380 



KNOWLEDGE. 



October. 1911. 



torn suns than they are upon one another. Conse- 

 quently the effective attraction is more than seven, is 

 realh" equivalent to about nine, that is one-and-a- 

 half-times as much as before the collision. If the 

 stars had anv proper motion before attraction 

 brought them together the orbit at impact would 

 have been a h\-perbola, generally \ery close to a 

 [Kirabola. After the collision the additional attraction 

 will in \erv manv cases convert the orbit into an 

 ellipse and consequently the two torn suns will 

 be wedded by the influence of the third body into 

 an orbitally connected pair, that is, into a double star 

 (see Figure 9). If the proper motion were large, and 

 the graze small, the capturing power of the third bod\- 

 would not suffice to wed the two stars into a binarw 

 The number of pairs of variable stars that are too 

 far se[)arated to be considered physically-connected 

 binaries seems to suggest that this must happen 

 frequentl\-. .\ list of some of these pairs of 

 variables tliat were found in less than a thirtieth of 

 the celestial vault is gi\'en on page 89 of "" The Birth 

 of Worlds and Systems." 



It has been argued by astronomers that the orbit 

 of double stars that had been wedded b}' impact, 

 must be such as to cause the stars to collide again at 

 everv periastrum. This, probably, as a rule, is not 

 the case, but occasionalh such recurrent impacts 

 apjiarentlv take place. Agencies, howex'er. come into 

 pla\' that tend to [irevent recurrent ini[)acts: these 

 agencies are fullv discussed in " The Birth of 

 W'orlds and Systems." We will give one \'er\- 

 efficient agent that tends to prevent recurrent impact. 

 The pair of torn suns are wedded b\- the attraction 

 of the third bod\-, but this third body is largely 

 dissipated before the two stars reach their greatest 

 distance. Hence it is not there to attract them back 

 again. Consequently they do not ap[)roach so near 

 one another at periastrum as they would have done 

 had the third star remained the same mass that it 

 possessed when first formed. The several factors 

 fullv account for the low eccentricitv of manv 

 binaries. There is great reason to imagine that 

 both No\a Persei and Mr. Espin's star are cases of 

 recurrent imjiact. In the case of Nova Persei a 

 peculiar concentric-looking nebula existed before the 

 impact took place that produced Nova Persei itself. 

 As exactl}' such a nebula might have been produced 

 by a previous impact, this idea of recurrent impacts 

 seems to furnish a sufficient explanation of its exist- 

 ence. The accomjianying Figure 10 is one of those 

 taken as the nebula was progressively lit up b\- the 

 flash of light that was produced b\- the exploding 

 third star — the flash we saw and named Nova 

 Persei. We have every reason to suppose that 

 this was the case, because, as the light was pro- 

 gressively reflected from the whorls of the nebula, 

 it showed the characteristic blaze band spectrum 

 that was so unique a peculiarit\' of the flash period 

 of Nova Persei. 



The reason why we may think of Nova Lacertae as 

 being the third body produced b\' a recurrent impact, 

 is that one of the two stars that came into impact in 



this case appeared to have been a variable star before 

 the partial imjxict occurred that produced this 

 apparition. It may be worth\- of note that one of 

 the dynamical deductions as to the properties of the 

 third bod\- is that it passes through the stage of 

 being a planetarv nebula. Nova Lacertae is now said 

 to have become a [)lanetarv nebula. 



I cannot enter into greater detail here because the 

 " theory of [lartial impact, and the third body " is 

 so wide as to reall}- constitute a new cosmogony. I 

 would refer students to the original papers in the 

 transactions of the New Zealand Institute, for the 

 years 1878, 1879 and 1880. These are in all the 

 important libraries of London. 



If double stars are produced by the grazing impact 

 of suns, in their earl}' days the\' must have exhibited 

 the scars produced by the encounter and sometimes 

 be doubly variable : but there are many reasons why 

 this scarred condition should last very much longer 

 in some stars than in others. Consequently we should 

 expect to find a comparatively large number of double 

 stars that exhibit variability and a few that show- 

 double variabilit}'. There are several double stars 

 actually known to be thus doubly variable, quite a 

 large number of double stars that are single variables, 

 and still more double stars that are coloured, a state 

 which has been deduced as being a final condition of 

 \ariability before the torn suns have healed com- 

 pletely and once more become normal stars. 



It would be an extremely important piece of work 

 were some of our amateur astronomers to find the 

 many \-ariable stars that are double, and especially 

 the man}' spectroscopic binaries that are variable and 

 in which the variability is not due to eclipse. Mr. 

 Espin and Mr. \'ictor Anestin. have already found 

 some most valuable and remarkable results in con- 

 nection with double stars, their association with the 

 various classes of variable stars and with density of 

 star distribution. 



Let us now examine in some detail the condition 

 of these torn suns with respect to the distribution oi 

 the material, and the character of the motions 

 developed in the suns that have come into grazing 

 impact. 



Nearh- the whole of the dynamical conclusions 

 regarding the grazing impact of suns were worked 

 out over thirty years ago, and the\' have been 

 subjected to endless debate and calculation on the 

 part of mathematicians and astronomers. If we 

 grant the fact that collisions must occur, as on the 

 grounds of the doctrine of probabilit}' seems 

 inevitable, we have an actual explanation of the 

 phenomena of temporar\-, variable and double stars, 

 for we cannot doubt the accuracy of the thermo- 

 dx'uamic deductions that are here presented. 



\\'hen these basic deductions are combined with 

 the fact that there appears to be scarcely a single 

 one of all these complex deductions but has been 

 confirmed hv obserx-ations. we must admit that 

 impact, if not of the supreme importance that is 

 suggested b\- these researches and observations, 

 must be a stujiendousl}- large factor in cosmic 



