328 Mr. L. Lownds on the Thermomagnetic and 



level so that there was also good contact between D and A, 

 and E and F. Finally, to improve the contact of the jaws 



Fiff. 3 



with the plate, and to prevent fracture of the crystal, narrow 

 strips of tinfoil were placed between the crystal and the 

 coppers before screwing firmly down. The coppers D and E 

 served as electrodes for measuring the longitudinal effect and 

 the thermoelectric force. The fall of temperature in the 

 plate was obtained by placing in A either liquid air, a mixture 

 of carbon dioxide, snow, and ether, ice or boiling water. 

 Water at room-temperature flowed through G. It is necessary 

 to know the temperatures at the upper and lower edges of the 

 crystal, and for this purpose small screws d and d! passed 

 through D and E, and pressed the junctions of thermocouples 

 of iron and constantan firmly on the face of the plate. The 

 temperatures given by these thermocouples were taken as 

 those of the contacts of the copper with the crystal. 



The electrodes for measuring the transversal effect were 



