700 PHILOSOPHICAL TKANSACTIONS. [ANNO 1785. 



inside of the cistern through a valve c in the brass cap e, which producing a 

 perfect vacuum in the gage b, formed an instrument peculiarly well adapted for 

 experiments of this kind. Things being thus adjusted (a small wire p having 

 been previously fixed on the inside of the cistern to form a communication be- 

 tween the brass cap e and the mercury g, into which the gage was inverted) the 

 coated end a was applied to the conductor of an electrical machine, and notwith- 

 standing every effort, neither the smallest ray of light, nor the slightest charge, 

 could ever be procured in this exhausted gage. I need not observe, that if the 

 vacuum on its inside had been a conductor of electricity, the latter at least must 

 have taken place, for it is well known (and I have myself often made the experi- 

 ment) that if a glass tube be exhausted by an air-pump, and coated on the out- 

 side, both light and a charge may very readily be procured. If the mercury in 

 the gage be imperfectly boiled, the experiment will not succeed ; but the colour 

 of the electric light, which in air rarefied by an exhauster is always violet or 

 purple, appears in this case of a beautiful green, and, what is very curious, the 

 degree of the air's rarefaction may be nearly determined by this means ; for I 

 have known instances, during the course of these experiments, where a small 

 particle of air having found its way into the tube b, the electric light became 

 visible, and as usual of a green colour ; but the charge being often repeated, 

 the gage has at length cracked at its sealed end, and in consequence the external 

 air, by being admitted into the inside, has gradually produced a change in the 

 electric light from green to blue, from blue to indigo, and so on to violet and 

 purple, till the medium has at length become so dense as no longer to be a con- 

 ductor of electricity. I think there can be little doubt from the above experi- 

 ments of the non-conducting power of a perfect vacuum ; and this fact is still 

 more strongly confirmed by the phenomena which appear on the admission of 

 a very minute particle of air into the inside of the gage. In this case the whole 

 becomes immediately luminous on the slightest application of electricity, and a 

 charge takes place, which continues to grow more and more powerful in propor- 

 tion as fresh air is admitted, till the density of the conducting medium arrives at 

 its maximum, which it always does when the colour of the electric light is indigo 

 or violet. Under these circumstances the charge may be so far increased as fre- 

 quently to break the glass. In some tubes, which have not been completely 

 boiled, I have observed, that they will not conduct the electric fluid when the 

 mercury is fallen very low in them ; yet on letting in air into the cistern h, so 

 that the mercury shall rise in the gage b, the electric fluid, which was before 

 latent in the inside, shall now become visible, and as the mercury continues to 

 rise, and of consequence the medium is rendered less rare, the light shall be- 

 come more and more visible, and the gage shall at last be charged, though it 

 has not been near an electrical machine for 1 or 3 days. This seems to prove, 



