550 Scientific Proceedings, Royal Dublin Society. 



about 4 inches deep and of 8 inches diameter, covered with a 

 wooden cover pierced to receive the tube. A little oil (pure sperm 

 oil seems to answer the purpose well) is floated on the surface of 

 the glycerine in the bath to preserve the latter from taking mois- 

 ture from the atmosphere. 



The mode of setting up the barometer is as follows: — The 

 mercury (26 J") is first poured into the tube. If all air is to be got 

 out a rubber stopper with a tube leading to an air-pump is inserted 

 in the open end and the pump worked. While the vacuum is 

 maintained the tube is cautiously inclined so that the column 

 becomes detached all down the upper surface, and this is done 

 several times, rotating the tube axially a little each time till all 

 bubbles are got out. The cylinder is then dropped in, float 

 downwards, and sufiicient mercury is poured in after it to sink by 

 its weight the float up to the disk. The glycerine is now slowly 

 poured down the tube. The glycerine used should be pure, and 

 must previously have been warmed and let stand under the ex- 

 hausted receiver of an air-pump for some hours to extract dissolved 

 air from it as recommended by Mr. Jordan. When quite full a 

 cork is inserted, squeezing out some of the glycerine and leaving no 

 air bubble beneath it. The tube is now inverted and fixed in its 

 place upon its board ; the corked end of the tube dipping below the 

 glycerine in the bath. The cork is now withdrawn slowly, so as not 

 to let the column descend too rapidly. The small quantity of 

 mercury used in sinking the float of course falls out into the glyce- 

 rine bath. The mercury column descending leaves a vacuum above 

 it. Some glycerine will finally descend upon the upper surface of 

 the mercury column from the wet walls of the tube. The proportion 

 of the float to the area and weight of the disk is important. Ebonite 

 and ivory are almost of the density of glycerine, so that these sub- 

 stances exert but little downward pull upon the float. On the 

 other hand, a certain force is required to preserve the surface 

 of the cylinder pressing against the mercury column, while too 

 much force may lead to mercury being pressed down past the 

 cylinder. With a three-quarter inch diameter ebonite cylinder, 

 having a length equal to its diameter, about, I found a cylindrical 

 float half an inch long and about one-fifth inch diameter suitable, 

 or a spherical float may be used as in the figure. In the air-ther- 

 mometer alluded to previously an adjustment was provided in 



