Page:My Airships.djvu/123

 tube to furnish a jet of the room's atmosphere from an air pump and light it with the electric spark. Similarly, should a pin-prick have made ever so slight a vent in my air-ship balloon, the interior pressure would have sent out into the atmosphere a long thin stream of hydrogen that might have ignited had there been any flame near enough to do it. But there was none. This was the problem. My motor did undoubtedly send out flames for, say, half-a-yard round about it. They were, however, mere flames; not still-burning products of incomplete combustion like the sparks of a coal - burning steam-engine. This admitted, how was the fact that I had a mass of hydrogen unmixed with air and well secured in a tight envelope so high above the motor to prove dangerous? Turning the matter over and over in my mind I could see but one dangerous possibility from fire. This was the possibility of the petroleum reservoir itself taking fire by a retour de flamme from the motor. During five years, I may here say in passing, I enjoyed complete immunity from the retour de flamme (sucking back of the flame). Then, in the same week in which Mr Vanderbilt burned himself so severely, 6th July 1903, the