the atmosphere surrounding the earth, so that the direction in which the light enters the telescope is somewhat different from the direction which the light had when it first encountered the atmosphere at an altitude of some hundreds of miles. The amount of this change in the direction of a ray by refraction is by no means inconsiderable, and the practical astronomer has always to allow for it. At sunset, for example, the light from the departing orb is bent to such an extent, that at the time when the sun has really sunk below the horizon it still appears to be above, on account of the curvature in the rays produced by atmospheric refraction. If, therefore, the light from a star had to traverse in its passage through the comet any quantity of vapour at all comparable in refractive power to the earth's atmosphere, that light would be deflected to an extent which could not be overlooked in the refined methods of modern astronomy.
Here then we have a delicate means for investigating the quantity of refractive matter in a comet. The observations are conducted in this manner. Two neighbouring stars are selected which are so placed with regard to the movements of the comet that it passes over one of the stars while leaving the other uncovered. Suppose that the apparent distance between the two stars upon the sky be measured, before the comet has come into their immediate neighbourhood. The measurement is to be repeated while one of the stars is behind the comet, and a third and concluding observation is to be made after the comet has passed on its way and left both stars behind. We have thus obtained the necessary materials for the investigation. If the body possessed any appreciable refractive power, then the apparent distance of the two stars would
