186 
DR. W. .T. RUSSELL OX THE FORMATIOX OF 
II tlie plate Fe tiiaiigular in fuim, tlien three rays are I’orined, again each starting 
troin an angle (fig. 2), and il the plate Ije an octagon, then a star with eight rays is 
produced (fig. 3). An angle in a plate always tends to give rise to a ray. This is 
often very fine at the ])ofnt, and tliickens consideraldy afterwards. If a flat circular 
plate is used, then no deposit takes place, hut if it is concave, a uniform dept»sit over 
tlie whole of it occurs, and if it )je convex, then little or no deposit is formed, if anv, 
it is in the form of a star. When an oblong rectangular plate is used, then the rays 
similar to those formed on a scpiare plate are produced, Imt they do not meet, but 
are, as it were, drawn asmuler, and remain at each end of the plates, being however 
often connected by a. thin .straiglit line (fig. 4). 
In all cases, tlie angles of the jdate determine the figure formed on it. With 
regard tf) othe)’ general ])oints connected with the formation of these figures. The 
nature of the dust used is not a matter of impoi'tance, it may be composed of organic 
or inorganic matter; the spores of a fungus, or magnesia, or the dust from ashes 
