228 Mr. W. C. Dampier Whetham. On the alleged [June 19, 



was increased by placing a partially exhausted space in connexion 

 with the lower end of the tube, the size of the bulb being also 

 increased, The pressure was read by a kathetometer on a gauge of 

 sulphuric acid or mercury, and a small correction (always less than 

 0'2 per cent.) made in order to compare the times of flow under the 

 same pressure. The temperature of the water was observed imme- 

 diately after it had passed the transpiration tube, by reading a ther- 

 mometer immersed in the cylinder C, at equal intervals, while the 

 water was running. The smallest tubes which could conveniently be 

 silvered had a diameter of rather less than a millimetre, and in this 

 series of experiments the tubes had diameters of 0'084 and 0'072 cm. 

 The results of four series of experiments showed that in three cases 

 the times for the silver and glass surfaces agreed to less than 0'5 per 

 cent., and in one case, where the silver was rather too thick, and 

 probably a little irregular, the time of flow for the silvered tube was 

 1 per cent, greater. 



Even if the slipping effect was only half that deduced by Helmholtz 

 and Piotrowski for gold, the times ol! flow should have been for these 

 small tubes from fifteen to twenty times less for silver than for glass. 

 The two series of experiments given above may therefore be considered 

 a quite satisfactory proof cf identity. 



Fia. 3. 



III. Girard's experiments on copper tubes give X = 0'3984 mm. 

 The times of flow for a tube of 1'83 mm. in diameter were about 

 three times less than that given by Poiseuille's law. These experi- 

 ments were repeated with some solid drawn tubes, kindly made for 

 me by Messrs. Elliott, of Selly Oak, Birmingham. The diameters were 

 estimated to about 01 per cent., by weighing the tube first empty, 

 and then full of water. By some subsidiary deter ruinations with glass 

 tubes, I showed that this gave the same result as the usual method 

 with mercury. The results in all cases showed agreement with 

 Poiseuille's observations. The times of flow were always a little 



