From Derek Lowe, regarding the four DuPont workers who died from a methyl mercaptan leak two weeks ago (I would add, also, the many others who’ve died in laboratory accidents):
So in memory of these four, here’s something that all of us who work in the lab can do today. Take a look around you. Remind yourself of where the fire extinguishers are (and there should be more than one kind). Think of how you’d get to the safety shower if you had to use it. And pick the door you’ll use if a situation get beyond that. It’s far easier to go over such details when things are quiet, and if you do that every so often you’ll have a much better chance of remembering where to go when you really need to.
And whenever you’re setting up an experiment that involve any noticeable hazard (pyrophoric reagent, toxic liquid or gas, potential exotherm), think for a moment about what might be most likely to go wrong, and also what the worst thing that could happen might be, and what you’d do about them. Is it dropping that bottle of phosgene solution on the floor? A fire started by your hydrogenation catalyst or your sodium hydride? An exotherm that sends your reaction pouring out over the hot plate or heating mantle? Picturing these things beforehand is never wasted time, because (as everyone with experience in the lab knows) such things do happen, and not on anyone’s schedule. Those four DuPont workers were getting ready to go home for the day when suddenly everything went wrong: in their memory, keep an eye out for what might go wrong in your own fume hood.