Stephenson followed up these experiments by others of a similar kind, with the view of ascertaining whether ordinary flame would pass through tubes of a small diameter, and with this object he filed off the barrels of several small keys. Placing these together, he held them perpendicularly over a strong flame, and ascertained that it did not pass upward. This was a farther proof to him of the soundness of the principle on which he had been proceeding. In order to correct the defect of his first lamp, he accordingly proceeded to alter it so as to admit the air to the flame by several tubes of reduced diameter instead of by a single tube. He inferred that a sufficient quantity of air would thus be introduced into the lamp for the purposes of combustion, while the smallness of the apertures would still prevent the explosion passing downward, at the same time that the "burnt air" (the cause, in his opinion, of the lamp going out) would be more effectually dislodged. The requisite alterations were made in the lamp by Mr. Matthews, a tinman in Newcastle, and it was so altered that the air was admitted by three small tubes inserted in the bottom, the openings of which were placed on the outside of the burner, instead of having (as in the original lamp) the one tube opening directly under the flame. This second or altered lamp was tried in the Killingworth Pit on the 4th of November, and was found to burn better than the first lamp, and to be perfectly safe. But, as it did not yet come up entirely to the inventor's expectations, he proceeded to contrive a third lamp, in which he proposed to surround the oil vessel with a number of capillary tubes. Then it struck him that if he cut off the middle of the tubes, or made holes in metal plates, placed at a distance from each other equal to the length of the tubes, the air would get in better, and the effect in preventing the communication of explosion would be the same.