ondelastic Fluids, and on the Measurement of Tempieralures. 377 



great liquid mass 200 or 300 degrees warmer than the surround- 

 jng air. We attain this rigorously when the temjjeratiire at 

 which we operate is, for example, that of the ebullition of the li- 

 (juid which we employ : then this temperature is necessarily fixed; 

 but in every other case the progress more or less rapid of the 

 lieating or cooling of i!ie different points of the mass, opposes the 

 necessary uniformity from taking place. We are of opinion, how- 

 ever, that the arrangement of our apparatus remedies in a great 

 measure this kind of inconvenience, and this is owing, on the one 

 hand, to the copper tub beiiig sunk in the furnace, and composing : 

 with it a considerable mass, which is cooled slow ly, particularly when 

 it is near its maximum of temperature ; and in the second place, 

 the liquid being continually agitated, the heat ouglit to spread 

 through it more uniforndy. In short, to remove all doubt*, we 

 inserted into this tub two thermometers situated horizontally at 

 thesame height ; we had raised the temperature in the same way 

 as in our ordinary experiments, and we ascertained that by 

 shaking the Hquid the thermometers never differed but by a few- 

 tenths of a degree. 



Besides, even supposlnsr that the different particles of the liquid 

 layer which surro\md the air-tube were not exactly at the same 

 temperature, the error will not be so great as at first believed ; 

 for, in consequence of the arrangement of the apparatus, the bulb 

 of the thermometer nearly answers the middle of the length of 

 the tube, and consequently this instrument ought in all cases to 

 indicate a temperature not far from the mean of the different parts 

 of the tube. We ought to recollect also, from what has been 

 said, the necessity which there is, in order to know tlie true indica- 

 tions of the thermometer, to sink it sufjiciontly in the liquid, that 

 the whole column of mercury should enter it. This precaution, 

 which appears to be unnecessary in low temperatures, ought not 

 to be omitted in high temperatures; for then the column of mer- 

 cury contained in the tube nuiy undergo an increment of very sen- 

 sible length. Thus we have remarked that, at the temperature of 

 300^ for instance, there was frequently more than 15° of differ- 

 ence between the indications of one and the same thermometer, 

 whether the whole of the mercury was in the liquid, or the bulb 

 only entered it. We mii;ht indeed, according to our knowledge 

 of the dilatation of the mercury, estimate the error which we com- 

 mit by plunging only part of the thermometer into it ; but the 

 correction to make in this case, bearing on considera!)!e numbers, 

 will occasion serious errors, l)ccause we never know exactly the 

 mean tenrperature of the mercury contained in the tube. It ap- 

 peared to us to be ])rcfcrable in all cases to place the thermome- 

 ters horizontally. In order to give a more precise idea of the 

 various operations of which each of our experiments is composed, 



Ave 



