204 PROCEEDINGS OF THE AMERICAN ACADEMY 



Plate II. The balance rests on the plank shelf of a wooden case, 

 whose glazed doors are shown open in the drawing ; and as this shelf 

 is firmly secured to a thick brick partition wall behind, great steadiness 

 is seemed tor the instrument. Around the inside of the wooden case 

 are iiuiig folds of drapery, which surround the balance case and protect 

 the beam from radiations, while a curtain rolling on a spring fixture 

 enables the experimenter to uncover the front as necessary. Fastened 

 by flanges and screws to the under side of the same shelf is a box 

 made of tinned sheet iron, of somewhat larger dimensions than the 

 balance case. The front of this box can be almost wholly uncovered 

 by means of two large doors made of the same metal sheet, and se- 

 cured when closed by turning buttons, but shown open in the draw- 

 ing. In the metal top are two circular holes about half an inch in 

 diameter, corresponding to holes through the shelf and the base of 

 the balance case, and directly under the centres of the balance pans. 

 From hooks soldered to the under side of the pans the globe and its 

 counterpoise are suspended by means of brass wires, and the lengths 

 of the wires are so adjusted that the two glass vessels shall hang mid- 

 way in the metal box. The globe is hung from a wire stirrup, Plate I., 

 which swings from an eye at the end of one wire ; while on the neck 

 of the counterpoise is cemented a brass cap and hook, and both it and 

 the small subsidiary bulb are hung on an eye at the end of the other 

 wire. Of course these details might be varied. It is only important 

 that the whole system should swing freely, with as little friction as 

 possible, and that the balloon should be easily removed and replaced. 



Through a third hole in the top of the metal box and shelf, placed 

 just in front of the balance case, passes a thermometer (not shown in 

 the drawing) by which the tem})erature may be watched, and the in- 

 side of the box is painted with lamp-black so as to secure a uniform 

 temperature throughout the interior. In our previous paper we spoke 

 of the disturbance sometimes caused by currents of air within the box, 

 but we found on further experience that these currents could be pre- 

 vented by keeping the interior of the box as nearly as possible at the 

 temperature of the external air. For this purpose it is important that 

 the air of the room should circulate as freely as possible around the 

 box, and the required condition is more easily secured if the box is 

 made of thin metal sheet. It is also important that the temperature 

 of the room should not be constantly changing, and that all circum- 

 stances should be avoided which would cause a flow of heat either into 

 or out from the box. Sitting in front of the box in the process of 

 weighing immediately causes a disturbance, and hence the division of 



