THE INDEX ERRORS OF THERMOMETER SCALES. 13 



upright, and giving it a slight jerk, the vacuum will be brought to 

 the aperture of the fine tube, and when the position of the ther- 

 mometer is again reversed, the column of mercury in the tube will 

 separate from that in the bulb, leaving a vacuum in the tube. 



The bulb is then to be gently heated until the remainder of the 

 mercury in it rises to the lowest point of the scale to be proved, and 

 by again holding the tube in an upright position, the columns run 

 together, A gentle shake will insure their joining, which is essential. 



Let the bulb be now immersed in snow, or ice and water, until the 

 mercury has nearly reached 32°, and it will be found that on again 

 reversing the position of the tube, the mercury will separate at the 

 same point of the scale, at which it had joined. 



The thermometer being placed in a horizontal or slightly inclined 

 position, with the bulb surrounded with wet snow, the separated 

 column of mercury can be made to move upwards, (this is best effected 

 by gentle taps at the end of the tube,) as required, while the observer 

 reads the lower and upper extremities at such intervals as he may 

 find necessary. The readings may be written down as follows : — 



From — 30 



to 



33.2 



— 29 



to 



34.2 



— 28 



to 



35.1 



— 29 



to 



36.1 



— 26 



to 



37.0 



32 p to 95.0 

 By applying the corrections previously known at 32° and 95°, we 

 find the true length of the column of mercury in degrees of the scale, 

 and the true reading at each of the lower points of the scale may be 

 obtained by deducting the number of degrees in the column from the 

 upper reading (corrected) thus : 



Correction at 95° +'5 — true reading, 95° 5 

 Do. 32 _ 0— Do. 32. 



Column of mercury 63° 5 



Upper reading, 32° 

 Col. 63 5 



Correct reading for — 30° is — 31 .5, — error at — 30 = — 1.5. 

 It is scarcely necessary to mention that, in applying the above method, 

 the greatest care must be taken that the length of the column of 

 mercury is not affected by change of temperature. 



