To answer that we must first see why the others sometimes refused to open. In whatever way an ordinary parachute may be packed it must, when coming into use, assume the state of a shut umbrella with a hole in the top. In this condition it is assumed that as it falls the air will find a way in through the lower end and will blow the parachute open in precisely the same way that a strong wind will sometimes blow out the folds of an umbrella. But, as a matter of fact, the loose folds of a parachute, when the edge of the gussets is gathered in, are sure to overlap and enfold each other more or less. Thus, when in the shut-umbrella state, it sometimes happens that air which is inside can escape upwards through the hole more easily than fresh air can get in from below. The parachute, in such a state, is, let us imagine, falling rapidly through the air. The result is just the same as if it were still and the air were rushing upwards past it. And the upward rush past the top hole tends to _suck air out_ through the hole faster than fresh air can find a way in at the bottom. This is the principle of the ejector, which engineers have put to many uses. For example, the vacuum brakes employed on many large railways owe all their power to stop a train to a vacuum caused by an ejector. There is a short tube or nozzle, placed in the centre of another tube through which steam blows. The action of the steam in the outer tube as it rushes past the end of the inner tube drags after it the air which is in the inner tube so effectively as to produce quite a good vacuum. And in precisely the same way, the upward rush of air past the parachute, or what is just the same, the falling of the parachute through stationary air, can suck the air from inside the latter and create a vacuum in it if the gussets gathered together at the mouth unfortunately overlap one another and are thus locked together by the pressure of the air striving to get in. Thus, instead of the downward fall causing the ordinary parachute to open, as in most cases it will do quite well, the fall under these particular conditions actually binds its folds together and prevents it from opening. It is true this does not often happen, but the risk is _always_ present at every drop, and this unreliability has cost the lives of brave men and women, and the knowledge of this constant risk has led others to write down the parachute a failure, by reason of its known unreliability to open instantly. Even when it does open the depth it falls before it opens is so variable, by reason of the fight between vacuum and pressure, that it may be one hundred feet one time and one thousand feet next time with the same parachute.