767 



that in the meantime the resistance has elianged, and (he resistance 

 in the resistance box must also be changed. 



It is simple lo determine the volume that was occupied by (he 

 mercury between the two i)laces of contact. Let G be the weight 

 of the expelled mercury, S^ and S^ the scalar heights which have 

 been read, and ƒ the weight of one scalar division of mercury, then 

 \G + ('Sj— aSJ/I : A is equal to the required volume. (A is the specific 

 weight at the temperature used). In this way the great advantage 

 is reached that in the measurement any cocks and other movable 

 parts are avoided. 



Thus the difFei'ent reservoirs arc calibrated. The ii[)most reservoir 

 is in an exceptional case. First of all tliere is no place of contact in 

 it. It would not be practicable to make one there. Besides, to clean 

 the tube after contamination the upmost point mirst be knocked off. 

 In order to enable us yet to accurately know the volume of the 

 upmost reservoir every time, the tube is drawn out very thin at the 

 top On this narrow part lines are etched at some millimeters' dis- 

 tance. When the tube is quite filled with mercury, the positions of 

 the mercury at the etched lines can every time be compared with 

 the position of the mercury along the scale. Thus the volumes can 

 be expressed and calculated from line to line in scalar divisions, 

 and also those from one of these lines to the upmost contact, after 

 a quantity of mercury has been expelled. 



To determine the weight of a scalar division of mercury, we 

 make use of one of the places of contact. When the deviation of the 

 galvanometer needle has been compared with the level of the mei'cury 

 along the scale, we press out a drop of mercury, and again compare 

 the mercury level with the same place of contact. The decrease of 

 height agrees with the expelled drop of mercury, which is weighed, 

 and then the scale is at least partially gauged. This can be done 

 for different parts of the scale. Care can further be taken always 

 to work within a certain, pretty small part of the scale. And the 

 tube being very narrow, the difTerences of position are only to be 

 taken into account as a correction. 



During the measurements the whole tube is of course placed in a 

 tiiermostat. On account of the length of the tube the thermostat is thus 

 constructed. A glass tube of -t 6 cm. diameter passes at the bottom into 

 a narrower tube, which is connected by a rubber tube with a large 

 copper tnixing vessel, where the water is kept at the desired temperature 

 by means of a toluol-lhernioregulator, slii-rer, and burner. The glass 

 tube is placed so high with respect to the liquid level in the mixijig 

 vessel that the measuring tube which is to be gauged, is quite 



