I^EFINITE EIGUEES BY THE DEPOSITION OF DUST. 
557 
greater thickness of clnstless air on each side than opposite the olrstruction. As the 
light travelled backwards and forwards the lower limit of the dust always seemed to 
dip just when the light was opposite the obstruction. As the plate cools, dust begins 
to fall behind the obstruction, while as yet there is a space of dustless air on each side 
of it. While it may be strange that so small an obstruction as a liair should produce 
these deposits, yet it is known that when stream-lines are interfered with, unexpected 
results frequently happen. 
Figs. 30, 31 and 32 call for no special remarks, as these cases are explained in the 
previous paragraph, the deposits being due to the rough edges of the plates interfering 
with the stream-lines. 
Fig. 33 is the same as fig. 23, already explained, only in the former the obstruction 
is placed on one of the diagonals, and not on the side of the square, as in fig. 23. 
It seems unnecessary to consider in detail the other'figures in Dr. Russell’s paper, 
which show the effects of different kinds of obstructions placed on or above the hot 
surface. The manner in which these figures are formed can be easily understood 
with the aid of what has been said. Only a few remarks may be made as to 
figs. 44 and 44a. When we examine by means of a beam of light the conditions 
in these two cases, it is seen that the dust-free film from underneath the plate rises 
and flows upwards past the edge of the top plate. In doing so, it seems to draw 
away with it some of the air from between the plates, witli the result that a very 
slight negative pressure is established in the space between them. The beam of light 
shows that the upward current does not move in a straight line Imt curves inwards, 
drawn in by the lower pressure between the plates. The amount of tliis in-curving is 
least at the middle of the sides of the plate and greatest at the corners, the reason 
for this being that at tlie middle of the sides the air currents are stronger, as they 
contain more hot air than the currents at the corners, with the result that the weak 
currents at the corneis are drawn more out of tlieir course than the stronger currents 
at the sides. As the temperature falls the currents weaken, and at last the currents 
at the corners yield and the air is drawn in there and lirings its dust with it, but 
its velocity being small, only the edges get any protection from the dust-free film 
and the dust settles on the plate before it travels far, giving the patches of dust 
shown in the corners of the plates in figs. 44 and 44a. 
Wliat perhaps surprises one most in the formation of these dust figures is the 
important part the dustless film from the under side of the plate plays in protecting 
the upper surface from deposits of dust; and we have seen that whatever tends to 
destroy this dustless film tends to bring about conditions favourable for the dust 
settling on the plate. 
\_Note hij Dr. PtUSSELL, July 23, 1903. — The interesting observations made by 
Mr. Aitken by means of his exploring beam of light have contributed substantially 
to the elucidation of the dust-figures. It was, of course, clear from the first that the 
