516 ■ Mr. Dugald aerh [Feb. 22, 



at the end of the first compression. The engine was running while 

 this diagram was taken at the rate of 180 revolutions per minute — 

 that is, each single stroke was performed in one-sixth of a second. 

 From this it is evident that the oscillation in this particular engine 

 lasted during a third of a second. The ordinary indicator very rarely 

 shows oscillations of this kind. The mechanism is too rough, and is 

 too much damped by the friction of the pencil to follow such rapid 

 changes. I have tested the period of the indicator, and find that it 

 is about 200 to the second, so far as ordinary piston displacement is 

 concerned. From this it follows that considerable pressure disturb- 

 ances within the cylinder must have occasioned the oscillation. In 

 this particular engine, the explosion was always accompanied by a 

 peculiar whistling sound. This is very common in many types of 

 engines. This whistling sound seems to start just about the time the 

 diagrams show the beginning of the oscillations— that is, imme- 

 diately after ignition. It is somewhat difficult to account for this 

 peculiar action, but it appears to have some connection with the dis- 

 continuous nature of combustion of a mixture of inflammable gas, or 

 vapour with air. An experiment in this open glass tube will show 

 clearly what I mean. Here I have a mixture of gas and air whicli is 

 slowly being filled into the tube by means of a Bunsen burner at its 

 bottom end. A flame is placed at the upper end. After a time you 

 will notice that the mixture ignites at the open end ; the flame 

 travels back to the glass tube, accompanied at first by a low roaring 

 sound, which increases in intensity as the end of the tube is reached, 

 terminating in a loud snap, sometimes a sort of whistle. When this 

 occurs, the flame flashes l)ack along the tul)e again, and there is 

 obvious oscillation of some kind proceeding. It is not known Avhy the 

 mixture burns in this way, but this particular roaring or whistling 

 seems to occur only when combustion is going on. It is noticed in 

 all pressure flames in the open air. Such flames, for instance, as gas 

 and air mixture flames, make a loud roaring sound, and sometimes a 

 loud wbistUng sound. In experimenting with flames for the igniter 

 I described some time ago, and for a gas furnace using such mixture, 

 I found loud whistling noises very often amounting to a shriek. It 

 appears highly probable then, that wherever tliis oscillation goes on, 

 as shown by the diagram, combustion is still proceeding. The com- 

 l)ustion may be small in amount, but the oscillation appears to be an 

 indication that combustion is continuing. 



After taking a diagram with my optical indicator, I mentioned 

 the matter to Professor Hopkinson, and he stated that he had the 

 same experience in a large Otto engine. He was good enough to 

 send me a diagram, taken with his optical indicator, on this large 

 engine. Here you see the diagram, and in it the whole expansion 

 line is obviously occupied by rapid oscillations of the indicator. 

 Many diagrams have been taken, one on the top of the other, so 



