0. N. Rood on the discharge of a Leyden jar. 161 
visible : it was followed by a faint flame-shaped tail, corres- 
ponding to the yellowish portion of the discharge. 
Having thus proved the existence of an isolated discharge, 
at once the first and most brilliant act in the explosion, a 
method was contrived for measuring its duration, or at a 
events for setting a limit to it on at least one side. Ordinary 
micrometric methods, conducted on the naked spark or on a 
strip of paper illuminated by it, were out of the question, and 
would have served no purpose except to lead the observer into 
error, so I contrived the following plan, which is moderately 
easy of application and safe. 
small piece of cardboard was ruled with two black lines ; 
their distance apart was ‘0067 inches, and they were separated 
by a space exactly equal to the width of a single line. A small 
dividing engine was used for their production ; they were tested 
with a microscope. These lines were illuminated by the spark, 
and their image examined by an eye-piece magnifying five di- 
ameters. ‘The breadth of the image of a single line on the 
ground glass was ‘011 inches, that is, each line subtended an 
angle of 2’ 24”, reckoning from the mirror. With a velocity 
of 223 per second the mirror is able to move the reflected image 
through 2’ 24” in the 00000024 of a second. If now this first 
discharge had actually lasted this minute portion of time, it is 
evident that the motion of the mirror would just have carri 
the image of one of the black lines forward, so that at the end 
of this infinitesimal period, it would have occupied the space 
where just before the white line had been traced : hence owing 
to the retention of impressions by the retina the white central 
line would have been obliterated, and in place of three lines, 
& gray band would have been seen. On the other hand if the 
duration had been only ;; or } of the above mentioned inter- 
val, the white line not having been much encroached on, would 
still have remained visible. The correctness of the above rea- 
