of Water at High Temperatures. 1 21 



fig. 7, and an oil-tube were placed. The heating took place by- 

 means of two Bunsen-burners placed below, as shown in fig. 1. 

 At the top of each Bun sen- tube there was fixed a spiral wire 

 (fig. 5), which effectually prevented the flame from jumping 

 downwards when at its lowest. With these and a lever with 

 graduated arc applied to the gas-cock s, fig. 1, the temperature 

 could be gradually raised and held in command if the day was 

 calm. In repeating such experiments it would be well to have 

 a gas-cock made specially, so as to be able to regulate the supply- 

 by uniform motion. 



In the lard-bath the convection currents were very rapid, and 

 the temperature throughout nearly the same. The standard 

 thermometer was kept moving up and down for some time 

 before making the comparison (which requires a perfectly steady- 

 temperature to be maintained). 



The value of the scale- readings of the oil-tubes were thus 

 determined, and their curves of expansion drawn to a large scale 

 up to about 280°. Beyond this the curve was carried on the 

 hypothesis that it was an arc of a hyperbola, which from pre- 

 vious experience I had found to answer well as an empirical 

 curve in liquid expansion. 



§ 9. Another method was employed to measure the higher 

 temperatures of the last water-tube. A tube thermometer con- 

 taining a 5^-inch column of sulphuric acid was prepared. The 

 value of its scale-readings above 210° depended on a very good 

 observation in acid-bath at 203° (the temperature having re- 

 mained steady for half an hour). To this and the reading at 

 atmospheric temperature the formula derived from the law of 

 expansion was applied, and the value of the higher scale-read- 

 ings computed. It was then lashed to the oil thermometer, and 

 simultaneous readings taken on both up to 350°. On comparing 

 results, it was found that 330° by the acid-tube was 343° by the 

 oil. This corresponds to the highest observation of the last water- 

 tube. A comparison was finally made with acid-tube and standard 

 in acid-bath at 270°, but this did not help to a nearer accordance. 



§ 10. Granting that the higher temperatures are uncertain 

 and doubtful in their exact position on the scale of temperature, 

 the method of curves generally employed will cause the diver- 

 gence to be gradual, so that the law of expansion, if it existed, 

 would not by this cause be masked. 



§ 11. Sulphuric acid has little more than half the expansibi- 

 lity of oil, but has the advantage of being colourless and easier 

 managed. Glycerine was not employed, although it may be the 

 best liquid upon the whole, being only one-fifth less expansible 

 than oil. In using oil-tube thermometers it is not proper to 

 regard observations under 100°. 



