When working on tobacco Rubisco in Los Angeles, we got the best crystals of the activated form of the protein after small earthquakes (4-5 on the richter scale). However, as people in the lab. noticed, this earthquake parameter is somewhat difficult to control. Handling the linbro plates sometimes caused massive nucleation: I would look at a plate with only clear drops and by the time I would put the plate back in the incubator there was so much nucleation that the plate was of no use any more. When the nucleation of your protein is sensitive to vibrations, I would worry more about handling than about the fire alarm.