( 22 ) 
pressure to which this air is subjected becomes, we must believe 
that, no matter how suddenly the form is applied, the air must 
undergo some •compression, yet we find that the indentations in 
the iron are often nearly as deep as those in the gun-cotton. 
In considering von Forster’s hypothesis, we are willing to ad¬ 
mit that the gases at the time of detonation possess the exact 
form and occupy the same space as the gun-cotton from which 
they are formed provided the change takes place instantaneously. 
But it does not; in fact it occupies so appreciable a period of 
time, that the rate of propagation of the detonation in it has been 
measured. Apart from this and even granting it, it will be ob¬ 
served that von Forster does not explain how the impression is to 
be produced by the gas. If the gas moves as a coherent mass, 
then the impressions should be the reverse of what we get. 
From my recent experiments, I am the more strongly convinced 
that the effect is a purely ballistic one and that while the base of 
the gun-cotton, or its products, are projected as a whole against 
the plate, where the intervening spaces are the greatest, there we 
have the greatest indentation. This is true, also, in the leaf ex¬ 
periment which has been exquisitely reproduced. The varying 
thicknesses of the leaf vary the distances through which the ma¬ 
terial is projected and hence the form and texture is reproduced 
in the impression. 
I take pleasure in exhibiting the specimens described and 
also two others which I have produced to-day. The first was 
made by placing a gun-cotton disk in a tin can, 5 inches in di¬ 
ameter, and just covering the disk with water. The smooth 
bottom of the can rested on an iron beam. After detonation, the 
beam was found ruptured and its face indented as shown in -Fig. 
1 where F represents the iron beam, B the original position of 
the gun-cotton, A B' the base of the can, B'B' the indentation 
corresponding to the base of the gun-cotton, B C a smooth space 
beyond, C D a space covered with ripple marks visible under a 
microscope and D A a space covered with ripple marks easily 
seen by the naked eye. Fig. 2 is a photostereotype of the face 
of this iron beam. 
The second specimen was produced by placing a gun-cotton 
disk in a tin can just large enough to admit it, and covering the 
