COMPARISON OF PLATINUM AND GAS THEBMOmTDB, 77 



cub. centim. 



(1) ...... 0-4371 



(2) 0-4503 



(3) 0-4469 



(4) 0-4380 



(5) 0-4385 



(6) 0-4385 



(7) . . . ... 0-4385 



Mean . . . v= 0-4411 



(e) Determination of the "Index Error." 



The readings made on the scale by means of the vernier do not represent exactly 

 the difference of level between the points in the barometer and the closed branch of 

 the manometer ; this is due to two causes. The first is that the point of the 

 barometer is not at the same level as the division, 0, of the scale. All the scale- 

 readings have, therefore, a correction applied which we will call the correction for 

 " index error," which must be determined by special measurements with a good 

 cathetometer. This correction, which would be constant if the plane surface of the 

 girder on which the barometer slides were absolutely true, varies slightly according to 

 the position of the barometer. 



The second cause is that the point in the manometer is not at the same level as the 

 division, 0, of the vernier. This latter correction may, perhaps, be considered con- 

 stant for a given position of the closed branch of the manometer. 



The correction for the " index error " has been determined frequently during the 

 course of the experiments, especially that relating to the lower point, the position of 

 which has been modified several times. The publication of the observations being of 

 no interest, we give simply the values of the constants relating to two positions of 

 the closed branch. From the observations of the 3rd and 4th of May, 1898, the 

 correction relating to the barometer was found to be 



Cj, = 9-588 millims., 



and the correction relating to the manometer point in the raised position (for observa- 

 tions at and 100) 



C w = + 20-130 millims., 



whence the total correction for " index error " is given by 



C = C + C m = 10-542 millims. 



In the lower position of the manometer (used in comparisons between 100 and 200) 

 the total correction had the slightly different value C = 10'552 millims. 



