40 Dr. J. A. Harker on 



threads, one, two, three, and six degrees in length, I could 

 obtain the value of the fundamental interval and the 

 correction for each whole degree of the scale. Except 

 the one for the fundamental interval, the corrections were 

 found to be so small that even in the measurement of 

 an interval, where the two corrections had to be added 

 together, they were almost negligible. 



The next question that might be raised is with regard 

 to the dryness of the steam formed in the boiler, and its 

 possible partial condensation on its way to the calorimeter. 



In order to maintain regularity of boiling in the glass 

 apparatus, in all the experiments a few pieces of crumpled 

 platinum foil were put into the generator, and in no case 

 was the flame allowed to impinge directly on to the glass, 

 the whole of the glass surface being surrounded by a 

 double layer of copper gauze. In order to test if liquid 

 particles, formed by the bursting of steam bubbles on the 

 surface of the water, could by any chance pass over into 

 the calorimeter, the following test was devised: — 



The boiler was filled with a 5 % solution of potassium 

 permanganate, and the distillate examined calorimetrically 

 in Nessler glasses. The experiment was repeated under 

 varying conditions and with different rates of boiling, in 

 some cases much faster than in any L determination, 

 but in no case could any certain colouration be detected 

 in the distillate, even though if one part in ten thousand 

 had been carried over in the liquid form it could easily 

 have been recognised. 



The further question, as to whether the steam, even 

 though dry on leaving the generator, might not on its 

 way to the calorimeter lose some heat by radiation, and 

 thus hold condensed water in suspension, was more 

 difficult to decide. One obvious thing to be done was 

 to make the path of the steam from the generator to 

 the condenser as short as possible. 



