554 
ME. JOHN AITKEN ON THE FOEMATION OF 
The importance of the dustless film from the under side of the plate is evidenced hv 
the fact observed by Dr. Russell that no figure is obtained unless the plate he 
supported above the bottom of the receiver. When we examine bv means of the 
beam of light the surface of a plate laid on the bottom of the receiver, we still find 
the dust-free film of air over the plate when the plate is pretty hot. This dustless 
film is veiy thin at the edges and thickens towards the centre, and a rising current 
can be seen fiowfing over it towards the centre, the rising current having a dustless 
core. This dustless film soon disappears as the temperature of the plate falls, and 
long befoie the jilate is cold dust falls all over it, but no definite figures are formed. 
Turning again to Dr. Russell’s figures, the effect of the velocity of the current is 
well shown m figs. 7 and 8. Fig. 7 was obtained Avith only a slight heating of the 
plate, and fig. 8 by a higher temjAerature. In the former figure the sIoav currents 
pioduced by the slight heating only kept the outside edges free from dust and 
allowed a large deposit to take place over the centre of the plate; while in the latter 
the highei temperature gave a current strong enough to prevent almost any deposit 
at the centre. 
The cause of the extension in the breadth of the arms of the cross in Dr. Russell’s 
fig. 9, Avhich AAms obtained by placing a hot cylinder some distance beloAv the plate, is 
not so evident, hut the probable explanation seems to be the folloAving; When the 
cuiients are due to the hot plate alone the circulation is mostly horizontal from 
centre to edge beloAv and from edge to centre above the plate, and the area AARere 
the cuiients meet is narroAv, but AARen there is a hot body under the plate there will 
be an upAvard current all round it of hot air. This upAvard current Avill preAmnt the 
horizontal movements above described being so markedly horizontal, and Avill cause 
them to turn upAAmrds at an easier curve, so broadening the dead dust-depositing area 
under the up-ciirAung air. 
In fig. 10 the extra deposit is jarohably due to some interference Avith the under 
air current produced by the piece of glass held under the plate. This subject will be 
referred to' later. 
Turning noAv to the effect on tlie figures of flames, &c., placed at a distance from 
the apparatus, as shoAvn in Dr. Russell’s figs. 11, 12, and 13. These alterations in 
the figures appear to he due not to any direct effect of the flames, &c., on the dust or 
on the plate, Imt to the heat radiated by them heating the receiver and so giving 
rise to convection currents at tlie side of the plate. These currents entirely change 
the symmetrical Aoav of tlie air over the plate and cause the centre of the current 
rising oyer it to move to one side, as shoAvn in figs. 11 and 12. When making 
observations Avith the apparatus described in this paper, it Avas not possible to get 
any of the figures quite regular; even the slight amount of heat given off by the 
small incandescent strip of mantle after passing through water interfered with the 
lesults, and air currents could be seen rising in the receiAmr on the side next the 
light. In support of this convection explanation, it may be further stated that the 
