Page:Popular Science Monthly Volume 8.djvu/654

636 the faintest and most difficult objects to observe in the solar system. Indeed, it is not wholly certain that they have ever been seen save in the telescopes of Mr, Lassell (their discoverer), Lord Rosse, and by the Washington refractor, although there are several telescopes now mounted both in Europe and in America which are adequate for their observation.

The satellite of Neptune, too, is a very difficult object, and hence it is extremely gratifying to find so many measures of these satellites as Prof. Newcomb has obtained. The telescope was mounted in November, 1873. From that time to April, 1875, there were made:

It must be remembered that Neptune was only observed from July to February, and Uranus from January to May.

From a consideration of all the measures of Uranus's satellites, the author assigns as the mass of that planet $1⁄22600$ of the mass of the sun, and he estimates the probable error of the denominator of this fraction at 100, so that we may say that this mass is not less than $1⁄22700$ and not more than $1⁄22500$; that is, the mass is determined within less than $1⁄200$ part of its value. To understand the nicety of such measurements as have been made, it must be remembered that any error in the measures of the distance of the satellite from the planet is shown in the resulting mass of the planet in an amount not proportional to this error directly, but to the third power of the error.

The times of revolution of the satellites have been determined with high accuracy by a comparison of Newcomb's observations with those of the elder Herschel—the uncertainty in the period of Titania $$=$$ 8days.705897, being not more than one second of time, or $1⁄1000000$ of the whole amount.

From the relative brightness of the satellites of Uranus, Prof. Newcomb concludes that they have masses not more than $1⁄60000$ of that of Uranus itself, i. e., vastly less than the mass of our own moon.

It is an interesting fact too that the author suspects that the nearest of the satellites of Uranus (Ariel) "belongs to that class of satellites of which the brilliancy is variable, and depends on its position in the orbit." With regard to the interesting question as to the number of satellites of Uranus, Prof. Newcomb's testimony is as follows:

The second section of the memoir deals with the Neptunian system. Three principal determinations of the mass of Neptune have been made:

From the work of the Washington telescope the mass results $1⁄194000$, which agrees most remarkably with Bond's previous determination.

No evidence for an elliptic form to the orbits of any of these satellites has been made out: "We are thus led to the remarkable conclusion that the orbits of all the satellites of the two outer planets are less eccentric than those of the planets of our system, and that, so far as observations have yet shown, they may be perfect circles. No trace of a second satellite of Neptune has ever been seen, though several times carefully looked for, under the finest atmospheric conditions, during July, 1874."

We have thus far spoken mainly of the