410 Prof. Rudberg o« the Variations "johich Temperature 



here compare them with these. In order to maintain a con- 

 stant temperature during the whole time lliat an experiment 

 lasted, it was necessary to have a particular apparatus, of 

 which 1 shall first give a description. 



A box, of the form of a parallelepiped, was made of white 

 iron, and so strong that four of its faces were double, and 

 formed with each other a shut-up space, which communi- 

 cated only at one side, and beneath one surface, with a small 

 steam boiler; and at the other side, above a second surface, 

 with the external air. The other two surfaces were formed with 

 plates of mica, so that there was an inclosed space which 

 contained only air, and which, when the prism was placed 

 in it, was heated by the steam, which, surrounding the four 



sides, circulated only in the space between the four double 

 surfaces, without being permitted to mix with the interior air. 

 The temperature of this space was indicated by a thermome- 

 ter put into a cork, which was introduced into a tube passing 

 through the two upper surfaces of the box, and which com- 

 pletely closed the tube, in oi'der to prevent the heated air 

 from escaping upwards. A similar tube passing through the 

 two lower surfaces of the box formed a free communication 

 between the interior and the exterior air, so that their elasti- 

 city remained always the same. In the middle of this tube, 

 without touching its sides, there rose from the centre of the 

 repeating circle a vertical copper rod, carrying on its summit 

 a plate, upon which the crj'stal was placed. This rod was 

 attached below to another plate of copper, which, instead of 



