72 Notes concerning the Phenomena 



nearest pith first of all recedes from the name, bnt very soon 

 turns and swings towards it in the usual manner.' 



"It occasionally happens, when the name is placed exactly 

 equidistant from the two pieces of pith, that the balance 

 remains in unstable equilibrium for four or five minutes, 

 until, probably, some little accident, such as a current of 

 outside air inpinging on the glass, gives one or other pith 

 the advantage. 



" The two opposing forces that suggested themselves 

 were : — Air currents within the shade, setting towards the 

 name and causing the pith to approach, and evaporation 

 from the surface of the pith or glass tending to withdraw the 

 pith from the flame. 



" Experiment 3. — With a view to connecting air currents 

 with the production of these motions, the following 

 experiment was devised. 



"A pith-ball balance was suspended within a dry flask, 

 and a piece of phosphorus introduced. A thin vertical 

 stream of fume rose from the phosphorus. When a flame 

 was then placed near, the fume took the direction of the 

 flame before the balance itself became sensibly affected, and 

 flowed in a circle rising on the side next the flame and 

 descending on the opposite side. When, again, the relative 

 position of the flame to any part was altered 180°, the 

 direction of the ascending current of phosphorus fume was 

 reversed, it then ascended on the side next the flame in its 

 new position, and soon the pith swung round to that part. 

 The setting up of the current was antecedent to the rotation 

 of the balance, and the reversal of the current was ante- 

 cedent to the reversal of the direction of the balance. 



" Experiment 4. — In order to examine further the action 

 of convection currents, another medium was employed. 



" A balance made of a glass rod drawn out so as to leave 

 heavy ends, was suspended horizontally in a vessel of water 

 by a fibre of silk. When a flame was placed near it and 

 allowed to remain for some time, the rod slowly but steadily 

 swung round, until it pointed straight at the flame, in which 

 position it remained stationary. 



" As evaporation has no place in this experiment, the 

 position taken up by the rod with reference to the flame is 

 quite what might have been expected — setting in the line of 

 least resistance. 



" The course of the current may be rendered visible by 

 dropping a little colouring matter into the liquid just over 



