Ixvi Proceedings of Royal Society of Edinburgh. 
notice such as the present to give any really adequate account of 
them, and for the best available information regarding them refer- 
ence may be made, once for all, to his Treatise on Lighthouse Con- 
struction and Illumination, of which the third edition was published 
in 1881. Yet the history of inventions has seldom failed to be 
interesting to a numerous class of readers ; and here it seems 
desirable at least to point, in chronological order, to Thomas 
Stevenson’s principal lighthouse improvements. 
Of these the earliest, made in 1849-50, was the invention of his 
Catadioptric Holophote. This consisted in the removal of the pos- 
terior portion of the parabolic reflector hitherto in use in light- 
houses, replacing it by a spherical reflector behind and concentric 
with the flame, so as to reflect the light incident on it back through 
the flame along with a lens in front, by which means he contrived 
that, while by far the greater portion of the original parabolic mir- 
ror was retained to fulfil its function of reflecting a beam of parallel 
rays, the very large portion of light which hitherto had wastefully 
escaped by natural divergence, was now also emitted in the same 
beam of parallel rays with that reflected by the parabolic surface. 
Thus, for the first time, the whole “ sphere of rays ” diverging from 
the flame was utilised by being combined into a single beam of 
parallel rays, with the least jDossible number of reflections or refrac- 
tions. The merit of this contrivance will best be appreciated by 
comparing it Avith previous attempts made to increase the efficiency 
of the parabolic lighthouse reflector. It is the first of various 
arrangements which Mr Stevenson devised' in order to intercept, 
and by the least possible number of optical agents to render parallel, 
all the rays proceeding from a focal point, and Avhich accordingly 
he termed “ holophotal.” 
In designing a lighthouse where a portion only of the horizon 
and not the whole all round AVas to be lit up, the catoptric system 
of illumination presented no difficulty. It was enough to fit up a 
system of reflectors embracing in their range the arc to be illuminated 
and no more. But it is otherAvise with the dioptric system. A 
portion only, instead of the entire Fresnel’s fixed-light apparatus, 
doubtless, can be constructed. But then those rays proceeding 
from its great central lamp, amounting it may be to more than one- 
half, which this partial apparatus fails to intercept, will be uselessly 
