156 Trans. Acad. Sci. of St. Lows. 
The strength of the field just outside of the flame, at 
the point occupied by the needle, was in every case in- 
creased by the flame. The commotion of the ionized 
molecules within the flame, decreased the permeability 
of the space occupied by the flame. The lines of force 
were deflected around this vertical column. The instant 
when the combustion began is represented by the arrow 
on the two diagrams. The effect upon the needle did not 
continue in any case for very much more than one min- 
ute, and it was generally somewhat less. 
The explosions of dynamite in open air produced a 
disturbance over a very large area. In some cases as 
many as 22 sticks of 40 per cent dynamite were laid upon 
a long bar of wood, having a cross section of 1x1.5 inches. 
Each stick was securely bound to the bar at each end, by 
means of heavy binding twine. The sticks were placed 
end on end and the explosive cap was placed in the stick 
at the end of the column. The bar was mounted upon 
supporting posts at a height of eight ft. above the ground. 
No trace of the bar of wood could be found after the ex- 
plosion, except the shattered ends projecting slightly be- 
yond the column of dynamite. The marsh grass and iris 
which was about 2 ft. in height was closely shaved from 
the surface of the ground, over a distance of about 30 ft. 
on either side of the column. It was all deposited in a 
long heap directly below the line of the column. Byi- 
dently the rebound of the compression wave of air and 
gas from the ground, had caused an inflow which swept 
the grass inward and upward. No disturbance of the 
tent, 275 ft. distant, could be detected. A screen of boards 
was erected in front of the end of the table to protect the 
needle, but fine fibers of silk suspended within the tent 
showed no appreciable deflection. A window glass in 
our cottage, having dimensions 29x30 inches, was broken 
by an explosion of 25 sticks. The cottage was distant 225 
ft. This does not necessarily mean that the breakage 
was due wholly to the explosion. A glass which replaced 
it also broke down a few weeks after being put in place, 
