164 MR JAMES DEWAR AND DR M‘KENDRICK ON THE 
that could be employed, yielding perfectly accurate results, would be to photo- 
graph on a sensitive surface, covering a cylinder rapidly revolving on a hori- 
zontal axis, the alteration of position of the spot of light reflected from the 
mirror, just as continuous magnetic observations are registered. As the appa- 
ratus required to execute these observations is very complicated, and will only 
become necessary in the second part of this memoir, where we must deal with 
the time occupied in the transmission of retinal impressions due to different 
coloured rays, we have, in the meantime, adopted two simple methods of regis- 
tration. The first plan is to note the position of the galvanometer at equal 
intervals of time before, during, and after the impact of light on the eye. In 
these observations we have used a seconds pendulum giving a loud beat. One 
observer reads aloud the galvanometer, the other marks every interval of two 
seconds, registers the numbers obtained, and regulates the supply of light. A 
little practice in the method above described has enabled us to obtain very 
satisfactory results, agreeing very closely in different observations, and showing 
in a decided way the salient points of the variation curve (see curves, Plate XI.). 
The second plan of registration which may be of service in many physical and 
physiological researches, consists in placing at proper distances from the gal- 
vanometer, instead of the ordinary graduated scale, the surface of a cylinder 
covered with paper, and moving on a horizontal axis by clockwork. The spot. 
of light reflected from the galvanometer mirror is rendered more precise by 
having the shade of the galvanometer lamp blackened over the entire surface, 
with the exception of a spot about three millimetres in breadth, in the centre of 
which a line or cross is made of soot. The image of this line or cross is, of 
course, reflected by the mirror upon the cylinder. When the cylinder is set in 
motion by the clockwork, the spot of light may be accurately followed by the 
hand of the observer after a little practice, with a fine brush moistened with 
ink. The cylinder we employed performed a complete revolution in eighty 
seconds. This time was divided into four equal parts, each representing twenty ~ 
seconds, by four lines drawn transversely at equal intervals across the paper on 
the cylinder. The first space, between lines one and two, represented twenty 
seconds, in which the eye was in the dark, and in which the electro-motive force 
is represented by a straight line ; the second space, between lines two and 
three, represented twenty seconds, during which the effect of the impact of 
light took place, and in which the variation of the electro-motive force is indi- 
cated, either by a curve to the right or to the left ; the third space, between 
lines three and four, represented twenty seconds of continued action of light, 
during which the electro-motive force gradually rises; and lastly, the fourth 
space between lines four and one (the point of starting), represented twenty 
seconds, during which the electro-motive force at first rises on the withdrawal 
of light, and afterwards sinks rapidly. Illustrations of curves registered by 

