170 Dr. Shaw and Mr. Garrett on 
12x 10-* ems. apart*. It might be supposed that the P.D. 
for the steady current between the surfaces, being small, will 
cause a spark to pass if the film separating them is so thin. 
But the P.D. is sometimes only 1 volt, and never more than 
6 volts. Until recently, the discharge distance for less than 
40 volts had not been measured; but one of us has been able 
to show that for 6 volts the distance is about 6 x 10—° cms., 
and for 1 volt is 1x10-%cem. These results show us what 
is the superior limit to the thickness of film existing when 
current passes through the contact. Suppose the P.D. 
between the contacts is 6 volts, then, as shown, the film 
between the contacts must be more than 6 x 10—* em. thick, 
and when discharge occurs the electric waves burst over this 
interval. 
(b) Now sunder the surfaces; clean anddry them. Bring 
them together gently as before: the current now passes readily. 
On producing a spark, it is found that coherence will either 
be slight or non-existent. Thus a high resistance of the con- 
tact is a necessary antecedent condition of coherence. We 
may conclude either (1) that the film acts as a flux assisting 
the fusion of the two surfaces, or (2) that the ionization of 
the film into positive and negative charges causes attraction of 
the surfaces, or (3) that there must be some gap between the 
surfaces; this gap provides a large resistance so that enough 
heat is developed by the passage of the waves to fuse the 
surfaces. The last theory can be proved experimentally to 
be correct, for it is possible after cleaning the surfaces to bring 
them together with such slight pressure that the current 
passing is small. On passing a spark near the coherer, 
coherence readily occurs, as shown below. 
TABLE I. 
| | Force in dynes to Voltmeter 
| sunder surfaces, F. | Reading, V. | 
| Bis eee as fy ee ae a 
| Before SPOT Bah awn | ‘Dd dyne 0°25 
After bh Sashnoioe Pei bb sas? a 0°10 
Before spark........ IO <yash 0:30 
After et ee PUD is 0°10 
| Before spark......... / erat fe, | 0°30 
| After as) ee eceed Lh eibelDaxiss 2 <,, 0:12 
Thus coherence occurs even when no film existed previously, 
provided there is a gap between the surfaces. In the former 
* Loe, cle, Pp, 24k; 
