230 Messrs. Lodge and Clark on Dusty Air in the 



free space of liquid beneath the tube. Under the influence of 

 the beam this becomes thinner, and two dust-free planes rise 

 from each side of the tube, ultimately meeting upon the top 

 of the rod, and often at first enclosing particles between them, 

 as occasionally observed in gases. A dust- free black coat 

 and plane are then seen, distinct but rather thin. This seems 

 to be a permanent state so long as these conditions are pre- 

 served unaltered. The convection-currents are fairly well 

 developed, and the velocity-distribution resembles that de- 

 scribed for a warm solid in dusty air. If now a current of 

 steam be sent through the brass tube, the convection-currents 

 become more rapid, and the coat so thin that we are not certain 

 of its existence. Moreover, the distinctly dust-free black 

 plane is no longer quite certainly there, although a casual 

 observation would probably lead to the conclusion that it was. 

 A long fine up-streaming layer of liquid, differing from the 

 surrounding liquid, is distinctly visible, and a very similar 

 result is obtained from a platinum wire strongly heated by 

 the passage of an electric current. There is no doubt that 

 liquids holding small particles in suspension are capable of 

 giving rise to appearances similar to those described in gases, 

 and it seems that in a liquid an increase of temperature 

 occasions a less rapid increase in the thickness of the coat 

 than is the case with a gas ; or, in other words, the coat is 

 thinner at high temperatures such as 100° C. than it is at a 

 lower because the convection-currents have more power to 

 destroy it. 



Examination of the Hypothetical Explanations already 

 suggested. 



At the commencement of the paper are given all the pro- 

 visional explanations which have been hitherto suggested as 

 accounting for the dark plane so far as we are aware. The 

 existence of the coat renders several of them improbable, but 

 they appear capable of direct disproof, thus : — 



Any centrifugal-force hypothesis is negatived by the points 

 of inflexion on the stream-lines of fig. 1, and more conclusively 

 by fig. 2, where the white space between the parallel curved 

 lines represents in section a hemicylinder of sheet-copper with 

 its black coats. The convection of the stream-lines is here 

 such as to whirl the particles into the coat rather than out of 

 it. The evaporation and combustion hypotheses are dis- 

 proved by the use of a non-volatile and incombustible smoke 

 like magnesic oxide; also by the phenomena being observable 

 in liquids. 



