WR + O binaries as probes of the first phase of mass transfer

Context. Wolf-Rayet (WR) and O-star binaries can be the progenitors of X-ray binaries and double black hole binaries. Their formation is not yet fully understood, however. For 21 observed WR+O systems, we aim to infer whether the mass transfer started on the main sequence (Case A) or later (Case B). We also calculated (limits on) the mass-transfer efficiency {\beta}, that is, the fraction of transferred mass that is accreted, and the parameter {\gamma}, which denotes the fraction of angular momentum of the binary that is lost per unit mass in units of the average angular momentum of the binary per unit mass. Aims. We inferred the possible values for the initial masses based on the observed WR masses and models for WR from the literature. With these initial primary masses, we created a grid of possible periods and secondary masses for which we determined the values that {\beta} and {\gamma} would have taken for either Case A or Case B mass transfer. Based on this, we also determined the case of mass transfer that is most likely for each system. Methods. Taking into account the progenitor distribution of WR+O binaries, we find that highly non-conservative Case A mass transfer seems to be the most likely scenario for the majority of systems as this can explain 14 out of 21 systems. The angular momentum loss is likely relatively high (typically {\gamma} > 1). Our finding that most systems in our sample experienced Case A mass transfer contradicts the expectation that most massive binaries go through Case B mass transfer. This suggests that post-case-B systems are significantly underrepresented in the observed WR+O binary population, either intrinsically or due to severe selection effects.