670 PROCEEDINGS OF THE AMERICAN ACADEMY. 



The liquids must evidently be stirred both before and after mixing. 

 Because of the small available space in the entrance tubes of the ther- 

 mostat-pipettes the thermometers were utilized for stirring the initial hot 

 and cold liquids. For this purpose small spiral-shaped pieces of silver 

 foil were attached to the lower end of the thermometers projecting 

 somewhat below the bulb. The upper end of each thermometer was pro- 

 vided with a bearing from which it hung, and which permitted the ther- 

 mometer to rotate about its long axis. A thin glass tube was snugly- 

 fitted over the shaft of each thermometer, and this, rubbing against the 

 upper tube of the pipette, furnished the second bearing. The spiral- 

 shaped piece of foil acted like a screw, and produced circulation of the 

 liquids in the pipettes. 



The stirring in the cold solution was discontinued as soon as the ther- 

 mometer immersed in it showed that the constant temperature of zero 

 had been reached. Tlie mixed liquids in the platinum can were stirred 

 by means of an inverted glass T-tube, provided W'ith holes along its ver- 

 tical arm, and rotating at a moderate speed. The belting which ran this 

 was so arranged as not to interfere with the free vertical motion of the 

 calorimeter (see diagram). A simple catch was also provided, such that 

 a single motion sufficed to start or stop this stirrer. 



Unfortunately a few drops of the liquid always fell from the pipettes 

 after the main body of the liquids had rushed out. We could not be at 

 all sure that these drops were at the same temperature as the rest of the 

 liquid. For that reason an arm was constructed which carried two glass 

 cups, and which could be swung beneath the ends of the pipettes the 

 moment the calorimeter had been lowered. 



For results accurate only to between 0.1 and 0.2 per cent, it would be 

 possible to neglect both these drippings and the liquid remaining in the 

 pipettes. For more accurate measurements, the amounts of these drain- 

 ings and drippings must be determined. At first it was planned to weigh 

 them, having collected the liquid left in the pi|>ettes in suitable absorp- 

 tion tubes after evaporation by means of a current of dry air. This 

 proved to be a very tedious process. 



A much more expeditious and satisfactory method was finally adopted 

 with water or aqueous solutions. Instead of dry air, the pipettes were 

 rinsed with dry acetone, the drippings added to this, and the whole made 

 up to 50 millilitres at 25°. The water in the mixture was then deter- 

 mined by means of specific gravity measurements, and was subtracted 

 from the total weight of water originally in the pipettes in order to find 

 the amount which actually found its way into the calorimeter. 



