The air valve AV was an exhaust valve similar to the two gas valves VV in the great balloon, with the one exception that it was weaker. In this way, when there happened to be too much fluid (i.e. gas or air, or both) distending the great balloon, all the air would leave the interior balloon before any of the gas would leave the great balloon.
The first trial of my "No. 2" was set for 11th May 1899. Unfortunately, the weather, which had been fine in the morning, grew steadily rainy in the afternoon. In those days I had no balloon house of my own. All the morning the balloon had been slowly filling with hydrogen gas at the captive balloon station of the Jardin d'Acclimatation. As there was no shed there for me the work had to be done in the open, and it was done vexatiously, with a hundred delays, surprises, and excuses.
When the rain came on, it wetted the balloon. What was to be done? I must either empty it and lose the hydrogen and all my time and trouble, or go on under the disadvantage of a rain-soaked balloon envelope, heavier than it ought to be.
I chose to go up in the rain. No sooner had I risen than the weather caused a great contraction of the hydrogen, so that the long cylindrical balloon shrunk visibly. Then before the air pump could remedy the fault, a strong wind gust of the rainstorm doubled it up worse than the "No. 1," and tossed it into the neighbouring trees.
My friends began at me again, saying:
"This time you have learned your lesson. You must understand that it is impossible to keep the shape of your cylindrical balloon rigid. You must not again risk your life by taking a petroleum motor up beneath it."
I said to myself:
"What has the rigidity of the balloon's form to do with danger from a petroleum motor? Errors do not count. I have learned my lesson, but it is not that lesson."
ACCIDENT TO "No. 2," MAY 11, 1899
(SECOND PHASE)