The terminal which passed through the insulator was also extended for something like half an inch beyond the porcelain and bent to a proper extent to co-act with a piece of platinum wire inserted in the interior wall of the plug which formed the other terminal. After making this improvement in the plugs practically no difficulty was experienced from short-circuits caused by the soot. In making the tests of this experimental engine it was found practically impossible to absorb the power by a Prony brake in a sufficiently uniform manner on account of the fact that the engine was being run without a fly wheel. The consequent variation in the torque and speed during each revolution caused such great fluctuations in the reading of the scales which measured the pull of the Prony brake that no confidence could be felt in the accuracy of the readings and, therefore, no confidence could be placed in the determinations of the effect which different changes in the engine produced. A water-absorption dynamometer consisting of a number of flat, circular discs fastened to a shaft and rotating between other parallel flat discs arranged in a circular drum which was filled to any desired extent with water was immediately planned, and the construction of two of them was begun so that power could be taken from both ends of the engine shaft, which, on account of its necessary lightness, was apt to be injured by being twisted when all the power was taken from one end of the shaft. In order to continue the tests on the engine while this dynamometer was being made it was decided to employ one of the propellers as a dynamometer. Although no accurate tests had been made to determine just how much power was required to drive these propellers at various speeds, yet the fundamental law was known that under the same conditions the power required to drive any propeller would vary as the cube of the number of revolutions, and since the Prony-brake tests had given an approximation as to the amount of power which the engine developed at certain speeds, the law of the propeller, and extrapolations from the data obtained in the tests of the smaller propellers in 1898, enabled further approximations to be made as to the amount of extra power which the engine developed when certain changes enabled it to drive the propeller at increased speeds. This method had also the great advantage that, since the power required varies as the cube of the number of revolutions, it is practically impossible for the engine to “run away” with the propeller and cause serious damage through the possible excessive strains introduced by high speed. This feature is also possessed by water-absorption dynamometers of the type which were built and used in the later tests.