(53 



Ti nililier tube. The tube is closed at the lower end and is some shorter than the 

 thermometer, so that a little pressure is recjuired to push it in far enough to 

 allow the cap to screw on firmly. This pressure is directly on lower end of bulb, 

 and is more than a person would at first think. By repeated tests I found it 

 e(juiv:).lent to 240 grams, or a little over a half pound. Such a pressure acting 

 continuously for some length of time would certainly change the shape of the 

 bulb, and conseciuently the zero mark. 



The length of the bulb is 25 mm. Its volume is approximately .3 cu. cm., as 

 near as can be ascertained by measurement of its dimensions. Tlie weight of the 

 thermometer is 43 grams. 



On the 16tli of November, after a period of five weeks, the pressure was 

 released, the thermometer placed in an ice and water bath and the exact position 

 of the mercury column noted. Observations were made twice per week from that 

 date, the last one being Saturday, December 21. Great care was taken in making 

 these observations. The bulb was placed in an ice and water bath, while the 

 stem for five inches above the zero mark was packed with finely broken ice. The 

 added water was to equalize the pressure on the bulb. An aperture in the side of 

 the vessel, through which passed a tube, the outer end of which guarded by a 

 plane glass window made it possible to readily observe the mercury column, and 

 yet have it completely surrounded with ice. Each observation extended over a 

 period of tiiree hours. To guard against jarring, the cathetometer and vessel 

 holding the thermometer were placed on a stone pier. 



The apparatus was allowed to stand for one hour before taking a reading, 

 after which readings were taken every half hour. It was observed that when 

 great accuracy was expected, all of an hour is required as the glass is very slow 

 to take up the temperature of the melting ice and adjust itself accordingly, 

 while the mercury takes up the temperature in a very few minutes. The first 

 readings, therefore, are always too low. Before taking a reading the stem was 

 jarred to facilitate the adjustment of the mercury. To prevent radiation the 

 vessel was covered with a towel. After putting ice in the vessel it was thoroughly 

 /washed with distilled water. This last precaution was at first overlooked and the 

 result was that the readings were far too low, i. e.. the melting mixture was made 

 •colder by its containing foreign substances. 



We would naturally expect that the pressure on the lower cud of the bulli 

 would considerably change its size, and that a pressure of over a lialf i)Ound could 

 not continue long without considerably altering the size, volume and accuracy of 

 the thermometer. In the case under discussion the volume of the bulb would be 

 increased by pressure on lower end. Also since the length of the bull) is 25 mm. 



