The experimental design that clinched the argument was the use of the swan-neck flask. In this experiment, fermentable juice was placed in a flask and after sterilization the neck was heated and drawn out as a thin tube taking a gentle downward then upward arc -- resembling the neck of a swan (Fig 4). The end of the neck was then sealed. As long as it was sealed, the contents remained unchanged. If the flask was opened by nipping off the end of the neck, air entered but dust was trapped on the walls of the neck. Under this condition, the fluid would remain forever sterile, showing that air alone could not trigger growth of microorganisms. If, however, the flask was tipped to allow the sterile liquid to touch the contaminated walls and this liquid was then returned to the broth, growth of microorganisms immediately began.
The long "swan-like" neck is open to air, but dust and air-borne microbes cannot reach the liquid. Some of Pasteur's preparations are at the Pasteur Institute, Paris where they continue to remain sterile for more than 100 years.