of Edinburgh, Session 1881-82. 
571 
electrification, the non-lnminous flame is not attracted powerfully, 
and it retains its original size and character, while the luminous one 
is thrown down into and over the mouth of the tube, and retains 
its luminosity. If there are convection currents they must be such 
as to throw oxygen into the flame in the former case, hut not in the 
latter. 
The observations recorded in Table II. were undertaken to test 
whether the difference of potential required for the spark depended 
largely upon the height of the flame. Where the conditions are in 
common with cases in the former series, the observations serve as a 
test of consistency ; the agreement is satisfactory. Diagram 2, 
representing the results, in addition to confirming the conclusions 
drawn from the preceding diagram, shows that the variation of the 
height of the flame produces little change in the electromotive 
force, when the flame is non-luminous ; also when the flame is 
luminous and the electrification positive, hut that it produces a great 
difference in the remaining case, when the flame is luminous and 
the electrification negative, so long as the flame is not sufficiently 
high to reach the disc. Though the flame is attracted so as to he 
in contact with the disc, the discharge still passes in the form of a 
spark. 
This investigation was made in one of the fine suite of rooms 
recently added to the Physical Laboratory, and in conducting it 
we had every facility given us by Professor Tait. 
Table I. 
