There was but one thing to do.
Pulling on the manœuvre valve I let out a sufficient quantity of hydrogen and came slowly down to the surface of the water, in which the air-ship sank.
Balloon, keel, and motor were successfully fished up the next day and shipped off to Paris for repairs. Thus abruptly ended my maritime experiments; but thus also I learned that, while a properly inflated balloon, furnished with the proper valves, has nothing to fear from gas displacement, it is best to be on the safe side and guard oneself against the possibility of such displacement, when by some neglect or other the balloon is allowed to go out imperfectly inflated.
For this reason, in all my succeeding air-ships, the balloon is divided into many compartments by vertical silk partitions, not varnished. The partitions remaining unvarnished, the hydrogen gas can slowly pass through their meshes from one compartment to another to ensure an equal pressure throughout. But as they are, nevertheless, partitions, they are always ready to guard against any precipitous rushing of gas toward either extremity of the balloon.
Indeed, the experimenter with dirigible balloons must be continually on his guard against little errors and neglects of his aids. I have four men who have now been with me four years. They are in their way experts, and I have every confidence in them. Yet this thing happened: the air-ship was allowed to leave the aerodrome imperfectly inflated. Imagine, then, what might be the danger of an experimenter with a set of inexperienced subordinates.
In spite of their great simplicity my air-ships require constant surveillance on a few capital heads:
Is the balloon properly filled?
Is there any possibility of a leak?
Is the rigging in condition?
Is the motor in condition?