After making a variety of experiments with aerial telescopes constructed of metallic specula of different focal lengths, I constructed a telescope on the same plan, with a concave glass mirror. Having obtained a fragment of a very large convex mirror which happened accidentally to have been broken, I caused the convex side to be foliated, or silverised, and found its focal length to be about 27 inches. This mirror, which was about 5 inches diameter, I placed in one of the aerial reflectors, instead of the metallic speculum, and tried its effects with different terrestrial eye-pieces. With a power of about 35 or 40 times, it gave a beautiful and splendid view of distant terrestrial objects--the quantity of light reflected from them, being considerably greater than when a metallic speculum was used, and they appeared on the whole well-defined. The only imperfection--as I had foreseen--consisted in a double image being formed of objects which were remarkably bright and white, such as a light-house whitened on the outside, and strongly illuminated by the sun. One of the images was bright and the other faint. This was obviously owing to the two reflections from the two surfaces of the mirror--one from the convex silverised side, and the other from the concave side next the eye, which produced the faint image--which circumstance has been generally considered as a sufficient reason for rejecting the use of glass specula in telescopes. But although very bright objects exhibited a double image, almost all the other objects in the terrestrial landscape appeared quite distinct and without any secondary image, so that a common observer could scarcely have noticed any imperfection. When the instrument, however, was directed to celestial objects, the secondary image was somewhat vivid, so that every object appeared double. Jupiter appeared with two bodies, at a little distance from each other, and his four satellites appeared increased to eight. The moon likewise appeared as a double orb, but the principal image was distinct and well-defined. Such a telescope, therefore, was not well-adapted for celestial observations, but might answer well enough for viewing terrestrial objects.