126 Mr. J. J. Waterston on the Expansion 



curve of the acid tube thermometer produced. I have drawn 

 similar curves with the temperatures by the curve of the oil tube 

 thermometer produced, and have projected the results on fig. 10, 

 the points being shown by crosses and the straight line B B drawn 

 through them. We have here a similar flat ogee range, and the 



value of — that comes out by a process of computation similar to 



the above is 2*69. The mean is 3*14, which is remarkably close 

 to 3-21. 



§ 19. That the deviation of the acid-tube curve should be so 

 nearly the same as that of the oil-tube and in the opposite 

 direction, is a curious coincidence if the true line really lies, as it 

 appears to do, halfway between A A and B B. While suspecting 

 that there may be a latent cause for this (though each curve was 

 as carefully determined as if the results depended on it alone), it 



cannot of course be said that - has been fully brought out. 



P 

 There may be a deviation in its value corresponding to the 



deviation in expansion that undoubtedly exists from about 185° 

 downwards. 



§ 20. With the view of obtaining further evidence on this 

 point, I have examined by the differential process (detailed in 

 § 27) the results of three series of observations of three of those 

 tubes first experimented upon in which the temperature was 



carried up to from 260° to 280°. Assuming -=3'21, the values 



dv 

 of 7 obtained by computing several of the values of -r- at the higher 



temperatures accord with 411-6, varying from 400° to 430°. 

 The thermometers employed with these water-tubes were linseed- 

 oil tubes, but different from the last (represented in fig. 3), and 

 somewhat longer and wider. 



§ 21. To arrive at the probably true value of the highest 

 temperature observed, we may divide the difference 12°*3 (be- 

 tween the scale of acid- and oil-tubes) in the ratio of the values 



of -. Thus 3-59-269=0-90, and 3'59-3-21=0'38; so 



P 

 •90 : -38 : : 12°'3 : 5°'2, which, added to 320°, gives 325°-2 C.A. 

 as the temperature at which the volume is 15363. We have to 



k 

 draw the curve v 3 ' 21 = through this point, touching or cut- 



ting the lower branch or verging into it. If we arrange the 

 terms so as to make the theoretical curve touch the parabola 

 which answers so exactly between 50° and 150°, we find that 

 contact is impossible. This shows that there is a gradual rise 



