10 



KNOWLEDGE 



[.Tancaby, 1902. 



clearly known as could be wished ; but they are not 

 entirely lost. On the Har\'ard photographs of the 22nd 

 the bright lines were beginning to appear; bnt on tiie 

 Potsdani plates of the 23rd " thero was not any 

 suggestion of emission lines or bands," || but the 

 hydrogen lines were broad, weak, and diffusive. From 

 this it would appear that on the night of the 23rd the 

 dai-k hydrogen lines were widened out together with the 

 bright lines, partly masking ono another, and extinguish- 

 ing the bright edges seen on the IIai-\'ard College plates 

 of the 22nd. The two scries were not yet separated, 

 and the scpai-ation took place between the nights of 

 the 23rd and 24th. And if this separation of the bright 

 and dark hydrogen lines is to be attributed to velocities 

 on Doppler's principle, and according to Lord Kelvin's 

 explanation of the high velocity of the cooler gases after 

 impact,^ the actual collision of the two masses must 

 have occurred between these two nights. Such an 

 occuiTence is not inconsistent with the previous short 

 history of the star. It was increasing in brightness 

 from the early morning of Febiiiai-y 22, civil time, up 

 to the night of the 23rd, and may have reached its 

 maximum on the 24th in the day-liglib hours of both 

 Europe and America. During the two or three days 

 of approaching one another, both bodies would suffer 

 tidal disturbances enough to account for the increasing 

 white heat, and brilliant continuous spectrum. And 

 after the crash, the temperature would rise suddenly, 

 vaporising the masses and converting much of the con- 

 tinuous spectrum into gaseous radiations, with corre- 

 sponding fall off in general brilliancy. 



But, to retiuTi to the first enquiry, the evidence of 

 the photogi-aphic plates of the 22nd and 23rd appears 

 to be against any transition of the specti-um through 

 the changes whicli are supposed to mark epochs in the 

 long history of the declining life of a peniianent stai". 

 The transition from the type of the Orion stars to the 

 composite bright and dai-k broad line spectiiim was 

 abrupt; the bright lines had already begun to appear 

 on the 22nd, and the dai-k lines were broad on the 23rd. 

 It should, therefore, be considered possible that a star 

 of the Orion type might die out without passing through 

 the gi-adations of the Solar type, \uiless the gap in the 

 succession of spectra of the Nova was covered by the 

 leap of temperature at the impact of two stars, already 

 suggested. After the first flash of brilliancy the hotter 

 gaseous radiation would be less luminous than that of 

 the previous solid or liquid state, and the scattered 

 fragments would cool down rapidly by freer radiation 

 into space. The pair of stais would be broken iip to 

 begin again their slow career of condensation. 



The new composite spectrum of Febi-uai-y 24, which 

 may be called the second spectrum, to distinguish it 

 from later fonns, remained substantially the same until 

 March 19, and is illustrated by the first four photographs 

 of the plate. But one very significant change was in pro- 

 gress diiring this time, and was completed between 

 March 21 and 22: the dark hydrogen lines were thin- 

 ning away, and were seen for the last time on March 21. 

 The importance of this lies in its connection with Lord 

 Kelvin's explanation of the apprnaching velocity of the 

 ahmrhing hydrogen. This temporary atmosphere might 

 be very dense at first, fomiing broad absorption lines, 

 and would thin away in its flight from the star, showing 

 thinner lines and finally vanishing. So that if the dark 

 hydrogen lines had not thinned and disappeared com- 



II Vogel, ihid. % Observatorv, No. 300, June, p. 223. 



paratively soon, the chief support of the impact 

 hypothesis would have broken down. 



On March 19 another great change was found in the 

 spectrum photographed at Hai-vard College. This is 

 tlie third spectrum of the star. It was photographed 

 again at Stonyhurst on the 22nd, 25th, and 28th, and 

 also at the Yerkes Observatory on the 22nd and 28th. 

 But it was not a pennanent change. The second 

 specti-um returned on the 21 sb, 26th, and 27th; and 

 up to April 26 the spectrum alternated between the two 

 foiTns. Tiio later, or third spectrum, is illustrated on 

 the plate already referred to in the pliotogi-aphs of 

 March 25 and 28. Its leading features are (1) a great 

 apparent extension of the first ultra violet hydrogen 

 line, H?, on the side of shorter wave lengths, occupying 

 nearly the whole interval between H? and H>) ; (2^ the 

 greater prominence of the blue bands, together with 

 new formations between H^ aud Hy ; and (3) the loss 

 of light in the continuous spectrum. 



At fii'st these alternations appeared to be connected 

 with the phases of the stai''s light cm-ve, which at this 

 time were quite periodic, the minimum recurring on 

 the 19th, 22nd, 25th, and 28th. But this three^day 

 period failed eai-ly in April ; and since then, a com- 

 parison of all the photographs taken at Stonyhurst, up 

 to April 26, with the magnitudes of the star as 

 measured at the Radcliffe Observatory, Oxford, has 

 shown that the connection was not with any phase of 

 the light variations, but with an absolute magnitude of 

 the star. All the spectra of the third fonn appeared 

 when the magnitude of the star was below 4.57 ; all 

 the returns of the second spectrum occurred when the 

 star was brighter ; and on April 9, when the magnitude 

 was at this critical figure, the spectrum had been already 

 noted as belonging partly to both forms.** 



The inference here is that at this period, when the 

 star was largely a glowing gaseous mass, the temperature 

 was oscillating with oscillations of the gaseous volume. 

 This volume of mixed heated gases might well have 

 been put into a state of gi'eat oscillation on being set 

 free at the impact of two stars, expanding and contract>- 

 ing, chilling and heating; and when the mean 

 temperature had fallen to the combining temperature of 

 the Constituents of the new baud adjoining H^, this 

 band would reappear and disappear with the oscillations 

 synchronously with the light variations. 



So far, therefore, there does not appear to be any 

 insuperable obstacle in the way of the direct collision 

 hypothesis. And the recent astonishing information 

 from America regai-ding the surroundings of Nova Pei-sei 

 seems to demand the greatest imaginable catastrophe, 

 and there is none greater than the clash of two cold 

 worlds. In these suiTouudings there is telescopic 

 evidence of matter drifting at a rate which leaves the 

 alanning velocities revealed by the spectroscope to be 

 reckoned amongst comparatively easy going speeds. And 

 the evidence seems to be complete, in tlie agreement of 

 foui" separate condensations of nebulosity. The positions 

 of these on a photogi-aphic plate exposed on November 

 7 at the Lick Observatoiy, compared with a photogi-.-iph 

 at the Yerkes Obsei-vatory on September 20, show 

 practically the same displacements of each, amounting 



*• These comparisons are confirmed by those of Harvard College, 

 in Ap. J., 1901, July, with one exception, viz., Ihat " on April 12 

 and 13 the magnitude was the same while the spectrum was different." 

 But the Oxford measures give different magnitudes for the two 

 nights, viz., 4,(i7 on the 12tli, when the third spectrum was photo- 

 graphed, and -t. 49 on the 13th.— M. N., May, 468. 



