OF ARTS AND SCIENCES. 203 



closed flasks beside the calorimeter until the following day, or perhaps 

 during a longer interval. When ready for the caloriraetric determi- 

 nation one of the prepared solutions was transferred with rinsing to 

 the measuring-flask, (care being taken not to change unduly its tem- 

 perature in the handling,) and filled to 500 c.c. with distilled water 

 that had stood a day ^r longer on the same table. It was then poured 

 into the calorimeter and its temperature taken, after thorough stirring, 

 at intervals of five minutes. The change of temperature during this 

 interval should not be more than .02"^. At the same time, the second 

 solution having been transferred to the sebond measuring-flask, its 

 temperature in the flask was taken by a second thermometer in similar 

 manner. Its change likewise should not be more than .02° per in- 

 terval of five minutes ; and it is also important that it should be close 

 to that of the room, which was read from an ordinary thermometer. 



When the temperature conditions were satisfactory, and the read- 

 ings duly recorded, the contents of the second flask were poured as 

 quickly as possible into the calorimeter, the mixture thoroughly stirred, 

 and the resultant temperature recorded. The interval between the 

 last reading of the initial temperatures and the first of the resullaut 

 temperature was in this series of experiments one and a half or two 

 minutes, and with few exceptions the maximum temperature was 

 reached in this time. Readings were then taken, always with stir- 

 ring, every minute for about five minutes, and then at an interval of 

 five minutes, in order to have a record of the rate of cooling. 



There seems to be one unsatisfactory point in this method of 

 manipulation ; namely, the difficulty of obtaining the true temperature 

 of the second volume of liquid while it still remains in the flask, and 

 of transferring it to the calorimeter without consequent change. 

 Experience will convince one of this difficulty in experiments where a 

 variation of .01° is of considerable importance. My habit is to close 

 the flask and shake it as thoroughly as possible before inserting the 

 thermometer, and to see that its temperature at the time shall be such 

 that it need not be left more than five or ten minutes before trans- 

 ferring to the calorimeter. Meantime the flask stands in a loose bag 

 of flannel, supported by a frame that serves as handle so that the 

 hand itself is distant several indies from the flask. Nevertheless, I 

 have not infrequently obtained apparently accidental results, which 

 varied from the mean notably more than the avei'age variation, and of 

 which the difficulty just described seems the probable cause. I hope 

 later on to reduce, if not to eliminate, this uncertainty by a simple 

 modification of apparatus. 



