**6. Conclusions**

GR 290 is is located in the outer spiral arm of the M 33 galaxy at a projected distance of about 4 kpc from the centre. Its spatial location, the proximity to the OB 88 and OB 89 associations, and the similarity of their ages (about 4–5 Myr) as well as a basic concept that a large fraction of all stars, including massive stars, forms in clusters sugges<sup>t</sup> the common origin of GR 290 and OB 89. It is tempting to sugges<sup>t</sup> that GR 290 may have escaped from the association.

The evolution of LBVs during the S Dor cycles seems to occur in most cases roughly at constant bolometric luminosity (see, e.g., [3,57]). However, a decrease of bolometric luminosity from minimum towards the light maximum of the S Dor cycle were observed for several LBVs (e.g., S Dor [58] and AG Car [40]). Lamers [58] interpreted it in terms of the radiative power being partially transformed into mechanical power in order to expand the outer layers of the star from minimum to maximum. In contrast, spectral monitoring of Romano's star during its recent peaks of activity, and the numerical simulation of its stellar atmosphere based on acquired spectra, demonstrated that its bolometric luminosity varies in correlation with its visual brightness, i.e., *L*bol increases during its visual luminosity maxima [10]. Guzik and Lovekin [59] discussed several mechanisms that could trigger the large outburst activity and variations in bolometric magnitude as observed in GR 290. An interesting possibility is that the interplay between pulsations and rotational mixing lead to an unstable transport of H-rich material to the nuclear burning core. In this context, GR 290 may be the ideal object for testing such theories.

The star is hotter than most other LBVs (Table 6), and lays outside of the LBV instability strip in the H-R diagram. On the other hand, the hydrogen abundance of the envelope appears higher than in late type WN stars, and therefore, from the evolutionary and structural point of view, GR 290 is less evolved than WN8h stars [10]. This suggests that Romano's star may be a post-LBV object, the transition phase between LBVs and Wolf-Rayet stars.

The century long light curve of Romano's star shows that until the 1960s the object was in a long lasting quasi-stationary state, a state to which it has returned in 2013, and since then displaying a WN8h spectrum. While the spectral type during the early "low" state (pre-1960) is unknown, from the observed correlation between the visual magnitude and spectral type, we may sugges<sup>t</sup> that it also was WN8h. The Galactic WN8 stars are known to be significantly more variable than the WRs with hotter spectral types [60]. Thus, it is tempting to speculate on the possibility that, in analogy with GR 290, other WN8s may have just recently passed through the LBV phase. Hence, a systematic

investigation of archival data and constructing century long light curves for WN8-WN9 stars using archival photographic plates will probably be able to uncover more objects similar to Romano's star.

**Table 6.** Comparison of Romano's star with other LBVs and LBV candidates which show WR like spectra.


**Author Contributions:** R.F.V., spectral analysis; M.C., C.R. and R.G., photometric monitoring and reduction of photometry data; and O.M., numerical modeling of stellar atmosphere and reduced the spectroscopic material and manuscript preparation. G.K. was PI of the 2016 and 2018 GranTeCan observations and performed spectral analysis. All authors discussed the results and commented on the manuscript.

**Funding:** This research was funded by CONACYT gran<sup>t</sup> 252499, UNAM/PAPIIT gran<sup>t</sup> IN103619, Russian Foundation for Basic Research gran<sup>t</sup> 19-02-00779 and Czech Science Foundation gran<sup>t</sup> GA18-05665S. This project received funding from the European Union's Framework Programme for Research and Innovation Horizon 2020 (2014-2020) under the Marie Skłodowska-Curie Grant Agreement No. 823734.

**Acknowledgments:** We express our enormous gratitude to V.F. Polcaro, who recently passed away, for having stimulated our interest and studies of this unique object. We thank Roman Zhuchkov, Oleg Egorov and Olga Vozyakova for obtaining the photometric observations on 1.5 m Russian–Turkish telescope and 2.5 m telescope of the Caucasian Mountain Observatory. We thank Thomas Szeifert and Philip Massey for the spectra obtained with Calar Alto/TWIN spectrograph in 1992 and with WIYN 3.5 m telescope in September 2006. We thank the GTC observatory staff for obtaining the spectra and Antonio Cabrera-Lavers for guidance in processing the observations. We thank Guest Editor Prof. Roberta M. Humphreys and our anonymous referees for providing helpful comments and suggestions. In this paper, we use data taken from the public archive of the SAO RAS. The work is partially based on the observation at 2.5-m CMO telescope that is supported by M.V. Lomonosov Moscow State University Program of Development.

**Conflicts of Interest:** The authors declare no conflict of interest.
