

1905. | On Colour- Vision by Very Weak Light. 211 
for every fresh note, an indiarubber band round those written on preventing 
the same page from being used twice. - An account of each experiment was 
written in the laboratory book, while the exact significance of the notes was 
fresh in my mind. | 
Time was recorded as follows :—A shaft carrying a grooved wheel 10 cm. 
in circumference was connected with the arbor of the hour-hand of a common 
Ansonia clock, so as to revolve with it. On a thread wound round this 
wheel, and attached to it, hung a small weight, to the bottom of which was 
soldered a toothed wheel of rather larger diameter than the weight. 
The whole was mounted on the top of an upright case about 18 inches 
high, inside which the weight hung. Before making an experiment, a strip 
of white blotting-paper was secured to the back of the case with drawing- 
pins, the weight wound up to the top by turning the hour-hand backwards, 
and a mark made on the blotting-paper by pressing the toothed edge of the 
weight against it. This mark was entered in the note-book as the zero of 
time. 
While in the dark room, before making each note, the weight was pressed 
against the blotting-paper so as to mark it. Immediately after the conclusion 
of the experiment, the weight was wound up to each mark in succession, and 
the corresponding time read off on the clock face. There is no difficulty in 
reading to 30 seconds, and there is ample space for a 3 hours’ record. Except 
in a vague way, the observer does not know how the time is going on. If it 
were desirable to keep him in entire ignorance of it, an electrical time-marker 
might be used to record on smoked paper. I used this clock during each of 
the above-described experiments, and during many others made solely for 
the purpose of investigating the time-relations of these subjective phenomena. 
The phenomena succeed one another in a well-marked order, but at. no 
fixed time. As in the case of the duration of artificial colour-blindness,* and 
also of the rate at which flickering ceases for the various colours in experi- 
ments by intermittent light,f not only the condition of the eye as regards 
previous fatigue, but also the state of the health seems to exert a considerable 
influence on the time required for the eye to become completely rested. 
There are, however, very noticeable time-relations between the various stages 
of the process which preserve a certain proportionality among themselves. 
Actual after-images—the ocular spectra of Newton—do not as a rule last 
more than 10 minutes or a quarter of an hour, unless they have been 
excited by looking for some time at a strong light, in which case they may 
* ¢Phil. Trans.,’ B, vol. 191, 1899, p. 1. 
t ‘ Journal of Physiology,’ vol. 21, 1897, p. 426. 
