DEFINITE FIGUEES BY THE DEPOSITION OF DUST. 553 
Fig. 1 represents the appearance of the plate at this stage. As the temperature of 
the plate falls, the rate of flow slackens and the repulsive action of the heat grows 
less, and at last a certain stage is reached when the current has become so slow that 
the under side of the dusty air comes close to the plate over the area where the 
Fig. I 
Fg 2 
currents meet, and as these currents here turn sharply upwards, there is a small space 
under their meeting point where the air is still and into which the dust collects and 
is seen showering down on the plate, as shown in fig. 2. 
As these air currents rising from beneath all flow round the edges of the plate and 
move horizontally in a direction at right angles to the edges, they thus meet over the 
diagonals of the square; hence the deposition of the dust in Dr. Bussell’s figures on 
the diagonal lines in his fig. 1 and on the lines bisecting the angles of the triangle in 
fig. 2 and the octagon fig. 3."^' In fig. 4, for evident reasons, the currents here meet 
about the same angle as in fig. I, and deposits take place on the lines Ifisecting tiie 
angles, but the stronger currents provided by the longer sides of the oblong plate pre¬ 
vent much deposit taking place where they meet in the centre of the plate. Further, 
m all these four cases, the currents meeting over the diagonals do not turn directly 
upwards, but flow also towards tlie centre of the plate, taking more or less of a horizontal 
moveinent along the diagonals, so tending to give time for the dust to fall. Further, 
when the currents have met and their direction has become partly vertical and partly 
horizontal, the repelling effect of the hot surface nearly ceases, as the temperature 
gradient perpendicular to the hot surface is in the rising current practically nil. The 
reason for the narrowing of the deposits as they approach the centre of the plate would 
appear to be due to the greater velocity of the currents where they meet over the 
centre of the plates caused by the union of all the currents from the different sides. 
* Eefereuce must be made to Dr. Russell’s paper for these dust figures, as they are not reproduced 
here. 
4 B 
VOL. CCI.—A, 
