34 PROFESSOR SWAN ON THE GRADUAL PRODUCTION 
trical arrangements, which are fully described in my paper.* By such means I 
succeeded in measuring the brightness of luminous impressions caused by light 
acting on the eye for short intervals of time, varying from a tenth to a thousandth 
of a second. 
There are obvious limits to the applicability of this method of experiment- 
ing when itis attempted to examine isolated impressions of excessively short 
duration. To shorten the impression, we may either diminish the angle of the 
sector or increase the velocity of the disc, but whichever of these methods we 
adopt, we very speedily arrive at limits which either cannot be overpassed at all, 
or can only be so at the cost of great inconvenience. The velocity of the disc is 
absolutely limited. If we attempt to drive it faster than about 10 revolutions per 
second, the successive flashes become blended into a single impression, more or less 
uniform ; and as the eye no longer distinguishes them as separate, so we can no 
longer perceive, or measure, their separate intensities. Again, if we shorten the 
impressions by diminishing the angle of the sector, we quickly arrive at a practical 
limit from the smallness of the angle required, even with a disc of considerable 
diameter. This difficulty may be overcome in theory by increasing the diameter 
of the disc, while the sector is made of the smallest practicable angle ; but a very 
large disc would obviously be an extremely inconvenient instrument to work 
with. 
I have from time to time thought of various mechanical expedients, more or 
less feasible, for obviating this difficulty, but I have never been sufficiently satis- 
fied with any of them to attempt their construction. Last autumn, however, I 
devised an arrangement so simple that I resolved to put it in practice. Suppose 
a disc with a sector cut in it to revolve between the eye and a light 100 times per 
second. The eye would in every second receive 100 impressions, and if the sector 
were —_ th of the circumference of the disc, each impression would be made in 
1000 
aaa th of a second; but, from the high velocity of rotation, the observer would 
be conscious only of an uniform light,—the aggregate of all the nearly instantane- 
ous flashes blended into a single impression. The instrument which I have to 
describe has for its object to select a s¢ngle impression out of the hundred, to con- 
vey it to the eye isolated and alone, and to measure its brightness; while the 
remaining 99 impressions which would have interfered with its effect are inter- 
cepted, and excluded from the field of view. From the disc revolving 100 times 
in a second, the eye will then receive, once a second, a single isolated impression, 
lasting for the extremely short period of 2th of a second. This result admits 
of being obtained by the very simplest possible means; in fact, I have found it 
unnecessary to employ any mechanical agent whatever, other than the train of 
wheels by which motion is communicated to the disc. 

* Hdinburgh Transactions, vol. xvi., pp. 583-587, 
