phil. trans. r. soc. lond. a-2000-heckman-2077-91

7/31/2019 Phil. Trans. R. Soc. Lond. a-2000-Heckman-2077-91

1/16

doi: 10.1098/rsta.2000.0631, 2077-20913582000Phil. Trans. R. Soc. Lond. A

Timothy Heckmangalaxies and active galactic nucleiThe role of starbursts in the formation of

Rapid response

/roypta;358/1772/2077http://rsta.royalsocietypublishing.org/letters/submit

Respond to this article

Email alerting serviceherethe article or click

ofarticle - sign up in the box at the top right-hand cornerReceive free email alerts when new articles cite this

http://rsta.royalsocietypublishing.org/subscriptionsgo to:Phil. Trans. R. Soc. Lond. ATo subscribe to

This journal is 2000 The Royal Society

on October 20, 2010rsta.royalsocietypublishing.orgDownloaded from
http://rsta.royalsocietypublishing.org/letters/submit/roypta;358/1772/2077http://rsta.royalsocietypublishing.org/cgi/alerts/ctalert?alertType=citedby&addAlert=cited_by&saveAlert=no&cited_by_criteria_resid=roypta;358/1772/2077&return_type=article&return_url=http://rsta.royalsocietypublishing.org/content/358/1772/2077.full.pdfhttp://rsta.royalsocietypublishing.org/cgi/alerts/ctalert?alertType=citedby&addAlert=cited_by&saveAlert=no&cited_by_criteria_resid=roypta;358/1772/2077&return_type=article&return_url=http://rsta.royalsocietypublishing.org/content/358/1772/2077.full.pdfhttp://rsta.royalsocietypublishing.org/cgi/alerts/ctalert?alertType=citedby&addAlert=cited_by&saveAlert=no&cited_by_criteria_resid=roypta;358/1772/2077&return_type=article&return_url=http://rsta.royalsocietypublishing.org/content/358/1772/2077.full.pdfhttp://rsta.royalsocietypublishing.org/subscriptionshttp://rsta.royalsocietypublishing.org/subscriptionshttp://rsta.royalsocietypublishing.org/subscriptionshttp://rsta.royalsocietypublishing.org/subscriptionshttp://rsta.royalsocietypublishing.org/subscriptionshttp://rsta.royalsocietypublishing.org/http://rsta.royalsocietypublishing.org/http://rsta.royalsocietypublishing.org/http://rsta.royalsocietypublishing.org/subscriptionshttp://rsta.royalsocietypublishing.org/cgi/alerts/ctalert?alertType=citedby&addAlert=cited_by&saveAlert=no&cited_by_criteria_resid=roypta;358/1772/2077&return_type=article&return_url=http://rsta.royalsocietypublishing.org/content/358/1772/2077.full.pdfhttp://rsta.royalsocietypublishing.org/letters/submit/roypta;358/1772/2077


2/16

10.1098/rsta.2000.0631

The role of starbursts in the formation of

galaxies and active galactic nuclei

B y T i m o t h y M. H e c k m a n

Department of Physics and Astronomy, Johns Hopkins University,Baltimore, MD 21028, USA

Starbursts are episodes of intense star formation in the central regions of galaxies,and are the sites of ca. 25% of the high-mass star formation in the local Universe. Inthis contribution, I review the role that starbursts play in the formation and evolu-

tion of galaxies, the intergalactic medium (IGM), and active galactic nuclei. First,I point out the empirical similarities between local starbursts and the Lyman-breakpopulation at high redshift, and emphasize the implied similarities in their basic phys-ical, dynamical and chemical properties. In the local Universe, more-massive galaxieshost more-luminous, more-metal-rich, and dustier (IR-dominated) starbursts. Thisunderscores the need for a pan-chromatic approach to documenting and understand-ing the cosmic history of star formation. Second, I review the systematic propertiesof starburst-driven galactic superwinds. These drive metal-rich dusty gas outward ata typical velocity of 400{800 km s1 (independent of the galaxy rotation speed) and

at several times the star-formation rate. They can be directly observed both in localstarbursts and high-redshift galaxies. They are probably responsible for establish-ing the strong mass{metallicity relation in spheroids, and for the metal-enrichmentand (pre)heating of the IGM. They may have also ejected cosmologically signi-cant amounts of intergalactic dust. Third, I discuss UV observations of the nucleiof type 2 Seyfert galaxies. These show that compact (on a scale of a few hundredpc) heavily reddened starbursts are the source of most of the `featureless contin-uum in UV-bright Seyfert 2 nuclei, and are an energetically signicant componentin these objects. Finally, I discuss the evolution of the host galaxies of radio-quiet

quasars. Rest-frame optical images imply that the hosts at z 2 are only as lumi-nous as present-day L galaxies, less massive than the hosts of similarly luminouslow-z quasars, similar to the Lyman-break galaxies, and much less luminous thanpowerful radio galaxies at the same redshift. These results are consistent with theidea of hierarchical galaxy assembly, and suggest that super-massive black holes maybe formed/fed before their host galaxy is fully assembled.

Keywords: starbursts; galactic winds; Seyfert galaxies; quasars

1. Introduction

Starbursts are short-lived episodes of intense star formation that usually occur in the`circum-nuclear (kpc scale) regions of galaxies, and dominate the integrated emissionfrom the `host galaxy (Leitherer et al. 1991). Starbursts are major components ofthe local Universe (see, for example, Gallego et al. 1995; Soifer et al. 1987), and

Phil. Trans. R. Soc. Lond. A (2000) 358, 2077{2091

2077

c 2000 The Royal Society

http://rsta.royalsocietypublishing.org/http://rsta.royalsocietypublishing.org/http://rsta.royalsocietypublishing.org/


3/16

2078 T. M. Heckman

are the sites of ca. 25% of the total (high-mass) star formation in the local Universe(Heckman 1997).

The cosmological relevance of starbursts has been dramatically underscored bythe spectacular discovery of populations of high-redshift (z > 2) star-forming eldgalaxies selected by their rest-frame UV continuum emission (Steidel et al. 1999;

Lowenthal et al. 1997), rest-frame far-infrared emission (see, for example, Hugheset al. 1998; Barger et al. 1999; Cowie, this issue), Ly emission (Hu et al. 1998),and H emission (Teplitz et al. 1998; Mannucci et al. 1998). The comoving spacedensity and luminosity of these galaxies imply that they almost certainly representprecursors of typical present-day galaxies and are responsible for the production of asignicant fraction of the stars and heavy elements in the present-day Universe (see,for example, Madau et al. 1996; Blain et al. 1998; Calzetti & Heckman 1999).

Starbursts may also play a vital energetic or evolutionary role in active galacticnuclei (AGN). There have been recurring suggestions to this eect in the Seyfert-

galaxy phenomenon (see, for example, Weedman 1983; Perry & Dyson 1985; Terlevich& Melnick 1998; Norman & Scoville 1988; Cid Fernandes & Terlevich 1995). On amore global scale, the rough proportionality between the mass of a super-massiveblack hole and that of the stellar spheroid in which it now resides (Magorrian etal. 1998; Van der Marel 1999) strongly suggests that the quasar phenomenon is anintimate part of the formation or early evolution of massive ellipticals and bulges.

In this contribution, I will therefore discuss the relevance of local starbursts tounderstanding the high-redshift Universe and its major baryonic components: star-forming galaxies, AGN and the intergalactic medium (IGM). In x 2 I will argue thatstarbursts are the only local analogues to the star-forming galaxies observed at highredshift, and will summarize some of the inferences that follow. In x 3 I will describethe systematic properties of starburst-driven galactic superwinds and summarize thecosmological implications of these outows. In x 4 I will present a status report oneorts to understand the energetic/evolutionary signicance of young stars in/nearthe nuclei of Seyfert galaxies. Then, in x 5 I will report new results on the propertiesof the host galaxies of high-redshift radio-quiet quasars, and compare these withtheoretical expectations.

2. Starbursts as analogues to high-z galaxies

Starburst galaxies are the only plausible local analogues to the population of star-forming galaxies at high redshift. Meurer et al. (1997) showed that local starburstsand high-redshift Lyman-break galaxies (LBGs) have similar rest-frame UV surfacebrightnesses and UV colours. Using the empirical correlation between UV colourand extinction for local starbursts, the implied bolometric surface brightnesses ofthe high-z galaxies are, thus, also very similar to local starbursts: typically ca. 1010{1011L- kpc

2. The high-redshift LBGs appear to be `scaled-up (larger and more-luminous) versions of local starbursts. The intrinsic UV surface brightnesses are

roughly three orders of magnitude higher than the discs of normal spirals, and indeednormal spirals at z 3 (if they exist) would be virtually undetectable in HubbleSpace Telescope (HST) rest-UV images due to their low surface brightness (see, forexample, Hibbard & Vacca 1997; Lanzetta et al. 1999).

The similarity in the surface brightnesses of local starbursts and LBGs immediatelyimplies that there are also strong similarities in their basic physical properties. A high

Phil. Trans. R. Soc. Lond. A (2000)



4/16

Starbursts, galaxy formation and active galactic nuclei 2079

UV surface brightness implies a high star-formation rate (SFR) per unit area (S FR ),and, thus, high surface mass densities in the stars () and in the interstellar gas (g)that fuels the star formation. A typical case would have S FR 10M- yr

1 kpc2

and g 109M- kpc

2. These are roughly 103 (S FR ), 102 (g) and 10

1 ()times larger than the corresponding values in the discs of normal galaxies.

The basic physical and dynamical properties of starbursts and the Lyman-breakobjects follow directly from the above. A gas surface mass density of 109M- kpc2

corresponds to an extinction of AB 102 for a Milky Way dust-to-gas ratio. The

characteristic dynamical time in the star-forming region will be short:

t d yn (G)1=2 (GtotH)

1=2 a few Myr;

where H 102 pc is the thickness of the disc. A surface brightness of a few 1010L- kpc

2 corresponds to a radiant energy density inside the star-forming regionthat is roughly 103 times the value in the interstellar medium (ISM) of the Milky

Way. Finally, simple considerations of hydrostatic equilibrium imply correspond-ingly high total pressures in the ISM: P Ggtot a few 109 dyn cm2

(P=k a few 107 K cm3, or several thousand times the value in the local ISM inthe Milky Way). The rate of mechanical energy deposition (supernova heating) perunit volume is also 103 or 104 times higher than in the ISM of our Galaxy.

To summarize, the strong empirical similarity between local starbursts and thehigh-z LBGs implies strong similarity in their basic physical properties. The conclu-sion seems inescapable: if we want to understand the LBGs in the early Universe,we need to understand local starbursts.

As discussed by Heckman et al. (1998), local starbursts occupy a very small frac-tional volume in the multi-dimensional manifold dened by such fundamental param-eters as the extinction, metallicity and vacuum-UV line strengths (both stellar andinterstellar) of the starburst and the rotation speed (mass) and absolute magni-tude of the starbursts `host galaxy. In particular, more-massive galaxies host more-luminous, more-metal-rich and dustier (more-heavily extincted) starbursts. Thereare simple physical explanations for these trends. Firstly, simple considerations ofcausality for a self-gravitating system with a gas mass Mgas = fgasMtot imply thatthe maximum possible star-formation rate is given by the conversion of all the gas intostars in one crossing time: SFRm ax Mgas =td yn fgas

3=G. Thus, more-massivegalaxies (with larger velocity dispersions ) can sustain bursts with higher star-formation rates and, therefore, larger luminosities. The physical basis for the strongobserved correlation between galaxy mass and metallicity will be discussed in x 3below. Finally, the trend for more-metal-rich starbursts to be more-heavily extinctedfollows, provided that neither the gas-column density towards the starburst nor thefraction of interstellar metals locked into dust grains are strong inverse functions ofmetallicity.

The result of these correlations is that a UV census of the local Universe wouldnot only underestimate the true star-formation rate, it would also systematically

under-represent the most-powerful, most-metal-rich starbursts occurring in the most-massive galaxies. This eect can clearly be seen in the recent comparison of thevacuum-UV and far-IR galaxy luminosity functions at low redshift by Buat & Bur-garella (1998). Estimates of star-formation rates at high redshift based on rest-framevacuum-UV sample selection probably suer the same bias, and, thus, may under-represent the ultra-luminous progenitors of the most-massive present-day spheroids.




5/16

2080 T. M. Heckman

It is tempting to speculate (based on the empirical properties of local starbursts) thatthe Submillimetre Common User Bolometer Array (SCUBA) far-IR-selected sourcesat high z may represent just such objects. This speculation is consistent with therelative space densities of the most-luminous LBGs and the SCUBA sources (Meureret al. 1999).

Clearly, we cannot understand the cosmic evolution of the star-formation rate andthe formation and evolution of galaxies without a panchromatic approach.

3. Starburst-driven superwinds

Over the last few years, observations have provided convincing evidence of the exis-tence (and even the ubiquity) of `superwinds: galactic-scale outows of gas drivenby the collective eect of multiple supernovae and stellar winds in a starburst (see,for example, Lehnert & Heckman 1996; Dahlem et al. 1998; Martin 1999).

In this section I will summarize the systematic properties of superwinds as gleanedfrom the analysis of their X-ray emission and their UV{optical absorption lines. TheX-ray data have proven particularly important, since they are the only direct probeof the energetically dominant `piston of hot gas that drives the ow. The interstellarabsorption lines trace cooler and denser gas that has probably been entrained intothe outowing hot gas (see, for example, Suchkov et al. 1994), and supply crucialcomplementary data. They provide unambiguous information about the magnitudeand sign of the radial velocities of the gas, more fully sample the whole range ofgas densities in the outow (rather than being strongly weighted in favour of thedensest material that may contain relatively little mass), and can be used to studyoutows in high-redshift galaxies, where the associated X-ray or optical emissionmay be undetectably faint (Franx et al. 1997; Pettini et al. 1998).

Martin (1999) empirically quantied the global eects of intense star formation onthe ISM by estimating the outow rates and velocities for a local sample of starburstand related galaxies. I will follow her approach, but update it by adding new X-rayresults on additional galaxies and improve upon it by including results from our(Heckman et al. 2000) analysis of the interstellar absorption lines in starbursts (adata type not considered by Martin). The principal results are as follows.

The outow speeds are, typically, 400{800 km s1, independent of the rotation speedof the `host galaxy. This result is shown in gure 1. Heckman et al. (2000) showthat the absorbing gas typically spans a velocity range from close to the galaxysystemic velocity up to some maximum blueshift. They argued that this can beunderstood if the hot outow is ablating material o of cold dense clouds and thenaccelerating it up to some terminal velocity. It is these inferred terminal velocitiesthat are plotted in gure 1. In the case of the X-ray data, outow velocities cannot

be observed directly. Instead, we have estimated the outow speed from the observedgas temperature following Chevalier & Cleggs (1985) solution for an adiabatic windfed by gas at a temperature T: v (5kT=)1=2, where is the mean mass perparticle. This is a conservative approach as it ignores any kinetic energy that theX-ray-emitting gas may already have. It is encouraging that the overall agreementbetween the two datasets is satisfactory.




6/16


1000

800

600

400

200

0 200100 300

v (km s )rot-

1

v

(kms

)

out

-1

Figure 1. Plot of the galaxy rotation speed versus the inferred terminal velocity of the outowfor starburst galaxies. The two diagonal lines show the relations vo u t = 2vr o t and vo u t = 3vr o t .Data points based on blueshifted interstellar absorption-line proles are indicated by lled dotsand the points based on the X-ray temperatures are indicated by empty dots. Note that the twodatasets are consistent with each other, imply that the outow speed is independent of the hostgalaxy potential well depth, and, thus, suggest that outows will preferentially escape from theleast-massive galaxies (see Heckman et al. (2000) for details).

The implied outow rates of metal-enriched material probably exceed the star-form-ation rate. I have calculated star-formation rates for a large sample of starburstgalaxies for which mass-outow rates can be estimated from either interstellar ab-sorption lines, X-ray emission, or H emission-line kinematics. The star-formationrates assume a Salpeter initial mass function extending from 0.3 to 100M-. For theoutow rates based on the absorption-line data, I follow Heckman et al. (2000) andset these equal to

_M 60rNHv(w=4)M- yr1;

where r (kpc) is the radius of the region of mass-injection (the starburst), NH (3

1021 cm2) is the total H column density in the absorbing material, v (200 km s1)is the mean outow velocity, and w is the solid angle lled by the wind. I adoptedthe values measured or estimated for these parameters by Heckman et al. (2000),Lehnert & Heckman (1995) and Armus et al. (1990). For the outow rates derivedfrom the X-ray data, I have simply taken the inferred mass of hot gas (assuminga volume lling factor of unity) and divided it by the outow time (approximately




7/16

2082 T. M. Heckman

equal to the sound-crossing time) for the X-ray emitting region. I also include theestimates for the outow rates in the warm ionized gas from Martin (1999), which arebased on the observed velocities in the H-emitting gas and the masses derived byassuming that this gas is in pressure balance with the ram pressure of the superwind(see Martin (1999) for details). While each method is crude, and makes simplifying

assumptions that may be unwarranted, it is gratifying that the three methods yieldsimilar results: the median value for _M =SFR = 1:4 for the absorption-line data,2.8 for the X-ray data, and 1.6 for the H data. In fact, since these three methodsmeasure dierent gas phases, the total outow rate could be approximated as thesum of the three rates. Thus, gas is being expelled from starbursts at a rate in excessof the rate that gas is being processed into stars.

The implied outow rates of kinetic/thermal energy are a signicant fraction of thetotal injection rate by supernovae and stellar winds. The rate at which superwinds

carry energy is uncertain. Simple estimates can be made from the X-ray data bycalculating the thermal energy content in the hot gas (assuming unit-volume ll-ing factor) and dividing this by the dynamical time for the X-ray-emitting region.As emphasized by Strickland (1998), this ignores the eects of clumping (which willreduce the thermal energy) and of the supersonic ow speeds (which represent signif-icant kinetic energy). Energy (more properly momentum) outow rates can also bededuced from the radial pressure gradients observed via standard optical emission-line diagnostics, assuming that these trace the sum of the ram and thermal pressurein the outow (see, for example, Heckman et al. 1990; Lehnert & Heckman 1996).

Both techniques suggest that the majority of the kinetic/thermal energy supplied bythe supernovae and stellar winds is being carried out in the ow (and has not beenradiated away). The key seems to be that the porosity of the starburst ISM is high,so that most supernovae detonate in a rareed preheated medium (see, for example,Heckman et al. 1990; Marlowe et al. 1995).

The outows are apparently quite dusty, with inferred reddenings of E(B V) 1over regions of a few to 10 kpc in size. Heckman et al. (2000) mapped the Na-D

interstellar absorption line over regions with sizes of a few to 10 kpc in a sampleof 18 starbursts. Figure 2 shows the strong correlation that they found betweenthe depth of the absorption-line prole (roughly proportional to the fraction of theemitting region covered by absorbing gas) and the line-of-sight reddening towardsthe emitting region. The observed reddening is substantial, with E(B V) 0:9 0:4 (corresponding to NH several 10

21 cm2 for normal galactic dust). Thisgeneralizes Phillipss (1993) discovery of a spectacularly dusty few-kpc-scale outowin the halo of the starburst galaxy NGC 1808.

The four results above have a variety of important implications. The lack of inde-

pendence of the outow speed upon the galaxy rotation speed strongly suggests thatthe outows selectively escape the potential wells of the less-massive galaxies, carry-ing metals with them. This process plausibly accounts for the strong mass{metallicityrelation in ellipticals and bulges. Lynden-Bell (1992) has proposed an appealinglysimple model in which the fraction of starburst-produced metals that are retainedby a galaxy experiencing an outow is proportional to the galaxy potential-well




8/16


0

0.8

0.2

0.4

0.6

1.0

-0.2 0 0.2 0.4

log (C65

/

C48)

I5890

Figure 2. Plot of the normalized residual intensity at the centre of the interstellar NaI5890transition (I5 8 9 0 ) versus the log of the colour of the optical continuum (the ratio of the valuesof F at rest wavelengths of 6560 and 4860 A). Points plotted as lled dots are the nucleiof powerful starbursts. Other points are o-nuclear locations. The deeper the NaI absorptionline, the more reddened the background starlight. An unreddened starburst population shouldhave log(C6 5 =C4 8 ) = 0:3. For a standard galactic reddening curve, the implied AV ranges upto roughly four magnitudes for the most-reddened sight lines (see Heckman et al. (2000) fordetails).

depth for galaxies with vesc < vou t, and asymptotes to full retention for the most-massive galaxies (vesc > vou t). For vou t in the measured range (gure 1), such a sim-ple prescription can reproduce the observed mass{metallicity relation for spheroids(Lynden-Bell 1992).

At the same time, superwinds could deposit the required amount of observed met-als in the intracluster medium (see, for example, Gibson et al. 1997). The kineticenergy carried out by superwinds at high redshift could be the agent that `preheatedthe proto-intracluster medium prior to the collapse/formation of rich clusters (Pon-man et al. 1999; Evrard, this issue). If the ratio of ejected metals to stellar spheroid

mass is the same globally as in clusters of galaxies, the present-day mass-weightedmetallicity of a general IGM with IGM = 0:015 will be approximately one-sixthsolar (see also Renzini 1997). These intergalactic metals could reconcile the apparentdecit in the inventory of metals contained by stars in the present Universe comparedwith the expectations based on the integrated amount of high-mass star formationover cosmic time (see, for example, Aguirre 1999). Pettini (this issue) notes a similar




9/16

2084 T. M. Heckman

problem with `missing metals at high z. Highly ionized superwind ejecta may bea plausible place to hide them. Superwinds may also help alleviate the problems inunderstanding the apparently wide distribution of metals in the Ly forest (Pettini,this issue; Efstathiou, this issue).

More speculatively, if the outowing dust survives its journey, the cumulative

eect of dusty superwinds could lead to a cosmologically signicant amount of inter-galactic dust (d u st = a few105), which would seriously aect the Hubble diagram

for type Ia supernovae (Aguirre 1999; Aguirre & Haiman 1999). Finally, Kronberget al. (1999) have argued that a substantial fraction of the IGM at high redshift canbe permeated with magnetic elds carried out of the rst generation of galaxies byearly superwinds.

4. The starburst{Seyfert connection

There have been many observational investigations into the possible presence ofstarbursts in Seyfert nuclei, and it is beyond the scope of this contribution to reviewthis literature. Instead, I want to report briey on eorts by our group to make directspectroscopic detections of hot, massive stars in a large, unbiased sample of Seyfertnuclei.

According to the standard `unied picture for radio-quiet AGN, the principalbuilding blocks for a Seyfert nucleus are as follows.

(1) A super-massive black hole and its associated accretion disc (the primary sourceof X-ray-through-optical continuum emission).

(2) An optically and geometrically thick circum-nuclear torus of dust and gas, withan inner radius of a few pc and an ill-dened outer radius (of order 10 2 pc). Itis viewed close to its equatorial plane (polar axis) in type 2 (type 1) Seyferts.

(3) A `mirror of dust and/or warm electrons located along the polar axis of thetorus.

Thus, the optimal targets in which to search for hot stars are the type 2 Seyfertnuclei, in which the torus providentially blocks out the blinding glare from a hiddentype 1 Seyfert nucleus. Indeed, type 2 Seyfert nuclei have long been known to exhibita `featureless continuum (FC) that produces most of the UV light and, typically,10{30% of the visible/NIR light (the rest appears to be light from an ordinary oldpopulation of stars). Until recently, it was thought that the optical/UV FC wasmostly light from the hidden type 1 Seyfert nucleus that had been reected into ourline of sight by the mirror.

Instead, if at least part of the FC is produced by a population of hot stars, itshould not actually be featureless! Instead, spectroscopy in the blue and near-UVshould show the high-order Balmer lines in absorption (unlike H and H-, which are

dominated by nebular emission) and weak HeI stellar photospheric lines (Gonzalez-Delgado et al. 1999). Spectra in the vacuum-UV should reveal strong stellar windlines and weaker stellar photospheric lines (cf. De Mello et al. 1999). To test this pos-sibility we have therefore undertaken a programme to obtain high-resolution vacuum-UV images and spectra (with HST) and near-UV spectra (with ground-based tele-scopes) of a representative sample of the brightest type 2 Seyfert nuclei.




10/16


The rst results have been presented in detail in Heckman et al. (1997) andGonzalez-Delgado et al. (1998). HST imaging shows that the UV continuum sourcein every case is spatially resolved (with a scale size of a few hundred pc or greater). Insome cases, the morphology is strikingly reminiscent of UV images of starbursts. Inother cases, a component of the UV continuum is roughly aligned with the inferred

polar axis of the obscuring torus (as expected for reected and/or reprocessed lightfrom the central engine).Of the original sample of 13 type 2 Seyferts with HST vacuum-UV images, only

four were bright enough for us to obtain spectra of adequate quality in the crucialUV spectral window from ca. 1200 to 1600 A. However, these spectra are decisive: allfour show the clear spectroscopic signature of a starburst population that dominatesthe UV continuum. In addition to classic, strong, stellar wind features (NV1240,SiIV1400 and CIV1550), we can also detect weaker and much narrower absorptionfeatures from highly excited transitions (which are, therefore, indisputably of stellar

origin).In each of the four cases, if we use the empirical `starburst attenuation law(Calzetti et al. 1994; Meurer et al. 1999) to correct the observed UV continuum fordust extinction, we nd that the bolometric luminosity of the nuclear (102 pc-scale)starburst is comparable with the estimated bolometric luminosity of the `hiddentype 1 Seyfert nucleus (of the order of 1010L-). Large-aperture UV spectra withinternal ultraviolet explore (IUE) imply the existence of a surrounding larger-scale(few kpc) and more powerful (a few 1010{1011L-) dusty starburst that is energet-ically capable of powering the bulk of the observed far-IR emission from the galaxy.Thus, starbursts are an energetically signicant (or even dominant) component of atleast some Seyfert galaxies.

However, we have HST spectra of only four type 2 Seyferts, and these are stronglybiased in favour of cases with high UV surface brightness. Can we say anything moregeneral? To address this, we have embarked on a programme to obtain spectra fromca. 3500 to 9000 A of a complete sample of the 25 brightest type 2 Seyfert nuclei inthe local Universe. These objects are selected from extensive lists of known Seyfertgalaxies on the basis of the ux of either the nebular line emission from the `narrowline region (the [OIII]5007 line) or from the nuclear radio source (Whittle 1992).

We are still analysing these spectra, but even a cursory inspection of the near-UVregion (below 4000 A) shows that about half have pronounced Balmer absorptionlines, whose strength is consistent with a population of late O or early B stars. Inseveral cases we can also detect those photospheric HeI absorption lines (4921,4387, 3819) that are not lled in by nebular emission. In most of the remainderof the sample, the FC is so weak relative to the light from a normal old-bulge stellarpopulation that its origin is still not clear. Recent work in a related vein has beenundertaken by Cid Fernandes et al. (1998) and Schmitt et al. (1999). Their resultsagree, at least qualitatively, with ours: they nd that most of the optical and near-UV FC in type 2 Seyfert nuclei is produced by young and intermediate-age stars

(age less than or equal to 100 Myr).Thus, it is clear that massive stars and starbursts play an important energetic

role in a signicant fraction of Seyfert nuclei. What is not yet clear is whether star-bursts are an essential component of the Seyfert phenomenon and what (if any) thecausal or evolutionary connection might be. Perhaps, as Cid Fernandes & Terlevich(1995) suggested, the starburst is an inevitable by-product of the dense molecu-




11/16

2086 T. M. Heckman

lar torus that is now believed to be a fundamental part of the inner machinery ofAGN (both obscuring the `central engine and serving as its fuel source). If true,this would have major implications for the relationship between quasars and galaxyformation.

5. The host galaxies of high-z radio-quiet quasars

The cosmic evolution of the population of powerful radio galaxies has been welldocumented over the past decade (see Rottgering et al. 1999). The uniformity ofthe K-band Hubble diagram out to z 4 and the evolution of the red envelopein the visible and near-IR colours are consistent with a large redshift of formationand the subsequent passive evolution of the progenitors of (some) present-day giantelliptical galaxies (see, for example, Lilly 1989; Eales & Rawlings 1996; McCarthy1999). Much less is known about the evolution of the population of the hosts of

radio-loud quasars, but the available data paint a broadly similar picture (see, forexample, Lehnert et al. 1992; Ridgway & Stockton 1997; McLure et al. 1999). Thisis consistent with the standard `unied model in which radio-loud quasars and radiogalaxies are drawn from the same parent population, but the quasars (radio galaxies)are viewed roughly along (perpendicular to) the polar axis of a dusty torus (Barthel1989).

While the notion of the `monolithic formation of apparently massive ellipticalgalaxies at high redshifts does not t comfortably into the standard cold-dark-matterpicture of hierarchical assembly at late times, powerful radio galaxies are exceedingly

rare objects. For H0 = 50 km s

1

Mpc

1

and = 1, 3CR radio galaxies at z = 2:5have a comoving space density of only 0.2 Gpc3 log P1rad

, while the fainter 6C(Eales 1999) and MRC (McCarthy 1999) radio galaxies have space densities roughly102 times larger (Dunlop & Peacock 1990). These values can be compared with thepresent-day space density of rst-ranked-cluster galaxies (roughly 5000 Gpc3; seeBahcall & Cen 1993). Even allowing for a short lifetime for the radio galaxy phase,the evolved descendants of powerful radio galaxies would account for only a smallminority of the rst-ranked-cluster elliptical galaxies.

In contrast, radio-quiet quasars are far more common and, therefore, more likelyto be progenitors of typical present-day early-type galaxies. Radio-quiet quasars withMB 6 23 have comoving space densities at z 2 of ca. 2 104 Gpc3 (Hartwick &Schade 1990). Now, for a quasar lifetime of the order of the Eddington growth time(a few per cent of the Hubble time at z = 2), the implied space density of the present-day descendants is of the order of 106 Gpc3, which is comparable with the spacedensity ofL Es and S0s (see, for example, Fukugita et al. 1998). This identication isquite consistent with the correlation between the masses of super-massive black holesand the spheroids in which they live today. A quasar with MB = 24 powered byaccretion at the Eddington rate requires MS MBH 210

8M-, and this super-massiveblack hole would live today in a spheroid with a mass of Msp h;z=0 5 10

10M-and V-band luminosity of ca. 1:5 1010L- 0:4L (Magorrian et al. 1998; Van derMarel 1999).

Thus, it is clearly important to document the properties of the host galaxies ofradio-quiet quasars over a broad range in redshifts. At low redshifts, the hosts ofthe most luminous radio-loud quasars, radio-quiet quasars and radio galaxies allappear to be similar: several-L E or S0 galaxies (McLure et al. 1999). However,




12/16


(a) (b)

(c) (d)

Figure 3. High-z radio-quiet quasar hosts, after the quasar has been subtracted and the imagesmoothed with a Gaussian kernel of 0.06 arcsec. Each panel has an area of 5.7 arcsec2 (orca. 45 kpc), N up, E left. (a) MZZ 9592, z 2:7, with an inset of the central region (unsmoothed).There is an o-centre residual host component. (b) MZZ 9744, z 1:8, (c) MZZ 1558, z 2:7,(d) MZZ 4935, z 1:8, host marginally detected. Note the apparent close (ca. 10 kpc) companiongalaxies in panels (b) and (c) (see Ridgway et al. (1999) for details).

ground-based near-IR imaging of small samples has already hinted that the situationmight be quite dierent at high redshift, with the hosts of radio-quiet quasars being

signicantly fainter than their radio-loud cousins (Lowenthal et al. 1995; but see alsoAretxaga et al. (1998)).

Recent analyses of HST NICMOSy images of radio-quiet quasars at z 2 have nowclaried the situation. We (Ridgway et al. 1999) have imaged ve faint (MB 23)radio-quiet quasars, complementing the analysis reported by Rix et al. (1999) of sixluminous (MB 26) gravitationally lensed cases. Some examples of the underlyinghosts galaxies in the Ridgway et al. (1999) sample are shown in gure 3.

While the samples are still modest in size, several conclusions can already be drawn(see gure 4; Ridgway et al. 1999; Rix et al. 1999).

(1) Typical radio-quiet quasars at z 2 are hosted by galaxies with rest-frameabsolute visual magnitudes similar to present-day L galaxies (MV = 20 to23 = M;V 1:5 mag).

y Near Infrared Camera and Multi-Object Spectrograph.




13/16

2088 T. M. Heckman

z = 0.4 z = 1

z = 2 z = 3

-26

-26

-24

-22

-20

-18

-24

-22

-20

-18

(a)

(c) (d)

(b)

-24 -26 -26-28 -24 -28

M quasarB M quasarB

M

host

V

M

host

V

Figure 4. The measured properties of the ve radio-quiet z 2 quasars and hosts in the Ridgwayet al. (1999) sample (empty symbols) overplotted on the theoretical predictions from Kaumann& Haehnelt (1999, their gure 12). The data have been plotted at both z = 2 (c) for comparisonwith the models (dots) and at z = 0:4 (a) to compare with observations of low-z hosts (indicatedby the triangular region) (see Ridgway et al. (1999) for details). Rix et al. (1999) found similarresults for more luminous z 2 radio-quiet quasars.

(2) As such, they are much fainter than radio galaxies at the same redshift (typi-

cally by ca. 2 mag).(3) These host galaxies are, if anything, less luminous than the hosts of similarly

powerful low-z radio-quiet quasars. Since the luminosity-weighted mean age ofthe stellar population in the high-z hosts is almost certainly younger than thatof the low-z hosts, the dierence in stellar mass will be even more pronounced.

(4) The rest-frame visual luminosities and sizes of the radio-quiet quasar hosts areroughly similar to those of the LBGs. Thus, the Lyman-break population mightrepresent the parent population of typical radio-quiet quasars. Our cycle 8 HST

observing programme will determine whether this similarity extends into therest-frame UV.

The potential implications of these results are quite tantalizing. First, they implythat the well-studied cosmic evolution of the hosts of the very-radio-loud AGN popu-lation is evidently not representative of the much-more-numerous radio-quiet popula-tion. Second, if we make the simplifying assumptions that the ratio ofLqu asar=MS MBH




14/16


is roughly independent of redshift and that MS MBH / Msp h;z=0, then it follows thatsuper-massive black holes form well before their host galaxies are fully assembled.This agrees qualitatively with the idea of the hierarchical assembly of massive galax-ies at late epochs. Indeed, as already pointed out by Rix et al. (1999) and Ridgwayet al. (1999), the observations (gure 4) agree rather well with the recent theoretical

predictions of Kaumann & Haehnelt (1999).

6. Summary

If the reader takes away only four ideas from my contribution, I hope that they arethe following.

(1) Starburst galaxies are good analogues (in fact, the only plausible local ana-logues) to the known population of star-forming galaxies at high redshift.

(2) Integrated over cosmic time, supernova-driven galactic winds (`superwinds)play an essential role in the evolution of galaxies and the IGM.

(3) Circum-nuclear starbursts are an energetically signicant component of theSeyfert phenomenon.

(4) The evolution of the population of the host galaxies of radio-quiet quasarsis signicantly dierent from that of powerful radio galaxies, and is at leastqualitatively consistent with the standard picture of the hierarchical assemblyof massive galaxies at relatively late times.

I thank all my collaborators on the work described above, and in particular Rosa Gonzalez-Delgado and Susan Ridgway. My research is supported in part by NASA LTSA grant NAG5-6400and HST grant GO-07864.

References

Aguirre, A. 1999 astro-ph/9904319.

Aguirre, A. & Haiman, Z. 1999 astro-ph/9907039.

Aretxaga, I., Le Mignant, D., Melnick, J., Terlevich, R. & Boyle, B. 1998 Mon. Not. R. Astr.Soc. 298, L13.

Armus, L., Heckman, T. & Miley, G. 1990 Astrophys. J. 364, 471.

Bahcall, N. & Cen, R. 1993 Astrophys. J. 407, L49.

Barger, A., Cowie, L. & Richards, E. 1999 astro-ph/9907022.

Barthel, P. 1989 Astrophys. J. 336, 606.

Blain, A., Smail, I., Ivison, R. & Kneib, J.-P. 1999 Mon. Not. R. Astr. Soc. 302, 632.

Buat, V. & Burgarella, D. 1998 Astron. Astrophys. 334, 772.

Calzetti, D. & Heckman, T. 1999 Astrophys. J. 519, 27.

Calzetti, D., Kinney, A. & Storchi-Bergmann, T. 1994 Astrophys. J. 429, 582.

Chevalier, R. & Clegg, A. 1985 Nature 317, 44.

Cid Fernandes, R. & Terlevich, R. 1995 Mon. Not. R. Astr. Soc. 272, 423.

Cid Fernandes, R., Storchi-Bergmann, T. & Schmitt, E. 1998 Mon. Not. R. Astr. Soc. 297, 579.

Dahlem, M., Weaver, K. & Heckman, T. 1998 Astrophys. J. Suppl. 118, 401.

De Mello, D., Leitherer, C. & Heckman, T. 1999 astro-ph/9909513.

Dunlop, J. & Peacock, J. 1990 Mon. Not. R. Astr. Soc. 247, 19.


http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29364L.471[aid=541119,doi=10.1086/169431]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29364L.471[aid=541119,doi=10.1086/169431]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29364L.471[aid=541119,doi=10.1086/169431]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29336L.606[aid=541120,doi=10.1086/167038]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29336L.606[aid=541120,doi=10.1086/167038]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29336L.606[aid=541120,doi=10.1086/167038]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29336L.606[aid=541120,doi=10.1086/167038]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29302L.632[aid=541121,doi=10.1046/j.1365-8711.1999.02178.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29302L.632[aid=541121,doi=10.1046/j.1365-8711.1999.02178.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29302L.632[aid=541121,doi=10.1046/j.1365-8711.1999.02178.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6361^28^29334L.772[aid=541122,springer=1]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6361^28^29334L.772[aid=541122,springer=1]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6361^28^29334L.772[aid=541122,springer=1]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29519L.27[aid=541123,doi=10.1086/307338]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29519L.27[aid=541123,doi=10.1086/307338]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29519L.27[aid=541123,doi=10.1086/307338]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29429L.582[aid=541124,doi=10.1086/174346]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29429L.582[aid=541124,doi=10.1086/174346]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29429L.582[aid=541124,doi=10.1086/174346]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29297L.579[aid=541127,doi=10.1046/j.1365-8711.1998.01519.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29297L.579[aid=541127,doi=10.1046/j.1365-8711.1998.01519.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29297L.579[aid=541127,doi=10.1046/j.1365-8711.1998.01519.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^29118L.401[aid=541128,doi=10.1086/313137]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^29118L.401[aid=541128,doi=10.1086/313137]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^29118L.401[aid=541128,doi=10.1086/313137]http://rsta.royalsocietypublishing.org/http://rsta.royalsocietypublishing.org/http://rsta.royalsocietypublishing.org/http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^29118L.401[aid=541128,doi=10.1086/313137]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29297L.579[aid=541127,doi=10.1046/j.1365-8711.1998.01519.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29429L.582[aid=541124,doi=10.1086/174346]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29519L.27[aid=541123,doi=10.1086/307338]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6361^28^29334L.772[aid=541122,springer=1]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29302L.632[aid=541121,doi=10.1046/j.1365-8711.1999.02178.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29336L.606[aid=541120,doi=10.1086/167038]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29364L.471[aid=541119,doi=10.1086/169431]


15/16

2090 T. M. Heckman

Eales, S. 1999 In The most distant radio galaxies (ed. H. Rottgering, P. Best & M. Lehnert),p. 33. Royal Netherlands Academy of Arts and Sciences.

Eales, S. & Rawlings, S. 1996 Astrophys. J. 460, 68.

Franx, M., Illingworth, G., Kelson, D., van Dokkum, P. & Tran, K.-V. 1997 Astrophys. J. 486,L75.

Fukugita, M., Hogan, C. & Peebles, J. 1998 Astrophys. J. 503, 518.Gallego, J., Zamorano, J., Aragon-Salamanca, A. & Rego, M. 1995 Astrophys. J. 445, L1.

Gibson, B., Loewenstein, M. & Mushotzky, R. 1997 Mon. Not. R. Astr. Soc. 290, 623.

Gonzalez-Delgado, R., Heckman, T., Leitherer, C., Meurer, G., Krolik, J., Wilson, A., Kinney,A. & Koratkar, A. 1998 Astrophys. J. 505, 174.

Gonzalez-Delgado, R., Leitherer, C. & Heckman, T. 1999 astro-ph/9907116.

Hartwick, F. & Schade, D. 1990 A. Rev. Astron. Astrophys. 28, 437.

Heckman, T. 1997 In Cosmic origins of galaxies, planets, and life (ed. J. M. Shull, C. Woodward& H. Thronson). Astronomical Society of the Pacic.

Heckman, T., Armus, L. & Miley, G. K. 1990 Astrophys. J. Suppl. 74, 833.

Heckman, T., Gonzalez-Delgado, R., Leitherer, C., Meurer, G., Krolik, J., Wilson, A., Koratkar,A. & Kinney, A. 1997 Astrophys. J. 482, 114.

Heckman, T., Robert, C., Leitherer, C., Garnett, D. & Van der Rydt, F. 1998 Astrophys. J.503, 646.

Heckman, T., Lehnert, M., Strickland, D. & Armus, L. 2000 Astrophys. J. (In the press.)

Hibbard, J. & Vacca, W. 1997 Astron. J. 114, 1741.

Hu, E., Cowie, L. & McMahon, R. 1998 Astrophys. J. 502, L99.

Hughes, D. (and 14 others) 1998 Nature 394, 241.

Kaumann, G. & Haehnelt, M. 1999 astro-ph/9906493.

Kronberg, P., Lesch, H. & Hopp, U. 1999 Astrophys. J. 511, 56.Lanzetta, K., Chen, H.-W., Fernandez-Soto, A., Pascarelle, S., Yahata, N. & Yahil, A. 1999

astro-ph/9910554.

Lehnert, M. D. & Heckman, T. 1995 Astrophys. J. Suppl. 97, 89.

Lehnert, M. D. & Heckman, T. M. 1996 Astrophys. J. 462, 651.

Lehnert, M. D., Heckman, T. M., Chambers, K. & Miley, G. 1992 Astrophys. J. 395, 466.

Leitherer, C., Walborn, N., Heckman, T. & Norman, C. 1991 Massive stars in starburst galaxies.Cambridge University Press.

Lilly, S. 1989 Astrophys. J. 340, 77.

Lowenthal, J., Heckman, T., Lehnert, M. & Elias, J. 1995 Astrophys. J. 451, 484.Lowenthal, J., Koo, D., Guzman, R., Gallego, J., Phillips, A., Faber, S., Vogt, N. & Illingworth,

G. 1997 Astrophys. J. 481, 673.

Lynden-Bell, D. 1992 In Elements and the cosmos (ed. M. Edmunds & R. Terlevich), p. 270.Cambridge University Press.

McCarthy, P. 1999 In The most distant radio galaxies (ed. H. Rottgering, P. Best & M. Lehnert),p. 5. Royal Netherlands Academy of Arts and Sciences.

McLure, R., Kukula, M., Dunlop, J., Baum, S., ODea, C. & Hughes, D. 1999 Mon. Not. R.Astr. Soc. 308, 377.

Madau, P., Ferguson, H., Dickinson, M., Giavalisco, M., Steidel, C. & Fruchter, A. 1996 Mon.Not. R. Astr. Soc. 283, 1388.

Magorrian, J., Tremaine, S., Richstone, D., Bender, R., Bower, G., Dressler, A., Faber, S.,Gebhardt, K., Green, R., Grillmair, C., Kormendy, J. & Lauer, T. 1998 Astron. J. 115, 2285.

Mannucci, F., Thompson, D., Beckwith, S. & Williger, G. 1998 Astrophys. J. 501, L11.

Marlowe, A., Heckman, T., Wyse, R. & Schommer, R. 1995 Astrophys. J. 438, 563.

Martin, C. 1999 Astrophys. J. 513, 156.


http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29460L.68[aid=541130,doi=10.1086/176953]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29460L.68[aid=541130,doi=10.1086/176953]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29460L.68[aid=541130,doi=10.1086/176953]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29503L.518[aid=541131,doi=10.1086/306025]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29503L.518[aid=541131,doi=10.1086/306025]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29503L.518[aid=541131,doi=10.1086/306025]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29290L.623[aid=541132]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29290L.623[aid=541132]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29290L.623[aid=541132]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29505L.174[aid=541133,doi=10.1086/306154]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29505L.174[aid=541133,doi=10.1086/306154]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29505L.174[aid=541133,doi=10.1086/306154]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29505L.174[aid=541133,doi=10.1086/306154]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0066-4146^28^2928L.437[aid=540873]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0066-4146^28^2928L.437[aid=540873]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0066-4146^28^2928L.437[aid=540873]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^2974L.833[aid=541134,doi=10.1086/191522]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^2974L.833[aid=541134,doi=10.1086/191522]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^2974L.833[aid=541134,doi=10.1086/191522]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^2974L.833[aid=541134,doi=10.1086/191522]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29482L.114[aid=541135,doi=10.1086/304139]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29482L.114[aid=541135,doi=10.1086/304139]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29482L.114[aid=541135,doi=10.1086/304139]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29503L.646[aid=541136,doi=10.1086/306035]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29503L.646[aid=541136,doi=10.1086/306035]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29503L.646[aid=541136,doi=10.1086/306035]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29503L.646[aid=541136,doi=10.1086/306035]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.1741[aid=541137,doi=10.1086/118603]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.1741[aid=541137,doi=10.1086/118603]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.1741[aid=541137,doi=10.1086/118603]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.1741[aid=541137,doi=10.1086/118603]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0028-0836^28^29394L.241[aid=540896,doi=10.1038/28328]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0028-0836^28^29394L.241[aid=540896,doi=10.1038/28328]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0028-0836^28^29394L.241[aid=540896,doi=10.1038/28328]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29511L.56[aid=541138,doi=10.1086/306662]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29511L.56[aid=541138,doi=10.1086/306662]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29511L.56[aid=541138,doi=10.1086/306662]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29511L.56[aid=541138,doi=10.1086/306662]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^2997L.89[aid=541139,doi=10.1086/192137]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^2997L.89[aid=541139,doi=10.1086/192137]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^2997L.89[aid=541139,doi=10.1086/192137]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29462L.651[aid=541140,doi=10.1086/177180]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29462L.651[aid=541140,doi=10.1086/177180]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29462L.651[aid=541140,doi=10.1086/177180]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29395L.466[aid=541141,doi=10.1086/171666]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29395L.466[aid=541141,doi=10.1086/171666]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29395L.466[aid=541141,doi=10.1086/171666]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29340L.77[aid=541142,doi=10.1086/167377]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29340L.77[aid=541142,doi=10.1086/167377]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29340L.77[aid=541142,doi=10.1086/167377]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29340L.77[aid=541142,doi=10.1086/167377]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29451L.484[aid=541143,doi=10.1086/176237]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29451L.484[aid=541143,doi=10.1086/176237]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29451L.484[aid=541143,doi=10.1086/176237]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29481L.673[aid=540969,doi=10.1086/304092]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29481L.673[aid=540969,doi=10.1086/304092]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29481L.673[aid=540969,doi=10.1086/304092]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29481L.673[aid=540969,doi=10.1086/304092]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29308L.377[aid=540879,doi=10.1046/j.1365-8711.1999.02676.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29308L.377[aid=540879,doi=10.1046/j.1365-8711.1999.02676.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29308L.377[aid=540879,doi=10.1046/j.1365-8711.1999.02676.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29308L.377[aid=540879,doi=10.1046/j.1365-8711.1999.02676.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29308L.377[aid=540879,doi=10.1046/j.1365-8711.1999.02676.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29115L.2285[aid=539052,doi=10.1086/300353]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29115L.2285[aid=539052,doi=10.1086/300353]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29115L.2285[aid=539052,doi=10.1086/300353]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29115L.2285[aid=539052,doi=10.1086/300353]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29438L.563[aid=541144,doi=10.1086/175101]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29438L.563[aid=541144,doi=10.1086/175101]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29438L.563[aid=541144,doi=10.1086/175101]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29513L.156[aid=541145,doi=10.1086/306863]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29513L.156[aid=541145,doi=10.1086/306863]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29513L.156[aid=541145,doi=10.1086/306863]http://rsta.royalsocietypublishing.org/http://rsta.royalsocietypublishing.org/http://rsta.royalsocietypublishing.org/http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29308L.377[aid=540879,doi=10.1046/j.1365-8711.1999.02676.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29503L.646[aid=541136,doi=10.1086/306035]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29513L.156[aid=541145,doi=10.1086/306863]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29438L.563[aid=541144,doi=10.1086/175101]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29115L.2285[aid=539052,doi=10.1086/300353]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29308L.377[aid=540879,doi=10.1046/j.1365-8711.1999.02676.x]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29481L.673[aid=540969,doi=10.1086/304092]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29451L.484[aid=541143,doi=10.1086/176237]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29340L.77[aid=541142,doi=10.1086/167377]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29395L.466[aid=541141,doi=10.1086/171666]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29462L.651[aid=541140,doi=10.1086/177180]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^2997L.89[aid=541139,doi=10.1086/192137]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29511L.56[aid=541138,doi=10.1086/306662]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0028-0836^28^29394L.241[aid=540896,doi=10.1038/28328]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.1741[aid=541137,doi=10.1086/118603]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29503L.646[aid=541136,doi=10.1086/306035]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29482L.114[aid=541135,doi=10.1086/304139]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0067-0049^28^2974L.833[aid=541134,doi=10.1086/191522]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0066-4146^28^2928L.437[aid=540873]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29505L.174[aid=541133,doi=10.1086/306154]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29290L.623[aid=541132]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29503L.518[aid=541131,doi=10.1086/306025]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29460L.68[aid=541130,doi=10.1086/176953]


16/16


Meurer, G., Heckman, T., Leitherer, C., Lowenthal, J. & Lehnert, M. 1997 Astron. J. 114, 54.

Meurer, G., Heckman, T. & Calzetti, D. 1999 Astrophys. J. 521, 64.

Norman, C. & Scoville, N. 1988 Astrophys. J. 332 , 134.

Perry, J. & Dyson, J. 1985 Mon. Not. R. Astr. Soc. 213, 665.

Pettini, M., Kellogg, M., Steidel, C., Dickinson, M., Adelberger, K. & Giavalisco, M. 1998

Astrophys. J. 508, 539.Phillips, A. 1993 Astron. J. 105 , 486.

Ponman, T., Cannon, D. & Navarro, J. 1999 Nature 397, 135.

Renzini, A. 1997 Astrophys. J. 488, 35.

Ridgway, S. & Stockton, A. 1997 Astron. J. 114, 511.

Ridgway, S., Heckman, T., Calzetti, D. & Lehnert, M. 1999 astro-ph/9911049.

Rix, H.-W., Falco, E., Impey, C., Kochanek, C., Lehar, J., McLeod, B., Munoz, J. & Peng, C. Y.1999 astro-ph/9910065.

Rottgering, H., Best, P. & Lehnert, M. 1999 The most distant radio galaxies. Royal NetherlandsAcademy of Arts and Sciences.

Schmitt, H., Storchi-Bergmann, T. & Cid Fernandes, R. 1999 Mon. Not. R. Astr. Soc. 303, 35.

Soifer, B. T., Sanders, D., Madore, B., Neugebauer, G., Lonsdale, C., Persson, S. E. & Rice, W.1987 Astrophys. J. 320, 238.

Steidel, C., Adelberger, K., Giavalisco, M., Dickinson, M. & Pettini, M. 1999 Astrophys. J.519, 1.

Strickland, D. 1998 X-ray emission from starburst-driven galactic winds. PhD thesis, The Uni-versity of Birmingham.

Suchkov, A. A., Balsara, D. S., Heckman, T. M. & Leitherer, C. 1994 Astrophys. J. 430, 511.

Teplitz, H., Malkan, M. & McLean, I. 1998 Astrophys. J. 506, 519.

Terlevich, R. & Melnick, J. 1998 Mon. Not. R. Astr. Soc. 213, 841.Van der Marel, R. 1999 Astron. J. 117, 744.

Weedman, D. 1983 Astrophys. J. 266, 479.

Whittle, M. 1992 Ap. J. Suppl. 79, 49.

Phil Trans R Soc Lond A (2000)

http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.54[aid=540971,doi=10.1086/118452]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.54[aid=540971,doi=10.1086/118452]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.54[aid=540971,doi=10.1086/118452]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29521L.64[aid=540972,doi=10.1086/307523]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29521L.64[aid=540972,doi=10.1086/307523]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29521L.64[aid=540972,doi=10.1086/307523]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29508L.539[aid=540900,doi=10.1086/306431]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29508L.539[aid=540900,doi=10.1086/306431]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29508L.539[aid=540900,doi=10.1086/306431]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29508L.539[aid=540900,doi=10.1086/306431]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29105L.486[aid=541148,doi=10.1086/116447]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29105L.486[aid=541148,doi=10.1086/116447]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29105L.486[aid=541148,doi=10.1086/116447]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0028-0836^28^29397L.135[aid=541149,doi=10.1038/16410]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0028-0836^28^29397L.135[aid=541149,doi=10.1038/16410]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0028-0836^28^29397L.135[aid=541149,doi=10.1038/16410]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29488L.35[aid=541150,doi=10.1086/304696]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29488L.35[aid=541150,doi=10.1086/304696]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29488L.35[aid=541150,doi=10.1086/304696]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.511[aid=541151,doi=10.1086/118490]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.511[aid=541151,doi=10.1086/118490]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.511[aid=541151,doi=10.1086/118490]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29303L.35[aid=541152]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29303L.35[aid=541152]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29303L.35[aid=541152]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29320L.238[aid=541153,doi=10.1086/165536]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29320L.238[aid=541153,doi=10.1086/165536]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29320L.238[aid=541153,doi=10.1086/165536]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29320L.238[aid=541153,doi=10.1086/165536]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29519L.1[aid=540907,doi=10.1086/307363]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29519L.1[aid=540907,doi=10.1086/307363]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29519L.1[aid=540907,doi=10.1086/307363]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29519L.1[aid=540907,doi=10.1086/307363]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29430L.511[aid=541154,doi=10.1086/174427]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29430L.511[aid=541154,doi=10.1086/174427]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29430L.511[aid=541154,doi=10.1086/174427]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29506L.519[aid=541155,doi=10.1086/306255]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29506L.519[aid=541155,doi=10.1086/306255]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29506L.519[aid=541155,doi=10.1086/306255]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29213L.841[aid=541156]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29213L.841[aid=541156]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29213L.841[aid=541156]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29117L.744[aid=541157,doi=10.1086/300730]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29117L.744[aid=541157,doi=10.1086/300730]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29117L.744[aid=541157,doi=10.1086/300730]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29266L.479[aid=541158,doi=10.1086/160796]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29266L.479[aid=541158,doi=10.1086/160796]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29266L.479[aid=541158,doi=10.1086/160796]http://rsta.royalsocietypublishing.org/http://rsta.royalsocietypublishing.org/http://rsta.royalsocietypublishing.org/http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29519L.1[aid=540907,doi=10.1086/307363]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29266L.479[aid=541158,doi=10.1086/160796]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29117L.744[aid=541157,doi=10.1086/300730]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29213L.841[aid=541156]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29506L.519[aid=541155,doi=10.1086/306255]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29430L.511[aid=541154,doi=10.1086/174427]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29519L.1[aid=540907,doi=10.1086/307363]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29320L.238[aid=541153,doi=10.1086/165536]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0035-8711^28^29303L.35[aid=541152]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.511[aid=541151,doi=10.1086/118490]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29488L.35[aid=541150,doi=10.1086/304696]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0028-0836^28^29397L.135[aid=541149,doi=10.1038/16410]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29105L.486[aid=541148,doi=10.1086/116447]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29508L.539[aid=540900,doi=10.1086/306431]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-637X^28^29521L.64[aid=540972,doi=10.1086/307523]http://gessler.ingentaselect.com/nw=1/rpsv/cgi-bin/linker?ext=a&reqidx=/0004-6256^28^29114L.54[aid=540971,doi=10.1086/118452]

phil. trans. r. soc. lond. a-2000-heckman-2077-91

Documents