March 27, 1890] 



NA TURK 



489 



The results obtained are given in the following table : — 

 Temperature 280°. 



Temperature 356° 



The last result at 356° is a little uncertain, owing to a breakage 

 of the apparatus. 



I may also mention that M. Guillaume has informed me that 

 M. Tonnelot has heated several thermometers to 450°, and that, 

 notwithstanding a considerable internal pressure, a rise of the 

 zero-point was observed in every case. 



All these results seem to lead unmistakably to the conclusion 

 that pressure has little or no effect on the rise of the zero-point. 



Three questions remain to be discussed — 



(i) Would the total rise of the zero-point be different if two 

 similar thermometers were subjected to sufficiently prolonged 

 heating at different temperatures ? At first sight, it would cer- 

 j tainly appear that at 356" the total rise with my thermometers 

 I must be greater than at 280°, but I do not feel satisfied that the 

 proof is sufficient. If we map the observations of zero-point 

 against the time of heating, curves are obtained which appear 

 as if they might become horizontal after a few weeks or, pos- 

 sibly, months ; but if, instead of the actual times, we take their 

 logarithms — as in the diagram — as abscissae, there is no appear- 

 ance of an approach to the final state at either temperature. 

 But while at 356° the curve has become almost a straight line, 

 at 280° there appears to be an increasing tendency towards the 

 vertical direction. I do not for a moment argue that the curves 

 indicate that the maximum rise would be the same at both tem- 

 peratures if the experiments were carried on for a sufficiently 

 long time ; but, at the same time, I do not think that they 

 afford any convincing proof that the total rise would be different. 

 The re.sults merely tend to increase my scepticism as to the 

 value of the determination of the maximum rise at 0° obtained 

 by extrapolation of the curve constructed from observations at 

 that temperature. It does not appear to me that it would be 

 justifiable to extrapolate these curves at all, and I am afraid that 

 they do not throw much light on the total rise of zero-point at 

 either temperature. Very much more prolonged heating would 

 be necessary before arriving at a definite conclusion. 



(2) With regard to the causes of the contraction of the bulb, 

 I have no hesitation in admitting that — as shown by M. Guil- 

 laume — the removal of the condition of strain caused by the 



■■■■■■■■■■■■■■ 



■■■■■■■■■■■■■■■■■■■■■■■■■■■■I 



!■■■■■■■■■■■■■■ 



■■■■ mmwam ■■■■■ 



■■■■BBBHI 



LOGARITHM OF TIME (IN HOURS) 



more rapid cooling of the outer parts of the glass, is insufficient 

 to account for the results. No doubt we must also take into 

 account the too rapid cooling of the glass as a whole, which 

 prevents the molecules from assuming the position of greatest 

 stability, perhaps In the same sort of way that the assumption by 

 sulphur of the monoclinic or the more stable rhombic form de- 

 pends on the rate at which solidification takes place. That 

 there are other causes besides these two does not at present 

 appear to me to be Jproved. 



(3) Lastly, there is the question raised by Mr. Tomlinson, as 

 to whether repeated heating and cooling between wide limits of 

 temperature is more effective in raising the zero-point than pro- 

 longed heating at the higher temperature. The points repre- 

 senting the individual observations fall very fairly on the curves 

 constructed from them, and do not seem to indicate any notice- 

 able difference in the effect of long or short heating. The results 

 can hardly, however, be regarded as decisive. 



University College, Bristol, March I. Sydney Young. 



