386 Dr. N. Ernest Dorsey on the Surface- Tension of 



from one of these positions to the other is over two centi- 

 metres ; and I can easily adjust the line to the centre of this 

 field to within one millimetre ; i. e. one-twentieth of the total 

 angle through which I turned the telescope. If the error 

 was proportional to the angle this would correspond to an 

 error of 01 per cent.; but the error varies almost as the 

 cosine of the angle. Hence the error introduced is negligible. 



Effect of Particles. 

 In the same article Prof. Tait shows that the effect of a 

 " uniform film of oxide or dust, in separate particles which 

 adhere to and move with the surface " is to increase the 

 wave-length if there is no viscosity. If terms involving the 

 first power of the viscosity must be taken into account, the 

 surface-layer will not affect the wave-length, but will aid 

 viscosity in causing the waves to subside as they advance. 



Water and Salts. 



The water used in these experiments was especially dis- 

 tilled by Dr. W. T. Mather, and was the kind he used for his 

 electrolytic work. He distilled it from chromic acid, and 

 from potassium permanganate, and condensed it in a block- 

 tin condenser. It was collected in a bottle used for no other 

 purpose, which was kept always covered or stoppered. It 

 was transferred from this to a larger bottle that had a hole 

 bored near its bottom so that the water could be drawn off 

 through a piece of black rubber tubing and a glass nozzle. 

 This tubing and nozzle always remained filled with water 

 which protected their inner surface from impurities in the air. 

 The glass nozzle was removed and cleaned before drawing 

 the water. In this way it was hoped to get water as pure as 

 possible. I might mention here that the water and solutions 

 were not free from air; but Rontgen and Schneider* have 

 shown that the effect of dissolved air is entirely negligible at 

 ordinary pressures. 



I used Eimer and Amend's chemically pure salts. 



The temperatures of the water and of the solutions were 

 determined by means of a Bender and Hobein thermometer 

 of " Normalglas " graduated to tenths of a degree Centigrade. 



The densities of the solutions of sodium chloride, of sodium 

 carbonate, and of zinc sulphate were calculated from the 

 values given by Kohlrausch and Hallwachs f ; those for 

 potassium chloride from the results obtained by Bender} ; 



* Wied. Ann. xvii. p. 207 (1882). 

 t Wied. Ann. liii. p. 14 (1894). 

 X Wied. Ann. xxii. p. 179 (1884). 



