tween the two adjacent surfaces, it may increase the extent in any greater proportion.
The last objection to which the author replies, relates to those po- sitions in which rings of colours, and other similar phenomena, are seen, but in which the colours produced by critical separation could not reach the eye. For instance, rings and bands of colour, which arise from contact at the under surface of a plate of glass terminated by parallel planes, are seen through the upper surface, although co- lours separated by critical reflection or intermission, evidently could not come to the eye under these circumstances. But Dr. Herschel reminds us, that he does not affirm critical separation to be the sole cause of the rings produced by contact of a plane and sphere, but that it only furnishes the colours, which are afterwards modified by the subjacent spherical surface; and next proceeds to several sets of experiments, which he considers decisive in support of the validity of his theory, in reply to this objection.
In the first set of experiments a series of prisms, of different forms, are successively placed within their bases, resting upon a spherical metallic surface.
When a right-angled prism was placed in this situation, and the eye was gradually raised from a level with its base, no colours were seen till it arrived at the elevation necessary for critical separation. At this point the blue bow became visible, and rings began to be-per— ceived at the same time. When the eye was lifted gradually higher and higher, till it arrived opposite to the vertex of the prism, the rings continued visible, without interruption, notwithstanding suc- cessive changes which occurred in their colours and size: and even when the eye was carried beyond the vertical position, the same rings continued visible, so as to be seen, upon the whole, through a range of at least 77°.
Instead of the right-angled prism, having a retracting angle of 45° on each side, Dr. Herschel afterwards substituted prisms with their vertical angles successively more obtuse, and with equal refracting angles on each side, first of 30°, then of 25", then of 20°, and lastly of 9° on each side.
In all these instances the phenomena were similar; but the range of visibility increased in proportion as the refracting angle was smaller, so that in the last instance the range within which the rings were visible from each surface, exceeded 138°. And hence might be inferred the still greater extent, in case of plain glass, which may be looked upon as a. prism with a vanishing retracting angle.
These experiments, in Dr. Herschel's estimation, establish the mo- difying power of spherical surfaces, whereby they render colours that have been entirely separated, visible in every direction.
In the next set of experiments, the author substitutes a cylindrical surface for that which in the former set was spherical; and by a similar series of prisms, successively more and more obtuse at their vertical angles, the coloured streaks, which in this case appeared instead of rings, were rendered visible to greater and greater extent, till with a plain glass they were seen as far as 170°.
