Tests to obtain the resistance of trains are usually made by placing a dynamometer car between the locomotive and the body of the train.

The coupler between the car and the locomotive includes a dynamometer attachment which automatically records at any instant the actual pull on the draw-bar. Apparently thisought to solve the problem very easily and accurately, but in practice it is found that the interpretation of the dynamometer records is not easy and is liable to misconstruction, unless care is taken to make several allowances. One of the practical difficulties in interpreting the results of dynamometer experiments is to determine the actual velocity, especially when the velocity is not regular. Speed-recorders are supposed to indicate the velocity at any instant, but they are not very accurate even when the velocity is uniform, and they are especially inaccurate when the velocity is fluctuating. When the velocity of the train is decreasing, the kinetic energy of the train is being turned into work, and a force transmitted through the dynamometer is less than the amount of the resistance which is actually being overcome. Therefore, unless the indication of the dynamometer is carefully corrected by adding to it the force calculated according to formula 4, which equals the force which is really assisting the train when its velocity is reduced from V2 to V1 in a distance s, the indication of the dynamometer will not represent the force required to overcome the resistances actually encountered by the train. On the other hand, when the velocity is increasing, the dynamometer indicates a larger force than is required to overcome the resistances, but the excess force is being stored up in the train as kinetic energy. In such a case, the force P1, calculated from formula 4 on the basis of the differences of velocities in any assumed distance s, must be subtracted from the dynamometer record in order to obtain the force necessary to overcome the train resistances. Grade has a similar effect, and the force indicated by the dynamometer may be greater or less than that required at the given velocity on a level by the force which is derived from, or is turned into, potential energy. Since grade, either ascending or descending, is usually found in the track, the actual grade of the road-bed must be known and allowed for at all points. Curvature must likewise be allowed for, as it has a constant retarding force. Usually the allowance per ton is 0.5 pound.