88 Prof. B. Hopkinson on the 



rapid ingress and egress through the valves. During the 

 compression and expansion strokes, the air is probably almost 

 at rest except in so far as it is moved by the piston. 



The wire formed one arm ot a Wheatstone's bridge, the 

 connexions being as in fig. 1 (Plate VI.). The rotating 

 contact-maker A is a block of hard wood soaked in paraffin 

 and carried on the valve-shaft of the engine. It had a slip 

 of brass J of an inch wide let into its periphery which made 

 contact with the brushes B, and so completed the galvano- 

 meter circuit once in two revolutions of the engine, that is, 

 once in the cycle described above. The moment of contact 

 could be adjusted to any desired point in the cycle by altering 

 the position of the brushes. The current in the wire with this 

 arrangement of resistances was limited to. about T) T n j ampere, 

 and it had no appreciable heating effect upon the wire. 



The resistance of the wire at 0° C. was P2(3 ohms, the 

 temperature coefficient *^0o43, and S (the factor for correcting 

 platinum to centigrade temperatures) was 1*57. The suc- 

 cession of temperatures determined in this way is shown in 

 fig. 2, curve A (Plate VI .). Of the actual observations, those 

 shown thus • were all taken on the same day, and the con- 

 ditions were approximately the same throughout. The tempe- 

 rature of the air in the air-pipe close to the engine was 

 17° C, the temperature in the exhaust-pipe 2G^° C, and the 

 temperature of the jacket-water 46° 0. The mean speed 

 throughout was 180 revs, per minute. The observations 

 shown thus Q were taken on another day when the conditions 

 would not be quite the same. 



The connexions were now altered as shown in fig. 3 (PL VI.) . 

 The conditions were kept exactly the same as before ; in fact, 

 the engine was not stopped. The current now passing in the 

 wire (measured by the ammeter M) is about 0'75 ampere 

 and produces a considerable heating effect. The external 

 resistance of 50 ohms sufficed to keep the current sub- 

 stantially constant in spite of the varying resistance of the 

 wire. The succession of wire temperatures under these 

 conditions is plotted in curve B, fig. 2 (PL VI.). 



From curves A and B the curve of gas temperature can be 

 at once deduced. Let 6 be the gas temperature, and 6" the 

 wire temperature given by curve B. Then if is the current 

 in the wire and li its resistance at the moment, heat is being 

 supplied to the wire electrically at the rate C 2 P, and by 

 conduction from the gas at the rate X(0 — 0"). Neglecting 

 radiation losses, we have therefore 



X(0-0") = &^--C 2 R, (1) 



