_Answer._ The most important point to be observed is, whether the pressure at the end of the stroke is reduced as low as possible, as at high speeds it is usually much more difficult to exhaust the steam from than to admit it into the cylinder. As already stated, the blast in the chimney makes it almost impossible to exhaust the steam to atmospheric pressure when the locomotive is running fast. If the steam is released too late in the stroke, as already explained, there will not be time enough nor sufficient opening of the port to allow the confined steam to escape from the cylinder before the end of the stroke, and this will be indicated on the diagram by the space between the line of back pressure and the atmospheric line during the commencement of the return stroke, as shown in fig. 138. QUESTION 222. _What should be observed regarding the line of back pressure?_ _Answer._ The most important point is, that it should approximate as closely as possible to the atmospheric line, as all the back pressure not only diminishes the efficiency of the engine, but is a total loss of energy. Too much inside lap will increase the amount of back pressure, but generally it is more influenced by the area of the blast orifices than by any other cause. Every effort should be made, therefore, to have them as large as possible and yet have the boiler make as much steam as is needed. When only one blast orifice is used for both cylinders, it often happens that when the steam is exhausted from the one cylinder it “blows” over into the other, and thus produces an additional amount of back pressure. This is shown by a rise or “hump” in the line of back pressure, as indicated in fig. 138. QUESTION 223. _Can the amount of compression which is needed be determined by calculation?_ _Answer._ Yes; but it involves more abstruse principles of mathematics than it is thought best to introduce here. Some of the reasons can, however, be given, which will make the subject clearer, and enable the reader, if he has sufficient knowledge of mathematics, to investigate the subject still further. [Illustration: _Fig. 140._]