Mr. F. D. Brown's Notes on Thermometry. 67 



the thread with a current of running water, and so, while pre- 

 serving it from the uncertain effects of conduction, radiation, 

 &c.j to render possible the observation of its exact temperature. 

 The value of (T— f) being thus correctly measured, that of m 

 is found to be constant for all values of !N~, and to differ but 

 little from 0*001545. It varies, however, with different ther- 

 mometers. 



The following experiments show most distinctly the truth 

 of this statement: — 



One of the standard thermometers mentioned in the first 

 section of this communication was partially surrounded by a 

 glass tube, a b (fig. 2), about an inch in diameter; this tube was 

 closed at the bottom with a piece of good cork, about 8 millim. 

 thick, through which the stem of the thermometer passed. The • 

 upper end of the tube a b was fitted with a cork, in which were 

 four holes — one for the stem of the chief thermometer, a second 

 for a thermometer to indicate the temperature of the water 

 contained in the tube, while through the two others passed the 

 tubes by means of which the current of water was maintained. 

 The thermometer thus furnished was fixed vertically in the 

 ordinary apparatus, A, for determining the 100°-point of 

 thermometers. The open end of A was closed with a thin 

 disk of brass, with a small central hole, through which the 

 thermometer passed. One degree was equal to about four 

 divisions of the millimetre-scale of the thermometer, the 

 readings of which were observed with a cathetometer, and the 

 fractions of a division measured with that instrument. It was 

 found that the readings of the thermometer under these con- 

 ditions were correct to "02 of a millimetre, or *005 of a degree. 

 The numbers given below are the means of three readings, 

 which, however, were nearly always identical. The thermo- 

 meter in the water was graduated to fifths of a degree, and had 

 been compared with the standard. 



The chief thermometer was first heated in the steam for an 

 hour, with two or three inches of the thread above the cork ; 

 it was then pushed down until the quicksilver was only just 

 visible above the cork, and the reading noted ; it was then 

 pulled up again, and readings taken in various positions, as 

 given in the following table ; finally the thermometer was 

 again pushed clown as far as possible, and the reading taken, 

 when it was found to be the same as before, showing that no 

 change in the 100°-point had supervened during the experi- 

 ment. Of several series of observations made in this manner, 

 the one contained in the following tat 

 they all led to precisely the same result. 

 F2 



