March 2, 1891.] 



KNO^A^LEDGE 



55 



much stress can be laid upon them. He, too, has recourse 

 in a general way to the tidal rationale ; and — what is more 

 siginificant — ranks A. Tauri as a transition instance between 

 Algol and (3 Lyra\ adding the remark that spectroscopic 

 measures of its radial movements may help to elucidate 

 the still imfathomed mystery of "short-period" varia- 

 bility. {Asti: Nad,., No.' 3,016.) 



A cause tending further to complicate tidal phenomena 

 in sun-like bodies has been adverted to by Wr. Eanyard. 

 It is this. A photosphere is probably a region of con- 

 densation, or the hottest region where matter can exist in 

 a non-gaseous form. Consequently the temperature of 

 the photospheric region is fixed. It may be regarded as 

 an isothermal surface changing its level with local varia- 

 tions of heat. The photospheres, accordingly, of two 

 adjacent radiating masses should bulge out somewhat, one 

 towards the other. Deformations arising in this way in 

 the Algol stars might be expected to become sensible to 

 our perception — if at all — after a similar fashion to tidal 

 deformations, namely, by maxima of lustre at elongations, 

 minima at conjunctions.* 



Neither source of disturbance, however, connects itself 

 naturally with the enigmatical pause in recovery charac- 

 teristic of this class of variables. And the late M. 

 Klinkerfues's supposition of an atmosisheric tidal wave 

 following in the wake of the satellite, and bringing about 



* If we suppose the larger star to have given birth to the smaller 

 eclipsing star, in a manner similar to that suggested by Prof. Geo. 

 Darwin with regard to the birth of the moon from the earth, we 

 should expect to find the larger star rotating on its axis faster than 

 the smaller star completes a revolution about it: and the longest 

 axis of the tidal ellipsoid would also, as in the case of the earth's tides, 

 travel in advance of the line joining the centres of the two stars. 

 Thus the larger star would present its minimum area before 

 the time of inferior conjunction or central eclipse. O 



But there is another possible cause of variation in the 

 light derived from the eclipsed star, which was not over- 

 looked by Prof. E. C. Pickering in his remarkable paper on *'~* 

 the " Dimensions of the Fixed Stars, with special reference ^ 

 to Binaries of the Algol type," published in the Proceedini/s .5P 

 of the American Academy, Vol. XVI., viz. the probable de- '" 

 crease of the brightness of the discs of stars towards their ^ 

 edges. Prof. Pickering says : '• The presence of lines in g 

 stellar spectra leads to the belief that the stars, like our j 

 sun, are surrounded by an absorbing atmosphere. They 

 also, therefore, probably resemble it in presenting a disc brighter in 

 the centre than at the edges, owing to the greater thickness of the 

 atmosphere and consequent greater absorption at the edges." 



Prof. Pickering seems to have assumed that the decrease of 

 brightness would be similar at both limbs, but with an egg-shaped 

 star with the longer axis inclined to the line of sight the rate of 

 increase of thickness of the absorbing layers would bo different at 

 the two limbs, and, under the conditions 

 assumed above, we should have the decrease 

 of light towards the preceding limli A more 

 rapid than towards the following limb B ; 

 consequently the light of the larger star 

 would recover its brightness more slowly 

 just after central eclipse than it decreased 

 before central eclipse. 



According to my theory, the photosphere 

 of a binary star would be intermediate in form between an isothermal 

 surface and a surface of equilibrium, for, as explained in a former 

 paper, wo cannot suppose the photospheric clouds to bo floating in an 

 atmosphere. Tho particles must bo falling under the action of 

 gravity retarded but slightly, if at all, by gaseous friction, such as 

 tliat which retards the fall of the particles composing a cloud in our 

 atmosphere, but they would be retarded by tho backward kicks of 

 molecules evaporated towards the heated centre. Such falling 

 pari ides would bo finally evaporated at a level which would depemi 

 on the temperature of the region, as well as on their rate of falling : 

 and since tho acceleration of gravity would be similar at all places 

 on a surface of equilibrium, and tho temperature similar at all places 

 on an isothermal surface, we should expect to find tho falling particles 

 glowing most vividly before their final dissolution in a stratum which 

 would extend around tho star as a thin shell intermediate in form 

 between an isothermal surface and a surface of equilibrium. — 

 A. C. RANVAur>. 



partial obscurations through increased absorption, receives 

 no countenance from the spectroscope. The light of U 

 Cephei, it is true, turns ruddy as it fades ; but not, it may 

 safely be asserted, owing to this cause. The variations 

 of its spectrum offer a tempting and hitherto unexplored 

 field of study. Indeed, the star has of late in every way, 

 especially in this country, been too much neglected. It is 

 remarkable for a prolonged minimum, originally explained 

 by Professor Pickering as due to a total eclipse by a large, 

 semiobscm-e satellite. But Mr. Chandler's observations 

 showing variations of about two-tenths of a magnitude 

 during the supposed totality, seem to compel recourse 

 to some other hypothesis, if not to replace, at any 

 rate to supplement the first. Two periods of the 

 star being nearly equal to five days, only every second 

 minimum can be followed at the same season of the 

 year. Those of which the average course is represented 

 in Fig. 2 occurred in the autumn. A shorter spring 

 series of observations, also by Mr. Chandler, giving 

 almost a dead level of least light, of close upon two hours' 

 duration, claimed an inferior degree of authority (Astyo- 

 nomical Jijiirnal, No. 199). There is probably no real 

 distinction in character between the alternating phases. 



^V /^ i 



8f 

 9v 



^ 



f 



Fig. 2. — Minimum of U Cephei. From 159 autumn observations bv 

 Mr. S. C. Chandler. 



The vicinity of this object to the pole renders it par- 

 ticularly suitable for '• trailing" experiments, which might 

 definitively settle the interesting question as to the true 

 form of its light-curve. 



^ 67 1 



6? ^ 



77 



C^ID ifO 6C So t>c no IV M)C 'fO 200 13£ i^ i^<> :&■ SCO 



Fig. 3. — Provisional MiNiMCM-CnEVE op S Astli-e (Yendell). 



Since tidal effects grow, ctrterix ptirilms, in the inverse 

 proportion of the cube of distance, they ought, so far as 

 they intluence luminosity, to be most apparent in 

 variables of the shortest periods, since those are proved 

 by the proportionate length of their eclipses to be made 

 up of the most closely contiguous pairs. But there 

 is no sign that the subsidiary changes of Algol stars obey 

 any such law. They are especially striking in S Cancri, 

 with its relatively long period of nine and a half days. 



