The theory that I advance to account for the several phenomena relating to comets' tails is, that comets are non-luminous, transparent bodies; that they transmit the light of the sun; that the transmitted light reflected by the particles of matter in space constitutes the tails of comets. "Like causes produce like effects." By contraries, then, like effects must be produced by similar causes; for, if an effect produced by a cause which is known is similar to an effect produced by a cause which is not known, the cause which is known must be similar to the cause which is not known. This is true or not. I submit the following experiments to substantiate the theory advanced. Partially fill a vial or a tumbler with water, hold it by the rim, and move it around a lighted candle placed upon a table. A shadow surrounding the transmitted light will be cast upon the table. As the tumbler approaches the light, the shadow follows the tumbler, and when receding the tumbler follows the shadow; and as the tumbler is moved around the light, the shadow will swing round from one side to the other. If the tumbler be held so that a puff of smoke can be blown into the transmitted rays, the particles of smoke will reflect the transmitted light, and will illustrate my idea of what constitutes a comet's tail. A dark band may be observed in this stream of light, as also in the light cast upon the table. In these experiments, we see the effects produced by a cause which is known; the effects are similar to those observed in the tails of comets, the cause of which we do not know; but is it not reasonable to assume that the cause is similar? Assuming now that comets are transparent, can any other phenomena peculiar to comets be accounted for upon this hypothesis? Next to the tail itself, the curve is the most noticeable feature, and if we consider the extraordinary length of these appendages, the astounding velocity at which comets move in their orbits, and the time that would elapse before a ray of light, emitted from the nucleus, would reach the end of the tail, perhaps the curve--which, if I am not deceived in my observations, always dips toward its orbit--can be accounted for. If a comet moved in a direct line toward the center of the sun, there would be no curve to the tail. But taking Donati's comet of 1858 as an example, the tail of which was said to be about 200,000,000 miles long, a ray of light traveling at the rate of 192,000 miles per second would be about twenty minutes in going from the nucleus to the end of the tail. But during that time the comet would move in its orbit, say, 50,000 miles, and as light moves in a straight line, and other rays are constantly emerging from the nucleus as it moves along in its course, the result is that the tail has a curved appearance.