As a comet contains transparent gaseous material we might expect that the place of a star would be deranged when the comet approached it. The refractive power of air is very considerable. When we look at the sunset, we see the sun appearing to pass below the horizon; yet the sun has actually sunk beneath the horizon before any part of its disk appears to have commenced its descent. The refractive power of the air bends the luminous rays round and shows the sun, though it is directly screened by the intervening obstacles. The refractive power of the material of comets has been carefully tested. A comet has been observed to approach two stars; one of which was seen through the comet, while the other could be observed directly. If the body had any appreciable quantity of gas in its composition the relative places of the two stars would be altered. This question has been more than once submitted to the test of actual measurement. It has sometimes been found that no appreciable change of position could be detected, and that accordingly in such cases the comet has no perceptible density. Careful measurements of the great comet in 1881 showed, however, that in the neighbourhood of the nucleus there was some refractive power, though quite insignificant in comparison with the refraction of our atmosphere. [Illustration: PLATE C. COMET A 1892, 1. SWIFT. _Photographed by E.E. Barnard, 7th April, 1892._] From these considerations it will probably be at once admitted that the _mass_ of a comet must be indeed a very small quantity in comparison with its bulk. When we attempt actually to weigh the comet, our efforts have proved abortive. We have been able to weigh the mighty planets Jupiter and Saturn; we have been even able to weigh the vast sun himself; the law of gravitation has provided us with a stupendous weighing apparatus, which has been applied in all these cases with success, but the same methods applied to comets are speedily seen to be illusory. No weighing machinery known to the astronomer is delicate enough to determine the weight of a comet. All that we can accomplish in any circumstances is to weigh one heavenly body in comparison with another. Comets seem to be almost imponderable when estimated by such robust masses as those of the earth, or any of the other great planets. Of course, it will be understood that when we say the weight of a comet is inappreciable, we mean with regard to the other bodies of our system. Perhaps no one now doubts that a great comet must really weigh tons; though whether those tons are to be reckoned in tens, in hundreds, in thousands, or in millions, the total seems quite insignificant when compared with the weight of a body like the earth.