IOWA ACADEMY OP SCIENCE 
275 
Obviously the conclusion just arrived at is inconsistent with Prof. 
Wood’s experiment where he obtained a current of several milli-amperes 
when parallel planes were only five wave lengths apart. I therefore 
thought it advisable to repeat the experiment using smaller quartz fibres, 
this time three in number. The sizes were, first, .00052cm : second, .00053 ; 
third, .00059 and the mean probable distance of separation was there- 
fore .00056cm. The electrical method by the mean of five observations, 
with 0.3 volt across condenser, gave a separating distance of .00054cm. 
There are two conclusions that should be drawn from this experiment. 
First there was almost perfect insulation when the two planes were 
separated by only ten wave lengths and second the conducting atmos- 
phere could not have extended as much as two wave lengths outside the 
mechanical surface. 
A further attempt to repeat the above experiments with yet finer 
quartz fibres met with failure owing to unsatisfactory insulation. I did 
not try heating the quartz fibres at this time because I feared that it was 
steel surfaces might vary by ten wave lengths. I observed that it was 
impossible by the electrical method to obtain insulation at distances less 
than those measured by the microscope. Unless some unknown factor 
is involved this observation obviously leads to a third conclusion, which 
is that conduction at short distances arises from conducting material 
and not from an electron atmosphere of unvarying depth. 
The question as to the imperfect insulation at less than 9 wave lengths 
was still unanswered. Possibly the insulating properties of certain 
dielectrics break down when a certain minimum thinness is reached. It 
was thought that a constant minimum thinness for any given material 
would strengthen this view. Or perhaps as Wood suggested moisture 
films about small particles might readily explain the imperfect insula- 
tion if not also the high conduction. We shall recur to this point in 
a succeeding paragraph. But the most opportune explanation, whose 
plausibility should be considered, was the possible irregularity of the 
steel surfaces. 
The question of the planeness was answered in two ways. First a 
glass plate with plane sides was moved over the surface of the plates, 
and the character of the interference fringes noted. By this test the 
extreme variations of the surface was not as much as two wave-lengths. 
Second two optical flats on speculum metal were obtained, whose sur- 
faces by optional test did not vary from a plane as much as one wave 
length. With these my surprise was certainly great when I found im- 
perfect insulation when the surfaces were separated by quartz fibres of 
24 wave lengths diameter. As my supply of quartz fibres, and also of 
