On the Formation of Giant Elliptical Galaxi Es And

Journal of The Korean Astronomical Society

36: 189 212, 2003

ON THE FORMATION OF GIANT ELLIPTICAL GALAXIES AND GLOBULAR

CLUSTERS

Myung Gyoon Lee

Astronomy Program, SEES, Seoul National University, Seoul 151-742, Korea

E-mail: [email protected]

(Received June 26, 2003; Accepted August 17, 2003)

ABSTRACT

I review the current status of understanding when, how long, and how giant elliptical galaxies formed,

fo cusing on the globular clusters. Several observational evidences show that massive elliptical galaxies

formed at z > 2 (> 10 Gyr ago). Giant elliptical galaxies show mostly a bimo dal color distribution

20 metallicity di erence b etween the two p eaks. The red of globular clusters, indicating a factor of

globular clusters (RGCs) are closely related with the stellar halo in color and spatial distribution, while

are not. The ratio of the numb er of the RGCs and that of the BGCs the blue globular clusters (BGCs)

varies dep ending on galaxies. It is concluded that the BGCs might have formed 12{13 Gyr ago, while

the RGCs and giant elliptical galaxies might have formed similarly 10-11 Gyr ago. It remains now

to explain the existence of a gap b etween the RGC formation ep o ch and the BGC formation ep o ch,

and the rapid metallicity increase during the gap ( t 2 Gyr). If hierarchical merging can form a

signi can t numb er of giant elliptical galaxies > 10 Gyr ago, several observational constraints from stars

and globular clusters in elliptical galaxies can b e explained.

words : galaxies:general | galaxies:formation | galaxies: globular clusters | galaxies: elliptical Key

I. INTRODUCTION:UNSOLVED MYSTERY clusters we had in the past b efore mo dern data came

out.

Galaxies are a gateway to understanding the for-

The history changed in the 1990's with the advent

mation and evolution of large-scale structures in the

of the Hubble Space Telescop e, large-format CCD cam-

universe. Morphological typ es of galaxies are diverse

eras and large ground-based telescop es. It turned out

from dwarf spheroidal galaxies to spiral galaxies with

that the elliptical galaxies and globular clusters are not

b eautiful arms, and masses of galaxies span a very large

that simple. Two of the most noteworthy ndings re-

12 7

M of dwarf galaxies to 10 M of gi- range from 10

lated to globular clusters in elliptical galaxies are 1)

ant elliptical galaxies. Here I will concentrate only on

that the color (and metallicity) distribution of globu-

< 20 mag) which are giant elliptical galaxies (M

lar clusters in elliptical galaxies is often bimo dal (e.g.,

often found in the centers of rich clusters.

& Geisler 1993, Whitmore et al. 1995, El- M87: Lee

3 6

Globular clusters (with masses of 10 to 10 M )

son & Santiago 1996; M49: Geisler, Lee, & Kim 1996),

are found in many of these galaxies (Ashman & Zepf

and 2) that there exist many blue bright clusters in the

1998). The brightest old globular clusters (M 10

systems of galaxies that are much interacting/merging

mag) are brighter than the lowest-mass dwarf galaxies,

brighter than the Galactic halo globular clusters. These

but the globular clusters (with half mass radii of order

can b e progenitors of globular clusters (e.g., Whitmore

2{10 p c) are incomparably smaller than dwarf galaxies

& Schweizer 1995, Whitmore 2000, Schweizer 2002a,b).

arf galaxies which are larger than a few 100 p c. Dw

These ndings showed that these simple-lo oking stel-

host globular clusters, but globular clusters cannot host

lar systems are complex in reality, and that globular

dwarf galaxies.

clusters are forming in the pro cess of galaxy merging,

Among the diverse galaxies, elliptical galaxies (and

providing critical constraints to mo dels of galaxy for-

the bulges of spiral galaxies as well) share some com- mation.

mon prop erties with globular clusters. At rst lo ok

When, how long, and how these galaxies and globu-

b oth lo ok monotonously simple (elliptical or globular

lar clusters formed (formation ep o ch, formation dura-

shap e in uniform color in the images). This leads us to

in mo d- tion, and formation pro cess) is a key question

consider that they must b e comp osed of one single old

ern cosmology. It is a long-standing question. Mo d-

stellar p opulation, and to conclude that they may b e

came out as early as els of formation of these ob jects

the rst stellar systems which formed in the universe.

in 1960's and 1970's (Eggen, Lynden-Bell, & Sandage

That is an old picture of elliptical galaxies and globular

1962, Partridge & Peebles 1967, Peebles & Dicke 1968,

Tinsley 1972, Larsen 1974, To omre 1977), and mo dels

based on numerical simulations with high resolution

Pro ceedings of the APCTP Workshop on Formation and Inter-

are coming out to day (e.g., Steinmetz & Navarro 2003,

action of Galaxies

{ 189 {

190 LEE

Fig. 1.| A color image of M49, the brightest galaxy (gE) in the Virgo cluster, created by combining C and T images

0 0

tak en at the KPNO 4m. The eld of view is 16 16 . A fraction of stellar halo light of M49 was subtracted from the original

oint sources are globular clusters. Several dwarf elliptical image to show b etter the globular clusters. Most of the faint p

galaxies are also seen around M49. One blue feature which lo oks like an anchor in the south-east of M49 is UGC 7636, a

dwarf irregular galaxy interacting with M49.

Bekki et al. 2002, Meza et al. 2003, Beasley et al. 2003, galaxies.

Kravtsov & Gnedin 2003). In spite of many e orts for

1 shows an example of a giant elliptical Figure

this problem over many years, this problem is still an

galaxy, M49, with its globular clusters. M49 (NGC

unsolved mystery.

clus- 4472) is the brightest galaxy (E2/S0) in the Virgo

ter which is at the distance of ab out 15 Mp c. The Most previous work on the formation of elliptical

image was created by combining C images ob- galaxies is based on integrated prop erties of unresolved and T

tained at the KPNO stars in the galaxies. However, globular clusters play 4m. A signi cant fraction of stel-

a signi cant role in our understanding how elliptical lar halo light of M49 was subtracted from the original

galaxies formed. There was only limited observational images to show b etter the p oint sources in the image.

information of globular clusters in elliptical galaxies Note the numerous p oint sources concentrated around

in the past, but new observational information ab out the center of M49, most of which are genuine globular

o ding these days, making mo delers busy. It them is o clusters. This gure shows that those globular clusters

is necessary to use b oth stars and globular clusters to can b e a critical to ol to investigate the nature of ellip-

solve the unknown mystery of the formation of elliptical tical galaxies. In Figure 1 one blue feature that lo oks

ELLIPTICAL GALAXY FORMATION 191

galaxies and globular clusters are photometric color like an anchor in the south-east of M49 is UGC 7636,

and/or sp ectral features. These two pieces of informa- a dwarf irregular galaxy interacting with M49. Several

tion are based on the integrated light from the various young clusters are found in this region and the color of

kinds of stellar p opulations. Sp ectra contain more in- this galaxy is blue, showing that clusters form during

formation than color, but colors are easier to get than the interaction of galaxies (Lee, Kim, & Geisler 1997).

sp ectra. So the information we have now is color for

There are numerous studies and reviews on the for-

most ob jects and sp ectra for a small numb er.

mation of elliptical galaxies and globular clusters (e.g.,

Burkert 1994, Renzini 1999, Whitmore 2000, Silk & A well-known problem in using integrated color is

Bouwens 2001, Ellis 2001, Harris 2001,2003, Gnedin, that it is dicult to distinguish b etween the metal-

Lahav, & Rees 2001, Gnedin 2002, Peebles 2002, Mat- p o or old systems and the metal-rich intermediate-age

ote 2002, teucci 2002, Steinmetz 2002, Conselice 2002, C^ systems, b ecause age and metallicity have similar ef-

Burkert & Naab 2003). A recent conference pro ceed- fects to color (and age and metallicity are often cor-

ing related to this topic is in Geisler et al. (2002). related). Increasing age and/or metallicity makes the

Even reviewing of reviews is not easy. Here I review color redder. Mo dern history in the studies of elliptical

the current status of understanding the formation of galaxies and globular clusters is full of e orts to break

elliptical galaxies (mostly giant elliptical galaxies with this age-metallicity degeneracy. Sp ectral information is

M > 10 M ), fo cusing on globular clusters. very imp ortant in breaking this age-metallicity degen-

eracy, but not yet enough to do it.

This pap er is comp osed as follows. Section II p oints

out key elements for solving the key question of galaxy So the problem of knowing `when' and `how long' is

formation, and Section III describ es brie y two proto- observationally equivalent to the problem of breaking

mo dels of formation of elliptical galaxies. Observa- the age-metallicity degeneracy. It is easy in principle,

tional constraints from stars in elliptical galaxies are but dicult in reality. However, we are getting closer

listed in Section IV, and observational constraints from to the goal.

globular clusters in elliptical galaxies are given in Sec-

One of the nagging factors in these studies is large

tion V. Section VI describ es brie y mo dels of formation

uncertainties of the parameters used for conversion b e-

of globular clusters. The nal section concludes with

tween the lo okback time and redshift. Astronomers

the present status, and presents a sketch for the forma-

working with age-dating of stellar systems almost al-

tion of giant elliptical galaxies.

ways used time, while astronomers working with dis-

tant ob jects always used redshift z . So it is not easy to

I I. KEY ELEMENTS FOR UNSOLVED MYS-

compare directly results given in terms of time and z .

TERY

The lo okback time is given as a function of redshift,z ,

as follows (Hogg 1999):

Primary goals of studying the formation of struc-

tures (dust, planets, solar systems, stars, star clusters,

9:78

galaxies, galaxy clusters, etc) are to nd when, how

(z ) = t

long, and how they formed. These goals are not easy

to achieve even in the case of stars (see Elmergreen

to gure out p os- 2002). In general it is often easy(?)

;

0 0 3 0 2

(1 + z ) (1 + z ) + (1 + z ) +

(formation sible ways to form structures, and `when'

M k

ep o ch) and `how long' (formation duration) are needed

where h is a normalized Hubble constant, H 100 =

to constrain the p ossible `how's (formation pro cesses).

1 1

s Mp c ], and , , and are, resp ec- [km

M k

However, `when' and `how long' are often dicult to

tively, density parameters for matter(including dark

know accurately and precisely, esp ecially when the tar-

matter and baryonic matter), dark energy, and cur-

gets are old and far away, like elliptical galaxies. With

vature.

the exception of NGC 5128, at the distance of ab out

Now several indep endent sources provide converging

4 Mp c, all of the giant elliptical galaxies are lo cated

values for these parameters so the direct comparison b e-

b eyond 10 Mp c. Most of the globular clusters in the

tween the time and z is p ossible. Sp ergel et al. (2003)

universe may b elong to these giant elliptical galaxies,

provide precise estimates of cosmological parameters

and they must contain critical information on the for-

combining WMAP results on microwave cosmic back-

mation of elliptical galaxies (although they o ccupy only

ground radiation, Sup ernovae Ia, and large-scale galaxy

a tiny fraction of mass in galaxies).

:04 +0

h = 0:0224 0:0009, redshifts: h = 0:71 ,

0:03

Figure 2 displays a schematic diagram showing the

+0:008

h = 0:135 , = 1:02 0:02, the equa-

M total

0:009 key elements in the study of formation of elliptical

tion of state of the dark energy w < 0:78, and the

galaxies and globular clusters. Galaxies form and

age of the universe, 13:7 0:2 Gyr. Tonry et al. (2003)

evolve chemically (photometrically and dynamically as

presented their estimates of some of these parameters

well), and we need to know the star formation history

based on high-z Sup ernovae Ia results, which are con-

and chemical enrichment history of these galaxies.

sistent with Sp ergel et al. (2003)'s. Sp ergel et al.'s

Two ma jor observational to ols obtained for distant

age estimate 13:7 0:2 Gyr is only slightly larger than

192 LEE

Fig. 2.| A schematic diagram showing the key elements in the study of formation of elliptical galaxies and globular

clusters. Photometric colors and sp ectral features in sp ectra are observational to ols to gure out star formation history and

chemical evolution in elliptical galaxies and globular clusters. Integrated colors and sp ectra are obtained from the mixed

p opulations in these stellar systems, and are dep endent on b oth age and metallicity. A core of the problem is in breaking

the age-metallicity degeneracy in colors and sp ectra to understand when and how long these ob jects formed.

the mo dels of elliptical galaxies followed the mo dels the mean age of the 17 metal-p o or globular clusters

of the halo of our Galaxy, a spiral galaxy. There are con - in the halo of our Galaxy, 12.6 Gyr with a 95%

two main comp eting mo dels to explain the formation dence level lower limit of 11.2 Gyr (Krauss & Chab oyer

of elliptical galaxies: monolithic collapse mo del and hi- 2003). According to the WMAP results, the reioniza-

= 17 3, or 11 < z < 30 (cor- erarchical merging mo del. tion ep o ch may b e z

r r

resp onding to times 100 < t < 400 Myr after the Big

Monolithic Collapse Mo dels (a)

Bang) ( Kogut et al. 2003). We adopt a current stan-

dard mo del of cosmology, concordance mo del to calcu-

The concept of monolithic collapse mo del (MCM)

late the lo okback time: = 0:3, = 0:7, = 0:0,

M k

was intro duced rst by Eggen, Lynden-Bell, & Sandage

and h = 0:71.

(1962) who tried to explain the radial orbits of the stars

in the halo of our Galaxy. This mo del was expanded

I I I. MODELS OF FORMATION OF ELLIP-

to the case of elliptical galaxies later (Partridge & Pee-

TICAL GALAXIES

bles 1967, Tinsley 1972, Larson 1974, Chiosi & Carraro

2002, Matteucci 2002, Kawata & Gibson 2003). In this

The formation of elliptical galaxies may b e closely

mo del, elliptical galaxies (and the spheroidal comp o-

related with the formation of bulges and halos in spiral

nents of disk galaxies as well) formed as a result of

galaxies. So the mo dels for the elliptical galaxies are

large protogalactic clouds collapsing and forming stars

also related with the mo dels for the bulges and halos in

7 8

in a very short time of ab out 10 10 yr very early

spiral galaxies. One motivates the other. Historically

ELLIPTICAL GALAXY FORMATION 193

(at z > 3). After the violent burst of star formation (b) Color-Magnitude Relation

leading to high metallicity very early, these galaxies

The brighter elliptical galaxies are, the redder they

evolved quiescently until to day (passive evolution).

are. It is called a color-magnitude relation (CMR) of

elliptical galaxies (and S0 galaxies). It has b een known

Hierarchical Merging Mo dels (b)

for long (Baum 1959). Baum (1959) p ointed out early

To omre (1977) prop osed an alternative mo del where a transition in color from Galactic globular clusters,

elliptical galaxies formed when two spiral galaxies mer-

through dwarf elliptical, to giant elliptical galaxies (see

ged, and Searle & Zinn (1978) suggested that the halo his Figure 5). Figure 3 shows the color-magnitude re-

of our Galaxy might have formed via accreting tran- lation of early typ e galaxies in Virgo and Coma given

sient proto-Galactic fragments with dwarf galaxy-mass

by Bower, Lucey, & Ellis (1992). The scatters in color

for an extended p erio d of time. These ideas were de- of elliptical galaxies (and S0's) in Virgo and Coma are

velop ed into hierarchical merging (or clustering) mo d- V ) 0:04). This found to b e remarkably small ((U

els (HMM) in the context of cold dark matter (CDM)

scatter is related with age disp ersion, d(U V )=dt

mo del of structure formation where small ob jects are 0:02 mag/Gyr for old age according to the evolutionary

progressively incorp orated into larger structures, and p opulation synthesis mo dels. This small scatter indi-

stars form at the baryonic cores emb edded in the dark

cates that the formation duration of these galaxies is

matter halo (Kau mann, White, & Guiderdoni 1993, shorter than 2 Gyr and the formation ep o ch is very

Benson, Ellis, & Menanteau 2002, Benson & Madau early at z > 3 (Bower, Lucey, & Ellis 1992, Terlevich,

2003, Steinmetz & Navarro 2003, Helly et al. 2003,

Caldwell, & Bower 2001, Peebles 2002).

Kho chfar & Burkert 2003, Meza et al. 2003). In this

The color-magnitude relations of early typ e galaxies

mo del, massive elliptical galaxies formed much later

in more distant clusters (up to z < 1:27) and in the

and much longer than those monolithic collapse mo d-

Hubble Deep Field (Williams et al. 1996) are consis-

els predict. Merging pro cess can b e roughly divided in

tent with those for z = 0, indicating the primary origin

two: ma jor merger (merging of large galaxies) and mi-

of the color-magnitude relation is metallicity, not age

nor merger (accretion/tidal stripping of low-mass sys-

(Ko dama & Arimoto 1997, Ko dama et al. 1998, Bower,

tems like dwarf galaxies, globular clusters, and dwarf

Ko dama, & Terlevich 1998, Ko dama, Bower, & Bell

protogalactic clouds). These mo dels predict that ab out

1999). This is strong evidence supp orting the MCM.

half of the massive elliptical galaxies formed at z < 1

The HMM also predicts this relation, but with an as-

(Peebles 2002).

sumption that elliptical galaxies merge other elliptical

galaxies of the same luminosity to keep the CMR.

IV. OBSERVATIONAL CONSTRAINTS

FROM STARS

Stellar systems can b e characterized by three global

Elliptical galaxies show two kinds of integrated prop-

physical parameters: central velo city disp ersion, e ec-

erties of stars: one that almost all elliptical galaxies

tive radius, and e ective surface brightness. Funda-

share (uniformity and family) and the other that only

mental planes of stellar systems are mathematical rela-

some elliptical galaxies show (p eculiarity). The for-

tions b etween these parameters. Elliptical galaxies in

mer includes the color-magnitude relation, fundamen-

clusters are known to o ccupy very narrow regions in the

tal planes, color gradients, and surface brightness pro-

fundmental planes (Kormendy 1985, Bender, Burstein,

les, and the latter includes morphological and kine-

& Fab er 1992, Burstein et al. 1997). The basic physics

matical p eculiar features (ripples, shells, and kinemati-

leading to the fundamental planes is b elieved to b e

cally decoupled cores). These are used as constraints of

virial theorem and galaxy evolution history.

mo dels of formation of elliptical galaxies (e.g., Bernardi

et al. 2003a,b).

-space Figure 4 displays fundamental planes in the

of elliptical galaxies, dwarf elliptical galaxies (dE), Lo-

and Rotation (a) Structure

cal Group dwarf spheroidal galaxies (dSph), and Galac-

tic globular clusters, using the data given by Burstein

The surface brightness pro les of elliptical galaxies

et al. (1997).

(and the bulges of disk galaxies) follow generally a de

The -space parameters are de ned as follows: =

1=4

p p

law (except for disky elliptical galax- Vaucouleurs r

2 2

2, 6, = (log + 2 log I (log + log r )= log r )=

2 e e e

c c

ies which follow exp onential law). Elliptical galaxies

3 (Burstein et al. and = (log log I log r )=

3 e e

are slow rotators, and are pressure-supp orted (dynam-

1997).

ically hot), while spiral galaxy disks are fast rotators,

2 1

and rotation-supp orted (dynamically cold). Giant el-

is the central velo city disp ersion in km s ,

liptical galaxies are attened by an anisotropic velo city

r is e ective radius in kp c, and I is the B -band

e e

disp ersion, while faint elliptical galaxies (and bulges in

surface brightness at r in solar luminosity p c (=

S B 27:0) 0:4(

spiral galaxies) are in general isotropic and rotationally

), where S B is the mean B -band sur- 10

attened (Bender, Burstein, & Fab er 1992).

face brightness in B mag arcsec within r . , ,

e 1 2

and are logarithmically related with galaxy mass 3

194 LEE

Fig. 3.| (a), (b) and (c) Color-magnitude relations of early typ e galaxies in Virgo (op en symb ols) and Coma ( lled

wer et al. (1992). Circles, triangles and star symb ols represent, resp ectively, Es, S0s, and symb ols) based on the data of Bo

galaxies with typ es later than S0. The solid lines represent the linear ts to the data of Es excluding a few outliers at the

faint end, and the dashed lines the ts for all galaxies given by Bower et al. (1992). (d) Log vs. total magnitude relation.

The scatters in color of Es are mostly due to observational errors, and the intrinsic disp ersions of color are remarkably small

0:04 mag). (source: Bower et al. 1992) (<

(M), surface brightness times mass-to-luminosity ratio, globular clusters are lo cated in a di erent region in the

and mass-to-luminosity ratio (M/L). space, but follow a similar trend, with o set in

1 2

, with E's in the space.

3 1 3

In Figure 4, the fundamental plane in space

1 3

de ned by the gEs in Virgo and Coma is given by

(d) Color Gradients

the solid line ( = 0:15 + 0:56, log(M =L ) =

3 1 e e

0:184 log M 1:25). Figure 4 shows that elliptical

Many bright elliptical galaxies show radial gradi-

galaxies show very tight fundamental planes in the

ents of color in the sense that the color gets bluer as

edge-on views, indicating these elliptical galaxies in var-

the galacto centric radius increases, with mean values

ious environments share a common origin (Renzini &

of (U R )= log r = 0:20 mag/dex and (B

Ciotti 1993). The scatters in the fundamental planes

R )= log r = 0:09 mag/dex (Peletier et al. 1990,

of elliptical galaxies in clusters at z < 0:6 lead to an es-

Kim, Lee, & Geisler 2000). This color gradient may

timate for formation ep o ch z > 2 (Peebles 2002). This

b e due to metallicity or age variation. Tamura et al.

is signi cant evidence supp orting the MCM. Galactic

ELLIPTICAL GALAXY FORMATION 195

Fig. 4.| Fundamental planes of elliptical galaxies, dwarf galaxies, and Galactic globular clusters (op en circles: isotropic

Es, lled circles: anisotropic Es, op en squares: Es without isotropy information, small red dots: GCs in our Galaxy, small

blue dots: dwarf spheroidal galaxies around our Galaxy). The solid line represent the fundamental planes for gEs in Virgo.

Note the tight fundamental planes of elliptical galaxies. (source: Burstein et al. 1997)

avalisco 2002, Blain et al. 2002, Cimatti et al. 2002a,b, (2000) investigated the origin of color gradients in ellip-

Cimatti 2003, Chapman et al. 2003). This shows that tical galaxies by comparing the mo dels with the sample

massive galaxies formed at z > 2, supp orting the MCM of seven red elliptical galaxies at z = 0:1 1:0 in the

(see also Sarasco et al. 2002, Zirm, Dickinson, & Dey Hubble Deep Field. They concluded that the color gra-

2003, Lo eb & Peebles 2003). There are several pieces dients in elliptical galaxies are due to metallicity e ect,

of evidence indicating a p ossibility that quasars might not due to age e ect, supp orting the MCM.

have formed together with the stellar p opulations of

Galaxies (e) High-redshift

early-typ e galaxies (Cattaneo & Bernardi 2003). If so,

quasars might have formed after the metal-p o or glob-

Several kinds of massive ob jects which can b e pro-

ular clusters.

genitors of massive elliptical galaxies are found at high

redshift z > 2: Lyman break galaxies, extremely red

ob jects (ERO), submillimeter galaxies and quasars (Gi-

196 LEE

Recent photometric studies of globular clusters in (f ) Morphological and Kinematical Peculiarity

elliptical galaxies can b e roughly divided into three

Many elliptical galaxies show p eculiar and remark-

classes: a) V I photometry of HST WPFC2 for a small

able features in their morphology and kinematics: shells,

eld often covering the centers of the galaxies (e.g., Lee,

ripples, dust, jets, and kinematically decoupled cores.

& Kim 2000, Kundu & Whitmore 2001a,b, Larsen et al.

These are considered as remnants of mergers (see Barnes

2001), b) wide- eld multi-color photometry obtained

1998), supp orting the HMM.

at ground-based telescop es (e.g., M87: Lee & Geisler

1993, Lee et al. 2003; M49: Geisler, Lee, & Kim 1996,

(g) Mixed Populations in the Nearest gE NGC

Lee, Kim, & Geisler 1998, Rho de & Zepf 2001; NGC

5128

1399: Dirsch et al. 2003), and c) near IR photometry

at the large ground-based telescop es to derive metal-

NGC 5128 (Cen A) is the nearest giant elliptical

licity and age of globular clusters (e.g., Kissler-Patig,

galaxy (at d 4 Mp c). It is an intermediate-age merger

Bro die, & Minniti 2002, Puzia et al. 2002).

remnant (Schweizer 2002a), and has a blue tidal stream

of young stars including a numerous young bright stars

Recent sp ectroscopic studies are trying to get the

(with an age of 350 Myr) in the outer area (Peng et metallicity, kinematics and age of globular clusters

al. 2002).

(e.g., Hanes et al. 2001, Cohen, Blakeslee, & C^ote

2003, Richtler et al. 2002, Strader et al. 2003). An

In the pioneering work of resolved stars in giant el-

excellent review of globular clusters in M87 and M49

liptical galaxies, Harris & Harris (2002, and early ref-

in the Virgo cluster was given by C^ote (2002).

erences therein) presented color-magnitude diagrams of

giant stars in the halo elds of NGC 5128 (at 8 kp c, 21

(a) Color-Magnitude Diagram

kp c, and 31 kp c from the center of the galaxy center).

These data show remarkable features: 1) The metal-

Figure 5 displays T (C T ) color-magnitude di-

1 1

licity (derived from (V I ) colors) distribution of the

agrams of p oint sources in six giant elliptical galaxies

stars is asymmetric and very broad, from [Fe/H] 2:2

in the Virgo cluster (Geisler, Lee, & Kim 1996, Lee

dex to +0.4 dex; 2) Most of the stars are metal-rich,

et al. 2003). The size of the eld for each galaxy is

with a p eak at [Fe/H] 0:4 dex. The mean metallic-

0 0

16 16 . Most of the ob jects inside the blue b oxes in

ity of these stars is only slightly smaller than that of

Figure 5 are probably genuine globular clusters. Most

the metal-rich globular clusters in the same galaxy; 3)

of the blue faint ob jects are background galaxies, and

A small fraction of stars are metal-p o or. Bekki, Harris,

bright ob jects outside the b oxes are mostly foreground

& Harris (2003) and Beasley et al. (2003) used merger

stars.

mo dels (merging of two spiral galaxies) to explain the

Figure 5 shows immediately several features of glob-

metallicity distribution of NGC 5128.

ular clusters in these galaxies: 1) there are two vertical

plumes inside the b oxes in most of these galaxies, which

(h) Summary

are, resp ectively, blue globular clusters (BGC) and red

There are several constraints based on observations

globular clusters (RGC); 2) there is a large variation in

of stars in elliptical galaxies, showing that massive el- the shap e of the two plumes among the galaxies. A re-

liptical galaxies might have formed very early at z > 2,

markable contrast is shown by M84 and NGC4636; and

supp orting the MCM. At the same time there are

3) there is a large variation in the numb er of globular

various observational constraints showing there were clusters among the galaxies.

merger events in elliptical galaxies, and the numeri-

cal simulations based on the HMM can repro duce im- (b) Color Distribution

pressively various features of large-scale structures from

Figure 6 illustrates (C T ) color distribution of

dwarf galaxies to giant galaxies and rich clusters of

bright globular cluster candidates with 19 < T < 23

galaxies (see, e.g., the gures in Steinmetz 2002, Stein-

mag in the same galaxies as shown in Figure 5. Ob jects

metz & Navarro 2003, Kravtsov & Gnedin 2003). One

with 1:0 < (C T ) < 2:1 are mostly genuine globular

critical problem in the current HMM based on the

clusters. Red, blue and black lines represent, resp ec-

CDM mo dels is that the HMM cannot make many

tively, the ob jects in the inner region (R < 4 :5), in the

massive clusters very early at z > 2. Another well-

outer region (R > 4 :5), and in the entire region, where

known problem is that the HMM predicts to o many

R is a pro jected radial distance from the center of each

dwarf galaxies in the Lo cal Group, by an order of mag-

galaxy. Metallicity scale based on (C T ) color is also

nitude or more than the observed (Mo ore et al. 1999,

shown ab ove the top panel.

C^ote, West, & Marzke 2002).

All six galaxies show clearly bimo dal color distri-

bution with p eaks at (C T ) 1:3 and 1:75.

V. OBSERVATIONAL CONSTRAINTS

Interestingly, the mean metallicity di erence b etween

FROM GLOBULAR CLUSTERS

the RGCs and BGCs in all six galaxies is found to

b e [Fe/H] 1:0 dex, showing the RGCs are 20 times

Critical information of globular clusters in elliptical

more metal-rich than the BGCs in all galaxies. Why

galaxies are derived from photometry and sp ectroscopy.

ELLIPTICAL GALAXY FORMATION 197

Fig. 5.| Color-magnitude diagrams of p oint sources in six giant elliptical galaxies in Virgo. The eld of view of the images

0 0

for each galaxy is 16 16 . The order of the panels is according to the total luminosity from the top: M49, M87, M60,

M86, M84, and NGC 4636. Most of the ob jects inside the blue b oxes are probably bright globular clusters. (source: Lee et

al. 2003)

it is signi cantly elongated (the ellipticity at r (= 120 is the di erence in metallicity (all ab out 20 times) so

arcsec = 10.1 kp c)=0.17), Kim, Lee, & Geisler 2000). similar? Below I use the BGCs for metal-p o or globular

M87 has three times more globular clusters than M49, clusters and the RGCs for metal-rich globular clusters.

but its ellipticity is almost zero (E0). To date M49 is

However, the ratio of the total numb er b etween the

the only giant elliptical galaxy, for which the details

BGCs and RGCs varies signi cantly among the galax-

of the spatial structure including the ellipticity of the

ies. Note the dramatic contrast b etween M84 and NGC

globular cluster systems were studied. Similar stud-

4636: the BGC is much stronger than the RGC in M84,

ies of other giant elliptical galaxies are b eing done now

while b oth are comparable in NGC 4636.

(Lee et al. 2003).

In addition, the ratio of the numb er b etween the

Figure 7 displays spatial distribution of globular

BGCs and RGCs changes signi cantly dep ending on

clusters in M49 in comparison with the stellar halo of

the galacto centric distance: relatively more RGCs are

M49 based on the ground-based Washington photom-

seen in the inner region than in the outer region.

etry (Lee, Kim, & Geisler 1998). Figure 7 shows a

T image of M49 (a) and the surface numb er density

maps for the entire GCs (b), the BGCs (c), and the

M49 is an ideal target to compare the spatial dis-

RGCs (d). Note the spatial structure of the RGC sys-

tribution of globular clusters and stellar halo, b ecause

tem is elongated very similarly to that of the stellar

198 LEE

Fig. 6.| (C T ) color distribution of bright globular cluster candidates with 19 < T < 23 mag in the giant elliptical

1 1

galaxies in Virgo. Ob jects with 1:0 < (C T ) < 2:1 are mostly genuine globular clusters. Red, blue and black lines

0 0

< 4 :5), in the outer region (R > 4 :5), and in the entire region, represent, resp ectively, the ob jects in the inner region (R

where R is a radial distance from the center of each galaxy. Metallicity scale based on (C T ) color is also shown on the

top panel. Note the variation in the bimo dality among the six galaxies that are all in the same Virgo cluster. (source: Lee

et al. 2003)

halo, while that of the BGCs is almost circular. The This shows the spatial structure of the RGC system

RGCs are more centrally concentrated than the BGCs, is very consistent with that of the stellar halo, while

which is also seen in Figure 8. that of the BGC system is di erent from b oth. So it

is the RGCs, not the BGCs, that go with halo stars in

Figure 8 plots the radial variation of the ratio of

M49.

the numb er of the BGCs to that to the RGCs. The

lled circles represent the data for globular clusters

(d) Color Gradients

with V < 23:9 mag based on the HST WFPC2 V I

photometry, and the op en circles the data for globular

Figures 9 and 10 display color gradients of the glob-

clusters with T < 23 mag based on the ground-based

ular clusters in M49 in comparison with the stellar halo

Washington photometry (Lee & Kim 2000). Figure

(Lee, Kim, & Geisler 1998, Lee & Kim 2000, Kim, Lee,

8 shows that the N(BGC)/N(RGC) increases as the

& Geisler 2000). Striking features seen in these g-

galacto centric radius increases. There are ab out equal

ures are: 1) all GCs show a strong radial color gradient

numb er of BGCs and RGCs in the inner region, but

(much stronger than the stellar halo); 2) the color pro-

the outer region is dominated by the BGCs.

le of the RGCs is almost the same as that of the stellar

ELLIPTICAL GALAXY FORMATION 199

is considered to b e due to dynamical evolution on the

low-mass clusters, which converts the p ower-law lumi-

nosity functions into a Gaussian form or two comp onent

p ower-law form (Fall & Zhang 2001, Smith & Burkert

2002). The bright part of the luminosity function is less

a ected than the faint part by the dynamical evolution

of clusters.

One thorny problem in the luminosity function is

the di erence in the p eak magnitude b etween the BGC

and RGC. The p eak magnitude of the globular clus-

ter luminosity function may dep end on metallicity, age,

and dynamical evolution. In the case of M49, Puzia et

al. (1999) found, using the same HST WFPC2 V I

images for three elds in M49 as used by Lee & Kim

(2000), that the p eak magnitude of the RGC is fainter

B GC ) = 0:51 than that of the BGC, by V (R GC

mag and I = 0:41 mag, and concluded, from this,

that the BGCs and the RGCs are co eval within the

errors of 3 Gyr. Later Larsen et al. (2001), us-

ing the same HST data, concluded that the p eak mag-

nitude of the RGC is fainter than that of the BGC,

by V = (24:21 0:23) (23:38 0:16) = 0:83

mag, and the p eak magnitude for the entire sample

+0:14

is V = 23:78 , which is 0.38 mag fainter than the

0:15

value given by Lee & Kim (2000).

Fig. 7.| Gray-scale map of a T CCD images of M49

On the other hand, Lee & Kim (2000) found, using

(a), and the surface numb er density maps for the globular

the HST WFPC2 V I images for four elds, that the

clusters with T < 23 mag for the entire GCs (b), the BGCs

V -band p eak magnitudes of the RGC and the BGC

(c), and the RGCs (d). Note the spatial distribution of the

RGCs is similar to that of the stellar halo, while that of

the BGCs is almost spherical. (source: Lee, Kim, & Geisler

1998)

halo (note, however, that the numb er density pro le of

the RGCs is much more extended than that of the stel-

lar halo, as seen in Figure 9(a)); and 3) the BGCs and

RGCs show, resp ectively, little radial color gradients.

The strong color gradient of all GCs is primarily due

to the stronger central concentration of the RGCs com-

pared with the BGCs. This evidence shows again that

the RGCs and the halo stars in M49 are of the same

origin.

(e) Luminosity Function

Luminosity functions of old globular clusters in ellip-

tical galaxies are approximately t by a Gaussian with

a p eak at M = 7:4 mag and a width 1:2 (or

function), which is often used as distance in- by a t

dicator. Figure 11 illustrates the luminosity functions

of globular clusters in M49 based on the HST WFPC2

V I photometry, showing a p eak at V = 23:50 0:16

and I = 22:40 0:14, leading to a distance estimate

of d = 14:7 1:3 Mp c (Lee& Kim 2000). The bright

Fig. 8.| Radial variation of the numb er of the BGCs to

part of the luminosity function of globular clusters in

that of the RGCs with V < 23:9 mag in M49 derived from

elliptical galaxies is approximately t by a p ower law,

S T WFPC2 VI data ( lled circles). The op en circles the H

and the luminosity functions of bright blue clusters in

represent the data for the globular clusters with T < 23

mag in the outer region of M49 derived from the Washington

the merging galaxies are also well t by a p ower law.

photometry (source: Lee & Kim 2000)

The di erence in the faint part b etween these two kinds

200 LEE

16 (a) 2 (b) Halo 18 GCs : -2.5 Logσ + 22 GC 1.8 GCs(Red) ) 20 1

µ 1.6

(C-T GCs(Total) Halo (T1) 22 1.4

GCs(Blue) 24 1.2

0 100 200 300 400 500 0 100 200 300 400 500

Reff [arcsec] Reff [arcsec]

.3 ] o 140 P.A. [ P.A.

Ellipticity .2 120 .1 100

0 0 100 200 300 400 500 0 100 200 300 400 500

Rmajor [arcsec] Rmajor [arcsec]

Fig. 9.| Radial variation of (a) the surface numb er density, (b) color, (c) ellipticity, and (d) the p osition angle of the

globular clusters in M49. The lled circles, op en circles, and crosses represent, resp ectively, the entire GCs, the BGCs

and the RGCs. The small squares with error bars represent the stellar halo of M49. The three lines in (a) represent after

background subtraction (the solid line for the entire GCs, the dotted line for the BGCs and the dashed line for the RGCs).

Note the color of the RGCs is remarkably similar to that of the stellar halo. (source: Lee, Kim, & Geisler 1998)

Geisler (1998) and the similar result for M87 by Kundu are similar within the errors: V (B GC ) = 23:53 0:16,

et al. (1999). This result indicates that the RGCs may V (R GC ) = 23:44 0:22, I (B GC ) = 22:63 0:17 and

b e several Gyr younger than the BGCs, if the same I (R GC ) = 22:30 0:26 mag, as shown in Figure 11. In

iso chrones as used by Puzia et al. (1999) are used. this case, the I -band magnitude di erence, I = 0:33

is consistent with the (V I ) di erence b etween the

It is puzzling that the two groups's results based on

RGC and BGC, (V I ) = 1:233 0:975 = 0:26 (i.e.,

the basically same data gave signi cantly di erent es-

the I -band p eak magnitude of the RGC is brighter than

timates for the p eak magnitudes of the BGC and RGC

that of the BGC, if the V -band p eak magnitudes of

in M49. The reason for this di erence app ears to b e,

the BGC and RGC are the same.). This result is also

in part, due to the di erence in the metho d of pho-

consistent with previous results based on the ground-

tometry. Comparison of the luminosity functions for

based Washington photometry given by Lee, Kim, &

ELLIPTICAL GALAXY FORMATION 201

the fact that the BGCs are relatively more abundant in

the outer region than the RGCs (Lee, Kim, & Geisler

1998).

(g) Kinematics

Figure 12 displays colors and radial velo cities of

globular cluster candidates in M87: a) the color distri-

bution of globular cluster candidates in M87 with 334

measured radial velo city (black histogram) of C^ote et

al. (2001) among the photometric sample (shaded his-

togram), b) the radial velo city distribution of globular

cluster candidates, and c) (C T ) vs. radial velo cities

of the globular clusters in M87. Figure 12 shows 1) 278

ob jects inside the dashed b ox (with 0:8 < (C T ) <

2:35) are genuine globular clusters b elonging to M87

(or at least to the Virgo cluster), 2) most ob jects with

colors (C T ) > 2:35 or (C T ) < 0:8 are fore-

1 1

ground stars, and 3) the fraction of ob jects with colors

0:8 < (C T ) < 2:35 is negligible compared with

the globular clusters with the same color range. This

Fig. 10.| Radial variation of the mean (closed symb ols)

con rms that most of the photometric globular cluster

and median (op en symb ols) color of the bright globular clus-

candidates based on the Washington photometry are

ters with V < 23:9 mag in M49. The circles, triangles, and

genuine globular clusters, showing the high eciency of

squares represent, resp ectively, the total sample, the BGCs

the Washington photometry in selecting globular clus-

and the RGCs. The solid line represents the mean color of

ter candidates.

the stellar halo of M49 given by Kim, Lee, & Geisler (2000).

Note the color of the RGCs is remarkably similar to that of

There are only three giant elliptical galaxies for

the stellar halo. (source: Lee & Kim 2000)

which the kinematics of globular clusters were stud-

ied: M87, M49 and NGC 1399 (Cohen & Ryzhov 1997,

the same elds given by two groups shows that Lee &

Kissler-Patig & Gebhardt 1998, Kissler-Patig et al.

Kim (2000)'s photometry is more complete in the faint

1998, Minniti et al. 1998, Zepf et al. 2000, C^ote et

end than Puzia et al.(1999)'s even b efore incomplete-

al. 2001, 2003). C^ote (2002) gave a nice summary of

ness correction (see Figure 9 in Lee & Kim 2000). This

kinematics of globular clusters in M49 and M87. Table

indicates that the careful analysis is needed for inves-

1 is a slightly up dated version of C^ote (2002)'s compila-

tigating the di erence in the p eak magnitude b etween

tion, including NGC 1399. Table 1 lists the pro jected

the BGCs and RGCs in distant galaxies.

>, the rotational velo city, velo city disp ersion, <

< R >, and the anisotropy parameter < R = >.

(f ) Sp eci c Frequency

What is surprising is that all three galaxies showed

di erent kinematics of globular clusters: 1) M87 GCs

The globular cluster sp eci c frequency is de ned as

(in total) show a larger velo city disp ersion than M49

the total numb er of globular clusters p er unit host

GCs, although M87 is somewhat fainter than M49.

0:4(M +15)

galaxy luminosity: S = N (GC )10 (Harris

This must b e related with the fact that M87 is lo cated

& van den Bergh 1981). S for dwarf and giant ellipti-

right at the center of the Virgo cluster, while M47 is

cal galaxies in rich clusters (S 4 6) is higher than

far from the center of the Virgo; 2) The BGCs have a

S for such galaxies in lo ose groups (S 2 3), and

N N

larger velo city disp ersion than the RGCs in M49, but

S for spiral galaxies is as low as ab out 1 (Harris 1991).

NGC 1399 GCs show an opp osite result. On the other

Several brightest cluster galaxies and cD galaxies show

hand, M87 GCs show a similar velo city disp ersion b e-

very high S = 10 20 (see Figure 8 in Beasley et al.

tween the BGC and the RGC; 3) The M87 GCs rotate

2002). The higher S for ellipticals compared with S

N N

faster than the M49 GCs, and there is little di erence

for spiral galaxies motivated a merger mo del (Ashman

in the rotation velo city b etween the BGC and RGCs

& Zepf 1992).

in M87. In contrast, there is signi cant di erence in

Recently C^ote et al. (2000) p ointed out that, if the

the rotation velo city b etween the BGC and RGCs in

bulge luminosity is used instead of total galaxy lumi-

M49. The BGCs in M49 rotate with a velo city of 93

nosity, S for 11 spiral galaxies (Sa to Sc) in the sample

km s , while the RGCs in M49 show little rotation.

of Kissler-Patig et al. (1999) b ecomes S = 3:8 2:9,

The kinematics of the M49 GCs is consistent with the

which is comparable to that for elliptical galaxies. The

gaseous merger mo del (Ashman & Zepf 1992), but that

high S (> 5) for giant elliptical galaxies in rich clus-

of the M87 GCs is not. 4) The BGCs in M87, M49, and

ters are primarily due to the BGCs, rather than due to

our Galaxy show similar anisotropy (< R = >), but

RGCs (Forb es et al. 1997). Also the lo cal S increases

smaller than that of the BGCs in M31. However, the

as the galacto centric distance increases, which is due to

202 LEE

Fig. 11.| V -band and I -band luminosity functions of the globular clusters in M49 derived from HST WFPC2 images of

four elds (histograms) t with the Gaussian functions (the curved lines). (source: Lee & Kim 2000)

and the RGC is exp ected to b e smaller than 3 Gyr (see RGCs in M87 and M49 show much smaller anisotropy

the nal section). to the RGCs of in our Galaxy and M31. One notable

feature is that the BGCs in M31 is rotating almost as

The sources of uncertainties are b oth observational

fast as the RGCs. In summary, any uniform feature is

data and the mo del iso chrones (Thomas, Maraston,

not yet emerging from the kinematics of giant elliptical

& Bender 2003, Thomas & Maraston 2003). Cohen,

galaxies.

Blakeslee, & C^ote (2003) derived, from the comparison

of their sp ectra for 47 globular clusters in M49 with

(h) Age

the mo dels of Worthey (1994) and Thomas, Maraston,

& Bender (2003), ages which are signi cantly di erent

Age determination of globular clusters in giant el-

b etween the two mo dels, as listed in Table 2. They

liptical galaxies is extremely dicult, and currest age

wever, the only safe statement in our p ointed out \Ho

estimates include a large uncertainty. There are only

view is that M49 GCs are in the mean older than 10

a few giant elliptical galaxies for which the ages of the

Gyr". Puzia et al (2002) also found a large di erence

globular clusters were estimated. Table 2 is an up-

in age estimates dep ending on iso chrone mo dels, when

dated version of C^ote (2002)'s compilation of the age

they estimates the age of globular clusters in NGC

estimates of M87 and M49, by including other refer-

) diagrams (note that this 3115 using (V I ){(V K

ences. Table 2 shows 1) the globular clusters in these

metho d was used also for estimates of age and metal-

galaxies are older than 10 Gyr, and 2) the errors of age

licity of early-typ e galaxies in galaxy clusters (Smail et

estimates of the globular clusters are large. The BGCs

al. 2001)). Better and more age estimates are needed.

and RGCs may b e said to b e co eval within the error of

3-7 Gyr. However, this uncertainty is to o large to dis-

tinguish di erent mo dels for the formation of globular

clusters, b ecause the age di erence b etween the BGC

ELLIPTICAL GALAXY FORMATION 203

Table 1. Kinematics of Globular Clusters in gEs

Galaxy GC < > < R > < R = > N(GC) Ref.

p p

+31 +39 +0:09

M87 All 383 171 0:45 278 1

7 30 0:09

+36 +58 +0:13

BGC 397 186 0:47 1

14 41 0:11

+38 +53 +0:14

RGC 365 155 0:43 1

18 37 0:12

+27 +52 +0:15

M49 All 316 48 0:15 263 2

8 26 0:08

+33 +69 +0:19

BGC 342 93 0:27 2

18 37 0:11

+0:27 +64 +34

2 0:10 26 RGC 265

0:25 79 13

NGC 1399 All 304 11 470 3,4

BGC 297 16 3,4

RGC 355 22 3,4

MWG BGC 0.32 5

RGC 1.05 5

M31 BGC 0.85 6

RGC 1.10 6

References: (1) C^ote et al. (2001); (2)C^ote et al. (2003); (3) Richtler et al. (2002); (4) Dirsch et al. (2002); (5)

C^ote (1999); (6) Perrett et al. (2002).

Table 2. Age estimates for GCs in gEs

Galaxy t(BGC) t(RGC) t Source Ref.

M87 13:7 1:8 12:7 2:2 1:0 3:3 Sp ectra 1

3 6 LF(V I ) 2

0:2 2:0 LF(uv by ) 3

> 15 > 15 Sp(+TMB02) 4

{a few (V I )-(V K ) 5

M49 0:6 3:2 LF(V I ) 6

3 6 LF(V I ) 7

14:5 4 13:8 6 0:7 7:2 Sp ectra 8

13:9 4:1 12:8 2:6 Sp ectra (+W94) 4

> 15 2 (5 10) (3 4) Sp (+TMB02) 4

NGC 1399 11(N=8) 2? (N=2) Sp ectra 9

NGC 3115 2 3 (V I )-(V K ) 10

References: (1) Cohen, Blakeslee, & Ryzhov (1998); (2) Kundu et al. (1999); (3) Jordan et al. (2002a); (4)

Cohen, Blakeslee, & C^ote (2003) used two mo dels: Worthey (1994, W94) and Thomas, Maraston, & Bender (2003,

TMB03); (5) Kissler-Patig, Bro die, & Minniti (2002); (6) Puzia et al. (1999); (7) Lee & Kim (2000); (8) Beasley et

al. (2000); (9) Forb es, D. A. et al. (2001) (the ages of the two metal-rich globular clusters may b e 2 Gyr or > 15

Gyr); (10) Puzia et al. (2002) who gave a mean of estimates based on four di erent mo dels.

204 LEE

Fig. 12.| (Upp er left panel) Color distribution of globular cluster candidates in M87 with 334 measured radial velo city

(black histogram) of C^ ote et al. (2001) among the photometric sample (shaded histogram). (Upp er right panel) Radial

velo city distribution of globular cluster candidates in M87. (Lower panel) (C T ) vs. radial velo cities of the globular

cluster candidates in M87. There are 278 globular clusters inside the dashed line. (source: C^ote et al. 2001)

There are 42 early-typ e galaxies (except for our Galaxy Host Galaxies (i) Relation with

and M31) in the sample, and 22 to 28 elliptical galax-

The correlation b etween the RGCs and their host

ies with HST observations among them were used for

galaxies has b een p ointed out by Forb es, Bro die, &

linear tting of the data (represented by the solid line).

Grillmair (1997), Forb es & Forte (2001) and Larsen et

Figure 13 shows several imp ortant features: 1) The

al. (2001), although Kundu & Whitmore (2001a) ar-

(V I ) colors of the RGCs are very similar to those of

gued that the correlation is weak at b est ,or disapp ears

the host galaxies. 2) The (V I ) colors of the RGCs

when the galaxies with strongest bimo dality are con-

show a correlation with those of their host galaxies,

sidered. Figure 13 displays the colors of the globular

while those of the BGCs do little. 3) The slop e of

clusters in elliptical galaxies versus the parameters for

the (V I ) colors of the RGCs vs. the color of their

their host galaxies. Our Galaxy and M31 are included

host galaxies (the solid line in Figure 13(c)) is some-

for comparison. The sources of the data are: Forb es

what atter than that for the color-magnitude relation

& Forte (2001), Kundu & Whitmore (2001a,b), Larsen

of the host galaxies (the dashed line in Figures 13(a)

et al. (2001), Lee & Kim (2000), Lee et al. (2002) for

and the solid line in Figure 13(f )). 4) The di erence

globular clusters, Tonry et al. (2001) and Jensen et

in color b etween the BGCs and the RGCs is shown

al. (2003) for distances of the galaxies, Golev & Prug-

in Figure 13(e). The color di erence is in the range

of the galaxies, and Barmby et al. niel (1998) for M g

of 0:15 < (V I ) < 0:27, which corresp onds to the

(2000) for the globular clusters in M31 and our Galaxy.

metallicity range of 0:63 < [Fe=H] < 1:14 dex, us-

ELLIPTICAL GALAXY FORMATION 205

Fig. 13.| Relations b etween the globular clusters and their host galaxies with early typ es. (a) (V I )(GC) vs. (V

I )(galaxy). (b) (V I )(GC) vs. Log (galaxy), velo city disp ersion. (c) (V I )(GC) vs. M (galaxy). (d) (V I )(GC)

v V

(galaxy). (e) (V I )(RGC-BGC) vs. (V I )(galaxy). (f ) (V I )(galaxy) vs. M (galaxy). The op en circles, vs. M g

V 2

lled circles, lled triangles represent, resp ectively, Es, S0s and galaxies with typ es later than S0. Op en square and star

e/H]. Solid lines symb ol represent, resp ectively, our Galaxy and M31 for which the (V I ) colors were predicted from [F

represent the linear ts to E's with HST WFPC2 photometry. The dashed line in (a) represent the one-to-one relation,

and the dashed line in (c) represents the color-magnitude relation of host galaxies which is also plotted in (f ) as a solid line.

Sources of the data: Barmby et al. (2000), Larsen et al. (2001), Forb es & Forte (2001), Kundu & Whitmore (2001a,b), Lee

et al. (2002), Golev & Prugniel (1998), and Tonry et al. (2001).

Young to Intermediate-Age Globular Clus- ing [Fe/H]= 4:22(0:39)(V I ) 5:39(0:35) derived (j)

Galaxies ters in Interacting/Merging from M31 globular clusters (Barmby et al. 2000). 5)

Similar trends are seen against the logarithmic value of

Star clusters brighter and younger than typical glob-

velo city disp ersion, the total magnitude, and the M g

ular clusters in our Galaxy have b een found in several

of the host galaxies (Figure 13(b), 13(c), and 13(d)).

merging galaxies with young to intermediate-age and

These results show that the RGCs show a close rela-

in a few dynamically young elliptical galaxies (such as

tion with the stellar halos of their host galaxies, while

NGC 3610 and NGC 1700) (Whitmore 2000, Schweizer

the BGCs do not.

206 LEE

2002b, Davies et al. 2001, Larsen et al. 2003). (b) The MCGCFM

They app ear to follow well the age sequence in the

Fall & Rees (1985) prop osed that globular clusters

V (V I ) diagram where V is a magni-

10 10

would form in the collapsing gas of a protogalaxy. Due

tude di erence b etween the 10th-bright 2nd-generation

to the thermal instability in hot gaseous halos of young

GC and its old counterpart (see Figure 6 in Schweizer

galaxies the cold comp onents are compressed into dis-

2002b). This evidence shows clearly that massive star

crete clouds with temp eratures near 10 K, mean den-

clusters formed during the merger of galaxies, and the

3 6

sities with 1{10 M p c , and masses of 10 M .

mering of galaxies happ en continuously until recently.

Harris & Pudritz (1994) noted that the mass func-

tions of massive globular clusters (> 10 M ) in giant

Summary (k)

elliptical galaxies and disk galaxies have a p ower law in-

Several observational evidences show that the RGCs

dex similar to that of the giant molecular clouds in the

1:7

in elliptical galaxies are closely related with the stellar

Milky Way (N M ), and prop osed that globular

halo of their host galaxies in color and spatial distri-

clusters formed out of dense cores in the (hyp othetical)

bution, while the BGCs are not. Three giant elliptical

sup ergiant molecular clouds (with masses of 10 M

galaxies show all di erent kinematics of the globular

and diameters of 1 kp c) which existed in the proto-

clusters. Present age estimates of the BGCs and the

galactic ep o ch. The masses of these clouds are similar

RGCs contain large uncertainty so that the di erence

to those of dwarf galaxies. They predicted that the

in age b etween the two are not yet known, but show

ep o ch of ma jor globular cluster formation is at z < 6.

that the BGCs and the RGCs may b e b oth older than

Numerical simulations based this concept were given

10 Gyr (see also Ashman & Zepf 2001).

by Weil & Pudritz (2001).

Forb es, Bro die, & Grillmair (1997) prop osed an

VI. MODELS OF FORMATION OF GLOB-

episo dic in situ formation plus tidal stripping mo del.

ULAR CLUSTERS

In this mo del, metal-p o or globular clusters are formed

rst at an early stage in the initial collapse of the pro-

Mo dels for the formation of globular clusters are of-

togalactic cloud with only minor star formation. Af-

ten classi ed according to the formation ep o ch of glob-

ter some time (ab out 4 Gyr) of quiescence, metal-rich

ular clusters: Primary, secondary and tertiary forma-

globular clusters are formed out of more enriched gas,

tion mo dels, or pre-galactic, proto-galactic, and p ost-

roughly contemp oraneously with most of the galaxy

galactic mo dels (see C^ote 2002). Here I divide the

stars, during the ma jor collapse of the protogalac-

mo dels according to the mo dels of elliptical galaxy for-

tic cloud, which then forms an elliptical galaxy with

mation (MCM and HMM): the primary origin mo d-

two globular cluster p opulations. Most of the glob-

els (only the BGC), the Monolithic Collapse Globular

ular clusters are formed in the rst formation episo de

Cluster Formation Mo dels (MCGCFM), and the Hier-

and should not b e structurally related to the halo light,

archical Merging Globular Cluster Formation Mo dels

while the newly formed GCs are closely coupled to the

(HMGCFM).

galaxy and share a common chemical enrichment his-

In the primary origin mo dels, the BGCs formed

tory and structural characteristics.

b efore galaxies formed, and the RGCs are not men-

tioned. In the MCGCFM, the BGCs and then the

RGCs formed during the multiple collapses of a large

There are two typ es of mo dels in the HMGCF. One

proto-galactic cloud. In the HMGCFM, the BGCs and

concentrates on the ma jor gaseous merger of two spi-

RGCs formed in the regime of the HMM.

ral galaxies at late stage of evolution, and the other

includes b oth early formation of globular clusters and

(a) Primary Origin Mo dels:b efore Galaxies

late merging/accretion pro cesses of smaller ob jects of

Peebles & Dicke (1968) prop osed, noting the little

a dwarf galaxy size (proto-galactic fragments).

variation of the luminosities of globular clusters, that

(Ma jor Gaseous Merger Mo dels)

globular clusters formed from the gravitationally b ound

Ashman & Zepf (1992) prop osed a gaseous merger

gas clouds b efore galaxies formed so on after the recom-

mo del (as also Zepf & Ashman (1993), Zepf et al.

bination era, when the temp erature go es down ab out

4 3

(2000)). In the gaseous merger mo del, an elliptical

4000 K and the density is ab out 10 atoms cm .

galaxy is formed by the merging of two or more gas-rich

Right after the recombination era, the Jeans radius is

5 6

spiral galaxies. The spiral galaxies have enriched gas in

ab out 5 p c, and the Jeans mass 10 10 M , which is

the disk and metal-p o or GCs in the halo (which need

similar to the mean mass of the globular clusters in the

to b e explained separately). New GCs are formed from

halo of our Galaxy. Newer versions of this mo del were

the enriched gas during the merger/interaction pro cess.

given by Peebles (1984) and Cen(2001). These globu-

The resulting elliptical galaxy then has two GC p opula-

lar clusters are exp ected to b e very old, metal-p o or and

tions: the younger, more metal-rich, and spatially more

everywhere.

concentrated clusters formed as a result of the merger,

and the original, metal-p o or, spatially more extended

ELLIPTICAL GALAXY FORMATION 207

ure 12), 2) they had to truncate the formation of the GCs formed in the progentor spirals.

BGCs at z 5 to avoid to o high metallicity at z = 0,

A key idea in this mo del was to explain the fact

the origin of which needs to b e explained, and 3) they

that the sp eci c frequencies in elliptical galaxies are a

used formation eciency of the RGCs, 0.7%. This

factor of two higher than those in spiral galaxies. In this

value for the RGCs is 3-4 times higher than those for

mo del, the BGCs are exp ected to show some rotation,

M87 and M49 (C^ote 2002).

while the RGCs are exp ected to show little rotation,

Kravtsov & Gnedin (2003) found, in their numeri- b ecause the angular moment would b e transp orted to

cal simulation for the formation of globular clusters in the outer region during the merging pro cess (Zepf et

the Milky Way-size galaxy, that the b est conditions for al. 2000).

globular cluster formation at the high density core of

Bekki et al. (2002) presented an extensive simu-

the sup ergiant molecular clouds in the gaseous disks of

lation of globular clusters formed in merging and in-

the proto-galaxy are at z 3 5, although the rst

teracting galaxies, assuming that the GCs formed at

globular clusters start to app ear at z 12. Their sim-

high redshift (z > 5) in protoglactic fragments, and

ulation covered z > 3.

during the subsequent gas-rich merging of these frag-

C^ote (2002) p ointed out that various names (proto- ments. They found that the BGCs should b e 3 Gyr

galactic fragment, sup ergiant molecular clouds, proto- older for gEs in clusters (5 Gyr for low-luminosity ellip-

galactic disks, Searle-Zinn fragments, and failed dwarfs) tical galaxies in the eld or groups) than the RGCs, and

are the same ob jects in their physical prop erties. that the age disp ersion of the RGCs is large (t 5 12

Gyr). They found also that the dissipative merging of

(d) Summary

present-day spirals would pro duce sup er-solar metal-

rich globular clusters which are not observed now. To

There are several mo dels which can explain well sev-

avoid this problem, they p ointed out that, if elliptical

eral observational constraints, but none of them is free

galaxies form by dissipative ma jor merger, then they

from a few ma jor limits at the moment. However, it

must do it at a very early ep o ch.

app ears that the concept `simple is b eautiful' is chang-

ing to `complexity and delicacy are reality'. Finally we

(Proto-galactic Fragment Mo dels)

may need a bibimbap mo del to explain the formation

C^ote, Marzke, & West (1998) prop osed a mo del

of giant elliptical galaxies and globular clusters (bibim-

showing a change of paradigm: the metal-rich glob-

bap is a very delicious Korean rice mixed with various

ular clusters are the intrinsic p opulation of a galaxy,

vegetables and some meat, and bibim (Korean) means

and the galaxy acquires later the metal-p o or globular

mixing several).

clusters via mergers or tidal stripping. Later C^ote et al.

(2000) suggested that the Galactic halo and its globu-

VI I. CONCLUSION AND FUTURE

lar cluster system were assembled via the accretion and

disruption of 10 metal-p o or, proto-Galactic frag-

Most giant elliptical galaxies show a bimo dal color

ments by a proto-bulge which owns metal-rich globular

distribution of globular clusters, indicating a factor of

cluster system (see also C^ote, West, & Marzke 2002).

20 metallicity di erence b etween the two. The RGCs

In this mo del no globular clusters formed during the

(metal-rich globular clusters) are closely related with

accretion/merger pro cess, in contrast to the gaseous

the stellar halo in color and spatial distribution, while

merger mo dels, and the total numb er of globular clus-

the BGCs (metal-p o or globular clusters) are not. The

ters and their metallicities are determined by the mass

ratio of the numb er of the RGC and the BGC varies

of the proto-galactic fragment where globular clusters

dep ending on galaxies. Considering the observational

formed.

constraints from stars and globular clusters in giant

Beasley et al. (2002) presented a simulation of the

elliptical galaxies, I draw the following conclusive con-

globular cluster systems of 450 elliptical galaxies in the

straints for understanding when, how long, and how

mo del, assuming that the globular clusters form CDM

elliptical galaxies and globular clusters formed, which

at two ep o chs, rst at z > 5 in protogalactic fragments,

are sketched in Figure 14.

and second during the subsequent gas-rich merging of

1. There are found to b e many giant elliptical galax-

these fragments. They found 1) the GCs in most galax-

ies which formed at z > 2 (> 10 Gyr ago).

ies show bimo dality in metallicity, 2) the RGCs (9 Gyr)

are on average younger than the BGCs (12 Gyr), 3) the

2. The RGCs and most stars in elliptical galaxies

RGCs show a large range in age (5 to 12 Gyr), which

formed at the similar time > 10 Gyr ago.

increases for low-luminosity galaxies, and for galax-

3. The BGCs formed b efore the RGCs, but after

ies with low circular velo city halos, and 4) the RGCs

reionization at z 10 30 (< 13 Gyr ago).

formed via gaseous merging, the bulk of which o ccurs

4. The BGCs all must have formed out of metal-

at 1 < z < 4.

p o or clouds, indep endently of most stars in their

A few remaining problems in this mo del are 1)

present host galaxy.

their mo del galaxies do not show the well-known color-

5. There were no ma jor bursts b etween the RGCs

magnitude relation of elliptical galaxies (see their Fig-

and the BGCs, but the globular clusters continued

208 LEE

Fig. 14.| A schematic diagram showing the formation time of gEs, BGCs and RGCs according to the Monolithic Collapse

) is based on the concordance cosmological Mo dels (MCM) and the Hierarchical Merging Mo dels (HMM). Lo okback time(z

mo del. Magenta solid line at the lo okback time 12.95 Gyr (z = 17) represents the reionization ep o ch, indicating the birth

of the rst stellar systems. Blue dashed line at the lo okback time 12.6 Gyr represents the mean age of the 17 metal-p o or

globular clusters in the halo of our Galaxy. Green arrow line represents a sketch for chemical enrichment history needed.

forming around the ep o chs of the BGC and RGC 9. Merging (ma jor, minor) should continue to o ccur,

p eaks. while keeping the color-magnitude relations, the

fundamental planes, and bimo dality of globular

6. Therefore there must have b een two ma jor bursts

clusters.

with a gap during the p erio d of 10{13 Gyr ago,

and the formation duration should b e 1 Gyr for

Some remaining problems are:

the BGCs, and 1 Gyr or little longer for the

RGCs. The ma jor ep o chs for the formation of the

1. Why there is a gap b etween the BGC formation

BGCs and the RGCs may b e 12.5 Gyr and 10.5

ep o ch and the RGC formation ep o ch, seen over

Gyr (presented by the blue and red bars in Figure

ab out 5 magnitude range of galaxies (by a factor

14). Note that 12.5 Gyr is similar to the mean age

of 100 in the total luminosity)?

of the metal-p o or globular clusters in the Galactic

2. How did the metallicity increase by a factor of 20

halo.

during the short gap (t 2 Gyr)?

7. From the ep o ch of the BGC formation to that

3. Why are kinematics of the globular clusters in

of the RGC formation, the metallicity should b e

M49, M87 and NGC 1399 so di erent? Is it just

increased by a factor of 20 ([Fe=H] = 1:0 dex).

due to a statistical e ect or due to intrinsic di er-

ence?

8. Both the MCM and HMM are needed, but the

HMM needs to have a way to make giant elliptical

4. How can the HMM make giant elliptical galaxies

galaxies form at > 10 Gyr ago (much earlier than

> 10 Gyr ago?

the present HMMs predict). The MCM by itself

cannot repro duce the large-scale structures. A list for the future is:

ELLIPTICAL GALAXY FORMATION 209

Beasley, M. A. et al, 2000, Ages and Metallicities of Glob- 1. Determination of the relative age of the BGCs and

ular Clusters in NGC 4472, MNRAS, 318, 1249

RGCs with an accuracy of 1 Gyr is desp erately

needed to solve the unknown mystery directly. For

Beasley, M. A. et al, 2002, On the Formation of Globu-

this we need to reduce the measurement errors of

lar Cluster Systems in a Hierarchical Universe, MNRAS,

sp ectral lines or photometric colors (optical and

333, 383

near-IR), and to design new ways of precise age

Beasley, M. A., Harris, W. E., Harris, G. L., & Forb es, D.

estimates. At the same time, we need to improve

A. 2003, Simultaneous Mo delling of the Stellar Halo and

iso chrone mo dels based on p opulation synthesis.

Globular Cluster System of NGC 5128, MNRAS, 340,

341

2. We need to understand the diversity in the kine-

Bekki, K., Forb es, D. A., Beasley, M. A., & Couch, W.

matics of the globular clusters in M87, M49 and

J. 2002, Globular Cluster Formation from Gravitational

NGC 1399, and need to investigate the kinemat-

Tidal E ects of Merging and Interacting Galaxies, MN-

ics of the globular clusters in more giant elliptical

RAS, 335, 1176

galaxies.

Bekki, K., Harris, W. E., & Harris, G. L. H. 2003, Origin of

3. Search for intracluster globular clusters and intra-

the metallicity distribution of the NGC 5128 stellar halo,

galactic globular cluster may pro duce an interest-

MNRAS, 338, 587

ing result . There are some rep orts related with

Bender, R., Burstein, D., & Fab er, S. 1992, Dynamically

this issue (Hilker 2002, Jordan et al. 2002b), but

hot galaxies. I - Structural prop erties, ApJ, 399, 462

more are needed.

Benson, A. J., Ellis, R. S., & Menanteau, F. 2002, On the

4. Ultimatum mo dels to explain b oth stars and glob-

Continuous Formation of Field Spheroidal Galaxies in

ular clusters in galaxies are needed. When will it

Hierarchical Mo dels of Structure Formation, MNRAS,

b e?

336, 564

Benson, A. J., & Madau, P. 2003, Early Preheating and

Galaxy Formation, MNRAS, in press (astro-ph/0303121)

The author is grateful to his collab orators on the

pro ject of extragalactic globular clusters, esp ecially,

Bernardi, M. et al. 2003a, Early-typ e Galaxies in the Sloan

Digital Sky Survey. IV. Colors and Chemical Evolution,

Doug Geisler, Ata Sara jedini, Brad Whitmore, Rupali

AJ, 125, 1882

Chandar, Nobuo Arimoto, Eunhyeuk Kim, Sang Chul

Kim, Hong So o Park and Ho Seong Hwang, and to

Bernardi, M. et al. 2003b, Early-typ e Galaxies and their

Prof.Hyung Mok Lee for inviting him to the exciting

Environment: Constraints on Mo dels of Galaxy Forma-

tion, BAAS, in press.

meeting on the formation and interaction of galaxies

during the Korean heatwave of Worldcup so ccer 2002.

Blain, A. W. et al. 2002, Submillimeter Galaxies, Physics

The author thanks the sta of the Department of Ter-

Rep ort, 369, 111

restrial Magnetism, Carnegie Institution of Washing-

Bower, R. G., Lucey, J. R., & Ellis, R. S. 1992, Preci-

ton, for their kind hospitality while writing this pap er

sion Photometry of Early-typ e Galaxies in the Coma and

as Visiting Investigator. Vera Rubin is thanked for her

Virgo Clusters: a Test of the Universality of the Colour-

kind and careful reading the draft, for p ointing out an

Magnitude Relation I I. Analysis, MNRAS, 254, 601

imp ortant reference by Baum (1959), and for showing

Bower, Ko dama, & Terlevich 1998, The Colour-magnitude

what an astronomer is. This pro ject was supp orted in

Relation as a Constraint on the Formation of Rich Clus-

part by the Korean Research Foundation Grant (KRF-

ter Galaxies, MNRAS, 299, 1193

2000-DP0450) and the BK21 program.

Burkert, A. 1994, On the Formation of Elliptical Galaxies,

Review of Mo dern Astronomy, 7, 191

REFERENCES

Burkert, A. & Naab, T. 2003, The Formation of Spheroidal

Stellar Systems, preprint (astro-ph/0305076)

S. E. 1992, The Formation of Glob- Ashman, K. M., & Zepf,

Burstein,D. et al. 1997, Global Relationships Among the

ular Clusters in Merging and Interacting Galaxies, ApJ,

Physical Prop erties of Stellar Systems, AJ, 114, 1365

384, 50

Cattaneo, A., & Bernardi, M. 2003, The Quasar Ep o ch

Ashman, K. M., & Zepf, S. E. 1998, Globular Cluster Sys-

and the Stellar Ages of Early-typ e Galaxies, MNRAS,

tems(Cambridge: Cambridge University Press)

in press (astro-ph/0305298)

Ashman, K. M., & Zepf, S. E. 2001, Some Constraints on

Cen, R. 2001, Synchronized Formation of Subgalactic Sys-

the Formation of Globular Clusters, ApJ, 122, 1888

tems at Cosmological Reionization: Origin of Halo Glob-

Barmby et al. 2000, M31 Globular Clusters: Colors and

ular Clusters, ApJ, 560, 592

Metallicities, AJ, 119, 727

Chapman, S. C., Blain, A. W., Ivison, R. J., & Smail, I.

Barnes, J. E. 1998, Mergers and Galaxy Assembly, astro-

2003, A Median Redshift of 2.4 for Galaxies Bright at

ph/9811242

Submillimetre Wavelengths, Nature, 422, 695

Baum, W. 1959, The Hertzsprung-Russell diagrams of old

Chiosi, C., & Carraro, G. 2002, Formation and Evolution

stellar Populations, in The Hertzsprung-Russell Dia-

of Galaxies, MNRAS, 335, 335

gram, IAU Symp o. No. 10, 23, J. Greenstein ed

210 LEE

Elson, R. A. W. & Santiago, B. X. 1996, The globular clus- Cimatti, A. et al. 2002a, The K20 Survey I. Disentangling

ters in M87: a bimo dal colour distribution, MNRAS, 280, old and dusty star-forming galaxies in the ERO p opula-

971 tion, A&A, 384, L68

Fall, S. M., & Rees, M. J. 1985, A Theory for the Origin of Cimatti, A. et al. 2002b, The K20 Survey IV. The Red-

Globular Clusters, ApJ, 298, 18 < 20 galaxies: A Test of Galaxy shift Distribution of K

Formation Mo dels, A&A, 391, L1

Fall, S. M., & Zhang, Q. 2001, Dynamical Evolution of the

Mass Function of Globular Star Clusters, ApJ, 561, 751 Cimatti, A. 2003, The Formation and Evolution of Field

Massive Galaxies, preprint(astro-ph/0303023)

Forb es, D. A., Bro die, J. P., & Grillmair, C. J. 1997, On the

Origin of Globular Clusters in Elliptical and cD Galaxies, Cohen, J. & Ryzhov, A. 1997, The Dynamics of the M87

AJ, 113, 1652 Globular Cluster System, ApJ, 486, 230

Forb es, D. A., & Forte, J. C. 2001, The Connection b e- Cohen, J. G., Blakeslee, J. P., & Ryzhov, A. 1998, The

tween Globular Cluster Systems and the Host Galaxies, Ages and Abundances of a Large Sample of M87 Globular

MNRAS, 322, 257 Clusters, ApJ, 496, 908

Forb es, D., Beasley, M. A., Bro die, J. P., & Kissler-Patig, Cohen, J. G., Blakeslee, J. P., & C^ote, P. 2003, The Ages

M. 2001, Age Estimates for Globular Clusters in NGC and Abundances of a Sample of Globular Clusters in M49

1399, ApJ, 565, L143 (NGC 4472), preprint (astro-ph/0304333)

Geisler, D., Lee, M. G., & Kim, E. 1996, Washington Pho- Conselice, C. J. 2002, Observing Massive Galaxy Forma-

tometry of the Globular Cluster System of NGC 4472. I. tion, preprint (astro-ph/0212468)

Analysis of the Metallic, AJ, 111, 1529

C^ote, P. 1999, Kinematics of the Galactic Globular Clus-

Geisler, D., Minniti, D., & Greb el, E. 2002, Extragalactic ter System: New Radial Velo cities for Clusters in the

Globular Clusters, IAU Symp. No. 207, ASPCS. Direction of the Inner Galaxy, AJ, 118, 406

Giavalisco, M. 2002, Lyman-break Galaxies, ARA&A, 40, C^ote, P. 2002, The Formation of Globular Cluster Systems,

579 astro-ph/0210582, in New Horizons in Globular Cluster

Astronomy, ASPCS, Piotto, G. et al. eds.

Gnedin, O. Y., Lahav, O., & Rees, M. J. 2001, Do Globular

Clusters Time the Universe?, C^ote, P., Marzke, R. O., & West, M. J. 1998, The Formation preprint (astro-ph/0108034)

of Giant Elliptical Galaxies and theire Globular Cluster

Gnedin, O. Y. 2002, Formation of Globular Clusters: In

Systems, ApJ, 501, 554

and Out of Dwarf Galaxies, astro-ph/0210556

C^ote, P., Marzke, R. O., West, M. J., & Minniti, D. 2000,

Golev, V., & Prugniel, Ph. 1998, A Catalogue of M g In-

Evidence for the Hierarchical Formation of the Galactic

dices of Galaxies and Globular Clusters, A&AS, 132, 255

Spheroid, ApJ, 533, 869

Hanes, D. et al. 2001, Dynamics of the Globular Cluster

C^ote, P., et al. 2001, Dynamics of the Globular Cluster

System Asso ciated with M87 (NGC 4486). I. New CFHT

System Asso ciated with M87 (NGC 4486): I I. Analysis,

MOS Sp ectroscopy and the Comp osite Database, ApJ,

ApJ, 559, 828

559, 812

C^ote, P., West, M. J., & Marzke, R. O. 2002, Globular

Harris, W. E. 1991, Globular Cluster Systems in Galaxies

Cluster Systems and the Missing Satellite Problem: Im-

Beyond the Lo cal Group, ARA&A, 29,543

plications for Cold Dark Matter Mo dels, ApJ, 567, 853

Harris, W. E., & Pudritz, R. E. 1994, Sup ergiant Molecular

C^ote, P., McLaughlin, D. E., Cohen, J. G., & Blakeslee, J.

Clouds and the Formation of Globular Cluster Systems,

P. 2003, Dynamics of the Globular Cluster System Asso-

ApJ, 420, 177

ciated with M49 (NGC 4472): Cluster Orbital Prop erties

Harris, W. E. 2001, in Star Clusters, Saas-Fee Advanced

and the Distribution of Dark Matter, ApJ, 591, 850

Course 28, Labhardt, L. et al. eds.

Davies, R. L. et al. 2001, Galaxy Mapping with the Sauron

Harris, W. E., & Harris, G. L. H. 2002,The Halo Stars in

Integral-Field Sp ectrograph: The Star Formation His-

NGC 5128. I I I. An Inner Halo Field and the Metallicity

tory of NGC 4365, ApJ, 548, L33

Distribution, AJ, 123, 3108

Dirsch, B. et al. 2002, The Mass Pro le of NGC 1399 De-

Harris, W. E. 2003, Globular Clusters: the View from HST,

termined from Globular Cluster Dynamics, RMxAC, 14,

preprint

Harris, W. E., & van den Bergh, S. 1981, Globular Clusters

Dirsch, B. et al. 2003, The Globular Cluster System of

in Galaxies Beyond the Lo cal Group. I - New Cluster

NGC 1399.I. A Wide-Field Photometric Study, AJ, 125,

Systems in Selected Northern Ellipticals, AJ, 86, 1629

1908

Hilker, M. 2002, Globular Clusters in Nearby Galaxy Clus-

Eggen, O. J., Lynden-Bell, D., & Sandage, A. R. 1962, Ev-

ters, preprint (astro-ph/0210466)

idence from the Motions of Old Stars that the Galaxy

Collapsed, ApJ, 136, 748

Helly, J. C., Cole, S., Frenk, C. S., Baugh, C. M., Benson,

A., Lacey, C. 2003, Galaxy formation using halo merger

Ellis, R. 2001, Galaxy Formation and Evolution: Recent

histories taken from N-b o dy simulations, MNRAS, 338,

Progress, astro-ph/0102056

903

Elmergreen, B. G. 2002, Star Formation from Galaxies to

Hogg, D. W. 1999, Distance Measures in Cosmology, astro-

Globules, ApJ, 577, 206 ph/9905116

ELLIPTICAL GALAXY FORMATION 211

Kundu, A. V., & Whitmore, B. C. 2001b, New Insights Jensen, J. B. et al. 2003, Measuring Distances and Prob-

from Hubble Space Telescop e Studies of Globular Cluster ing the Unresolved Stellar Populations of Galaxies Using

Systems. I I. Analysis of 29 S0 Systems AJ, 122, 1251 Infrared Surface Brightness Fluctuations, ApJ, 583, 712

an, A., C^ote, P., West, M. J., & Marzke, R. O. 2002a, Jord

Kundu, A. V. et al. 1999, The Globular Cluster System in

The Relative Ages of the Globular Cluster Subp opula-

the Inner Region of M87, ApJ, 513, 733

tions in M87, ApJ, 576, L113

Larsen, S. S. et al. 2001, Prop erties of Globular Cluster

Jordan, A., West, M. J., C^ote, P., & Marzke, R. O. 2002b,

Systems in Nearby Early-Typ e Galaxies, AJ, 121, 2974

A Point Source Excess in Ab ell 1185: Intergalactic Glob-

Larsen, S. S. et al. 2003, Evidence for an Intermediate-Age

ular Clusters?, AJ, 125, 1642

Metal-rich Population of Globular Clusters in NGC 4365,

Kau mann, G., White, S. D. M., & Guiderdoni, B. 1993,

ApJ, 585, 767

The Formation and Evolution of Galaxies within Merging

Larson, R. B. 1974, Dynamical Mo dels for the Formation

Dark Matter Halos, MNRAS, 264, 201

and Evolution of Spherical Galaxies, MNRAS, 166, 585

Kawada, D., & Gibson, B. K. 2003, GCD+: a New Chemo-

Lee, M. G., & Geisler, D. 1993, Metal abundances for a

dynamical Approach to Mo delling Sup ernovae and Chem-

large sample of globular clusters in M87, AJ, 106, 493

ical Enrichment in Elliptical Galaxies, MNRAS, 340, 908

Lee, M. G., Kim, E.,& Geisler, D. 1997, Young Star Clusters

Kho chfar, S., & Burkert, A. 2003, The Imp ortance of Dry

in the Dwarf Irregular Galaxy, UGC 7636, Interacting

and Mixed Mergers for Earl-typ e Galaxy Formation,

with the Giant Elliptical Galaxy NGC 4472, AJ, 114,

preprint (astro-ph/0303529)

1824

Kim, E., Lee, M. G., & Geisler, D. 2000, Wide- eld CCD

Lee, M. G., Kim, E.,& Geisler, D. 1998, Washington Pho-

Surface Photometry of the Giant Elliptical Galaxy NGC

tometry of the Globular Cluster System of NGC 4472.I I.

4472 in the Virgo cluster, MNRAS, 314, 307

The Luminosity Function and Spatial Structure, AJ, 115,

Kissler-Patig, M., & Gebhardt, K. 1998, The Spin of M87

947

as Measured from the Rotation of its Globular Clusters,

Lee, M. G., & Kim, E. 2000, The Globular Cluster System

AJ, 116, 2237

in the Inner Region of the Giant Elliptical Galaxy NGC

Kissler-Patig, M., Forb es, D. A., & Minniti, D. 1998, Con-

4472, AJ, 120, 260

straints on the Merger Mo dels of Elliptical Galaxies from

Lee, M. G., Kim, E., Geisler, D, Bridges, T., & Ashman, K.

their Globular Cluster Systems, MNRAS, 298, 1123

2002, A Comparative Study of Globular Clusters in UGC

Kissler-Patig, M., et al. 1999, Toward an Understanding of

9799 and NGC 1129, in Extragalactic Star Clusters, IAU

the Globular Cluster Overabundance around the Central

Symp. No. 207, E. K. Greb el, D. Geisler, & D. Minniti

Giant Elliptical Galaxy NGC 1399, AJ, 117, 1206

eds., 330

Kissler-Patig, M., Bro die, J. P., & Minniti, D. 2002, Ex-

Lee, M. G., Park, H.S., Kim, E., & Geisler, D. 2003, in

tragalactic Globular Clusters in the Near Infrared, I. A

preparation

Comparison b etween M87 and NGC 4478, A&A, 391,

441

Lo eb, A., & Peebles, P. J. E. 2003, Cosmological Origin

of the Stellar Velo city Disp ersions in Massive Early-typ e

Ko dama, T., & Arimoto, N. 1997, Origin of the Colour-

Galaxies, ApJ, in press (astro-ph/0211465)

magnitude Relation of Elliptical Galaxies, A&Ap, 320,

Matteucci, F. 2002, What Determines Galactic Evolution?,

preprint (astro-ph/0210540)

Ko dama, T., et al. 1998, Evolution of the Colour-magnitude

Relation of Early-typ e Galaxies in Distant Clusters,

Meza, A., Navarro, J. F., Steinmetz, M., & Icke, V. R. 2003,

A&Ap, 334, 99

Simulations of Galaxy Formation in a Lamb da CDM

Universe I I I: The Dissipative Formation of an Elliptical

Ko dama, T., Bower, R. G., & Bell, E. F. 1999, The Colour-

Galaxy, ApJ, preprint (astro-ph/0301224)

magnitude Relation of Early-typ e Galaxies in the Hubble

Deep Field, MNRAS, 306, 561

Minniti, D., Kissler-Patig, M., Goudfro oij, P., & Meylan,

G. 1998, Radical Velo cities of Globular Clusters in the

Kogut, A. et al. 2003, First Year Wilkinson Microwave

Giant Elliptical Galaxy NGC 1399, AJ, 115, 121

Anisotropy Prob e (WMAP) Observations: TE Polariza-

tion, ApJ, in press (astro-ph/0302213)

Mo ore, B. et al. 1999, Dark Matter Substructure within

Galactic Halos, ApJ, 524, L19 Kormendy, J. 1985, Families of Ellipsoidal Stellar Systems

and the Formation of Dwarf Elliptical Galaxies, ApJ,

Partridge, R. B., & Peebles, P. J. E. 1967, Are Young Galax-

295, 73

ies Visible?, ApJ, 147, 868

Krauss, L. M., & Chab oyer, B. 2003, Age Estimates of Glob-

Peebles, P. J. E. 2002, When Did the Large Elliptical Galax-

ular Clusters in the Milky Way: Constaints on Cosmol-

ies Form?, preprint (astro-ph/0201015), in Metcalfe, N.,

ogy, Science, 299, 65

& Shanks, T. eds, ASPCS, A New Era in Cosmology

Kravtsov, A. V., & Gnedin, O. Y. 2003, Formation of Glob-

Peebles, P. J. E., & Dicke, R. H. 1968, Origin of the Glob-

ular Clusters in Hierarchical Cosmology, preprint(astro-

ular Star Clusters, ApJ, 154, 891

ph/0305199)

Peebles, P. J. E. 1984, Dark Matter and the Origin of Galax-

Kundu, A. V., & Whitmore, B. C. 2001a, New Insights from

ies and Globular Clusters, ApJ, 277, 470

HST Studies of Globular Cluster Systems. I. Colors, Dis-

Peng, W., Ford, H. C., Freeman, K. C., White, R. L.,2002, tances, and Sp eci c Frequencies of 28 Elliptical Galaxies,

A Young Blue Tidal Stream in NGC5128, AJ, 124, 3144 AJ, 121, 1250

212 LEE

Thomas, D., & Maraston, C. 2003, The Impact of /Fe Peletier, R. F. et a. 1990, CCD Surface Photometry of

Enhanced Stellar Evolutionary Tracks on the Ages of El- Galaxies with Dynamical Data. II - UBR Photometry of

liptical Galaxies, A&A, 401, 429 39 elliptical galaxies, AJ, 100, 1091

Perrett et al. 2002, The Kinematics and Metallicity of the

Thomas, D., Maraston, C., & Bender, R. 2003, Stellar Pop-

M31 Globular Cluster System, AJ, 123, 2490

ulation Mo dels of Lick Indices with Variable Element

Abundance Ratios, MNRAS, 339, 897

Puzia, T. H., Kissler-Patig, M., Bro die, J. P., & Huchra, J.

P. 1999, The Age Di erence b etween the Globular Clus-

Tinsley, B. M. 1972, Stellar Evolution in Elliptical Galax-

ter Subp opulations in NGC 4472, AJ, 118, 2734

ies,ApJ, 178, 319

Puzia, T. H. et al. 2002, Extragalactic Globular Clusters

Tonry, J. et al. 2001, The SBF Survey of Galaxy Distances.

in the Near Infrared, I. The Globular Cluster Systems of

IV. SBF Magnitudes, Colors, and Distances, ApJ, 546,

NGC 3115 and NGC 4365, A&A, 391, 453

681

Renzini, A. 1999, Origin of Bulges, astro-ph/9902108

Tonry, J. et al. 2003, Cosmological Results from High-z

Sup ernovae, AJ, in press (astro-ph/0305008)

Renzini, A., & Ciotti, L. 1993, Transverse Dissections of the

Fundamental Planes of Elliptical Galaxies and Clusters

To omre, A. 1977, in The Evolution of Galaxies and Galaxy

of Galaxies, ApJ, 416, L49

Populations, ed. B. M. Tinsley & R. B. Larsen (New

Haven: Yale University), 401

Rho de, K. L., & Zepf, S. E. 2001, The Globular Cluster

System in the Outer Regions of NGC 4472, AJ, 121, 210

Weil, M. L., & Pudritz, R. E. 2001, Cosmological Evolution

of Sup ergiant Star-forming Clouds, ApJ, 556, 164 Richtler, T. et al. 2002, Sp ectroscopy of Globular Clus-

ters in NGC 1399-A Progress Rep ort, Extragalactic Star

Whitmore, B. C., & Schweizer, F. 1995, Hubble Space Tele-

Clusters, IAU Symp. No. 207, E. K.Greb el & D. Minniti

scop e Observations of Young Star Clusters in NGC4038/4039,

eds., 263

\The Antennae" Galaxies, ApJ, 109, 960

Evolved Galaxies Saracco, P. et al. 2002, Massive z{1.3

Whitmore, B. C., Sparks, W. B., Lucas, R. A., Macchetto,

Revealed, A&A,, preprint (astro-ph/0211394)

F. D., & Biretta, J. A. 1995, Hubble Space Telescop e Ob-

Schweizer, F. 2002a, Evolution of Globular Clusters Formed

servations of Globular Clusters in M87 and an Estimate

in Mergers, in Extragalactic Star Clusters, IAU Symp.

of H0, ApJ, 454, L73

No. 207, E. K. Greb el, D. Geisler, & D. Minniti eds.,

Whitmore, B. C. 2000, The Formation of Star Clusters,

630

preprint (astro-ph/0012546)

Schweizer, F. 2002b, Formation of Globular Clusters in Mer-

Williams, R. E. et al. 1996, The Hubble Deep Field: Obser-

gin Galaxies, in New Horizon in Globular Cluster Astron-

vations, Data Reduction, and Galaxy Photometry, AJ,

omy, ASPCS, G. Piotto et al. eds.

112, 1335

Searle, L., & Zinn, R. 1978, Comp ositions of Halo Clusters

Worthey, G. 1994, Comprehensive Stellar Population Mo d-

and the Formation of the Galactic Halo, ApJ, 225, 357

els and the Disentanglement of Age and Metallicity Ef-

Silk, J., & Bowwens, R. 2001, The Formation of Galaxies,

fects, ApJS, 95, 107

New Astronomy Reviews, 45, 337

Zepf, S. E., & Ashman, K. M. 1993, Globular Cluster Sys-

Smail, I. et al. 2001, A Photometric Study of the Ages and

tems Formed in Galaxy Mergers, MNRAS, 264, 611

Metallicities of Early-typ e Galaxies in A2218, MNRAS,

Zepf, S. E. et al. 2000, Dynamical Constraints on the For-

323, 839

mation of NGC 4472 and its Globular Clusters, AJ, 120,

Smith, G. H., & Burkert, A. 2002, On the Low-Mass End

2928

of the Globular Cluster Luminosity Function, ApJ, 578,

Zirm, A. W., Dickinson, M., & Dey, A. 2003, Massive El-

L51

liptical Galaxies at High Redshift: NICMOS Imaging of

Sp ergel, D. N. et al. 2003, First Year Wilkinson Microwave

1 Radio Galaxies, ApJ, 585, 90 z

Anisotropy Prob e (WMAP) Observations: Determina-

tion of Cosmological Parameters, ApJ, in press (astro-

ph/0302209)

Steinmetz, M. 2002, Galaxy Formation Now and Then,

preprint (astro-ph/0208252), in Hubble's Science Legacy:

Future Optical-Ultraviolet Astronomy from Space, ASP

Conf. Series.

Steinmetz, M., & Navarro, J. F. 2003, The Hierarchical

Origin of Galaxy Morphologies, New Astronomy, 8, 557

Strader, J., Bro die, J. P., Schweizer, F., Larsen, S. S., &

Seitzer, P. 2003, Keck Sp ectroscopy of Globular Clusters

in the Elliptical Galaxy NGC 3610, AJ, 125, 626

Tamura, N. et al. 2000, Origin of Color Gradients in Ellip-

tical Galaxies, AJ, 119, 2134

Terlevich, A. I., Caldwell, N., & Bower, R. G. 2001, The

Colour-magnitude Relation for Galaxies in the Coma

Cluster, MNRAS, 326, 1547