TRANSACTIONS OF THE SECTIONS. 37 



Witli patterns on a liorizontal surface, such as a carpet, the results were verj' cmioua. 

 On comtDiniug; pairs of patterns with the optic axes crossed, I appeared to be stand- 

 ing in a hole with the level of the floor up to my waist, while on combining pairs 

 of patterns with the optic axes uncrossed, I was apparently standing on a pedestal 

 with the widely expanded floor far below me ; and so strong was the delusion, that 

 I could scarcely venture to move for fear of falling over. 



Colours I found could be fairly well combined by painting two patterns diflereut 

 colours and then causing the two to coalesce, with or without crossing the optic 

 axes. 



I have also succeeded in combining two solid bodies of the same size and shape, 

 but of different colour, both with the optic axes crossed and with them uncrossed. 

 Perhaps one of the most curious experiments I have made of this kind is to opti- 

 cally combine the heads of two persons, thereby producing' a combinational figure 

 of the two. 



On a Refraction-Spectrum without a Prism. 

 By Professor J. D. Everett, F.R.S.E. 



It was pointed out by WoUaston in the Philosophical Transactions for 1800, that 

 triple images can be obtained by looking at real objects through the stratum of 

 intermixture of two liquids of different refractive powers, one of which has been 

 gently poured on the top of the other. 



Having set up an arrangement of this kind last spring, in a cubical vessel (raea- 

 siu-ing 6 inches each way) with plate-glass sides, a strong solution of common salt 

 beiug the lower liquid and pure water the upper, I observed such decided colour- 

 effects, that the idea occurred to me of trying whether a spectrum could be obtained. 

 I accordingly placed the vessel of liquid on a high stool in the centre of a dark room, 

 and looked through the stratum of intermixture at a horizontal slit in the window- 

 shutter, which was about 10 feet distant, and was below the level of the said stra- 

 tum. The following were some of the phenomena observed, about a week having 

 elapsed since the liquids were placed in the vessel. When the eye was at any dis- 

 tance less tlian about 3 feet from the vessel, one image of the slit was seen. It was 

 highly coloured, forming a very impure spectrum, with blue above and red below. 

 Its apparent position was above the real slit, and at about the same distance from 

 the observer. 



When the eye was at a distance of 3^ feet or upwards from the vessel, three 

 images of the slit were visible. At some distances they could all be seen at once. 

 At other distances two could be seen at once, and the third came into A'iew on rais- 

 ing or lowering the eye. AU three of them were above the true direction of the 

 slit, and all were highly coloured. The highest and the lowest were virtual images, 

 and were almost precisely alike and similar to the single image above described. 

 They were erect images, and had accordingly blue above and red below. Between 

 them, when the eye was at a proper height, was seen another image with more 

 colour than either, and with the colours in inverted order, that is to say, with 

 red above and blue below. It was in fact a real and inverted image, formed at 

 the distance of about 3 feet from the vessel ; and a screen held,there received the 

 image in the form of a horizontal line of light with coloured edges, the action of 

 the liquid being somewhat similar to that of a cylindrical lens. All the images 

 were very impure in colour, being nearly white in their central portions. 



The colours were improved by lowering the eye so as to make the middle image 

 move up to the highest. Red was the first colour that appeared at the junction ; 

 and it showed extremely well. Violet (or when the light was feeble, blue) was 

 the last colour tliat was seen before both images became extinct by the descent of 

 the observer's eye. 



The largest sheets of colour were seen when the eye was exactly at the place 

 where the real image was formed. It was easy to obtain a long vertical strip of 

 blue by holding the eye at the distance of about 3 feet, and a long vertical strip of 

 red by holding the ej'e at the distance of about 4 feet. A long vertical strip of rich 

 yellow could be obtained at an intermediate distance. 



The experiment was varied by holding close in front of the eye a card with a 



