Its more rapid eastward motion causes it to overtake and pass Mars on the morning of May 6th. At the time of nearest approach the planets are only 5' apart, a distance less than twice the minimum separable by the naked eye. Unfortunately this very close approach occurs while the planets are below our horizon, but may be seen in the eastern hemisphere. Venus overtakes and passes Jupiter on the evening of May 11th. The minimum distance is a little more than 1°, about two diameters of the Moon, and occurs at a time when the planets are above our horizon. Then on the morning of May 18th Mars overtakes and passes Jupiter, the least distance being about the same as that at the conjunction of Venus and Mars.

Mars is still visible in the evening sky, but the Sun is drawing nearer to it. On May 1st it sets about two hours after the Sun; on June 1st, about an hour, and on July 1st, less than half an hour. The planet has now reached nearly its minimum of brightness, about that of the Pole Star, and it will not be easy to see it in the evening sky after June 1st. It will reach conjunction with the Sun about the middle of July, becoming a morning star, but will not be at all conspicuous for several months.

Jupiter during the early part of the May-June period is still conspicuous in the evening twilight. On May 1st it does not set until after 9 o'clock, but by June 1st, it remains above the horizon only about half an hour after sunset. It may be possible to see the planet on that date owing to its great brightness, but it soon draws too near the Sun for naked-eye view. It passes conjunction with the Sun on the morning of June 10th and becomes a morning star. By the end of June it rises about an hour before sunrise, and may be easily seen as a morning star. Between May 1st and June 30th it moves about 15° eastward in the constellation Taurus.

Saturn rises at about 3 A. M. on May 1st, and at about 11 P. M. on June 30th. It is therefore in fair position for earlymorning observation. It is in the eastern part of the constellation Aquarius and moves about 3° eastward and northward, with constantly diminishing motion, until June 27th. It then ceases its eastward motion and begins to move westward. During the first half of the year the Earth has been continually

drawing nearer the plane of the rings, and the apparent minor axis is in June, only about one eighteenth of the major. The motion of the Earth during the latter half of the year will cause a slight increase in the apparent minor axis, but the Earth will pass the plane of the rings in 1907, and the apparent ellipse will reduce to a mere line.

Uranus rises before midnight on May 1st and shortly after 7 P. M. on June 30th. It retrogrades (moves westward) about 2° in the constellation Sagittarius during this period, and is a few degrees north of the "milk-dipper." No bright star is very near.

Neptune is in Gemini, and is in the western sky in the evening. It will reach conjunction with the Sun on July 2d.

NOTES FROM PACIFIC COAST OBSERVATORIES.

OBSERVATIONS OF ECLIPSE SHADOW-BANDS OF AUGUST 30, 1905.

I regret exceedingly to say that through my oversight the valuable observations of shadow-bands made under my direction at the eclipse of August 30th, by Dr. VIGGO STROYBERG, of Copenhagen, were omitted from the account of the Lick Observatory-Crocker Eclipse Expedition to Spain published in No. 106. I trust that Dr. STROYBERG will pardon the great injustice thereby done to his skill and enthusiasm. Following is quoted from his record of observation:

"My place of observation was about 10 meters immediately north of the instruments. I had spread on the ground two white sheets of cloth, each between two and three meters square.

"Nearly one minute before totality the shadow-bands began. Their distance apart was about 15cm, and the dark lines were 3cm to 4cm broad. The velocity was so great that it was impossible to estimate it. I saw the lines straight and not undulating. I fixed the position of the lines with a long wooden bar, wnich I laid on the eastern sheet parallel to the lines. Their motion was southeastward and perpendicular to the bar. About 15 seconds before totality the shadow-bands disappeared.

"About 20 seconds after totality was over the shadow-bands came again. They had nearly the same motion and direction as before totality. The direction was fixed with another long wooden bar on the western sheet. They were very small and only a few of them could be seen. Their distances were nearly the same as before totality."

The azimuths of the two bars were measured and found to be, for the first, 35° south of west, and, for the second, 32° south of west. The directions of motion were therefore 55° and 58° east of south, respectively.

I do not venture an interpretation of these phenomena, nor a criticism of existing theories as to their cause; but it is of interest to note that their direction of motion was very nearly that of the Moon's shadow. W. W. CAMPBELL.

THE SYSTEM OF CASTOR.

(Abstract from L. O. Bulletin No. 98.)

Both components of this interesting visual double star are of the Sirian type of spectrum. The absorption is more complete in the fainter star (a, Geminorum, magn.=3.7) than in the brighter (a2 Geminorum, magn.=2.7). As a result the lines are more distinct in a1 and more lines can be measured. But the less distinct lines in a admit of slightly more precise settings being made.

The fainter component was shown to be a spectroscopic binary, with a period of nearly three days, by Professor BELOPOLSKY, of Pulkova, in 1895. The binary character of the brighter member of the system was discovered by the writer in October, 1904, from plates taken with the Milis spectrograph. The definitive orbits of both systems have been computed. The number of plates used in the case of a was thirty-two, and an average of thirty-four lines was measured on each plate. The elements derived for a depend upon forty-eight plates. The average number of measurable lines for this component was twenty-four. The full details of the calculations are given in the original paper. The resulting definitive elements for the two systems are as follows:

a, Geminorum.

2

Geminorum.

(Brighter Component.)

[pov] = 37.6

± 5°.120

+0.220

± 0.15km

1,485.000km
29.0

0.0112

±1°.730

0.22 -0.17km

For an Geminorum the eccentricity is shown to be much smaller than supposed by BELOPOLSKY. In particular, BELOPOLSKY's assumption of any rotation in the line of apsides is shown to be without foundation. The maximum effect of such a rotation for elements with an eccentricity of 0.01 would be but 0.3km, an amount too small to detect with certainty, in stars of this spectral type.

With future determinations of the elements of the visual

system it will eventually be possible to determine with considerable accuracy the parallax, masses, and orbital dimensions of this unique quadruple system. But the visual elements are absolutely indeterminate as yet, and assumptions as to the values of the parallax and other physical constants have no value.

It is not unreasonable to postulate, however, that the inclination of the orbital planes of both spectroscopic systems is roughly that of the main system. This is indeterminate at present. If we assume i=63° as given by DOBERCK1 in the elements which he regards as most probable (period=347 years), then the values for the semi-major axes of the spectroscopic systems are:

a1 Geminorum, a =1,435,000km

a Geminorum, a =1,667,000km

2

These values are mere hypotheses, but it seems quite probable that the orbital dimensions of both systems are of the same order of magnitude. We would have in this case the interesting combination of two spectroscopic systems of approximately the same linear dimensions, one of which has the very great eccentricity of 0.50, while the orbit of the other is nearly circular. By the commonly accepted theories of stellar evolution this would seem to indicate that the brighter pair is much the older of the two systems. It has already approximated to the eccentricity of the main system, while its fainter companion system, of three times greater mass, still revolves in orbits almost circular.

2

Such an eccentricity as shown in a is rarely met with except in those spectroscopic binaries which have as well a variation in their light. The tidal action in such a system, where at periastron the stars are but one third of their apastron distance apart, must be enormous. The attempt was accordingly made to test the constancy of the light of the brighter star by a series of observations with the smaller of the two Bruce polarizing double-image photometers belonging to the Lick Observatory. The observations were made with the photometer attached to the twelve-inch equatorial. There is no evidence of any variation from the present series. If it

'A. N., Vol. 166. p. 145.