Temperature of Mercury in a Siphon Barometer. 251 



posite to each other, about four tenths of an inch wide, parallel 

 to the axis, and extending 9 or 10 inches from the extremities to- 

 wards the centre. The design of this is to expose to view the 

 extremities of the column, and to allow the requisite movement 

 to the vernier. The remaining central portion, which is about 

 15 inches in length, and has no perforation through its surface, 

 embraces the attached thermometer. This arrangement evident- 

 ly exposes the bulb and the tube to different influences. The 

 bulb is comparatively small, and its connection with the mount- 

 ing is more intimate than it is with the tube, and consequently 

 it tends to indicate the temperature of the former more truly than 

 that of the latter ; and more especially so as a like connection 

 between the tube and the mounting is unattainable with a due 

 regard to the safety of the instrument. And besides, allowing 

 the attached thermometer to mark truly the temperature of that 

 part of the column in its vicinity, and which, like it, is protected 

 by the mounting, it might materially err in respect to the remain- 

 ing portions which are exposed to the direct and variable influen- 

 ces of the atmosphere. These remarks, suggested by the con- 

 struction of the instrument, I verified experimentally in the fol- 

 lowing manner. I filled with mercury a tube 14 inches long and 

 of a diameter not much larger than that of the barometric tube, 

 and inserted into the open end a tube of less diameter ; joining 

 the two firmly with sealing wax. This was introduced into the 

 central part of the brass mounting. I took a series of observa- 

 tions with this instrument in a cellar where the temperature was 

 low and uniform, and another series in a room where the temper- 

 ature was high and also uniform. In both these series of obser- 

 vations the temperature indicated by the attached thermometer 

 must have been very nearly the same as that of the mercury in 

 the tube. Denoting the means of these temperatures by {t,) 

 {t") respectively, and the mean altitudes of the upper surface of 

 the mercury by (a,) {a",) we evidently have t"=t-\-A.{a" — a); 

 A being a constant depending upon the volume of mercury and 

 the diameter of the lesser tube. For the same reason, if (f) is 

 the temperature of the mercury at any other time, and (aQ the 

 reading of the upper surface, we have f=t-\-A{a' — a). Elimi- 

 nating the constant A between these two equations, and resolving 



for (f,) there results i'=i-ZL(a' -a)-\-t (1). I then made the 



following observations. 



