1877.] 



Phenomena connected with Vision. 



489 



perimeuts are susceptible, on the afternoon of November 29 I made 

 24 consecutive experiments in two sets. In each case an assistant 

 moved one of the candles until I told him to stop ; he then marked the 

 table with a piece of chalk. After each three experiments I took the 

 measurements ; these only differed by about 2 to 4 centimetres when 

 the furthest candle was from 1 to 3 metres distant. I calculated the re- 

 sults after all the experiments were finished, and found, in the first 

 case, the shadow had the value of 



86 86^ _85 8£ _84 8£ 

 100' 100' 100' 100' 100' 100* 



I was disappointed at this result, as I had had the paper ruled very 



75 



carefully to represent -jqq. On examining it afterwards with the micro- 

 scope I found the ruling had failed to give the proportion it should, 

 as the lines were too narrow, and the engraver had removed the 

 centre of each line with a diamond, so that, as accurately as I could 



85 



measure it, it represented jqq, the mean of my measurements. This 



measurement is far the largest source of error. 



In the second set I obtained ^ and ^ as my results ; and again I 



found the ruling, which should have been | black, was only, as nearly 



as I could measure, § black, or and in some parts still less. I think, 



therefore, I may say the error in different cases, eliminating the difficulty 

 of estimating the ruling, is only about 5 per cent, or less, probably not 

 more than 2 when a mean is taken. With regard to the ruliug its 

 estimation is more difficult, and I have hitherto found it impossible to 

 get it done accurately. The ruling-machine does its work well enough, 

 but the printing always spoils it. I think, however, the accuracy is 

 sufficient to establish my law. 



In my earlier experiments I assumed that the number of retinal 

 elements stimulated by any given surface vary as the amount of the 

 surface left uncovered by the black lines ruled upon it. Of course this 

 is only the case when an accurate picture is made upon the retina of 

 such a size that the lines and spaces fall on physiologically distinct 

 elements ; the lines need not be mentally distinguished, but are, I believe, 

 always capable of being distinguished as lines by a mental act. 



Professor Stokes first pointed out to me the necessity of proving that 

 in my experiments such a picture is actually formed, and of investigating 

 the effect when the lines no longer produce a perfect picture, but be- 

 come diffused so as to give what is physiologically equivalent to a 

 shadow. He pointed out that if I were right, such a surface should be 

 fainter in shade — that is, it should appear brighter than a ruled surface. 



I found this a by no means easy question to settle ; but I easily con- 

 vinced myself that a ruled surface seen slightly out of focus, or by an 



