Recent phylogenetic analyses reveal that many eukaryotic nuclear genes whose prokaryotic ancestry can be pinned down are of bacterial origin. Among them are genes whose products function exclusively in cytosolic metabolism. The results are surprising: we had come to believe that the eukaryotic nuclear genome shares a most recent common ancestor with archaeal genomes, thus most of its gene should be 'archaeal' (loosely speaking). Some genes of bacterial origin were expected as the result of transfer from mitochondria, of course, but these were thought to be relatively few, and limited to producing proteins reimported into mitochondria. Here, I suggest that the presence of many bacterial genes with many kinds of functions should not be a surprise. The operation of a gene transfer ratchet would inevitably result in the replacement of nuclear genes of early eukaryotes by genes from the bacteria taken by them as food.