296 Measurement of Cyclically Varying Temperature. 



engine. In every other case the wires had been exposed to 

 the action of the cylinder gases for about half an hour before 

 the determination of the fixed points. In this case the wire 

 had only been exposed to the flame of the blowpipe used in the 

 soldering. The fixed points were 110=0*7596 ; 1^ = 1*0536; 



5? = 1-32 ; R 1 -E =0*294 ; 3 = 2*93. 





 This value of S seemed so high that I repeated the whole 



of the points without detecting any error in the previous 

 work. The low value of -p- 1 together with the high 8 seems 



SXq 



to indicate that a gold-platinum alloy formed at the junction 

 of the main lead to the fine wire and on the surface of the 

 latter. The thermometer was then placed in the engine on 

 March 16th and gave the following results : — 



Percentage of Observed Eesist- . Temperature on 



stroke. ance, ohms. " ' the air-scale. 



10 3*527 948 1200 



After running in the engine some time the fixed points were 

 re-determined with the results given below : — 



R = 0*786. Ba=l-075. I^-R^O'289. 



^ = 1*37. 3=1-95. 

 ±t 



From this it will be seen that the action of the oases on 

 the cylinder had been such as to bring the wire into a normal 

 condition, most probably by sweeping off any alloy which 

 had been formed. 



I have plotted the results of the experiments, and the curves 

 are given in Plate II. figs. 5, 6, & 7, the corresponding pressure- 

 volume curves being given in figs. 8, 9, & 10. In figs. 5, 6, & 7 

 the horizontal ordinates show the percentage of the stroke, and 

 the vertical ordinates the temperature in degrees C. on the 

 air-scale, the observed points being marked by small circles. 

 I have divided the experiments up into three groups — the first 

 being from Jan. 1 to Jan. 4, the second from Jan. 31 to 

 March 7, and the last from March 16 to March 20. It will 

 be noticed that those of Jan. 1-4 (fig. 5) show a distinct 

 break in the regularity of the curve, the line falling but little 

 between the 3rd and 4th points on the 0*0025 wire and between 

 the 4th and 5th points on the 0*003 wire, instead of falling 

 regularly throughout as would have been expected. The 

 curves of the second group of experiments (fig. 6) show no 

 trace of irregularity ; while those of the third group (fig. 7) 



