AND KATK) OF THKIK ATOMlf WKIGHTS. 



23 



comparison witli the level of the merciiry in tlie tubes behind them when all were 

 open to the atmosphei-e. The tubes wei'e of nineteen millimeti'es in internal diarae- 

 tei-, so that tliflferences of capillary depression could not be gi'eat, but the scales 

 were adjusted so as to eliminate any diflference wliicli ni;i\ liaxe existed. 





ii-^ 





i^^r 



^^' 



^^tf^*s . 



Fig. S. — Barometer antl reading microscopes. 



Tiiese four tubes all stood in a cast-iron box filled with mercurv ; connected 

 to this was an iron cylinder open to the air and filled with mercury so that the mer- 

 cury in the tubes stood a few centimetres above the bottom of the casing full of 

 water which surrounded them. 



Before the mano-barometer stood an iron column cairying the reading micro- 

 scopes. As is well known, a cathetometer with a scale somewhat removed from 

 the length to be measured cannot be trusted like one in which the scale is neai- this 



