may ensue if sufficiently large quantities of opposite electricity have been accumulated. By reason of these discharges the gases will become luminescent, although their temperature may be very low. Stark has even shown that low temperatures are favorable for the display of a strong luminosity in electric discharges. We have stated that Kepler, as early as the beginning of the seventeenth century, came to the conclusion that the tails of comets were repelled by the sun. Newton indicated how we might, from the shape of the comets’ tails, calculate their velocity. The best way, however, is to determine this velocity by direct observation. The comets’ tails are not so uniform in appearance as they are generally represented in illustrations, but they often contain several luminous nuclei (Fig. 33), whose motions can be directly ascertained. [Illustration: Fig. 33.—Photograph of Roerdam’s comet (1893 II.), suggesting several strong nuclei in the tail] From a study of the movements of comets’ tails, Olbers concluded, about the beginning of the last century, that the repulsion of the comets’ tails by the sun is inversely proportional to the square of their distance—that is to say, that the force of the repulsion is subject to the same law as the force of gravitation. We can, therefore, express the repulsion effect in units of solar gravitation, and this has generally been done. That the radiation pressure will in the same manner change with the distance is only natural. For the radiation against the same surface is also inversely proportional to the square of the distance from the radiating body, the sun. [Illustration: Fig. 34.—Photograph of Swift’s comet (1892 I.)] In the latter part of the past century the Russian astronomer Bredichin conducted a great many measurements on the magnitude of the forces with which comets’ tails are repelled by the sun. He considered himself, on the strength of these measurements, justified in dividing comets’ tails into three classes. In the first class the repulsion was 19 times stronger than gravitation; in the second class, from 3.2 to 1.5 times stronger; and in the third class, from 1.3 to 1 times stronger. Still higher values have, however, been deduced for several comets. Thus Hussey found for the comet of 1893 (Roerdam’s comet, 1893 II., Fig. 33) a repulsion 37 times as strong as gravitation; and Swift’s comet (1892 I.) yields the still higher value of 40.5 (Fig. 34). Some comets show several tails of different kinds, as the famous comet of Donati (Fig. 35). Its two almost straight tails would belong to the first class, and the more strongly developed and curved third tail to the second class. [Illustration: Fig. 35.—Donati’s comet at its greatest brilliancy in 1858]