September 1, 1890.] 



KNOWLEDGE 



203 



^^ AN ILLUSTRATED "^^ 



MAGAZINE OF SCIENCE 



SIMPLY WORDED— EXACTLY DESCRIBED 



LONDON: SEPTEMBER 1, 1890. 



CONTENTS. 



Variable Double Stars. By J. E. Gohe, F.R.A.S., M.R.I.A. 203 



Horns and Antlers. By R. Ltdekker, B.A.Cantab. ... 205 



Old Hindu Algebra. By Frederic Pixiott 207 



Why Is the Sea Salt ? By W. Mattied Williams 207 



The Bed-Bug.— I. By E. A. Butler 209 



On Some Recent Advances in the Mapping of the Solar 



Spectrum. By A C Rantard 211 



Letters:— R. Chartre.s ; S. E. Peal; W. W. E 213 



Notices of Bool<s 214 



The Mariner's Compass: Its Errors and their Causes. 



By R. Betnon 21G 



Harmless Parasites and Uninvited Companions. — II. 



By Henry J. Slack, F.G.S.; F.R.M.S 218 



The Face of the Sky for September. Bv Herbert 



Sadler, F.R.A.S ' 219 



Whist Column. By W. Montagu Gattie 220 



Chess Column. By I. Gunsbero 221 



VARIABLE DOUBLE STARS. 



By J. E. Gore, F.R.A.S. 



VARIABLE stars are not as a rule close doubles, 

 neither do the wider double stars include many 

 examples of the variable class. Among the known 

 binary or revohing double stars I do not know of 

 a single instance — with the possible exception 

 of 3G AndromediE — of undoubted variation, although in 

 some cases there exists a strong suspicion of inconstancy 

 of light. 



In speaking of double stars I refer to those which are 

 actually seen to be double in good telescopes, for on the 

 eclipsing satellite theory the variables of the Algol type 

 must bo considered as binary stars with very short periods, 

 and with tlie component stars so close that even the largest 

 telescopes yet constructed fail to show them as anything 

 but single stars. 



To the above rules there are, however, some notable 

 exceptions. Perhaps the most remarkable instance of a 

 variable star being also a close double is r) (Tominorum. 

 The variability of this 8rd magnitude star w'as detected by 

 Schmidt in 1865, and the reality of its fluctuations has 

 been confirmed by Schonfeld and other observers, including 

 the present writer. The total variation amounts to about 

 one magnitude in a period which is somewhat irregular, 

 varying in length from 185 to 151 days, according to 

 Schmidt. Occasionally, however, its light remains con- 

 stant, or nearly so, for several weeks at a time. The colour 

 of the star is reddish yellow, and its spectrum a tine one of 

 Secchi's 3rd type, a type to which many of the long period 

 variables belong. In 1881 Mr. l^uniham. the eminent 



American astronomer and discoverer of so many double 

 stars, foimd the star to be a close double, the companion 

 being of about the 10th magnitude and distant less than 

 one second of arc from its comparatively brilliant primary. 

 The difference of light is therefore about 7 magnitudes, 

 indicating that the bright star is about 630 times brighter 

 than the faint one. Mr. Burnham speaks of it as " a splendid 

 unequal pair, and likely to prove an interesting system." 

 Time will of course be requu-ed to prove the accuracy 

 of this prediction, but should its binary character be 

 established it will form a most interesting object, especially 

 as hitherto no binary star has been found with a spectrum 

 of the 3rd type. 



Another variable star with a close companion is the 

 short period variable S (15) Monocerotis. This star is 

 the principal one of a small cluster k-nown to astronomers 

 as Hersehel VIII. 5. The variable star, which fluctuates 

 from about 5th magnitude to 5|, has two faint companions 

 of about the 9th and 11th magnitude. The 9th magnitude 

 is distant about 3 seconds of arc fi'om the brighter star, 

 and the 11th magnitude about 17^ seconds, and both may 

 possibly have a physical connection with their primary. 

 If the close pair are really in motion, the period of revolu- 

 tion must be very long, as a measure by Mr, Tarrant in 

 1888 shows an angular change of only 6 degrees since its 

 discovery by Struve in 1832. The colours have been noted 

 as greenish, and blue or pale grey. The spectrum is of the 

 first or Sirian typo. Additional interest is attached to this 

 object from the fact that a distant companion (at about 

 76 seconds of arc) has been strongly suspected of variation 

 in light from 8| to 12th magnitude. Observations by Mr. 

 Tarrant in March and April 1888 tend to confirm this 

 suspicion, as they show a difl'erence in the estimates of its 

 brightness of about li magnitude in about three weeks. 

 This subject deserves more careful observation than it has 

 hitherto received. A variable star with a variable com- 

 panion would indeed be an interesting object. 



A similar suspicion of variabUity is attached to a distant 

 companion to the famous variable star Algol. This faint 

 companion lies about 80 seconds of arc to the south of Algol. 

 It was discovered in 1 787 by Schroter, who suspected varia- 

 tion in its light ; and observations in recent years tend to 

 the same conclusion. In the early part of 1871 one observer 

 faiJed to see any trace of it with a 7-inch refi-actor ; 

 but on September 9 of the same year, it was distinctly 

 visible in the same instrument." Sadler considers it pro- 

 bably variable from the 10th to the 11th magnitude in 

 some short period. Three other faint companions were 

 foimd by Burnham, one of them being distant less than 

 11 seconds from the suspected variable, which was 

 estimated of the 10th magnitude by Burnham in 

 September 1877, the companion being rated about 12^ on 

 the same evening. The others are of about the 13th 

 magnitude, and form good comparison stai's for the 

 suspected variable, which Franks foimd •' easy enough " 

 with an lli-inch reflector on-Tanuary 11, 1885, and about 

 2 magnitudes brighter than Burnham's companion. 



The well-known variable star 68 (;/) Herculis has also a 

 tolerably close companion of about the 10th magnitude, 

 ihstant about 1 seconds in 1878. Here we have also a 

 ditfercnce of brightness of about 5 magnitudes, or a ratio 

 of 100 to 1 in the relative brilliancy of the components. 



Another interesting case is that of the well-known 

 variable a Herculis. This is a double star with com- 

 ponents of about the 3rd and 6th magnitudes, at a distance 

 of about li seconds of arc. Measures of position from 1782 

 to 1876 show little or no change, and indicate no physical 



* Nature, Feb. 20, 1879. 



