228 Proceedings of the Royal Society of Edinburgh. [Sess. 
rent, through which at any point a thin card could be passed — but fortu- 
nately the damage was confined to that. Lehmann has observed the same 
phenomenon when working on discharges through gases in large vessels : * 
in his case, however, working with an “ egg ” 60 cm. long and 45 cm. broad, 
whose two parts had been separately tested beforehand to make sure of 
their strength, “ after it had stood untouched for about five hours it was 
suddenly shattered, with a loud report, into countless little pieces. These 
were driven as far as 15 metres off, some of them going through the window- 
panes. All breakables in the neighbourhood were damaged, and some 
splinters of glass even forced themselves so violently into the wooden doors 
of cupboards and other wooden fittings that they stuck there.” 
The danger of breakage is, however, not entirely confined to large vessels. 
After the bell-jar had broken, it was decided to abandon the idea of getting 
the special spark-gap bodily into the vessel. A cylindrical glass tube, 
14 cm. in length and 5 cm. inside diameter, was taken instead. The ends 
were open, but could be closed by two plane slabs of glass, bored to allow 
stout pieces of copper wire (which were to do duty as the electrodes) to 
enter, and the whole was rendered air-tight with sealing-wax. This stood 
successfully the first and second discharges, but on the third the lower end 
of the tube was completely shattered. 
In such a series of experiments as described above, the vagaries common 
to most*electrostatic experiments performed in the free air of a laboratory 
are of course prominent, especially during adverse atmospheric conditions ; 
and results on occasion display such consistent inconsistency as to baffle 
* 0. Lehmann, Annalen d. Physik (4), Bd. vii., 1902, p. 1. 
