Good morning,

    Yesterday afternoon we ( Bill Strelo, Derek Shuman, myself ) did a
breakdown test in water of the wire samples that we did the first
voltage tests on ref. ( email Jan. 28, 2002). the test was done in the
same beaker of water as the earlier test. A conductivity meter has not
been used on the water as of this time. Looking into that still.
The finds were great ;    sample: 
 #1  broke down	 @ 5kv
  #2      "		@ 7kv
  #3      "		@ 3kv
  #4      "		@ 11kv

The breakdowns do not appear to be directly related to the over
stretching of the wire or the bending of the wire. The breakdown points
are random in nature, but could be related to a combination of events.

Also noted by D. Shuman:

Three of the four samples broke under water on the outside of the bend
 (where the coating is in tension)
approx 30-45 degrees into the bend from where it deviates from straight. 
The fourth specimen broke on the straight section, again, well below the waterline, 
on the one sample that had been previously stretched to breaking point
 in the tensile test machine. 
	Thus the entire coating was under a residual tensile stress, and the bend
 (which was formed around the tensile test pulley) may not have had a substantially different 
residual tensile stress than the straight section; indeed, due to friction from the pulley,
 it may well have had less. 
	The elongation of the coating from bending around the 1" radius pulley is 9%, 
whereas the elongation from stretching , in the straight section, was over 30%.  
One would expect to see the greatest coating elongation in a section of the bend
 very close to the start of the straight section.

	There was one instance of (liquid) surface flashover, at 10kV on specimen #4, 
about a second before it broke at 11 kV, under water. This indicates the liquid line is not 
a serious electrical stress concentrator. No corona could be heard or seen 
(in the normal room light) at the liquid line. 

This is mystifying; I'll bet the reason is quite interesting, electrochemically.