*2.1. The Milky Way*

The first survey for Wolf–Rayet stars (inadvertently) began in 1867 when Charles Wolf and Georges Rayet were examining spectra of stars in Cygnus using a visual spectrometer on the 40-cm Foucault telescope at the Paris Observatory. They came across three very unusual stars. While the spectra of most stars are dominated by absorption lines, these stars had mysterious strong, broad emission lines. (These stars were later designated and classified as HD 191765, WN5; HD 192103, WC8; and HD 192641, WC7.)

The correct identification of the spectral features was lacking for nearly 60 years after their discovery: it was Carlyle Beals, a Canadian astronomer, who correctly identified the lines as due to ionized helium, nitrogen, and carbon [13]. The width of these lines were understood as being due to

Doppler broadening of thousands of km s<sup>−</sup>1, a result of the outflow rates of the strong stellar winds in the formation region of these lines [14–16]. Example spectra are shown in Figure 1.

**Figure 1.** The spectra of two of the first discovered WR stars. Left: HD 191765 is a WN star, with unusually broad and strong lines. Its classification is a "WN5" subtype. Right: HD 192103 is a WC star, with a "WC8" subtype.

WN-type Wolf–Rayet stars are further classified primarily by the relative strengths of N III *λ*4634,42, N IV *λ*4058, and N V *λ*4603,19, while the classification of WC-type WRs is based upon the relative strengths of O V *λ*5592, C III *λ*5696, and C IV *λ*5806,12. The system was first proposed by Lindsey Smith [17], although some extension to earlier and later types of WNs have been made by others [18,19]; a classification scheme for WO stars was developed by Paul Crowther and collaborators [20]. As with normal stars, a lower number is indicative of higher excitation, i.e., WN2 (hotter) vs. WN9 (cooler), WC4 (hotter) vs. WC9 (cooler), WO1 (hotter) vs. WO4 (cooler).

The late-type WNs are morphologically similar to O-type supergiants, known as "Of-type" type stars, in that the latter show N III *λ*4634,42 and He II *λ*4686 emission, also the result of stellar winds. The late-type WNs are more extreme, however, with stronger lines. In general, WNs (and WRs in general) do not show absorption lines; rather, all of the lines are formed in the stellar winds. There are, however, exceptions, such as HD 92740, a singled-lined WR binary in which the emission and absorption move together in phase [21]. It was the similarity between Of-type and WNs that led in part to the Conti scenario [2].

As summarized in [22], a total of 52 similar stars were discovered by Copeland, Fleming, Pickering, and Respighi in the 25 years that followed Wolf and Rayet's discovery. These findings, and early visual work by Vogel in 1885, and photographic studies of their spectra by Pickering in 1890, are discussed in the contemporary review by Julius Scheiner and Edwin Frost in their 1894 publication *A Treatise on Astronomical Spectroscopy* [23]. William Campbell (who served as director of Lick Observatory 1901–1930) published the first catalog of these 55 Galactic WRs in 1894 [24]. Additional WRs were discovered as by Williamina Fleming, Annie J. Cannon and coworkers as part of the Henry Draper catalog project, and accidental discoveries continued to be made over the years. The first modern catalog of Galactic Wolf–Rayet stars compiled by Karel van der Hucht and collaborators in 1981 [18]. Titled "The VIth Catalog" (Campbell's was considered the first), the work included extensive bibliographies and references to earlier studies. This catalog contained 168 WRs. The next edition, in 2001 [19], listed 227 spectroscopically confirmed Galactic WRs, with an "annex" in 2006 [25] bringing the number known to 298. The most-up-to-date catalog of Milky Way WRs is maintained online by Paul Crowther1, which contained 661 entries as of of this writing, June 2019.

<sup>1</sup> http://www.pacrowther.staff.shef.ac.uk/WRcat/

Systematic searches for WRs in the Milky Way are stymied by the vast angular extent that needs to be examined (the entire sky!), and variable and sometimes high reddening. The Henry Draper catalog is probably complete down to an apparent magnitude of 10th or 11th, except in regions of crowding. Spectroscopic surveys of young clusters or OB associations reveal additional WR finds yearly; possibly the most extreme example is that of Westerlund 1 and various open clusters near the Galactic Center; see [25] and references therein. However, the large increase in the number of WR stars known in the Galaxy in the past 15 years has has come about primarily from the use near- and mid-IR colors to identify WR candidates [26–30], a method first pioneered by Schuyler van Dyk and Pat Morris, plus the use of narrow-band IR imaging in the K-band [31,32], pioneered by Mike Shara. Optical or near-IR spectroscopy is then used to confirm the color-selected candidates.

With the advent of *Gaia*, it is now possible for the first time to actually derive distances to many of these Wolf–Rayet stars. However, difficulties of constructing meaningful volume-limited samples remain for Galactic studies. As discussed later, WN-type WRs are harder to find than WC-type due to their weaker lines; at the same time, WC stars may be dustier (and thus fainter) than WN stars in the same location. They also cover a limited range in metallicity compared to what can be achieved by using the non-MW members of the Local Group. Finally, observations of Galactic WRs may be more difficult due to reddening than those in much further, but less reddened, regions. Thus, Galactic studies still have limited value for testing models of stellar evolution theory. Thus, for the rest of this review, we will focus on the WR content of galaxies outside our own.

### *2.2. Early Searches for Extra-Galactic WRs*
