676 



NATURE 



[February 18, 19 15 



so that within the limits of uncertainty in every 

 case the distribution of the stars is consistent with 

 a solution of the above differential equation when 

 y is assigned the value i'2. 



The analogy between the distribution of stars in 

 a condensed cluster and the density in a spherical 

 mass of gas of a particular type in adiabatic equi- 

 librium thus seems to be fairly established. Even 

 if it be supposed that the cluster is the outcome 

 of an original nebula the question still remains 

 why the distribution of matter should persist long 

 after its condition has completely changed, or why 

 the arrangement should resemble what might be 

 expected of certain vapours {e.g., chloroform). 

 The answer given by v. Zeipel on the basis of a 

 strict mathematical analysis is that this is in con- 

 formity wMth a kinetic theory which applies to an 

 aggregate containing a high proportion of 

 Keplerian binaries. This may be a bold application 

 of the law of large numbers, but it is certainly an 

 interesting conception. Since there is every reason 

 to believe that all short period variables are binary 

 systems the observed occurrence of these in 

 clusters lends support to the view, though they can 

 only represent the exceptionally close systems. 

 The investigations here described refer exclusively 

 to the highly condensed clusters. But there exist 

 also clusters showing states of concentration in 

 varying degree until probably all visible trace of 

 organic connection is lost. In Stromgren's view 

 the whole series represents an order of evolution 

 by which the dense clusters grow out of more 

 scattered forms. Whether the results will throw 

 light on the wider problems of the structure of the 

 sidereal universe seems doubtful in view of certain 

 conclusions drawn by Polncare, Jeans and Ed- 

 dington as to the relevance of the kinetic theory. 

 But taken by themselves they present questions of 

 the highest interest which are likely to repay 

 further study. H. C. Plummer. 



ON COLOUR SENSITISED PLATES. 

 I. — In General and Orthochromatic Plates. 



IT used to be customary to draw three curves 

 above a diagrammatic spectrum, heat, lumin- 

 osity, and actinism curves, the last representing 

 the power of light to produce or facilitate chemical 

 change independently of the temperature change. 

 This custom survives to a certain extent, though 

 only one of the curves, namely, the heat curve, is 

 definite. The luminosity curve depends upon the 

 human eye, and eyes vary, sometimes even in the 

 same Individual, with regard to their sensitive- 

 ness to light and colour. Still, it is possible to 

 draw practically useful luminosity -curves in a 

 general sense, and by taking an average human 

 eye, In perhaps almost an absolute sense. 



But the " actinism " curve is essentially different, 

 for here we may be concerned, not with a single 

 organ and its possible variations or degrees of per- 

 fection, but with every substance that exists on 

 the face of the earth or that can be prepared by 

 artificial means. And If we limit our considera- 

 tions to the very few substnnces that are practically 

 NO. 2^64. VOL. 94! 



utilised in photography, we find that " actinism " 

 extends from well into the Infra-red down to the 

 Rontgen rays, which are far below what is gener- 

 ally known as the ultra-violet. " Actinism " ex- 

 tends over a range of eleven or twelve octaves for 

 practical photographic purposes, while luminosity 

 extends over scarcely one octave, and for practical 

 purposes even less than this, and yet some people 

 speak of the photographic plate as colour-blind ! 



The whole of this eleven or twelve octaves has 

 not yet been dealt with photographically, because 

 In the extreme ultra-violet (the " Schumann 

 region ") at wave-lengths a little less than 

 200 nil, the absorbing power of air and gelatine 

 prevents the passage of radiations through t!hem. 

 But this appears to be due to absorption bands,, 

 as radiations of still shorter wave-length (Rontgen 

 rays) pass freely through these media. By getting 

 rid as far as possible of air and gelatine, the photo- 

 graphy of the ordinary spectrum has been extended 

 down to wave-length loo/x/x, or even less. There 

 are other difficulties than the air and gelatine to 

 contend with In investigations of this region, but 

 with these we are not Immediately concerned. 



Although it is necessary sometimes to bear in 

 mind the enormous range of sensitiveness of photo- 

 graphic materials, even from a purely practical 

 point of view, if we exclude the Rontgen region, 

 and regard only those circumstances that concern 

 the photography of objects, whether terrestrial 

 or celestial, and whether by daylight or artlficla! 

 light, we have to consider only about two octaves 

 of radiations, or rather more if the far infra-red 

 is taken into account. This range may be still 

 further curtailed when daylight or glass apparatus 

 is used, on account of the absorptive power of 

 glass and the atmosphere, and what remains may 

 often be sufficiently described by Indicating 

 five regions, namely, ultra-violet, blue, green, 

 red, and infra-red. The "blue" will in- 

 clude the indigo and violet, and the " red " will 

 Include the orange, and the yellow Is negligible as 

 in a good spectrum it is represented by little 

 more than the sodium D lines. 



In order to photograph coloured objects so that 

 their luminosities shall be correctly represented in 

 the print, we want to get the curve that represents 

 the action of the spectrum on the plate to coincide 

 with the luminosity curve of the spectrum, and 

 then we want a printing method that will preserve 

 these tone values. The alternative of getting equal 

 and opposite errors in the negative and the print 

 so that the one shall correct the other, may have 

 a degree of possibility about It. The fact to be 

 emphasised Is that the getting of a correct nega- 

 tive Is not the whole business. Indeed, the getting 

 of the two curves to correspond is not the whole 

 business so far as the negative is concerned, for 

 they may correspond at one exposure of the 

 plate to the spectrum and not at another, because 

 the steepness of the gradation of the deposits 

 produced on the plate by equivalent ranges of 

 exposures to the various parts of the spectrum is 

 not the same. These difficulties are mentioned tO' 

 show that, from a practical point of view, " ortho- 



