**3. Massive Star Content**

The integrated nebular properties of the Tarantula Nebula and other selected giant H II regions in the Local Group is presented in Table 2. The ionizing output of the Tarantula Nebula corresponds to the equivalent of over a thousand O7V stars, each with 10<sup>49</sup> ph s<sup>−</sup>1. In reality, the Tarantula hosts somewhat fewer O-type stars since the most extreme examples—early O stars and luminous Wolf-Rayet (WR) stars—each produce an order of magnitude more Lyman continuum photons. Nevertheless, this population represents an order of magnitude more O and WR stars than any Milky Way or SMC giant HII regions, and is not likely to be improved upon until extremely large telescopes are capable of resolving the massive stellar content of more extreme giant HII regions, such as NGC 5461, 5462 and 5471 in M 101 [1].

**Table 2.** Integrated nebular properties of nearby giant HII regions, adapted from Kennicutt [20] for an assumed O7V Lyman continuum output of 10<sup>49</sup> ph s<sup>−</sup>1.


Historically, there have been several photometric and spectroscopic surveys of early-type stars in the Tarantula Nebula, each of which have employed contemporary (Galactic) temperature calibrations to produce Hertzsprung-Russell diagrams. Parker [21,22] obtained the first extensive study of the entire region. Subsequently, Massey and Hunter [23] obtained high-spatial resolution HST spectroscopy of early-type stars in the central, crowded R136 region, revealing an exceptional population of very early O stars, whilst Melnick and colleagues [24,25] obtained spectroscopy for a large number of early-type stars in the NGC 2070 region. The advent of efficient multi-object spectrographs on 8–10 m telescopes, has permitted the most comprehensive optical spectroscopic survey of massive stars in the Tarantula Nebula to date through the VLT FLAMES Tarantula Survey (VFTS) [2]. Multi-object, multi-epoch spectroscopy of ∼800 massive stars across the entire region, for which detailed spectroscopic analyses have been undertaken for over 500 O and early B stars, such that temperature calibrations are no longer necessary. Although this represents the most extensive of early-type stars in a single star-forming region to date, this survey is incomplete owing the fibre-placement limitations, sampling ∼70% of massive stars exterior to the dense R136 cluster from comparison with photometric surveys [12].

Two additional surveys have recently provided complete optical spectroscopic observations of all bright sources within the central crowded region of the Tarantula: (a) the central 4 arcsec (1 pc) of the R136 cluster exploiting the high spatial resolution of HST/STIS [5]; (b) the central 2 × 2 arcmin (30 × 30 pc) region of NGC 2070 using the VLT/MUSE integral field spectrograph [10]. Although these lack the multi-epoch capabilities of VFTS, they are complementary since the richest stellar populations of the Tarantula are found within NGC 2070/R136, and provide extended wavelength coverage (yellow and red for MUSE, ultraviolet for STIS), albeit at reduced spectral resolution. A summary of these three contemporary surveys is provided in Table 3, together with literature results. In total, approximately 1100 early-type massive stars have been spectroscopically observed. Our discussion of the massive star content of the Tarantula will largely draw upon results from VFTS, but include others where appropriate. It should be noted that analyses based on spectroscopic fibre-fed observations of early-type stars in regions of strong, highly variable nebulosity, is inherently problematic owing to the lack of local sky subtraction. Such issues do no affect long-slit STIS or integral field MUSE observations.


**Table 3.** Summary of stellar content of the Tarantula Nebula from recent spectroscopic surveys (excluding sources in common, although including individual components within SB2 binaries).

VFTS and other surveys have revealed that the Tarantula hosts extreme examples of stellar exotica, including the most massive stars known [9], a Luminous Blue Variable, R143 [27], a very massive runaway [28], the fastest rotating stars [29], a massive overcontact binary [30], and a supernova remnant N157B [31], with SN 1987A located ∼300 pc to the south west (Figure 1).

**Figure 1.** (**left**) Optical image of the Tarantula Nebula from the MPG/ESO 2.2m WFI, with NGC 2060 and SN1987A indicated; (**centre**) Optical VLT/FORS2 image centred on NGC 2070, with Hodge 301 to the upper right; (**right**) an infrared VLT/MAD image of the central R136 region, with the massive colliding wind binary Mk 34 indicated. Credit: ESO/P. Crowther/C.J. Evans.

R136a, the central cluster, merits special consideration since it was considered by some to be a supermassive star as recently as the early 1980s [32]. Speckle interferometry resolved R136a into multiple sources [33], and it was subsequently established as a compact star cluster [34]. Massey and Hunter [23] established that dozens of the brightest sources within the central parsec were hot, early O stars. Spectroscopic studies of the brightest components R136a1, a2, a3, with nitrogen-sequence Wolf-Rayet spectral types, indicated masses of ∼100 *M* [23,35]. They established that these relatively weak-lined WN stars are luminous main-sequence stars close to their Eddington limits, rather than classical Wolf-Rayet stars. Subsequent analyses of the WN stars in R136 indicated significantly higher masses of 150–300 *M* [9] as a result of increased spectroscopic luminosities, owing to higher stellar temperatures and IR photometry less affected by dust extinction. Indirectly inferred masses of massive stars are notoriously imprecise, and if binarity were established for individual stars their inferred luminosities and masses would be reduced. To date, faint companions to members of R136a have been detected with extreme adaptive optics imaging [36]. Melnick 34 is spectroscopically similar to the WN5-stars in R136a, and has recently been shown to be a colliding wind binary system comprising two WN5 components, with a total mass exceeding 250 *M* [37]. Figure 2 shows that Melnick 34 is an order of magnitude brighter than R136a in X-rays, indicating that there are no colliding wind binaries comparable to Melnick 34 within R136a [38].

**Figure 2.** (**left**) Chandra ACIS X-ray logarithmic intensity image of the core of NGC 2070 from T-ReX, centred on R136c, adapted from [38], showing the relative brightness of the colliding wind binary Melnick 34 (WNh5 + WN5h) [37] to the R136a star cluster (hosting multiple WN5h stars) and R136c (WN5h+?); (**right**) HST WFC3/F555W logarithmic intensity image of the same 19 × 19 arcsec region, highlighting the rich stellar population of R136a with respect to R136c and Melnick 34.

Overall, the Tarantula hosts a remarkable number of ∼50 early-type stars with bolometric luminosities exceeding <sup>10</sup>6*L*. For reference, the Milky Way's Carina Nebula (NGC 3372) hosts ≈5 massive stars with such extreme properties [39]. As such, the Tarantula Nebula represents our best opportunity to study the highest mass stars known, both individually and collectively. Schneider et al. [40] analysed VFTS spectroscopic results to establish an excess of massive stars with respect to a standard Salpeter Initial Mass Function (|MF), indicating 1/3 more stars with ≥30 *M* in 30 Doradus compared to expectations from a standard IMF. Finally, although we focus primarily on high mass early-type stars in the Tarantula Nebula, it also hosts red supergiants (RSG). Since RSG are the evolved descendants of moderately massive stars, and star formation in the Tarantula has peaked relatively recently, of order ∼10 RSG are known, most of which are associated with mature star clusters Hodge 301 and SL 639 [6].
