be short lived because moistening often becomes sufficiently large to swamp this
differential drying process.
Generally, head cracks are more disruptive than radial ones. Should an insulator
string become so severely polluted that a flashover occurs, the flow of such a large
current through a crack in the head of the porcelain shell will cause a very large rise
in pressure. Then the cap usually splits. Occasionally it does not; instead, the whole
contents of the insulator are ejected like shot from a canon. In this situation, it seems
very likely that the current path was essentially along the axis of the insulator. In both
cases, the insulator string breaks. In contrast, an arc within a radial crack in the shell
outside the cap can very rapidly escape into free air.
It would thus seem wise that, from time to time, sample units – especially those
from the high voltage end of the string – should be removed and subjected to a high
voltage test, particularly for those lines in highly polluted areas.