Radiation in Incandescence Lamps. 49 
good contact. If this precaution is omitted, a loose contact 
may be formed, the result of which will be a generation of 
heat at that point. 
In the ordinary working the life-history of a carbon fila- 
ment is something as follows: — 
At some point or other the filament is probably thinner 
than at other places. At this place there will be a greater 
generation of heat and a higher temperature; volatilization of 
the carbon ensues, and the vapour condenses on the sides of 
the glass bulb, as far as I have observed, uniformly. If, how- 
ever, the point of greatest resistance occurs on the copper 
clamp, then it is found that copper volatilizes and deposits on 
the inside of the glass. 
But what is most curious is, that in this case an examina- 
tion of the glass envelope shows that there is a narrow line 
along which no copper has been deposited. This is seen best 
by holding the lamp up before the light and slowly turning it 
round. In one particular position, easily found, it is best 
seen. Now, on examining carefully the position of the line 
of no deposit as compared with the position of the carbon fila- 
ment, it will be seen that it lies in the plane of the loop, and 
on the opposite side to that nearest to which the break of the 
loop has occurred. It is in fact a shadow of the loop. 
The conclusion which must be arrived at, then, is that the 
copper molecules are shot off in straight lines; otherwise it is 
impossible that there should be this line of no deposit. 
The most noticeable thing is, that it occurs only when the 
deposition of copper takes place; I have never noticed it in 
an ordinary carbon deposit. 
Hence there must be some essential difference between the 
vaporization of the carbon and that of the copper. The car- 
bon deposit resembles more the condensation of a vapour and 
is uniformly distributed ; but the copper deposit exhibits the 
character of a molecular radiation or shower taking place from 
a certain point. 
The whole phenomenon calls at once to mind the beautiful 
researches of Mr. Crookes with vacuum-tubes. Here, how- 
ever, we are dealing not with an induction-coil discharge, but 
with a comparatively low potential. 
I have never failed to see the effect in any lamp which has 
had a deposition of copper on its interior. 
It is interesting to note how nearly the colour of transpa- 
rent copper resembles that of transparent gold. The simi- 
larity of the surface-colour of pure unoxidized copper and of 
gold is accompanied by a near resemblance in colour of the 
two metals in thin films. 
Phil, Mag. S. 5. Vol. 16. No. 97. July 1883. E 
