*3.4. Magellanic Clouds*

Thanks to previous surveys, such as the BAT99 catalog, the population of WRs in the MCs was thought to be complete. However, over the years, a few unexpected discoveries were made. Perhaps the most surprising of which was of a rare strong-lined WO discovered in the LMC in the rich OB association of Lucke–Hodge 41 [90]. Since the BAT99 catalog, six new WRs were discovered before the addition of this new WO suggesting that perhaps our knowledge of the WR content of the LMC was still not complete. Thus, a new search for WRs in the MCs was launched [79,91,92]. A summary of the results can be found in [38].

The overall process of this survey was similar to finding WRs in M31 and M33. The entire optical disks of both the LMC and SMC were observed using the 1-m Swope telescope on Las Campanas, with the three-filter interference system and then a combination of image subtraction and photometry was used to detect candidate WRs before they were spectroscopically confirmed.

In the SMC, no new WRs were discovered. However, this isn't too surprising given that there are only 12 known WRs in the entire galaxy [49], and that the Massey and Duffy survey had covered the entire galaxy. All of them are of WN type except one binary WO. Further characteristics, such as their physical properties and binary status, are discussed later.

The LMC, however, held many surprises. Overall, the new study found 15 new WRs bringing the total number of WRs in the LMC up to 152. Five of them were normal WNs that had been missed due to crowded fields and faint emission lines. However, ten of them were unlike any WR we had seen before.

The spectra of these stars contain absorption lines like that of a O3 star with emission lines like that of a WN3, thus leading to a designation of WN3/O3s [93]. A spectrum of one such star showing both the narrow absorption lines and broad emission lines is shown in Figure 9. While their spectra initially suggests binarity, these stars are simply too faint to be WN3 + O3V binaries. The absolute magnitude of an O3V by itself is *MV* ∼ −5.5 while the absolute magnitudes of these WN3/O3s are around *MV* ∼ −2.5. Thus, they could not be in systems with even brighter O3Vs. For this, and other reasons detailed in [93], these stars are single in nature. A further description of their physical parameters and hypothesized place in massive star evolution is discussed in Section 6.

In Figure 10, we now show the effect that the recent work of ourselves and others have made in our knowledge of the WC/WN ratio as a function of metallicity. Clearly, the biggest improvements have come about for M31 and IC10. However, even for IC 10, the results are still very uncertain, with [68] finding many additional candidates that have not ye<sup>t</sup> been certified by spectroscopy, and [69] finding a small number that have also not ye<sup>t</sup> been observed. For the Milky Way (MW), we took the current 661 in Paul Crowther's online catalog, and selected only those with Gaia distances <3 kpc using the (model-dependent) catalog of Bailer-Jones et al. [94]. This found 99 WRs. Despite the vast improvement in the distances available since the estimate of the MW's WC/WN by Massey and Johnson [73], the value for the WC/WN ratio is essentially unchanged. Still, as emphasized earlier, construction of a volume-limited sample for the MW is fraught with difficulties.

**Figure 9.** Spectrum of LMC170-2, one of our newly discovered WN3/O3 type stars. The WN3 classification comes from the star's N V emission (*λλ* 4603,19 and *λ*4945), but lack of N IV. The O3 classification comes from the strong HeII absorption lines but lack of HeI; this figure is from [93].

**Figure 10.** Updated comparison of WC/WN ratio of observed results vs. Geneva Evolutionary models. Notice the drastic changes between the old and new values for IC10 and M31. However, the lack of agreemen<sup>t</sup> between the models and observations at high metallicities remains.

### **4. Wolf–Rayets beyond the Local Group**
