228 Dr. J. Monckman. Occluded Gases and [May 31, 



Fig. 6a. 



from the platinum wire, F, F platinum wire insulated ; W a vessel of 

 water containing a brass tube E, closed at one end, in which the 

 carbon rod B is placed after each contact. 



During the first series of experiments the temperature of W, and 

 hence of B, was 16° C, that of A was changed in each contact, rising to 

 480° C. and higher. At about 480° the deflection changed ; decreasing 

 on approaching that temperature, and changing sign above it. I am 

 sorry to say that the difficulty of obtaining the same perfection in each 

 contact was so great that the deflections, although increasing above 

 480°, were not sufficiently consistent to allow a curve to be drawn. 



Therefore, assuming that the nentral point was midway between 

 that of the two rods when no current was produced (i.e., 16° C. and 

 480° C.) we get 248° C. for the temperature of that point. 



B being kept in the second series at 50°, in the third at 100°, and in 

 the fourth at 200°, and the same assumption made in the calculation 

 as before, 255° C. was given as the neutral point. If we now rule a 

 line such that any two points being taken in it, the current shall be 

 equal to the vertical distance between them, and shall flow from the 

 higher point to the lower, it will have its lowest point afc from 248° to 

 255°, rising to 0° and 480° and above (see fig. 7). This assumes that 

 the two lines are equally inclined, and from the experiment with a 

 platinum- carbon couple we judge them to be so, and their turning 

 point to be 250° C. 



From the preceding experiments I was led to expect that the line of 

 carbon in a thermo-electric diagram, in which the area of the space 

 between the lines is proportional to the electromotive force, would 

 show a bend of some kind, and as no researches were known showing 

 such a bend, it appeared desirable to test it carefully. 



There is a paper by E. Becquerel in which he gives an account of a 



