The length of time occupied by the anticipatory pumping was often increased by the difficulty of arranging proper telegraphic communication on the South Devon Railway, and by the absence of it on the Bristol and Exeter Railway. The Electric Telegraph was in its infancy, and though Mr. Brunel had been the first to apply it in connection with railways, namely, between London and Slough on the Great Western Railway in the year 1839, it was not brought into perfect working order on the South Devon Railway till the Atmospheric System was on the point of being abandoned. The result of the defects in the telegraph was that, when a train was late, warning was not received at the several engine-houses;[71] and thus, when this was the case, the pumping-engines which had been started at the right interval of time before the train was due, had to be kept at work for a needlessly long period pumping out the air, which was all the while leaking in through the deteriorated valve. This inefficiency of the telegraphic apparatus would have been of trifling importance but for the defects of the valve. Had the valve been as perfect as it was expected to be, the vacuum, after it had been formed, could have been maintained by an expenditure of power very moderate in comparison with that which was actually required. As regards the relative cost of the power consumed, it appears that, owing to imperfections in the engines, their expenditure of fuel per indicated horse-power was more than double that of the best of the Cornish pumping-engines, to which they were analogous; while the cost of working was more than three times as great. The defects in the engines were for the most part such as might have been remedied; and this would have been done, had not the excessive duty imposed on them by the leakage of the longitudinal valve prevented their being stopped for repairs and alterations. In this way the defects of the different parts of the apparatus mutually aggravated each other. It appears then that, chiefly owing to the defective longitudinal valve, the engines were expending nearly three times the power which they should have done for a given tractive duty, according to previous experience, and the results obtained on the Dalkey line, and that they cost per horse-power at least three times as much as was expected. The cost of traction, nearly nine times as much as had been calculated on, was between two and a half and three times what it would have been with locomotive power; and this was on a level part of the line, where the comparative advantages of the Atmospheric System were not exhibited as they would have been in the part which had steep gradients.