330, 755
\bibitem[Farouki \& Shapiro(1980)]{1980ApJ...241..928F} Farouki, R.~\& Shapiro, S.~L.\ 1980, \apj, 241, 928
\bibitem[Fasano et al.(2000)]{2000ApJ...542..673F} Fasano, G.,
Poggianti, B.~M., Couch, W.~J.,
Bettoni, D., Kj{\ae}rgaard, P., \&
Moles, M. 2000, \apj, 542, 673
\bibitem{}Finoguenov, A., Briel, U.G., Henry, J.P., 2003, submitted to
A\&A
\bibitem[Fujita(1998)]{1998ApJ...509..587F} Fujita, Y.\ 1998, \apj, 509,
587
\bibitem[Fujita \& Nagashima(1999)]{1999ApJ...516..619F} Fujita, Y.~\& Nagashima, M.\ 1999, \apj, 516, 619
\bibitem[Fujita(2001)]{2001ApJ...550..612F} Fujita, Y.\ 2001, \apj, 550,
612
\bibitem[Fukugita, Shimasaku, \& Ichikawa(1995)]{1995PASP..107..945F}
Fukugita, M., Shimasaku, K., \& Ichikawa,
T.\ 1995, \pasp, 107, 945
\bibitem[Fukugita et al.(1996)]{1996AJ....111.1748F} Fukugita, M.,
Ichikawa, T., Gunn, J.~E., Doi, M., Shimasaku, K., \& Schneider, D.~P.\
1996, \aj, 111, 1748.
\bibitem[Gal et al.(2000)]{2000AJ....120..540G} Gal, R.~R., de Carvalho,
R.~R., Brunner, R., Odewahn, S.~C., \& Djorgovski, S.~G.\ 2000, \aj, 120,
540
\bibitem[Garilli et al.(1996)]{1996ApJS..105..191G} Garilli, B., Bottini,
D., Maccagni, D., Carrasco, L., \& Recillas, E.\ 1996, \apjs, 105, 191
\bibitem[Goto et al.(2002)]{2002AJ....123.1807G} Goto, T.~et al.\ 2002a,
\aj, 123, 1807.
\bibitem{}Goto, T., Okamura, S. \& Brinkman, J. 2002b, PASJ, 54, 4
\bibitem[Gladders \& Yee(2000)]{2000AJ....120.2148G} Gladders, M.~D.~\& Yee, H.~K.~C.\ 2000, \aj, 120, 2148
\bibitem[Gunn \& Gott(1972)]{1972ApJ...176....1G} Gunn, J.~E.~\& Gott,
J.~R.~I.\ 1972, \apj, 176, 1
\bibitem{}Gunn, J.E., Carr, M., Rockosi, C., Sekiguchi, M., Berry, K., Elms, B., de Haas, E., Ivezic , Z. et al.\ 1998, AJ, 116, 3040 % camera paper
\bibitem{}Gunn, J.E., Hoessel, J.G., \& Oke, J.B.\ 1986, \apj, 306, 30.
\bibitem[Hammer et al.(1997)]{1997ApJ...481...49H} Hammer, F., Flores, H.,
Lilly, S.~J., Crampton, D., Le Fevre, O., Rola, C., Mallen-Ornelas, G.,
Schade, D., \& Tresse, L.\ 1997, \apj, 481, 49
\bibitem[Henry \& Arnaud(1991)]{1991ApJ...372..410H} Henry, J.~P.~\&
Arnaud, K.~A.\ 1991, \apj, 372, 410
\bibitem[Hogg, Finkbeiner, Schlegel, \& Gunn(2001)]{2001AJ....122.2129H}
Hogg, D.~W., Finkbeiner, D.~P., Schlegel, D.~J., \& Gunn, J.~E.\ 2001, \aj,
122, 2129
\bibitem[Icke(1985)]{1985A&A...144..115I} Icke, V.\ 1985, \aap, 144, 115
\bibitem[Jones \& Forman(1978)]{1978ApJ...224....1J} Jones, C.~\& Forman,
W.\ 1978, \apj, 224, 1
\bibitem[Jones \& Forman(1999)]{1999ApJ...511...65J} Jones, C.~\& Forman,
W.\ 1999, \apj, 511, 65
\bibitem[Kauffmann(1995)]{1995MNRAS.274..153K} Kauffmann, G.\ 1995, \mnras,
274, 153
\bibitem[Kent(1981)]{1981ApJ...245..805K} Kent, S.~M.\ 1981, \apj, 245, 805
\bibitem[Kim et al.(2002)]{2002AJ....123...20K} Kim, R.~S.~J.~et al.\ 2002,
\aj, 123, 20
\bibitem[Kodama et al.(2001)]{2001ApJ...562L...9K} Kodama, T., Smail, I.,
Nakata, F., Okamura, S., \& Bower, R.~G.\ 2001, \apjl, 562, L9
\bibitem[Kodama \& Bower(2001)]{2001MNRAS.321...18K} Kodama, T.~\& Bower,
R.~G.\ 2001, \mnras, 321, 18
\bibitem[Kodama \& Smail(2001)]{2001MNRAS.326..637K} Kodama, T.~\& Smail,
I.\ 2001, \mnras, 326, 637
\bibitem[Lahav et al.(1995)]{1995Sci...267..859L} Lahav, O.~et al.\ 1995,
Science, 267, 859
\bibitem[Larson, Tinsley, \& Caldwell(1980)]{1980ApJ...237..692L} Larson,
R.~B., Tinsley, B.~M., \& Caldwell, C.~N.\ 1980, \apj, 237, 692
\bibitem[Lavery \& Henry(1988)]{1988ApJ...330..596L} Lavery, R.~J.~\&
Henry, J.~P.\ 1988, \apj, 330, 596
\bibitem[Lea \& Henry(1988)]{1988ApJ...332...81L} Lea, S.~M.~\& Henry,
J.~P.\ 1988, \apj, 332, 81
\bibitem[Lilly, Le Fevre, Hammer, and Crampton(1996)]{1996ApJ...460L...1L}
Lilly, S.~J., Le Fevre, O., Hammer, F., \& Crampton, D.\ 1996, \apjl, 460,
L1
\bibitem[Lilly et al.(1998)]{1998ApJ...500...75L} Lilly, S., Schade, D.,
Ellis, R., Le Fevre, O., Brinchmann, J., Tresse, L., Abraham, R., Hammer,
F., et al.\ 1998, \apj, 500, 75
\bibitem[Lin et al.(1999)]{1999ApJ...518..533L} Lin, H., Yee, H.~K.~C.,
Carlberg, R.~G., Morris, S.~L., Sawicki, M., Patton, D.~R., Wirth, G., \& Shepherd, C.~W.\ 1999, \apj, 518, 533
\bibitem[Loveday, Peterson, Efstathiou, \& Maddox(1992)]{1992ApJ...390..338L} Loveday, J., Peterson, B.~A.,
Efstathiou, G., \& Maddox, S.~J.\ 1992, \apj, 390, 338
\bibitem[Lupton, Gunn, \& Szalay(1999)]{1999AJ....118.1406L} Lupton, R.~H., Gunn, J.~E., \& Szalay, A.~S.\ 1999, \aj, 118, 1406
\bibitem[Lupton et al.(2001)]{2001adass..10..269L} Lupton, R.~H., Gunn, J.~E., Ivezi{\' c}, Z., Knapp, G.~R., Kent, S., \& Yasuda, N.\ 2001, Astronomical Data Analysis Software and Systems X, ASP Conference Proceedings, Vol.~238.~Edited by F.~R.~Harnden, Jr., Francis A.~Primini, and Harry E.~Payne.~San Francisco: Astronomical Society of the Pacific, ISSN: 1080-7926, 2001., p.269, 10, 269
\bibitem{}Lupton, R. et al. 2002, $in$ $preparation$
\bibitem[Madau et al.(1996)]{1996MNRAS.283.1388M} Madau, P., Ferguson,
H.~C., Dickinson, M.~E., Giavalisco, M., Steidel, C.~C., \& Fruchter, A.\
1996, \mnras, 283, 1388
\bibitem[Makino \& Hut(1997)]{1997ApJ...481...83M} Makino, J.~\& Hut, P.\
1997, \apj, 481, 83
\bibitem[Mamon(1992)]{1992ApJ...401L...3M} Mamon, G.~A.\ 1992, \apjl, 401,
L3
\bibitem[Margoniner \& de Carvalho(2000)]{2000AJ....119.1562M} Margoniner,
V.~E.~\& de Carvalho, R.~R.\ 2000, \aj, 119, 1562
\bibitem[Margoniner, de Carvalho, Gal, \& Djorgovski(2001)]{2001ApJ...548L.143M} Margoniner, V.~E., de Carvalho,
R.~R., Gal, R.~R., \& Djorgovski, S.~G.\ 2001, \apjl, 548, L143
\bibitem[Markevitch(1998)]{1998ApJ...504...27M} Markevitch, M.\ 1998, \apj,
504, 27
\bibitem[Marzke, Geller, Huchra, \& Corwin(1994)]{1994AJ....108..437M}
Marzke, R.~O., Geller, M.~J., Huchra, J.~P., \& Corwin, H.~G.\ 1994, \aj,
108, 437
\bibitem[Marzke \& da Costa(1997)]{1997AJ....113..185M} Marzke, R.~O.~\& da
Costa, L.~N.\ 1997, \aj, 113, 185
\bibitem[Marzke et al.(1998)]{1998ApJ...503..617M} Marzke, R.~O., da Costa,
L.~N., Pellegrini, P.~S., Willmer, C.~N.~A., \& Geller, M.~J.\ 1998, \apj,
503, 617
\bibitem[Metevier, Romer, \& Ulmer(2000)]{2000AJ....119.1090M} Metevier,
A.~J., Romer, A.~K., \& Ulmer, M.~P.\ 2000, \aj, 119, 1090
\bibitem[Mihos(1995)]{1995ApJ...438L..75M} Mihos, J.~C.\ 1995, \apjl, 438,
L75
\bibitem{}Miller, C. J. et al. astro-ph/9912362
\bibitem{}Miller, C. J. et al. 2002 $in$ $preparation$
\bibitem[Mitchell, Ives, and Culhane(1977)]{1977MNRAS.181P..25M} Mitchell,
R.~J., Ives, J.~C., \& Culhane, J.~L.\ 1977, \mnras, 181, 25P
\bibitem[Mo \& Mao(2002)]{2002MNRAS.333..768M} Mo, H.~J.~\& Mao, S.\
2002, \mnras, 333, 768
\bibitem[Moore et al.(1996)]{1996Natur.379..613M} Moore, B., Katz, N.,
Lake, G., Dressler, A., \& Oemler, A.\ 1996, \nat, 379, 613
\bibitem[Moore, Lake, Quinn, \& Stadel(1999)]{1999MNRAS.304..465M} Moore,
B., Lake, G., Quinn, T., \& Stadel, J.\ 1999, \mnras, 304, 465
\bibitem[Morgan(1958)]{1958PASP...70..364M} Morgan, W.~W.\ 1958, \pasp, 70,
364
\bibitem[Morgan(1959)]{1959PASP...71..394M} Morgan, W.~W.\ 1959, \pasp, 71,
394
\bibitem[Newberry, Kirshner, \& Boroson(1988)]{1988ApJ...335..629N}
Newberry, M.~V., Kirshner, R.~P., \& Boroson, T.~A.\ 1988, \apj, 335,
629
\bibitem{}Oh, S.P. \& Benson, A.J. 2003, submitted to MNRAS
\bibitem[Ostriker(1980)]{1980ComAp...8..177O} Ostriker, J.~P.\ 1980,
Comments on Astrophysics, 8, 177
% \bibitem{}Pier, J.R., et al 2002 AJ, submitted
\bibitem[Pier et al.(2003)]{2003AJ....125.1559P} Pier, J.~R., Munn, J.~A.,
Hindsley, R.~B., Hennessy, G.~S., Kent, S.~M., Lupton, R.~H., \& Ivezi{\'
c}, {\v Z}.\ 2003, \aj, 125, 1559
\bibitem[Poggianti et al.(1999)]{1999ApJ...518..576P} Poggianti, B.~M.,
Smail, I., Dressler, A., Couch, W.~J., Barger, A.~J., Butcher, H., Ellis,
R.~S., \& Oemler, A.~J.\ 1999, \apj, 518, 576
\bibitem[Quilis, Moore, \& Bower(2000)]{2000Sci...288.1617Q} Quilis, V.,
Moore, B., \& Bower, R.\ 2000, Science, 288, 1617
\bibitem[Rakos \& Schombert(1995)]{1995ApJ...439...47R} Rakos, K.~D.~\&
Schombert, J.~M.\ 1995, \apj, 439,
47
\bibitem[Reichart, Castander, \& Nichol(1999)]{1999ApJ...516....1R}
Reichart, D.~E., Castander, F.~J., \& Nichol, R.~C.\ 1999, \apj, 516, 1
\bibitem[Schade et al.(1996)]{1996ApJ...464...79S} Schade, D., Lilly,
S.~J., Le Fevre, O., Hammer, F., \& Crampton, D.\ 1996, \apj, 464, 79
\bibitem[Schlegel, Finkbeiner, \& Davis(1998)]{1998ApJ...500..525S}
Schlegel, D.~J., Finkbeiner, D.~P., \& Davis, M.\ 1998, \apj, 500, 525
\bibitem{}Scranton, R. et al. 2002 $in$ $preparation$
\bibitem[Sellwood \& Carlberg(1984)]{1984ApJ...282...61S} Sellwood,
J.~A.~\& Carlberg, R.~G.\ 1984, \apj, 282, 61
\bibitem[Shimasaku et al.(2001)]{2001AJ....122.1238S} Shimasaku, K.~et
al.\ 2001, \aj, 122, 1238
\bibitem[Shioya, Bekki, Couch, \& De Propris(2002)]{2002ApJ...565..223S}
Shioya, Y., Bekki, K., Couch, W.~J., \& De Propris, R.\ 2002, \apj, 565, 223
\bibitem[Smail et al.(1997)]{1997ApJS..110..213S} Smail, I., Dressler, A.,
Couch, W.~J., Ellis, R.~S., Oemler, A.~J., Butcher, H., \& Sharples, R.~M.\
1997, \apjs, 110, 213
\bibitem[Smail, Edge, Ellis, \& Blandford(1998)]{1998MNRAS.293..124S}
Smail, I., Edge, A.~C., Ellis, R.~S., \& Blandford, R.~D.\ 1998, \mnras,
293, 124
\bibitem[Smith et al.(2002)]{2002AJ....123.2121S} Smith, J.~A.~et al.\
2002, \aj, 123, 2121
\bibitem[Spitzer \& Baade(1951)]{1951ApJ...113..413S} Spitzer, L.~J.~\& Baade, W.\ 1951, \apj, 113, 413
\bibitem[Stevens, Acreman, \& Ponman(1999)]{1999MNRAS.310..663S} Stevens,
I.~R., Acreman, D.~M., \& Ponman, T.~J.\ 1999, \mnras, 310, 663
% \bibitem{}Strauss, M.A., et al 2002, AJ, in press
\bibitem[Strauss et al.(2002)]{2002AJ....124.1810S} Strauss, M.~A.~et al.\
2002, \aj, 124, 1810
\bibitem[Strateva et al.(2001)]{2001AJ....122.1861S} Strateva, I.~et al.\ 2001, \aj, 122, 1861
\bibitem[Stoughton et al.(2002)]{2002AJ....123..485S} Stoughton, C.~et al.\
2002, \aj, 123, 485
\bibitem[Sullivan et al.(2000)]{2000MNRAS.312..442S} Sullivan, M., Treyer,
M.~A., Ellis, R.~S., Bridges, T.~J., Milliard, B., \& Donas, J.~; 2000,
\mnras, 312, 442
\bibitem[Treyer et al.(1998)]{1998MNRAS.300..303T} Treyer, M.~A., Ellis,
R.~S., Milliard, B., Donas, J., \& Bridges, T.~J.\ 1998, \mnras, 300, 303
\bibitem[Valotto, Moore, \& Lambas(2001)]{2001ApJ...546..157V} Valotto,
C.~A., Moore, B., \& Lambas, D.~G.\ 2001, \apj, 546, 157
\bibitem[Valluri(1993)]{1993ApJ...408...57V} Valluri, M.\ 1993, \apj, 408,
57
\bibitem[van den Bergh(1976)]{1976ApJ...206..883V} van den Bergh, S.\ 1976,
\apj, 206, 883
\bibitem[Vollmer, Cayatte, Balkowski, \& Duschl(2001)]{2001ApJ...561..708V}
Vollmer, B., Cayatte, V., Balkowski, C., \& Duschl, W.~J.\ 2001, \apj, 561,
708
\bibitem[Wang \& Ulmer(1997)]{1997MNRAS.292..920W} Wang, Q.~D.~\& Ulmer,
M.~P.\ 1997, \mnras, 292, 920
\bibitem[White, Jones, and Forman(1997)]{1997MNRAS.292..419W} White, D.~A.,
Jones, C., \& Forman, W.\ 1997, \mnras, 292, 419
% \bibitem{}Wilson, G., Cowie, L.L., Barger,A.J.,Burke, D.J. 2002,
% submitted to AJ
\bibitem[Wilson, Cowie, Barger, \& Burke(2002)]{2002AJ....124.1258W}
Wilson, G., Cowie, L.~L., Barger, A.~J., \& Burke, D.~J.\ 2002, \aj, 124,
1258
\bibitem[Wu, Xue, and Fang(1999)]{1999ApJ...524...22W} Wu, X., Xue, Y., \&
Fang, L.\ 1999, \apj, 524, 22
\bibitem[Xue \& Wu(2000)]{2000ApJ...538...65X} Xue, Y.~\& Wu, X.\ 2000,
\apj, 538, 65
\bibitem[Yagi et al.(2002)]{2002AJ....123...87Y} Yagi, M., Kashikawa, N.,
Sekiguchi, M., Doi, M., Yasuda, N., Shimasaku, K., \& Okamura, S.\ 2002,
\aj, 123, 87
\bibitem[Yagi et al.(2002)]{2002AJ....123...66Y} Yagi, M., Kashikawa, N.,
Sekiguchi, M., Doi, M., Yasuda, N., Shimasaku, K., \& Okamura, S.\ 2002,
\aj, 123, 66
\bibitem[Yee, Ellingson, \& Carlberg(1996)]{1996ApJS..102..269Y} Yee,
H.~K.~C., Ellingson, E., \& Carlberg, R.~G.\ 1996, \apjs, 102, 269
\bibitem[York et al.(2000)]{2000AJ....120.1579Y} York, D.~G.~et al.\ 2000, \aj, 120, 1579
%\bibitem[Zabludoff et al.(1996)]{1996ApJ...466..104Z} Zabludoff, A.~I.,
%Zaritsky, D., Lin, H., Tucker, D., Hashimoto, Y., Shectman, S.~A., Oemler,
%A., \& Kirshner, R.~P.\ 1996, \apj, 466, 104
\end{thebibliography}
%\end{document}
\newpage
\begin{figure}[h]
\begin{center}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020424_bo4_error.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020510_bo.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020519_bo_photoz.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020520_bo_photoz.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020525_bo_photoz.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020609_bo_photoz.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020616_bo_photoz_yagi.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020623_bo_photoz.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020902_bo_global.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020909_bo_global.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/021031_bo_annuli_yagi.ps}
\includegraphics[scale=0.7]{030321_bo_annuli_yagi.ps}
\end{center}
\caption{\label{fig:bo}
Photometric and morphological Butcher-Oemler effect from the 514 SDSS Cut \&
Enhance clusters.
$f_b$, $f_{Cin}$, $f_{exp}$ and $f_{u-r}$ are plotted against redshift.
The dashed lines show the weighted least-squares fit to the data. The stars
and solid lines show the median values. The median values of errors
are shown in the upper left corners of each panel. The Spearman's correlation coefficients are shown in Table \ref{tab:correlation}.
}
\end{figure}
\newpage
\begin{figure}[h]
\begin{center}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020609_bo_ks.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020909_bo_ks_normalize.ps}
\includegraphics[scale=0.7]{021031_bo_ks_normalize.ps}
\end{center}
\caption{\label{fig:kstest}
Normalized distributions of late type fractions ($f_b$, $f_{Cin}$, $f_{exp}$ and $f_{u-r}$).
Dashed lines show distributions of lower redshift clusters
(z$\leq$0.15) and solid lines show ones of higher redshift clusters
(0.15$<$z$\leq$0.3).
The results of Kolomogorov-Smirnov tests are shown in Table \ref{tab:kstest}. In all
cases, Kolomogorov-Smirnov tests show the two distributions are significantly different.
}
\end{figure}
\newpage
\begin{figure}[h]
\begin{center}
\includegraphics[scale=0.7]{020610_z_rich.ps}
\end{center}
\caption{
\label{fig:z_rich}
Richness distribution as a function of redshift. Richnesses are measure
as the number of galaxies brighter than $M_r=-19.44$ within 0.7 Mpc after
fore/background subtraction.
}
\end{figure}
\newpage
\begin{figure}[h]
\begin{center}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020609_okamura_rich.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020909_okamura_rich.ps}
\includegraphics[scale=0.7]{021031_okamura_rich.ps}
\end{center}
\caption{
\label{fig:okamura_rich}
The difference of the late type fractions from the best
fit lines as a function of redshift are plotted against cluster richnesses. Solid lines and stars
show the median values.
}
\end{figure}
\newpage
\begin{figure}[h]
\begin{center}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020609_okamura_elong.ps}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020909_okamura_elong.ps}
\includegraphics[scale=0.7]{021031_okamura_elong.ps}
\end{center}
\caption{
\label{fig:okamura_elong}
The difference of late type fractions from the best
fit lines as a function of redshift are plotted against cluster elongation, which was measured as
a ratio of major axis to minor axis on a cluster detection map of Goto
et al. (2002a). Solid lines and stars show median values.
}
\end{figure}
\newpage
\begin{figure}[h]
\begin{center}
\includegraphics[scale=0.7]{020525_cin_g_r.ps}
\end{center}
\caption{
\label{fig:cin_g_r}
The distribution of the inverse of the concentration index ($C_{in}$), defined
as the ratio of Petrosian 50\% flux radius to Petrosian 90\% flux
radius. The solid line shows the distribution of $C_{in}$ measured in
the $g$ band image. The dashed line shows the distribution of $C_{in}$ measured in
the $r$ band image. The difference between the $g$ band and $r$ band is
marginal, assuring our usage of $r$ band $C_{in}$ in the right upper
panel of Figure \ref{fig:bo} from $z$=0.02 to $z$=0.3. The
statistics are summarized in Table \ref{tab:cin_g_r}.
}
\end{figure}
\newpage
\begin{figure}[h]
\begin{center}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020611_cin_seeing.ps}
\includegraphics[scale=0.7]{020613_cin_seeing.ps}
\end{center}
\caption{
\label{fig:seeing_cin}
The dependence of $f_{Cin}$ on seeing. Open squares and solid lines show
the distribution and medians of low z clusters (z$\leq$0.15). Filled
triangles and dashed lines show the distribution and medians of high z
clusters (0.24$<$z$\leq$0.3). The median bins are chosen so that equal
numbers of galaxies are included in each bin.
}
\end{figure}
\newpage
\begin{figure}[h]
\begin{center}
\includegraphics[scale=0.7]{020609_seeing_hist.ps}
\end{center}
\caption{
\label{fig:seeing_hist}
The seeing distribution of all galaxies brighter than $r$=21.5. 87\% of all
galaxies have seeing better than 2.0 arcsec.
}
\end{figure}
\newpage
\begin{figure}[h]
\begin{center}
\includegraphics[scale=0.7]{020613_exp_seeing.ps}
\end{center}
\caption{
\label{fig:seeing_exp}
The dependence of $f_{exp}$ on seeing. Open squares and solid lines show
the distribution and medians of low z clusters (z$\leq$0.15). Filled
triangles and dashed lines show the distribution and medians of high z
clusters (0.24$<$z$\leq$0.3). Median bins are chosen so that equal
numbers of galaxies are included in each bin.
1 $\sigma$ errors shown as vertical bars are more dominant.
There is no significant trend with seeing.
}
\end{figure}
\newpage
%
%\begin{figure}[h]
%\begin{center}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020609_seeing_ratio.ps}
%\end{center}
%\caption{
%\label{fig:seeing_ratio_cin}
% Dependence on seeing of $f_{Cin}$. All galaxies brighter than
% $r^*$=21.5 are included in the plot. The relative numbers of galaxies
% in each bin are shown in Figure \ref{fig:seeing_hist}.
%}
%\end{figure}
%
%\newpage
%
%\begin{figure}[h]
%\begin{center}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020609_seeing_ratio_exp.ps}
%\end{center}
%\caption{
%\label{fig:seeing_ratio_exp}
% Dependence of $f_{exp}$ on seeing. All galaxies brighter than
% $r^*$=21.5 are included in the plot. The relative numbers of galaxies
% in each bin are shown in Figure \ref{fig:seeing_hist}.
%}
%\end{figure}
%
\newpage
\begin{figure}[h]
\begin{center}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020609_bg_test2.ps}
\includegraphics[scale=0.7]{020612_bg_test2.ps}
\end{center}
\caption{
\label{fig:bg_test}
Various systematic tests. Solid lines show distributions for
2.1-2.21 Mpc annular fore/background subtraction. Dashed lines show
distributions for a global background subtraction. Dotted lines show
distributions using 0.7/(1+z) Mpc radius assuming a standard cold dark matter cosmology.
Long dashed lines show distributions using the brightest galaxy position as a cluster center.
In none of the cases does a Kolomogorov-Smirnov test show
significant difference between the distributions (significance to be different is less than 26\% in all cases).
}
\end{figure}
\newpage
\begin{figure}[h]
\begin{center}
\includegraphics[scale=0.7]{020629_imaging_vs_spectra_ned.ps}
\end{center}
\caption{
\label{fig:comparison_with_spectroscopy}
Comparison with late-type fraction from spectroscopy.
Late-type fractions measured using spectroscopic data are plotted
against that from imaging data for three clusters with z$<$0.06 (ABELL 295, RXC
J0114.9+0024, and ABELL 957). Dashed lines are drawn to guide
eyes. All points agree with each other within the error.
}
\end{figure}
%
%\newpage
%\begin{figure}[h]
%\begin{center}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/020609_okamura_concent.ps}
%\end{center}
%\caption{
%\label{fig:okamura_concent}
%Concentration is measured as a ratio of numbers of galaxies within 0.5
% Mpc and 0.7 Mpc. Should this figure be included?
%}
%\end{figure}
%
\newpage
\begin{figure}[h]
\begin{center}
\includegraphics[scale=0.7]{030315_new_rad.ps}
\end{center}
\caption{
Distribution of varying radius to measure blue/spiral fractions. It has
a peak at 0.7 Mpc.
}\label{fig:new_rad}
\end{figure}
\newpage
\begin{figure}[h]
\begin{center}
%\includegraphics[scale=0.7]{/home/tomo/public_html/research/out/030317_bo_vary_rad.ps}
\includegraphics[scale=0.7]{030329_bo_vary_rad.ps}
\end{center}
\caption{
The same as figure \ref{fig:bo}, but measured with varying radius.
}\label{fig:bo_new}
\end{figure}
\newpage
\begin{figure}[h]
\begin{center}
\includegraphics[scale=0.7]{030317_okamura_rich.ps}
\end{center}
\caption{
The same as figure \ref{fig:okamura_rich}, but measured with varying radius.
}\label{fig:okamura_rich_new}
\end{figure}
\newpage
\begin{figure}[h]
\begin{center}
\includegraphics[scale=0.7]{030317_okamura_elong.ps}
\end{center}
\caption{
The same as figure \ref{fig:okamura_elong}, but measured with varying radius.
}\label{fig:okamura_elong_new}
\end{figure}
\newpage
%
%
%
%\begin{table}[h]
%\caption{
%\label{tab:correlation}
%Correlation coefficients between $z$ and fraction of late type galaxeis.
%}
%\begin{center}
%\begin{tabular}{llll}
%\hline
% & Correlation coefficient & Significance & N clusters \\
%\hline
%\hline
% $f_{exp}$ & 0.257915 & 0.0165066 & 86\\
% $f_{Cin}$ & 0.426798 & 4.16980e-05 & 86\\
% $f_b$ & 0.227926 & 0.0348036 & 86\\
% $f_{u-r}$ & 0.237315 & 0.0278002 & 86\\
%\hline
%\end{tabular}
%\end{center}
%\end{table}
%
%\begin{table}[h]
%\caption{
%\label{tab:correlation}
% Spearman's correlation coefficients between $z$ and fractions of late type
% galaxeis. 514 clusters with richness$>$25 are chosen as a sample.
%}
%\begin{center}
%\begin{tabular}{llll}
%\hline
% & Correlation coefficient & Significance & N clusters \\
%\hline
%\hline
% $f_b$ & 0.206 & 7.3$\times$10$^{-5}$ & 514\\
% $f_{u-r}$ & 0.196 & 1.6$\times$10$^{-4}$ & 514\\
% $f_{exp}$ & 0.191 & 2.5$\times$10$^{-4}$ & 514\\
% $f_{Cin}$ & 0.233 & 6.7$\times$10$^{-6}$ & 514\\
%% $f_{exp}$ & 0.190527 & 0.000251100 & 514\\
%% $f_{Cin}$ & 0.233304 & 6.65916e-06 & 514\\
%% $f_b$ & 0.206153 & 7.25556e-05& 514\\
%% $f_{u-r}$ & 0.196099 & 0.000163062 & 514\\
%\hline
%\end{tabular}
%\end{center}
%\end{table}
\begin{table}[h]
\caption{
\label{tab:correlation}
Spearman's correlation coefficients between $z$ and fractions of late type
galaxeis. 514 clusters with richness$>$25 are chosen as a sample.
}
\begin{center}
\begin{tabular}{llll}
\hline
& Correlation coefficient & Significance & N clusters \\
\hline
\hline
$f_b$ & 0.238 & 4.4$\times$10$^{-8}$ & 514\\
$f_{u-r}$ & 0.234 & 7.6$\times$10$^{-8}$ & 514\\
$f_{exp}$ & 0.194 & 9.6$\times$10$^{-6}$ & 514\\
$f_{Cin}$ & 0.223 & 2.9$\times$10$^{-7}$ & 514\\
\hline
\end{tabular}
\end{center}
\end{table}
\begin{table}[h]
\caption{
\label{tab:kstest}
Significances in Kolomogorov-Smirnov tests between distributions
for z$\leq$0.15 and 0.15$<$z$<$0.3. In all cases, Kolomogorov-Smirnov tests show the distributions for
the lower redshift sample and the higher redshift sample are significantly different.
}
\begin{center}
\begin{tabular}{ll}
\hline
& Significance \\
\hline
\hline
$f_b$ & 2.9$\times$10$^{-3}$ \\
$f_{u-r}$ & 1.0$\times$10$^{-3}$ \\
$f_{exp}$ & 3.4$\times$10$^{-4}$ \\
$f_{Cin}$ & 2.9$\times$10$^{-3}$ \\
%1.5891 0.0128124
%1.87803 0.00172791
%1.87803 0.00172791
%1.87803 0.0017279
\hline
\end{tabular}
\end{center}
\end{table}
\begin{table}[h]
\caption{
\label{tab:error_comparison}
Scatters in late-type fractions around the best-fit line are compared
with median errors of late-type fraction calculated with equation (\ref{frac_err})
}
\begin{center}
\begin{tabular}{lllll}
\hline
& $f_b$ & $f_{u-r}$ & $f_{exp}$ & $f_{Cin}$\\
\hline
\hline
Real scatter (1$\sigma$) & 0.169 & 0.183 & 0.171 & 0.163 \\
% Error estimate & 0.111 & 0.137 & 0.140 & 0.160 \\
Error estimate & 0.078 & 0.069 & 0.050 & 0.089 \\
\hline
\end{tabular}
\end{center}
\end{table}
\begin{table}[h]
\caption{
\label{tab:cin_g_r}
Change in the fraction of galaxies with $C_{in}>$0.4 (late type) in two different filters($g,r$).
}
\begin{center}
\begin{tabular}{llll}
\hline
band & N($C_{in}>$0.4) & N(total) & Percentage(\%) \\
\hline
\hline
$g$ & 802 & 1336 & 60.0 \\
$r$ & 787 & 1336 & 58.9 \\
\hline
Difference & 15 & 1336 & 1.1\\
\hline
\end{tabular}
\end{center}
\end{table}
\begin{table}[h]
\caption{
\label{tab:exp_g_r}
Change in the fraction of galaxies with exponential fit likelihood
greater than de Vaucouleur likelihood (late type) in two different
filters($g,r$). Since we discard the galaxies
with the same likelihood in this analysis, the total number of galaxies
in the sample are different in $g$ and $r$.
}
\begin{center}
\begin{tabular}{llll}
\hline
band & N(late) & N(late + early) & Percentage(\%) \\
\hline
\hline
$g$ & 503 & 804 & 62.6 \\
$r$ & 476 & 792 & 60.1 \\
\hline
Difference & - & - & 2.5\\
\hline
\end{tabular}
\end{center}
\end{table}
\end{document}