[turnbullI a subjective PHENOMENON OF VISION IS 



11. The image is now, so to speak, "on call"; that is, the writer 

 can at any time of the day or night see this hazy image; and the oscill- 

 ations, so far as he has observed, never cease. 



12. Only, however, when certain conditions are complied with is 

 the image seen when light is entering the eyes; for instance the image 

 may be usually seen when the eyes are "fixed" upon some non-luminous 

 and non-lustrous body, or upon the blue sky, etc. — i.e., when reflected 

 light is entering the eyes. It is usually not seen when a luminous 

 or lustrous body is fixed, although in this case suggestions of bright 

 points are sometimes seen. 



13. AVith opened eyes the image always appears at the point of 

 direct vision and seems to oscillate iipon the focussed object. Unless the 

 writer looks for the imago it usually remains unnoticed, but it sometimes 

 appears unbidden when dark or shadowed objects are seen. 



14. In complete darkness the image seems to be projected in the 

 centre of the visual field or, rather, what would be the fixation point. 



15. The oscillations can generally be counted and with more or 

 less accuracy according as the image is more or less distinct. It must 

 be borne in mind that the more carefully and continuously such phen- 

 omena as this are studied, the more one is able to see distinctly and 

 recall subjective images once seen, and it is only by long practice that 

 the best results are obtained. 



Apparatus and Methods of Observation. 



16. In order to obtain the periodicity of the oscillations, a stop 

 ^^■atch, reading to tenths of a second, was at first employed and a chrono- 

 graph reading to hundredths of a second was used in the latter part of 

 the experiments. 



17. It was found that the ticking of the watch or any other sound 

 having a distinct tempo was very distracting, when counting the oscilla- 

 tions, so the watch was set up in a room other than that in which the ob- 

 sen^ations were made. 



18. The watch was opei-^ted hj an electro-magnetic gravit}^ drop 

 and the readings were taken by an assistant. The electro-magnet was 

 operated by a key, at the observer's right hand, which broke a l^attery cir- 

 cuit, "when pressed, allowing an iron rod pivoted at one end to fall on a 

 lever operating the spring stop of the watch. The rod was at once 

 replaced by the assistant, the current having been l^roken but an instant. 

 The sources of error of this instrument were the inertia of the watch's 

 balance wheel and the variable residual magnetism of the electro- 



