VOL. XC.] PHILOSOPHICAL TRANSACTIONS. 669 



verted. The magnitude would depend on the relative distances of the eye and 

 object. Below the point d, where the curve terminates, vision would be direct, 

 so that an object might be situated so as to be seen in all the 3 ways at the same 

 time, direct at o, inverted at i, and erect again at a. 



I consider the foregoing propositions as applicable to all cases of varying density, 

 whether occasioned by mutual solution of different fluids, or partial rarefaction of 

 the same fluid; and by trial of various fluids, however different in density, or even 

 in viscidity, I find that the refractions observe a law agreeable to the theory, as will 

 appear by the following experiments. 



Exper. 1. Into a square phial containing a small quantity of clear syrup, I put 

 about an equal quantity of water, in such a way that it floated on the surface of 

 the syrup, without mixing. For a short time, the stratum of union was so thin 

 that nothing could be distinctly seen through it. But soon, by mutual penetration 

 of the water and the syrup, the effects represented at a, fig. 7, were produced. 

 Through the syrup, a word written on a card placed behind was seen erect, and in 

 its place; through the adjacent variable medium, an inverted image was visible 

 above the trite place; and also above that a 2d image of the same object appeared 

 erect. When these appearances are first discernible, the variations of density are 

 so great, that the object to be looked at may be in contact with the phial; but 

 when the variations of density become more gradual, and so the focus more distant, 

 any object so near is only elongated, and requires to be removed an inch or 2, to 

 be seen inverted. 



Exper. 2. Over the surface of the water, in the same phial, I next put about 

 the same measure of rectified spirit of wine. At the stratum where the water and 

 spirit united, the appearances were the same; but since the refractive power of 

 spirit exceeds that of water, the true place of the object was seen uppermost; the 

 inverted and erect images are below. Fig. 7> b. When an oblique line der is 

 viewed through any variable medium so made, it appears bent into different forms, 

 according to its situation with respect to focal distance. If it be at the distance of 

 the principal focus, one point of it is dilated into a vertical line, as lm. Fig. 8, a. 

 If beyond that focus, the portion lm is inclined backwards, being an inverted image 

 of dl; and mn is another image of ihe same portion seen erect. Fig. 8, b. On 

 this account, it becomes a convenient object for ascertaining the state of any 

 medium under examination. 



In these experiments, the appearances continue many hours, even with spirit of 

 wine; with syrup, 2 or 3 days; with acid of vitriol, 4 or 5; with a solution of 

 gum arabic, much longer; but though their disposition to unite is so different, yet 

 the appearances produced are the same in all. The refraction is greatest nearly in 

 the plain of original contact of the fluids, diminishing from thence both upwards 

 and downwards. The exact rate of diminution above or below this point, I had 

 no means of measuring, with the accuracy that would be requisite for determining 

 the nature of the curves of density formed according to the first proposition. But 



