Application of the Electric Arc to Photomicrography. 171 
With regard to the effect of inconstancy of the length of the arc, 
our results showed that so long as the carbons were maintained at a 
C3nstant critical relative distance the intensely incandescent crater, 
from which practically the whole of the useful light is derived, 
remained absolutely motionless and of constant intensity, affording 
from a photomicrographic point of view an almost perfect illuminant. 
So soon, however, as the slightest lengthening of the arc occurred, 
the incandescent crater is lost, owing to the arc at once travelling 
round the positive carbon, playing successively upon different parts, 
and never remaining long enough in one place to allow the high 
incandescence and constancy of position secured when running under 
Fig. 24 . Fig. 25 . 
the critical condition we have described. In order to more directly 
test the effect of the small variations we had observed, two photo- 
graphs were taken : one with the crater of the positive carbon truly 
central (fig. 24), and the other (fig. 2’5) with the crater decentred by 
an amount, indicated by projection, corresponding to the fluctuations 
of the automatic lamp. 
We were thus led to realise the importance of higher accuracy in 
the adjustment of the position and length of the arc ; and to consider 
means whereby it could be secured. The necessarily intermittent 
action of any automatic mechanism, the action of which can only be 
started as the result of the very factor it was our object to avoid, 
viz. lengthening of the arc, forced us to the conclusion that such 
