SECT. XL. 4. OCULAR SPECTRA. 61; 



2. Look through a dark tube, about half a yard 

 long, at the area of a yellow circle of half an inch 

 diameter, lying upon a blue area of double that di- 

 ameter, for half a minute ; and on clofing your eyes 

 the colours of the fpectrum will appear fimilar to 

 the two areas, as in fig. 3. ; but if the eye is kept 

 too long upon them, the colours of the fpectrum 

 will be the reverfe of thofe upon the paper, that 

 is, the internal circle will become blue, and the ex- 

 ternal area yellow ; hence fome attention is required 

 in making this experiment. 



3. Place the bright flame of a fpermaceti candle 

 before a black object in the night ; look fteadilyat it 

 for a fhort time, till it is obferved to become fome- 

 what paler ; and on clofing the eyes, and covering 

 them carefully, but not fo as to comprefs them, the 

 image of the blazing candle will continue diftindtly 

 to be vifible. 



4. Look fteadily, for a fhort time, at a window 

 in a black day, as in Exp. 2. Se&.III. and then clofing 

 your eyes, and covering them with your hands, an 

 exaft delineation of the window remains for fome 

 time vifible in the eye. This experiment requires a 

 little practice to make it fucceed well ; lince, if the 

 eyes are fatigued by looking too long on the win- 

 dow, or the day be too bright, the luminous parts 

 of the window will appear dark in the fpectrum, and 

 the dark parts of the frame-work will appear lumin- 

 ous, as in Exp. 2. Sect. III. And it is even diffi- 

 cult for many, who firft try this experiment, to 

 perceive the fpectrum at all ; for any hurry of mind,, 

 or even too great attention to the fpectrum itfelf, 

 will difappoint them, till they have had a little ex- 

 perience in attending to fuch finall fenfations. 



The fpectra defcribed in this fedtion, termed direct 

 ocular fpectra, are produced without much fatigue 

 of the eye ; the irritation of the luminous object be- 

 ing foon withdrawn, or its quantity of light being 



VOL. I, S s not 



