HALL. — THERMAL AND ELECTRICAL EFFECTS IN SOFT IRON. 51 



over, the ends of the uppermost guard-ring bars were not all the time 

 fully covered by the water in the end vessels. In fact, these two bars 

 were chosen to carry the junctions X, T, Z, etc., for the reason that they 

 seemed fairly representative, one of conditions in the upper half of the 

 guard-ring, the other of conditions in the lower half. 



The observations of December 22 indicate that the mean tempera- 

 ture of the guard-ring bars, for the 17 cm. length extending from the 

 vertical plane of the point X to the plane of the point X', was at the 

 time of these observations about 5° below the temperature of the boiling 

 water in which the ends of the bars were immersed ; but a close exam- 

 ination of the observations shows that the apparatus as a whole had not 

 yet reached a perfectly stable condition of temperature. The thermo- 

 electric observations began when the water at the ends of the bars had 

 been boiling not much more than two hours and ended about two hours 

 later ; and the earlier observations of the series showed a slightly lower 

 temperature for the bars than the later ones. It is not unlikely that, if 

 the experiment had continued some hours longer, the mean difference of 

 temperature between the boiling water and that part of the guard-ring 

 which lies between the planes of X and X' would have fallen as low as 

 4.5° ; and this is taken as the final value. See p. 35. 



On December 9, 1904, with all the usual conditions of an experiment 

 on the Thomson effect, observations were made for the purpose of deter- 

 mining, approximately, the temperatures of the mid points of the couples 

 KK' , N N', etc., on the bars a and (3. The outer junctions belonging 

 to E, F, G, etc., were placed in oil tubes, in ice and water, and after 

 the terminal thermal conditions of the main apparatus had been estab- 

 lished, and the main electric current had been running nearly eight 

 hours, observations on the current from the couples ./and G., joined in 

 series with each other, were made. They were followed by observa- 

 tions on couples / and F, joined in series; and so on. The general 

 course of these observations was very satisfactory, and it indicated the 

 following values : 



Mean temperatures of the points J and G = 90.2° 



" " " I " F — 51.0° 



" H " E = 13.0° 



On the same day, and after the observations just described, similar 

 observations were made with the couples X, Y, Z, etc., taken singly. 

 The following values were indicated : 



