ALMOST the only real difficulty in preparing such articles as these arises when a new comet is discovered. To give a detailed account of its motions during the ensuing month, the writer must have before him an ephemeris of the comets place for the next five or six weeks; and this is not always forthcoming. An approximate orbit is indeed usually calculated by the time a comet has been in sight for a week, and its motion predicted by the end of this for two or three weeks more. But these preliminary orbits are often not very accurate. The difficulty may be illustrated by that of finding the exact size and position of a circle which passes through three points, A, B, and C. close together, and nearly but not quite in line. If we suppose that the position of the point B has been observed, and that a slight error has been made placing it at B instead of its real position, our calculations will lead us to the circle A.B.C.D instead of the correct circle, A B CD. Within the region covered by the original observations, the two circles will be quite close together, but outside this region they will deviate from one another at an ever-increasing rate. The mathematical processes involved . in calculating a cometa orbit are closely analogous to this simple case, though much more complicated. It is, therefore, easy to see that a small error in one of the observed positions of the comet, on which the calculations are- based, may give rise to . an enormously greater error when the attempt is made to predict, by means of these calculations, where the comet will be after an interval of time ten times as great, or more, than that covered by the original observations. It is therefore not worth while to use the preliminary orbit .of a comet to predict its motion for more than two or three weeks. Before the end of this time, many more observations can be secured. and an improved determination of the orbit made (just as, in our diagram, it is evident that, if we had observed the positions of two or three mere points on the circle between C and D. even with errors as large as that made in the case of point B, the whole set cf such observations would suffice to fix the position of the true circle with considerable . accuracy). For these reasons it is the exception, rather than the rule, when an ephemeris predicting a comets position till the end of a given month is available by the 15th or 20th of the preceding month; and this explains why the predictions concerning newly discovered comets, appearing in these columns, may sometimes appear vague. Brooks comet, mentioned last month, is a case in point. In this instance, too, the early estimates cf brightness seem to have been too low. The comet has not only become visible to the naked eye, but fairly conspicuous. Its orbit is now pretty well determined, perihelion passage occurring near the . end of October, at a distance of some 40 million miles from the sun. During the early days of the month it will be visible in the evening sky, though the moon, then in her second quarter, will interfere with observation. There ought, however, to be no difficulty in identifying the large diffuse head of the comet with a field-glass. Later in the month it passes between the pole a nil the sun, and for a few days will be visible both in the evening and the morning, low in the northwest after dark, and low in the northeast before dawn. At the time- of perihelion, toward the end of the month, it will be in Virgo, rising more than two hours before Slight shift of B makes very different curve the sun, and should be a fairly conspicuous object after new moon, when the morning skies are dark. THE HEAVENS. At the hour for which our map portrays the appearance of the sky the Milky Way forms a vast arch passing almost through the zenith. Along this are many fine constellations Sagittarius, low in the At 9 V. o'clock: September 29. NIGHT SKY: SEPTEMBER AND OCTOBER southwest; Aquila, well up in the same quarter; Cyg-nus, almost overhead, but a little west of the zenith; Lyra, west of this; Cassiopeia, high in the northeast; the less conspicuous Cepheus, just above the Pole star; Perseus, at a moderate altitude in the northeast; and finally Auriga, just rising below him. The Great Bear is low on the northwestern horizon, and the Dragon, coiling around the Little Bear, is higher up. Bootes has almost set in the northwest, and Corona is very low. Hercules is a little higher. on the way toward Lyra. Ophiuchus and Serpens are also represented by only a few stars in the west. Capricornus and Aquarius fill a large, dull region in the southern sky. Below them is an isolated star, which from its very loneliness seems all the brighter. This is Fomalhaut, the only conspicuous object in the constellation of the Southern Fish. A nearby horizontal line of five stars of the fourth magnitude, not shown on our map. but visible easily enough on a clear night, form the remainder of the constellation. Fomalhaut is of some interest as being within a measurable distance, though not one of the very nearest stars. Its parallax is 0.14 second, corresponding to a distance of 26 light years, and its actual brightness is about fifteen times that of the sun. Below Fomalhaut, and now on the meridian, is the far southern group of Grus. the Crane. Its two brightest stars, of the second magnitude, barely rise above the horizon of New York, but it is a conspicuous constellation in lower latitudes. The great square of Pegasus is dose to the zenith, on the southeast. Andromeda is just to the left with Triangulum and Aries below her. Taurus is rising, a little north of east. Cetus and Pisces fill the southeastern sky. THE PLANETS. Mercury is morning star until the 23rd when he passes behind the sun, and becomes evening star. He rises about an hour and a quarter earlier than the sun at the beginning of the month, and should be fairly easy to see. By the 15th he. is lost in the sun's rays, and remans so until November is half over. Venus is likewise a morning star, but is much farther from the sun, and exceedingly conspicuous— reaching her greatest brilliancy on the. 21st. She rises about 4:15 A. M. on the 1st, and just before 3 A. M. on . the 31st. At I P. M. on the 18th she is in conjunction with the moon. Though de- grees south of the latter. she can then be easily found with the naked eye —her great brightness making her clearly visible, even at noon. Mars is in Taurus, not far from Aldebaran, rising about 8:30 P. M. on the 1st, and 6:30 on the 31st. He is steadily approaching the earth, and by the end of the month is less than fifty m i Hion miles distant, which makes him appear very bright. Jupiter is evening star at the beginning of the month, setting about 7 P. M.. and too near the sun to be easily seen. As he passes through conjunction on the 23rd, be is practically invisible later in the month. Saturn is in Aries, approaching opposition. He rises before 7 P. M. in the middle of the month, and is well seen before midnight. Uranus is in quadrature with the sun on the 19th, and can be observed in ;he west in the early evening. Neptune is in a similar configuration, on the other side of the sun, on the 17th, and is visible. in the morning sky. The moon is full at 11 P. M. on October 7th, in her last quarter at 7 P. M. on the 14th, new at 11 P. M. on the- 21st. and in her first quarter at 2 A. M. on the 30th. She is nearest . us on the 11th, and farthest away on the 27th. The present new moon is marked by an annular eclipse of the sun, visible along a tract passing from the Aral Sea through Turkestan, Tibet, and southern China. thence through the southern part of the Philippines, and over .New Guinea. As a partial eclipse it will be visible over all southern Asia and Australasia from India to Siberia, and from Japan to southern Australia. Besides the conjunction with Venus on the 18th, already mentioned, the moon passes near Saturn on the 10th, Mars on the 11th, Neptune on the 14th, Mercury on the 21st, Jupiter on the 23rd, and Uranus on the 29th, none of the observable .conjunctions being close. Princeton University Observatory