rays would be apt to shine upon the lens, which would make the picture appear with a confused lustre. An east window is best in the afternoon, and a western in the morning; but a north window is in most cases to be preferred, especially in the forenoon, when the sun is shining with his greatest strength and splendour. In general, that window ought to be used which looks to the quarter opposite to that in which the sun is shining. The picture should be received upon a very white surface, as the finest and whitest paper, or a painted cloth, bordered with black; as white bodies reflect most copiously the incident rays, while black surfaces absorb them. If the screen could be bent into the concave segment of a sphere, of which the focal distance of the double convex lens which is used, is the radius, the parts of the picture adjacent to the extremities would appear most distinct. Sir D. Brewster informs us that, having tried a number of white substances of different degrees of smoothness, and several metallic surfaces, on which to receive the image, he happened to receive the picture on the silvered back of a looking-glass, and was surprised at the brilliancy and distinctness with which external objects were represented. To remove the spherical protuberances of the tin foil, he ground the surface very carefully with a bed of hones which he had used for working the plane specula of Newtonian telescopes. By this operation, which may be performed without injuring the other side of the mirror, he obtained a surface finely adapted for the reception of images. The minute parts of the landscape were formed with so much precision, and the brilliancy of colouring was so uncommonly fine, as to equal, if not exceed the images that are formed in the air by means of concave specula.