Notes. Because macroturbulent velocities cannot be determined reliably for fast rotators (Sect. 4.2), all values in Col. 3 are upper limits. Column 4 provides the multiplicity status (see Sect. 4.1 for the classification criterion and Appendix C for the RV studies of each individual object). The runaway status is based on literature studies (references are given on a star-to-star basis in Appendix C). Columns 9–11 list the assumed wind parameters. For stars with the lowest temperatures (typically B0.5 stars), the carbon abundance cannot be firmly determined due to the weakness of the C iii lines at these temperatures. Besides, N ii lines may also be very weak for the hottest stars studied with DETAIL/SURFACE. In these cases, we provide upper limits for both carbon and nitrogen abundances. They correspond to predicted lines becoming detectable, i.e. having a depth significantly exceeding the local noise. Similarly, CMFGEN fits may converge towards very high or very low CNO abundances. In both cases, the upper or lower limits were determined from the χ² curves and correspond to the limit of their flat minimum. Since the CNO abundances are measured relative to the hydrogen content (assumed to be constant in our study), a correction should in principle be applied to the CNO abundances of stars that exhibit a very high helium abundance (and therefore have a reduced hydrogen abundance). However, we found this correction to be negligible (≲0.1 dex) even for the most He-rich stars. The last row of each table section gives the typical 1σ errors on the parameters. Stars studied with both DETAIL/SURFACE and CMFGEN (see Sect. 6.3) are flagged with †.