686 PROCEEDINGS OP THE AMERICAN ACADEMY. 



5. Variation in source of energy. — The energy was supplied from a 

 battery of storage cells. A rheostat with a very fine adjustment was 

 connected in series, so that any change might be counterbalanced. Ad- 

 justment of this rheostat was made every minute if there was any change 

 in the current. In this way the movement of the needle of the ammeter 

 was less than one twentieth of a division. 



6. Energy supplied by agitating fans in the cavity. — These fans 

 rotated about 800 times per minute. The energy supplied was deter- 

 mined by placing the coil with fans in a calorimeter containing oil, rotat- 

 ing the fans at the required rate, and determining the rate of increase of 

 temperature not exceeding 1° C. Energy was then supplied to the coil 

 until the rate of increase was approximately double. 



Three determinations resulted as follows : 



0.366 watts 



0.344 



0.385 



Mean, 0.365 ± 0.10. 



Since the total energy supplied was about 100 watts, this correction was 

 determined with sufficient accuracy. 



7. Correction on dimensions of shell. — The shell was measured by a 

 special micrometer caliper made by Brown and Sharpe for this work. 

 This was compared with a standard meter, which was compared by the 

 Societe Genevoise with their standard. 



8. Variation in the thickness of the nickel. — When the nickel was 

 being plated to the lead it was first plated on the outside, frequent 

 measurements being taken to insure its being of a uniform thickness, the 

 electrodes being adjusted when it was necessary to increase the thickness 

 at any particular place. 



During the plating of the inside, measurements were taken in the 

 same way, so that the thickness of the nickel probably does not vary 

 over 0.2 mm. However, in the manipulation of the hemispheres small 

 portions of the nickel cracked off, from the outside near the edge of the 

 hemispheres. The amount cracked off" in this way is approximately 

 5 sq. cm., the largest amount in one place being perhaps 2 sq. cm. 



9. Conduction of the heat from the shell at the poi?its where the nickel 

 wires were electroplated. — In electroplating the wires to the shell the 

 nickel on the shell was filed slightly and a small depression made in 

 which to put the wire. After plating over the wire, the thickness of the 



