132 THE ROYAL SOCIETY OF CANADA 



about 5 c.c. capacity is blown. E is a wire with a platinum end which 

 is fused through the side of the tube. A is another wire with a plat- 

 inum tip at the bottom, which is held in position by an adjustable 

 collar resting at the top of the tube, and centred by a bead fused on to 

 the wire. The collar and the bead must allow the free passage of air. 



It is necessary to fill the grid with toluene through the end which 

 is to be closed. Great care must be taken to remove all air by heating, 

 both from the toluene and the mercury. After the closed end is 

 sealed the mercury is let in through D and B. The elimination of air 

 inside is of great importance — even an invisible quantity may be 

 sufficient to affect the regulation. 



The grid should be mounted near the bottom but above the 

 stirring blades, the shaft of the stirrer passing upwards between the 

 bends of the tube at the centre. The part D B should be arranged 

 to pass over the edge of the container, and the mounting should be 

 slightly loose so that the vibration of the stirrer may be imparted 

 to the thermoregulator, as this helps in preventing the mercury from 

 sticking. 



The electrical arrangements are simple. A certain amount of 

 current (sample numerical data given below) passes through some 

 lamps on an ordinary lamp board, from the lighting circuit into the 

 heating coils. This should be adjusted so that the bath is maintained 

 several degrees below the temperature desired. The regulation is 

 obtained by the automatic turning off and on of additional lamps in 

 parallel with the first lot. These are operated by an ordinary telegraph 

 relay which is connected to A and E, with a storage cell in series, 

 and is thus operated by the expansion and contraction of the toluene 

 and mercury. To insure good contact between the wire A, and the 

 moving mercury, a little light oil should be dropped through C. 

 To avoid excessively carbonising the oil, or colloidalising the mercury 

 b}- sparking, the voltage used in the relay circuit should be low, and a 

 small condenser (e.g. made of metal paper and mica) should be placed 

 in parallel with the "break." It is advisable also to make the wire 

 A negative. 



The thermoregulator insures constancy in temperature at any one 

 spot, but the elimination of variations between different parts of the 

 bath is obtained by rapid stirring, although within certain limits this 

 appreciably increases the radiation from the bath. A rapidity of 

 stirring just insufficient to produce wavelets and floating bubbles was 

 found in our apparatus to lead to a maximum difference between 

 places in the bath, of less than 0-002°C. (This does not include the 

 small region below the heating coils). The matter can always be 



