of Edinburgh, Session 1879-80. 
539 
“ central position ” being that in wbicb B and B' are equidistant 
from the centre of r b. 
(7.) A lamp and scale, with proper focussing lens if the 
mirror is not concave, are applied to show and measure small 
deflections as in my mirror galvanometers and electrometer. 
Use of the Thermoscope. 
(8.) Place the instrument with the needles approximately per- 
pendicular to the magnetic meridian, turning it so as to bring 
b and V to the south side of the vertical plane bisecting the small 
angle between the projections of r b , r' b , and r and r' to the 
north side of it. 
(9.) By aid of the micrometer screw bring the luminous image 
to its middle position on the scale, 
(10.) Cause R B, B' R' to have different temperatures. The 
luminous image is seen to move in such a direction as is due to r 
approaching the cooler, and receding from the warmer of the two 
deflectors B R, B' R/ 
5. On a Constant Pressure Gas Thermometer. 
By Sir William Thomson, F.R.S. 
In the article on “Heat” published in the eleventh volume 
of the Encyclopaedia Britannica , referred to in my previous com- 
munications to the Royal Society on Steam Pressure Thermometers, 
it is shown that the Constant Pressure Air Thermometer is the 
proper form of expansional thermometer to give temperature on the 
absolute thermodynamic scale, with no other data as to physical 
properties of the fluid than the thermal effect which it experiences 
in being forced through a porous plug, as in the experiment of 
Joule and myself on this subject;* and the thermal capacity of the 
fluid under constant pressure. These data for air, hydrogen, and 
nitrogen have all been obtained with considerable accuracy, and there- 
fore it becomes an important object towards promoting accurate ther- 
mometry, to make a practical working thermometer directly adapted 
to show temperature on the absolute thermodynamic scale through 
the whole range of temperature, from the lowest attainable by any 
* “Thermal Effects of Fluids in Motion,” Trans. Roy. Soc. Lond., June 
1853, June 1854, June 1860, and June 1862. 
