Nowadays it is difficult to realise that such a system was seriously proposed for railway work and actually adopted by an engineer of such eminence as Brunel. But in point of fact it was recommended by two Board of Trade experts in 1842 and by a Select Committee appointed in 1845 to consider several Bills for atmospheric railways. It was tried at Dalkey and Croydon, and it was installed under Brunel's supervision on a six-mile line in Devon. The carrier in the tube was connected to the train through a longitudinal slit at the top of the tube. The slit was closed by a leather flap, except when momentarily lifted by the passage of the train. A great deal of ingenuity was exhausted in attempting to make this 'longitudinal valve' efficient, but it was found that heat, moisture, and frost made the leather deteriorate so rapidly as to render it hopelessly ineffective in a short time. After a series of misfortunes the atmospheric railway became a mere curiosity in the history of invention. Stephenson was right in regarding the atmospheric railway as 'only the fixed engine and ropes over again, in another form.' He was also right in his belief that the steam locomotive was more economical than either of its rivals. But the stationary engine idea had the germ of an even sounder principle than that of the locomotive. Both in electric tramways and electric railways the power is obtained from stationary engines. The main difference between the electric system and the old rope and atmospheric systems lies in the superior economy with which the power is conveyed electrically to the trains. There are other important differences; but the essential point is that both rope traction and pneumatic propulsion wasted so much power between the engine and the train that their other advantages were annulled, and it was found cheaper to put the engine on wheels and make it drag itself as well as the train. Brunel's reasons for his faith in the atmospheric railway are well worth quoting for the light they throw indirectly upon the advantages of electric traction. He argued that stationary power, if freed from incumbrances such as the friction and dead weight of a rope, was superior to locomotive power, on the following grounds: (_a_) A given amount of power may be supplied by a stationary engine at a less cost than if supplied by a locomotive. (_b_) The dead weight of a locomotive forms a large proportion of the whole travelling load, and thus inherently involves a proportionate waste of power--a waste which is enhanced by the steepness of the gradients and the speed of the trains.