By reckoning backward, it was found to have been seen and mapped several times as a star,—no less than twelve times by Lemonnier alone,—and yet its planetary character had slipped through his fingers. It can even be seen with the naked eye as a star of the 6th magnitude, and its course is said to have been watched by savage tribes in Polynesia long before Sir William Herschel discovered it. Its greenish blue disk indicates that it is about thirty-two thousand miles in diameter, and its mass that its density is about 0.22 of the Earth’s or, like Jupiter’s, somewhat greater than water. Of its surface we probably see nothing. Indeed, it is very doubtful if it have any surface properly so called, being but a ball of vapors. Its flattening, ¹/₁₁ according to Schiaparelli, which is probably the best determination, agrees with the density given above, indicating its substance to be very light. Belts have faintly been descried traversing its disk after the analogy of Jupiter and Saturn. These would be much better known than they are but for the great tilt of the planet’s axis to the ecliptic, so that during a part of its immense annual sweep its poles are pointed nearly at the Earth, and its tropical features, the places where the belts lie, are wholly hidden or greatly foreshortened from our point of view. As the planet’s year is eighty-four of our years long, it is only at intervals of forty odd years that the disk is well enough displayed to bring the belts into observable position. The planet is attended by four satellites,—Ariel, Umbriel, Titania, and Oberon,—a midsummer night’s dream to a watcher of the skies. They travel in a plane inclined 98° to the ecliptic, so that their motion is nearly up and down to that plane and even a little backward. Whether their plane is also the equatorial plane of the planet, we do not know for certain. The observations as yet are not conclusive one way or the other. If the two planes should turn out not to coincide, it will open up some new fields in celestial mechanics. The belts have been thought to indicate divergence, but the most recent observations by Perrotin on them minimize this. They suggest, too, a rotation period of about ten hours, which is what we should expect. Its albedo, or intrinsic brightness, is, according to Müller, 0.73, or almost exactly that of cloud. This tallies with the lack of pronouncement of the belts and is another argument against the reality of the recent diametral measurements, as all Müller’s values are got by dividing the amount of light received by the amount of surface sending it. If the diameter were much less than thirty-two thousand miles, the resulting albedo would become impossibly high.