466 TRANS. ST. LOUIS ACAD. SCIENCE. 



substantially as in 1S7S. The work was done on the south pier of 

 the clock room. The temperature of the room varied slowly from day 

 to day, and for most of the work advantage was taken of these fluc- 

 tuations 'in order to observe at different temperatures. Beginning 

 with Mar, 23d, however, the temperature of the magnet was raised 

 to between 90° and 103° F. by means of a water-jacket of sheet cop- 

 per which surrounded the magnetometer except at the bottom, which 

 rested upon the flat stone with which the pier is capped. The front 

 wall of the jacket was provided with a window to admit of telescopic 

 observation, and the sides were provided with small windows for 

 illumination of the scale, and to allow the reading of the thermome- 

 ter which was placed inside the magnet box. These windows were 

 closed with double walls of thin mica. The top of the jacket was 

 pierced to admit the passage of the torsion head, the annular space 

 around the suspension tube being packed with cotton to prevent the 

 hot air of the interior from escaping. The jacket was of course ex- 

 amined to determine whether it acted magnetically, but no such 

 effect could be detected. It was, of course, necessary to remove 

 the water-jacket for a few minutes in order to adjust the inertia 

 ring, but care was always taken to allow time enough to elapse to 

 secure a constant temperature. 



The time of oscillation was determined by means of a Bond chro- 

 nograph. No. 171, connected in circuit with the astronomical clock 

 Howard No. 214. 



Part of these observations were made by Mr. W. R. Vickroy of 

 the Junior class, who also made all of the calculations, which were 

 afterwards revised by myself. In the reduction the observations for 

 Feb. 16 were rejected. 



The observations and reductions are given in the following tables, 

 where / is the temperature of the magnet during the loaded series, 

 T' the time of vibration of the magnet loaded with inertia ring Y 

 and corrected for torsion, T the time of vibration of the magnet 

 alone, corrected for torsion, and reduced to the temperature of the 

 loaded series. These differences of temperature were always small. 

 k\ is the moment of inertia of ring Y at the temperature /,* and 

 k^ is the moment of inertia of the magnet at a temperature /, while 

 kf^ is the moment of inertia of the magnet at the mean temperature 

 of 64.7 F. 



* Trans, vol. iv. No. i, p. 86. 



