In order to make obligate sex, as well as crossing-over in a multichromosomal genome, advantageous, a very strong selection, incurring the genetic load of at least ~50%, must operate more or less permanently (Kondrashov 1984). It is not clear what forces can lead to such selection. The two most realistic candidates are negative selection against deleterious mutations and positive selection for beneficial mutations. Recently, it became obvious that genomic deleterious mutation rates in multicellular eukaryotes are high (~1) but generally not much higher (Haag-Liautard et al. 2007). Such rates are enough to cause a substantial deterministic disadvantage of obligate asexual reproduction, under synergistic epistasis, but not to favor obligate sex. In addition, the available data, although far from being conclusive, do not indicate that synergistic epistasis is common (Kouyos et al. 2006) Also, I failed to reproduce a recent claim by Keightley and Otto (2006) that sex and recombination can be enjoy a short-term advantage under recurrent deleterious mutation even without epistasis, due to stochastic Hill-Robertson interference. As far as beneficial mutations are concerned, sex can obviously be advantageous only if multiple selection-driven allele replacements proceed more or less simultaneously. Despite a lot or recent attention to the subject, it is still unclear what is the typical per generation number of adaptive allele replacements in an evolving lineage. However, it appears that this number may be as low as 1/1000, and that it can hardly be above 1/100 in most cases (Eyre-Walker 2006). Other forces that could make obligate sex advantageous could be imagined, but none of them looks too promising as a general explanation. Thus, the ubiquity of obligate sex and crossing-over remains a mystery.