COlJ Prof. Rankine and Mr. Smith on the Viscosity of 



been made in order to make comparable the times at definite 

 tempeiatures. It will be seen that although these times at 

 15° C. vary by as much as 2 per cent., the corrected times 

 in the last column are consistent to within 0'15 per cent. 

 The variations in the capillary correction x are also worthy 

 of" notice. The changes of this quantity ate presumably 

 attributable to variations of temperature, and the substitution 

 of other gases for air ; and, although the exact relation 

 between cause and effect has not yet been fully examined, it 

 appears to be clear that, generally speaking, rise of temper- 

 ature produces a considerable reduction in the correction, 

 and that the temporary exposure to a new gas may effect a 

 semi-permanent alteration. For example, the low v;due of 

 x corresponding to observation (7/) seems to be definitely 

 attributable to the fact that the apparatus contained phosphine 

 just previously. In spile of this variability, however, the 

 above procedure enables the capillary effect to be practically 

 eliminated, and ratios of gaseous viscosities to be determined 

 with an accuracy of 01 per cent., or thereabouts. 



The mean corrected time for air derived from the above 

 figures is at 15° C. 1 25*28 sees., whence the ratio 



^7io 1^0-28 

 This gives for Sutherland's Constant, 

 0=117, 

 which is very close to the recognized value. 



Ammonia. 



The sample of ammonia was obtained from a solution of 

 ammonia in water, by raising the temperature to about 

 30° C. and reducing the pressure in the generating apparatus 

 to about half an atmosphere, by means of a water-pump. 

 The ammonia was rapidly evolved, and dried by passing it 

 over soda-lime and caustic potash. It was then solidified in 

 a tube surrounded by liquid air. A sample was collected 

 over mercury by removing- the liquid air and allowing the 

 ammonia to evaporate Before transferring the sample to 

 the viscometer it was further purified by being fractionated 

 at liquid air temperature, all permanent gas being pumped 

 off by means of a Toepler pump. The ammonia was then 

 allowed to evaporate and fill the viscometer at atmospheric 

 pressure. Since after this operation had been completed, 

 some liquid ammonia still remained in the bulb where it had 



