September, 1903.] 



KNOWLEDGE. 



201 



Admiral Smyth remarked, with i-eference to the double 

 star Struve 946, " the possibility of the comes being 

 variable awakens considerations of peculiar interest ; it 

 having been surmised that certain small acolyte stars 

 shiue by reflected light." '• But it may be easily shown 

 that this is highly improbable, if not impossible. Let us 

 take the system of Sirius. In this case the satellite — 

 although very faint for its computed mass — certainly does 

 not shine by reflected light from Sirius. This will appear 

 from the following considerations which I have carefully 

 worked out. Assuming for a moment that the satellite 

 shines merely by reflected light, let us see what its 

 brightness would be as seen from the earth. According to 

 the computed orbit and parallax of Sirius — which are 

 probably as reliable as those of any binary star hitherto 

 computed — the mean distance of the satellite from the 

 bright star is a little more than the distance of Uranus 

 from the sun. Let us assume this distance. (A greater 

 distance would strengthen my argument.) As the com 

 puted mass of the satellite is about the same as that of 

 the sun, let us assume that it has the same diameter or 

 866,000 miles (a smaller diameter would of course 

 strengthen my argument), and let us take the diameter of 

 Uranus at 33,000 miles — which is very near the truth. 

 Now assuming the same '' albedo," or reflective power for 

 Uranus and the satellite of Sii'ius (the albedo of Uranus 

 is very high), we have the satellite, as seen from Sirius, 

 shining with a greater brilliancy than Uranus as seen 

 from the sun in the proportion of 866,000 squared to 

 33,000 squared, or as 688 to 1. This is on the assumption 

 that Sirius and the sun are of equal brightness. But 

 from the photometric measures of Sirius, and its known 

 distance from the earth, I find that Sirius is at least 20 

 times brighter than our sun. We must, therefore, increase 

 the above ratio 2<) times to obtain the illumination of the 

 satellite by the light of Sirius. This gives 688x20= 

 13,760. That is, the satellite as seen from Sirius would 

 be about 13,760 times brighter than Uranus as seen from 

 the sun. This number corresponds to 10'3 stellar magni- 

 tudes. Now, taking the magnitude of Uranus, as seen 

 from the sun, at 58 (which must be very near the truth), 

 we have the brightness of the Siriau satellite, as seen from 

 Sirius, equal to -SS —10-3, or -4-5 magnitude, that 

 is, 4i magnitudes brighter than a star of zero magnitude, 

 like Arcturus, or slightly brighter than Venus appears at 

 her greatest brilliancy as seen from the earth. Now, the 

 the simple problem is this : if a body shines with a stellar 

 magnitude of — 4'5 as seen at the distance of Uranus, 

 what would be its magnitude if jilaced at the distance of 

 Sirius ? Taking the parallax of Sirius at 0""38, we have 

 its distance from the earth equal to 542,800 times the sun's 

 distance from the earth. Hence the light of a body at the 

 distance of Uranus would, if removed to the distance of 

 Sirius, be reduced in the proportion of the srjuare of 

 542,800 to the square of 19, or as 816,244,900 to 1. This 

 corresponds to 223 stellar magnitudes. Hence the mag- 

 nitude of the satellite of Sirius, as seen from the earth — 

 (/ shining only by rejh'clcl liijht from Sirius— would be 

 22'3 — 4'5, or 17'8 magnitude, and it would therefore be 

 quite invisible in the great 40-iuch telescope of theYerkes 

 Observatory, even if seen on a dark sky, the smallest star 

 visible in that telescope being about the 17th magnitude. 

 As its actual brightness is about the 10th magnitude, it 

 follows that it is about loOO times brighter than if it 

 shone merely by reflected light, and it is evident that it 

 must have some inherent light of its own. I have shown 

 in the beginning of this paper that the actual brightness 

 of the satellite, as seen from Sirius, is equal to that of full 



moonlight on the earth. We shovdd obtain a very similar 

 result if we assumed that Sirius is very much brighter 

 than 20 times the brightuess of the sun. If we assume it 

 to be lO times brighter than this, or 200 times the sun's 

 brightness — a very improbable sujiposition — we should 

 still have the satellite reduced to about the 15th magni- 

 tude, and placed as it is so close to such a brilliant star as 

 Sirius, it would probably still remain invisible in our 

 largest telescopes. The assumption I have just made is, 

 however, quite inadmissible, for if we increase the light of 

 Sirius we must increase its distance also, and this would 

 further diminish the computed light of the satellite. We 

 may, therefore, dismiss the idea that the satellite of Sirius 

 could possibly shine merely by reflected light from its 

 primary. The same considerations will apply to the case 

 of Procyon and its satellite, and with greater force, as 

 the satellite of Procyon is about three magnitudes fainter 

 than the Sirian satellite, and Procyon is a less luminous 

 sun than " the monarch of the skies." 



Bedford Catalui/ue, p. 155. 



PHOTOGRAPHS OF COMET c 1903 (BORELLY). 



By Isaac Egberts, d.sc, f.e.s. 



The comet was discovered on the 21st of June by M. 

 Borelly, at the Observatory of Marseilles, and the fact was 

 immediately communicated to the Central-Slelle at Kiel, 

 from which centre Prof. Kreutz distributed the information 

 to all the principal observatories in Europe and in America. 

 Ephemerides of the positions and orbit of the comet were 

 quickly issued, by aid of which astronomers could 

 follow its movements in the sky and make their observa- 

 tions upon it. 



The three photographs annexed hereto were taken on 

 the 24th and 26th July, when the comet was seven days 

 past that part of its orbit where it was computed to attain 

 the maximum brightness, namely, 14 6 times brighter 

 than on the day of its discovery. On the 24th July, wheu 

 the photograph. Fig. 1, was taken, the brightness had 

 diminished to about 110 times, and on the 26th. wheu 

 the photos Pigs. 2 and 3 were taken, the brightness had 

 further diminished to about 98 times, and the light will 

 continue to decrease till the comet again will lie invisible. 

 Figs. 1 and 3 on the j'late were taken with the Cooke 

 5-inch lens, the exposures being (1) 45 minutes, (3) 60 

 minutes ; scale about 122" of arc to 1 millimetre ; the 

 length of the part shown of the tail is about six degrees, 

 but if the plate had been larger the tail would have been 

 shown to be considerably greater iu length. 



If we compare Fig. 1 with Fig. 3 it will be seen that 

 the tail has imdergone a remarkable change in outline, iu 

 structure, and in density, during the interval of two days, 

 and that the gap of about one and a half degrees in length 

 from the head had almost closed vqi, and the tail had 

 assumed a more symmetrical form of straight streamei-s 

 and of apparent nebulous matter, but they are too faint to 

 be fully shown on the print. 



Fig. 2. This photograph was taken with the 20-iuch 

 reflector sinuiltaneously with No. 3 ; the scale in this case 

 being about 43" of arc to 1 millimetre. It shows more 

 clearly the structure of the tail than can be seen with the 

 5-inch lens. The length of the tail is about 2} degi-ees, 

 for the plate was too small to show more of it, but the 

 structure is well seen, consisting of streamers and vacant 

 spaces between them. Unfortunately the bad weather and 

 moonlight nights have prevented more photographic evi- 

 dence being obtained at Startield of this interesting comet. 



The nucleus was strongly stellar, Init the surrounding 

 comn ilid not indicate any structural details. If the three- 

 fold classification nf comets should be reliable, the three 



