Some studies have claimed a universal stellar upper-mass limit of 150 Msun. A factor that is often overlooked is that there might be a difference between the current and initial masses of the most massive stars, as a result of mass loss. We present Monte Carlo mass-loss predictions for very massive stars in the range 40-300 Msun, with large luminosities and Eddington factors Gamma. Using our new dynamical approach, we find an upturn in the mass-loss vs. Gamma dependence, at the point where the winds become optically thick. This coincides with the location where wind efficiency numbers surpass the single-scattering limit of Eta = 1, reaching values up to Eta = 2.5. Our modelling suggests a transition from common O-type winds to Wolf-Rayet characteristics at the point where the winds become optically thick. This transitional behaviour is also revealed with respect to the wind acceleration parameter beta, which starts at values below 1 for the optically thin O-stars, and naturally reaches values as high as 1.5-2 for the optically thick Wolf-Rayet models. An additional finding concerns the transition in spectral morphology of the Of and WN characteristic He II line at 4686 Angstrom. When we express our mass-loss predictions as a function of the electron scattering Gamma_e (=L/M) only, we obtain a mass-loss Gamma dependence that is consistent with a previously reported power-law Mdot propto Gamma^5 (Vink 2006) that was based on our semi-empirical modelling approach. When we express Mdot in terms of both Gamma and stellar mass, we find Mdot propto M^0.8 Gamma^4.8 for our high Gamma models. Finally, we confirm that the Gamma-effect on the mass-loss predictions is much stronger than that of an increased helium abundance, calling for a fundamental revision in the way mass loss is incorporated in evolutionary models of the most massive stars.

We analyse ~ 360 ks of archival data from the Rossi X-Ray Timing Explorer (RXTE) of the 21 hr orbital period dipping low-mass X-ray binary 4U 1624-49. We find that outside the dips the tracks in the colour-colour and hardness-intensity diagrams (CDs and HIDs) are reminiscent of those of atoll sources in the middle and upper parts of the banana branch. The tracks show secular shifts up to ~ 10%. We study the power spectrum of 4U 1624-49 as a function of the position in the CD. This is the first time power spectra of this source are presented. No quasi-periodic oscillations (QPOs) are found. The power spectra are dominated by very low frequency noise (VLFN), characteristic for atoll sources in the banana state, and band limited noise (BLN) which is not reliably detected but may, uncharacteristically, strengthen and increase in frequency with spectral hardness. The VLFN fits to a power law, which becomes steeper when the source moves to the harder part of the CD. We conclude that 4U 1624-49 is an atoll source which in our observations is in the upper banana branch. Combining this with the high (0.5-0.7 L_Edd) luminosity, the long-term flux stability of the source as seen with the RXTE All-Sky Monitor (ASM), and with the fact that it is an X-ray dip source, we conclude that 4U 1624-49 is most likely a GX atoll source such as GX 3+1 and GX 9+9, but seen edge on.