Our prediction of levofloxacin resistance was based partially on the behavior of ciprofloxacin with Staphylococcus aureus. When this organism was treated with ciprofloxacin, resistance developed very rapidly as resistant strains spread through hospitals (1). Although the selection window hypothesis does not bear on dissemination of resistant strains, those strains had to come from somewhere. The window hypothesis predicts that ciprofloxacin-resistant S. aureus would arise readily because of the location of the pharmacokinetic curve relative to the window (filled circles). Notice that data for levofloxacin with S. pneumoniae are similar to data for ciprofloxacin and S. aureus. This correlation does not mean that levofloxacin resistance in S. pneumoniae will develop as rapidly as ciprofloxacin resistant S. aureus since dissemination factors are probably different. However, this correspondence is very unsettling.