June 30, 1910J 



NATURE 



527 



bright nebulae, suggest the presence of obstructing 

 nebulous masses such as that which was only revealed 

 by the outburst of Nova Persei. It is thus reasonable 

 to suppose that a sensible amount of absorbing matter 

 exists in space. But to form a quantitative estimate 

 of its effect is a task of which our increased knowledge 

 makes us only the more appreciate the difficulty. 



The problem would be simpler if we could accept 

 the conditions supposed by the earlier astronomers. 

 For them the extent of the universe was indefinitelv 

 large, and the distribution of the stars roughly uni- 

 form. Moreover, they imagined that the intrinsic 

 brightness of all stars was nearly constant, and that 

 the observed differences of magnitude were almost 

 entirely attributable to the effect of varying distance. 

 But for an effective general absorption we ought, 

 in these circumstances, to expect the whole sky 

 to shine with the average brilliance of a stellar sur- 

 face, and Halley, in supposing otherwise, was guilty 

 of a simple error. The truth of this was perceived by 

 Cheseaux (1744), and later by Olbers (1823), and both 

 astronomers inferred an extinction of light in space 

 without estimating its amount, or even supporting it 

 by direct evidence, rather than admit that the universe 

 was finite. The first estimate based on these premises 

 was giv'en by W. Struve in his "Etudes d'Astronomie 

 Stellaire " (1847), a work of great historical interest. 

 Using the data provided by Herschel's gauges, and 

 the counts of Bessel and Argelander, Struve concluded 

 that light coming from the mean distance of sixth- 

 magnitude stars suffered a loss of 8 per cent, of its 

 intensity. But we have ceased to regard as valid the 

 premises on which this conclusion was based. We 

 know now that the variability of the intrinsic light 

 of the stars is so great that distance can no longer 

 be considered as the chief factor in determining their 

 apparent magnitudes. There are also grave difficulties 

 in the way of assuming that the universe extends with 

 finite density to an infinite distance. Seeliger has 

 pointed out that unless the Newtonian law of gravita- 

 tion be modified, an infinite strain will exist at ever}; 

 point ; and even if the smallness of the total light 

 of the sky be accounted for by some kind of absorp- 

 tion, a thermal difficulty remains; for any part of 

 an infinite and eternal universe will be, as it were, 

 within an isothermal enclosure, and the temperature 

 at every point will be at least 6000° C. Such con- 

 siderations, and the facts of observation, have led 

 us to abandon the idea of an infinite universe, and 

 Newcomb asserted, not only that the system of the 

 stars v^-as finite, but that there was no evidence that 

 any extinction of light in space occurred. That will 

 not hinder us from seeking for evidence. If we 

 nossessed a knowledge of the spatial relations and the 

 inosities of the stars, if, in a word, we held the 

 to the sidereal problem, we should be in a position 

 to assess the absorption of light in space. But with- 

 out assessing the loss of light according to distance, 

 the sidereal problem cannot be solved. In fact, the 

 two problems are interdependent, and it may be long 

 before a satisfactory solution is reached. 



There is, however, a subsidiary line of attack possi- 

 ble. The absorption may be selective in its character, 

 or, in other words, its amount may be a function of 

 the wave-length. This will be the case if it is due 

 to scattering by particles the dimensions of which are 

 of the order of a wave-length of light. It is not abso- 

 lutely clear that a similar effect may not be produced by 

 the aether itself. In either case a corresponding dis- 

 persion is to be expected, and the rate of propagation 

 svill depend on the colour of the light. If any celestial 

 phenomenon be carefully observed which is strictly 

 localised in space and in time, the relative rate of 

 transmission for different parts of the spectrum can 



NO. 2122, VOL. 83] 



be detected. Newton approached the subject from 

 this point of view, and suggested the eclipses of 

 Jupiter's satellites as suitable phenomena for investi- 

 gation. It is well understood now that the conditions 

 of a gradual eclipse are quite unfavourable for the 

 detection of subtle colour changes, and the distance 

 in this case is altogether too small. The circum- 

 stances of stellar aberration have also been invoked 

 to set a limit to the possible dispersion. But the 

 sensitiveness of this test is also too small, for a differ- 

 ence of as much as one-half per cent, in the rate of 

 transmission would at best produce a spectrum o'l' 

 in length (and this is about the estimated width of 

 the fine micrometer wire of the largest retractors). 

 As nothing of this order is to be looked for, a finer 

 test must be sought. Arago conceived the possibility 

 of detecting a change of colour in variable stars 

 according to the light phase. Contenting himself 

 with the simple inspection of certain variables, he 

 concluded that there was no such effect. As a matter 

 of fact, more careful observers have noticed a change 

 of tint accompanying the change of brightness ; but 

 even so the natural explanation is to be found in the 

 physical character of the stars. This illustrates the 

 need for a cautious interpretation of results, as well 

 as for the most careful and refined methods of 

 observation. 



A great advance in practical methods has been 

 made recently by M. Nordmann, of Paris. His plan 

 has been to study the light curves of certain variable 

 stars, using the light from different regions of their 

 spectra, in accordance with the belief that a sensible 

 dispersion in space must produce a want of simul- 

 taneity in the respective curves. With this object in 

 view he designed an ingenious modification of the 

 Zollner type of photometer. By its means the light of 

 the star examined can be compared with an artificial 

 star produced by condensing on a small hole the 

 light of an Osram lamp fed by a constant current. 

 Before entering the eyepiece, the light from both 

 images passes through one of three liquid light 

 filters, and thus the comparison is made in red, green, 

 or blue light as desired. Whatever opinion may be 

 formed of the validity of M. Nordmann 's conclusions, 

 it is fair to say that his apparatus has been admirably 

 designed, and that much is to be expected from the 

 systematic application of his method to the study of 

 coloured and variable stars. Finding the atmosphere 

 of Paris unsuitable for delicate researches of this 

 kind, he took his apparatus to Biskra, in Algeria, 

 where he spent several months in 1907-8. Unfor- 

 tunately, the climate of this station did not fulfil 

 expectations, owing to the prevalence of sand storms, 

 and this fact may account for a certain want of con- 

 tinuity in the observations in a research which 

 demanded continuity as a necessar\' condition of com- 

 plete success. 



M. Nordmann studied chiefly the stars /3 Persei and 

 X Tauri. But before alluding to his results, we may 

 refer to the nearly contemporaneous work of M. 

 Tikhoff, of Pulkowa. M. Tikhoff has conceived more 

 than one ingenious method of attacking the problem 

 of dispersion in space. One of these is in principle 

 the same as that of M. Nordmann, but differs from it 

 in employing photography instead of direct visual 

 estimates. By using bathed plates, the region of the 

 star examined is photographed through certain 

 screens, which are prepared in such a way as to allow 

 only light belonging to restricted ranges in the spec- 

 trum to be effective. Thus, an orange screen may be 

 expected to give results comparable with those 

 obtained by visual methods, while a blue screen will 

 give photometric estimates in the ordinary photo- 

 graphic region of the spectrum. By this method M. 



