260 
Mr Sears, On the Longitudinal Impact 
( b ) When the rods separate, a spark may be produced, so that 
the current does not fall instantaneously to zero. 
With the high resistance and small self-induction used, the 
latter effect was found to be negligible, while the correction for 
the former was quite small (2'5 x 10 _6 secs.). The method of 
determining these corrections, together with other details of the 
experimental work, is fully described at the end of the paper. 
The rods used were all in diameter, the radius of the 
spherical ends being 1". It was only possible to obtain con- 
cordant results when these ends were kept scrupulously clean, the 
presence of the least trace of grease or dirt producing great 
variations in the readings. The spherical surfaces were therefore 
washed, first with KHO, and then with water, and carefully dried. 
They were then polished with a clean piece of fine emery paper 
(No. 0000). This process was then repeated, and sometimes it 
was even found necessary to repeat it a third time before satis- 
factory readings could be obtained. The rods were suspended 
from two parallel wooden beams by means of “V”s of fine cord 
(fishing line), being thus compelled to swing always in the same 
plane, and with their axes always horizontal (Fig. 2). In their 
lowest position, the rods were collinear, and just in contact at their 
ends. The radius of the arc which they described was o'. Several 
preliminary experiments were made, first with three rods in line, 
