Mon.Not.R.Astron.Soc.000,1–??(2002) Printed5February2008 (MNLATEXstylefilev2.2) A Census of the Carina Nebula. I: Cumulative Energy Input from Massive Stars Nathan Smith⋆ Centerfor Astrophysics and Space Astronomy, University of Colorado, 389 UCB, Boulder, CO 80309, USA 6 0 0 Accepted 0000,Received0000,inoriginalform0000 2 n a ABSTRACT J 4 TheCarinaNebula(NGC3372)isourrichestnearbylaboratoryinwhichtostudy feedback through UV radiation and stellar winds from very massive stars during the 1 formation of an OB association, at an early phase in the evolution of the surround- v ing proto-superbubble before supernova explosions have influenced the environment. 0 This feedback is triggering successive generations of new star formation around the 6 periphery ofthe nebula, while simultaneouslyevaporating the gasanddust reservoirs 0 out of whichyoung stars are trying to accrete material.This paper takes inventoryof 1 the combined effect from all the known massive stars that power the Carina Nebula 0 through their total ionizing flux and integrated mechanical energy from their stellar 6 winds.Carinais closeenoughandaccessibleenoughthatspectraltypes forindividual 0 stars are available, and many close binary and multiple systems have recently been / h spatially resolved, so that one can simply add them. Adopting values from the litera- p tureforcorrespondingspectraltypes,thepresent-daytotalionizingphotonluminosity o- produced by the 65 O stars and 3 WNL stars in Carina is QH ≃1051 s−1, the total r bolometric luminosity of all stars earlier than B2 is 2.5×107 L⊙, and the total me- st chanical luminosity of stellar winds is LSW ≃105 L⊙. The total QH was about 25% a higher when η Carinae was on the main sequence, before it and its companion were : surroundedby its obscuringdust shell;for the first3 Myr,the net ionizing flux of the v i 70 O stars in Carina was about 150 times greater than in the Orion Nebula. About X 400-500M⊙ has been contributed to the H ii region by stellar wind mass loss during r the past 3 Myr. Values for QH and LSW are also given for the individual clusters a Tr14, 15, and 16, and Bo10 and 11, which are more relevant on smaller spatial scales than the total values for the whole nebula. Key words: H ii regions — ISM: bubbles — ISM: individual (NGC 3372) — stars: early-type — stars: formation — stars: mass-loss — stars: winds, outflows 1 INTRODUCTION Hunter1998), or objects in ourown Galaxy liketheArches clusterneartheGalacticcenter(Najarroetal.2004;Figeret Feedback from young massive stars may play an integral al.1999,2002)andNGC3603(Moffatetal.2002).However, role in star and planet formation. Most stars are born in these other regions are too distant for detailed studies of the vicinity of the hottest and most massive stars spawned small-scale phenomena like irradiated protoplanetary disks only from giant molecular clouds, and the effects of feed- andjets,andtheirstudyishamperedbyconsiderably more back from these massive stars cannot be studied in nearby extinction.ThelowextinctiontowardCarinacombinedwith quiescent regions of star formation. Even the nearest H ii itsproximityandrichnebularcontentprovideaworthwhile region, the Orion Nebula, may not be representative of the trade-off for its less concentrated star clusters. While this extremeenvironmentswheremoststarsareborn,sinceitis may be a liability for investigating the upper end of the dominated byjust a single O6 dwarf. initialmassfunction,thefactthatCarina’sstarsaremoreof Amuchmoreextremecollectionofstarscanbefoundin alooseaggregatemaybeanadvantageforstudyingfeedback theCarina Nebula(NGC3372). Itsstarclustersarealmost and triggered star formation. Carina provides a snapshot as spectacular as those of 30 Dor in the LMC (Massey & of an OB association in the making, which may be more representativeoftheenvironmentsinwhichmoststarsform than are bound super star clusters. It may also provide an ⋆ HubbleFellow;[email protected] 2 N. Smith early analog of thelocal Galactic environment that evolved remarkable stellar content of the Carina Nebula has been into Gould’s Belt. discussed many times in the literature (e.g., Walborn 1995, Recent studies suggest that at the present epoch there 2002,2005;Feinstein1995;Massey&Johnson1993),thereis exists an upper mass limit to the most massive stars of nocompleteandup-to-datecensusofthemassivestars,and ∼150 M⊙ (Figer 2005, Weidner & Kroupa 2004, Kroupa especially there has been no estimate of their collective UV & Weidner 2005; Oey & Clarke 2005). While the star clus- radiationandmechanical(wind)luminositythatpowersthe ters that power Carina may not be as dense as the Arches region.Thisisthemainpurposeofthepresentinvestigation. cluster orNGC 3603, Carina doescontain several examples A second paper in this series will compare the total energy of stars that probably began their lives with 100–150 M⊙. input from stars with the apparent energy budget inferred Amongtheseareη Carinae, theprototypicalO2supergiant from observations of thesurrounding Carina Nebula. HD93129A, threelate-typehydrogen-richWolf-Rayetstars (WNL stars), and the remaining original members of the O3 spectral class for which this spectral type was first in- 2 ADOPTED PARAMETERS troduced(Walborn1973;althoughseeWalbornetal.2002). Forthesecolossal stars,lifetimes of ∼3Myrbefore theyex- Thebasicstrategy inthispaperwill betousetheobserved plodeassupernovae(SNe)areshorterthanthetimeittakes spectral typeof each massive star (spectral typesas late as to clear away their natal molecular material. Consequently, B2 will be considered, although their influence is minimal) we already have a situation in Carina where the most mas- to assign a corresponding value for the hydrogen-ionizing sive members like η Car are approaching their imminent photonluminosity QH,as wellas themass-loss rate M˙ and demise while new stars are being born from dense molecu- terminal wind speed V∞, and hence, the wind mechanical lar gas only 5–20 pc away. In the next 1–2 Myr, there will luminosity LSW=(1/2)M˙V2∞. Table 1 lists the values that be several very energetic SNe in the Carina Nebula, which willbeadoptedforstarslaterthanO3.InTable1,valuesof will carve out an even larger cavity in the ISM and form a logLandQH foreachspectraltypearetakenfromthe“ob- giant superbubble in the Galactic plane, and may pollute served” temperature-scale models of Martins, Schaerer, & protoplanetary disks with nuclear-processed ejecta. In the Hillier (2005). Values for M˙ are taken from Repolust,Puls, mean time, studying this rich region provides a snapshot & Herrero (2004), and terminal velocities for each spectral of a young proto-superbubble(Smith et al. 2000) energized typearetakenbyaroughfittotheaveragevalueslistedby onlybyUVradiationandstellarwinds,immediatelybefore Prinja, Barlow, & Howarth (1990).1 thedisruptiveSN-dominated phase. The stellar content of the Carina Nebula is truly re- In just the past decade, the Carina Nebula has been markable and accessible, and several of the more spectac- recognized as a hotbed of ongoing active star formation. It ular stars have received enough attention that the desired provides a laboratory to study several star formation phe- parametershavebeenderivedindividuallyforthem.Insome nomenaingreatdetail,allofwhichhaverecentlybeeniden- cases, when noted below, individual studies may supersede tified in this region: (1) evaporating protoplanetary disks the average values adopted in Table 1. Many of these stars (the so-called “proplyds”), small cometary clouds, or glob- are,infact,benchmarksfortheirspectraltypesinthethree ules (Smith et al. 2003a; Smith 2002b; Brooks et al. 2000, papers just mentioned. In addition, the three WNL stars 2006; Cox & Bronfman 1995), (2) the erosion of large dust in Carina are of particular interest, as are HD 93129A and pillars and the triggering of a second generation of embed- η Car (when it was on the main sequence), since these are ded star formation within them (Smith et al. 2000, 2005; themostluminoussourcesthathavethemostinfluence.The Rathborne et al. 2004; Megeath et al. 1996), (3) irradiated specificstellarcontentofthenebulaisdescribedinmorede- Herbig-Haro(HH)jetsthatareasignpostofactiveaccretion tail in thefollowing section. (Smithetal.2004),(4)photodissociationregions(PDRs)on the surfaces of molecular clouds across the region (Brooks et al. 2003, 2006, 1998; Rathborne et al. 2002; Smith et al. 3 STELLAR CONTENT 2000; Mizutani et al. 2004), and on the largest scales, (5) theearlyformationofafledgelingsuperbubble(Smithetal. Tables 2–6 compile the stellar content powering the Carina 2000). Thesephenomenatraceasecondgeneration ofstars, Nebula.Referencesforthestellarcontentofeachclusterare perhapstriggeredbyfeedbackfromthefirstgenerationclus- noted in each table. Several reviews exist in the literature ters.Carinaisnearenoughthatanupcomingprogramwith (especially Walborn 1995, 2005; Feinstein 1995); these were theHubbleSpaceTelescopewillundertakethesametypesof invaluable in compiling the tables. Every effort has been detailed studies of feedback that have been done in Orion made to provide the most recent information with regard atvisualandnear-IRwavelengths,butsamplingamoreex- to binarity/multiplicity through spectral types or through treme environment. high spatial resolution techniques, especially for Tr14 and 16. This, of course, does not preclude the possibility that All of these phenomena respond to the cumulative UV additionalcompanionsmaybeidentified,andtheirinfluence radiationand/orstellarwindsfromstarclustersthatpower theHiiregion,andacensusofthisenergyinputisessential should be considered by readers in that case. Following Walborn (1995) and Tapia et al. (1988), it in order to understand the relationship between them. Es- timating the total stellar energy input in Carina is a more difficult problem than in smaller H ii regions like Orion, 1 Given the spread in terminal velocities listed by Prinja et al. where a single star dominates thephotoevaporation of pro- 1990, we adopted the same velocities for all three luminosity plyds and the irradiation of HH jets. Instead, Carina has classes in Table 1. This is a reasonable approximation for the dozens of massive O stars overmany parsecs. Although the earliestspectraltypes thatdominatetheenergybudget. A Census of the Carina Nebula 3 Table 1.Adopted ParametersforEachSpectralType Spect. Lum. logL logQH M˙ V∞ LSW Type Class (L⊙) (s−1) (10−6 M⊙ yr−1) (kms−1) (L⊙) O3 V 5.84 49.64 3.1 3160 2450 O3.5 V 5.76 49.54 2.5 3080 1870 O4 V 5.67 49.44 2.0 2990 1410 O4.5 V 5.58 49.33 1.6 2870 1040 O5 V 5.49 49.22 1.3 2760 780 O5.5 V 5.41 49.10 1.0 2650 550 O6 V 5.32 48.99 0.8 2560 410 O6.5 V 5.23 48.88 0.6 2460 290 O7 V 5.14 48.75 0.4 2320 170 O7.5 V 5.05 48.61 0.35 2210 135 O8 V 4.96 48.44 0.26 2100 91 O8.5 V 4.86 48.27 0.22 1960 67 O9 V 4.77 48.06 0.18 1820 47 O9.5 V 4.68 47.88 0.15 1660 33 B0 V 4.57 47.70 0.12 1510 22 B0.5 V 4.47 47.50 0.10 1460 17 B1 V 4.37 47.28 0.08 1180 8.8 B1.5 V 4.28 47.05 0.06 960 4.4 B2 V 4.19 46.80 0.05 750 2.2 O3 III 5.96 49.77 6.5 3160 5130 O3.5 III 5.91 49.71 5.5 3080 4120 O4 III 5.85 49.64 4.7 2990 3320 O4.5 III 5.79 49.56 4.0 2870 2600 O5 III 5.73 49.48 3.4 2760 2050 O5.5 III 5.67 49.40 2.7 2650 1500 O6 III 5.61 49.32 2.2 2560 1140 O6.5 III 5.54 49.23 1.8 2460 860 O7 III 5.48 49.13 1.5 2320 640 O7.5 III 5.42 49.01 1.2 2210 460 O8 III 5.35 48.88 0.8 2100 280 O8.5 III 5.28 48.75 0.7 1960 210 O9 III 5.21 48.65 0.5 1820 130 O9.5 III 5.15 48.42 0.4 1660 87 B0 III 5.08 48.28 0.3 1510 54 B0.5 III 5.00 48.10 0.25 1460 42 B1 III 4.93 47.90 0.20 1180 22 B1.5 III 4.86 47.68 0.15 960 11 B2 III 4.78 47.44 0.11 750 4.9 O3 I 5.99 49.78 13.0 3160 10300 O3.5 I 5.96 49.74 11.3 3080 8470 O4 I 5.93 49.70 10.0 2990 7060 O4.5 I 5.90 49.66 8.9 2870 5790 O5 I 5.87 49.62 8.0 2760 4810 O5.5 I 5.84 49.58 7.1 2650 3940 O6 I 5.81 49.52 6.4 2560 3310 O6.5 I 5.78 49.46 5.7 2460 2730 O7 I 5.75 49.41 5.0 2320 2130 O7.5 I 7.72 49.31 4.3 2210 1660 O8 I 5.68 49.25 3.6 1960 1090 O8.5 I 5.65 49.19 3.0 1820 790 O9 I 5.61 49.11 2.5 1660 540 O9.5 I 5.57 49.00 1.9 1510 340 References. — L,QH: Martinset al. (2005); V∞:Prinjaet al.(1990); M˙: Repolustetal.(2004).LandQHforBstarswereextrapolatedfromMartins etal.(2005)followingCrowther(2005). is assumed that all these clusters share a common distance requiringthatTr15beseveralhundredpcbehindtherestof to theCarina Nebula, taken hereto be 2.3±0.1 kpc(Smith theCarinanebula.Thisseemsimplausible,althoughnotim- 2002a;Walborn1995;Allen&Hillier1993).Thedistanceof possible, for several reasons, including the fact that Tr15 is 2.3kpcdeterminedfor η Carinaeitself isveryreliable,hav- seenalongthesamelineofsightasdensegasandduststruc- ingbeenderivedfrompropermotionsandDopplervelocities tures,someofwhicharedustpillarsthatpointtowardTr16. oftheexpandingHomunculus(Smith2002a;Allen&Hillier Yet,starsinTr15haveanaverageE(B−V)=0.5(Feinstein 1993). Furthermore, η Car is known to be at the same dis- et al. 1980) – roughly the same as Tr14 and 16 (Walborn tance as the Keyhole nebula, since its reflected spectrum is 1995; Feinstein et al. 1973) –which could not bethecaseif seen across the Keyhole (Walborn & Liller 1977; Lopez & Tr15 were far behind the rest of the nebula and obscured. Meaburn1986). Differencesinphotometricdistancemoduli Similarly, relatively large distances to Tr14 (Vazquez et al. given in the literature are probably due to anomalous or 1996)areimplausible.EventhoughTr14mustbesomewhat varying reddening laws, patchy extinction, small age differ- moredistantthanTr16(perhapsasmuchas20pc)because ences, or other factors. For example, photometric distances silhouetteobjects projected infront ofitpointtowardTr16 ashighas3.7kpchavebeengivenforTr15(Walborn1973), (Smith et al. 2003a), it cannot be several hundred parsecs 4 N. Smith Table 2.StellarContentofTrumpler16andCollinder228 Name Sp. logL logQH M˙ V∞ LSW Type (L⊙) (s−1) (10−6 M⊙ yr−1) (kms−1) (L⊙) ηCar-now LBV 6.67 ... 1000 600 28700 ηCar-MS O2If(?) 6.57 50.36 8.32 3200 6780 ηCar-B O5V(?) 5.49 49.2 1.3 2760 780 HD93162 WN6ha 6.22 50.0 10.5 2480 5140 HD93131 WN6ha 5.94 49.7 13.8 2160 5130 HD92740A WN7ha 6.01 49.8 14.5 1785 3680 HD92740B O8-9V 4.86 48.27 0.22 1960 67 HD93205A O3.5V((f)) 5.76 49.54 2.5 3080 1870 HD93205B O8V 4.96 48.44 0.26 2100 91 HD93250 O3.5V((f)) 5.76 49.54 2.5 3080 1870 HDE303308 O4V((f)) 5.67 49.44 2.0 2990 1410 HD93204 O5V((f)) 5.49 49.22 1.3 2760 780 Tr16-244 O4If 5.93 49.70 10.0 2990 7060 CPD-592600 O6V((f)) 5.32 48.99 0.8 2560 410 CPD-592603A O7V((f)) 5.14 48.75 0.4 2320 170 CPD-592603B O9.5V 4.68 47.88 0.15 1660 33 CPD-592603C B0.2V 4.57 47.70 0.12 1510 22 CPD-592628A O9.5V 4.68 47.88 0.15 1660 33 CPD-592628B B0.3V 4.57 47.70 0.12 1510 22 CPD-592635A O8V 4.96 48.44 0.26 2100 91 CPD-592635B O9.5V 4.68 47.88 0.15 1660 33 CPD-592636A O7V 5.14 48.75 0.4 2320 170 CPD-592636B O8V 4.96 48.44 0.26 2100 91 CPD-592636C O9V 4.77 48.06 0.18 1820 47 CPD-592641 O5V 5.49 49.22 1.3 2760 780 HDE303311 O5V 5.49 49.22 1.3 2760 780 HDE305536 O9V 4.77 48.06 0.18 1820 47 HD93027 O9.5V 4.68 47.88 0.15 1660 33 HD93028 O9V 4.77 48.06 0.18 1820 47 HD93130 O6III(f) 5.61 49.32 2.2 2560 1140 HD93146 O6.5V((f)) 5.23 48.88 0.6 2460 290 HD93160 O6III((f)) 5.61 49.32 2.2 2560 1140 HD93161Aa O8V 4.96 48.44 0.26 2100 91 HD93161Ab O9V 4.77 48.06 0.18 1820 47 HD93161B O6.5V((f)) 5.23 48.88 0.6 2460 290 HD93206A O9.7Ib:(n) 5.57 49.00 1.9 1510 340 HD93206B O9III 5.21 48.65 0.5 1820 130 HD93222 O7III(f) 5.48 49.13 1.5 2320 640 HD93343 O7V(n) 5.14 48.75 0.4 2320 170 HD93403A O5III(f)var 5.73 49.48 3.4 2760 2050 HD93403B O7V 5.14 48.75 0.4 2320 170 Tr16-2 B1V 4.37 47.28 0.08 1180 8.8 Tr16-3 O8.5V 4.86 48.27 0.22 1960 67 Tr16-4 B1V 4.37 47.28 0.08 1180 8.8 Tr16-5 B1V 4.37 47.28 0.08 1180 8.8 Tr16-8 B0.5V 4.47 47.50 0.10 1460 17 Tr16-9 O9.5V 4.68 47.88 0.15 1660 33 Tr16-10 B0V 4.57 47.70 0.12 1510 22 Tr16-11 B1.5V 4.28 47.05 0.06 960 4.4 Tr16-12 B1V 4.37 47.28 0.08 1180 8.8 Tr16-13 B1V 4.37 47.28 0.08 1180 8.8 Tr16-14 B0.5V 4.47 47.50 0.10 1460 17 Tr16-16 B1V 4.37 47.28 0.08 1180 8.8 Tr16-17 B1V 4.37 47.28 0.08 1180 8.8 Tr16-19 O9.5V 4.68 47.88 0.15 1660 33 Tr16-20 B1V 4.37 47.28 0.08 1180 8.8 Tr16-21 O8V 4.96 48.44 0.26 2100 91 Tr16-22 O8.5V 4.86 48.27 0.22 1960 67 Tr16-23 O7V 5.14 48.75 0.4 2320 170 Tr16-24 B2V 4.19 46.80 0.05 750 2.2 Tr16-27 B1V? 4.37 47.28 0.08 1180 8.8 Tr16-28 B2V 4.19 46.80 0.05 750 2.2 Tr16-29 B2V 4.19 46.80 0.05 750 2.2 Tr16-31 B0.5V 4.47 47.50 0.10 1460 17 Tr16-37 B1V? 4.37 47.28 0.08 1180 8.8 Tr16-74 B1V 4.37 47.28 0.08 1180 8.8 Tr16-76 B2V 4.19 46.80 0.05 750 2.2 Tr16-94 B1.5V 4.28 47.05 0.06 960 4.4 Tr16-115 O8.5V 4.86 48.27 0.22 1960 67 Tr16-122 B1.5V 4.28 47.05 0.06 960 4.4 Tr16-124 B1V 4.37 47.28 0.08 1180 8.8 Tr16-126 O9V 4.77 48.06 0.18 1820 47 Tr16-127 O9V 4.77 48.06 0.18 1820 47 Tr16-245 B0V 4.57 47.70 0.12 1510 22 Note — η Car’s hypothetical value for QH on the main sequence is taken by adopting the sameQH/LBolratioasforHD93129Aa.References.—Nelanetal.(2004);Massey&Johnson (1993);Walborn(1973,1995;2002,2005);Pennyetal.(1993);Repolustetal.(2004);Naz´eet al.(2005;forHD93161);Smithetal.(1996);Crowtheretal.(1995,2002;althoughQH andM˙ valuesfortheWNLstarsarefromCrowther2005,privatecomm.,modifiedtoincludeclumping inthewind). A Census of the Carina Nebula 5 farther away. The outermost filaments of the Carina region denseoptically-thickstellarwind2 (Hillieretal.2001;Smith spanadistanceof∼120pc(Smithetal.2000),yetTr14and etal.2003b)iseventuallyabsorbedbydustinthesurround- 16areprojectedonlyabout5pcfromeachotheronthesky, ing Homunculus nebula and converted into infrared radia- and even the KeyholeNebulanear Tr16 shows signs of ion- tion (Smith et al. 2003c; Cox et al. 1995). However, more ization from thenorthwest (i.e. from Tr14). than 104 years ago, before it entered its presently-observed Additional clusters that are sometimes associated with luminous blue variable (LBV) phase and was on the main Carina, such as NGC 3324 and NGC 3293, are ignored in sequence—orevenmorethan160yearsagobeforetheHo- the present analysis. They do not appear to lie within the munculusexisted—ηCarwasamuchstrongersourceofUV nebular structures that define the outermost boundaries of radiation. Indeed, it probably contributed ∼20% of the to- theCarinaNebula,andthestarsofNGC3324areassociated talionizingphotonsfortheentireregion(seebelow).Bythe with their own separate circular H iiregion. sametoken,itsmain-sequencestellarwindwasweakerthan its presently-observed LBV wind. Its spectral type prior to theLBV phaseis of course unknown,but given itsextreme luminosity,it is reasonable toassume that it was an O2su- pergiantlikeHD93129A whileonthemainsequence.From stellar evolution tracks for very massive stars (e.g., Young 3.1 Trumpler 16 2005; Maeder & Meynet 2000), it is likely that η Car was EnergyinputinCarinaisdominatedbythemassivecluster ∼0.1 dex less luminous at that time. The value for QH at Tr16,ofwhichtheinfamousmassivestarη Caristhemost that time is then calculated assuming this lower LBol, and luminous member. This is clear even before a detailed look assuming the same ratio of QH/LBol as for HD93129A. at its stellar content, since optical images of Carina show ItisalsolikelythatηCarhasaclosecompanionstarin numerous sculpted dust pillars and other elongated struc- a5.5yreccentricorbit(Daminelietal.2000;Corcoran2005; tures located across several degrees of the sky, and nearly Duncanetal.1995).WhileitsUVradiationdoesnotescape all of them point back toward η Car and thecenterof Tr16 theHomunculusatthepresenttime,itdidcontributetothe (Smith et al. 2000, 2003a, 2004, 2005; Walborn 2002). ionization of the larger region throughout the 2-3 Myr life- AsnotedbyWalborn(1995)andothers,Cr228andCr timeofTr16.ThenatureofηCar’scompanionstarishighly 232areprobablypartofTr16,wheretheirapparentsepara- uncertainandamatterofcurrentdebate(Pittard&Corco- tionontheskyisanartifactcausedbydustlanesonthenear ran2002;Ishibashietal.1999;Davidson1999;Duncanetal. side of the nebula. Thus, Table 2 includes Cr 228 and 232 1995);atentativespectraltypeofO5VisadoptedinTable as members of a larger Tr16 cluster. Table 2 also includes 2.Thebolometricluminosity ofη Car(theprimarystar)in the O5 V star HDE 303311, which is part of Cr232, as a Table 2 has thecontribution of its companion subtracted. member of Tr16, even though its membership is somewhat uncertain (Walborn 1995). Inthis and othercases, whether thisstarisconsideredpartofTr16orTr14hasnoneteffect 3.2 Trumpler 14 ontheglobalenergeticsoftheregion,butitmayinflatethe Trumpler 14 is a smaller, much more compact, and per- relative importance of Tr16. haps somewhat younger cluster than Tr16. Some members AlsoincludedwithinTr16inTable2arethethreeWNL of Tr16, like HD 93161, 93160, and 93250, are projected stars in theregion. HD93162 is almost certainly amember nearby on the sky (Walborn 1973) and are often mistaken ofTr16.ThetwootherWNLstars,HD92740andHD93131, as members of Tr14. As noted above, it is likely that Tr14 arequiteabitfartherawayfromη Caronthesky.Walborn is perhaps 10–20 pc more distant than Tr16, since silhou- (1995) notes that they are both roughly equidistant (∼20 ette objects seen in front of it point toward Tr16 (Smith et pc) from η Car, and that they could easily have traveled al. 2003), but it is still part of the central engine that pow- thisfar in 2 Myrwith a kickof ∼10 km s−1. Previously,on ers theCarina Nebula. Its most conspicuous member is the the main-sequence, they presumably had O2 spectral types prototypical O2 If* supergiant HD 93129A. likeHD93129A,andsomewhatdifferentQH values.Inorder Because of its compact size, its relatively faint upper to account for this stellar evolution, the cumulative “main- main sequence, and certain spectral diagnostics, Tr14 may sequence” values in Table 7 (see §4) were calculated using be 1-2 Myr younger than Tr16 (Walborn 1995). Penny et QH values lowered by 0.1 dex for each WNL star. This al.(1993)estimateagesyoungerthan1MyrforHD93129A is roughly correct if they were ∼0.1 dex less luminous on andHD93128,whileVazquezetal.(1996)findaclusterage themain-sequence(Maeder&Meynet2000),withthesame of ∼1.5 Myr. In addition, Tr14 remains closer to its natal QH/LBol ratio as HD93129A. molecular cloud than Tr16 (e.g., Brooks et al. 2003). TheageofTr16isprobablyabout2-3Myr.Thiscomes The stellar content summarized in Table 3 is compiled from thefact thatit containsH-richWNLstars butnoHe- from several references listed at the end of the table, with richWRstars,andthatηCarhasalikelyZAMSmass>100 particular attention to recent results on the binarity and M⊙andhasalreadyevolvedoffthemain-sequence—aswell multiplicity of stars in the cluster (see Nelan et al. 2004; as the fact that there is no clear evidence for a supernova Walborn 2005), including a new companion to HD 93129A havingexploded yet. (Nelanetal.2004).Thegreatestsourceofuncertaintyinthe When considering the cumulative effect of the massive netoutputofthisclusterisintheparametersforHD93129A stars in Tr16, we must bear in mind the peculiar present stateofitsmost luminousmember,η Car.Asweobserveit today,η CarisapitifulsourceofUVphotonsforthelarger 2 Themass-lossrateforηCarlistedinTable2isfromHillieret region.Whateverionizingradiationisabletoescapeitsvery al.(2001), andincludestheeffects ofclumpinginthewind. 6 N. Smith Table 3.StellarContentofTrumpler14 Name Sp. logL logQH M˙ V∞ LSW Type (L⊙) (s−1) (10−6 M⊙ yr−1) (kms−1) (L⊙) HD93129Aa O2If* 6.17 49.96 8.32 3200 6780 HD93129Ab O3.5V 5.76 49.54 2.5 3080 1870 HD93129B O3.5V((f+)) 5.76 49.54 2.5 3080 1870 HD93128 O3.5V((f+)) 5.76 49.54 2.5 3080 1870 CPD-582611 O6V((f)) 5.32 48.99 0.8 2560 410 CPD-582620 O6.5V((f)) 5.23 48.88 0.6 2460 290 Tr14-3 B0.5IV-V 4.50 47.70 0.15 1460 26 Tr14-4 B0V 4.57 47.70 0.12 1510 22 Tr14-5 O9V 4.77 48.06 0.18 1820 47 Tr14-6 B1V 4.37 47.28 0.08 1180 8.8 Tr14-9 O8V 4.96 48.44 0.26 2100 91 Tr14-18 B0V 4.57 47.70 0.12 1510 22 Tr14-21 O9V 4.77 48.06 0.18 1820 47 Tr14-27 O9V 4.77 48.06 0.18 1820 47 Tr14-30 B0III-IV 4.90 48.10 0.25 1510 45 References. — Nelan et al. (2004); Massey & Johnson (1993); Walborn (1973, 1995; 2005); Penny et al. (1993);Repolustetal.(2004). Table 4.StellarContentofTrumpler15 Name Sp. logL logQH M˙ V∞ LSW Type (L⊙) (s−1) (10−6 M⊙ yr−1) (kms−1) (L⊙) HDE303304 O8V 4.96 48.44 0.26 2100 91 HD93249 O8II/O9III 5.28 48.75 0.7 1960 210 HD93342 O9III 5.21 48.65 0.5 1820 130 Tr15-18 O9I-II 5.61 49.11 2.5 1660 540 Tr15-2 O9III 5.21 48.65 0.5 1820 130 Tr15-19 O9V 4.77 48.06 0.18 1820 47 HD93190 B0IV 4.83 48.00 0.21 1510 38 Tr15-3 B2V 4.19 46.80 0.05 750 2.2 Tr15-4 B1V 4.37 47.28 0.08 1180 8.8 Tr15-7 B2V 4.19 46.80 0.05 750 2.2 Tr15-9 B1V 4.37 47.28 0.08 1180 8.8 Tr15-10 B2V 4.19 46.80 0.05 750 2.2 Tr15-13 B1V 4.37 47.28 0.08 1180 8.8 Tr15-14 B2V 4.19 46.80 0.05 750 2.2 Tr15-15 B0.5IV-V 4.47 47.50 0.10 1460 17 Tr15-20 B1V? 4.37 47.28 0.08 1180 8.8 Tr15-21 B0III 5.08 48.28 0.3 1510 54 Tr15-23 B0V 4.57 47.70 0.12 1510 22 Tr15-26 B0.5V 4.47 47.50 0.10 1460 17 References.—Feinsteinetal.(1980);Morrelletal.(1988). itself. Thevalueof log L =6.17 is adopted from Repolust 3.4 Bochum 10 Bol et al. (2004), and the value for QH was calculated from Bo10isarelativelymeagerclusterlocatedroughly40′north- the luminosity using the revised T calibration scale of eff westofTr16,withahandfulofOBstarsandaprobableage Martins et al. (2005). of ∼7 Myr(Feinstein 1981; Fitzgerald et al. 1987). Thus, it may be significantly older than Tr 14 and 16, but perhaps not much older than Tr15. Feinstein found a large distance of3.6 pc,butalateranalysis byFitzgerald etal. (1987) fa- 3.3 Trumpler 15 voredacloserdistanceof2.5kpc,placingitinsidetheCarina Tr15 is an open cluster located about 20′ north of Tr16 on complexassociatedwithsomenearbynebulosity.Becauseof the sky. It is thought to be at the same distance and has itslargerageof∼7Myr,itislikelythatthenearbyWRstar roughly the same reddening as Tr16, but it is somewhat HD92809maybeassociatedwithBo10,oratleastwiththe older, with a likely age of 6±3 Myr (Feinstein et al. 1980; larger Carina complex. Therefore, HD 92809 is included in Morrell et al. 1988). It is seen near a very bright red star Table5alongwiththeothermembersofBo10.Forthetotal on thesky – the M2 Ia supergiant RT Car, which is proba- FUV luminosity of HD92809 in Table 7 (see §4), we adopt bly not a memberof thecluster. Table 4 adopts thecluster LFUV/LBol=0.39 (for 912–3650 ˚A) following the study of members listed by Feinstein et al. (1980) and Morrell et al. Smartt et al. (2001). (1988),includingtheO8VstarHD303304whichwaslisted as a possible nonmember by Feinstein et al. (1980), as well theO9giantHD93342andthehighly-reddenedO9I-IIstar 3.5 Bochum 11 Tr15-18,whichwerelistedaspossiblemembers.Table4does notincludethenon-memberO5IIIstarHD93403,whichis Bo11isalooseopenclusterinthesoutheasternCarinaNeb- taken to bea member of Tr16. ula.Fitzgeraldetal.(1987)findanageforBo11oflessthan A Census of the Carina Nebula 7 Table 5.StellarContentofBochum10 Name Sp. logL logQH M˙ V∞ LSW Type (L⊙) (s−1) (10−6 M⊙ yr−1) (kms−1) (L⊙) HD92809 WC6 5.32 49.20 16 2280 6620 HD92725 O9.5III 5.15 48.42 0.4 1660 87 HD92964 B2Ia 5.30 48.35 0.5 750 22 HD92759 B0III 5.08 48.28 0.3 1510 54 HD92894 B0IV 4.83 48.00 0.21 1510 38 HD93002 B0IV 4.83 48.00 0.21 1510 38 HDE302989 B2V 4.19 46.80 0.05 750 2.2 HDE303190 B1V 4.37 47.28 0.08 1180 8.8 HDE303296 B1V 4.37 47.28 0.08 1180 8.8 HDE303297 B1V 4.37 47.28 0.08 1180 8.8 Bo10-2 B0.5IV 4.74 47.80 0.18 1460 30 Bo10-7 B2IV 4.49 47.12 0.08 750 3.5 Bo10-15 B1V 4.37 47.28 0.08 1180 8.8 References.—Feinstein(1981);Fitzgeraldetal.(1987);Prinjaetal.(1990);Smarttetal.(2001). Table 6.StellarContentofBochum11 Name Sp. logL logQH M˙ V∞ LSW Type (L⊙) (s−1) (10−6 M⊙ yr−1) (kms−1) (L⊙) HD93632 O4-5III(f) 5.79 49.56 4.0 2870 2600 Bo11-2 O9V 4.77 48.06 0.18 1820 47 Bo11-3 B2V 4.19 46.80 0.05 750 2.2 Bo11-4 B0.5V 4.47 47.50 0.10 1460 17 Bo11-5 O9V 4.77 48.06 0.18 1820 47 Bo11-9 O9IV 4.99 48.36 0.34 1820 89 Bo11-10 O9IV 4.99 48.36 0.34 1820 89 References.—Fitzgeraldetal.(1987);Walborn(1973). 3 Myr. They also find that the cluster includes several pre- models typically have LFUV/LBol values of 0.39–0.85 for main-sequencestars,consistent with anage asyoungas0.3 912–3650 ˚A (e.g., Smartt et al. 2001; Crowther et al. 2002; Myr. This very young age is interesting in light of the fact Crowther 2005, private comm.). Three cases are given for that Bo11 is located amid what is currently the most ac- Tr16 and for thetotal of all clusters: tiveregionofstarformationinCarina–theso-calledSouth ThefirstcasecorrepondstothehistoryofCarinaupun- Pillars (Smith et al. 2000; Rathborne et al. 2004). The age tilrecenttimes,whenηCaranditscompanionwerenotsur- and stellar content is similar to the nearby embedded clus- roundedbyadustshellanddidnothaveadenseLBVwind terknown astheTreasure Chest (Smithet al. 2005; H¨agele chokingoff theLymancontinuumflux,and whentheWNL et al. 2004).3 There is considerable discrepancy as to the stars in Tr16 were O2 stars. For this first case, there was a spectral type of the most luminous star in Bo 11, which is totalof70O-typestarsinCarina,producingQH=1.15×1051 HD93632orBo11-1.Fitzgeraldetal.(1987)listthisstaras s−1. This would be the appropriate number to adopt when O8 Iap, whereas Walborn (1973) classifies it as O4-5 III(f). consideringthehistoryandformationofthenebula,thelife- Walborn(1973) notedthatthespectraltypewassomewhat times of evaporating proplyds, globules, and dust pillars, variable, while Fitzgerald et al. (1987) noted a few prob- triggered star formation by radiative driven implosion, and lems with their photometric classification of stars in Bo11. the growth of the cavity that will blow out of the Galactic A spectral typeof O4-5 III(f) is adopted in Table 6. plane as a bipolar superbubble. The second case is the LBV phase of η Car during the last∼104 yr,whenηCarhadadensewindthatreprocessed 4 CUMULATIVE EFFECTS most of the Lyman continuum radiation into the Balmer continuum (FUV). During this phase, η Car was not sur- Table 7 lists the total luminosity, ionizing flux, total FUV roundedbya densedust shell, allowing itsFUV luminosity (Balmer continuum) luminosity, mass loss, and mechanical to escape. Since the UV radiation of its companion could luminosity for each cluster, as well as the cumulative total also escape, and since the WNL stars were no longer on of all clusters for each parameter. For the FUV luminosity the main sequence, the total number of O stars is listed as (in all cases except where noted above), L ≃0.5L is FUV Bol 66 in Table 7. This brief phase may be important in un- assumed, which isadequateconsidering theinherentuncer- derstandingtherecentphotoevaporationofneutralglobules taintyinthebolometricluminosityasafunctionofspectral and photodissociation regions throughout the nebula, since type.Brooksetal.(2003)foundasimilarresultfortheFUV η Car was such an incredibly luminous FUV source. luminosity of Tr14 alone. More detailed stellar stmosphere The third case represents the currently-observed state of the Carina Nebula, when η Car and its companion are 3 Although it is not discussed in detail here, the O9.5 V star surrounded by an obscuring dust shell, blocking all their CPD-59◦2661 inthe Treasure Chest cluster is included inTable contribution to the total ionizing flux and FUV luminos- 7. ity. With η Car’s companion blocked by the Homunculus, 8 N. Smith Table 7.Total StellarEnergyInput Cluster Numberof logL logQH logL(FUV) M˙ LSW Ostars (L⊙) (s−1) (L⊙) (10−6 M⊙ yr−1) (L⊙) Tr16(MS) 47 7.215 50.91 6.91 91 45400 Tr16(LBV) 43 7.240 50.78 7.05 1083 67000 Tr16(now) 42 7.240 50.77 6.79 1083 67000 Tr14 10 6.61 50.34 6.31 18.7 13500 Tr15 6 6.18 49.56 5.88 5.9 1300 Bo10 1 6.00 49.42 5.69 18.3 7120 Bo11 5 6.00 49.64 5.70 5.2 2900 CPD-59◦2661 1 4.68 47.88 4.38 0.15 33 TOTAL(MS) 70 7.38 51.06 7.08 139 70200 TOTAL(LBV) 66 7.40 50.97 7.18 1131 91800 TOTAL(now) 65 7.40 50.96 7.00 1131 91800 the effective number of O stars is reduced to 65. For this third case, the cumulative ionization source is QH=9×1050 s−1. This would be the appropriate number to adopt when considering the current UV flux incident upon evaporating proplyds,globules,andirradiatedjetsovermostoftheneb- ula. For most of its lifetime, gas and dust in the Carina Nebulahasbeenexposedtoanionizingluminosityabout150 timesstrongerthanthat oftheOrion nebula.This changed 160 years ago when η Car ejected a thick dust shell that cut off essentially all of its UV output,and thetotal QH of theCarinaNebuladroppedbyabout20%duetothelossof ionization from theregion’s most luminous member.4 This change illustrates the dominant role of the most luminous star in any region, and highlights a truly unique propertyoftheCarinaNebula—itoffersalaboratorywhere wecanaddressvariabilityinQH overshorttimescales.This behaviorhasnotbeenwitnessedanywhereelseinourGalaxy to a similar degree. We can observe some of the effects of this variable UV output — for example, neutral cometary globules and dust pillars that point at η Car are seen only in silhouettetoday,buttheirshapesuggests thattheywere Figure1.IonizingphotonfluxΦasafunctionofradius.Thetwo formed by a strong UV flux from η Car in the past (Smith solidcurvesareforCase1and3(i.e.ηCaronthemainsequence et al. 2003a). It is interesting to note that the outer edges andatthepresenttime). of theCarina Nebula, beyond a radius of 160 light yror 50 pc (1.5×1020 cm) from η Car, have not yet seen the Great Eruption of η Car, and still see its pre-eruption UV out- sincetheircollectivecontributionislessthantheuncertainty put.Also,ifthecomplexnestedcircumstellarejectaaround η Carareafairindication,thistypeofUVcutoffhasprob- in the value for QH from the most luminous members. For theotherluminousmembers,inCase1(whenη Carwason ably happened multiple times in the past due to previous outbursts, and may yet happen again. the main sequence), about 8% of the total QH came from theO2supergiantHD93129Aaalone,15%camefromthe3 The main uncertainty in deriving the pre-eruption UV WNL stars, and 15% from the remaining fiveO3.5 V stars. outputofCarinaisinthespectraltypeandionizingluminos- Forcase3,correspondingtothepresentstateofaffairs,these ityofη Caritself, which hasalikelyuncertaintyofperhaps contribution have changed to about 10% of the total from ±15%(roughly3%ofthetotalfortheregion)ifthemodels HD 93129Aa alone, 23% from the 3 WNL stars, and 19% at lower luminosity and later spectral types are correct. In from theremaining fiveO3.5V stars. terms of the cumulative effect of the entire stellar popula- tion, then, dwarf stars later than O9 can be safely ignored, Figure 1 shows the ionizing photon flux ΦH as a func- tionofradius.ThetwosolidcurvesareforCase1and3(i.e. η Car on the main sequence and at the present time). Be- yondaradiusofroughly10pcfromthecenterofthenebula, 4 Actually, the change might not have been so sudden, since theionizingsourcecanbetreatedroughlyasapointsource, muchoftheLymancontinuumfluxfromηCarmayhavealready droppedwhenitdevelopedathickstellarwinduponenteringthe so that the photon flux drops as r−2. Within a few parsecs LBVevolutionaryphaseinthelast104 yr.Duringthistime,the of any cluster, however, the local radiation from that clus- majorityofη Car’sradiationescaped asFUVphotons. ter becomes more important than the diminished collective A Census of the Carina Nebula 9 sametimewasover350timesmorethanthemechanicalen- ergy.Thissituationwillchangedramaticallywhenthemost massivestarsbegintoexplodeassupernovaeinthenextfew 105–106 yr, each injecting an additional 1051–1052 ergs in a veryshort time. With a total mass-loss rate of M˙ ≈ 1.4 ×10−4 M⊙ yr−1,themassejectedbystellarwindsoverthepast3Myris about420M⊙.Thisisonlyabout10%ofthemassofionized gasestimatedtobefillingtheinterioroftheHiiregion(e.g., Walborn2005). Thus,themajorityofionizedgasinsidethe H ii region cavity probably comes from photoevaporative flows(e.g.,Bertoldi1989)offthesurfacesofmolecularclouds andevaporatingglobulesembeddedwithintheregion.Note that the present value of M˙ for η Car’s brief LBV wind phaseismuchgreaterthanthecumulativemassloss ofrest of thestars in Carina combined. 5 COMPARISON WITH OTHER MASSIVE CLUSTERS Finally, it is interesting to see how the stellar energy input Figure 2. Same as figure 1, but for FUV radiation, plotted in in Carina stacks up to other young massive star clusters. unitsoftheHabingflux,G0=1.6×10−3ergss−1cm−2.Thethree As noted earlier, in general, Carina’s overall stellar con- casesfromTable7areshown. tentissimilar tothemostwell-studied massiveclusterslike NGC3603, Arches, and 30 Dor, except that these clusters are more compact than Carina. For example, Crowther & effectsfromallclusters.Thisisduetothefactthatthestel- Dessart (1998) list log QH=51.19 and LSW=1.5×105 L⊙ for NGC3603. The total energy input in Carina is essen- lar population in Carina is a loose aggregate, rather than a tiallythesameasthesevalues,becauseCrowther&Dessart single dense cluster. The specific case of Tr14 is shown in Figure1,butwithinafewparsecsofanyindividualcluster, usedanolderTeff scale, leadingtohigher QH valuesfor O stars, and did not account for wind clumping in their anal- itsownionizationwilldominate,whiletheinfluenceofmore ysis; if those same assumption were made here for Carina, distant clusters diminishes. the results would be identical to NGC3603 within the un- Figure 2 shows the radial dependence of the FUV ra- certainties. Carina has about 25–30% of the total ionizing diation field intheCarinaNebula,compared totheHabing flux,G =1.6×10−3 ergss−1 cm−2(Habing1968).Thethree fluxofthemoremassiveArchesclusterintheGalactic cen- 0 ter (Figer et al. 2002), and roughly 25% of both the total curves correspond to η Car on the main sequence, in the LBVphase,andinitspresentstateobscuredbyitsowndust QH and LSW of R136 in 30 Doradus given by Crowther & Dessart (1998). Given the similar cumulative energy input shell.ThestrongestFUVfluxesintheregionarisewhenthe andstellarcontent,itisinterestingthatCarina’sclustersare LBVphaseofηCarisconsidered,whenmostofitsbolomet- lesstightlyboundthantheothers,especiallyifthisgivesrise ric luminosity escaped in the Balmer continuum. Through- todifferent effectson thenearbystar-forming environment. outtheCarinaNebula,thisresultedinalocalFUVfluxthat A future paper will investigate how the cumulative stellar was roughly 75% higher than the present state when η Car energy input assessed directly from the observed massive is surrounded by dust. Thus, for extremely massive stars, stars in Carina compares with more indirect tracers of the theirchangingUVoutputinvariousevolutionarystatescan energy budget of thelarger Carina Nebula. profoundly effect the radiation field of their environments, and consequently, the evaporation of protoplanetary disks ACKNOWLEDGMENTS and the erosion of dust pillars. Again, however, individual IthankPaulCrowtherforadetailedrefereereportandseveralhelpful clustersdominateoverthecollectiveradiationfieldwithina suggestionsregardingmodelparameters.IalsothankNolanWalbornand few parsecs of any cluster core. Kate Brooks for useful discussionsand helpfulcomments and corrections The present total mechanical luminosity from stellar on the manuscript. Support was provided by NASA through grant HF- 01166.01A from the Space Telescope Science Institute, which is operated winds is just under 105 L⊙, or 7×104 L⊙ during the main- bytheAssociationofUniversitiesforResearch inAstronomy,Inc.,under sequence phase, having contributed a total energy of about NASAcontractNAS5-26555. 2.6×1052 ergssofarduringthe3Myrlifetimeoftheregion. Thisishigherthanapreviousestimateofseveral×1051 ergs (Walborn 1982) because it includes the influence of many REFERENCES more stars. The estimated kinetic energy of the expanding Allen,D.A.,&Hillier,D.J.1993,PASA,10,338 proto-superbubble(Smithetal.2000;Walborn1982)isonly Bertoldi,F.1989,ApJ,346,735 5–30%ofthismechanicalenergyfromstellarwinds,suggest- Brooks,K.J.,Burton,M.G.,Rathborne,J.M.,Ashley,M.C.B.,& ing that most of the energy injected by stellar winds is ra- Storey,J.W.V.2000,MNRAS,319,95 diated away, even in the early phases of a bubble’s growth. Brooks, K.J., Cox, P., Schneider, N., Storey, J.W.V., Poglitsch, The total energy radiated by early-type stars during the A.,Geis,N.,&Bronfman,L.2003,A&A,412,751 10 N. Smith Brooks,K.J.,Garay,G.,Nielbock,M.,Smith,N.,&Cox,P.2006, Rathborne,J.M.,Brooks,K.J.,Burton,M.G.,Cohen,M.,&Bon- ApJ,inpress temps,S.2004,A&A,418,563 Brooks,K.J.,Whiteoak,J.B.,&Storey,J.W.V.1998,PASA,15, Repolust,T.,Puls,J.,&Herrero,A.2004,A&A,415,349 202 Smartt,S.,etal.2001,MNRAS,325,257 Corcoran,M.F.2005,AJ,129,2018 Smith,L.F.,Shara,M.M.,&Moffat,A.F.J.1996, MNRAS,281, Cox,P.,&Bronfman,L.1995,A&A,299,583 163 Cox, P., Mezger, P.G., Sievers, A., Najarro, F., Bronfman, L., Smith,N.2002a, MNRAS,337,1252 Kreysa,E.,&Haslam,G.1995,A&A,297,168 Smith,N.2002b, MNRAS,331,7 Crowther, P.A. 2005, IAU Symp. 227, Massive Star Birth: A Smith, N., Egan, M.P., Carey, S., Price, S.D., Morse, J.A., & Crossroads of Astrophysics, ed. R. Cesaroni, E. Churchwell, Price,P.A.2000,ApJ,532,L145 M.Felli,&C.M.Walmsley(SanFrancisco:ASP),inpress Smith,N.,Bally,J.,&Morse,J.A.2003a,ApJ,587,L105 Crowther,P.A.,&Dessart,L.1998,MNRAS,296,622 Smith,N.,Davidson,K.,Gull,T.R.,Ishibashi,K.,&Hillier,D.J. Crowther, P.A., Hillier, D.J., Evans, C.J., Fullerton, A.W., De 2003b,ApJ,586,432 Marco,O.,&Willis,A.J.2002,ApJ,579,774 Smith,N.,Gehrz,R.D.,Hinz,P.M.,Hoffmann,W.F.,Hora,J.L., Crowther, P.A., Smith, L.J., Hillier,D.J., & Schmutz, W. 1995, Mamajek,E.E.,&Meyer,M.R.2003c,AJ,125,1458 A&A,293,427 Smith,N.,Bally,J.,&Brooks,K.J.2004,AJ,127,2793 Davidson,K.1999,inASPConf.Ser.179,304 Smith,N.,Stassun,K.G.,&Bally,J.2005,AJ,129,888 Tapia,M.,Roth,M.,Marraco,H.G.,&Ruiz,M.T.1988,MNRAS, Damineli,A.,Kaufer, A.,Wolf, B.,Stahl, O.,Lopes, D.F.,& de 232,661 Araujo,F.X.2000,ApJ,528,L101 Va´zquez,R.A.,Baume,G.,Feinstein,A.,&Prado,P.1996,A&A, Duncan, R.A., White, S.M., Lim, J., Nelson, G.J., Drake, S.A., 116,75 &Kundu,M.R.1995,ApJ,441,L73 Walborn,N.R.1973,ApJ,179,517 Feinstein,A.1981,PASP,93,202 Walborn,N.R.1976,ApJ,204,L17 Feinstein,A.1995,RevMexAASer.Conf.,2,57 Walborn,N.R.1982,ApJS,48,145 Feinstein, A., Marraco, H.G., & Muzzio, J.C. 1973, A&AS, 12, Walborn,N.R.1995,RevMexAASer.Conf.,2,51 331 Walborn, N.R. 2002, in ASP Conf. Ser. 267, Hot Star Work- Feinstein, A., Fitzgerald, M.P., & Moffat, A.F.J. 1980, AJ, 85, shop III: The EarliestPhases of MassiveStar Birth, ed. P.A. 708 Crowther(SanFrancisco:ASP),111 Figer,D.F.2005, Nature,34,192 Walborn, N.R. 2005, in the online proceedings of Figer, D.F., Kim, S.S., Morris, M., Serabyn, E., Rich, R.M., & the workshop, Eta Car: Reading the Legend, McLean,I.S.1999,ApJ,525,750 http://www.astro.washington.edu/balick/eta conf. Figer,D.F.,Najarro,F.,Gilmore,D.,Morris,M.,Kim,S.S.,Ser- Walborn,N.R.,&Liller,M.H.1977,ApJ,211,181 abyn, E.,McLean, I.S.,Gilbert,A.M.,Graham, J.R.,Larkin, Walborn,N.R.,etal.2002,AJ,123,2754 J.E.,Levenson, N.A.,&Teplitz,H.I.2002,ApJ,581,258 Weidner,C.,&Kroupa,P.2004, MNRAS,348,187 Fitzgerald,M.P.,&Mehta,S.1987,MNRAS,228,545 Young,P.2005,ASPConf.Ser.332,190 Habing,H.J.1968, Bull.Astron.Neth.,19,421 Ha¨gele, G.F., Albacete Colombo, J.F., Barba, R.H., & Bosch, G.L.2004, MNRAS,355,1237 Hillier,D.J.,Davidson,K.,Ishibashi,K.,&Gull,T.R.2001,ApJ, 553,837 Ishibashi,K.,etal.1999,ApJ,524,983 Kroupa, P., & Weidner, C. 2005, IAU Symp. 227, Massive Star Birth: A Crossroads of Astrophysics, ed. R. Cesaroni, E. Churchwell,M.Felli,&C.M.Walmsley(SanFrancisco:ASP), inpress Lopez,J.A.,&Meaburn,J.1986,RevMexAA,13,27 Maeder,A.,&Meynet, G.2000, ARAA,38,143 Martins,F.,Schaerer,D.,&Hillier,D.J.2005, A&A,436,1049 Massey,P.,&Hunter,D.A.1998,ApJ,493,180 Massey,P.,&Johnson,J.1993, AJ,105,980 Megeath, S.T., Cox, P., Bronfman, L., & Roelfsema, P.R. 1996, A&A,305,296 Mizutani,M.,Onaka,T.,&Shibai,H.2004,A&A,423,579 Moffat,A.F.J.,etal.2002, ApJ,573,191 Morrell,N.,Garcia,B.,&Levato, H.1988,PASP,100,1431 Najarro, F., Figer, D.F., Hillier, D.J., & Kudritzki, R.P. 2004, ApJ,611,L105 Naz´e,Y.,Antokhin,I.I.,Sana,H.,Gosset,E.,&Rauw,G.2005, MNRAS,359,688 Nelan,E.P.,Walborn,N.R.,Wallace,D.J.,Moffat,A.F.J.,Maki- don,R.B.,Gies,D.R.,&Panagia,N.2004,AJ,128,323 Oey,M.S.,&Clarke,C.J.2005, ApJ,620,430 Penny,L.R.,Gies,D.R.,Hartkopf,W.I.,Mason,B.D.,&Turner, N.H.1993,PASP,105,588 Pittard,J.M.,&Corcoran,M.F.2002,A&A,383,636 Prinja,R.K.,Barlow,M.J.,&Howarth,I.D.1990,ApJ,361,607 Rathborne,J.M.,Burton,M.G.,Brooks,K.J.,Cohen,M.,Ashley, M.C.B.,&Storey,J.W.V.2002, MNRAS,331,85