In the "why didn't I think of that?" department: Menzies et al performed pleural fluid sampling on 62 patients with clinical pleural infection, and sent samples to the lab in both standard tubes and in blood culture bottles inoculated at the bedside. 20 standard samples were culture-positive; adding the positive inoculated samples identified an additional 11 infections (an absolute increase in yield of 21%). They suggest we all do this routinely. I'm going to. Why not?