278 Mr. Dugald Clerk [Jan. 2'J, 



when the pressure rises 70 lbs. per square inch above atmosphere, 

 and the maximum pressure is not attained until 82 lbs. per square 

 inch is reached. Combustion is thus proceeding for about O'OS 

 second after the whole vessel is filled with the flame. This experi- 

 ment of Prof. Hopkinson thus proves in a conclusive manner the 

 correctness of the earlier deduction that part of the pressure rise in a 

 closed vessel was accomplished after the flame had spread completely 

 through the vessel. The same experiment proves that in this mixture 

 the velocity of the spread of the flame between the points B and D is 

 5 feet per second. 



In these experiments it was invariably found that the platinum- 

 wire thermometer B at the centre of the vessel always melted before 

 maximum pressure was attained. As the melting point of platinum 

 is 1750° C, it follows that in the centre the temperature always rises 

 above that point. The thermometers C and D were never fused with 

 this particular mixture, so that the temperatures there were necessarily 

 lower. 



In considering a gaseous explosion it has been long recognized 

 that the core of gas would be hotter than the outer layers which are 

 in actual contact with cold metal walls, but Hopkinson's experiments 

 showed a discrepancy in temperature which could not be accounted 

 for in this way. As the results of his measurements he considers the 

 distribution of temperature in his vessel at the moment of maximum 

 pressure to be roughly as follows : — 



Degrees C. 

 Mean temperature (inferred from pressure) . . . 1600 

 Temperature at centre of vessel, thermometer B . . 1900 

 Temperature 10 cm, (4 in.) within wall, thermometer C 1700 

 Temperature 1 cm. (0-4 in.) from wall at end, thermo- 

 meter D 1100 to 1300 



Temperature 1 cm. (0-4 in.) within wall at side . . 850 



The mean temperature throughout the whole vessel inferred from 

 the change of pressure is thus 1600° C, and the estimated temperature 

 at the centre of the vessel, thermometer B, 1900° C, while the tem- 

 perature -4 inches within the wall at thermometer C is 1700° C, and 

 the temperature * 4 inches from the end of the wall thermometer D 

 from 1100 to 1300°, while another thermometer, not shown in the 

 diagram, • 4 inches from the wall at the side showed a temperature 

 of 850° C. This experiment thus shows very great variations of 

 temperature in the flame filling the cylindrical vessel. 



With regard to thermometers B, C, and D, it is obvious that tem- 

 perature difference is not due to difference in heat loss. But little 

 heat, for example, can have been lost at the central thermometer B, as 

 no part of the flame there generated can have touched the walls much 

 before maximum pressure. Although the flame has lasted longest at 

 B, yet that thermometer being right in the centre of the ignited 

 gases is protected against loss. The thermometer C also is 4 inches 



