682 PROCEEDINGS OF THE AMERICAN ACADEMY. 



exactly identical with the conditions under which the electromotive force 

 is to be employed ; as, for instance, the nickel wires will extend in each 

 case from a bath of hot oil to the cold air, and the difference of tempera- 

 ture between the oil and the air at any one particular temperature will 

 be practically identical in the two cases. Again, the nickel wires are to 

 terminate in mercury cups, which will in both cases be at the same tem- 

 perature with respect to the oil. Also the galvanometer will always be 

 connected to the nickel leads in exactly the same way. Again, the ter- 

 minals of the nickel wires which dip into the mercury were treated alike 

 in each case, — that is, polished with emery cloth, etc. 



5. Tliermo-electromotive forces in the galvanometer circuit. — Reversed 

 readings were taken which would eliminate any electromotive force in 

 the galvanometer circuit. In order that the galvanometer might he 

 as free as possible from any effects produced by changes in tempera- 

 ture, it was placed in a box 30 X 30 X 46 cm., lined with pressed hair 

 5 cm. thick ; a small opening, covered with glass, gave a view of the 

 mirror. A thermometer was placed inside to determine the changes of 

 temperature. When taking a series of readings under the best condi- 

 tions this thermometer would indicate no change whatever. The room 

 was maintained at a constant temperature to within 2° C, and during the 

 time necessary for a set of readings the variation in temperature was 

 usually much less than 0°.5. 



6. Electromotive force at the contact of the galvanometer terminals in 

 the mercury cups. — That an electromotive force of considerable magni- 

 tude might exist at this point was discovered by connecting the two 

 mercury cups in which the nickel wires terminated by a single nickel 

 wire, then placing an electric light near the mercury cups ; they were 

 heated unequally and a marked deflection of the galvanometer was 

 obtained. While both cups were maintained at the same temperature no 

 deflection was noted, but it is possible that the Thomson effect caused by 

 the wires coming from the air into the mercury would be different unless 

 the mercury cups were maintained at approximately the same tempera- 

 ture for different sets of readings, consequently they were surrounded with 

 compressed hair in a wooden box 25 X 25 X 25 cm., and this was main- 

 tained at room temperature. When the room temperature differed 

 greatly from the normal, no attempt was made to take readings. 



7. Resistances. — All resistances were accurate to .1 of 1 per cent or 

 better. All corrections for changes of temperature, etc., were made. 



8. Standard electromotire force. — The electromotive force used to 

 calibrate the galvanometer was obtained from a number of storage cells, 



