518 



Mr. E. C. C. Balv on the 



refilled during 



the 



and the thermometer was several times 

 work. 



The correction for the contraction of the glass bulb was 

 also determined, and an attempt was made to mea- 

 sure the coefficient of expansion of glass between Fig-. 1. 

 the limits -190° C. and 20° 0., the usual limits of 

 the experiments. The apparatus was a very simple 

 one, as shown in fig. 1, and depended on the con- 

 traction of a glass tube of known length. A tube 

 A was taken, which just slipped through the con- 

 tracted neck of the tube B. Both tubes were 

 sealed at their lower ends, and A rested on the 

 bottom of B by a fine glass point, to minimise heat- 

 conduction as far as possible. The whole apparatus 

 was dipped into liquid air, while water was kept 

 running through the inner tube A, thereby cooling 

 B while A was kept warm. The contraction suf- 

 fered by B was measured by observing with a 

 micrometer the distance between two lines, etched 

 one on A and the other on B, before and during 

 the cooling. The apparatus worked very well, and 

 a satisfactory measurement was obtained in which 

 the total length of B cooled was 89'0 cm. The mean 

 linear coefficient found was •0000073, 

 measurement between 0° and 10° being 



The measurements of temperature were corrected 

 accordance with the above found coefficient, the volume of 

 the thermometer-bulb being taken as 28*655 (1 — '0000219 /). 



For the temperatures dealt with in these experiments the 

 correction was about 0°'37. 



The probable accuracy of the results with the thermometer 

 may be judged from the volume of the thermometer-bulb, 

 28'655 c. c, which caused the probable error to be very 

 small. The reading of the volume of the gas in the burette 

 was certainly accurate to *02 c. c, which means a maximum 

 error of about 0°'02. In practice, however, the results agree 

 better than this ; for two determinations of the boiling-point 

 of oxygen, and two of the boiling-point of nitrogen, differ 

 respectively by only o, 01, the measurements being made 

 several days apart. The volumes of the various portions 

 of the thermometer were most accurately determined, and 

 there seems every reason to trust the measurements made 

 with it. As regards the practical use of the instrument, too 

 much cannot be said in praise of it. It is extremely quick 

 to take up temperature-changes and very rapidly worked, 

 and generally a most convenient and simple apparatus to use. 



^O 



Regnault's 

 '0000085. 

 were corrected 



