While Professor KAPTEYN contends that with the same expenditure of labor one may attain to nearly as small a probable error by this method of differential transits as by the use of the heliometer, he also concedes in his closing remarks that this method is best adapted to the investigation of parallax by the wholesale, rather than to that of individual parallaxes. He outlines a very attractive program whereby a few observers, in a comparatively short time, would be able to determine the parallaxes of all the stars of the fifth magnitude or brighter. According to the Uranometria Argentina these, including both hemispheres, number 1212. Assuming that, at the two extremes of parallax in Right Ascension, it is possible on the average to secure an available factor of 0.89 in the observations (measuring 1.76 in each case), eight observations at each extreme would give for the probable error of each parallax, o".052. All parallaxes of o". 20 or greater would be certainly detected; 90 per cent. of those greater than o". 15 would be revealed; and of those which come out by observation o". 10 or more, three-fourths would turn out to have parallaxes of o". 10 or greater. But, as he maintains, the important result would be that the mean parallax of each class (order of magnitude) would be determined with great exactness.

There seems to be nothing unreasonable, or unattainable, in the enthusiastic picture which Professor KAPTEYN presents. In the present state of stellar astronomy, it might be suggested that the stars for such a Parallax-Durchmusterung ought not to be selected according to their magnitudes, but according to their proper motions. Let the 1000 stars having largest proper motions be selected for this proposed investigation. If such an investigation were carried on with the vigor and skill which is manifest in the work of Professor KAPTEYN here under consideration by four observers, two in each hemisphere, there is scarcely a doubt but that the necessary observations could be completed on a liberal scale within from six to ten years, and that the results would be of inestimable value in founding the stellar astronomy of the future.

In selecting these stars having large proper motions, it might be well to leave out all those of the fifth magnitude or brighter. These would be but a small fraction of the whole; and there is not much doubt that more reliable results for these brighter stars would be attainable by the use of the heliometer. The stars fainter than the ninth magnitude should be attended to by pho

tography. In selecting the comparison stars for use with the parallax stars, care should be exercised, so far as possible, to employ no stars for comparison that have a sensible proper motion. Furthermore, the magnitudes of the comparison stars should be as nearly as possible the same as those of the respective parallax stars. For stars below the 6.5 magnitude these conditions would not be difficult of attainment; for brighter stars the difficulty would increase very rapidly with the magnitude, and this constitutes sufficient reason for turning those stars over to the heliometers. With such an arrangement of comparison stars and with no observing appliances further than those which are already available, we appear to have the means of attacking the general problem of star-parallax more effectively than ever before. It would be difficult to conceive of a method which offers better security against the influence of systematic error.

The work of Professor KAPTEYN well illustrates the possibility of putting new life into a hackneyed subject without the aid of startling novelties either in the details of methods, or in the apparatus employed.

OBSERVATIONS OF THE DARK TRANSIT OF JUPITER'S THIRD SATELLITE ON AUGUST 20, 1891, AT WINDSOR, NEW SOUTH WALES.

BY JOHN TEBBUTT, F. R. A. S.

The times of the various phases are local sidereal.

h. m.

S.

18 33 34; the satellite had advanced considerably on the disc of its primary and was visible as an oval dark spot on the north edge of one of the slender belts in the south equatorial zone. The major axis of the oval was par

18 50 59;

19 24 59;

20

allel to the belt. The satellite was not nearly so dark
as its shadow, which was also visible.
satellite still visible.

it was now at about mid-transit, and the dark phase
had slightly increased in depth.

3 59; the preceding observations were made with the 44-inch equatorial and a power of 120. I was now enabled to

turn the 8-inch telescope on the planet, and with a

power of 74, found the satellite to be quite conspicuous as a dark spot.

The following measures were made of the satellite's position on the disc by turning the distance-threads of the micrometer parallel to the belts, and placing one on the satellite and the other alternately across each pole of the planet. distance from south pole distance from north pole

=

12".23 (3 measures).
34".63 (3 measures).

20 II 54;

20 13 14;

20 21 59;

20 25 59;

satellite had obviously faded.
still plain, but fainter.

20 47 59;

not certainly visible.

satellite had been suspected for some time, and was now certainly seen as a faint bright spot just within the limb with powers 74 and 112.

20 50 39; distinctly seen as a bright spot with a power of 112 at

Clouds prevented observation of the transit of the 27th of the same month.

WINDSOR, N. S. WALES, Sept. 4, 1891.

(SEVENTH) AWARD OF THE DONOHOE COMET MEDAL.

The Comet Medal of the Astronomical Society of the Pacific has been awarded to E. E. BARNARD, Astronomer of the Lick Observatory, for his discovery of a Comet at o" 55" G. M. T., October 3, 1891. This is the nineteenth comet discovered by Mr. BARNARD.

The Committee on the Comet Medal,

EDWARD S. HOLDEN,
J. M. SCHAEBERLE,

CHARLES BURCKHALTER.

(Dated) October 12, 1891.

THE FORMS OF JUPITER'S SATELLITES.

BY J. M. SCHAEberle and W. W. Campbell.

While we were engaged in observing the markings on the third satellite of Jupiter with the 36-inch equatorial, in 1891, August, September and October, we noticed that the first satellite. was not round. Careful observations of all the satellites on several nights have led us to conclude that the first satellite of Jupiter is ellipsoidal, that its longest axis is directed towards the centre of Jupiter, and that the other satellites appear to be spherical. (See Astronomical Journal, no. 247.) The results of the observations on the several nights are given below.

1891, Sept. 6. Seeing very fine. [The spiral or double-ring structure of the nebula in Draco, G. C. 4373, shown beautifully While waiting for eclipse (reapseen that I was elongated in the

under all powers up to 3000.] pearance) of satellite III, it was position angle 70°-250°, in the ratio 5:4. I, II and IV were successively brought to the same collimation axis without changing the focus, using all powers up to 3000, but 2000 was used to best advantage. I was decidedly elliptical under all tests; II and IV were round. After the reappearance of III, the observations were repeated: I elongated; II, III and IV round.

At the extremities of the major axis of I were bright regions. which probably caused a part of the apparent elongation, but cannot account for it all. Later, it was noticed that the position angle 70°-250° was equivalent to the longer axis pointing towards Jupiter's centre.

1891, Sept. 16. Seeing, weight 3.

The highest power used was 1000. I was decidedly elongated, probably 5:4, in a direction parallel to Jupiter's equator, the end towards Jupiter being very bright. II appeared very slightly elongated in position angle 45°-225° with reference to Jupiter's polar axis, probably owing to the very bright region at the southern end of the major axis. III was round, but a very bright region on the northeast limb gave it the appearance of being elongated, when the seeing was poor, in the direction 30°-210° with reference to Jupiter's polar axis. This probably accounts

for SECCHI's observed ellipticity of this satellite (see Astr. Nach., no. 1017, pp. 135-142). IV was round; the end towards Jupiter bright.

1891, Sept. 20. Seeing, weight 2.

While I was in transit, near egress, it appeared round, or very nearly so. Its shadow was likewise round or only very slightly elongated the whole time. Powers greater than 700 could not be used on account of poor seeing. After the egress of I it was certainly elongated, though not so much as had been observed before, and there were bright regions at its preceding and followends. II was either round or elongated only very slightly as before. III and IV were round.

1891, Sept. 26. Seeing, weight 4.

I was certainly elongated in direction of Jupiter's equator, but less than when it was observed near quadrature. The bright regions on I were not at all prominent. II was either perfectly round, or very slightly elongated nearly at right angles to Jupiter's equator. III and IV were round. The northeast polar cap on III was scarcely visible. Powers 700 and 1000.

[W. W. C. alone.]

1891, Sept. 27. Seeing, weight 3.

Before and after transit, I was certainly elongated, apparently not so much as when seen near quadrature. During the visible part of its transit it was continually elongated in the direction of Jupiter's equator. When nearest the centre, it was almost round. At 10 27 P. s. t., just before egress, it was elongated 2:1, and two condensations of light in it gave it the appearance of being double. The shadow of I was elongated 3:2 just after ingress, and when nearest the centre was only very slightly elongated. II was very slightly elongated nearly at right angles to Jupiter's equator. III and IV were round.

I

1891, Oct. 4. Seeing, weight 3.

I

Before and after the transit of I, it was certainly elongated. During the transit the elongations of I and its shadow varied from 12:1 or even 134:1, just after ingress, to nearly round when they were nearest the centre. II appeared elongated nearly 5:4 in the direction 150-195° with reference to Jupiter's polar axis, but a very bright region was observed at its southern extremity. III

and IV were round.