Lessons from Surveys of The Galaxy

0.19 MB·English

by  Rosemary F. G. Wyse

Mem. S.A.It. Vol. XXXX, 1
(cid:13)c SAIt 2004 Memorie della
Lessons from Surveys of The Galaxy
Rosemary F.G. Wyse
The Johns Hopkins University, Department of Physics & Astronomy, Baltimore,
MD 21218, USA e-mail: [email protected]
7
0
0 Abstract. I present a brief survey of surveys, and their results and implications,
2 intended to set thecontext for theensuing discussion.

n
Key words. Galaxy – disk: Galaxy – solar neighbourhood: Galaxy – structure:
a
J Galaxy: abundances– Cosmology: observations

9
2
1. Introduction oneshouldnotforgettheimportanceofthe

1 interstellar medium.
v These are exciting times to study local
2 galaxies,due to the confluence of three ap-
3 proaches: 2. Early Surveys
8 • Advances in technology have allowed
1 2.1. Star Counts
large, high-resolution simulations of struc-
0
ture formation to model Galaxy formation Since the work of Kapteyn in the early
7
0 in a cosmological context 20th centuries, star counts, particularly at
/ • Large observational surveys of stars high Galactic latitude, have been utilised
h
in Local Group galaxies are now possi- todefineGalacticstructure.However,their
p
- ble, using wide-field imagers and multi- shortcomings, when taken alone, have also
o object spectroscopy, complemented by been long known. The apparent magni-
r space-based imaging and spectroscopy,fol- tudedistributionofstarsdepends onmany
t
s lowedinthenearfuturebytheGAIAsatel- factors, not only their density distribu-
a
lite and full phase space information tion–theirluminosityfunctiondependson
:
v •High-redshiftsurveysarenowquanti- metallicity, their birth-rate and the under-
i lying(invariant?)initialmassfunction(see
X fyingthestellarpopulationsandmorpholo-
e.g. Gilmore & Wyse 1987 for a review).
gies of galaxies at high look-back times
r
Refining star counts to include colour al-
a I will here focus on the second approach,
lows more stringent testing of models, but
while acknowledging the synergy with the
again the result is critically dependent
others. I will not attempt a full histor-
on the adopted luminosity function and
ical review, but highlight advances with
colour-magnituderelationfordifferentpop-
what I consider appropriate examples. I
ulations/components in the model (this
will also focus on stellar components, but
may seem obvious, but the use of inappro-
priate choices was the source of much con-
Send offprint requests to: R.Wyse tentious debate in the 1980s).
2 Wyse: Surveysof TheGalaxy
Bearing their limitations in mind, star try remainsaneffective tool,mostrecently
counts in selected lines-of-sight proved ex- illustrated by the work of Nordstro¨m et
tremely useful for delineating the overall al. (2004), providing the definitive analy-
large-scale structure of the stellar com- sis of the metallicity distribution of nearby
ponents of the Galaxy. This is usually F/G dwarfs.
achieved not by direct inversion of the
star counts, but by comparisons of the
2.2. Star Counts plus Kinematics
observations with model predictions (van
den Bergh 1980; Bahcall & Soneira 1980;
Photometric observations repeated after a
Gilmore 1981). In particular, the stellar
sufficientlylongbaselineallowforthecom-
density laws (radial and vertical) of the
binationofstarcountspluspropermotions.
thin disk were derived (Sandage & Katem
The reduced proper motion diagram is a
1977;Bahcall&Soneira1981;Yoshii1982);
useful discriminant of different kinematic
the density profile and shape of the stellar
populations, in the absence of reliable
halo were estimated from a variety of trac-
distances. Chiu (1980) applied this tech-
ers (e.g. Hartwick 1987; Wyse & Gilmore
nique to his database of proper-motions in
1989; Reid & Majewski 1993; Kinman,
Cardinal directions1 for a faint (V ∼< 20.5)
Suntzeff & Kraft 1994); the thick disk was
magnitude-limited sample, based on deep
defined as a component with exponential
photographic plates with a 25 year base-
scale heightsome 3–4times that of the old
line. He concluded that ‘Population I’ far
thin disk (Gilmore & Reid 1983; Fenkart
from the disk plane was of lower metal-
1988; Larsen & Humphreys 2003).
licity, and had higher velocity dispersions,
The interpretation of early star counts
thandidPopulationIlocally.Are-analysis
was complicated by several factors.
of his data, without the requirement that
Degeneracies in reddening–age–metallicity
there be only two components (the classi-
were exacerbated by the fact that the
cal Populations I and II), showedthe pres-
counts were based on photographic
ence of the intermediate-kinematics thick
photometry, in only a limited range of
disk (Wyse & Gilmore 1986).
bandpasses.Poorstar-galaxyseparationat
faint magnitudes (V ∼> 20)cancause prob- Distances derived from photometric
parallaxespluspropermotions,forstarsse-
lems, particularly for blue objects (see the
lected purely as a magnitude-limited sam-
discussion in Reid & Majewski 1993). The
ple, i.e. not kinematically selected, in at
small number of lines-of-sight in any one
least one of the Cardinal directions, can
survey and limited areal coverage further
be used to probe one or more components
made it difficult to isolate the underlying
of the space motion (e.g. Majewski 1992),
cause(s) of discrepancies between different
as can radial velocities (Sandage & Fouts
investigations.
1987b). Reliable distances based on pho-
Determination of metallicities greatly
tometry need good metallicity and grav-
aidstheinterpretation.Thosederivedfrom
ity estimates, which are not always avail-
photometry can of course be obtained for
able. The Hipparcos/Tycho sample (with
more stars with less investment of tele-
trigonometric parallaxes) allowed an anal-
scope time, compared with spectroscopic
ysis of very local space motions, such as
determinations. The broad-band (UBV)
the velocity dispersion tensor as a func-
based metallicity distributions of faint
tion of colour, on the scale of less than a
F/G-dwarfs by Gilmore & Wyse (1985),
hundredparcsecs,ofcourseformostly disk
combined with density laws derived from
stars (e.g. Dehnen & Binney 1998).
star counts, were critical in ascertaining
that the thick disk was indeed a distinct
component.Intermediateandnarrow-band 1 The Galactic poles, the anti-center/center
photometry such as Stro¨mgren photome- line and towards/away from Galactic rotation
Wyse: Surveysof TheGalaxy 3
The combination of star counts plus 3. Motivation for Surveys:
spectroscopy–toprovideametallicityesti- Cosmology
mateinadditiontoradialvelocity–ismore
3.1. Early Surveys
powerful than photometry alone, allowing
joint analyses of metallicity and kinemat-
The idea that the stellar populations of
ics/dynamics. For example, this combina-
the Milky Way Galaxy have critical im-
tion allows a robust dynamical analysis of
portance for understanding largerissues in
local vertical motions to derive the verti-
cosmology has been a major motivation
cal acceleration and associated mass den-
for decades. Much fundamental early work
sity and surface density (K ; Kuijken &
z on stellar populations and Galaxy forma-
Gilmore 1989). The conclusion from this
tion by Sandage (e.g. Eggen, Lynden-Bell
analysis is that there is no ‘extra’ dissipa-
& Sandage 1962; Sandage 1970) was cen-
tive dark matter, confined to the disk, in
tered around the questions of ‘how old are
additional to the dissipationless dark mat-
the oldest stars in the Galaxy, how long
ter in the (dark) halo.
after the Big Bang did galaxies initiate
their collapse, and what was the duration
of that collapse?’ These are crucial in con-
Radial velocities, plus distances and straining the age of the Universe – obvi-
proper motions, allow full phase-space in- ously as least as old as the oldest stars –
vestigations.Earlysurveysusedphotomet- and hence the values of cosmological pa-
ric parallaxes for distances, plus a mini- rameters, particularly when these are esti-
mum proper motion selection criterion to mated through comparison with measure-
define the samples,in orderto increasethe ments of the present value of the Hubble
‘yield’ of non-thin disk stars in these nec- constant.Thefactthattheearly,rapidcol-
essarily localsample. This kinematic selec- lapse model developed in Eggen, Lynden-
tion requiresanunderstandingof, andcor- Bell & Sandage (1962) is still used as a
rection for, the kinematic bias introduced. paradigm for the formation of the Milky
The analysis is complicated, but fruitful Way Galaxy is testament to its simplicity
(Sandage&Fouts1987a;Carney&Latham and power.
1987; Carney, Laird, Latham & Aguilar Significant impetus to use the Milky
1996). Again the Hipparcos/Tycho sample Way as a template for testing theories
provided the opportunity for the deriva- of galaxy evolution also came from in-
tion of space motions for local stars, with- conclusive attempts to derive cosmologi-
out proper-motion selection, once metal- cal parameters from observations of galax-
licities and radial velocities were obtained ies, such as the Hubble diagram (appar-
(Nordstr¨om et al. 2004). ent magnitude vs redshift), galaxy num-
ber counts etc. It was then realised that
the evolution of galaxies must be under-
The modern era of star counts de- stood first (summarised in Tinsley 1977),
rived from very wide-field CCD photome- and the Milky Way was potentially an
trysuchasavailablefromtheSloanDigital ideal testbed. The drive to understand the
Sky Survey (SDSS, e.g. Chen et al. 2001; age distributions and metallicity distribu-
Ivezi´c’s and Newberg’s talks in these pro- tions of the different stellar components
ceedings) was preceded by pencil-beam of the Galaxy led to elegant analyses of
CCD photometry in selected areas, pro- the metallicity distribution of long-lived
viding multi-band, deep, data overseveral, stars, manifest in the local G-dwarf metal-
to many, square degrees (e.g. Phleps et licity distribution (Pagel & Patchett 1975,
al. 2000; Siegel et al. 2002). These pointed van den Bergh 1962; Schmidt 1963) and
to tantalizing inconsistencies in different predictions for the chemical evolution of
fields. the Galaxy beyond the local disk. The
4 Wyse: Surveysof TheGalaxy
simple, closed-box model of chemical evo- The merging history of a typical
lution had been developed (e.g. Schmidt massive-galaxydark halo is fairly straight-
1963; Searle & Sargent 1972) and the ap- forward to calculate, since only gravity is
plication to the local disk revealed a ‘G- involved. However, most simulations lack
dwarf problem’ in that the model signif- the resolution to follow how far inside a
icantly over-predicted the metal-poor tail ‘parent’haloamergingsatellitepenetrates,
of the metallicity distribution of long-lived andthisiscrucialtodeterminetheeffecton
stars. Analytic and numerical models of the baryonic disk. During mergers the or-
chemicalevolutionshowedtheseveralways bital energy goes into the internal degrees
in which the G-dwarf ‘problem’ could be of freedom of the merging systems, thus
solved(e.g.Tinsley1975;Pagel&Patchett ‘heating’ them. A corollary is that surviv-
1975). These models included such cur- ingdarksubstructure,aspredictedinCDM
rently topical aspects as inhomogeneities, simulations (Moore et al. 1999; Klypin et
and the interpretation of the pattern of al. 1999), can also heat thin disks. Thin
elemental abundances, showing how they disks are thus fattened, and while gas can
trace the paststellar Initial Mass Function cool and re-settle to the disk plane, stellar
and star formation history (Tinsley 1976; disksremain‘hot’.Duringa‘minor’merger
1979). The data did not merit a full ex- (mass ratio of less than ∼ 1 : 4), the (rel-
ploitation of these insights. atively) low density, outer regions of the
smaller system are removedby tides, to be
absorbed into the larger system. Orbital
3.2. The Modern Era
angular momentum is also absorbed and
redistributed, with in general outer parts
SignificantmotivationforstudyofGalactic
gainingangularmomentunandinnerparts
populationsstillcomesfromcosmology,but
losing. In the process gas and stars are
now not so much the estimation of the
driven to the center, perhaps helped by
present values of cosmological parameters
a bar that is often predicted to form as
such as the deceleration parameter q0 or
a result of instabilities in the disk. The
HubbleconstantH0,butthetestingofpre-
diskformedsubsequentlyhasashortscale-
dictions of galaxy formation in the con-
length: the corollary is that detailed angu-
textofparticularcosmologicalmodels.The
favouredmodelatpresentisΛCDM(ΩΛ ≃ lar momentum conservation is required in
0.7, Ωmatter ≃ 0.3, H0 ≃ 70 km/s/Mpc), order to form extended disks as observed
(Fall & Efstathiou 1980). Various schemes
based on the excellent agreement of its
have been developed to suppress angular
predictions with measurements of large-
momentum transport and redistribution,
scale structure, such as the fluctuations in
usuallyinvokingsome‘feedback’processto
the cosmicmicrowavebackground(Spergel
maintain the baryons in a diffuse gaseous
et al. 2006) and the galaxy power spec-
state for as long as possible (e.g. Weil, Eke
trum (Sanchez et al. 2006; Eisenstein et
& Efstathiou 1998; Maller & Dekel 2002;
al. 2005). As is well-known, such a model
Robertson et al. 2006), until the epoch of
predicts that large galaxies such as the
active (major) merging is complete, per-
Milky Way form and evolve through the
hapsevenasrecentlyasaredshiftofunity.
merging and accretion of smaller systems,
withthe‘firstobjects’havingamassofper-
haps ∼ 106 M⊙ (the characteristic mass, Abadietal.(2003)presentarecentsim-
andtherelationoftheseobjectstopresent- ulation of the formation of a present-day
day dwarf galaxies is the subject of much disk galaxythat demonstratesmany ofthe
on-going work).As is also well-known,this important aspects,including the outstand-
model faces severalchallenges,particularly ing problem of how to include star forma-
concerning its predictions on the scales of tion and gas physics. Generic predictions
groups of galaxies and below. for disk galaxies include the following:
Wyse: Surveysof TheGalaxy 5
– Extended disks settle and form late, lookbacktimesequivalenttoredshiftsof1.5
after the last major mergers, typically and greater (see Figure 1).
(for a dark halo of mass 1012M⊙) cor-
responding to a redshift of unity (e.g.
The clues to galaxy evolution that one
Maller et al. 2006), or a lookback time
of ∼8 Gyr might wish to extract from the local fossil
record include the star formation history,
– A large disk galaxy should have hun-
theformofthestellarInitialMassFunction
dreds of surviving satellite dark haloes
and whether or not it varied between then
at the present day; these may well
and now, chemical evolution, and the rela-
provide observable signatures through
tive importances of dissipative gas physics
theirgravitationalinteractionswiththe
versus dissipationless processes. The over-
baryonic galaxy, such as heating of the
alldarkhalopotentialwelldepthandshape
thin disk, disruption of wide binaries,
canbe inferredfromstellar (andgas)kine-
disturbance of extended tidal streams
matics.Thereareaspectsofthestellarpop-
etc.
ulations that are less sensitive to details of
– The stellar halo is formed from dis-
baryonicphysics–suchasthe agesofstars
rupted satellite galaxies
– Minor mergers (a mass ratio of ∼ 20% in the thick disk – and these can be used
to constrain the merging history – is this
between the satellite and the disk) into
compatiblewithΛCDM?IstheMilkyWay
a disk continue after the last ‘major
a typical galaxy?
merger’, and heat it, forming a thick
disk outofapre-existingthin disk,and
create torques that drive gas into the Most galaxies in the local Universe are
central (bulge?) regions observed to cluster in loose groups like the
– More significant mergers transform a Local Group (which in itself is unusual
disk galaxy into an S0 or even an el- in CDM models, Governato et al. 1997).
liptical While lacking a giant elliptical, the Local
– Subsequent accretion of gas can reform Group hosts a reasonably diverse selec-
a thin disk tion, with large disk galaxies of a range
– Stars canbe accretedinto the thin disk of bulge-to-disk ratios (The Milky Way,
from suitably massive satellites (dy- M31,M33),gas-richandgas-poorsatellites
namical friction must be efficient) and ranging from the compact elliptical M32
if to masquerade as stars formed in the through the numerous extremely low sur-
thin disk, must be on suitable high an- face brightness dwarf spheroidals (dSph).
gular momentum, prograde orbits Do trends in inferred merging history etc
forthe LocalGroupgalaxiesmatchpredic-
tions?
3.3. The Fossil Record: Tests of
Predictions
We can address these questions with
current and planned capabilities, with
Stars of mass like the Sun, and lower,
which the motions, spatial distributions,
live for essentially the age of the Universe,
ages and chemical elemental compositions
and retain memory of many aspects of
canbemeasured(withvaryingaccuracies!)
the conditions at early times. Studying old
for individual stars in galaxies throughout
stars nearby thus allows us to study cos-
the Local Group, plus additional comple-
mology locally, a very complementary ap-
mentary tracers such as HII regions and
proach to direct study of high-redshift ob-
planetary nebulae.
jects. There are copious numbers of stars
in Local Group galaxies that have ages of
greater than 10 Gyr, and thus formed at What have we learnt so far?
6 Wyse: Surveysof TheGalaxy
4. Milky Way Large Scale Structure
4.1. The Thin Disk
The large-scale structure of the thin stel-
lar disk is reasonably well modelled by
a double exponential with scalelength of
∼ 3 kpc and scaleheight of ∼ 300 pc (for
stars older than a few Gyr). Extrapolating
this smooth structure with a local nor-
malization for stellar surface density of
Σ∗ ∼ 35M⊙ pc−2 (Kuijken & Gilmore
1989; Flynn et al. 2006) gives a total mass
of around 6 × 1010M⊙. The interstellar
medium contributes ∼ 10M⊙ pc−2 locally, Fig.1. Plot of lookback time versus redshift
for two cosmological models, as labelled, each
and has a rather different radial profile
with present value of the Hubble constant of
from the stars, with atomic and molecular
70km/s/Mpc.Thehorizontallinesindicatees-
gas each having a distinct spatial distribu-
timates of the ages of the oldest stars in the
tion.
thin disk, from Binney et al. (2000).
Stellarmetallicityandagedistributions
are best-known at present for the local
disk, within around one kpc of the so-
lar circle. As noted above, Stro¨mgren pho-
tometry has proven a robust technique of
There is certainly no lack of old stars in
metallicity determination for large sam-
the local thin disk, a location that is some
ples of F/G dwarfs, confirming the ‘G-
3 scalelengths from the Galactic center.
dwarf problem’ in the local disk i.e. a
Assuming these stars formed in the thin
narrow metallicity distribution, with few
disk, one concludes that an extended thin
stars significantly more metal-poor than
disk was in place at a redshift of around
the peak, in contradiction to the large
2 (corresponding to the look-back time of
metal-poor tail predicted by the simplest
10–12 Gyr estimated for the onset of star
chemical evolution models (e.g. Wyse &
formationinthelocaldisk;Binney,Dehnen
Gilmore1995;Rocha-Pinto&Maciel1996;
& Bertelli 2000). This is significantly ear-
Nordstro¨met al.2004).The peak metallic-
lierthanatypicalextendedthindiskwould
ity of long-lived stars in the solar neigh-
form in CDM models.
bourhood is somewhat below the solar
value, ∼ −0.15 dex, with good agree- Complementary data for external disk
ment between G-dwarfs and lower-mass galaxiesofsimilarscale-lengthtotheMilky
K-dwarfs (Kotoneva et al. 2002). High- Way (half-light radii of between 5kpc and
resolution spectroscopic studies of neces- 7kpc) show little evolution in size or num-
sarily smallersamples providesa peak iron ber back to a redshift of unity (the limit
abundance of ∼ −0.1 dex (Allende-Prieto of the data, for the COSMOS survey sam-
et al. 2004). pleofSargentetal.2006).Hence,extended
The star formation history of the local disksdonotseemtostart formingatz ∼1,
stellardiskhasbeenestimatedthroughvar- but rather to be well-established by then.
ioustechniques,andthegeneralconclusion Modelsinwhichasignificantpartoftheold
isforanearlyonset,anapproximatelycon- thin stellar disk is formed by the later ad-
stant overall rate, and with low-amplitude dition of old stars by satellite accretion di-
(factor of two) bursts on (few?) Gyr rectly into high-angular momentum orbits
timescales (e.g. Hernandez, Valls-Gabaud in the disk plane (e.g. Abadi et al. 2003)
& Gilmore 2000; Rocha-Pinto et al. 2000). need to address this.
Wyse: Surveysof TheGalaxy 7
4.2. The Thick Disk of these two components, the thick disk
having a (significantly) shorter duration of
The large-scale structure of the thick stel- star formation. A well-defined separation
lardiskis(probably!)reasonablywellmod- of populations on the basis of elemental
elled by a double exponential, with scale- abundancepatternsholdsmuchpromisefor
length of ∼ 3 kpc and scaleheight of ∼ identification of substructure and tracing
1 kpc, giving an axial ratio that is a factor the history of the Galaxy (cf. Freeman &
of three or so ‘fatter’ than the thin disk. Bland-Hawthorn 2002).
Extrapolating this smooth structure with
The thick disk has kinematics that are
a local normalization of around 5% gives
intermediatebetweenthoseofthethindisk
a total mass of around 15% of that of the
and stellar halo: a typical local thick disk
thin stellar disk.
star is on a fairly high angular momen-
Again the metallicity and age distribu- tum orbit, with a lag behind the mean az-
tions are best determined at present only imuthal(rotational)velocityoftheoldthin
fairly locally, within a few kpc of the Sun, diskofonly∼30−50km/s.Theverticalve-
bothverticallyandradially.Thetailsofthe locity dispersion is around45 km/s, hotter
derived kinematic and metallicity distribu- than can be achieved through heating the
tions overlap with those of the thin disk, thin disk by local gravitational perturba-
so there is the danger in very local sam- tions such as Giant Molecular Cloud com-
ples of being overwhelmed by the much plexes and/or transient spiral arms.
more numerous thin disk stars. Defining
The dominant old age of stars in the
a thick disk sample in situ, for example thick disk, ∼ 11 Gyr, combined with the
above ∼ 1 kpc vertically from the disk largeagerangeofstarsinthethindisk,ar-
plane,providesaneffectivefilter.Suchsam-
guesagainstmodelsinwhichthethickdisk
ples find a thick disk peak metallicity of
forms from the thin disk by a heating pro-
∼ −0.6 dex, and that essentially all the cessthatoccursoveranextendedperiod.If
stars have an age as old as the globular
the heating is merger-induced (the minor-
cluster ofthe same metallicity asthe stars,
merger scenario for formation of the thick
or some 10-12 Gyr (e.g. Gilmore & Wyse
diskfromapre-existingthindisk),thenthe
1985;Gilmore,Wyse&Jones1995;seealso
last significant merger into the thin disk
Ratnatunga & Freeman 1989; Morrison, waslongago,ataredshift∼2,correspond-
Flynn & Freeman 1990). Local proper- ingto alookbacktime of∼11Gyr.This is
motionselectedsamplesfindsimilarresults
unusually long ago in ΛCDM models, par-
(e.g.Carney,Latham&Laird1989),albeit
ticularly when one remembers that for suf-
withalternativeinterpretations(e.g.Norris
ficient heating of the thin disk, the ‘signifi-
& Ryan 1991).
cant’ merger need only have mass equal to
The extension of the metallicity distri- 20% of the disk mass, not the total mass.
butionofthe thick disktometallicities sig- In any merger-model for the formation
nificantly below ∼ −1 dex (e.g. Morrison, of the thick disk, there will be a contri-
Flynn & Freeman 1990) remains topical, bution to ‘the thick disk’ from stars re-
with its most robust detection in samples movedfromthe culpritsatellite(s).Indeed,
selectedbyhavinglowmetallicity,andthus in some models tidal debris from shredded
with uncertain normalization to the main satellite galaxies is a very significant part
peak (e.g. Chiba & Beers 2000). of ‘the thick disk’ (e.g. Abadi et al. 2003).
As discussed in Sofia Felzing’s contri- However, the high peak metallicity of the
bution, the stars in the local thick disk thick disk stars suggests that these stars
follow a distinct elemental abundance pat- formed within a fairly deep potential well,
tern, offset from stars in the thin disk (de- particularly given their old age; as an ex-
fined kinematically). This presumably re- ample, while the LMC has managed to
flects the different star formation histories self-enrichto asimilarmetallicity,[Fe/H]∼
8 Wyse: Surveysof TheGalaxy
−0.6 dex, this is for stars only a few Gyr & Mack 1998). The total stellar mass of
old. The putative satellites in which the the bulgeis∼1010 M⊙,andthecentralre-
majority of thick disk stars formed would gionsareverybaryon-dominated(Bissantz,
have to be extremely different from those Debattista & Gerhard 2004).
surviving satellites.
Thestellarpopulationshavebeenstud-
Theevidencefromobservationsofhigh-
ied mostly in ‘windows’ of low optical ex-
redshift systems is limited, but there has
tinction. The peak spectroscopic metal-
been a recent detection of what appears to
licity from samples of K-giants is some-
be a kinematically hot (i.e. expected to be
whatbelowthesolarvalue(e.g.McWilliam
thick)stellardiskforminginaburstofstar
& Rich 1994; Ibata & Gilmore 1995;
formation at a redshift of greater than 2
Fulbright, McWilliam & Rich 2006a), sim-
(Genzel et al. 2006). More nearby galax-
ilar to the long-lived stars in the local thin
ies too appear to contain old thick disks
disk. The dominant age is old, again 10-
(Mould 2005;Yoachim& Dalcanton2005),
12Gyr,withyoungerstarsinlowerlatitude
not dissimilar to that of the Milky Way.
central regions (e.g. Ortolani et al. 1995;
Identifying the analogue (if one exists)
Feltzing & Gilmore 2000; Kuijken & Rich
of the Milky Way thick disk in M31 is
2002; van Loon et al. 2003). The coinci-
complex, due in part to the pervasive in-
dence in age with the thick disk may point
homogeneities in stellar surface densities
toonemergereventtosetthephysicalcon-
(Ferguson et al. 2002) and disparate lines-
ditionsforbothcomponents:thegasdriven
of-sightwithspectroscopicanddeepphoto-
tothecentreduringthemergerthatheated
metric information. Is the ‘spheroid’ com-
the thin disk to form the thick disk, would
ponent with [Fe/H]∼ −0.6 the thick disk
form the bulge (e.g. Wyse 2001).
in M31 (e.g. Wyse & Gilmore 1988)? Or
is it more associated with the outer disk Determinations of elemental abun-
(Brown et al. 2006), and contains stars of dances are limited to small samples of
a wide range of ages, thereby compatible K-giants (typically around 50 stars), and
with a more extended merger history than are consistent with enrichment by (normal
the Milky Way? IMF)TypeIIsupernovaeonly,i.e.thestars
The lower-mass spiral galaxy M33 ap- formed in only a short duration of star
pearstohavehadaveryquiescentlife,with formation (e.g. Fulbright, McWilliam &
little evidence forsignificantmergersorin- Rich2006b;Zoccalietal.2006).Thisshort
teractions. While a trend between merg- duration agrees with earlier inferences
ing history and total mass is expected in from more limited data (e.g. Matteucci &
ΛCDM, such that lower mass dark haloes Brocato1990;Rich1999;Ferreras,Wyse&
have fewer recent mergers (e.g. Maller et Silk 2003).
al.2006),itremainstobe seeniftheMilky The low-mass end of the IMF in the
Way,M31andM33canbeproducedeasily. bulge can be studied by direct star counts,
with the result (Zoccali et al. 2000) that it
is indistinguishable from that in (dynami-
4.3. The Central Bulge
callyunevolved)metal-poorglobulars.The
The smooth structure of the central bulge low-mass IMF in the Ursa Minor dwarf
is mildly triaxial, i.e. barred, with axial Spheroidalgalaxy(Wyseetal.2002)isalso
ratios of ∼ 1 : 0.35 : 0.3 (Bissantz & indistinguishable from that in metal-poor
Gerhard 2002). The profile is reasonably globular clusters, and again the elemen-
well-fit by anexponential, with scaleheight tal abundances in dSph do not require any
∼300pc,andthus the Milky Way bulge is variations in massive-star IMF. The IMF
not a classical ‘r1/4-bulge’ but rather per- seems remarkably invariant with metallic-
haps a‘pseudo-bulge’,oftenfoundinlater- ity, epoch of star formation, (present) stel-
type spiral galaxies (e.g. Carollo, Stiavelli lar density etc.
Wyse: Surveysof TheGalaxy 9
Models of how the bulge in the Milky halo stars with extremely high velocities,
Way formed generally appeal either to its probing the outer halo, have lower values
‘pseudo-bulge’ density profile and triaxial of [α/Fe], more similar to the stars in the
shape to argue for an instability in the in- dSph, but the overall abundance pattern
nerdisk(inwhichcasetheformationofthe remain different (Fulbright 2002).
bulgecouldhaveoccuredsignificantlyafter
theformationofthestarsthemselves),orto The abundance ratios of lighter met-
itsrapidenrichmentandhighdensitytoar- als in the field halo stars show remarkably
gueforanin situstarburst(e.g.Elmegreen little scatter down to the lowest metallici-
1999; see the review in Wyse 1999). ties (e.g.Cayrelet al.2004),defining a flat
‘Type II plateau’ in [α/Fe] and indicating
that the stars formed in systems with only
4.4. The Stellar Halo
a short duration of star formation, allow-
The large-scale structure of the inner re- ingenrichmentbyonlythe short-livedpro-
gions, within ∼ 15 kpc of the Galactic genitors of Type II supernovae. The value
center is the best-constrained at present. of a predicted ‘Type II plateau’ in [α/Fe]
Thedominantpopulationisoldandmetal- depends on the massive-star IMF (see e.g.
poor, with the stars on low angular mo- Wyse&Gilmore1992),andthelowampli-
mentum orbits. The overall density profile tude of scatter indicates an invariant IMF.
(tracedbyRRLyraestars)showsasmooth
power-law fall-off with distance (measured Themeanmetallicity ofthe stellarhalo
in the disk plane) of ρ ∝ R−3.1 out to is around −1.5 dex (e.g. Ryan & Norris
RRL
∼ 50 kpc (Vivas & Zinn 2006). The stel- 1991), significantly lower than the local
lar halo as traced by main sequence F/G (gas-rich) disk. With a fixed stellar initial
stars and RR Lyrae stars is not spherical, mass function, and no gas flows, one ex-
but can be reasonablywell fit by an oblate pects a system of low gas fraction, such
spheroid, with flattening at around the so- as the stellar halo, to be more chemically
lar distance of c/a ∼ 0.5 (Hartwick 1987; evolvedthanasystemwithhighergasfrac-
Wyse & Gilmore 1989) becoming rounder tion. Hartwick (1976) provided an elegant
withdistance,andapproximatelyspherical explanation to this conundrum: gas out-
at R∼> 20 kpc (Vivas & Zinn 2006). The flows from active star-forming regions in
total stellar mass in this smooth distribu- the proto-halo. The chemical evolution re-
tion is ∼2×109 M⊙ (e.g. Carney,Latham quirements are such that for a fixed stellar
& Laird 1990).Hints of triaxiality are seen IMF, one that matches the local thin disk
in deep imaging data, as discussed in the mean metallicity of just below the solar
meeting by Heidi Newberg. value,theoutflowsmustoccurataround10
Both the age distributions (Unavane, timestherateofstarformation.Anattrac-
Wyse & Gilmore 1996) and the elemen- tivecorollarytothispictureisthatonecan
tal abundance patterns (Fulbright 2002; tiethegasoutflowfromlow-masshalostar-
Stephens & Boesgaard 2002; Tolstoy et formingregionsto gasinflow tothe central
al. 2003; Venn et al. 2004) of the bulk regions to form the bulge; the low angular
of the field halo stars are very different momentum of halo material means that it
from those in the present satellite galax- will only come into centrifugal equilibrium
ies of the Milky Way (with the abundance after collapsing in radius by a significant
pattern consistent with the expectations factor. The mass ratio of bulge to halo is
from the extended star formation histo- around a factor of ten, just as would be
ries). Accretion of stars from systems like expected, and the specific angular momen-
the satellite galaxies, into the field halo, is tum distributions of stellar halo and bulge
limited to less than 10% since a redshift match(Wyse&Gilmore1992;seeFigure2
of unity (Unavane et al. 1996). The rare here).
10 Wyse: Surveysof TheGalaxy
eral (of order 10) candidate new dSph,
globularclusters and streams(including at
least one stream from a disrupting glob-
ular cluster – not all streams indicate
accretion from an external source) have
been announced this year, all exploiting
the SloanDigital Sky Survey imaging data
(e.g. Belokurov et al. 2006a,b; Grillmair
2006). Determining the masses of the new
putative satellite galaxies is crucial before
one can say what impact these discoveries
have on the ‘satellite problem’ of ΛCDM
models – namely the over-prediction by
Fig.2. AdaptedfromWyse&Gilmore1992, the models of satellite-galaxy mass dark
their Figure 1. Angular momentum distribu- haloes. At present, the available radial ve-
tions of the bulge (solid curve), the stellar locity data internal to the dSph compan-
halo (short-dashed/dotted curve), the thick ions of the Milky Way are consistent with
disk (long-dashed/dotted curve) and the thin each dSph being embedded in a dark halo
disk(long-dashedcurve).Thebulgeandstellar of fixed mass, of ∼4×107 M⊙ (Wilkinson
halohavesimilardistributions,asdothethick
et al. 2006), out to the extent of the stars
and thin disks.
(saying nothing about the extent, or mass,
at radii beyond available stellar kinematic
5. Small-Scale Structure data).Whileitisreasonablyeasy,theoreti-
cally,tomodify thebaryoncontentofshal-
5.1. The Outer Stellar Halo
lowpotentialwellsystemssuchasthesatel-
litehaloesandthuschangetheirluminosity
The outer halo, with dynamical timescales
function, it is much harder to conceive of
of > 1 Gyr, is the best location to search
modifications to the predicted mass func-
for structure. Indeed, several streams have
tion of satellite haloes to match a narrow
been found, in both coordinate space and
range in mass.
in kinematics. Most of these appear to be
due to the Sagittarius Dwarf Spheroidal
(Sgr dSph), a galaxy that was discovered 5.2. The Thick Disk
serendipitouslyduringasurveyofthekine-
maticsandmetallicitydistributionsofstars As noted above, a popular model for the
in the Milky Way bulge (Ibata, Gilmore & formation of the thick disk is based on a
Irwin 1994, 1995). Extended tidal streams minor merger into a pre-existing thin disk,
from the Sgr dSph are now detected across with the orbital energy in large part go-
the sky (Ibata et al. 2001; Majewski et ing into heating the thin disk. The satel-
al.2003;Belokurovetal.2006a).Theseare lite galaxy (or galaxies)responsible for the
potentially extremely useful in constrain- heating does not survive unscathed, but
ing the shape and smoothness of the dark will in general be tidally disrupted, the
halo potential, although we are in the in- mass of any surviving remnant being set
teresting situation of contradictoryconclu- by how deeply it penetrates and the rela-
sions from different datasets (e.g. Helmi tive density compared that of to the larger
2004; Johnston, Law & Majewski, 2005; (Milky Way) galaxy. The ‘shredded satel-
Fellhauer et al. 2006). lite’starswilllikelycontributetothestellar
The detection and characterization of populations above the thin disk plane, and
structure in the stellar halo is a very thus be included in samples of ‘thick disk’
fast-moving field! As described further in stars. The spatial distribution of satellite
Heidi Newberg’s talk in this volume, sev- debris will reflect the orbit of the satellite,